History

Merge branch 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf updates from Ingo Molnar: "Kernel side changes: - Clean up the x86 instruction decoder (Masami Hiramatsu) - Add new uprobes optimization for PUSH instructions on x86 (Yonghong Song) - Add MSR_IA32_THERM_STATUS to the MSR events (Stephane Eranian) - Fix misc bugs, update documentation, plus various cleanups (Jiri Olsa) There's a large number of tooling side improvements: - Intel-PT/BTS improvements (Adrian Hunter) - Numerous 'perf trace' improvements (Arnaldo Carvalho de Melo) - Introduce an errno code to string facility (Hendrik Brueckner) - Various build system improvements (Jiri Olsa) - Add support for CoreSight trace decoding by making the perf tools use the external openCSD (Mathieu Poirier, Tor Jeremiassen) - Add ARM Statistical Profiling Extensions (SPE) support (Kim Phillips) - libtraceevent updates (Steven Rostedt) - Intel vendor event JSON updates (Andi Kleen) - Introduce 'perf report --mmaps' and 'perf report --tasks' to show info present in 'perf.data' (Jiri Olsa, Arnaldo Carvalho de Melo) - Add infrastructure to record first and last sample time to the perf.data file header, so that when processing all samples in a 'perf record' session, such as when doing build-id processing, or when specifically requesting that that info be recorded, use that in 'perf report --time', that also got support for percent slices in addition to absolute ones. I.e. now it is possible to ask for the samples in the 10%-20% time slice of a perf.data file (Jin Yao) - Allow system wide 'perf stat --per-thread', sorting the result (Jin Yao) E.g.: [root@jouet ~]# perf stat --per-thread --metrics IPC ^C Performance counter stats for 'system wide': make-22229 23,012,094,032 inst_retired.any # 0.8 IPC cc1-22419 692,027,497 inst_retired.any # 0.8 IPC gcc-22418 328,231,855 inst_retired.any # 0.9 IPC cc1-22509 220,853,647 inst_retired.any # 0.8 IPC gcc-22486 199,874,810 inst_retired.any # 1.0 IPC as-22466 177,896,365 inst_retired.any # 0.9 IPC cc1-22465 150,732,374 inst_retired.any # 0.8 IPC gcc-22508 112,555,593 inst_retired.any # 0.9 IPC cc1-22487 108,964,079 inst_retired.any # 0.7 IPC qemu-system-x86-2697 21,330,550 inst_retired.any # 0.3 IPC systemd-journal-551 20,642,951 inst_retired.any # 0.4 IPC docker-containe-17651 9,552,892 inst_retired.any # 0.5 IPC dockerd-current-9809 7,528,586 inst_retired.any # 0.5 IPC make-22153 12,504,194,380 inst_retired.any # 0.8 IPC python2-22429 12,081,290,954 inst_retired.any # 0.8 IPC <SNIP> python2-22429 15,026,328,103 cpu_clk_unhalted.thread cc1-22419 826,660,193 cpu_clk_unhalted.thread gcc-22418 365,321,295 cpu_clk_unhalted.thread cc1-22509 279,169,362 cpu_clk_unhalted.thread gcc-22486 210,156,950 cpu_clk_unhalted.thread <SNIP> 5.638075538 seconds time elapsed [root@jouet ~]# - Improve shell auto-completion of perf events (Jin Yao) - 'perf probe' improvements (Masami Hiramatsu) - Improve PMU infrastructure to support amp64's ThunderX2 implementation defined core events (Ganapatrao Kulkarni) - Various annotation related improvements and fixes (Thomas Richter) - Clarify usage of 'overwrite' and 'backward' in the evlist/mmap code, removing the 'overwrite' parameter from several functions as it was always used it as 'false' (Wang Nan) - Fix/improve 'perf record' reverse recording support (Wang Nan) - Improve command line options documentation (Sihyeon Jang) - Optimize sample parsing for ordering events, where we don't need to parse all the PERF_SAMPLE_ bits, just the ones leading to the timestamp needed to reorder events (Jiri Olsa) - Generalize the annotation code to support other source information besides objdump/DWARF obtained ones, starting with python scripts, that will is slated to be merged soon (Jiri Olsa) - ... and a lot more that I failed to list, see the shortlog and changelog for details" * 'perf-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (262 commits) perf trace beauty flock: Move to separate object file perf evlist: Remove fcntl.h from evlist.h perf trace beauty futex: Beautify FUTEX_BITSET_MATCH_ANY perf trace: Do not print from time delta for interrupted syscall lines perf trace: Add --print-sample perf bpf: Remove misplaced __maybe_unused attribute MAINTAINERS: Adding entry for CoreSight trace decoding perf tools: Add mechanic to synthesise CoreSight trace packets perf tools: Add full support for CoreSight trace decoding pert tools: Add queue management functionality perf tools: Add functionality to communicate with the openCSD decoder perf tools: Add support for decoding CoreSight trace data perf tools: Add decoder mechanic to support dumping trace data perf tools: Add processing of coresight metadata perf tools: Add initial entry point for decoder CoreSight traces perf tools: Integrating the CoreSight decoding library perf vendor events intel: Update IvyTown files to V20 perf vendor events intel: Update IvyBridge files to V20 perf vendor events intel: Update BroadwellDE events to V7 perf vendor events intel: Update SkylakeX events to V1.06 ...

Linus Torvalds committed 8 years ago

Commit d8b91dd2 parents 5e7481a e4c1091270 files · +18523 −15234Browse files

Unified · Split

270 files
  1. Documentation/process/kernel-enforcement-statement.rst +1 −0
  2. MAINTAINERS +2 −1
  3. arch/x86/events/msr.c +43 −27
  4. arch/x86/include/asm/uprobes.h +4 −0
  5. arch/x86/kernel/uprobes.c +103 −4
  6. arch/x86/tools/Makefile +6 −6
  7. arch/x86/tools/insn_decoder_test.c +25 −18
  8. arch/x86/tools/objdump_reformat.awk +2 −2
  9. include/uapi/linux/perf_event.h +7 −3
  10. kernel/events/callchain.c +0 −15
  11. kernel/events/core.c +39 −15
  12. kernel/events/internal.h +0 −4
  13. tools/arch/alpha/include/uapi/asm/errno.h +128 −0
  14. tools/arch/mips/include/asm/errno.h +17 −0
  15. tools/arch/mips/include/uapi/asm/errno.h +130 −0
  16. tools/arch/parisc/include/uapi/asm/errno.h +128 −0
  17. tools/arch/powerpc/include/uapi/asm/errno.h +10 −0
  18. tools/arch/s390/include/uapi/asm/unistd.h +412 −0
  19. tools/arch/sparc/include/uapi/asm/errno.h +118 −0
  20. tools/arch/x86/include/asm/cpufeatures.h +3 −1
  21. tools/arch/x86/include/asm/disabled-features.h +7 −1
  22. tools/arch/x86/include/uapi/asm/errno.h +1 −0
  23. tools/build/Makefile.feature +3 −1
  24. tools/build/feature/Makefile +10 −1
  25. tools/build/feature/test-all.c +10 −0
  26. tools/build/feature/test-libopencsd.c +8 −0
  27. tools/build/feature/test-pthread-barrier.c +12 −0
  28. tools/include/uapi/asm-generic/errno-base.h +40 −0
  29. tools/include/uapi/asm-generic/errno.h +123 −0
  30. tools/include/uapi/linux/perf_event.h +7 −3
  31. tools/lib/traceevent/event-parse.c +51 −11
  32. tools/lib/traceevent/event-plugin.c +9 −15
  33. tools/lib/traceevent/kbuffer-parse.c +2 −2
  34. tools/lib/traceevent/parse-filter.c +14 −8
  35. tools/perf/Build +2 −2
  36. tools/perf/Documentation/perf-buildid-cache.txt +3 −0
  37. tools/perf/Documentation/perf-evlist.txt +4 −0
  38. tools/perf/Documentation/perf-inject.txt +4 −0
  39. tools/perf/Documentation/perf-lock.txt +4 −0
  40. tools/perf/Documentation/perf-probe.txt +17 −1
  41. tools/perf/Documentation/perf-record.txt +3 −0
  42. tools/perf/Documentation/perf-report.txt +36 −1
  43. tools/perf/Documentation/perf-sched.txt +4 −0
  44. tools/perf/Documentation/perf-script.txt +46 −1
  45. tools/perf/Documentation/perf-timechart.txt +3 −1
  46. tools/perf/Documentation/perf-top.txt +6 −0
  47. tools/perf/Documentation/perf-trace.txt +14 −6
  48. tools/perf/Documentation/perf.data-file-format.txt +27 −0
  49. tools/perf/Documentation/tips.txt +2 −0
  50. tools/perf/Makefile.config +55 −10
  51. tools/perf/Makefile.perf +14 −3
  52. tools/perf/arch/arm/util/auxtrace.c +71 −6
  53. tools/perf/arch/arm/util/pmu.c +6 −0
  54. tools/perf/arch/arm64/util/Build +4 −1
  55. tools/perf/arch/arm64/util/arm-spe.c +225 −0
  56. tools/perf/arch/arm64/util/header.c +65 −0
  57. tools/perf/arch/arm64/util/sym-handling.c +22 −0
  58. tools/perf/arch/common.c +3 −41
  59. tools/perf/arch/common.h +0 −1
  60. tools/perf/arch/powerpc/util/header.c +1 −1
  61. tools/perf/arch/powerpc/util/sym-handling.c +8 −0
  62. tools/perf/arch/s390/Makefile +21 −0
  63. tools/perf/arch/s390/annotate/instructions.c +2 −1
  64. tools/perf/arch/s390/entry/syscalls/mksyscalltbl +36 −0
  65. tools/perf/arch/x86/tests/perf-time-to-tsc.c +1 −1
  66. tools/perf/arch/x86/util/header.c +2 −2
  67. tools/perf/arch/x86/util/unwind-libunwind.c +1 −1
  68. tools/perf/bench/futex-hash.c +12 −8
  69. tools/perf/bench/futex-lock-pi.c +14 −9
  70. tools/perf/bench/futex-requeue.c +13 −9
  71. tools/perf/bench/futex-wake-parallel.c +35 −11
  72. tools/perf/bench/futex-wake.c +11 −7
  73. tools/perf/builtin-buildid-cache.c +2 −2
  74. tools/perf/builtin-c2c.c +7 −9
  75. tools/perf/builtin-help.c +1 −1
  76. tools/perf/builtin-inject.c +1 −2
  77. tools/perf/builtin-kvm.c +8 −5
  78. tools/perf/builtin-record.c +49 −14
  79. tools/perf/builtin-report.c +267 −16
  80. tools/perf/builtin-script.c +229 −37
  81. tools/perf/builtin-stat.c +147 −83
  82. tools/perf/builtin-top.c +7 −3
  83. tools/perf/builtin-trace.c +67 −55
  84. tools/perf/check-headers.sh +10 −1
  85. tools/perf/perf-completion.sh +43 −4
  86. tools/perf/perf.c +2 −2
  87. tools/perf/pmu-events/arch/arm64/cavium/thunderx2-imp-def.json +62 −0
  88. tools/perf/pmu-events/arch/arm64/mapfile.csv +15 −0
  89. tools/perf/pmu-events/arch/powerpc/mapfile.csv +2 −10
  90. tools/perf/pmu-events/arch/powerpc/power9/cache.json +0 −5
  91. tools/perf/pmu-events/arch/powerpc/power9/frontend.json +1 −6
  92. tools/perf/pmu-events/arch/powerpc/power9/marked.json +6 −21
  93. tools/perf/pmu-events/arch/powerpc/power9/other.json +73 −203
  94. tools/perf/pmu-events/arch/powerpc/power9/pipeline.json +7 −7
  95. tools/perf/pmu-events/arch/powerpc/power9/pmc.json +1 −1
  96. tools/perf/pmu-events/arch/powerpc/power9/translation.json +0 −5
  97. tools/perf/pmu-events/arch/x86/broadwell/cache.json +378 −177
  98. tools/perf/pmu-events/arch/x86/broadwell/floating-point.json +51 −57
  99. tools/perf/pmu-events/arch/x86/broadwell/frontend.json +69 −69
  100. tools/perf/pmu-events/arch/x86/broadwell/memory.json +205 −5
  101. tools/perf/pmu-events/arch/x86/broadwell/other.json +10 −10
  102. tools/perf/pmu-events/arch/x86/broadwell/pipeline.json +613 −603
  103. tools/perf/pmu-events/arch/x86/broadwell/virtual-memory.json +75 −75
  104. tools/perf/pmu-events/arch/x86/broadwellde/cache.json +212 −177
  105. tools/perf/pmu-events/arch/x86/broadwellde/floating-point.json +51 −57
  106. tools/perf/pmu-events/arch/x86/broadwellde/frontend.json +69 −69
  107. tools/perf/pmu-events/arch/x86/broadwellde/memory.json +4 −5
  108. tools/perf/pmu-events/arch/x86/broadwellde/other.json +10 −10
  109. tools/perf/pmu-events/arch/x86/broadwellde/pipeline.json +612 −602
  110. tools/perf/pmu-events/arch/x86/broadwellde/virtual-memory.json +75 −75
  111. tools/perf/pmu-events/arch/x86/broadwellx/cache.json +203 −180
  112. tools/perf/pmu-events/arch/x86/broadwellx/floating-point.json +51 −57
  113. tools/perf/pmu-events/arch/x86/broadwellx/frontend.json +69 −69
  114. tools/perf/pmu-events/arch/x86/broadwellx/memory.json +35 −5
  115. tools/perf/pmu-events/arch/x86/broadwellx/other.json +10 −10
  116. tools/perf/pmu-events/arch/x86/broadwellx/pipeline.json +612 −602
  117. tools/perf/pmu-events/arch/x86/broadwellx/virtual-memory.json +75 −75
  118. tools/perf/pmu-events/arch/x86/goldmont/cache.json +1081 −163
  119. tools/perf/pmu-events/arch/x86/goldmont/memory.json +270 −10
  120. tools/perf/pmu-events/arch/x86/goldmont/other.json +42 −12
  121. tools/perf/pmu-events/arch/x86/goldmont/pipeline.json +264 −242
  122. tools/perf/pmu-events/arch/x86/goldmont/virtual-memory.json +30 −30
  123. tools/perf/pmu-events/arch/x86/haswell/cache.json +194 −171
  124. tools/perf/pmu-events/arch/x86/haswell/floating-point.json +10 −10
  125. tools/perf/pmu-events/arch/x86/haswell/frontend.json +66 −66
  126. tools/perf/pmu-events/arch/x86/haswell/memory.json +21 −0
  127. tools/perf/pmu-events/arch/x86/haswell/other.json +10 −10
  128. tools/perf/pmu-events/arch/x86/haswell/pipeline.json +570 −561
  129. tools/perf/pmu-events/arch/x86/haswell/virtual-memory.json +106 −106
  130. tools/perf/pmu-events/arch/x86/haswellx/cache.json +200 −177
  131. tools/perf/pmu-events/arch/x86/haswellx/floating-point.json +10 −10
  132. tools/perf/pmu-events/arch/x86/haswellx/frontend.json +66 −66
  133. tools/perf/pmu-events/arch/x86/haswellx/memory.json +28 −0
  134. tools/perf/pmu-events/arch/x86/haswellx/other.json +10 −10
  135. tools/perf/pmu-events/arch/x86/haswellx/pipeline.json +571 −562
  136. tools/perf/pmu-events/arch/x86/haswellx/virtual-memory.json +106 −106
  137. tools/perf/pmu-events/arch/x86/ivybridge/cache.json +111 −132
  138. tools/perf/pmu-events/arch/x86/ivybridge/frontend.json +61 −61
  139. tools/perf/pmu-events/arch/x86/ivybridge/memory.json +12 −12
  140. tools/perf/pmu-events/arch/x86/ivybridge/other.json +10 −10
  141. tools/perf/pmu-events/arch/x86/ivybridge/pipeline.json +412 −410
  142. tools/perf/pmu-events/arch/x86/ivybridge/virtual-memory.json +30 −30
  143. tools/perf/pmu-events/arch/x86/ivytown/cache.json +112 −124
  144. tools/perf/pmu-events/arch/x86/ivytown/frontend.json +61 −61
  145. tools/perf/pmu-events/arch/x86/ivytown/memory.json +12 −12
  146. tools/perf/pmu-events/arch/x86/ivytown/other.json +10 −10
  147. tools/perf/pmu-events/arch/x86/ivytown/pipeline.json +412 −410
  148. tools/perf/pmu-events/arch/x86/ivytown/virtual-memory.json +30 −30
  149. tools/perf/pmu-events/arch/x86/mapfile.csv +1 −4
  150. tools/perf/pmu-events/arch/x86/silvermont/cache.json +2 −1
  151. tools/perf/pmu-events/arch/x86/skylake/cache.json +515 −3875
  152. tools/perf/pmu-events/arch/x86/skylake/floating-point.json +2 −3
  153. tools/perf/pmu-events/arch/x86/skylake/frontend.json +121 −111
  154. tools/perf/pmu-events/arch/x86/skylake/memory.json +212 −1906
  155. tools/perf/pmu-events/arch/x86/skylake/other.json +38 −2
  156. tools/perf/pmu-events/arch/x86/skylake/pipeline.json +492 −481
  157. tools/perf/pmu-events/arch/x86/skylake/virtual-memory.json +137 −125
  158. tools/perf/pmu-events/arch/x86/skylakex/cache.json +167 −90
  159. tools/perf/pmu-events/arch/x86/skylakex/floating-point.json +1 −2
  160. tools/perf/pmu-events/arch/x86/skylakex/frontend.json +24 −24
  161. tools/perf/pmu-events/arch/x86/skylakex/memory.json +154 −77
  162. tools/perf/pmu-events/arch/x86/skylakex/other.json +93 −1
  163. tools/perf/pmu-events/arch/x86/skylakex/pipeline.json +22 −22
  164. tools/perf/pmu-events/arch/x86/skylakex/virtual-memory.json +21 −21
  165. tools/perf/pmu-events/jevents.c +38 −1
  166. tools/perf/scripts/python/bin/mem-phys-addr-record +19 −0
  167. tools/perf/scripts/python/bin/mem-phys-addr-report +3 −0
  168. tools/perf/scripts/python/mem-phys-addr.py +95 −0
  169. tools/perf/tests/attr.c +6 −0
  170. tools/perf/tests/backward-ring-buffer.c +3 −3
  171. tools/perf/tests/bp_signal.c +1 −1
  172. tools/perf/tests/bpf-script-example.c +2 −2
  173. tools/perf/tests/bpf.c +32 −36
  174. tools/perf/tests/builtin-test.c +5 −5
  175. tools/perf/tests/code-reading.c +1 −1
  176. tools/perf/tests/dwarf-unwind.c +1 −0
  177. tools/perf/tests/keep-tracking.c +1 −1
  178. tools/perf/tests/mmap-basic.c +1 −1
  179. tools/perf/tests/openat-syscall-tp-fields.c +4 −1
  180. tools/perf/tests/parse-events.c +0 −1
  181. tools/perf/tests/perf-record.c +1 −1
  182. tools/perf/tests/sample-parsing.c +1 −1
  183. tools/perf/tests/shell/trace+probe_vfs_getname.sh +3 −4
  184. tools/perf/tests/sw-clock.c +1 −1
  185. tools/perf/tests/switch-tracking.c +1 −1
  186. tools/perf/tests/task-exit.c +1 −1
  187. tools/perf/tests/thread-map.c +1 −1
  188. tools/perf/trace/beauty/Build +1 −0
  189. tools/perf/trace/beauty/arch_errno_names.c +1 −0
  190. tools/perf/trace/beauty/arch_errno_names.sh +100 −0
  191. tools/perf/trace/beauty/beauty.h +5 −0
  192. tools/perf/trace/beauty/flock.c +5 −5
  193. tools/perf/trace/beauty/futex_val3.c +18 −0
  194. tools/perf/ui/browsers/annotate.c +197 −202
  195. tools/perf/ui/gtk/annotate.c +13 −12
  196. tools/perf/util/Build +9 −1
  197. tools/perf/util/annotate.c +340 −312
  198. tools/perf/util/annotate.h +42 −34
  199. tools/perf/util/arm-spe-pkt-decoder.c +462 −0
  200. tools/perf/util/arm-spe-pkt-decoder.h +43 −0
  201. tools/perf/util/arm-spe.c +231 −0
  202. tools/perf/util/arm-spe.h +31 −0
  203. tools/perf/util/auxtrace.c +5 −3
  204. tools/perf/util/auxtrace.h +1 −0
  205. tools/perf/util/bpf-loader.c +2 −2
  206. tools/perf/util/callchain.c +10 −0
  207. tools/perf/util/callchain.h +2 −0
  208. tools/perf/util/cgroup.c +3 −0
  209. tools/perf/util/cs-etm-decoder/Build +1 −0
  210. tools/perf/util/cs-etm-decoder/cs-etm-decoder.c +513 −0
  211. tools/perf/util/cs-etm-decoder/cs-etm-decoder.h +105 −0
  212. tools/perf/util/cs-etm.c +1023 −0
  213. tools/perf/util/cs-etm.h +18 −0
  214. tools/perf/util/data.c +0 −10
  215. tools/perf/util/dso.c +1 −1
  216. tools/perf/util/env.c +47 −0
  217. tools/perf/util/env.h +2 −0
  218. tools/perf/util/event.c +8 −0
  219. tools/perf/util/event.h +2 −2
  220. tools/perf/util/evlist.c +43 −27
  221. tools/perf/util/evlist.h +9 −6
  222. tools/perf/util/evsel.c +163 −64
  223. tools/perf/util/evsel.h +7 −5
  224. tools/perf/util/generate-cmdlist.sh +1 −1
  225. tools/perf/util/header.c +128 −2
  226. tools/perf/util/header.h +8 −1
  227. tools/perf/util/intel-bts.c +1 −5
  228. tools/perf/util/intel-pt-decoder/Build +15 −9
  229. tools/perf/util/intel-pt.c +3 −8
  230. tools/perf/util/machine.c +2 −2
  231. tools/perf/util/map.c +1 −1
  232. tools/perf/util/metricgroup.c +6 −4
  233. tools/perf/util/mmap.c +25 −48
  234. tools/perf/util/mmap.h +2 −2
  235. tools/perf/util/ordered-events.c +1 −2
  236. tools/perf/util/ordered-events.h +1 −1
  237. tools/perf/util/parse-events.c +3 −0
  238. tools/perf/util/path.c +14 −0
  239. tools/perf/util/path.h +3 −0
  240. tools/perf/util/pmu.c +71 −16
  241. tools/perf/util/pmu.h +1 −1
  242. tools/perf/util/probe-event.c +61 −24
  243. tools/perf/util/python-ext-sources +1 −0
  244. tools/perf/util/python.c +1 −1
  245. tools/perf/util/rblist.c +12 −7
  246. tools/perf/util/rblist.h +1 −0
  247. tools/perf/util/scripting-engines/trace-event-python.c +2 −1
  248. tools/perf/util/session.c +26 −28
  249. tools/perf/util/session.h +1 −1
  250. tools/perf/util/sort.c +12 −8
  251. tools/perf/util/srcline.c +5 −4
  252. tools/perf/util/srcline.h +3 −2
  253. tools/perf/util/stat-shadow.c +258 −168
  254. tools/perf/util/stat.c +11 −4
  255. tools/perf/util/stat.h +60 −3
  256. tools/perf/util/string.c +46 −0
  257. tools/perf/util/string2.h +2 −0
  258. tools/perf/util/symbol.c +5 −0
  259. tools/perf/util/symbol.h +1 −0
  260. tools/perf/util/syscalltbl.c +4 −0
  261. tools/perf/util/target.h +7 −0
  262. tools/perf/util/thread_map.c +20 −7
  263. tools/perf/util/thread_map.h +2 −1
  264. tools/perf/util/time-utils.c +289 −12
  265. tools/perf/util/time-utils.h +8 −0
  266. tools/perf/util/tool.h +1 −0
  267. tools/perf/util/unwind-libunwind-local.c +3 −6
  268. tools/perf/util/unwind-libunwind.c +2 −2
  269. tools/perf/util/util.c +1 −1
  270. tools/perf/util/util.h +10 −0
Documentation/process/kernel-enforcement-statement.rst+1−0
@@ -118,6 +118,7 @@ we might work for today, have in the past, or will in the future. · 117 unmodified lines
118118 - Mike Marshall
119119 - Chris Mason
120120 - Paul E. McKenney
121+ - Arnaldo Carvalho de Melo
121122 - David S. Miller
122123 - Ingo Molnar
123124 - Kuninori Morimoto
MAINTAINERS+2−1
@@ -1327,7 +1327,8 @@ F: tools/perf/arch/arm/util/pmu.c · 1326 unmodified lines
13271327 F: tools/perf/arch/arm/util/auxtrace.c
13281328 F: tools/perf/arch/arm/util/cs-etm.c
13291329 F: tools/perf/arch/arm/util/cs-etm.h
1330−F: tools/perf/util/cs-etm.h
1330+F: tools/perf/util/cs-etm.*
1331+F: tools/perf/util/cs-etm-decoder/*
13311332
13321333 ARM/CORGI MACHINE SUPPORT
13331334 M: Richard Purdie <rpurdie@rpsys.net>
arch/x86/events/msr.c+43−27
@@ -10,7 +10,9 @@ enum perf_msr_id { · 9 unmodified lines
1010 PERF_MSR_SMI = 4,
1111 PERF_MSR_PTSC = 5,
1212 PERF_MSR_IRPERF = 6,
13−
13+ PERF_MSR_THERM = 7,
14+ PERF_MSR_THERM_SNAP = 8,
15+ PERF_MSR_THERM_UNIT = 9,
1416 PERF_MSR_EVENT_MAX,
1517 };
1618
@@ -29,6 +31,11 @@ static bool test_irperf(int idx) · 12 unmodified lines
2931 return boot_cpu_has(X86_FEATURE_IRPERF);
3032 }
3133
34+static bool test_therm_status(int idx)
35+{
36+ return boot_cpu_has(X86_FEATURE_DTHERM);
37+}
38+
3239 static bool test_intel(int idx)
3340 {
3441 if (boot_cpu_data.x86_vendor != X86_VENDOR_INTEL ||
@@ -95,22 +102,28 @@ struct perf_msr { · 60 unmodified lines
95102 bool (*test)(int idx);
96103 };
97104
98−PMU_EVENT_ATTR_STRING(tsc, evattr_tsc, "event=0x00");
99−PMU_EVENT_ATTR_STRING(aperf, evattr_aperf, "event=0x01");
100−PMU_EVENT_ATTR_STRING(mperf, evattr_mperf, "event=0x02");
101−PMU_EVENT_ATTR_STRING(pperf, evattr_pperf, "event=0x03");
102−PMU_EVENT_ATTR_STRING(smi, evattr_smi, "event=0x04");
103−PMU_EVENT_ATTR_STRING(ptsc, evattr_ptsc, "event=0x05");
104−PMU_EVENT_ATTR_STRING(irperf, evattr_irperf, "event=0x06");
105+PMU_EVENT_ATTR_STRING(tsc, evattr_tsc, "event=0x00" );
106+PMU_EVENT_ATTR_STRING(aperf, evattr_aperf, "event=0x01" );
107+PMU_EVENT_ATTR_STRING(mperf, evattr_mperf, "event=0x02" );
108+PMU_EVENT_ATTR_STRING(pperf, evattr_pperf, "event=0x03" );
109+PMU_EVENT_ATTR_STRING(smi, evattr_smi, "event=0x04" );
110+PMU_EVENT_ATTR_STRING(ptsc, evattr_ptsc, "event=0x05" );
111+PMU_EVENT_ATTR_STRING(irperf, evattr_irperf, "event=0x06" );
112+PMU_EVENT_ATTR_STRING(cpu_thermal_margin, evattr_therm, "event=0x07" );
113+PMU_EVENT_ATTR_STRING(cpu_thermal_margin.snapshot, evattr_therm_snap, "1" );
114+PMU_EVENT_ATTR_STRING(cpu_thermal_margin.unit, evattr_therm_unit, "C" );
105115
106116 static struct perf_msr msr[] = {
107− [PERF_MSR_TSC] = { 0, &evattr_tsc, NULL, },
108− [PERF_MSR_APERF] = { MSR_IA32_APERF, &evattr_aperf, test_aperfmperf, },
109− [PERF_MSR_MPERF] = { MSR_IA32_MPERF, &evattr_mperf, test_aperfmperf, },
110− [PERF_MSR_PPERF] = { MSR_PPERF, &evattr_pperf, test_intel, },
111− [PERF_MSR_SMI] = { MSR_SMI_COUNT, &evattr_smi, test_intel, },
112− [PERF_MSR_PTSC] = { MSR_F15H_PTSC, &evattr_ptsc, test_ptsc, },
113− [PERF_MSR_IRPERF] = { MSR_F17H_IRPERF, &evattr_irperf, test_irperf, },
117+ [PERF_MSR_TSC] = { 0, &evattr_tsc, NULL, },
118+ [PERF_MSR_APERF] = { MSR_IA32_APERF, &evattr_aperf, test_aperfmperf, },
119+ [PERF_MSR_MPERF] = { MSR_IA32_MPERF, &evattr_mperf, test_aperfmperf, },
120+ [PERF_MSR_PPERF] = { MSR_PPERF, &evattr_pperf, test_intel, },
121+ [PERF_MSR_SMI] = { MSR_SMI_COUNT, &evattr_smi, test_intel, },
122+ [PERF_MSR_PTSC] = { MSR_F15H_PTSC, &evattr_ptsc, test_ptsc, },
123+ [PERF_MSR_IRPERF] = { MSR_F17H_IRPERF, &evattr_irperf, test_irperf, },
124+ [PERF_MSR_THERM] = { MSR_IA32_THERM_STATUS, &evattr_therm, test_therm_status, },
125+ [PERF_MSR_THERM_SNAP] = { MSR_IA32_THERM_STATUS, &evattr_therm_snap, test_therm_status, },
126+ [PERF_MSR_THERM_UNIT] = { MSR_IA32_THERM_STATUS, &evattr_therm_unit, test_therm_status, },
114127 };
115128
116129 static struct attribute *events_attrs[PERF_MSR_EVENT_MAX + 1] = {
@@ -161,9 +174,9 @@ static int msr_event_init(struct perf_event *event) · 44 unmodified lines
161174 if (!msr[cfg].attr)
162175 return -EINVAL;
163176
164− event->hw.idx = -1;
165− event->hw.event_base = msr[cfg].msr;
166− event->hw.config = cfg;
177+ event->hw.idx = -1;
178+ event->hw.event_base = msr[cfg].msr;
179+ event->hw.config = cfg;
167180
168181 return 0;
169182 }
@@ -184,7 +197,7 @@ static void msr_event_update(struct perf_event *event) · 14 unmodified lines
184197 u64 prev, now;
185198 s64 delta;
186199
187− /* Careful, an NMI might modify the previous event value. */
200+ /* Careful, an NMI might modify the previous event value: */
188201 again:
189202 prev = local64_read(&event->hw.prev_count);
190203 now = msr_read_counter(event);
@@ -193,17 +206,22 @@ again: · 2 unmodified lines
193206 goto again;
194207
195208 delta = now - prev;
196− if (unlikely(event->hw.event_base == MSR_SMI_COUNT))
209+ if (unlikely(event->hw.event_base == MSR_SMI_COUNT)) {
197210 delta = sign_extend64(delta, 31);
198−
199− local64_add(delta, &event->count);
211+ local64_add(delta, &event->count);
212+ } else if (unlikely(event->hw.event_base == MSR_IA32_THERM_STATUS)) {
213+ /* If valid, extract digital readout, otherwise set to -1: */
214+ now = now & (1ULL << 31) ? (now >> 16) & 0x3f : -1;
215+ local64_set(&event->count, now);
216+ } else {
217+ local64_add(delta, &event->count);
218+ }
200219 }
201220
202221 static void msr_event_start(struct perf_event *event, int flags)
203222 {
204− u64 now;
223+ u64 now = msr_read_counter(event);
205224
206− now = msr_read_counter(event);
207225 local64_set(&event->hw.prev_count, now);
208226 }
209227
@@ -250,9 +268,7 @@ static int __init msr_init(void) · 40 unmodified lines
250268 for (i = PERF_MSR_TSC + 1; i < PERF_MSR_EVENT_MAX; i++) {
251269 u64 val;
252270
253− /*
254− * Virt sucks arse; you cannot tell if a R/O MSR is present :/
255− */
271+ /* Virt sucks; you cannot tell if a R/O MSR is present :/ */
256272 if (!msr[i].test(i) || rdmsrl_safe(msr[i].msr, &val))
257273 msr[i].attr = NULL;
258274 }
arch/x86/include/asm/uprobes.h+4−0
@@ -53,6 +53,10 @@ struct arch_uprobe { · 52 unmodified lines
5353 u8 fixups;
5454 u8 ilen;
5555 } defparam;
56+ struct {
57+ u8 reg_offset; /* to the start of pt_regs */
58+ u8 ilen;
59+ } push;
5660 };
5761 };
5862
arch/x86/kernel/uprobes.c+103−4
@@ -528,11 +528,11 @@ static int default_pre_xol_op(struct arch_uprobe *auprobe, struct pt_regs *regs) · 527 unmodified lines
528528 return 0;
529529 }
530530
531−static int push_ret_address(struct pt_regs *regs, unsigned long ip)
531+static int emulate_push_stack(struct pt_regs *regs, unsigned long val)
532532 {
533533 unsigned long new_sp = regs->sp - sizeof_long();
534534
535− if (copy_to_user((void __user *)new_sp, &ip, sizeof_long()))
535+ if (copy_to_user((void __user *)new_sp, &val, sizeof_long()))
536536 return -EFAULT;
537537
538538 regs->sp = new_sp;
@@ -566,7 +566,7 @@ static int default_post_xol_op(struct arch_uprobe *auprobe, struct pt_regs *regs · 27 unmodified lines
566566 regs->ip += correction;
567567 } else if (auprobe->defparam.fixups & UPROBE_FIX_CALL) {
568568 regs->sp += sizeof_long(); /* Pop incorrect return address */
569− if (push_ret_address(regs, utask->vaddr + auprobe->defparam.ilen))
569+ if (emulate_push_stack(regs, utask->vaddr + auprobe->defparam.ilen))
570570 return -ERESTART;
571571 }
572572 /* popf; tell the caller to not touch TF */
@@ -655,7 +655,7 @@ static bool branch_emulate_op(struct arch_uprobe *auprobe, struct pt_regs *regs) · 82 unmodified lines
655655 *
656656 * But there is corner case, see the comment in ->post_xol().
657657 */
658− if (push_ret_address(regs, new_ip))
658+ if (emulate_push_stack(regs, new_ip))
659659 return false;
660660 } else if (!check_jmp_cond(auprobe, regs)) {
661661 offs = 0;
@@ -665,6 +665,16 @@ static bool branch_emulate_op(struct arch_uprobe *auprobe, struct pt_regs *regs) · 3 unmodified lines
665665 return true;
666666 }
667667
668+static bool push_emulate_op(struct arch_uprobe *auprobe, struct pt_regs *regs)
669+{
670+ unsigned long *src_ptr = (void *)regs + auprobe->push.reg_offset;
671+
672+ if (emulate_push_stack(regs, *src_ptr))
673+ return false;
674+ regs->ip += auprobe->push.ilen;
675+ return true;
676+}
677+
668678 static int branch_post_xol_op(struct arch_uprobe *auprobe, struct pt_regs *regs)
669679 {
670680 BUG_ON(!branch_is_call(auprobe));
@@ -703,6 +713,10 @@ static const struct uprobe_xol_ops branch_xol_ops = { · 32 unmodified lines
703713 .post_xol = branch_post_xol_op,
704714 };
705715
716+static const struct uprobe_xol_ops push_xol_ops = {
717+ .emulate = push_emulate_op,
718+};
719+
706720 /* Returns -ENOSYS if branch_xol_ops doesn't handle this insn */
707721 static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
708722 {
@@ -750,6 +764,87 @@ static int branch_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn) · 41 unmodified lines
750764 return 0;
751765 }
752766
767+/* Returns -ENOSYS if push_xol_ops doesn't handle this insn */
768+static int push_setup_xol_ops(struct arch_uprobe *auprobe, struct insn *insn)
769+{
770+ u8 opc1 = OPCODE1(insn), reg_offset = 0;
771+
772+ if (opc1 < 0x50 || opc1 > 0x57)
773+ return -ENOSYS;
774+
775+ if (insn->length > 2)
776+ return -ENOSYS;
777+ if (insn->length == 2) {
778+ /* only support rex_prefix 0x41 (x64 only) */
779+#ifdef CONFIG_X86_64
780+ if (insn->rex_prefix.nbytes != 1 ||
781+ insn->rex_prefix.bytes[0] != 0x41)
782+ return -ENOSYS;
783+
784+ switch (opc1) {
785+ case 0x50:
786+ reg_offset = offsetof(struct pt_regs, r8);
787+ break;
788+ case 0x51:
789+ reg_offset = offsetof(struct pt_regs, r9);
790+ break;
791+ case 0x52:
792+ reg_offset = offsetof(struct pt_regs, r10);
793+ break;
794+ case 0x53:
795+ reg_offset = offsetof(struct pt_regs, r11);
796+ break;
797+ case 0x54:
798+ reg_offset = offsetof(struct pt_regs, r12);
799+ break;
800+ case 0x55:
801+ reg_offset = offsetof(struct pt_regs, r13);
802+ break;
803+ case 0x56:
804+ reg_offset = offsetof(struct pt_regs, r14);
805+ break;
806+ case 0x57:
807+ reg_offset = offsetof(struct pt_regs, r15);
808+ break;
809+ }
810+#else
811+ return -ENOSYS;
812+#endif
813+ } else {
814+ switch (opc1) {
815+ case 0x50:
816+ reg_offset = offsetof(struct pt_regs, ax);
817+ break;
818+ case 0x51:
819+ reg_offset = offsetof(struct pt_regs, cx);
820+ break;
821+ case 0x52:
822+ reg_offset = offsetof(struct pt_regs, dx);
823+ break;
824+ case 0x53:
825+ reg_offset = offsetof(struct pt_regs, bx);
826+ break;
827+ case 0x54:
828+ reg_offset = offsetof(struct pt_regs, sp);
829+ break;
830+ case 0x55:
831+ reg_offset = offsetof(struct pt_regs, bp);
832+ break;
833+ case 0x56:
834+ reg_offset = offsetof(struct pt_regs, si);
835+ break;
836+ case 0x57:
837+ reg_offset = offsetof(struct pt_regs, di);
838+ break;
839+ }
840+ }
841+
842+ auprobe->push.reg_offset = reg_offset;
843+ auprobe->push.ilen = insn->length;
844+ auprobe->ops = &push_xol_ops;
845+ return 0;
846+}
847+
753848 /**
754849 * arch_uprobe_analyze_insn - instruction analysis including validity and fixups.
755850 * @mm: the probed address space.
@@ -771,6 +866,10 @@ int arch_uprobe_analyze_insn(struct arch_uprobe *auprobe, struct mm_struct *mm, · 15 unmodified lines
771866 if (ret != -ENOSYS)
772867 return ret;
773868
869+ ret = push_setup_xol_ops(auprobe, &insn);
870+ if (ret != -ENOSYS)
871+ return ret;
872+
774873 /*
775874 * Figure out which fixups default_post_xol_op() will need to perform,
776875 * and annotate defparam->fixups accordingly.
arch/x86/tools/Makefile+6−6
@@ -13,28 +13,28 @@ else · 12 unmodified lines
1313 posttest_64bit = -n
1414 endif
1515
16−distill_awk = $(srctree)/arch/x86/tools/distill.awk
16+reformatter = $(srctree)/arch/x86/tools/objdump_reformat.awk
1717 chkobjdump = $(srctree)/arch/x86/tools/chkobjdump.awk
1818
1919 quiet_cmd_posttest = TEST $@
20− cmd_posttest = ($(OBJDUMP) -v | $(AWK) -f $(chkobjdump)) || $(OBJDUMP) -d -j .text $(objtree)/vmlinux | $(AWK) -f $(distill_awk) | $(obj)/test_get_len $(posttest_64bit) $(posttest_verbose)
20+ cmd_posttest = ($(OBJDUMP) -v | $(AWK) -f $(chkobjdump)) || $(OBJDUMP) -d -j .text $(objtree)/vmlinux | $(AWK) -f $(reformatter) | $(obj)/insn_decoder_test $(posttest_64bit) $(posttest_verbose)
2121
2222 quiet_cmd_sanitytest = TEST $@
2323 cmd_sanitytest = $(obj)/insn_sanity $(posttest_64bit) -m 1000000
2424
25−posttest: $(obj)/test_get_len vmlinux $(obj)/insn_sanity
25+posttest: $(obj)/insn_decoder_test vmlinux $(obj)/insn_sanity
2626 $(call cmd,posttest)
2727 $(call cmd,sanitytest)
2828
29−hostprogs-y += test_get_len insn_sanity
29+hostprogs-y += insn_decoder_test insn_sanity
3030
3131 # -I needed for generated C source and C source which in the kernel tree.
32−HOSTCFLAGS_test_get_len.o := -Wall -I$(objtree)/arch/x86/lib/ -I$(srctree)/arch/x86/include/uapi/ -I$(srctree)/arch/x86/include/ -I$(srctree)/arch/x86/lib/ -I$(srctree)/include/uapi/
32+HOSTCFLAGS_insn_decoder_test.o := -Wall -I$(objtree)/arch/x86/lib/ -I$(srctree)/arch/x86/include/uapi/ -I$(srctree)/arch/x86/include/ -I$(srctree)/arch/x86/lib/ -I$(srctree)/include/uapi/
3333
3434 HOSTCFLAGS_insn_sanity.o := -Wall -I$(objtree)/arch/x86/lib/ -I$(srctree)/arch/x86/include/ -I$(srctree)/arch/x86/lib/ -I$(srctree)/include/
3535
3636 # Dependencies are also needed.
37−$(obj)/test_get_len.o: $(srctree)/arch/x86/lib/insn.c $(srctree)/arch/x86/lib/inat.c $(srctree)/arch/x86/include/asm/inat_types.h $(srctree)/arch/x86/include/asm/inat.h $(srctree)/arch/x86/include/asm/insn.h $(objtree)/arch/x86/lib/inat-tables.c
37+$(obj)/insn_decoder_test.o: $(srctree)/arch/x86/lib/insn.c $(srctree)/arch/x86/lib/inat.c $(srctree)/arch/x86/include/asm/inat_types.h $(srctree)/arch/x86/include/asm/inat.h $(srctree)/arch/x86/include/asm/insn.h $(objtree)/arch/x86/lib/inat-tables.c
3838
3939 $(obj)/insn_sanity.o: $(srctree)/arch/x86/lib/insn.c $(srctree)/arch/x86/lib/inat.c $(srctree)/arch/x86/include/asm/inat_types.h $(srctree)/arch/x86/include/asm/inat.h $(srctree)/arch/x86/include/asm/insn.h $(objtree)/arch/x86/lib/inat-tables.c
4040
arch/x86/tools/insn_decoder_test.crenamed from arch/x86/tools/test_get_len.c+25−18
@@ -9,10 +9,6 @@ · 8 unmodified lines
99 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
1010 * GNU General Public License for more details.
1111 *
12− * You should have received a copy of the GNU General Public License
13− * along with this program; if not, write to the Free Software
14− * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
15− *
1612 * Copyright (C) IBM Corporation, 2009
1713 */
1814
@@ -21,6 +17,7 @@ · 2 unmodified lines
2117 #include <string.h>
2218 #include <assert.h>
2319 #include <unistd.h>
20+#include <stdarg.h>
2421
2522 #define unlikely(cond) (cond)
2623
@@ -33,7 +30,7 @@ · 6 unmodified lines
3330 * particular. See if insn_get_length() and the disassembler agree
3431 * on the length of each instruction in an elf disassembly.
3532 *
36− * Usage: objdump -d a.out | awk -f distill.awk | ./test_get_len
33+ * Usage: objdump -d a.out | awk -f objdump_reformat.awk | ./insn_decoder_test
3734 */
3835
3936 const char *prog;
@@ -42,8 +39,8 @@ static int x86_64; · 2 unmodified lines
4239
4340 static void usage(void)
4441 {
45− fprintf(stderr, "Usage: objdump -d a.out | awk -f distill.awk |"
46− " %s [-y|-n] [-v]\n", prog);
42+ fprintf(stderr, "Usage: objdump -d a.out | awk -f objdump_reformat.awk"
43+ " | %s [-y|-n] [-v]\n", prog);
4744 fprintf(stderr, "\t-y 64bit mode\n");
4845 fprintf(stderr, "\t-n 32bit mode\n");
4946 fprintf(stderr, "\t-v verbose mode\n");
@@ -52,10 +49,21 @@ static void usage(void) · 2 unmodified lines
5249
5350 static void malformed_line(const char *line, int line_nr)
5451 {
55− fprintf(stderr, "%s: malformed line %d:\n%s", prog, line_nr, line);
52+ fprintf(stderr, "%s: error: malformed line %d:\n%s",
53+ prog, line_nr, line);
5654 exit(3);
5755 }
5856
57+static void pr_warn(const char *fmt, ...)
58+{
59+ va_list ap;
60+
61+ fprintf(stderr, "%s: warning: ", prog);
62+ va_start(ap, fmt);
63+ vfprintf(stderr, fmt, ap);
64+ va_end(ap);
65+}
66+
5967 static void dump_field(FILE *fp, const char *name, const char *indent,
6068 struct insn_field *field)
6169 {
@@ -153,21 +161,20 @@ int main(int argc, char **argv) · 91 unmodified lines
153161 insn_get_length(&insn);
154162 if (insn.length != nb) {
155163 warnings++;
156− fprintf(stderr, "Warning: %s found difference at %s\n",
157− prog, sym);
158− fprintf(stderr, "Warning: %s", line);
159− fprintf(stderr, "Warning: objdump says %d bytes, but "
160− "insn_get_length() says %d\n", nb,
161− insn.length);
164+ pr_warn("Found an x86 instruction decoder bug, "
165+ "please report this.\n", sym);
166+ pr_warn("%s", line);
167+ pr_warn("objdump says %d bytes, but insn_get_length() "
168+ "says %d\n", nb, insn.length);
162169 if (verbose)
163170 dump_insn(stderr, &insn);
164171 }
165172 }
166173 if (warnings)
167− fprintf(stderr, "Warning: decoded and checked %d"
168− " instructions with %d warnings\n", insns, warnings);
174+ pr_warn("Decoded and checked %d instructions with %d "
175+ "failures\n", insns, warnings);
169176 else
170− fprintf(stdout, "Success: decoded and checked %d"
171− " instructions\n", insns);
177+ fprintf(stdout, "%s: success: Decoded and checked %d"
178+ " instructions\n", prog, insns);
172179 return 0;
173180 }
arch/x86/tools/objdump_reformat.awkrenamed from arch/x86/tools/distill.awk+2−2
@@ -1,7 +1,7 @@
11 #!/bin/awk -f
22 # SPDX-License-Identifier: GPL-2.0
3−# Usage: objdump -d a.out | awk -f distill.awk | ./test_get_len
4−# Distills the disassembly as follows:
3+# Usage: objdump -d a.out | awk -f objdump_reformat.awk | ./insn_decoder_test
4+# Reformats the disassembly as follows:
55 # - Removes all lines except the disassembled instructions.
66 # - For instructions that exceed 1 line (7 bytes), crams all the hex bytes
77 # into a single line.
include/uapi/linux/perf_event.h+7−3
@@ -612,9 +612,12 @@ struct perf_event_mmap_page { · 611 unmodified lines
612612 */
613613 #define PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT (1 << 12)
614614 /*
615− * PERF_RECORD_MISC_MMAP_DATA and PERF_RECORD_MISC_COMM_EXEC are used on
616− * different events so can reuse the same bit position.
617− * Ditto PERF_RECORD_MISC_SWITCH_OUT.
615+ * Following PERF_RECORD_MISC_* are used on different
616+ * events, so can reuse the same bit position:
617+ *
618+ * PERF_RECORD_MISC_MMAP_DATA - PERF_RECORD_MMAP* events
619+ * PERF_RECORD_MISC_COMM_EXEC - PERF_RECORD_COMM event
620+ * PERF_RECORD_MISC_SWITCH_OUT - PERF_RECORD_SWITCH* events
618621 */
619622 #define PERF_RECORD_MISC_MMAP_DATA (1 << 13)
620623 #define PERF_RECORD_MISC_COMM_EXEC (1 << 13)
@@ -864,6 +867,7 @@ enum perf_event_type { · 243 unmodified lines
864867 * struct perf_event_header header;
865868 * u32 pid;
866869 * u32 tid;
870+ * struct sample_id sample_id;
867871 * };
868872 */
869873 PERF_RECORD_ITRACE_START = 12,
kernel/events/callchain.c+0−15
@@ -178,21 +178,6 @@ put_callchain_entry(int rctx) · 177 unmodified lines
178178 put_recursion_context(this_cpu_ptr(callchain_recursion), rctx);
179179 }
180180
181−struct perf_callchain_entry *
182−perf_callchain(struct perf_event *event, struct pt_regs *regs)
183−{
184− bool kernel = !event->attr.exclude_callchain_kernel;
185− bool user = !event->attr.exclude_callchain_user;
186− /* Disallow cross-task user callchains. */
187− bool crosstask = event->ctx->task && event->ctx->task != current;
188− const u32 max_stack = event->attr.sample_max_stack;
189−
190− if (!kernel && !user)
191− return NULL;
192−
193− return get_perf_callchain(regs, 0, kernel, user, max_stack, crosstask, true);
194−}
195−
196181 struct perf_callchain_entry *
197182 get_perf_callchain(struct pt_regs *regs, u32 init_nr, bool kernel, bool user,
198183 u32 max_stack, bool crosstask, bool add_mark)
kernel/events/core.c+39−15
@@ -5824,19 +5824,11 @@ void perf_output_sample(struct perf_output_handle *handle, · 5823 unmodified lines
58245824 perf_output_read(handle, event);
58255825
58265826 if (sample_type & PERF_SAMPLE_CALLCHAIN) {
5827− if (data->callchain) {
5828− int size = 1;
5829−
5830− if (data->callchain)
5831− size += data->callchain->nr;
5832−
5833− size *= sizeof(u64);
5827+ int size = 1;
58345828
5835− __output_copy(handle, data->callchain, size);
5836− } else {
5837− u64 nr = 0;
5838− perf_output_put(handle, nr);
5839− }
5829+ size += data->callchain->nr;
5830+ size *= sizeof(u64);
5831+ __output_copy(handle, data->callchain, size);
58405832 }
58415833
58425834 if (sample_type & PERF_SAMPLE_RAW) {
@@ -5989,6 +5981,26 @@ static u64 perf_virt_to_phys(u64 virt) · 146 unmodified lines
59895981 return phys_addr;
59905982 }
59915983
5984+static struct perf_callchain_entry __empty_callchain = { .nr = 0, };
5985+
5986+static struct perf_callchain_entry *
5987+perf_callchain(struct perf_event *event, struct pt_regs *regs)
5988+{
5989+ bool kernel = !event->attr.exclude_callchain_kernel;
5990+ bool user = !event->attr.exclude_callchain_user;
5991+ /* Disallow cross-task user callchains. */
5992+ bool crosstask = event->ctx->task && event->ctx->task != current;
5993+ const u32 max_stack = event->attr.sample_max_stack;
5994+ struct perf_callchain_entry *callchain;
5995+
5996+ if (!kernel && !user)
5997+ return &__empty_callchain;
5998+
5999+ callchain = get_perf_callchain(regs, 0, kernel, user,
6000+ max_stack, crosstask, true);
6001+ return callchain ?: &__empty_callchain;
6002+}
6003+
59926004 void perf_prepare_sample(struct perf_event_header *header,
59936005 struct perf_sample_data *data,
59946006 struct perf_event *event,
@@ -6011,9 +6023,7 @@ void perf_prepare_sample(struct perf_event_header *header, · 16 unmodified lines
60116023 int size = 1;
60126024
60136025 data->callchain = perf_callchain(event, regs);
6014−
6015− if (data->callchain)
6016− size += data->callchain->nr;
6026+ size += data->callchain->nr;
60176027
60186028 header->size += size * sizeof(u64);
60196029 }
@@ -10740,6 +10750,19 @@ inherit_event(struct perf_event *parent_event, · 4720 unmodified lines
1074010750 if (IS_ERR(child_event))
1074110751 return child_event;
1074210752
10753+
10754+ if ((child_event->attach_state & PERF_ATTACH_TASK_DATA) &&
10755+ !child_ctx->task_ctx_data) {
10756+ struct pmu *pmu = child_event->pmu;
10757+
10758+ child_ctx->task_ctx_data = kzalloc(pmu->task_ctx_size,
10759+ GFP_KERNEL);
10760+ if (!child_ctx->task_ctx_data) {
10761+ free_event(child_event);
10762+ return NULL;
10763+ }
10764+ }
10765+
1074310766 /*
1074410767 * is_orphaned_event() and list_add_tail(&parent_event->child_list)
1074510768 * must be under the same lock in order to serialize against
@@ -10750,6 +10773,7 @@ inherit_event(struct perf_event *parent_event, · 4 unmodified lines
1075010773 if (is_orphaned_event(parent_event) ||
1075110774 !atomic_long_inc_not_zero(&parent_event->refcount)) {
1075210775 mutex_unlock(&parent_event->child_mutex);
10776+ /* task_ctx_data is freed with child_ctx */
1075310777 free_event(child_event);
1075410778 return NULL;
1075510779 }
kernel/events/internal.h+0−4
@@ -201,10 +201,6 @@ arch_perf_out_copy_user(void *dst, const void *src, unsigned long n) · 200 unmodified lines
201201
202202 DEFINE_OUTPUT_COPY(__output_copy_user, arch_perf_out_copy_user)
203203
204−/* Callchain handling */
205−extern struct perf_callchain_entry *
206−perf_callchain(struct perf_event *event, struct pt_regs *regs);
207−
208204 static inline int get_recursion_context(int *recursion)
209205 {
210206 int rctx;
tools/arch/alpha/include/uapi/asm/errno.h+128−0
@@ -0,0 +1,128 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+#ifndef _ALPHA_ERRNO_H
3+#define _ALPHA_ERRNO_H
4+
5+#include <asm-generic/errno-base.h>
6+
7+#undef EAGAIN /* 11 in errno-base.h */
8+
9+#define EDEADLK 11 /* Resource deadlock would occur */
10+
11+#define EAGAIN 35 /* Try again */
12+#define EWOULDBLOCK EAGAIN /* Operation would block */
13+#define EINPROGRESS 36 /* Operation now in progress */
14+#define EALREADY 37 /* Operation already in progress */
15+#define ENOTSOCK 38 /* Socket operation on non-socket */
16+#define EDESTADDRREQ 39 /* Destination address required */
17+#define EMSGSIZE 40 /* Message too long */
18+#define EPROTOTYPE 41 /* Protocol wrong type for socket */
19+#define ENOPROTOOPT 42 /* Protocol not available */
20+#define EPROTONOSUPPORT 43 /* Protocol not supported */
21+#define ESOCKTNOSUPPORT 44 /* Socket type not supported */
22+#define EOPNOTSUPP 45 /* Operation not supported on transport endpoint */
23+#define EPFNOSUPPORT 46 /* Protocol family not supported */
24+#define EAFNOSUPPORT 47 /* Address family not supported by protocol */
25+#define EADDRINUSE 48 /* Address already in use */
26+#define EADDRNOTAVAIL 49 /* Cannot assign requested address */
27+#define ENETDOWN 50 /* Network is down */
28+#define ENETUNREACH 51 /* Network is unreachable */
29+#define ENETRESET 52 /* Network dropped connection because of reset */
30+#define ECONNABORTED 53 /* Software caused connection abort */
31+#define ECONNRESET 54 /* Connection reset by peer */
32+#define ENOBUFS 55 /* No buffer space available */
33+#define EISCONN 56 /* Transport endpoint is already connected */
34+#define ENOTCONN 57 /* Transport endpoint is not connected */
35+#define ESHUTDOWN 58 /* Cannot send after transport endpoint shutdown */
36+#define ETOOMANYREFS 59 /* Too many references: cannot splice */
37+#define ETIMEDOUT 60 /* Connection timed out */
38+#define ECONNREFUSED 61 /* Connection refused */
39+#define ELOOP 62 /* Too many symbolic links encountered */
40+#define ENAMETOOLONG 63 /* File name too long */
41+#define EHOSTDOWN 64 /* Host is down */
42+#define EHOSTUNREACH 65 /* No route to host */
43+#define ENOTEMPTY 66 /* Directory not empty */
44+
45+#define EUSERS 68 /* Too many users */
46+#define EDQUOT 69 /* Quota exceeded */
47+#define ESTALE 70 /* Stale file handle */
48+#define EREMOTE 71 /* Object is remote */
49+
50+#define ENOLCK 77 /* No record locks available */
51+#define ENOSYS 78 /* Function not implemented */
52+
53+#define ENOMSG 80 /* No message of desired type */
54+#define EIDRM 81 /* Identifier removed */
55+#define ENOSR 82 /* Out of streams resources */
56+#define ETIME 83 /* Timer expired */
57+#define EBADMSG 84 /* Not a data message */
58+#define EPROTO 85 /* Protocol error */
59+#define ENODATA 86 /* No data available */
60+#define ENOSTR 87 /* Device not a stream */
61+
62+#define ENOPKG 92 /* Package not installed */
63+
64+#define EILSEQ 116 /* Illegal byte sequence */
65+
66+/* The following are just random noise.. */
67+#define ECHRNG 88 /* Channel number out of range */
68+#define EL2NSYNC 89 /* Level 2 not synchronized */
69+#define EL3HLT 90 /* Level 3 halted */
70+#define EL3RST 91 /* Level 3 reset */
71+
72+#define ELNRNG 93 /* Link number out of range */
73+#define EUNATCH 94 /* Protocol driver not attached */
74+#define ENOCSI 95 /* No CSI structure available */
75+#define EL2HLT 96 /* Level 2 halted */
76+#define EBADE 97 /* Invalid exchange */
77+#define EBADR 98 /* Invalid request descriptor */
78+#define EXFULL 99 /* Exchange full */
79+#define ENOANO 100 /* No anode */
80+#define EBADRQC 101 /* Invalid request code */
81+#define EBADSLT 102 /* Invalid slot */
82+
83+#define EDEADLOCK EDEADLK
84+
85+#define EBFONT 104 /* Bad font file format */
86+#define ENONET 105 /* Machine is not on the network */
87+#define ENOLINK 106 /* Link has been severed */
88+#define EADV 107 /* Advertise error */
89+#define ESRMNT 108 /* Srmount error */
90+#define ECOMM 109 /* Communication error on send */
91+#define EMULTIHOP 110 /* Multihop attempted */
92+#define EDOTDOT 111 /* RFS specific error */
93+#define EOVERFLOW 112 /* Value too large for defined data type */
94+#define ENOTUNIQ 113 /* Name not unique on network */
95+#define EBADFD 114 /* File descriptor in bad state */
96+#define EREMCHG 115 /* Remote address changed */
97+
98+#define EUCLEAN 117 /* Structure needs cleaning */
99+#define ENOTNAM 118 /* Not a XENIX named type file */
100+#define ENAVAIL 119 /* No XENIX semaphores available */
101+#define EISNAM 120 /* Is a named type file */
102+#define EREMOTEIO 121 /* Remote I/O error */
103+
104+#define ELIBACC 122 /* Can not access a needed shared library */
105+#define ELIBBAD 123 /* Accessing a corrupted shared library */
106+#define ELIBSCN 124 /* .lib section in a.out corrupted */
107+#define ELIBMAX 125 /* Attempting to link in too many shared libraries */
108+#define ELIBEXEC 126 /* Cannot exec a shared library directly */
109+#define ERESTART 127 /* Interrupted system call should be restarted */
110+#define ESTRPIPE 128 /* Streams pipe error */
111+
112+#define ENOMEDIUM 129 /* No medium found */
113+#define EMEDIUMTYPE 130 /* Wrong medium type */
114+#define ECANCELED 131 /* Operation Cancelled */
115+#define ENOKEY 132 /* Required key not available */
116+#define EKEYEXPIRED 133 /* Key has expired */
117+#define EKEYREVOKED 134 /* Key has been revoked */
118+#define EKEYREJECTED 135 /* Key was rejected by service */
119+
120+/* for robust mutexes */
121+#define EOWNERDEAD 136 /* Owner died */
122+#define ENOTRECOVERABLE 137 /* State not recoverable */
123+
124+#define ERFKILL 138 /* Operation not possible due to RF-kill */
125+
126+#define EHWPOISON 139 /* Memory page has hardware error */
127+
128+#endif
tools/arch/mips/include/asm/errno.h+17−0
@@ -0,0 +1,17 @@
1+/*
2+ * This file is subject to the terms and conditions of the GNU General Public
3+ * License. See the file "COPYING" in the main directory of this archive
4+ * for more details.
5+ *
6+ * Copyright (C) 1995, 1999, 2001, 2002 by Ralf Baechle
7+ */
8+#ifndef _ASM_ERRNO_H
9+#define _ASM_ERRNO_H
10+
11+#include <uapi/asm/errno.h>
12+
13+
14+/* The biggest error number defined here or in <linux/errno.h>. */
15+#define EMAXERRNO 1133
16+
17+#endif /* _ASM_ERRNO_H */
tools/arch/mips/include/uapi/asm/errno.h+130−0
@@ -0,0 +1,130 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+/*
3+ * This file is subject to the terms and conditions of the GNU General Public
4+ * License. See the file "COPYING" in the main directory of this archive
5+ * for more details.
6+ *
7+ * Copyright (C) 1995, 1999, 2001, 2002 by Ralf Baechle
8+ */
9+#ifndef _UAPI_ASM_ERRNO_H
10+#define _UAPI_ASM_ERRNO_H
11+
12+/*
13+ * These error numbers are intended to be MIPS ABI compatible
14+ */
15+
16+#include <asm-generic/errno-base.h>
17+
18+#define ENOMSG 35 /* No message of desired type */
19+#define EIDRM 36 /* Identifier removed */
20+#define ECHRNG 37 /* Channel number out of range */
21+#define EL2NSYNC 38 /* Level 2 not synchronized */
22+#define EL3HLT 39 /* Level 3 halted */
23+#define EL3RST 40 /* Level 3 reset */
24+#define ELNRNG 41 /* Link number out of range */
25+#define EUNATCH 42 /* Protocol driver not attached */
26+#define ENOCSI 43 /* No CSI structure available */
27+#define EL2HLT 44 /* Level 2 halted */
28+#define EDEADLK 45 /* Resource deadlock would occur */
29+#define ENOLCK 46 /* No record locks available */
30+#define EBADE 50 /* Invalid exchange */
31+#define EBADR 51 /* Invalid request descriptor */
32+#define EXFULL 52 /* Exchange full */
33+#define ENOANO 53 /* No anode */
34+#define EBADRQC 54 /* Invalid request code */
35+#define EBADSLT 55 /* Invalid slot */
36+#define EDEADLOCK 56 /* File locking deadlock error */
37+#define EBFONT 59 /* Bad font file format */
38+#define ENOSTR 60 /* Device not a stream */
39+#define ENODATA 61 /* No data available */
40+#define ETIME 62 /* Timer expired */
41+#define ENOSR 63 /* Out of streams resources */
42+#define ENONET 64 /* Machine is not on the network */
43+#define ENOPKG 65 /* Package not installed */
44+#define EREMOTE 66 /* Object is remote */
45+#define ENOLINK 67 /* Link has been severed */
46+#define EADV 68 /* Advertise error */
47+#define ESRMNT 69 /* Srmount error */
48+#define ECOMM 70 /* Communication error on send */
49+#define EPROTO 71 /* Protocol error */
50+#define EDOTDOT 73 /* RFS specific error */
51+#define EMULTIHOP 74 /* Multihop attempted */
52+#define EBADMSG 77 /* Not a data message */
53+#define ENAMETOOLONG 78 /* File name too long */
54+#define EOVERFLOW 79 /* Value too large for defined data type */
55+#define ENOTUNIQ 80 /* Name not unique on network */
56+#define EBADFD 81 /* File descriptor in bad state */
57+#define EREMCHG 82 /* Remote address changed */
58+#define ELIBACC 83 /* Can not access a needed shared library */
59+#define ELIBBAD 84 /* Accessing a corrupted shared library */
60+#define ELIBSCN 85 /* .lib section in a.out corrupted */
61+#define ELIBMAX 86 /* Attempting to link in too many shared libraries */
62+#define ELIBEXEC 87 /* Cannot exec a shared library directly */
63+#define EILSEQ 88 /* Illegal byte sequence */
64+#define ENOSYS 89 /* Function not implemented */
65+#define ELOOP 90 /* Too many symbolic links encountered */
66+#define ERESTART 91 /* Interrupted system call should be restarted */
67+#define ESTRPIPE 92 /* Streams pipe error */
68+#define ENOTEMPTY 93 /* Directory not empty */
69+#define EUSERS 94 /* Too many users */
70+#define ENOTSOCK 95 /* Socket operation on non-socket */
71+#define EDESTADDRREQ 96 /* Destination address required */
72+#define EMSGSIZE 97 /* Message too long */
73+#define EPROTOTYPE 98 /* Protocol wrong type for socket */
74+#define ENOPROTOOPT 99 /* Protocol not available */
75+#define EPROTONOSUPPORT 120 /* Protocol not supported */
76+#define ESOCKTNOSUPPORT 121 /* Socket type not supported */
77+#define EOPNOTSUPP 122 /* Operation not supported on transport endpoint */
78+#define EPFNOSUPPORT 123 /* Protocol family not supported */
79+#define EAFNOSUPPORT 124 /* Address family not supported by protocol */
80+#define EADDRINUSE 125 /* Address already in use */
81+#define EADDRNOTAVAIL 126 /* Cannot assign requested address */
82+#define ENETDOWN 127 /* Network is down */
83+#define ENETUNREACH 128 /* Network is unreachable */
84+#define ENETRESET 129 /* Network dropped connection because of reset */
85+#define ECONNABORTED 130 /* Software caused connection abort */
86+#define ECONNRESET 131 /* Connection reset by peer */
87+#define ENOBUFS 132 /* No buffer space available */
88+#define EISCONN 133 /* Transport endpoint is already connected */
89+#define ENOTCONN 134 /* Transport endpoint is not connected */
90+#define EUCLEAN 135 /* Structure needs cleaning */
91+#define ENOTNAM 137 /* Not a XENIX named type file */
92+#define ENAVAIL 138 /* No XENIX semaphores available */
93+#define EISNAM 139 /* Is a named type file */
94+#define EREMOTEIO 140 /* Remote I/O error */
95+#define EINIT 141 /* Reserved */
96+#define EREMDEV 142 /* Error 142 */
97+#define ESHUTDOWN 143 /* Cannot send after transport endpoint shutdown */
98+#define ETOOMANYREFS 144 /* Too many references: cannot splice */
99+#define ETIMEDOUT 145 /* Connection timed out */
100+#define ECONNREFUSED 146 /* Connection refused */
101+#define EHOSTDOWN 147 /* Host is down */
102+#define EHOSTUNREACH 148 /* No route to host */
103+#define EWOULDBLOCK EAGAIN /* Operation would block */
104+#define EALREADY 149 /* Operation already in progress */
105+#define EINPROGRESS 150 /* Operation now in progress */
106+#define ESTALE 151 /* Stale file handle */
107+#define ECANCELED 158 /* AIO operation canceled */
108+
109+/*
110+ * These error are Linux extensions.
111+ */
112+#define ENOMEDIUM 159 /* No medium found */
113+#define EMEDIUMTYPE 160 /* Wrong medium type */
114+#define ENOKEY 161 /* Required key not available */
115+#define EKEYEXPIRED 162 /* Key has expired */
116+#define EKEYREVOKED 163 /* Key has been revoked */
117+#define EKEYREJECTED 164 /* Key was rejected by service */
118+
119+/* for robust mutexes */
120+#define EOWNERDEAD 165 /* Owner died */
121+#define ENOTRECOVERABLE 166 /* State not recoverable */
122+
123+#define ERFKILL 167 /* Operation not possible due to RF-kill */
124+
125+#define EHWPOISON 168 /* Memory page has hardware error */
126+
127+#define EDQUOT 1133 /* Quota exceeded */
128+
129+
130+#endif /* _UAPI_ASM_ERRNO_H */
tools/arch/parisc/include/uapi/asm/errno.h+128−0
@@ -0,0 +1,128 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+#ifndef _PARISC_ERRNO_H
3+#define _PARISC_ERRNO_H
4+
5+#include <asm-generic/errno-base.h>
6+
7+#define ENOMSG 35 /* No message of desired type */
8+#define EIDRM 36 /* Identifier removed */
9+#define ECHRNG 37 /* Channel number out of range */
10+#define EL2NSYNC 38 /* Level 2 not synchronized */
11+#define EL3HLT 39 /* Level 3 halted */
12+#define EL3RST 40 /* Level 3 reset */
13+#define ELNRNG 41 /* Link number out of range */
14+#define EUNATCH 42 /* Protocol driver not attached */
15+#define ENOCSI 43 /* No CSI structure available */
16+#define EL2HLT 44 /* Level 2 halted */
17+#define EDEADLK 45 /* Resource deadlock would occur */
18+#define EDEADLOCK EDEADLK
19+#define ENOLCK 46 /* No record locks available */
20+#define EILSEQ 47 /* Illegal byte sequence */
21+
22+#define ENONET 50 /* Machine is not on the network */
23+#define ENODATA 51 /* No data available */
24+#define ETIME 52 /* Timer expired */
25+#define ENOSR 53 /* Out of streams resources */
26+#define ENOSTR 54 /* Device not a stream */
27+#define ENOPKG 55 /* Package not installed */
28+
29+#define ENOLINK 57 /* Link has been severed */
30+#define EADV 58 /* Advertise error */
31+#define ESRMNT 59 /* Srmount error */
32+#define ECOMM 60 /* Communication error on send */
33+#define EPROTO 61 /* Protocol error */
34+
35+#define EMULTIHOP 64 /* Multihop attempted */
36+
37+#define EDOTDOT 66 /* RFS specific error */
38+#define EBADMSG 67 /* Not a data message */
39+#define EUSERS 68 /* Too many users */
40+#define EDQUOT 69 /* Quota exceeded */
41+#define ESTALE 70 /* Stale file handle */
42+#define EREMOTE 71 /* Object is remote */
43+#define EOVERFLOW 72 /* Value too large for defined data type */
44+
45+/* these errnos are defined by Linux but not HPUX. */
46+
47+#define EBADE 160 /* Invalid exchange */
48+#define EBADR 161 /* Invalid request descriptor */
49+#define EXFULL 162 /* Exchange full */
50+#define ENOANO 163 /* No anode */
51+#define EBADRQC 164 /* Invalid request code */
52+#define EBADSLT 165 /* Invalid slot */
53+#define EBFONT 166 /* Bad font file format */
54+#define ENOTUNIQ 167 /* Name not unique on network */
55+#define EBADFD 168 /* File descriptor in bad state */
56+#define EREMCHG 169 /* Remote address changed */
57+#define ELIBACC 170 /* Can not access a needed shared library */
58+#define ELIBBAD 171 /* Accessing a corrupted shared library */
59+#define ELIBSCN 172 /* .lib section in a.out corrupted */
60+#define ELIBMAX 173 /* Attempting to link in too many shared libraries */
61+#define ELIBEXEC 174 /* Cannot exec a shared library directly */
62+#define ERESTART 175 /* Interrupted system call should be restarted */
63+#define ESTRPIPE 176 /* Streams pipe error */
64+#define EUCLEAN 177 /* Structure needs cleaning */
65+#define ENOTNAM 178 /* Not a XENIX named type file */
66+#define ENAVAIL 179 /* No XENIX semaphores available */
67+#define EISNAM 180 /* Is a named type file */
68+#define EREMOTEIO 181 /* Remote I/O error */
69+#define ENOMEDIUM 182 /* No medium found */
70+#define EMEDIUMTYPE 183 /* Wrong medium type */
71+#define ENOKEY 184 /* Required key not available */
72+#define EKEYEXPIRED 185 /* Key has expired */
73+#define EKEYREVOKED 186 /* Key has been revoked */
74+#define EKEYREJECTED 187 /* Key was rejected by service */
75+
76+/* We now return you to your regularly scheduled HPUX. */
77+
78+#define ENOSYM 215 /* symbol does not exist in executable */
79+#define ENOTSOCK 216 /* Socket operation on non-socket */
80+#define EDESTADDRREQ 217 /* Destination address required */
81+#define EMSGSIZE 218 /* Message too long */
82+#define EPROTOTYPE 219 /* Protocol wrong type for socket */
83+#define ENOPROTOOPT 220 /* Protocol not available */
84+#define EPROTONOSUPPORT 221 /* Protocol not supported */
85+#define ESOCKTNOSUPPORT 222 /* Socket type not supported */
86+#define EOPNOTSUPP 223 /* Operation not supported on transport endpoint */
87+#define EPFNOSUPPORT 224 /* Protocol family not supported */
88+#define EAFNOSUPPORT 225 /* Address family not supported by protocol */
89+#define EADDRINUSE 226 /* Address already in use */
90+#define EADDRNOTAVAIL 227 /* Cannot assign requested address */
91+#define ENETDOWN 228 /* Network is down */
92+#define ENETUNREACH 229 /* Network is unreachable */
93+#define ENETRESET 230 /* Network dropped connection because of reset */
94+#define ECONNABORTED 231 /* Software caused connection abort */
95+#define ECONNRESET 232 /* Connection reset by peer */
96+#define ENOBUFS 233 /* No buffer space available */
97+#define EISCONN 234 /* Transport endpoint is already connected */
98+#define ENOTCONN 235 /* Transport endpoint is not connected */
99+#define ESHUTDOWN 236 /* Cannot send after transport endpoint shutdown */
100+#define ETOOMANYREFS 237 /* Too many references: cannot splice */
101+#define ETIMEDOUT 238 /* Connection timed out */
102+#define ECONNREFUSED 239 /* Connection refused */
103+#define EREFUSED ECONNREFUSED /* for HP's NFS apparently */
104+#define EREMOTERELEASE 240 /* Remote peer released connection */
105+#define EHOSTDOWN 241 /* Host is down */
106+#define EHOSTUNREACH 242 /* No route to host */
107+
108+#define EALREADY 244 /* Operation already in progress */
109+#define EINPROGRESS 245 /* Operation now in progress */
110+#define EWOULDBLOCK EAGAIN /* Operation would block (Not HPUX compliant) */
111+#define ENOTEMPTY 247 /* Directory not empty */
112+#define ENAMETOOLONG 248 /* File name too long */
113+#define ELOOP 249 /* Too many symbolic links encountered */
114+#define ENOSYS 251 /* Function not implemented */
115+
116+#define ENOTSUP 252 /* Function not implemented (POSIX.4 / HPUX) */
117+#define ECANCELLED 253 /* aio request was canceled before complete (POSIX.4 / HPUX) */
118+#define ECANCELED ECANCELLED /* SuSv3 and Solaris wants one 'L' */
119+
120+/* for robust mutexes */
121+#define EOWNERDEAD 254 /* Owner died */
122+#define ENOTRECOVERABLE 255 /* State not recoverable */
123+
124+#define ERFKILL 256 /* Operation not possible due to RF-kill */
125+
126+#define EHWPOISON 257 /* Memory page has hardware error */
127+
128+#endif
tools/arch/powerpc/include/uapi/asm/errno.h+10−0
@@ -0,0 +1,10 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+#ifndef _ASM_POWERPC_ERRNO_H
3+#define _ASM_POWERPC_ERRNO_H
4+
5+#include <asm-generic/errno.h>
6+
7+#undef EDEADLOCK
8+#define EDEADLOCK 58 /* File locking deadlock error */
9+
10+#endif /* _ASM_POWERPC_ERRNO_H */
tools/arch/s390/include/uapi/asm/unistd.h+412−0
@@ -0,0 +1,412 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+/*
3+ * S390 version
4+ *
5+ * Derived from "include/asm-i386/unistd.h"
6+ */
7+
8+#ifndef _UAPI_ASM_S390_UNISTD_H_
9+#define _UAPI_ASM_S390_UNISTD_H_
10+
11+/*
12+ * This file contains the system call numbers.
13+ */
14+
15+#define __NR_exit 1
16+#define __NR_fork 2
17+#define __NR_read 3
18+#define __NR_write 4
19+#define __NR_open 5
20+#define __NR_close 6
21+#define __NR_restart_syscall 7
22+#define __NR_creat 8
23+#define __NR_link 9
24+#define __NR_unlink 10
25+#define __NR_execve 11
26+#define __NR_chdir 12
27+#define __NR_mknod 14
28+#define __NR_chmod 15
29+#define __NR_lseek 19
30+#define __NR_getpid 20
31+#define __NR_mount 21
32+#define __NR_umount 22
33+#define __NR_ptrace 26
34+#define __NR_alarm 27
35+#define __NR_pause 29
36+#define __NR_utime 30
37+#define __NR_access 33
38+#define __NR_nice 34
39+#define __NR_sync 36
40+#define __NR_kill 37
41+#define __NR_rename 38
42+#define __NR_mkdir 39
43+#define __NR_rmdir 40
44+#define __NR_dup 41
45+#define __NR_pipe 42
46+#define __NR_times 43
47+#define __NR_brk 45
48+#define __NR_signal 48
49+#define __NR_acct 51
50+#define __NR_umount2 52
51+#define __NR_ioctl 54
52+#define __NR_fcntl 55
53+#define __NR_setpgid 57
54+#define __NR_umask 60
55+#define __NR_chroot 61
56+#define __NR_ustat 62
57+#define __NR_dup2 63
58+#define __NR_getppid 64
59+#define __NR_getpgrp 65
60+#define __NR_setsid 66
61+#define __NR_sigaction 67
62+#define __NR_sigsuspend 72
63+#define __NR_sigpending 73
64+#define __NR_sethostname 74
65+#define __NR_setrlimit 75
66+#define __NR_getrusage 77
67+#define __NR_gettimeofday 78
68+#define __NR_settimeofday 79
69+#define __NR_symlink 83
70+#define __NR_readlink 85
71+#define __NR_uselib 86
72+#define __NR_swapon 87
73+#define __NR_reboot 88
74+#define __NR_readdir 89
75+#define __NR_mmap 90
76+#define __NR_munmap 91
77+#define __NR_truncate 92
78+#define __NR_ftruncate 93
79+#define __NR_fchmod 94
80+#define __NR_getpriority 96
81+#define __NR_setpriority 97
82+#define __NR_statfs 99
83+#define __NR_fstatfs 100
84+#define __NR_socketcall 102
85+#define __NR_syslog 103
86+#define __NR_setitimer 104
87+#define __NR_getitimer 105
88+#define __NR_stat 106
89+#define __NR_lstat 107
90+#define __NR_fstat 108
91+#define __NR_lookup_dcookie 110
92+#define __NR_vhangup 111
93+#define __NR_idle 112
94+#define __NR_wait4 114
95+#define __NR_swapoff 115
96+#define __NR_sysinfo 116
97+#define __NR_ipc 117
98+#define __NR_fsync 118
99+#define __NR_sigreturn 119
100+#define __NR_clone 120
101+#define __NR_setdomainname 121
102+#define __NR_uname 122
103+#define __NR_adjtimex 124
104+#define __NR_mprotect 125
105+#define __NR_sigprocmask 126
106+#define __NR_create_module 127
107+#define __NR_init_module 128
108+#define __NR_delete_module 129
109+#define __NR_get_kernel_syms 130
110+#define __NR_quotactl 131
111+#define __NR_getpgid 132
112+#define __NR_fchdir 133
113+#define __NR_bdflush 134
114+#define __NR_sysfs 135
115+#define __NR_personality 136
116+#define __NR_afs_syscall 137 /* Syscall for Andrew File System */
117+#define __NR_getdents 141
118+#define __NR_flock 143
119+#define __NR_msync 144
120+#define __NR_readv 145
121+#define __NR_writev 146
122+#define __NR_getsid 147
123+#define __NR_fdatasync 148
124+#define __NR__sysctl 149
125+#define __NR_mlock 150
126+#define __NR_munlock 151
127+#define __NR_mlockall 152
128+#define __NR_munlockall 153
129+#define __NR_sched_setparam 154
130+#define __NR_sched_getparam 155
131+#define __NR_sched_setscheduler 156
132+#define __NR_sched_getscheduler 157
133+#define __NR_sched_yield 158
134+#define __NR_sched_get_priority_max 159
135+#define __NR_sched_get_priority_min 160
136+#define __NR_sched_rr_get_interval 161
137+#define __NR_nanosleep 162
138+#define __NR_mremap 163
139+#define __NR_query_module 167
140+#define __NR_poll 168
141+#define __NR_nfsservctl 169
142+#define __NR_prctl 172
143+#define __NR_rt_sigreturn 173
144+#define __NR_rt_sigaction 174
145+#define __NR_rt_sigprocmask 175
146+#define __NR_rt_sigpending 176
147+#define __NR_rt_sigtimedwait 177
148+#define __NR_rt_sigqueueinfo 178
149+#define __NR_rt_sigsuspend 179
150+#define __NR_pread64 180
151+#define __NR_pwrite64 181
152+#define __NR_getcwd 183
153+#define __NR_capget 184
154+#define __NR_capset 185
155+#define __NR_sigaltstack 186
156+#define __NR_sendfile 187
157+#define __NR_getpmsg 188
158+#define __NR_putpmsg 189
159+#define __NR_vfork 190
160+#define __NR_pivot_root 217
161+#define __NR_mincore 218
162+#define __NR_madvise 219
163+#define __NR_getdents64 220
164+#define __NR_readahead 222
165+#define __NR_setxattr 224
166+#define __NR_lsetxattr 225
167+#define __NR_fsetxattr 226
168+#define __NR_getxattr 227
169+#define __NR_lgetxattr 228
170+#define __NR_fgetxattr 229
171+#define __NR_listxattr 230
172+#define __NR_llistxattr 231
173+#define __NR_flistxattr 232
174+#define __NR_removexattr 233
175+#define __NR_lremovexattr 234
176+#define __NR_fremovexattr 235
177+#define __NR_gettid 236
178+#define __NR_tkill 237
179+#define __NR_futex 238
180+#define __NR_sched_setaffinity 239
181+#define __NR_sched_getaffinity 240
182+#define __NR_tgkill 241
183+/* Number 242 is reserved for tux */
184+#define __NR_io_setup 243
185+#define __NR_io_destroy 244
186+#define __NR_io_getevents 245
187+#define __NR_io_submit 246
188+#define __NR_io_cancel 247
189+#define __NR_exit_group 248
190+#define __NR_epoll_create 249
191+#define __NR_epoll_ctl 250
192+#define __NR_epoll_wait 251
193+#define __NR_set_tid_address 252
194+#define __NR_fadvise64 253
195+#define __NR_timer_create 254
196+#define __NR_timer_settime 255
197+#define __NR_timer_gettime 256
198+#define __NR_timer_getoverrun 257
199+#define __NR_timer_delete 258
200+#define __NR_clock_settime 259
201+#define __NR_clock_gettime 260
202+#define __NR_clock_getres 261
203+#define __NR_clock_nanosleep 262
204+/* Number 263 is reserved for vserver */
205+#define __NR_statfs64 265
206+#define __NR_fstatfs64 266
207+#define __NR_remap_file_pages 267
208+#define __NR_mbind 268
209+#define __NR_get_mempolicy 269
210+#define __NR_set_mempolicy 270
211+#define __NR_mq_open 271
212+#define __NR_mq_unlink 272
213+#define __NR_mq_timedsend 273
214+#define __NR_mq_timedreceive 274
215+#define __NR_mq_notify 275
216+#define __NR_mq_getsetattr 276
217+#define __NR_kexec_load 277
218+#define __NR_add_key 278
219+#define __NR_request_key 279
220+#define __NR_keyctl 280
221+#define __NR_waitid 281
222+#define __NR_ioprio_set 282
223+#define __NR_ioprio_get 283
224+#define __NR_inotify_init 284
225+#define __NR_inotify_add_watch 285
226+#define __NR_inotify_rm_watch 286
227+#define __NR_migrate_pages 287
228+#define __NR_openat 288
229+#define __NR_mkdirat 289
230+#define __NR_mknodat 290
231+#define __NR_fchownat 291
232+#define __NR_futimesat 292
233+#define __NR_unlinkat 294
234+#define __NR_renameat 295
235+#define __NR_linkat 296
236+#define __NR_symlinkat 297
237+#define __NR_readlinkat 298
238+#define __NR_fchmodat 299
239+#define __NR_faccessat 300
240+#define __NR_pselect6 301
241+#define __NR_ppoll 302
242+#define __NR_unshare 303
243+#define __NR_set_robust_list 304
244+#define __NR_get_robust_list 305
245+#define __NR_splice 306
246+#define __NR_sync_file_range 307
247+#define __NR_tee 308
248+#define __NR_vmsplice 309
249+#define __NR_move_pages 310
250+#define __NR_getcpu 311
251+#define __NR_epoll_pwait 312
252+#define __NR_utimes 313
253+#define __NR_fallocate 314
254+#define __NR_utimensat 315
255+#define __NR_signalfd 316
256+#define __NR_timerfd 317
257+#define __NR_eventfd 318
258+#define __NR_timerfd_create 319
259+#define __NR_timerfd_settime 320
260+#define __NR_timerfd_gettime 321
261+#define __NR_signalfd4 322
262+#define __NR_eventfd2 323
263+#define __NR_inotify_init1 324
264+#define __NR_pipe2 325
265+#define __NR_dup3 326
266+#define __NR_epoll_create1 327
267+#define __NR_preadv 328
268+#define __NR_pwritev 329
269+#define __NR_rt_tgsigqueueinfo 330
270+#define __NR_perf_event_open 331
271+#define __NR_fanotify_init 332
272+#define __NR_fanotify_mark 333
273+#define __NR_prlimit64 334
274+#define __NR_name_to_handle_at 335
275+#define __NR_open_by_handle_at 336
276+#define __NR_clock_adjtime 337
277+#define __NR_syncfs 338
278+#define __NR_setns 339
279+#define __NR_process_vm_readv 340
280+#define __NR_process_vm_writev 341
281+#define __NR_s390_runtime_instr 342
282+#define __NR_kcmp 343
283+#define __NR_finit_module 344
284+#define __NR_sched_setattr 345
285+#define __NR_sched_getattr 346
286+#define __NR_renameat2 347
287+#define __NR_seccomp 348
288+#define __NR_getrandom 349
289+#define __NR_memfd_create 350
290+#define __NR_bpf 351
291+#define __NR_s390_pci_mmio_write 352
292+#define __NR_s390_pci_mmio_read 353
293+#define __NR_execveat 354
294+#define __NR_userfaultfd 355
295+#define __NR_membarrier 356
296+#define __NR_recvmmsg 357
297+#define __NR_sendmmsg 358
298+#define __NR_socket 359
299+#define __NR_socketpair 360
300+#define __NR_bind 361
301+#define __NR_connect 362
302+#define __NR_listen 363
303+#define __NR_accept4 364
304+#define __NR_getsockopt 365
305+#define __NR_setsockopt 366
306+#define __NR_getsockname 367
307+#define __NR_getpeername 368
308+#define __NR_sendto 369
309+#define __NR_sendmsg 370
310+#define __NR_recvfrom 371
311+#define __NR_recvmsg 372
312+#define __NR_shutdown 373
313+#define __NR_mlock2 374
314+#define __NR_copy_file_range 375
315+#define __NR_preadv2 376
316+#define __NR_pwritev2 377
317+#define __NR_s390_guarded_storage 378
318+#define __NR_statx 379
319+#define __NR_s390_sthyi 380
320+#define NR_syscalls 381
321+
322+/*
323+ * There are some system calls that are not present on 64 bit, some
324+ * have a different name although they do the same (e.g. __NR_chown32
325+ * is __NR_chown on 64 bit).
326+ */
327+#ifndef __s390x__
328+
329+#define __NR_time 13
330+#define __NR_lchown 16
331+#define __NR_setuid 23
332+#define __NR_getuid 24
333+#define __NR_stime 25
334+#define __NR_setgid 46
335+#define __NR_getgid 47
336+#define __NR_geteuid 49
337+#define __NR_getegid 50
338+#define __NR_setreuid 70
339+#define __NR_setregid 71
340+#define __NR_getrlimit 76
341+#define __NR_getgroups 80
342+#define __NR_setgroups 81
343+#define __NR_fchown 95
344+#define __NR_ioperm 101
345+#define __NR_setfsuid 138
346+#define __NR_setfsgid 139
347+#define __NR__llseek 140
348+#define __NR__newselect 142
349+#define __NR_setresuid 164
350+#define __NR_getresuid 165
351+#define __NR_setresgid 170
352+#define __NR_getresgid 171
353+#define __NR_chown 182
354+#define __NR_ugetrlimit 191 /* SuS compliant getrlimit */
355+#define __NR_mmap2 192
356+#define __NR_truncate64 193
357+#define __NR_ftruncate64 194
358+#define __NR_stat64 195
359+#define __NR_lstat64 196
360+#define __NR_fstat64 197
361+#define __NR_lchown32 198
362+#define __NR_getuid32 199
363+#define __NR_getgid32 200
364+#define __NR_geteuid32 201
365+#define __NR_getegid32 202
366+#define __NR_setreuid32 203
367+#define __NR_setregid32 204
368+#define __NR_getgroups32 205
369+#define __NR_setgroups32 206
370+#define __NR_fchown32 207
371+#define __NR_setresuid32 208
372+#define __NR_getresuid32 209
373+#define __NR_setresgid32 210
374+#define __NR_getresgid32 211
375+#define __NR_chown32 212
376+#define __NR_setuid32 213
377+#define __NR_setgid32 214
378+#define __NR_setfsuid32 215
379+#define __NR_setfsgid32 216
380+#define __NR_fcntl64 221
381+#define __NR_sendfile64 223
382+#define __NR_fadvise64_64 264
383+#define __NR_fstatat64 293
384+
385+#else
386+
387+#define __NR_select 142
388+#define __NR_getrlimit 191 /* SuS compliant getrlimit */
389+#define __NR_lchown 198
390+#define __NR_getuid 199
391+#define __NR_getgid 200
392+#define __NR_geteuid 201
393+#define __NR_getegid 202
394+#define __NR_setreuid 203
395+#define __NR_setregid 204
396+#define __NR_getgroups 205
397+#define __NR_setgroups 206
398+#define __NR_fchown 207
399+#define __NR_setresuid 208
400+#define __NR_getresuid 209
401+#define __NR_setresgid 210
402+#define __NR_getresgid 211
403+#define __NR_chown 212
404+#define __NR_setuid 213
405+#define __NR_setgid 214
406+#define __NR_setfsuid 215
407+#define __NR_setfsgid 216
408+#define __NR_newfstatat 293
409+
410+#endif
411+
412+#endif /* _UAPI_ASM_S390_UNISTD_H_ */
tools/arch/sparc/include/uapi/asm/errno.h+118−0
@@ -0,0 +1,118 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+#ifndef _SPARC_ERRNO_H
3+#define _SPARC_ERRNO_H
4+
5+/* These match the SunOS error numbering scheme. */
6+
7+#include <asm-generic/errno-base.h>
8+
9+#define EWOULDBLOCK EAGAIN /* Operation would block */
10+#define EINPROGRESS 36 /* Operation now in progress */
11+#define EALREADY 37 /* Operation already in progress */
12+#define ENOTSOCK 38 /* Socket operation on non-socket */
13+#define EDESTADDRREQ 39 /* Destination address required */
14+#define EMSGSIZE 40 /* Message too long */
15+#define EPROTOTYPE 41 /* Protocol wrong type for socket */
16+#define ENOPROTOOPT 42 /* Protocol not available */
17+#define EPROTONOSUPPORT 43 /* Protocol not supported */
18+#define ESOCKTNOSUPPORT 44 /* Socket type not supported */
19+#define EOPNOTSUPP 45 /* Op not supported on transport endpoint */
20+#define EPFNOSUPPORT 46 /* Protocol family not supported */
21+#define EAFNOSUPPORT 47 /* Address family not supported by protocol */
22+#define EADDRINUSE 48 /* Address already in use */
23+#define EADDRNOTAVAIL 49 /* Cannot assign requested address */
24+#define ENETDOWN 50 /* Network is down */
25+#define ENETUNREACH 51 /* Network is unreachable */
26+#define ENETRESET 52 /* Net dropped connection because of reset */
27+#define ECONNABORTED 53 /* Software caused connection abort */
28+#define ECONNRESET 54 /* Connection reset by peer */
29+#define ENOBUFS 55 /* No buffer space available */
30+#define EISCONN 56 /* Transport endpoint is already connected */
31+#define ENOTCONN 57 /* Transport endpoint is not connected */
32+#define ESHUTDOWN 58 /* No send after transport endpoint shutdown */
33+#define ETOOMANYREFS 59 /* Too many references: cannot splice */
34+#define ETIMEDOUT 60 /* Connection timed out */
35+#define ECONNREFUSED 61 /* Connection refused */
36+#define ELOOP 62 /* Too many symbolic links encountered */
37+#define ENAMETOOLONG 63 /* File name too long */
38+#define EHOSTDOWN 64 /* Host is down */
39+#define EHOSTUNREACH 65 /* No route to host */
40+#define ENOTEMPTY 66 /* Directory not empty */
41+#define EPROCLIM 67 /* SUNOS: Too many processes */
42+#define EUSERS 68 /* Too many users */
43+#define EDQUOT 69 /* Quota exceeded */
44+#define ESTALE 70 /* Stale file handle */
45+#define EREMOTE 71 /* Object is remote */
46+#define ENOSTR 72 /* Device not a stream */
47+#define ETIME 73 /* Timer expired */
48+#define ENOSR 74 /* Out of streams resources */
49+#define ENOMSG 75 /* No message of desired type */
50+#define EBADMSG 76 /* Not a data message */
51+#define EIDRM 77 /* Identifier removed */
52+#define EDEADLK 78 /* Resource deadlock would occur */
53+#define ENOLCK 79 /* No record locks available */
54+#define ENONET 80 /* Machine is not on the network */
55+#define ERREMOTE 81 /* SunOS: Too many lvls of remote in path */
56+#define ENOLINK 82 /* Link has been severed */
57+#define EADV 83 /* Advertise error */
58+#define ESRMNT 84 /* Srmount error */
59+#define ECOMM 85 /* Communication error on send */
60+#define EPROTO 86 /* Protocol error */
61+#define EMULTIHOP 87 /* Multihop attempted */
62+#define EDOTDOT 88 /* RFS specific error */
63+#define EREMCHG 89 /* Remote address changed */
64+#define ENOSYS 90 /* Function not implemented */
65+
66+/* The rest have no SunOS equivalent. */
67+#define ESTRPIPE 91 /* Streams pipe error */
68+#define EOVERFLOW 92 /* Value too large for defined data type */
69+#define EBADFD 93 /* File descriptor in bad state */
70+#define ECHRNG 94 /* Channel number out of range */
71+#define EL2NSYNC 95 /* Level 2 not synchronized */
72+#define EL3HLT 96 /* Level 3 halted */
73+#define EL3RST 97 /* Level 3 reset */
74+#define ELNRNG 98 /* Link number out of range */
75+#define EUNATCH 99 /* Protocol driver not attached */
76+#define ENOCSI 100 /* No CSI structure available */
77+#define EL2HLT 101 /* Level 2 halted */
78+#define EBADE 102 /* Invalid exchange */
79+#define EBADR 103 /* Invalid request descriptor */
80+#define EXFULL 104 /* Exchange full */
81+#define ENOANO 105 /* No anode */
82+#define EBADRQC 106 /* Invalid request code */
83+#define EBADSLT 107 /* Invalid slot */
84+#define EDEADLOCK 108 /* File locking deadlock error */
85+#define EBFONT 109 /* Bad font file format */
86+#define ELIBEXEC 110 /* Cannot exec a shared library directly */
87+#define ENODATA 111 /* No data available */
88+#define ELIBBAD 112 /* Accessing a corrupted shared library */
89+#define ENOPKG 113 /* Package not installed */
90+#define ELIBACC 114 /* Can not access a needed shared library */
91+#define ENOTUNIQ 115 /* Name not unique on network */
92+#define ERESTART 116 /* Interrupted syscall should be restarted */
93+#define EUCLEAN 117 /* Structure needs cleaning */
94+#define ENOTNAM 118 /* Not a XENIX named type file */
95+#define ENAVAIL 119 /* No XENIX semaphores available */
96+#define EISNAM 120 /* Is a named type file */
97+#define EREMOTEIO 121 /* Remote I/O error */
98+#define EILSEQ 122 /* Illegal byte sequence */
99+#define ELIBMAX 123 /* Atmpt to link in too many shared libs */
100+#define ELIBSCN 124 /* .lib section in a.out corrupted */
101+
102+#define ENOMEDIUM 125 /* No medium found */
103+#define EMEDIUMTYPE 126 /* Wrong medium type */
104+#define ECANCELED 127 /* Operation Cancelled */
105+#define ENOKEY 128 /* Required key not available */
106+#define EKEYEXPIRED 129 /* Key has expired */
107+#define EKEYREVOKED 130 /* Key has been revoked */
108+#define EKEYREJECTED 131 /* Key was rejected by service */
109+
110+/* for robust mutexes */
111+#define EOWNERDEAD 132 /* Owner died */
112+#define ENOTRECOVERABLE 133 /* State not recoverable */
113+
114+#define ERFKILL 134 /* Operation not possible due to RF-kill */
115+
116+#define EHWPOISON 135 /* Memory page has hardware error */
117+
118+#endif
tools/arch/x86/include/asm/cpufeatures.h+3−1
@@ -197,11 +197,12 @@ · 196 unmodified lines
197197 #define X86_FEATURE_CAT_L3 ( 7*32+ 4) /* Cache Allocation Technology L3 */
198198 #define X86_FEATURE_CAT_L2 ( 7*32+ 5) /* Cache Allocation Technology L2 */
199199 #define X86_FEATURE_CDP_L3 ( 7*32+ 6) /* Code and Data Prioritization L3 */
200+#define X86_FEATURE_INVPCID_SINGLE ( 7*32+ 7) /* Effectively INVPCID && CR4.PCIDE=1 */
200201
201202 #define X86_FEATURE_HW_PSTATE ( 7*32+ 8) /* AMD HW-PState */
202203 #define X86_FEATURE_PROC_FEEDBACK ( 7*32+ 9) /* AMD ProcFeedbackInterface */
203204 #define X86_FEATURE_SME ( 7*32+10) /* AMD Secure Memory Encryption */
204−
205+#define X86_FEATURE_PTI ( 7*32+11) /* Kernel Page Table Isolation enabled */
205206 #define X86_FEATURE_INTEL_PPIN ( 7*32+14) /* Intel Processor Inventory Number */
206207 #define X86_FEATURE_INTEL_PT ( 7*32+15) /* Intel Processor Trace */
207208 #define X86_FEATURE_AVX512_4VNNIW ( 7*32+16) /* AVX-512 Neural Network Instructions */
@@ -340,5 +341,6 @@ · 132 unmodified lines
340341 #define X86_BUG_SWAPGS_FENCE X86_BUG(11) /* SWAPGS without input dep on GS */
341342 #define X86_BUG_MONITOR X86_BUG(12) /* IPI required to wake up remote CPU */
342343 #define X86_BUG_AMD_E400 X86_BUG(13) /* CPU is among the affected by Erratum 400 */
344+#define X86_BUG_CPU_MELTDOWN X86_BUG(14) /* CPU is affected by meltdown attack and needs kernel page table isolation */
343345
344346 #endif /* _ASM_X86_CPUFEATURES_H */
tools/arch/x86/include/asm/disabled-features.h+7−1
@@ -50,6 +50,12 @@ · 49 unmodified lines
5050 # define DISABLE_LA57 (1<<(X86_FEATURE_LA57 & 31))
5151 #endif
5252
53+#ifdef CONFIG_PAGE_TABLE_ISOLATION
54+# define DISABLE_PTI 0
55+#else
56+# define DISABLE_PTI (1 << (X86_FEATURE_PTI & 31))
57+#endif
58+
5359 /*
5460 * Make sure to add features to the correct mask
5561 */
@@ -60,7 +66,7 @@ · 4 unmodified lines
6066 #define DISABLED_MASK4 (DISABLE_PCID)
6167 #define DISABLED_MASK5 0
6268 #define DISABLED_MASK6 0
63−#define DISABLED_MASK7 0
69+#define DISABLED_MASK7 (DISABLE_PTI)
6470 #define DISABLED_MASK8 0
6571 #define DISABLED_MASK9 (DISABLE_MPX)
6672 #define DISABLED_MASK10 0
tools/arch/x86/include/uapi/asm/errno.h+1−0
@@ -0,0 +1 @@
1+#include <asm-generic/errno.h>
tools/build/Makefile.feature+3−1
@@ -56,6 +56,7 @@ FEATURE_TESTS_BASIC := \ · 55 unmodified lines
5656 libunwind-arm \
5757 libunwind-aarch64 \
5858 pthread-attr-setaffinity-np \
59+ pthread-barrier \
5960 stackprotector-all \
6061 timerfd \
6162 libdw-dwarf-unwind \
@@ -65,7 +66,8 @@ FEATURE_TESTS_BASIC := \ · 3 unmodified lines
6566 bpf \
6667 sched_getcpu \
6768 sdt \
68− setns
69+ setns \
70+ libopencsd
6971
7072 # FEATURE_TESTS_BASIC + FEATURE_TESTS_EXTRA is the complete list
7173 # of all feature tests
tools/build/feature/Makefile+10−1
@@ -37,6 +37,7 @@ FILES= \ · 36 unmodified lines
3737 test-libunwind-debug-frame-arm.bin \
3838 test-libunwind-debug-frame-aarch64.bin \
3939 test-pthread-attr-setaffinity-np.bin \
40+ test-pthread-barrier.bin \
4041 test-stackprotector-all.bin \
4142 test-timerfd.bin \
4243 test-libdw-dwarf-unwind.bin \
@@ -51,7 +52,8 @@ FILES= \ · 8 unmodified lines
5152 test-cxx.bin \
5253 test-jvmti.bin \
5354 test-sched_getcpu.bin \
54− test-setns.bin
55+ test-setns.bin \
56+ test-libopencsd.bin
5557
5658 FILES := $(addprefix $(OUTPUT),$(FILES))
5759
@@ -79,6 +81,9 @@ $(OUTPUT)test-hello.bin: · 21 unmodified lines
7981 $(OUTPUT)test-pthread-attr-setaffinity-np.bin:
8082 $(BUILD) -D_GNU_SOURCE -lpthread
8183
84+$(OUTPUT)test-pthread-barrier.bin:
85+ $(BUILD) -lpthread
86+
8287 $(OUTPUT)test-stackprotector-all.bin:
8388 $(BUILD) -fstack-protector-all
8489
@@ -100,6 +105,10 @@ $(OUTPUT)test-sched_getcpu.bin: · 15 unmodified lines
100105 $(OUTPUT)test-setns.bin:
101106 $(BUILD)
102107
108+$(OUTPUT)test-libopencsd.bin:
109+ $(BUILD) # -lopencsd_c_api -lopencsd provided by
110+ # $(FEATURE_CHECK_LDFLAGS-libopencsd)
111+
103112 DWARFLIBS := -ldw
104113 ifeq ($(findstring -static,${LDFLAGS}),-static)
105114 DWARFLIBS += -lelf -lebl -lz -llzma -lbz2
tools/build/feature/test-all.c+10−0
@@ -118,6 +118,10 @@ · 117 unmodified lines
118118 # include "test-pthread-attr-setaffinity-np.c"
119119 #undef main
120120
121+#define main main_test_pthread_barrier
122+# include "test-pthread-barrier.c"
123+#undef main
124+
121125 #define main main_test_sched_getcpu
122126 # include "test-sched_getcpu.c"
123127 #undef main
@@ -158,6 +162,10 @@ · 34 unmodified lines
158162 # include "test-setns.c"
159163 #undef main
160164
165+#define main main_test_libopencsd
166+# include "test-libopencsd.c"
167+#undef main
168+
161169 int main(int argc, char *argv[])
162170 {
163171 main_test_libpython();
@@ -187,6 +195,7 @@ int main(int argc, char *argv[]) · 23 unmodified lines
187195 main_test_sync_compare_and_swap(argc, argv);
188196 main_test_zlib();
189197 main_test_pthread_attr_setaffinity_np();
198+ main_test_pthread_barrier();
190199 main_test_lzma();
191200 main_test_get_cpuid();
192201 main_test_bpf();
@@ -194,6 +203,7 @@ int main(int argc, char *argv[]) · 1 unmodified line
194203 main_test_sched_getcpu();
195204 main_test_sdt();
196205 main_test_setns();
206+ main_test_libopencsd();
197207
198208 return 0;
199209 }
tools/build/feature/test-libopencsd.c+8−0
@@ -0,0 +1,8 @@
1+// SPDX-License-Identifier: GPL-2.0
2+#include <opencsd/c_api/opencsd_c_api.h>
3+
4+int main(void)
5+{
6+ (void)ocsd_get_version();
7+ return 0;
8+}
tools/build/feature/test-pthread-barrier.c+12−0
@@ -0,0 +1,12 @@
1+// SPDX-License-Identifier: GPL-2.0
2+#include <stdint.h>
3+#include <pthread.h>
4+
5+int main(void)
6+{
7+ pthread_barrier_t barrier;
8+
9+ pthread_barrier_init(&barrier, NULL, 1);
10+ pthread_barrier_wait(&barrier);
11+ return pthread_barrier_destroy(&barrier);
12+}
tools/include/uapi/asm-generic/errno-base.h+40−0
@@ -0,0 +1,40 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+#ifndef _ASM_GENERIC_ERRNO_BASE_H
3+#define _ASM_GENERIC_ERRNO_BASE_H
4+
5+#define EPERM 1 /* Operation not permitted */
6+#define ENOENT 2 /* No such file or directory */
7+#define ESRCH 3 /* No such process */
8+#define EINTR 4 /* Interrupted system call */
9+#define EIO 5 /* I/O error */
10+#define ENXIO 6 /* No such device or address */
11+#define E2BIG 7 /* Argument list too long */
12+#define ENOEXEC 8 /* Exec format error */
13+#define EBADF 9 /* Bad file number */
14+#define ECHILD 10 /* No child processes */
15+#define EAGAIN 11 /* Try again */
16+#define ENOMEM 12 /* Out of memory */
17+#define EACCES 13 /* Permission denied */
18+#define EFAULT 14 /* Bad address */
19+#define ENOTBLK 15 /* Block device required */
20+#define EBUSY 16 /* Device or resource busy */
21+#define EEXIST 17 /* File exists */
22+#define EXDEV 18 /* Cross-device link */
23+#define ENODEV 19 /* No such device */
24+#define ENOTDIR 20 /* Not a directory */
25+#define EISDIR 21 /* Is a directory */
26+#define EINVAL 22 /* Invalid argument */
27+#define ENFILE 23 /* File table overflow */
28+#define EMFILE 24 /* Too many open files */
29+#define ENOTTY 25 /* Not a typewriter */
30+#define ETXTBSY 26 /* Text file busy */
31+#define EFBIG 27 /* File too large */
32+#define ENOSPC 28 /* No space left on device */
33+#define ESPIPE 29 /* Illegal seek */
34+#define EROFS 30 /* Read-only file system */
35+#define EMLINK 31 /* Too many links */
36+#define EPIPE 32 /* Broken pipe */
37+#define EDOM 33 /* Math argument out of domain of func */
38+#define ERANGE 34 /* Math result not representable */
39+
40+#endif
tools/include/uapi/asm-generic/errno.h+123−0
@@ -0,0 +1,123 @@
1+/* SPDX-License-Identifier: GPL-2.0 WITH Linux-syscall-note */
2+#ifndef _ASM_GENERIC_ERRNO_H
3+#define _ASM_GENERIC_ERRNO_H
4+
5+#include <asm-generic/errno-base.h>
6+
7+#define EDEADLK 35 /* Resource deadlock would occur */
8+#define ENAMETOOLONG 36 /* File name too long */
9+#define ENOLCK 37 /* No record locks available */
10+
11+/*
12+ * This error code is special: arch syscall entry code will return
13+ * -ENOSYS if users try to call a syscall that doesn't exist. To keep
14+ * failures of syscalls that really do exist distinguishable from
15+ * failures due to attempts to use a nonexistent syscall, syscall
16+ * implementations should refrain from returning -ENOSYS.
17+ */
18+#define ENOSYS 38 /* Invalid system call number */
19+
20+#define ENOTEMPTY 39 /* Directory not empty */
21+#define ELOOP 40 /* Too many symbolic links encountered */
22+#define EWOULDBLOCK EAGAIN /* Operation would block */
23+#define ENOMSG 42 /* No message of desired type */
24+#define EIDRM 43 /* Identifier removed */
25+#define ECHRNG 44 /* Channel number out of range */
26+#define EL2NSYNC 45 /* Level 2 not synchronized */
27+#define EL3HLT 46 /* Level 3 halted */
28+#define EL3RST 47 /* Level 3 reset */
29+#define ELNRNG 48 /* Link number out of range */
30+#define EUNATCH 49 /* Protocol driver not attached */
31+#define ENOCSI 50 /* No CSI structure available */
32+#define EL2HLT 51 /* Level 2 halted */
33+#define EBADE 52 /* Invalid exchange */
34+#define EBADR 53 /* Invalid request descriptor */
35+#define EXFULL 54 /* Exchange full */
36+#define ENOANO 55 /* No anode */
37+#define EBADRQC 56 /* Invalid request code */
38+#define EBADSLT 57 /* Invalid slot */
39+
40+#define EDEADLOCK EDEADLK
41+
42+#define EBFONT 59 /* Bad font file format */
43+#define ENOSTR 60 /* Device not a stream */
44+#define ENODATA 61 /* No data available */
45+#define ETIME 62 /* Timer expired */
46+#define ENOSR 63 /* Out of streams resources */
47+#define ENONET 64 /* Machine is not on the network */
48+#define ENOPKG 65 /* Package not installed */
49+#define EREMOTE 66 /* Object is remote */
50+#define ENOLINK 67 /* Link has been severed */
51+#define EADV 68 /* Advertise error */
52+#define ESRMNT 69 /* Srmount error */
53+#define ECOMM 70 /* Communication error on send */
54+#define EPROTO 71 /* Protocol error */
55+#define EMULTIHOP 72 /* Multihop attempted */
56+#define EDOTDOT 73 /* RFS specific error */
57+#define EBADMSG 74 /* Not a data message */
58+#define EOVERFLOW 75 /* Value too large for defined data type */
59+#define ENOTUNIQ 76 /* Name not unique on network */
60+#define EBADFD 77 /* File descriptor in bad state */
61+#define EREMCHG 78 /* Remote address changed */
62+#define ELIBACC 79 /* Can not access a needed shared library */
63+#define ELIBBAD 80 /* Accessing a corrupted shared library */
64+#define ELIBSCN 81 /* .lib section in a.out corrupted */
65+#define ELIBMAX 82 /* Attempting to link in too many shared libraries */
66+#define ELIBEXEC 83 /* Cannot exec a shared library directly */
67+#define EILSEQ 84 /* Illegal byte sequence */
68+#define ERESTART 85 /* Interrupted system call should be restarted */
69+#define ESTRPIPE 86 /* Streams pipe error */
70+#define EUSERS 87 /* Too many users */
71+#define ENOTSOCK 88 /* Socket operation on non-socket */
72+#define EDESTADDRREQ 89 /* Destination address required */
73+#define EMSGSIZE 90 /* Message too long */
74+#define EPROTOTYPE 91 /* Protocol wrong type for socket */
75+#define ENOPROTOOPT 92 /* Protocol not available */
76+#define EPROTONOSUPPORT 93 /* Protocol not supported */
77+#define ESOCKTNOSUPPORT 94 /* Socket type not supported */
78+#define EOPNOTSUPP 95 /* Operation not supported on transport endpoint */
79+#define EPFNOSUPPORT 96 /* Protocol family not supported */
80+#define EAFNOSUPPORT 97 /* Address family not supported by protocol */
81+#define EADDRINUSE 98 /* Address already in use */
82+#define EADDRNOTAVAIL 99 /* Cannot assign requested address */
83+#define ENETDOWN 100 /* Network is down */
84+#define ENETUNREACH 101 /* Network is unreachable */
85+#define ENETRESET 102 /* Network dropped connection because of reset */
86+#define ECONNABORTED 103 /* Software caused connection abort */
87+#define ECONNRESET 104 /* Connection reset by peer */
88+#define ENOBUFS 105 /* No buffer space available */
89+#define EISCONN 106 /* Transport endpoint is already connected */
90+#define ENOTCONN 107 /* Transport endpoint is not connected */
91+#define ESHUTDOWN 108 /* Cannot send after transport endpoint shutdown */
92+#define ETOOMANYREFS 109 /* Too many references: cannot splice */
93+#define ETIMEDOUT 110 /* Connection timed out */
94+#define ECONNREFUSED 111 /* Connection refused */
95+#define EHOSTDOWN 112 /* Host is down */
96+#define EHOSTUNREACH 113 /* No route to host */
97+#define EALREADY 114 /* Operation already in progress */
98+#define EINPROGRESS 115 /* Operation now in progress */
99+#define ESTALE 116 /* Stale file handle */
100+#define EUCLEAN 117 /* Structure needs cleaning */
101+#define ENOTNAM 118 /* Not a XENIX named type file */
102+#define ENAVAIL 119 /* No XENIX semaphores available */
103+#define EISNAM 120 /* Is a named type file */
104+#define EREMOTEIO 121 /* Remote I/O error */
105+#define EDQUOT 122 /* Quota exceeded */
106+
107+#define ENOMEDIUM 123 /* No medium found */
108+#define EMEDIUMTYPE 124 /* Wrong medium type */
109+#define ECANCELED 125 /* Operation Canceled */
110+#define ENOKEY 126 /* Required key not available */
111+#define EKEYEXPIRED 127 /* Key has expired */
112+#define EKEYREVOKED 128 /* Key has been revoked */
113+#define EKEYREJECTED 129 /* Key was rejected by service */
114+
115+/* for robust mutexes */
116+#define EOWNERDEAD 130 /* Owner died */
117+#define ENOTRECOVERABLE 131 /* State not recoverable */
118+
119+#define ERFKILL 132 /* Operation not possible due to RF-kill */
120+
121+#define EHWPOISON 133 /* Memory page has hardware error */
122+
123+#endif
tools/include/uapi/linux/perf_event.h+7−3
@@ -612,9 +612,12 @@ struct perf_event_mmap_page { · 611 unmodified lines
612612 */
613613 #define PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT (1 << 12)
614614 /*
615− * PERF_RECORD_MISC_MMAP_DATA and PERF_RECORD_MISC_COMM_EXEC are used on
616− * different events so can reuse the same bit position.
617− * Ditto PERF_RECORD_MISC_SWITCH_OUT.
615+ * Following PERF_RECORD_MISC_* are used on different
616+ * events, so can reuse the same bit position:
617+ *
618+ * PERF_RECORD_MISC_MMAP_DATA - PERF_RECORD_MMAP* events
619+ * PERF_RECORD_MISC_COMM_EXEC - PERF_RECORD_COMM event
620+ * PERF_RECORD_MISC_SWITCH_OUT - PERF_RECORD_SWITCH* events
618621 */
619622 #define PERF_RECORD_MISC_MMAP_DATA (1 << 13)
620623 #define PERF_RECORD_MISC_COMM_EXEC (1 << 13)
@@ -864,6 +867,7 @@ enum perf_event_type { · 243 unmodified lines
864867 * struct perf_event_header header;
865868 * u32 pid;
866869 * u32 tid;
870+ * struct sample_id sample_id;
867871 * };
868872 */
869873 PERF_RECORD_ITRACE_START = 12,
tools/lib/traceevent/event-parse.c+51−11
@@ -1094,7 +1094,7 @@ static enum event_type __read_token(char **tok) · 1093 unmodified lines
10941094 if (strcmp(*tok, "LOCAL_PR_FMT") == 0) {
10951095 free(*tok);
10961096 *tok = NULL;
1097− return force_token("\"\%s\" ", tok);
1097+ return force_token("\"%s\" ", tok);
10981098 } else if (strcmp(*tok, "STA_PR_FMT") == 0) {
10991099 free(*tok);
11001100 *tok = NULL;
@@ -3970,6 +3970,11 @@ static void print_str_arg(struct trace_seq *s, void *data, int size, · 2869 unmodified lines
39703970 val &= ~fval;
39713971 }
39723972 }
3973+ if (val) {
3974+ if (print && arg->flags.delim)
3975+ trace_seq_puts(s, arg->flags.delim);
3976+ trace_seq_printf(s, "0x%llx", val);
3977+ }
39733978 break;
39743979 case PRINT_SYMBOL:
39753980 val = eval_num_arg(data, size, event, arg->symbol.field);
@@ -3980,6 +3985,8 @@ static void print_str_arg(struct trace_seq *s, void *data, int size, · 4 unmodified lines
39803985 break;
39813986 }
39823987 }
3988+ if (!flag)
3989+ trace_seq_printf(s, "0x%llx", val);
39833990 break;
39843991 case PRINT_HEX:
39853992 case PRINT_HEX_STR:
@@ -4293,6 +4300,26 @@ static struct print_arg *make_bprint_args(char *fmt, void *data, int size, struc · 307 unmodified lines
42934300 goto process_again;
42944301 case 'p':
42954302 ls = 1;
4303+ if (isalnum(ptr[1])) {
4304+ ptr++;
4305+ /* Check for special pointers */
4306+ switch (*ptr) {
4307+ case 's':
4308+ case 'S':
4309+ case 'f':
4310+ case 'F':
4311+ break;
4312+ default:
4313+ /*
4314+ * Older kernels do not process
4315+ * dereferenced pointers.
4316+ * Only process if the pointer
4317+ * value is a printable.
4318+ */
4319+ if (isprint(*(char *)bptr))
4320+ goto process_string;
4321+ }
4322+ }
42964323 /* fall through */
42974324 case 'd':
42984325 case 'u':
@@ -4345,6 +4372,7 @@ static struct print_arg *make_bprint_args(char *fmt, void *data, int size, struc · 46 unmodified lines
43454372
43464373 break;
43474374 case 's':
4375+ process_string:
43484376 arg = alloc_arg();
43494377 if (!arg) {
43504378 do_warning_event(event, "%s(%d): not enough memory!",
@@ -4949,21 +4977,27 @@ static void pretty_print(struct trace_seq *s, void *data, int size, struct event · 598 unmodified lines
49494977 else
49504978 ls = 2;
49514979
4952− if (*(ptr+1) == 'F' || *(ptr+1) == 'f' ||
4953− *(ptr+1) == 'S' || *(ptr+1) == 's') {
4980+ if (isalnum(ptr[1]))
49544981 ptr++;
4982+
4983+ if (arg->type == PRINT_BSTRING) {
4984+ trace_seq_puts(s, arg->string.string);
4985+ break;
4986+ }
4987+
4988+ if (*ptr == 'F' || *ptr == 'f' ||
4989+ *ptr == 'S' || *ptr == 's') {
49554990 show_func = *ptr;
4956− } else if (*(ptr+1) == 'M' || *(ptr+1) == 'm') {
4957− print_mac_arg(s, *(ptr+1), data, size, event, arg);
4958− ptr++;
4991+ } else if (*ptr == 'M' || *ptr == 'm') {
4992+ print_mac_arg(s, *ptr, data, size, event, arg);
49594993 arg = arg->next;
49604994 break;
4961− } else if (*(ptr+1) == 'I' || *(ptr+1) == 'i') {
4995+ } else if (*ptr == 'I' || *ptr == 'i') {
49624996 int n;
49634997
4964− n = print_ip_arg(s, ptr+1, data, size, event, arg);
4998+ n = print_ip_arg(s, ptr, data, size, event, arg);
49654999 if (n > 0) {
4966− ptr += n;
5000+ ptr += n - 1;
49675001 arg = arg->next;
49685002 break;
49695003 }
@@ -5532,8 +5566,14 @@ void pevent_print_event(struct pevent *pevent, struct trace_seq *s, · 562 unmodified lines
55325566
55335567 event = pevent_find_event_by_record(pevent, record);
55345568 if (!event) {
5535− do_warning("ug! no event found for type %d",
5536− trace_parse_common_type(pevent, record->data));
5569+ int i;
5570+ int type = trace_parse_common_type(pevent, record->data);
5571+
5572+ do_warning("ug! no event found for type %d", type);
5573+ trace_seq_printf(s, "[UNKNOWN TYPE %d]", type);
5574+ for (i = 0; i < record->size; i++)
5575+ trace_seq_printf(s, " %02x",
5576+ ((unsigned char *)record->data)[i]);
55375577 return;
55385578 }
55395579
tools/lib/traceevent/event-plugin.c+9−15
@@ -120,12 +120,12 @@ char **traceevent_plugin_list_options(void) · 119 unmodified lines
120120 for (op = reg->options; op->name; op++) {
121121 char *alias = op->plugin_alias ? op->plugin_alias : op->file;
122122 char **temp = list;
123+ int ret;
123124
124− name = malloc(strlen(op->name) + strlen(alias) + 2);
125− if (!name)
125+ ret = asprintf(&name, "%s:%s", alias, op->name);
126+ if (ret < 0)
126127 goto err;
127128
128− sprintf(name, "%s:%s", alias, op->name);
129129 list = realloc(list, count + 2);
130130 if (!list) {
131131 list = temp;
@@ -290,17 +290,14 @@ load_plugin(struct pevent *pevent, const char *path, · 158 unmodified lines
290290 const char *alias;
291291 char *plugin;
292292 void *handle;
293+ int ret;
293294
294− plugin = malloc(strlen(path) + strlen(file) + 2);
295− if (!plugin) {
295+ ret = asprintf(&plugin, "%s/%s", path, file);
296+ if (ret < 0) {
296297 warning("could not allocate plugin memory\n");
297298 return;
298299 }
299300
300− strcpy(plugin, path);
301− strcat(plugin, "/");
302− strcat(plugin, file);
303−
304301 handle = dlopen(plugin, RTLD_NOW | RTLD_GLOBAL);
305302 if (!handle) {
306303 warning("could not load plugin '%s'\n%s\n",
@@ -391,6 +388,7 @@ load_plugins(struct pevent *pevent, const char *suffix, · 84 unmodified lines
391388 char *home;
392389 char *path;
393390 char *envdir;
391+ int ret;
394392
395393 if (pevent->flags & PEVENT_DISABLE_PLUGINS)
396394 return;
@@ -421,16 +419,12 @@ load_plugins(struct pevent *pevent, const char *suffix, · 24 unmodified lines
421419 if (!home)
422420 return;
423421
424− path = malloc(strlen(home) + strlen(LOCAL_PLUGIN_DIR) + 2);
425− if (!path) {
422+ ret = asprintf(&path, "%s/%s", home, LOCAL_PLUGIN_DIR);
423+ if (ret < 0) {
426424 warning("could not allocate plugin memory\n");
427425 return;
428426 }
429427
430− strcpy(path, home);
431− strcat(path, "/");
432− strcat(path, LOCAL_PLUGIN_DIR);
433−
434428 load_plugins_dir(pevent, suffix, path, load_plugin, data);
435429
436430 free(path);
tools/lib/traceevent/kbuffer-parse.c+2−2
@@ -24,8 +24,8 @@ · 23 unmodified lines
2424
2525 #include "kbuffer.h"
2626
27−#define MISSING_EVENTS (1 << 31)
28−#define MISSING_STORED (1 << 30)
27+#define MISSING_EVENTS (1UL << 31)
28+#define MISSING_STORED (1UL << 30)
2929
3030 #define COMMIT_MASK ((1 << 27) - 1)
3131
tools/lib/traceevent/parse-filter.c+14−8
@@ -287,12 +287,10 @@ find_event(struct pevent *pevent, struct event_list **events, · 286 unmodified lines
287287 sys_name = NULL;
288288 }
289289
290− reg = malloc(strlen(event_name) + 3);
291− if (reg == NULL)
290+ ret = asprintf(&reg, "^%s$", event_name);
291+ if (ret < 0)
292292 return PEVENT_ERRNO__MEM_ALLOC_FAILED;
293293
294− sprintf(reg, "^%s$", event_name);
295−
296294 ret = regcomp(&ereg, reg, REG_ICASE|REG_NOSUB);
297295 free(reg);
298296
@@ -300,13 +298,12 @@ find_event(struct pevent *pevent, struct event_list **events, · 1 unmodified line
300298 return PEVENT_ERRNO__INVALID_EVENT_NAME;
301299
302300 if (sys_name) {
303− reg = malloc(strlen(sys_name) + 3);
304− if (reg == NULL) {
301+ ret = asprintf(&reg, "^%s$", sys_name);
302+ if (ret < 0) {
305303 regfree(&ereg);
306304 return PEVENT_ERRNO__MEM_ALLOC_FAILED;
307305 }
308306
309− sprintf(reg, "^%s$", sys_name);
310307 ret = regcomp(&sreg, reg, REG_ICASE|REG_NOSUB);
311308 free(reg);
312309 if (ret) {
@@ -1634,6 +1631,7 @@ int pevent_filter_clear_trivial(struct event_filter *filter, · 1321 unmodified lines
16341631 case FILTER_TRIVIAL_FALSE:
16351632 if (filter_type->filter->boolean.value)
16361633 continue;
1634+ break;
16371635 case FILTER_TRIVIAL_TRUE:
16381636 if (!filter_type->filter->boolean.value)
16391637 continue;
@@ -1879,17 +1877,25 @@ static const char *get_field_str(struct filter_arg *arg, struct pevent_record *r · 239 unmodified lines
18791877 struct pevent *pevent;
18801878 unsigned long long addr;
18811879 const char *val = NULL;
1880+ unsigned int size;
18821881 char hex[64];
18831882
18841883 /* If the field is not a string convert it */
18851884 if (arg->str.field->flags & FIELD_IS_STRING) {
18861885 val = record->data + arg->str.field->offset;
1886+ size = arg->str.field->size;
1887+
1888+ if (arg->str.field->flags & FIELD_IS_DYNAMIC) {
1889+ addr = *(unsigned int *)val;
1890+ val = record->data + (addr & 0xffff);
1891+ size = addr >> 16;
1892+ }
18871893
18881894 /*
18891895 * We need to copy the data since we can't be sure the field
18901896 * is null terminated.
18911897 */
1892− if (*(val + arg->str.field->size - 1)) {
1898+ if (*(val + size - 1)) {
18931899 /* copy it */
18941900 memcpy(arg->str.buffer, val, arg->str.field->size);
18951901 /* the buffer is already NULL terminated */
tools/perf/Build+2−2
@@ -25,7 +25,7 @@ perf-y += builtin-data.o · 24 unmodified lines
2525 perf-y += builtin-version.o
2626 perf-y += builtin-c2c.o
2727
28−perf-$(CONFIG_AUDIT) += builtin-trace.o
28+perf-$(CONFIG_TRACE) += builtin-trace.o
2929 perf-$(CONFIG_LIBELF) += builtin-probe.o
3030
3131 perf-y += bench/
@@ -50,6 +50,6 @@ libperf-y += util/ · 18 unmodified lines
5050 libperf-y += arch/
5151 libperf-y += ui/
5252 libperf-y += scripts/
53−libperf-$(CONFIG_AUDIT) += trace/beauty/
53+libperf-$(CONFIG_TRACE) += trace/beauty/
5454
5555 gtk-y += ui/gtk/
tools/perf/Documentation/perf-buildid-cache.txt+3−0
@@ -24,6 +24,9 @@ OPTIONS · 23 unmodified lines
2424 -a::
2525 --add=::
2626 Add specified file to the cache.
27+-f::
28+--force::
29+ Don't complain, do it.
2730 -k::
2831 --kcore::
2932 Add specified kcore file to the cache. For the current host that is
tools/perf/Documentation/perf-evlist.txt+4−0
@@ -20,6 +20,10 @@ OPTIONS · 19 unmodified lines
2020 --input=::
2121 Input file name. (default: perf.data unless stdin is a fifo)
2222
23+-f::
24+--force::
25+ Don't complain, do it.
26+
2327 -F::
2428 --freq=::
2529 Show just the sample frequency used for each event.
tools/perf/Documentation/perf-inject.txt+4−0
@@ -60,6 +60,10 @@ include::itrace.txt[] · 59 unmodified lines
6060 found in the jitdumps files captured in the input perf.data file. Use this option
6161 if you are monitoring environment using JIT runtimes, such as Java, DART or V8.
6262
63+-f::
64+--force::
65+ Don't complain, do it.
66+
6367 SEE ALSO
6468 --------
6569 linkperf:perf-record[1], linkperf:perf-report[1], linkperf:perf-archive[1]
tools/perf/Documentation/perf-lock.txt+4−0
@@ -42,6 +42,10 @@ COMMON OPTIONS · 41 unmodified lines
4242 --dump-raw-trace::
4343 Dump raw trace in ASCII.
4444
45+-f::
46+--force::
47+ Don't complan, do it.
48+
4549 REPORT OPTIONS
4650 --------------
4751
tools/perf/Documentation/perf-probe.txt+17−1
@@ -170,7 +170,7 @@ Probe points are defined by following syntax. · 169 unmodified lines
170170 or,
171171 sdt_PROVIDER:SDTEVENT
172172
173−'EVENT' specifies the name of new event, if omitted, it will be set the name of the probed function. You can also specify a group name by 'GROUP', if omitted, set 'probe' is used for kprobe and 'probe_<bin>' is used for uprobe.
173+'EVENT' specifies the name of new event, if omitted, it will be set the name of the probed function, and for return probes, a "\_\_return" suffix is automatically added to the function name. You can also specify a group name by 'GROUP', if omitted, set 'probe' is used for kprobe and 'probe_<bin>' is used for uprobe.
174174 Note that using existing group name can conflict with other events. Especially, using the group name reserved for kernel modules can hide embedded events in the
175175 modules.
176176 'FUNC' specifies a probed function name, and it may have one of the following options; '+OFFS' is the offset from function entry address in bytes, ':RLN' is the relative-line number from function entry line, and '%return' means that it probes function return. And ';PTN' means lazy matching pattern (see LAZY MATCHING). Note that ';PTN' must be the end of the probe point definition. In addition, '@SRC' specifies a source file which has that function.
@@ -182,6 +182,14 @@ Note that before using the SDT event, the target binary (on which SDT events are · 5 unmodified lines
182182 For details of the SDT, see below.
183183 https://sourceware.org/gdb/onlinedocs/gdb/Static-Probe-Points.html
184184
185+ESCAPED CHARACTER
186+-----------------
187+
188+In the probe syntax, '=', '@', '+', ':' and ';' are treated as a special character. You can use a backslash ('\') to escape the special characters.
189+This is useful if you need to probe on a specific versioned symbols, like @GLIBC_... suffixes, or also you need to specify a source file which includes the special characters.
190+Note that usually single backslash is consumed by shell, so you might need to pass double backslash (\\) or wrapping with single quotes (\'AAA\@BBB').
191+See EXAMPLES how it is used.
192+
185193 PROBE ARGUMENT
186194 --------------
187195 Each probe argument follows below syntax.
@@ -277,6 +285,14 @@ Add a USDT probe to a target process running in a different mount namespace · 89 unmodified lines
277285
278286 ./perf probe --target-ns <target pid> -x /usr/lib/jvm/java-1.8.0-openjdk-1.8.0.121-0.b13.el7_3.x86_64/jre/lib/amd64/server/libjvm.so %sdt_hotspot:thread__sleep__end
279287
288+Add a probe on specific versioned symbol by backslash escape
289+
290+ ./perf probe -x /lib64/libc-2.25.so 'malloc_get_state\@GLIBC_2.2.5'
291+
292+Add a probe in a source file using special characters by backslash escape
293+
294+ ./perf probe -x /opt/test/a.out 'foo\+bar.c:4'
295+
280296
281297 SEE ALSO
282298 --------
tools/perf/Documentation/perf-record.txt+3−0
@@ -430,6 +430,9 @@ Configure all used events to run in user space. · 429 unmodified lines
430430 --timestamp-filename
431431 Append timestamp to output file name.
432432
433+--timestamp-boundary::
434+Record timestamp boundary (time of first/last samples).
435+
433436 --switch-output[=mode]::
434437 Generate multiple perf.data files, timestamp prefixed, switching to a new one
435438 based on 'mode' value:
tools/perf/Documentation/perf-report.txt+36−1
@@ -402,6 +402,26 @@ OPTIONS · 401 unmodified lines
402402 stop time is not given (i.e, time string is 'x.y,') then analysis goes
403403 to end of file.
404404
405+ Also support time percent with multiple time range. Time string is
406+ 'a%/n,b%/m,...' or 'a%-b%,c%-%d,...'.
407+
408+ For example:
409+ Select the second 10% time slice:
410+
411+ perf report --time 10%/2
412+
413+ Select from 0% to 10% time slice:
414+
415+ perf report --time 0%-10%
416+
417+ Select the first and second 10% time slices:
418+
419+ perf report --time 10%/1,10%/2
420+
421+ Select from 0% to 10% and 30% to 40% slices:
422+
423+ perf report --time 0%-10%,30%-40%
424+
405425 --itrace::
406426 Options for decoding instruction tracing data. The options are:
407427
@@ -437,8 +457,23 @@ include::itrace.txt[] · 29 unmodified lines
437457 will be printed. Each entry is function name or file/line. Enabled by
438458 default, disable with --no-inline.
439459
460+--mmaps::
461+ Show --tasks output plus mmap information in a format similar to
462+ /proc/<PID>/maps.
463+
464+ Please note that not all mmaps are stored, options affecting which ones
465+ are include 'perf record --data', for instance.
466+
467+--stats::
468+ Display overall events statistics without any further processing.
469+ (like the one at the end of the perf report -D command)
470+
471+--tasks::
472+ Display monitored tasks stored in perf data. Displaying pid/tid/ppid
473+ plus the command string aligned to distinguish parent and child tasks.
474+
440475 include::callchain-overhead-calculation.txt[]
441476
442477 SEE ALSO
443478 --------
444−linkperf:perf-stat[1], linkperf:perf-annotate[1]
479+linkperf:perf-stat[1], linkperf:perf-annotate[1], linkperf:perf-record[1]
tools/perf/Documentation/perf-sched.txt+4−0
@@ -74,6 +74,10 @@ OPTIONS · 73 unmodified lines
7474 --dump-raw-trace=::
7575 Display verbose dump of the sched data.
7676
77+-f::
78+--force::
79+ Don't complain, do it.
80+
7781 OPTIONS for 'perf sched map'
7882 ----------------------------
7983
tools/perf/Documentation/perf-script.txt+46−1
@@ -117,7 +117,7 @@ OPTIONS · 116 unmodified lines
117117 Comma separated list of fields to print. Options are:
118118 comm, tid, pid, time, cpu, event, trace, ip, sym, dso, addr, symoff,
119119 srcline, period, iregs, uregs, brstack, brstacksym, flags, bpf-output, brstackinsn,
120− brstackoff, callindent, insn, insnlen, synth, phys_addr.
120+ brstackoff, callindent, insn, insnlen, synth, phys_addr, metric, misc.
121121 Field list can be prepended with the type, trace, sw or hw,
122122 to indicate to which event type the field list applies.
123123 e.g., -F sw:comm,tid,time,ip,sym and -F trace:time,cpu,trace
@@ -217,6 +217,32 @@ OPTIONS · 93 unmodified lines
217217
218218 The brstackoff field will print an offset into a specific dso/binary.
219219
220+ With the metric option perf script can compute metrics for
221+ sampling periods, similar to perf stat. This requires
222+ specifying a group with multiple metrics with the :S option
223+ for perf record. perf will sample on the first event, and
224+ compute metrics for all the events in the group. Please note
225+ that the metric computed is averaged over the whole sampling
226+ period, not just for the sample point.
227+
228+ For sample events it's possible to display misc field with -F +misc option,
229+ following letters are displayed for each bit:
230+
231+ PERF_RECORD_MISC_KERNEL K
232+ PERF_RECORD_MISC_USER U
233+ PERF_RECORD_MISC_HYPERVISOR H
234+ PERF_RECORD_MISC_GUEST_KERNEL G
235+ PERF_RECORD_MISC_GUEST_USER g
236+ PERF_RECORD_MISC_MMAP_DATA* M
237+ PERF_RECORD_MISC_COMM_EXEC E
238+ PERF_RECORD_MISC_SWITCH_OUT S
239+
240+ $ perf script -F +misc ...
241+ sched-messaging 1414 K 28690.636582: 4590 cycles ...
242+ sched-messaging 1407 U 28690.636600: 325620 cycles ...
243+ sched-messaging 1414 K 28690.636608: 19473 cycles ...
244+ misc field ___________/
245+
220246 -k::
221247 --vmlinux=<file>::
222248 vmlinux pathname
@@ -274,6 +300,9 @@ OPTIONS · 51 unmodified lines
274300 Display context switch events i.e. events of type PERF_RECORD_SWITCH or
275301 PERF_RECORD_SWITCH_CPU_WIDE.
276302
303+--show-lost-events
304+ Display lost events i.e. events of type PERF_RECORD_LOST.
305+
277306 --demangle::
278307 Demangle symbol names to human readable form. It's enabled by default,
279308 disable with --no-demangle.
@@ -321,6 +350,22 @@ include::itrace.txt[] · 41 unmodified lines
321350 stop time is not given (i.e, time string is 'x.y,') then analysis goes
322351 to end of file.
323352
353+ Also support time percent with multipe time range. Time string is
354+ 'a%/n,b%/m,...' or 'a%-b%,c%-%d,...'.
355+
356+ For example:
357+ Select the second 10% time slice:
358+ perf script --time 10%/2
359+
360+ Select from 0% to 10% time slice:
361+ perf script --time 0%-10%
362+
363+ Select the first and second 10% time slices:
364+ perf script --time 10%/1,10%/2
365+
366+ Select from 0% to 10% and 30% to 40% slices:
367+ perf script --time 0%-10%,30%-40%
368+
324369 --max-blocks::
325370 Set the maximum number of program blocks to print with brstackasm for
326371 each sample.
tools/perf/Documentation/perf-timechart.txt+3−1
@@ -50,7 +50,9 @@ TIMECHART OPTIONS · 49 unmodified lines
5050 -p::
5151 --process::
5252 Select the processes to display, by name or PID
53−
53+-f::
54+--force::
55+ Don't complain, do it.
5456 --symfs=<directory>::
5557 Look for files with symbols relative to this directory.
5658 -n::
tools/perf/Documentation/perf-top.txt+6−0
@@ -268,6 +268,12 @@ INTERACTIVE PROMPTING KEYS · 267 unmodified lines
268268 [S]::
269269 Stop annotation, return to full profile display.
270270
271+[K]::
272+ Hide kernel symbols.
273+
274+[U]::
275+ Hide user symbols.
276+
271277 [z]::
272278 Toggle event count zeroing across display updates.
273279
tools/perf/Documentation/perf-trace.txt+14−6
@@ -86,18 +86,18 @@ comma-separated list with no space: 0,1. Ranges of CPUs are specified with -: 0- · 85 unmodified lines
8686 In per-thread mode with inheritance mode on (default), Events are captured only when
8787 the thread executes on the designated CPUs. Default is to monitor all CPUs.
8888
89−--duration:
89+--duration::
9090 Show only events that had a duration greater than N.M ms.
9191
92−--sched:
92+--sched::
9393 Accrue thread runtime and provide a summary at the end of the session.
9494
95−-i
96−--input
95+-i::
96+--input::
9797 Process events from a given perf data file.
9898
99−-T
100−--time
99+-T::
100+--time::
101101 Print full timestamp rather time relative to first sample.
102102
103103 --comm::
@@ -117,6 +117,10 @@ the thread executes on the designated CPUs. Default is to monitor all CPUs. · 13 unmodified lines
117117 Show tool stats such as number of times fd->pathname was discovered thru
118118 hooking the open syscall return + vfs_getname or via reading /proc/pid/fd, etc.
119119
120+-f::
121+--force::
122+ Don't complain, do it.
123+
120124 -F=[all|min|maj]::
121125 --pf=[all|min|maj]::
122126 Trace pagefaults. Optionally, you can specify whether you want minor,
@@ -159,6 +163,10 @@ the thread executes on the designated CPUs. Default is to monitor all CPUs. · 36 unmodified lines
159163 Implies '--call-graph dwarf' when --call-graph not present on the
160164 command line, on systems where DWARF unwinding was built in.
161165
166+--print-sample::
167+ Print the PERF_RECORD_SAMPLE PERF_SAMPLE_ info for the
168+ raw_syscalls:sys_{enter,exit} tracepoints, for debugging.
169+
162170 --proc-map-timeout::
163171 When processing pre-existing threads /proc/XXX/mmap, it may take a long time,
164172 because the file may be huge. A time out is needed in such cases.
tools/perf/Documentation/perf.data-file-format.txt+27−0
@@ -238,6 +238,33 @@ struct auxtrace_index { · 237 unmodified lines
238238 struct auxtrace_index_entry entries[PERF_AUXTRACE_INDEX_ENTRY_COUNT];
239239 };
240240
241+ HEADER_STAT = 19,
242+
243+This is merely a flag signifying that the data section contains data
244+recorded from perf stat record.
245+
246+ HEADER_CACHE = 20,
247+
248+Description of the cache hierarchy. Based on the Linux sysfs format
249+in /sys/devices/system/cpu/cpu*/cache/
250+
251+ u32 version Currently always 1
252+ u32 number_of_cache_levels
253+
254+struct {
255+ u32 level;
256+ u32 line_size;
257+ u32 sets;
258+ u32 ways;
259+ struct perf_header_string type;
260+ struct perf_header_string size;
261+ struct perf_header_string map;
262+}[number_of_cache_levels];
263+
264+ HEADER_SAMPLE_TIME = 21,
265+
266+Two uint64_t for the time of first sample and the time of last sample.
267+
241268 other bits are reserved and should ignored for now
242269 HEADER_FEAT_BITS = 256,
243270
tools/perf/Documentation/tips.txt+2−0
@@ -32,3 +32,5 @@ Order by the overhead of source file name and line number: perf report -s srclin · 31 unmodified lines
3232 System-wide collection from all CPUs: perf record -a
3333 Show current config key-value pairs: perf config --list
3434 Show user configuration overrides: perf config --user --list
35+To add Node.js USDT(User-Level Statically Defined Tracing): perf buildid-cache --add `which node`
36+To report cacheline events from previous recording: perf c2c report
tools/perf/Makefile.config+55−10
@@ -22,6 +22,7 @@ include $(srctree)/tools/scripts/Makefile.arch · 21 unmodified lines
2222 $(call detected_var,SRCARCH)
2323
2424 NO_PERF_REGS := 1
25+NO_SYSCALL_TABLE := 1
2526
2627 # Additional ARCH settings for ppc
2728 ifeq ($(SRCARCH),powerpc)
@@ -33,7 +34,8 @@ endif · 5 unmodified lines
3334 ifeq ($(SRCARCH),x86)
3435 $(call detected,CONFIG_X86)
3536 ifeq (${IS_64_BIT}, 1)
36− CFLAGS += -DHAVE_ARCH_X86_64_SUPPORT -DHAVE_SYSCALL_TABLE -I$(OUTPUT)arch/x86/include/generated
37+ NO_SYSCALL_TABLE := 0
38+ CFLAGS += -DHAVE_ARCH_X86_64_SUPPORT -I$(OUTPUT)arch/x86/include/generated
3739 ARCH_INCLUDE = ../../arch/x86/lib/memcpy_64.S ../../arch/x86/lib/memset_64.S
3840 LIBUNWIND_LIBS = -lunwind-x86_64 -lunwind -llzma
3941 $(call detected,CONFIG_X86_64)
@@ -55,12 +57,18 @@ endif · 15 unmodified lines
5557
5658 ifeq ($(ARCH),s390)
5759 NO_PERF_REGS := 0
60+ NO_SYSCALL_TABLE := 0
61+ CFLAGS += -fPIC -I$(OUTPUT)arch/s390/include/generated
5862 endif
5963
6064 ifeq ($(NO_PERF_REGS),0)
6165 $(call detected,CONFIG_PERF_REGS)
6266 endif
6367
68+ifneq ($(NO_SYSCALL_TABLE),1)
69+ CFLAGS += -DHAVE_SYSCALL_TABLE
70+endif
71+
6472 # So far there's only x86 and arm libdw unwind support merged in perf.
6573 # Disable it on all other architectures in case libdw unwind
6674 # support is detected in system. Add supported architectures
@@ -97,6 +105,16 @@ FEATURE_CHECK_LDFLAGS-libunwind = $(LIBUNWIND_LDFLAGS) $(LIBUNWIND_LIBS) · 30 unmodified lines
97105 FEATURE_CHECK_CFLAGS-libunwind-debug-frame = $(LIBUNWIND_CFLAGS)
98106 FEATURE_CHECK_LDFLAGS-libunwind-debug-frame = $(LIBUNWIND_LDFLAGS) $(LIBUNWIND_LIBS)
99107
108+ifdef CSINCLUDES
109+ LIBOPENCSD_CFLAGS := -I$(CSINCLUDES)
110+endif
111+OPENCSDLIBS := -lopencsd_c_api -lopencsd
112+ifdef CSLIBS
113+ LIBOPENCSD_LDFLAGS := -L$(CSLIBS)
114+endif
115+FEATURE_CHECK_CFLAGS-libopencsd := $(LIBOPENCSD_CFLAGS)
116+FEATURE_CHECK_LDFLAGS-libopencsd := $(LIBOPENCSD_LDFLAGS) $(OPENCSDLIBS)
117+
100118 ifeq ($(NO_PERF_REGS),0)
101119 CFLAGS += -DHAVE_PERF_REGS_SUPPORT
102120 endif
@@ -265,6 +283,10 @@ ifeq ($(feature-pthread-attr-setaffinity-np), 1) · 162 unmodified lines
265283 CFLAGS += -DHAVE_PTHREAD_ATTR_SETAFFINITY_NP
266284 endif
267285
286+ifeq ($(feature-pthread-barrier), 1)
287+ CFLAGS += -DHAVE_PTHREAD_BARRIER
288+endif
289+
268290 ifndef NO_BIONIC
269291 $(call feature_check,bionic)
270292 ifeq ($(feature-bionic), 1)
@@ -341,6 +363,21 @@ ifeq ($(feature-setns), 1) · 70 unmodified lines
341363 $(call detected,CONFIG_SETNS)
342364 endif
343365
366+ifndef NO_CORESIGHT
367+ ifeq ($(feature-libopencsd), 1)
368+ CFLAGS += -DHAVE_CSTRACE_SUPPORT $(LIBOPENCSD_CFLAGS)
369+ LDFLAGS += $(LIBOPENCSD_LDFLAGS)
370+ EXTLIBS += $(OPENCSDLIBS)
371+ $(call detected,CONFIG_LIBOPENCSD)
372+ ifdef CSTRACE_RAW
373+ CFLAGS += -DCS_DEBUG_RAW
374+ ifeq (${CSTRACE_RAW}, packed)
375+ CFLAGS += -DCS_RAW_PACKED
376+ endif
377+ endif
378+ endif
379+endif
380+
344381 ifndef NO_LIBELF
345382 CFLAGS += -DHAVE_LIBELF_SUPPORT
346383 EXTLIBS += -lelf
@@ -519,14 +556,18 @@ ifndef NO_LIBUNWIND · 172 unmodified lines
519556 EXTLIBS += $(EXTLIBS_LIBUNWIND)
520557 endif
521558
522−ifndef NO_LIBAUDIT
523− ifneq ($(feature-libaudit), 1)
524− msg := $(warning No libaudit.h found, disables 'trace' tool, please install audit-libs-devel or libaudit-dev);
525− NO_LIBAUDIT := 1
526− else
527− CFLAGS += -DHAVE_LIBAUDIT_SUPPORT
528− EXTLIBS += -laudit
529− $(call detected,CONFIG_AUDIT)
559+ifeq ($(NO_SYSCALL_TABLE),0)
560+ $(call detected,CONFIG_TRACE)
561+else
562+ ifndef NO_LIBAUDIT
563+ ifneq ($(feature-libaudit), 1)
564+ msg := $(warning No libaudit.h found, disables 'trace' tool, please install audit-libs-devel or libaudit-dev);
565+ NO_LIBAUDIT := 1
566+ else
567+ CFLAGS += -DHAVE_LIBAUDIT_SUPPORT
568+ EXTLIBS += -laudit
569+ $(call detected,CONFIG_TRACE)
570+ endif
530571 endif
531572 endif
532573
@@ -768,7 +809,7 @@ else · 235 unmodified lines
768809 NO_PERF_READ_VDSOX32 := 1
769810 endif
770811
771−ifdef LIBBABELTRACE
812+ifndef NO_LIBBABELTRACE
772813 $(call feature_check,libbabeltrace)
773814 ifeq ($(feature-libbabeltrace), 1)
774815 CFLAGS += -DHAVE_LIBBABELTRACE_SUPPORT $(LIBBABELTRACE_CFLAGS)
@@ -935,6 +976,10 @@ define print_var_code · 160 unmodified lines
935976 endef
936977
937978 ifeq ($(VF),1)
979+ # Display EXTRA features which are detected manualy
980+ # from here with feature_check call and thus cannot
981+ # be partof global state output.
982+ $(foreach feat,$(FEATURE_TESTS_EXTRA),$(call feature_print_status,$(feat),))
938983 $(call print_var,prefix)
939984 $(call print_var,bindir)
940985 $(call print_var,libdir)
tools/perf/Makefile.perf+14−3
@@ -77,7 +77,7 @@ include ../scripts/utilities.mak · 76 unmodified lines
7777 #
7878 # Define NO_ZLIB if you do not want to support compressed kernel modules
7979 #
80−# Define LIBBABELTRACE if you DO want libbabeltrace support
80+# Define NO_LIBBABELTRACE if you do not want libbabeltrace support
8181 # for CTF data format.
8282 #
8383 # Define NO_LZMA if you do not want to support compressed (xz) kernel modules
@@ -98,6 +98,8 @@ include ../scripts/utilities.mak · 14 unmodified lines
9898 # When selected, pass LLVM_CONFIG=/path/to/llvm-config to `make' if
9999 # llvm-config is not in $PATH.
100100
101+# Define NO_CORESIGHT if you do not want support for CoreSight trace decoding.
102+
101103 # As per kernel Makefile, avoid funny character set dependencies
102104 unexport LC_ALL
103105 LC_COLLATE=C
@@ -462,6 +464,13 @@ prctl_option_tbl := $(srctree)/tools/perf/trace/beauty/prctl_option.sh · 358 unmodified lines
462464 $(prctl_option_array): $(prctl_hdr_dir)/prctl.h $(prctl_option_tbl)
463465 $(Q)$(SHELL) '$(prctl_option_tbl)' $(prctl_hdr_dir) > $@
464466
467+arch_errno_name_array := $(beauty_outdir)/arch_errno_name_array.c
468+arch_errno_hdr_dir := $(srctree)/tools
469+arch_errno_tbl := $(srctree)/tools/perf/trace/beauty/arch_errno_names.sh
470+
471+$(arch_errno_name_array): $(arch_errno_tbl)
472+ $(Q)$(SHELL) '$(arch_errno_tbl)' $(CC) $(arch_errno_hdr_dir) > $@
473+
465474 all: shell_compatibility_test $(ALL_PROGRAMS) $(LANG_BINDINGS) $(OTHER_PROGRAMS)
466475
467476 $(OUTPUT)python/perf.so: $(PYTHON_EXT_SRCS) $(PYTHON_EXT_DEPS) $(LIBTRACEEVENT_DYNAMIC_LIST)
@@ -565,7 +574,8 @@ prepare: $(OUTPUT)PERF-VERSION-FILE $(OUTPUT)common-cmds.h archheaders $(drm_ioc · 97 unmodified lines
565574 $(vhost_virtio_ioctl_array) \
566575 $(madvise_behavior_array) \
567576 $(perf_ioctl_array) \
568− $(prctl_option_array)
577+ $(prctl_option_array) \
578+ $(arch_errno_name_array)
569579
570580 $(OUTPUT)%.o: %.c prepare FORCE
571581 $(Q)$(MAKE) -f $(srctree)/tools/build/Makefile.build dir=$(build-dir) $@
@@ -847,7 +857,8 @@ clean:: $(LIBTRACEEVENT)-clean $(LIBAPI)-clean $(LIBBPF)-clean $(LIBSUBCMD)-clea · 275 unmodified lines
847857 $(OUTPUT)$(kcmp_type_array) \
848858 $(OUTPUT)$(vhost_virtio_ioctl_array) \
849859 $(OUTPUT)$(perf_ioctl_array) \
850− $(OUTPUT)$(prctl_option_array)
860+ $(OUTPUT)$(prctl_option_array) \
861+ $(OUTPUT)$(arch_errno_name_array)
851862 $(QUIET_SUBDIR0)Documentation $(QUIET_SUBDIR1) clean
852863
853864 #
tools/perf/arch/arm/util/auxtrace.c+71−6
@@ -22,6 +22,42 @@ · 21 unmodified lines
2222 #include "../../util/evlist.h"
2323 #include "../../util/pmu.h"
2424 #include "cs-etm.h"
25+#include "arm-spe.h"
26+
27+static struct perf_pmu **find_all_arm_spe_pmus(int *nr_spes, int *err)
28+{
29+ struct perf_pmu **arm_spe_pmus = NULL;
30+ int ret, i, nr_cpus = sysconf(_SC_NPROCESSORS_CONF);
31+ /* arm_spe_xxxxxxxxx\0 */
32+ char arm_spe_pmu_name[sizeof(ARM_SPE_PMU_NAME) + 10];
33+
34+ arm_spe_pmus = zalloc(sizeof(struct perf_pmu *) * nr_cpus);
35+ if (!arm_spe_pmus) {
36+ pr_err("spes alloc failed\n");
37+ *err = -ENOMEM;
38+ return NULL;
39+ }
40+
41+ for (i = 0; i < nr_cpus; i++) {
42+ ret = sprintf(arm_spe_pmu_name, "%s%d", ARM_SPE_PMU_NAME, i);
43+ if (ret < 0) {
44+ pr_err("sprintf failed\n");
45+ *err = -ENOMEM;
46+ return NULL;
47+ }
48+
49+ arm_spe_pmus[*nr_spes] = perf_pmu__find(arm_spe_pmu_name);
50+ if (arm_spe_pmus[*nr_spes]) {
51+ pr_debug2("%s %d: arm_spe_pmu %d type %d name %s\n",
52+ __func__, __LINE__, *nr_spes,
53+ arm_spe_pmus[*nr_spes]->type,
54+ arm_spe_pmus[*nr_spes]->name);
55+ (*nr_spes)++;
56+ }
57+ }
58+
59+ return arm_spe_pmus;
60+}
2561
2662 struct auxtrace_record
2763 *auxtrace_record__init(struct perf_evlist *evlist, int *err)
@@ -29,22 +65,51 @@ struct auxtrace_record · 1 unmodified line
2965 struct perf_pmu *cs_etm_pmu;
3066 struct perf_evsel *evsel;
3167 bool found_etm = false;
68+ bool found_spe = false;
69+ static struct perf_pmu **arm_spe_pmus = NULL;
70+ static int nr_spes = 0;
71+ int i;
72+
73+ if (!evlist)
74+ return NULL;
3275
3376 cs_etm_pmu = perf_pmu__find(CORESIGHT_ETM_PMU_NAME);
3477
35− if (evlist) {
36− evlist__for_each_entry(evlist, evsel) {
37− if (cs_etm_pmu &&
38− evsel->attr.type == cs_etm_pmu->type)
39− found_etm = true;
78+ if (!arm_spe_pmus)
79+ arm_spe_pmus = find_all_arm_spe_pmus(&nr_spes, err);
80+
81+ evlist__for_each_entry(evlist, evsel) {
82+ if (cs_etm_pmu &&
83+ evsel->attr.type == cs_etm_pmu->type)
84+ found_etm = true;
85+
86+ if (!nr_spes)
87+ continue;
88+
89+ for (i = 0; i < nr_spes; i++) {
90+ if (evsel->attr.type == arm_spe_pmus[i]->type) {
91+ found_spe = true;
92+ break;
93+ }
4094 }
4195 }
4296
97+ if (found_etm && found_spe) {
98+ pr_err("Concurrent ARM Coresight ETM and SPE operation not currently supported\n");
99+ *err = -EOPNOTSUPP;
100+ return NULL;
101+ }
102+
43103 if (found_etm)
44104 return cs_etm_record_init(err);
45105
106+#if defined(__aarch64__)
107+ if (found_spe)
108+ return arm_spe_recording_init(err, arm_spe_pmus[i]);
109+#endif
110+
46111 /*
47− * Clear 'err' even if we haven't found a cs_etm event - that way perf
112+ * Clear 'err' even if we haven't found an event - that way perf
48113 * record can still be used even if tracers aren't present. The NULL
49114 * return value will take care of telling the infrastructure HW tracing
50115 * isn't available.
tools/perf/arch/arm/util/pmu.c+6−0
@@ -20,6 +20,7 @@ · 19 unmodified lines
2020 #include <linux/perf_event.h>
2121
2222 #include "cs-etm.h"
23+#include "arm-spe.h"
2324 #include "../../util/pmu.h"
2425
2526 struct perf_event_attr
@@ -30,7 +31,12 @@ struct perf_event_attr · 4 unmodified lines
3031 /* add ETM default config here */
3132 pmu->selectable = true;
3233 pmu->set_drv_config = cs_etm_set_drv_config;
34+#if defined(__aarch64__)
35+ } else if (strstarts(pmu->name, ARM_SPE_PMU_NAME)) {
36+ return arm_spe_pmu_default_config(pmu);
37+#endif
3338 }
39+
3440 #endif
3541 return NULL;
3642 }
tools/perf/arch/arm64/util/Build+4−1
@@ -1,6 +1,9 @@
1+libperf-y += header.o
2+libperf-y += sym-handling.o
13 libperf-$(CONFIG_DWARF) += dwarf-regs.o
24 libperf-$(CONFIG_LOCAL_LIBUNWIND) += unwind-libunwind.o
35
46 libperf-$(CONFIG_AUXTRACE) += ../../arm/util/pmu.o \
57 ../../arm/util/auxtrace.o \
6− ../../arm/util/cs-etm.o
8+ ../../arm/util/cs-etm.o \
9+ arm-spe.o
tools/perf/arch/arm64/util/arm-spe.c+225−0
@@ -0,0 +1,225 @@
1+// SPDX-License-Identifier: GPL-2.0
2+/*
3+ * Arm Statistical Profiling Extensions (SPE) support
4+ * Copyright (c) 2017-2018, Arm Ltd.
5+ */
6+
7+#include <linux/kernel.h>
8+#include <linux/types.h>
9+#include <linux/bitops.h>
10+#include <linux/log2.h>
11+#include <time.h>
12+
13+#include "../../util/cpumap.h"
14+#include "../../util/evsel.h"
15+#include "../../util/evlist.h"
16+#include "../../util/session.h"
17+#include "../../util/util.h"
18+#include "../../util/pmu.h"
19+#include "../../util/debug.h"
20+#include "../../util/auxtrace.h"
21+#include "../../util/arm-spe.h"
22+
23+#define KiB(x) ((x) * 1024)
24+#define MiB(x) ((x) * 1024 * 1024)
25+
26+struct arm_spe_recording {
27+ struct auxtrace_record itr;
28+ struct perf_pmu *arm_spe_pmu;
29+ struct perf_evlist *evlist;
30+};
31+
32+static size_t
33+arm_spe_info_priv_size(struct auxtrace_record *itr __maybe_unused,
34+ struct perf_evlist *evlist __maybe_unused)
35+{
36+ return ARM_SPE_AUXTRACE_PRIV_SIZE;
37+}
38+
39+static int arm_spe_info_fill(struct auxtrace_record *itr,
40+ struct perf_session *session,
41+ struct auxtrace_info_event *auxtrace_info,
42+ size_t priv_size)
43+{
44+ struct arm_spe_recording *sper =
45+ container_of(itr, struct arm_spe_recording, itr);
46+ struct perf_pmu *arm_spe_pmu = sper->arm_spe_pmu;
47+
48+ if (priv_size != ARM_SPE_AUXTRACE_PRIV_SIZE)
49+ return -EINVAL;
50+
51+ if (!session->evlist->nr_mmaps)
52+ return -EINVAL;
53+
54+ auxtrace_info->type = PERF_AUXTRACE_ARM_SPE;
55+ auxtrace_info->priv[ARM_SPE_PMU_TYPE] = arm_spe_pmu->type;
56+
57+ return 0;
58+}
59+
60+static int arm_spe_recording_options(struct auxtrace_record *itr,
61+ struct perf_evlist *evlist,
62+ struct record_opts *opts)
63+{
64+ struct arm_spe_recording *sper =
65+ container_of(itr, struct arm_spe_recording, itr);
66+ struct perf_pmu *arm_spe_pmu = sper->arm_spe_pmu;
67+ struct perf_evsel *evsel, *arm_spe_evsel = NULL;
68+ bool privileged = geteuid() == 0 || perf_event_paranoid() < 0;
69+ struct perf_evsel *tracking_evsel;
70+ int err;
71+
72+ sper->evlist = evlist;
73+
74+ evlist__for_each_entry(evlist, evsel) {
75+ if (evsel->attr.type == arm_spe_pmu->type) {
76+ if (arm_spe_evsel) {
77+ pr_err("There may be only one " ARM_SPE_PMU_NAME "x event\n");
78+ return -EINVAL;
79+ }
80+ evsel->attr.freq = 0;
81+ evsel->attr.sample_period = 1;
82+ arm_spe_evsel = evsel;
83+ opts->full_auxtrace = true;
84+ }
85+ }
86+
87+ if (!opts->full_auxtrace)
88+ return 0;
89+
90+ /* We are in full trace mode but '-m,xyz' wasn't specified */
91+ if (opts->full_auxtrace && !opts->auxtrace_mmap_pages) {
92+ if (privileged) {
93+ opts->auxtrace_mmap_pages = MiB(4) / page_size;
94+ } else {
95+ opts->auxtrace_mmap_pages = KiB(128) / page_size;
96+ if (opts->mmap_pages == UINT_MAX)
97+ opts->mmap_pages = KiB(256) / page_size;
98+ }
99+ }
100+
101+ /* Validate auxtrace_mmap_pages */
102+ if (opts->auxtrace_mmap_pages) {
103+ size_t sz = opts->auxtrace_mmap_pages * (size_t)page_size;
104+ size_t min_sz = KiB(8);
105+
106+ if (sz < min_sz || !is_power_of_2(sz)) {
107+ pr_err("Invalid mmap size for ARM SPE: must be at least %zuKiB and a power of 2\n",
108+ min_sz / 1024);
109+ return -EINVAL;
110+ }
111+ }
112+
113+
114+ /*
115+ * To obtain the auxtrace buffer file descriptor, the auxtrace event
116+ * must come first.
117+ */
118+ perf_evlist__to_front(evlist, arm_spe_evsel);
119+
120+ perf_evsel__set_sample_bit(arm_spe_evsel, CPU);
121+ perf_evsel__set_sample_bit(arm_spe_evsel, TIME);
122+ perf_evsel__set_sample_bit(arm_spe_evsel, TID);
123+
124+ /* Add dummy event to keep tracking */
125+ err = parse_events(evlist, "dummy:u", NULL);
126+ if (err)
127+ return err;
128+
129+ tracking_evsel = perf_evlist__last(evlist);
130+ perf_evlist__set_tracking_event(evlist, tracking_evsel);
131+
132+ tracking_evsel->attr.freq = 0;
133+ tracking_evsel->attr.sample_period = 1;
134+ perf_evsel__set_sample_bit(tracking_evsel, TIME);
135+ perf_evsel__set_sample_bit(tracking_evsel, CPU);
136+ perf_evsel__reset_sample_bit(tracking_evsel, BRANCH_STACK);
137+
138+ return 0;
139+}
140+
141+static u64 arm_spe_reference(struct auxtrace_record *itr __maybe_unused)
142+{
143+ struct timespec ts;
144+
145+ clock_gettime(CLOCK_MONOTONIC_RAW, &ts);
146+
147+ return ts.tv_sec ^ ts.tv_nsec;
148+}
149+
150+static void arm_spe_recording_free(struct auxtrace_record *itr)
151+{
152+ struct arm_spe_recording *sper =
153+ container_of(itr, struct arm_spe_recording, itr);
154+
155+ free(sper);
156+}
157+
158+static int arm_spe_read_finish(struct auxtrace_record *itr, int idx)
159+{
160+ struct arm_spe_recording *sper =
161+ container_of(itr, struct arm_spe_recording, itr);
162+ struct perf_evsel *evsel;
163+
164+ evlist__for_each_entry(sper->evlist, evsel) {
165+ if (evsel->attr.type == sper->arm_spe_pmu->type)
166+ return perf_evlist__enable_event_idx(sper->evlist,
167+ evsel, idx);
168+ }
169+ return -EINVAL;
170+}
171+
172+struct auxtrace_record *arm_spe_recording_init(int *err,
173+ struct perf_pmu *arm_spe_pmu)
174+{
175+ struct arm_spe_recording *sper;
176+
177+ if (!arm_spe_pmu) {
178+ *err = -ENODEV;
179+ return NULL;
180+ }
181+
182+ sper = zalloc(sizeof(struct arm_spe_recording));
183+ if (!sper) {
184+ *err = -ENOMEM;
185+ return NULL;
186+ }
187+
188+ sper->arm_spe_pmu = arm_spe_pmu;
189+ sper->itr.recording_options = arm_spe_recording_options;
190+ sper->itr.info_priv_size = arm_spe_info_priv_size;
191+ sper->itr.info_fill = arm_spe_info_fill;
192+ sper->itr.free = arm_spe_recording_free;
193+ sper->itr.reference = arm_spe_reference;
194+ sper->itr.read_finish = arm_spe_read_finish;
195+ sper->itr.alignment = 0;
196+
197+ return &sper->itr;
198+}
199+
200+struct perf_event_attr
201+*arm_spe_pmu_default_config(struct perf_pmu *arm_spe_pmu)
202+{
203+ struct perf_event_attr *attr;
204+
205+ attr = zalloc(sizeof(struct perf_event_attr));
206+ if (!attr) {
207+ pr_err("arm_spe default config cannot allocate a perf_event_attr\n");
208+ return NULL;
209+ }
210+
211+ /*
212+ * If kernel driver doesn't advertise a minimum,
213+ * use max allowable by PMSIDR_EL1.INTERVAL
214+ */
215+ if (perf_pmu__scan_file(arm_spe_pmu, "caps/min_interval", "%llu",
216+ &attr->sample_period) != 1) {
217+ pr_debug("arm_spe driver doesn't advertise a min. interval. Using 4096\n");
218+ attr->sample_period = 4096;
219+ }
220+
221+ arm_spe_pmu->selectable = true;
222+ arm_spe_pmu->is_uncore = false;
223+
224+ return attr;
225+}
tools/perf/arch/arm64/util/header.c+65−0
@@ -0,0 +1,65 @@
1+#include <stdio.h>
2+#include <stdlib.h>
3+#include <api/fs/fs.h>
4+#include "header.h"
5+
6+#define MIDR "/regs/identification/midr_el1"
7+#define MIDR_SIZE 19
8+#define MIDR_REVISION_MASK 0xf
9+#define MIDR_VARIANT_SHIFT 20
10+#define MIDR_VARIANT_MASK (0xf << MIDR_VARIANT_SHIFT)
11+
12+char *get_cpuid_str(struct perf_pmu *pmu)
13+{
14+ char *buf = NULL;
15+ char path[PATH_MAX];
16+ const char *sysfs = sysfs__mountpoint();
17+ int cpu;
18+ u64 midr = 0;
19+ struct cpu_map *cpus;
20+ FILE *file;
21+
22+ if (!sysfs || !pmu || !pmu->cpus)
23+ return NULL;
24+
25+ buf = malloc(MIDR_SIZE);
26+ if (!buf)
27+ return NULL;
28+
29+ /* read midr from list of cpus mapped to this pmu */
30+ cpus = cpu_map__get(pmu->cpus);
31+ for (cpu = 0; cpu < cpus->nr; cpu++) {
32+ scnprintf(path, PATH_MAX, "%s/devices/system/cpu/cpu%d"MIDR,
33+ sysfs, cpus->map[cpu]);
34+
35+ file = fopen(path, "r");
36+ if (!file) {
37+ pr_debug("fopen failed for file %s\n", path);
38+ continue;
39+ }
40+
41+ if (!fgets(buf, MIDR_SIZE, file)) {
42+ fclose(file);
43+ continue;
44+ }
45+ fclose(file);
46+
47+ /* Ignore/clear Variant[23:20] and
48+ * Revision[3:0] of MIDR
49+ */
50+ midr = strtoul(buf, NULL, 16);
51+ midr &= (~(MIDR_VARIANT_MASK | MIDR_REVISION_MASK));
52+ scnprintf(buf, MIDR_SIZE, "0x%016lx", midr);
53+ /* got midr break loop */
54+ break;
55+ }
56+
57+ if (!midr) {
58+ pr_err("failed to get cpuid string for PMU %s\n", pmu->name);
59+ free(buf);
60+ buf = NULL;
61+ }
62+
63+ cpu_map__put(cpus);
64+ return buf;
65+}
tools/perf/arch/arm64/util/sym-handling.c+22−0
@@ -0,0 +1,22 @@
1+/*
2+ * This program is free software; you can redistribute it and/or modify
3+ * it under the terms of the GNU General Public License, version 2, as
4+ * published by the Free Software Foundation.
5+ *
6+ * Copyright (C) 2015 Naveen N. Rao, IBM Corporation
7+ */
8+
9+#include "debug.h"
10+#include "symbol.h"
11+#include "map.h"
12+#include "probe-event.h"
13+#include "probe-file.h"
14+
15+#ifdef HAVE_LIBELF_SUPPORT
16+bool elf__needs_adjust_symbols(GElf_Ehdr ehdr)
17+{
18+ return ehdr.e_type == ET_EXEC ||
19+ ehdr.e_type == ET_REL ||
20+ ehdr.e_type == ET_DYN;
21+}
22+#endif
tools/perf/arch/common.c+3−41
@@ -1,12 +1,10 @@
11 // SPDX-License-Identifier: GPL-2.0
22 #include <stdio.h>
3−#include <sys/utsname.h>
43 #include "common.h"
4+#include "../util/env.h"
55 #include "../util/util.h"
66 #include "../util/debug.h"
77
8−#include "sane_ctype.h"
9−
108 const char *const arm_triplets[] = {
119 "arm-eabi-",
1210 "arm-linux-androideabi-",
@@ -120,55 +118,19 @@ static int lookup_triplets(const char *const *triplets, const char *name) · 107 unmodified lines
120118 return -1;
121119 }
122120
123−/*
124− * Return architecture name in a normalized form.
125− * The conversion logic comes from the Makefile.
126− */
127−const char *normalize_arch(char *arch)
128−{
129− if (!strcmp(arch, "x86_64"))
130− return "x86";
131− if (arch[0] == 'i' && arch[2] == '8' && arch[3] == '6')
132− return "x86";
133− if (!strcmp(arch, "sun4u") || !strncmp(arch, "sparc", 5))
134− return "sparc";
135− if (!strcmp(arch, "aarch64") || !strcmp(arch, "arm64"))
136− return "arm64";
137− if (!strncmp(arch, "arm", 3) || !strcmp(arch, "sa110"))
138− return "arm";
139− if (!strncmp(arch, "s390", 4))
140− return "s390";
141− if (!strncmp(arch, "parisc", 6))
142− return "parisc";
143− if (!strncmp(arch, "powerpc", 7) || !strncmp(arch, "ppc", 3))
144− return "powerpc";
145− if (!strncmp(arch, "mips", 4))
146− return "mips";
147− if (!strncmp(arch, "sh", 2) && isdigit(arch[2]))
148− return "sh";
149−
150− return arch;
151−}
152−
153121 static int perf_env__lookup_binutils_path(struct perf_env *env,
154122 const char *name, const char **path)
155123 {
156124 int idx;
157− const char *arch, *cross_env;
158− struct utsname uts;
125+ const char *arch = perf_env__arch(env), *cross_env;
159126 const char *const *path_list;
160127 char *buf = NULL;
161128
162− arch = normalize_arch(env->arch);
163−
164− if (uname(&uts) < 0)
165− goto out;
166−
167129 /*
168130 * We don't need to try to find objdump path for native system.
169131 * Just use default binutils path (e.g.: "objdump").
170132 */
171− if (!strcmp(normalize_arch(uts.machine), arch))
133+ if (!strcmp(perf_env__arch(NULL), arch))
172134 goto out;
173135
174136 cross_env = getenv("CROSS_COMPILE");
tools/perf/arch/common.h+0−1
@@ -7,6 +7,5 @@ · 6 unmodified lines
77 extern const char *objdump_path;
88
99 int perf_env__lookup_objdump(struct perf_env *env);
10−const char *normalize_arch(char *arch);
1110
1211 #endif /* ARCH_PERF_COMMON_H */
tools/perf/arch/powerpc/util/header.c+1−1
@@ -35,7 +35,7 @@ get_cpuid(char *buffer, size_t sz) · 34 unmodified lines
3535 }
3636
3737 char *
38−get_cpuid_str(void)
38+get_cpuid_str(struct perf_pmu *pmu __maybe_unused)
3939 {
4040 char *bufp;
4141
tools/perf/arch/powerpc/util/sym-handling.c+8−0
@@ -64,6 +64,14 @@ int arch__compare_symbol_names_n(const char *namea, const char *nameb, · 63 unmodified lines
6464
6565 return strncmp(namea, nameb, n);
6666 }
67+
68+const char *arch__normalize_symbol_name(const char *name)
69+{
70+ /* Skip over initial dot */
71+ if (name && *name == '.')
72+ name++;
73+ return name;
74+}
6775 #endif
6876
6977 #if defined(_CALL_ELF) && _CALL_ELF == 2
tools/perf/arch/s390/Makefile+21−0
@@ -3,3 +3,24 @@ PERF_HAVE_DWARF_REGS := 1 · 2 unmodified lines
33 endif
44 HAVE_KVM_STAT_SUPPORT := 1
55 PERF_HAVE_ARCH_REGS_QUERY_REGISTER_OFFSET := 1
6+
7+#
8+# Syscall table generation for perf
9+#
10+
11+out := $(OUTPUT)arch/s390/include/generated/asm
12+header := $(out)/syscalls_64.c
13+sysdef := $(srctree)/tools/arch/s390/include/uapi/asm/unistd.h
14+sysprf := $(srctree)/tools/perf/arch/s390/entry/syscalls/
15+systbl := $(sysprf)/mksyscalltbl
16+
17+# Create output directory if not already present
18+_dummy := $(shell [ -d '$(out)' ] || mkdir -p '$(out)')
19+
20+$(header): $(sysdef) $(systbl)
21+ $(Q)$(SHELL) '$(systbl)' '$(CC)' $(sysdef) > $@
22+
23+clean::
24+ $(call QUIET_CLEAN, s390) $(RM) $(header)
25+
26+archheaders: $(header)
tools/perf/arch/s390/annotate/instructions.c+2−1
@@ -18,7 +18,8 @@ static struct ins_ops *s390__associate_ins_ops(struct arch *arch, const char *na · 17 unmodified lines
1818 if (!strcmp(name, "br"))
1919 ops = &ret_ops;
2020
21− arch__associate_ins_ops(arch, name, ops);
21+ if (ops)
22+ arch__associate_ins_ops(arch, name, ops);
2223 return ops;
2324 }
2425
tools/perf/arch/s390/entry/syscalls/mksyscalltbl+36−0
@@ -0,0 +1,36 @@
1+#!/bin/sh
2+# SPDX-License-Identifier: GPL-2.0
3+#
4+# Generate system call table for perf
5+#
6+#
7+# Copyright IBM Corp. 2017
8+# Author(s): Hendrik Brueckner <brueckner@linux.vnet.ibm.com>
9+#
10+
11+gcc=$1
12+input=$2
13+
14+if ! test -r $input; then
15+ echo "Could not read input file" >&2
16+ exit 1
17+fi
18+
19+create_table()
20+{
21+ local max_nr
22+
23+ echo 'static const char *syscalltbl_s390_64[] = {'
24+ while read sc nr; do
25+ printf '\t[%d] = "%s",\n' $nr $sc
26+ max_nr=$nr
27+ done
28+ echo '};'
29+ echo "#define SYSCALLTBL_S390_64_MAX_ID $max_nr"
30+}
31+
32+
33+$gcc -m64 -E -dM -x c $input \
34+ |sed -ne 's/^#define __NR_//p' \
35+ |sort -t' ' -k2 -nu \
36+ |create_table
tools/perf/arch/x86/tests/perf-time-to-tsc.c+1−1
@@ -84,7 +84,7 @@ int test__perf_time_to_tsc(struct test *test __maybe_unused, int subtest __maybe · 83 unmodified lines
8484
8585 CHECK__(perf_evlist__open(evlist));
8686
87− CHECK__(perf_evlist__mmap(evlist, UINT_MAX, false));
87+ CHECK__(perf_evlist__mmap(evlist, UINT_MAX));
8888
8989 pc = evlist->mmap[0].base;
9090 ret = perf_read_tsc_conversion(pc, &tc);
tools/perf/arch/x86/util/header.c+2−2
@@ -66,11 +66,11 @@ get_cpuid(char *buffer, size_t sz) · 65 unmodified lines
6666 }
6767
6868 char *
69−get_cpuid_str(void)
69+get_cpuid_str(struct perf_pmu *pmu __maybe_unused)
7070 {
7171 char *buf = malloc(128);
7272
73− if (__get_cpuid(buf, 128, "%s-%u-%X$") < 0) {
73+ if (buf && __get_cpuid(buf, 128, "%s-%u-%X$") < 0) {
7474 free(buf);
7575 return NULL;
7676 }
tools/perf/arch/x86/util/unwind-libunwind.c+1−1
@@ -1,7 +1,7 @@
11 // SPDX-License-Identifier: GPL-2.0
22
3−#ifndef REMOTE_UNWIND_LIBUNWIND
43 #include <errno.h>
4+#ifndef REMOTE_UNWIND_LIBUNWIND
55 #include <libunwind.h>
66 #include "perf_regs.h"
77 #include "../../util/unwind.h"
tools/perf/bench/futex-hash.c+12−8
@@ -24,9 +24,9 @@ · 23 unmodified lines
2424 #include <subcmd/parse-options.h>
2525 #include "bench.h"
2626 #include "futex.h"
27+#include "cpumap.h"
2728
2829 #include <err.h>
29−#include <sys/time.h>
3030
3131 static unsigned int nthreads = 0;
3232 static unsigned int nsecs = 10;
@@ -118,11 +118,12 @@ static void print_summary(void) · 85 unmodified lines
118118 int bench_futex_hash(int argc, const char **argv)
119119 {
120120 int ret = 0;
121− cpu_set_t cpu;
121+ cpu_set_t cpuset;
122122 struct sigaction act;
123− unsigned int i, ncpus;
123+ unsigned int i;
124124 pthread_attr_t thread_attr;
125125 struct worker *worker = NULL;
126+ struct cpu_map *cpu;
126127
127128 argc = parse_options(argc, argv, options, bench_futex_hash_usage, 0);
128129 if (argc) {
@@ -130,14 +131,16 @@ int bench_futex_hash(int argc, const char **argv) · 1 unmodified line
130131 exit(EXIT_FAILURE);
131132 }
132133
133− ncpus = sysconf(_SC_NPROCESSORS_ONLN);
134+ cpu = cpu_map__new(NULL);
135+ if (!cpu)
136+ goto errmem;
134137
135138 sigfillset(&act.sa_mask);
136139 act.sa_sigaction = toggle_done;
137140 sigaction(SIGINT, &act, NULL);
138141
139142 if (!nthreads) /* default to the number of CPUs */
140− nthreads = ncpus;
143+ nthreads = cpu->nr;
141144
142145 worker = calloc(nthreads, sizeof(*worker));
143146 if (!worker)
@@ -163,10 +166,10 @@ int bench_futex_hash(int argc, const char **argv) · 19 unmodified lines
163166 if (!worker[i].futex)
164167 goto errmem;
165168
166− CPU_ZERO(&cpu);
167− CPU_SET(i % ncpus, &cpu);
169+ CPU_ZERO(&cpuset);
170+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);
168171
169− ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu);
172+ ret = pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset);
170173 if (ret)
171174 err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
172175
@@ -217,6 +220,7 @@ int bench_futex_hash(int argc, const char **argv) · 44 unmodified lines
217220 print_summary();
218221
219222 free(worker);
223+ free(cpu);
220224 return ret;
221225 errmem:
222226 err(EXIT_FAILURE, "calloc");
tools/perf/bench/futex-lock-pi.c+14−9
@@ -15,6 +15,7 @@ · 14 unmodified lines
1515 #include <errno.h>
1616 #include "bench.h"
1717 #include "futex.h"
18+#include "cpumap.h"
1819
1920 #include <err.h>
2021 #include <stdlib.h>
@@ -32,7 +33,7 @@ static struct worker *worker; · 11 unmodified lines
3233 static unsigned int nsecs = 10;
3334 static bool silent = false, multi = false;
3435 static bool done = false, fshared = false;
35−static unsigned int ncpus, nthreads = 0;
36+static unsigned int nthreads = 0;
3637 static int futex_flag = 0;
3738 struct timeval start, end, runtime;
3839 static pthread_mutex_t thread_lock;
@@ -113,9 +114,10 @@ static void *workerfn(void *arg) · 74 unmodified lines
113114 return NULL;
114115 }
115116
116−static void create_threads(struct worker *w, pthread_attr_t thread_attr)
117+static void create_threads(struct worker *w, pthread_attr_t thread_attr,
118+ struct cpu_map *cpu)
117119 {
118− cpu_set_t cpu;
120+ cpu_set_t cpuset;
119121 unsigned int i;
120122
121123 threads_starting = nthreads;
@@ -130,10 +132,10 @@ static void create_threads(struct worker *w, pthread_attr_t thread_attr) · 8 unmodified lines
130132 } else
131133 worker[i].futex = &global_futex;
132134
133− CPU_ZERO(&cpu);
134− CPU_SET(i % ncpus, &cpu);
135+ CPU_ZERO(&cpuset);
136+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);
135137
136− if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
138+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
137139 err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
138140
139141 if (pthread_create(&w[i].thread, &thread_attr, workerfn, &worker[i]))
@@ -147,19 +149,22 @@ int bench_futex_lock_pi(int argc, const char **argv) · 7 unmodified lines
147149 unsigned int i;
148150 struct sigaction act;
149151 pthread_attr_t thread_attr;
152+ struct cpu_map *cpu;
150153
151154 argc = parse_options(argc, argv, options, bench_futex_lock_pi_usage, 0);
152155 if (argc)
153156 goto err;
154157
155− ncpus = sysconf(_SC_NPROCESSORS_ONLN);
158+ cpu = cpu_map__new(NULL);
159+ if (!cpu)
160+ err(EXIT_FAILURE, "calloc");
156161
157162 sigfillset(&act.sa_mask);
158163 act.sa_sigaction = toggle_done;
159164 sigaction(SIGINT, &act, NULL);
160165
161166 if (!nthreads)
162− nthreads = ncpus;
167+ nthreads = cpu->nr;
163168
164169 worker = calloc(nthreads, sizeof(*worker));
165170 if (!worker)
@@ -180,7 +185,7 @@ int bench_futex_lock_pi(int argc, const char **argv) · 14 unmodified lines
180185 pthread_attr_init(&thread_attr);
181186 gettimeofday(&start, NULL);
182187
183− create_threads(worker, thread_attr);
188+ create_threads(worker, thread_attr, cpu);
184189 pthread_attr_destroy(&thread_attr);
185190
186191 pthread_mutex_lock(&thread_lock);
tools/perf/bench/futex-requeue.c+13−9
@@ -22,6 +22,7 @@ · 21 unmodified lines
2222 #include <errno.h>
2323 #include "bench.h"
2424 #include "futex.h"
25+#include "cpumap.h"
2526
2627 #include <err.h>
2728 #include <stdlib.h>
@@ -40,7 +41,7 @@ static bool done = false, silent = false, fshared = false; · 12 unmodified lines
4041 static pthread_mutex_t thread_lock;
4142 static pthread_cond_t thread_parent, thread_worker;
4243 static struct stats requeuetime_stats, requeued_stats;
43−static unsigned int ncpus, threads_starting, nthreads = 0;
44+static unsigned int threads_starting, nthreads = 0;
4445 static int futex_flag = 0;
4546
4647 static const struct option options[] = {
@@ -83,19 +84,19 @@ static void *workerfn(void *arg __maybe_unused) · 36 unmodified lines
8384 }
8485
8586 static void block_threads(pthread_t *w,
86− pthread_attr_t thread_attr)
87+ pthread_attr_t thread_attr, struct cpu_map *cpu)
8788 {
88− cpu_set_t cpu;
89+ cpu_set_t cpuset;
8990 unsigned int i;
9091
9192 threads_starting = nthreads;
9293
9394 /* create and block all threads */
9495 for (i = 0; i < nthreads; i++) {
95− CPU_ZERO(&cpu);
96− CPU_SET(i % ncpus, &cpu);
96+ CPU_ZERO(&cpuset);
97+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);
9798
98− if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
99+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
99100 err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
100101
101102 if (pthread_create(&w[i], &thread_attr, workerfn, NULL))
@@ -116,19 +117,22 @@ int bench_futex_requeue(int argc, const char **argv) · 14 unmodified lines
116117 unsigned int i, j;
117118 struct sigaction act;
118119 pthread_attr_t thread_attr;
120+ struct cpu_map *cpu;
119121
120122 argc = parse_options(argc, argv, options, bench_futex_requeue_usage, 0);
121123 if (argc)
122124 goto err;
123125
124− ncpus = sysconf(_SC_NPROCESSORS_ONLN);
126+ cpu = cpu_map__new(NULL);
127+ if (!cpu)
128+ err(EXIT_FAILURE, "cpu_map__new");
125129
126130 sigfillset(&act.sa_mask);
127131 act.sa_sigaction = toggle_done;
128132 sigaction(SIGINT, &act, NULL);
129133
130134 if (!nthreads)
131− nthreads = ncpus;
135+ nthreads = cpu->nr;
132136
133137 worker = calloc(nthreads, sizeof(*worker));
134138 if (!worker)
@@ -156,7 +160,7 @@ int bench_futex_requeue(int argc, const char **argv) · 21 unmodified lines
156160 struct timeval start, end, runtime;
157161
158162 /* create, launch & block all threads */
159− block_threads(worker, thread_attr);
163+ block_threads(worker, thread_attr, cpu);
160164
161165 /* make sure all threads are already blocked */
162166 pthread_mutex_lock(&thread_lock);
tools/perf/bench/futex-wake-parallel.c+35−11
@@ -7,7 +7,17 @@ · 6 unmodified lines
77 * for each individual thread to service its share of work. Ultimately
88 * it can be used to measure futex_wake() changes.
99 */
10+#include "bench.h"
11+#include <linux/compiler.h>
12+#include "../util/debug.h"
1013
14+#ifndef HAVE_PTHREAD_BARRIER
15+int bench_futex_wake_parallel(int argc __maybe_unused, const char **argv __maybe_unused)
16+{
17+ pr_err("%s: pthread_barrier_t unavailable, disabling this test...\n", __func__);
18+ return 0;
19+}
20+#else /* HAVE_PTHREAD_BARRIER */
1121 /* For the CLR_() macros */
1222 #include <string.h>
1323 #include <pthread.h>
@@ -15,12 +25,11 @@ · 1 unmodified line
1525 #include <signal.h>
1626 #include "../util/stat.h"
1727 #include <subcmd/parse-options.h>
18−#include <linux/compiler.h>
1928 #include <linux/kernel.h>
2029 #include <linux/time64.h>
2130 #include <errno.h>
22−#include "bench.h"
2331 #include "futex.h"
32+#include "cpumap.h"
2433
2534 #include <err.h>
2635 #include <stdlib.h>
@@ -42,8 +51,9 @@ static bool done = false, silent = false, fshared = false; · 15 unmodified lines
4251 static unsigned int nblocked_threads = 0, nwaking_threads = 0;
4352 static pthread_mutex_t thread_lock;
4453 static pthread_cond_t thread_parent, thread_worker;
54+static pthread_barrier_t barrier;
4555 static struct stats waketime_stats, wakeup_stats;
46−static unsigned int ncpus, threads_starting;
56+static unsigned int threads_starting;
4757 static int futex_flag = 0;
4858
4959 static const struct option options[] = {
@@ -64,6 +74,8 @@ static void *waking_workerfn(void *arg) · 14 unmodified lines
6474 struct thread_data *waker = (struct thread_data *) arg;
6575 struct timeval start, end;
6676
77+ pthread_barrier_wait(&barrier);
78+
6779 gettimeofday(&start, NULL);
6880
6981 waker->nwoken = futex_wake(&futex, nwakes, futex_flag);
@@ -84,6 +96,8 @@ static void wakeup_threads(struct thread_data *td, pthread_attr_t thread_attr) · 14 unmodified lines
8496
8597 pthread_attr_setdetachstate(&thread_attr, PTHREAD_CREATE_JOINABLE);
8698
99+ pthread_barrier_init(&barrier, NULL, nwaking_threads + 1);
100+
87101 /* create and block all threads */
88102 for (i = 0; i < nwaking_threads; i++) {
89103 /*
@@ -96,9 +110,13 @@ static void wakeup_threads(struct thread_data *td, pthread_attr_t thread_attr) · 6 unmodified lines
96110 err(EXIT_FAILURE, "pthread_create");
97111 }
98112
113+ pthread_barrier_wait(&barrier);
114+
99115 for (i = 0; i < nwaking_threads; i++)
100116 if (pthread_join(td[i].worker, NULL))
101117 err(EXIT_FAILURE, "pthread_join");
118+
119+ pthread_barrier_destroy(&barrier);
102120 }
103121
104122 static void *blocked_workerfn(void *arg __maybe_unused)
@@ -119,19 +137,20 @@ static void *blocked_workerfn(void *arg __maybe_unused) · 14 unmodified lines
119137 return NULL;
120138 }
121139
122−static void block_threads(pthread_t *w, pthread_attr_t thread_attr)
140+static void block_threads(pthread_t *w, pthread_attr_t thread_attr,
141+ struct cpu_map *cpu)
123142 {
124− cpu_set_t cpu;
143+ cpu_set_t cpuset;
125144 unsigned int i;
126145
127146 threads_starting = nblocked_threads;
128147
129148 /* create and block all threads */
130149 for (i = 0; i < nblocked_threads; i++) {
131− CPU_ZERO(&cpu);
132− CPU_SET(i % ncpus, &cpu);
150+ CPU_ZERO(&cpuset);
151+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);
133152
134− if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
153+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
135154 err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
136155
137156 if (pthread_create(&w[i], &thread_attr, blocked_workerfn, NULL))
@@ -205,6 +224,7 @@ int bench_futex_wake_parallel(int argc, const char **argv) · 67 unmodified lines
205224 struct sigaction act;
206225 pthread_attr_t thread_attr;
207226 struct thread_data *waking_worker;
227+ struct cpu_map *cpu;
208228
209229 argc = parse_options(argc, argv, options,
210230 bench_futex_wake_parallel_usage, 0);
@@ -217,9 +237,12 @@ int bench_futex_wake_parallel(int argc, const char **argv) · 6 unmodified lines
217237 act.sa_sigaction = toggle_done;
218238 sigaction(SIGINT, &act, NULL);
219239
220− ncpus = sysconf(_SC_NPROCESSORS_ONLN);
240+ cpu = cpu_map__new(NULL);
241+ if (!cpu)
242+ err(EXIT_FAILURE, "calloc");
243+
221244 if (!nblocked_threads)
222− nblocked_threads = ncpus;
245+ nblocked_threads = cpu->nr;
223246
224247 /* some sanity checks */
225248 if (nwaking_threads > nblocked_threads || !nwaking_threads)
@@ -259,7 +282,7 @@ int bench_futex_wake_parallel(int argc, const char **argv) · 33 unmodified lines
259282 err(EXIT_FAILURE, "calloc");
260283
261284 /* create, launch & block all threads */
262− block_threads(blocked_worker, thread_attr);
285+ block_threads(blocked_worker, thread_attr, cpu);
263286
264287 /* make sure all threads are already blocked */
265288 pthread_mutex_lock(&thread_lock);
@@ -297,3 +320,4 @@ int bench_futex_wake_parallel(int argc, const char **argv) · 31 unmodified lines
297320 free(blocked_worker);
298321 return ret;
299322 }
323+#endif /* HAVE_PTHREAD_BARRIER */
tools/perf/bench/futex-wake.c+11−7
@@ -22,6 +22,7 @@ · 21 unmodified lines
2222 #include <errno.h>
2323 #include "bench.h"
2424 #include "futex.h"
25+#include "cpumap.h"
2526
2627 #include <err.h>
2728 #include <stdlib.h>
@@ -89,19 +90,19 @@ static void print_summary(void) · 61 unmodified lines
8990 }
9091
9192 static void block_threads(pthread_t *w,
92− pthread_attr_t thread_attr)
93+ pthread_attr_t thread_attr, struct cpu_map *cpu)
9394 {
94− cpu_set_t cpu;
95+ cpu_set_t cpuset;
9596 unsigned int i;
9697
9798 threads_starting = nthreads;
9899
99100 /* create and block all threads */
100101 for (i = 0; i < nthreads; i++) {
101− CPU_ZERO(&cpu);
102− CPU_SET(i % ncpus, &cpu);
102+ CPU_ZERO(&cpuset);
103+ CPU_SET(cpu->map[i % cpu->nr], &cpuset);
103104
104− if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpu))
105+ if (pthread_attr_setaffinity_np(&thread_attr, sizeof(cpu_set_t), &cpuset))
105106 err(EXIT_FAILURE, "pthread_attr_setaffinity_np");
106107
107108 if (pthread_create(&w[i], &thread_attr, workerfn, NULL))
@@ -122,6 +123,7 @@ int bench_futex_wake(int argc, const char **argv) · 14 unmodified lines
122123 unsigned int i, j;
123124 struct sigaction act;
124125 pthread_attr_t thread_attr;
126+ struct cpu_map *cpu;
125127
126128 argc = parse_options(argc, argv, options, bench_futex_wake_usage, 0);
127129 if (argc) {
@@ -129,7 +131,9 @@ int bench_futex_wake(int argc, const char **argv) · 1 unmodified line
129131 exit(EXIT_FAILURE);
130132 }
131133
132− ncpus = sysconf(_SC_NPROCESSORS_ONLN);
134+ cpu = cpu_map__new(NULL);
135+ if (!cpu)
136+ err(EXIT_FAILURE, "calloc");
133137
134138 sigfillset(&act.sa_mask);
135139 act.sa_sigaction = toggle_done;
@@ -161,7 +165,7 @@ int bench_futex_wake(int argc, const char **argv) · 25 unmodified lines
161165 struct timeval start, end, runtime;
162166
163167 /* create, launch & block all threads */
164− block_threads(worker, thread_attr);
168+ block_threads(worker, thread_attr, cpu);
165169
166170 /* make sure all threads are already blocked */
167171 pthread_mutex_lock(&thread_lock);
tools/perf/builtin-buildid-cache.c+2−2
@@ -325,8 +325,8 @@ int cmd_buildid_cache(int argc, const char **argv) · 324 unmodified lines
325325 "file", "kcore file to add"),
326326 OPT_STRING('r', "remove", &remove_name_list_str, "file list",
327327 "file(s) to remove"),
328− OPT_STRING('p', "purge", &purge_name_list_str, "path list",
329− "path(s) to remove (remove old caches too)"),
328+ OPT_STRING('p', "purge", &purge_name_list_str, "file list",
329+ "file(s) to remove (remove old caches too)"),
330330 OPT_STRING('M', "missing", &missing_filename, "file",
331331 "to find missing build ids in the cache"),
332332 OPT_BOOLEAN('f', "force", &force, "don't complain, do it"),
tools/perf/builtin-c2c.c+7−9
@@ -27,13 +27,10 @@ · 26 unmodified lines
2727 #include "sort.h"
2828 #include "tool.h"
2929 #include "data.h"
30−#include "sort.h"
3130 #include "event.h"
3231 #include "evlist.h"
3332 #include "evsel.h"
34−#include <asm/bug.h>
3533 #include "ui/browsers/hists.h"
36−#include "evlist.h"
3734 #include "thread.h"
3835
3936 struct c2c_hists {
@@ -2224,9 +2221,9 @@ static int perf_c2c__browse_cacheline(struct hist_entry *he) · 2184 unmodified lines
22242221 struct hist_browser *browser;
22252222 int key = -1;
22262223 const char help[] =
2227− " ENTER Togle callchains (if present) \n"
2228− " n Togle Node details info \n"
2229− " s Togle full lenght of symbol and source line columns \n"
2224+ " ENTER Toggle callchains (if present) \n"
2225+ " n Toggle Node details info \n"
2226+ " s Toggle full length of symbol and source line columns \n"
22302227 " q Return back to cacheline list \n";
22312228
22322229 /* Display compact version first. */
@@ -2303,7 +2300,7 @@ static int perf_c2c__hists_browse(struct hists *hists) · 70 unmodified lines
23032300 int key = -1;
23042301 const char help[] =
23052302 " d Display cacheline details \n"
2306− " ENTER Togle callchains (if present) \n"
2303+ " ENTER Toggle callchains (if present) \n"
23072304 " q Quit \n";
23082305
23092306 browser = perf_c2c_browser__new(hists);
@@ -2393,9 +2390,10 @@ static int setup_callchain(struct perf_evlist *evlist) · 83 unmodified lines
23932390 enum perf_call_graph_mode mode = CALLCHAIN_NONE;
23942391
23952392 if ((sample_type & PERF_SAMPLE_REGS_USER) &&
2396− (sample_type & PERF_SAMPLE_STACK_USER))
2393+ (sample_type & PERF_SAMPLE_STACK_USER)) {
23972394 mode = CALLCHAIN_DWARF;
2398− else if (sample_type & PERF_SAMPLE_BRANCH_STACK)
2395+ dwarf_callchain_users = true;
2396+ } else if (sample_type & PERF_SAMPLE_BRANCH_STACK)
23992397 mode = CALLCHAIN_LBR;
24002398 else if (sample_type & PERF_SAMPLE_CALLCHAIN)
24012399 mode = CALLCHAIN_FP;
tools/perf/builtin-help.c+1−1
@@ -439,7 +439,7 @@ int cmd_help(int argc, const char **argv) · 438 unmodified lines
439439 #ifdef HAVE_LIBELF_SUPPORT
440440 "probe",
441441 #endif
442−#ifdef HAVE_LIBAUDIT_SUPPORT
442+#if defined(HAVE_LIBAUDIT_SUPPORT) || defined(HAVE_SYSCALL_TABLE)
443443 "trace",
444444 #endif
445445 NULL };
tools/perf/builtin-inject.c+1−2
@@ -536,8 +536,7 @@ found: · 535 unmodified lines
536536 sample_sw.period = sample->period;
537537 sample_sw.time = sample->time;
538538 perf_event__synthesize_sample(event_sw, evsel->attr.sample_type,
539− evsel->attr.read_format, &sample_sw,
540− false);
539+ evsel->attr.read_format, &sample_sw);
541540 build_id__mark_dso_hit(tool, event_sw, &sample_sw, evsel, machine);
542541 return perf_event__repipe(tool, event_sw, &sample_sw, machine);
543542 }
tools/perf/builtin-kvm.c+8−5
@@ -26,6 +26,9 @@ · 25 unmodified lines
2626 #include <sys/timerfd.h>
2727 #endif
2828 #include <sys/time.h>
29+#include <sys/types.h>
30+#include <sys/stat.h>
31+#include <fcntl.h>
2932
3033 #include <linux/kernel.h>
3134 #include <linux/time64.h>
@@ -741,20 +744,20 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx, · 709 unmodified lines
741744 u64 *mmap_time)
742745 {
743746 union perf_event *event;
744− struct perf_sample sample;
747+ u64 timestamp;
745748 s64 n = 0;
746749 int err;
747750
748751 *mmap_time = ULLONG_MAX;
749752 while ((event = perf_evlist__mmap_read(kvm->evlist, idx)) != NULL) {
750− err = perf_evlist__parse_sample(kvm->evlist, event, &sample);
753+ err = perf_evlist__parse_sample_timestamp(kvm->evlist, event, &timestamp);
751754 if (err) {
752755 perf_evlist__mmap_consume(kvm->evlist, idx);
753756 pr_err("Failed to parse sample\n");
754757 return -1;
755758 }
756759
757− err = perf_session__queue_event(kvm->session, event, &sample, 0);
760+ err = perf_session__queue_event(kvm->session, event, timestamp, 0);
758761 /*
759762 * FIXME: Here we can't consume the event, as perf_session__queue_event will
760763 * point to it, and it'll get possibly overwritten by the kernel.
@@ -768,7 +771,7 @@ static s64 perf_kvm__mmap_read_idx(struct perf_kvm_stat *kvm, int idx, · 7 unmodified lines
768771
769772 /* save time stamp of our first sample for this mmap */
770773 if (n == 0)
771− *mmap_time = sample.time;
774+ *mmap_time = timestamp;
772775
773776 /* limit events per mmap handled all at once */
774777 n++;
@@ -1044,7 +1047,7 @@ static int kvm_live_open_events(struct perf_kvm_stat *kvm) · 269 unmodified lines
10441047 goto out;
10451048 }
10461049
1047− if (perf_evlist__mmap(evlist, kvm->opts.mmap_pages, false) < 0) {
1050+ if (perf_evlist__mmap(evlist, kvm->opts.mmap_pages) < 0) {
10481051 ui__error("Failed to mmap the events: %s\n",
10491052 str_error_r(errno, sbuf, sizeof(sbuf)));
10501053 perf_evlist__close(evlist);
tools/perf/builtin-record.c+49−14
@@ -51,7 +51,6 @@ · 50 unmodified lines
5151 #include <signal.h>
5252 #include <sys/mman.h>
5353 #include <sys/wait.h>
54−#include <asm/bug.h>
5554 #include <linux/time64.h>
5655
5756 struct switch_output {
@@ -79,6 +78,7 @@ struct record { · 21 unmodified lines
7978 bool no_buildid_cache_set;
8079 bool buildid_all;
8180 bool timestamp_filename;
81+ bool timestamp_boundary;
8282 struct switch_output switch_output;
8383 unsigned long long samples;
8484 };
@@ -301,7 +301,7 @@ static int record__mmap_evlist(struct record *rec, · 216 unmodified lines
301301 struct record_opts *opts = &rec->opts;
302302 char msg[512];
303303
304− if (perf_evlist__mmap_ex(evlist, opts->mmap_pages, false,
304+ if (perf_evlist__mmap_ex(evlist, opts->mmap_pages,
305305 opts->auxtrace_mmap_pages,
306306 opts->auxtrace_snapshot_mode) < 0) {
307307 if (errno == EPERM) {
@@ -372,6 +372,8 @@ try_again: · 64 unmodified lines
372372 ui__error("%s\n", msg);
373373 goto out;
374374 }
375+
376+ pos->supported = true;
375377 }
376378
377379 if (perf_evlist__apply_filters(evlist, &pos)) {
@@ -408,8 +410,15 @@ static int process_sample_event(struct perf_tool *tool, · 30 unmodified lines
408410 {
409411 struct record *rec = container_of(tool, struct record, tool);
410412
411− rec->samples++;
413+ if (rec->evlist->first_sample_time == 0)
414+ rec->evlist->first_sample_time = sample->time;
415+
416+ rec->evlist->last_sample_time = sample->time;
412417
418+ if (rec->buildid_all)
419+ return 0;
420+
421+ rec->samples++;
413422 return build_id__mark_dso_hit(tool, event, sample, evsel, machine);
414423 }
415424
@@ -434,9 +443,11 @@ static int process_buildids(struct record *rec) · 18 unmodified lines
434443
435444 /*
436445 * If --buildid-all is given, it marks all DSO regardless of hits,
437− * so no need to process samples.
446+ * so no need to process samples. But if timestamp_boundary is enabled,
447+ * it still needs to walk on all samples to get the timestamps of
448+ * first/last samples.
438449 */
439− if (rec->buildid_all)
450+ if (rec->buildid_all && !rec->timestamp_boundary)
440451 rec->tool.sample = NULL;
441452
442453 return perf_session__process_events(session);
@@ -477,7 +488,7 @@ static struct perf_event_header finished_round_event = { · 34 unmodified lines
477488 };
478489
479490 static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evlist,
480− bool backward)
491+ bool overwrite)
481492 {
482493 u64 bytes_written = rec->bytes_written;
483494 int i;
@@ -487,18 +498,18 @@ static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evli · 3 unmodified lines
487498 if (!evlist)
488499 return 0;
489500
490− maps = backward ? evlist->backward_mmap : evlist->mmap;
501+ maps = overwrite ? evlist->overwrite_mmap : evlist->mmap;
491502 if (!maps)
492503 return 0;
493504
494− if (backward && evlist->bkw_mmap_state != BKW_MMAP_DATA_PENDING)
505+ if (overwrite && evlist->bkw_mmap_state != BKW_MMAP_DATA_PENDING)
495506 return 0;
496507
497508 for (i = 0; i < evlist->nr_mmaps; i++) {
498509 struct auxtrace_mmap *mm = &maps[i].auxtrace_mmap;
499510
500511 if (maps[i].base) {
501− if (perf_mmap__push(&maps[i], evlist->overwrite, backward, rec, record__pushfn) != 0) {
512+ if (perf_mmap__push(&maps[i], overwrite, rec, record__pushfn) != 0) {
502513 rc = -1;
503514 goto out;
504515 }
@@ -518,7 +529,7 @@ static int record__mmap_read_evlist(struct record *rec, struct perf_evlist *evli · 13 unmodified lines
518529 if (bytes_written != rec->bytes_written)
519530 rc = record__write(rec, &finished_round_event, sizeof(finished_round_event));
520531
521− if (backward)
532+ if (overwrite)
522533 perf_evlist__toggle_bkw_mmap(evlist, BKW_MMAP_EMPTY);
523534 out:
524535 return rc;
@@ -690,8 +701,8 @@ perf_evlist__pick_pc(struct perf_evlist *evlist) · 165 unmodified lines
690701 if (evlist) {
691702 if (evlist->mmap && evlist->mmap[0].base)
692703 return evlist->mmap[0].base;
693− if (evlist->backward_mmap && evlist->backward_mmap[0].base)
694− return evlist->backward_mmap[0].base;
704+ if (evlist->overwrite_mmap && evlist->overwrite_mmap[0].base)
705+ return evlist->overwrite_mmap[0].base;
695706 }
696707 return NULL;
697708 }
@@ -784,6 +795,28 @@ static int record__synthesize(struct record *rec, bool tail) · 86 unmodified lines
784795 perf_event__synthesize_guest_os, tool);
785796 }
786797
798+ err = perf_event__synthesize_extra_attr(&rec->tool,
799+ rec->evlist,
800+ process_synthesized_event,
801+ data->is_pipe);
802+ if (err)
803+ goto out;
804+
805+ err = perf_event__synthesize_thread_map2(&rec->tool, rec->evlist->threads,
806+ process_synthesized_event,
807+ NULL);
808+ if (err < 0) {
809+ pr_err("Couldn't synthesize thread map.\n");
810+ return err;
811+ }
812+
813+ err = perf_event__synthesize_cpu_map(&rec->tool, rec->evlist->cpus,
814+ process_synthesized_event, NULL);
815+ if (err < 0) {
816+ pr_err("Couldn't synthesize cpu map.\n");
817+ return err;
818+ }
819+
787820 err = __machine__synthesize_threads(machine, tool, &opts->target, rec->evlist->threads,
788821 process_synthesized_event, opts->sample_address,
789822 opts->proc_map_timeout, 1);
@@ -1598,6 +1631,8 @@ static struct option __record_options[] = { · 808 unmodified lines
15981631 "Record build-id of all DSOs regardless of hits"),
15991632 OPT_BOOLEAN(0, "timestamp-filename", &record.timestamp_filename,
16001633 "append timestamp to output filename"),
1634+ OPT_BOOLEAN(0, "timestamp-boundary", &record.timestamp_boundary,
1635+ "Record timestamp boundary (time of first/last samples)"),
16011636 OPT_STRING_OPTARG_SET(0, "switch-output", &record.switch_output.str,
16021637 &record.switch_output.set, "signal,size,time",
16031638 "Switch output when receive SIGUSR2 or cross size,time threshold",
@@ -1781,8 +1816,8 @@ int cmd_record(int argc, const char **argv) · 177 unmodified lines
17811816 goto out;
17821817 }
17831818
1784− /* Enable ignoring missing threads when -u option is defined. */
1785− rec->opts.ignore_missing_thread = rec->opts.target.uid != UINT_MAX;
1819+ /* Enable ignoring missing threads when -u/-p option is defined. */
1820+ rec->opts.ignore_missing_thread = rec->opts.target.uid != UINT_MAX || rec->opts.target.pid;
17861821
17871822 err = -ENOMEM;
17881823 if (perf_evlist__create_maps(rec->evlist, &rec->opts.target) < 0)
tools/perf/builtin-report.c+267−16
@@ -15,6 +15,7 @@ · 14 unmodified lines
1515 #include "util/color.h"
1616 #include <linux/list.h>
1717 #include <linux/rbtree.h>
18+#include <linux/err.h>
1819 #include "util/symbol.h"
1920 #include "util/callchain.h"
2021 #include "util/values.h"
@@ -51,6 +52,7 @@ · 30 unmodified lines
5152 #include <sys/types.h>
5253 #include <sys/stat.h>
5354 #include <unistd.h>
55+#include <linux/mman.h>
5456
5557 struct report {
5658 struct perf_tool tool;
@@ -60,6 +62,9 @@ struct report { · 3 unmodified lines
6062 bool show_threads;
6163 bool inverted_callchain;
6264 bool mem_mode;
65+ bool stats_mode;
66+ bool tasks_mode;
67+ bool mmaps_mode;
6368 bool header;
6469 bool header_only;
6570 bool nonany_branch_mode;
@@ -69,7 +74,9 @@ struct report { · 3 unmodified lines
6974 const char *cpu_list;
7075 const char *symbol_filter_str;
7176 const char *time_str;
72− struct perf_time_interval ptime;
77+ struct perf_time_interval *ptime_range;
78+ int range_size;
79+ int range_num;
7380 float min_percent;
7481 u64 nr_entries;
7582 u64 queue_size;
@@ -162,12 +169,28 @@ static int hist_iter__branch_callback(struct hist_entry_iter *iter, · 86 unmodified lines
162169 struct hist_entry *he = iter->he;
163170 struct report *rep = arg;
164171 struct branch_info *bi;
172+ struct perf_sample *sample = iter->sample;
173+ struct perf_evsel *evsel = iter->evsel;
174+ int err;
175+
176+ if (!ui__has_annotation())
177+ return 0;
178+
179+ hist__account_cycles(sample->branch_stack, al, sample,
180+ rep->nonany_branch_mode);
165181
166182 bi = he->branch_info;
183+ err = addr_map_symbol__inc_samples(&bi->from, sample, evsel->idx);
184+ if (err)
185+ goto out;
186+
187+ err = addr_map_symbol__inc_samples(&bi->to, sample, evsel->idx);
188+
167189 branch_type_count(&rep->brtype_stat, &bi->flags,
168190 bi->from.addr, bi->to.addr);
169191
170− return 0;
192+out:
193+ return err;
171194 }
172195
173196 static int process_sample_event(struct perf_tool *tool,
@@ -186,8 +209,10 @@ static int process_sample_event(struct perf_tool *tool, · 12 unmodified lines
186209 };
187210 int ret = 0;
188211
189− if (perf_time__skip_sample(&rep->ptime, sample->time))
212+ if (perf_time__ranges_skip_sample(rep->ptime_range, rep->range_num,
213+ sample->time)) {
190214 return 0;
215+ }
191216
192217 if (machine__resolve(machine, &al, sample) < 0) {
193218 pr_debug("problem processing %d event, skipping it.\n",
@@ -312,9 +337,10 @@ static int report__setup_sample_type(struct report *rep) · 118 unmodified lines
312337
313338 if (symbol_conf.use_callchain || symbol_conf.cumulate_callchain) {
314339 if ((sample_type & PERF_SAMPLE_REGS_USER) &&
315− (sample_type & PERF_SAMPLE_STACK_USER))
340+ (sample_type & PERF_SAMPLE_STACK_USER)) {
316341 callchain_param.record_mode = CALLCHAIN_DWARF;
317− else if (sample_type & PERF_SAMPLE_BRANCH_STACK)
342+ dwarf_callchain_users = true;
343+ } else if (sample_type & PERF_SAMPLE_BRANCH_STACK)
318344 callchain_param.record_mode = CALLCHAIN_LBR;
319345 else
320346 callchain_param.record_mode = CALLCHAIN_FP;
@@ -377,6 +403,9 @@ static size_t hists__fprintf_nr_sample_events(struct hists *hists, struct report · 56 unmodified lines
377403 if (evname != NULL)
378404 ret += fprintf(fp, " of event '%s'", evname);
379405
406+ if (rep->time_str)
407+ ret += fprintf(fp, " (time slices: %s)", rep->time_str);
408+
380409 if (symbol_conf.show_ref_callgraph &&
381410 strstr(evname, "call-graph=no")) {
382411 ret += fprintf(fp, ", show reference callgraph");
@@ -567,6 +596,174 @@ static void report__output_resort(struct report *rep) · 184 unmodified lines
567596 ui_progress__finish();
568597 }
569598
599+static void stats_setup(struct report *rep)
600+{
601+ memset(&rep->tool, 0, sizeof(rep->tool));
602+ rep->tool.no_warn = true;
603+}
604+
605+static int stats_print(struct report *rep)
606+{
607+ struct perf_session *session = rep->session;
608+
609+ perf_session__fprintf_nr_events(session, stdout);
610+ return 0;
611+}
612+
613+static void tasks_setup(struct report *rep)
614+{
615+ memset(&rep->tool, 0, sizeof(rep->tool));
616+ if (rep->mmaps_mode) {
617+ rep->tool.mmap = perf_event__process_mmap;
618+ rep->tool.mmap2 = perf_event__process_mmap2;
619+ }
620+ rep->tool.comm = perf_event__process_comm;
621+ rep->tool.exit = perf_event__process_exit;
622+ rep->tool.fork = perf_event__process_fork;
623+ rep->tool.no_warn = true;
624+}
625+
626+struct task {
627+ struct thread *thread;
628+ struct list_head list;
629+ struct list_head children;
630+};
631+
632+static struct task *tasks_list(struct task *task, struct machine *machine)
633+{
634+ struct thread *parent_thread, *thread = task->thread;
635+ struct task *parent_task;
636+
637+ /* Already listed. */
638+ if (!list_empty(&task->list))
639+ return NULL;
640+
641+ /* Last one in the chain. */
642+ if (thread->ppid == -1)
643+ return task;
644+
645+ parent_thread = machine__find_thread(machine, -1, thread->ppid);
646+ if (!parent_thread)
647+ return ERR_PTR(-ENOENT);
648+
649+ parent_task = thread__priv(parent_thread);
650+ list_add_tail(&task->list, &parent_task->children);
651+ return tasks_list(parent_task, machine);
652+}
653+
654+static size_t maps__fprintf_task(struct maps *maps, int indent, FILE *fp)
655+{
656+ size_t printed = 0;
657+ struct rb_node *nd;
658+
659+ for (nd = rb_first(&maps->entries); nd; nd = rb_next(nd)) {
660+ struct map *map = rb_entry(nd, struct map, rb_node);
661+
662+ printed += fprintf(fp, "%*s %" PRIx64 "-%" PRIx64 " %c%c%c%c %08" PRIx64 " %" PRIu64 " %s\n",
663+ indent, "", map->start, map->end,
664+ map->prot & PROT_READ ? 'r' : '-',
665+ map->prot & PROT_WRITE ? 'w' : '-',
666+ map->prot & PROT_EXEC ? 'x' : '-',
667+ map->flags & MAP_SHARED ? 's' : 'p',
668+ map->pgoff,
669+ map->ino, map->dso->name);
670+ }
671+
672+ return printed;
673+}
674+
675+static int map_groups__fprintf_task(struct map_groups *mg, int indent, FILE *fp)
676+{
677+ int printed = 0, i;
678+ for (i = 0; i < MAP__NR_TYPES; ++i)
679+ printed += maps__fprintf_task(&mg->maps[i], indent, fp);
680+ return printed;
681+}
682+
683+static void task__print_level(struct task *task, FILE *fp, int level)
684+{
685+ struct thread *thread = task->thread;
686+ struct task *child;
687+ int comm_indent = fprintf(fp, " %8d %8d %8d |%*s",
688+ thread->pid_, thread->tid, thread->ppid,
689+ level, "");
690+
691+ fprintf(fp, "%s\n", thread__comm_str(thread));
692+
693+ map_groups__fprintf_task(thread->mg, comm_indent, fp);
694+
695+ if (!list_empty(&task->children)) {
696+ list_for_each_entry(child, &task->children, list)
697+ task__print_level(child, fp, level + 1);
698+ }
699+}
700+
701+static int tasks_print(struct report *rep, FILE *fp)
702+{
703+ struct perf_session *session = rep->session;
704+ struct machine *machine = &session->machines.host;
705+ struct task *tasks, *task;
706+ unsigned int nr = 0, itask = 0, i;
707+ struct rb_node *nd;
708+ LIST_HEAD(list);
709+
710+ /*
711+ * No locking needed while accessing machine->threads,
712+ * because --tasks is single threaded command.
713+ */
714+
715+ /* Count all the threads. */
716+ for (i = 0; i < THREADS__TABLE_SIZE; i++)
717+ nr += machine->threads[i].nr;
718+
719+ tasks = malloc(sizeof(*tasks) * nr);
720+ if (!tasks)
721+ return -ENOMEM;
722+
723+ for (i = 0; i < THREADS__TABLE_SIZE; i++) {
724+ struct threads *threads = &machine->threads[i];
725+
726+ for (nd = rb_first(&threads->entries); nd; nd = rb_next(nd)) {
727+ task = tasks + itask++;
728+
729+ task->thread = rb_entry(nd, struct thread, rb_node);
730+ INIT_LIST_HEAD(&task->children);
731+ INIT_LIST_HEAD(&task->list);
732+ thread__set_priv(task->thread, task);
733+ }
734+ }
735+
736+ /*
737+ * Iterate every task down to the unprocessed parent
738+ * and link all in task children list. Task with no
739+ * parent is added into 'list'.
740+ */
741+ for (itask = 0; itask < nr; itask++) {
742+ task = tasks + itask;
743+
744+ if (!list_empty(&task->list))
745+ continue;
746+
747+ task = tasks_list(task, machine);
748+ if (IS_ERR(task)) {
749+ pr_err("Error: failed to process tasks\n");
750+ free(tasks);
751+ return PTR_ERR(task);
752+ }
753+
754+ if (task)
755+ list_add_tail(&task->list, &list);
756+ }
757+
758+ fprintf(fp, "# %8s %8s %8s %s\n", "pid", "tid", "ppid", "comm");
759+
760+ list_for_each_entry(task, &list, list)
761+ task__print_level(task, fp, 0);
762+
763+ free(tasks);
764+ return 0;
765+}
766+
570767 static int __cmd_report(struct report *rep)
571768 {
572769 int ret;
@@ -598,12 +795,24 @@ static int __cmd_report(struct report *rep) · 25 unmodified lines
598795 return ret;
599796 }
600797
798+ if (rep->stats_mode)
799+ stats_setup(rep);
800+
801+ if (rep->tasks_mode)
802+ tasks_setup(rep);
803+
601804 ret = perf_session__process_events(session);
602805 if (ret) {
603806 ui__error("failed to process sample\n");
604807 return ret;
605808 }
606809
810+ if (rep->stats_mode)
811+ return stats_print(rep);
812+
813+ if (rep->tasks_mode)
814+ return tasks_print(rep, stdout);
815+
607816 report__warn_kptr_restrict(rep);
608817
609818 evlist__for_each_entry(session->evlist, pos)
@@ -760,6 +969,9 @@ int cmd_report(int argc, const char **argv) · 150 unmodified lines
760969 OPT_BOOLEAN('q', "quiet", &quiet, "Do not show any message"),
761970 OPT_BOOLEAN('D', "dump-raw-trace", &dump_trace,
762971 "dump raw trace in ASCII"),
972+ OPT_BOOLEAN(0, "stats", &report.stats_mode, "Display event stats"),
973+ OPT_BOOLEAN(0, "tasks", &report.tasks_mode, "Display recorded tasks"),
974+ OPT_BOOLEAN(0, "mmaps", &report.mmaps_mode, "Display recorded tasks memory maps"),
763975 OPT_STRING('k', "vmlinux", &symbol_conf.vmlinux_name,
764976 "file", "vmlinux pathname"),
765977 OPT_STRING(0, "kallsyms", &symbol_conf.kallsyms_name,
@@ -907,6 +1119,9 @@ int cmd_report(int argc, const char **argv) · 141 unmodified lines
9071119 report.symbol_filter_str = argv[0];
9081120 }
9091121
1122+ if (report.mmaps_mode)
1123+ report.tasks_mode = true;
1124+
9101125 if (quiet)
9111126 perf_quiet_option();
9121127
@@ -921,13 +1136,6 @@ int cmd_report(int argc, const char **argv) · 8 unmodified lines
9211136 return -EINVAL;
9221137 }
9231138
924− if (report.use_stdio)
925− use_browser = 0;
926− else if (report.use_tui)
927− use_browser = 1;
928− else if (report.use_gtk)
929− use_browser = 2;
930−
9311139 if (report.inverted_callchain)
9321140 callchain_param.order = ORDER_CALLER;
9331141 if (symbol_conf.cumulate_callchain && !callchain_param.order_set)
@@ -1014,6 +1222,13 @@ repeat: · 80 unmodified lines
10141222 perf_hpp_list.need_collapse = true;
10151223 }
10161224
1225+ if (report.use_stdio)
1226+ use_browser = 0;
1227+ else if (report.use_tui)
1228+ use_browser = 1;
1229+ else if (report.use_gtk)
1230+ use_browser = 2;
1231+
10171232 /* Force tty output for header output and per-thread stat. */
10181233 if (report.header || report.header_only || report.show_threads)
10191234 use_browser = 0;
@@ -1021,6 +1236,12 @@ repeat: · 1 unmodified line
10211236 report.tool.show_feat_hdr = SHOW_FEAT_HEADER;
10221237 if (report.show_full_info)
10231238 report.tool.show_feat_hdr = SHOW_FEAT_HEADER_FULL_INFO;
1239+ if (report.stats_mode || report.tasks_mode)
1240+ use_browser = 0;
1241+ if (report.stats_mode && report.tasks_mode) {
1242+ pr_err("Error: --tasks and --mmaps can't be used together with --stats\n");
1243+ goto error;
1244+ }
10241245
10251246 if (strcmp(input_name, "-") != 0)
10261247 setup_browser(true);
@@ -1043,7 +1264,8 @@ repeat: · 16 unmodified lines
10431264 ret = 0;
10441265 goto error;
10451266 }
1046− } else if (use_browser == 0 && !quiet) {
1267+ } else if (use_browser == 0 && !quiet &&
1268+ !report.stats_mode && !report.tasks_mode) {
10471269 fputs("# To display the perf.data header info, please use --header/--header-only options.\n#\n",
10481270 stdout);
10491271 }
@@ -1077,9 +1299,36 @@ repeat: · 27 unmodified lines
10771299 if (symbol__init(&session->header.env) < 0)
10781300 goto error;
10791301
1080− if (perf_time__parse_str(&report.ptime, report.time_str) != 0) {
1081− pr_err("Invalid time string\n");
1082− return -EINVAL;
1302+ report.ptime_range = perf_time__range_alloc(report.time_str,
1303+ &report.range_size);
1304+ if (!report.ptime_range) {
1305+ ret = -ENOMEM;
1306+ goto error;
1307+ }
1308+
1309+ if (perf_time__parse_str(report.ptime_range, report.time_str) != 0) {
1310+ if (session->evlist->first_sample_time == 0 &&
1311+ session->evlist->last_sample_time == 0) {
1312+ pr_err("HINT: no first/last sample time found in perf data.\n"
1313+ "Please use latest perf binary to execute 'perf record'\n"
1314+ "(if '--buildid-all' is enabled, please set '--timestamp-boundary').\n");
1315+ ret = -EINVAL;
1316+ goto error;
1317+ }
1318+
1319+ report.range_num = perf_time__percent_parse_str(
1320+ report.ptime_range, report.range_size,
1321+ report.time_str,
1322+ session->evlist->first_sample_time,
1323+ session->evlist->last_sample_time);
1324+
1325+ if (report.range_num < 0) {
1326+ pr_err("Invalid time string\n");
1327+ ret = -EINVAL;
1328+ goto error;
1329+ }
1330+ } else {
1331+ report.range_num = 1;
10831332 }
10841333
10851334 sort__setup_elide(stdout);
@@ -1092,6 +1341,8 @@ repeat: · 6 unmodified lines
10921341 ret = 0;
10931342
10941343 error:
1344+ zfree(&report.ptime_range);
1345+
10951346 perf_session__delete(session);
10961347 return ret;
10971348 }
tools/perf/builtin-script.c+229−37
@@ -22,9 +22,11 @@ · 21 unmodified lines
2222 #include "util/cpumap.h"
2323 #include "util/thread_map.h"
2424 #include "util/stat.h"
25+#include "util/color.h"
2526 #include "util/string2.h"
2627 #include "util/thread-stack.h"
2728 #include "util/time-utils.h"
29+#include "util/path.h"
2830 #include "print_binary.h"
2931 #include <linux/bitmap.h>
3032 #include <linux/kernel.h>
@@ -40,6 +42,7 @@ · 9 unmodified lines
4042 #include <sys/param.h>
4143 #include <sys/types.h>
4244 #include <sys/stat.h>
45+#include <fcntl.h>
4346 #include <unistd.h>
4447
4548 #include "sane_ctype.h"
@@ -90,6 +93,8 @@ enum perf_output_field { · 44 unmodified lines
9093 PERF_OUTPUT_SYNTH = 1U << 25,
9194 PERF_OUTPUT_PHYS_ADDR = 1U << 26,
9295 PERF_OUTPUT_UREGS = 1U << 27,
96+ PERF_OUTPUT_METRIC = 1U << 28,
97+ PERF_OUTPUT_MISC = 1U << 29,
9398 };
9499
95100 struct output_option {
@@ -124,6 +129,8 @@ struct output_option { · 28 unmodified lines
124129 {.str = "brstackoff", .field = PERF_OUTPUT_BRSTACKOFF},
125130 {.str = "synth", .field = PERF_OUTPUT_SYNTH},
126131 {.str = "phys_addr", .field = PERF_OUTPUT_PHYS_ADDR},
132+ {.str = "metric", .field = PERF_OUTPUT_METRIC},
133+ {.str = "misc", .field = PERF_OUTPUT_MISC},
127134 };
128135
129136 enum {
@@ -215,12 +222,20 @@ struct perf_evsel_script { · 85 unmodified lines
215222 char *filename;
216223 FILE *fp;
217224 u64 samples;
225+ /* For metric output */
226+ u64 val;
227+ int gnum;
218228 };
219229
230+static inline struct perf_evsel_script *evsel_script(struct perf_evsel *evsel)
231+{
232+ return (struct perf_evsel_script *)evsel->priv;
233+}
234+
220235 static struct perf_evsel_script *perf_evsel_script__new(struct perf_evsel *evsel,
221236 struct perf_data *data)
222237 {
223− struct perf_evsel_script *es = malloc(sizeof(*es));
238+ struct perf_evsel_script *es = zalloc(sizeof(*es));
224239
225240 if (es != NULL) {
226241 if (asprintf(&es->filename, "%s.%s.dump", data->file.path, perf_evsel__name(evsel)) < 0)
@@ -228,7 +243,6 @@ static struct perf_evsel_script *perf_evsel_script__new(struct perf_evsel *evsel · 1 unmodified line
228243 es->fp = fopen(es->filename, "w");
229244 if (es->fp == NULL)
230245 goto out_free_filename;
231− es->samples = 0;
232246 }
233247
234248 return es;
@@ -423,11 +437,6 @@ static int perf_evsel__check_attr(struct perf_evsel *evsel, · 188 unmodified lines
423437 PERF_OUTPUT_CPU, allow_user_set))
424438 return -EINVAL;
425439
426− if (PRINT_FIELD(PERIOD) &&
427− perf_evsel__check_stype(evsel, PERF_SAMPLE_PERIOD, "PERIOD",
428− PERF_OUTPUT_PERIOD))
429− return -EINVAL;
430−
431440 if (PRINT_FIELD(IREGS) &&
432441 perf_evsel__check_stype(evsel, PERF_SAMPLE_REGS_INTR, "IREGS",
433442 PERF_OUTPUT_IREGS))
@@ -588,7 +597,8 @@ static int perf_sample__fprintf_uregs(struct perf_sample *sample, · 154 unmodified lines
588597
589598 static int perf_sample__fprintf_start(struct perf_sample *sample,
590599 struct thread *thread,
591− struct perf_evsel *evsel, FILE *fp)
600+ struct perf_evsel *evsel,
601+ u32 type, FILE *fp)
592602 {
593603 struct perf_event_attr *attr = &evsel->attr;
594604 unsigned long secs;
@@ -618,6 +628,47 @@ static int perf_sample__fprintf_start(struct perf_sample *sample, · 23 unmodified lines
618628 printed += fprintf(fp, "[%03d] ", sample->cpu);
619629 }
620630
631+ if (PRINT_FIELD(MISC)) {
632+ int ret = 0;
633+
634+ #define has(m) \
635+ (sample->misc & PERF_RECORD_MISC_##m) == PERF_RECORD_MISC_##m
636+
637+ if (has(KERNEL))
638+ ret += fprintf(fp, "K");
639+ if (has(USER))
640+ ret += fprintf(fp, "U");
641+ if (has(HYPERVISOR))
642+ ret += fprintf(fp, "H");
643+ if (has(GUEST_KERNEL))
644+ ret += fprintf(fp, "G");
645+ if (has(GUEST_USER))
646+ ret += fprintf(fp, "g");
647+
648+ switch (type) {
649+ case PERF_RECORD_MMAP:
650+ case PERF_RECORD_MMAP2:
651+ if (has(MMAP_DATA))
652+ ret += fprintf(fp, "M");
653+ break;
654+ case PERF_RECORD_COMM:
655+ if (has(COMM_EXEC))
656+ ret += fprintf(fp, "E");
657+ break;
658+ case PERF_RECORD_SWITCH:
659+ case PERF_RECORD_SWITCH_CPU_WIDE:
660+ if (has(SWITCH_OUT))
661+ ret += fprintf(fp, "S");
662+ default:
663+ break;
664+ }
665+
666+ #undef has
667+
668+ ret += fprintf(fp, "%*s", 6 - ret, " ");
669+ printed += ret;
670+ }
671+
621672 if (PRINT_FIELD(TIME)) {
622673 nsecs = sample->time;
623674 secs = nsecs / NSEC_PER_SEC;
@@ -1437,13 +1488,16 @@ struct perf_script { · 813 unmodified lines
14371488 bool show_mmap_events;
14381489 bool show_switch_events;
14391490 bool show_namespace_events;
1491+ bool show_lost_events;
14401492 bool allocated;
14411493 bool per_event_dump;
14421494 struct cpu_map *cpus;
14431495 struct thread_map *threads;
14441496 int name_width;
14451497 const char *time_str;
1446− struct perf_time_interval ptime;
1498+ struct perf_time_interval *ptime_range;
1499+ int range_size;
1500+ int range_num;
14471501 };
14481502
14491503 static int perf_evlist__max_name_len(struct perf_evlist *evlist)
@@ -1477,6 +1531,88 @@ static int data_src__fprintf(u64 data_src, FILE *fp) · 27 unmodified lines
14771531 return fprintf(fp, "%-*s", maxlen, out);
14781532 }
14791533
1534+struct metric_ctx {
1535+ struct perf_sample *sample;
1536+ struct thread *thread;
1537+ struct perf_evsel *evsel;
1538+ FILE *fp;
1539+};
1540+
1541+static void script_print_metric(void *ctx, const char *color,
1542+ const char *fmt,
1543+ const char *unit, double val)
1544+{
1545+ struct metric_ctx *mctx = ctx;
1546+
1547+ if (!fmt)
1548+ return;
1549+ perf_sample__fprintf_start(mctx->sample, mctx->thread, mctx->evsel,
1550+ PERF_RECORD_SAMPLE, mctx->fp);
1551+ fputs("\tmetric: ", mctx->fp);
1552+ if (color)
1553+ color_fprintf(mctx->fp, color, fmt, val);
1554+ else
1555+ printf(fmt, val);
1556+ fprintf(mctx->fp, " %s\n", unit);
1557+}
1558+
1559+static void script_new_line(void *ctx)
1560+{
1561+ struct metric_ctx *mctx = ctx;
1562+
1563+ perf_sample__fprintf_start(mctx->sample, mctx->thread, mctx->evsel,
1564+ PERF_RECORD_SAMPLE, mctx->fp);
1565+ fputs("\tmetric: ", mctx->fp);
1566+}
1567+
1568+static void perf_sample__fprint_metric(struct perf_script *script,
1569+ struct thread *thread,
1570+ struct perf_evsel *evsel,
1571+ struct perf_sample *sample,
1572+ FILE *fp)
1573+{
1574+ struct perf_stat_output_ctx ctx = {
1575+ .print_metric = script_print_metric,
1576+ .new_line = script_new_line,
1577+ .ctx = &(struct metric_ctx) {
1578+ .sample = sample,
1579+ .thread = thread,
1580+ .evsel = evsel,
1581+ .fp = fp,
1582+ },
1583+ .force_header = false,
1584+ };
1585+ struct perf_evsel *ev2;
1586+ static bool init;
1587+ u64 val;
1588+
1589+ if (!init) {
1590+ perf_stat__init_shadow_stats();
1591+ init = true;
1592+ }
1593+ if (!evsel->stats)
1594+ perf_evlist__alloc_stats(script->session->evlist, false);
1595+ if (evsel_script(evsel->leader)->gnum++ == 0)
1596+ perf_stat__reset_shadow_stats();
1597+ val = sample->period * evsel->scale;
1598+ perf_stat__update_shadow_stats(evsel,
1599+ val,
1600+ sample->cpu,
1601+ &rt_stat);
1602+ evsel_script(evsel)->val = val;
1603+ if (evsel_script(evsel->leader)->gnum == evsel->leader->nr_members) {
1604+ for_each_group_member (ev2, evsel->leader) {
1605+ perf_stat__print_shadow_stats(ev2,
1606+ evsel_script(ev2)->val,
1607+ sample->cpu,
1608+ &ctx,
1609+ NULL,
1610+ &rt_stat);
1611+ }
1612+ evsel_script(evsel->leader)->gnum = 0;
1613+ }
1614+}
1615+
14801616 static void process_event(struct perf_script *script,
14811617 struct perf_sample *sample, struct perf_evsel *evsel,
14821618 struct addr_location *al,
@@ -1493,7 +1629,8 @@ static void process_event(struct perf_script *script, · 10 unmodified lines
14931629
14941630 ++es->samples;
14951631
1496− perf_sample__fprintf_start(sample, thread, evsel, fp);
1632+ perf_sample__fprintf_start(sample, thread, evsel,
1633+ PERF_RECORD_SAMPLE, fp);
14971634
14981635 if (PRINT_FIELD(PERIOD))
14991636 fprintf(fp, "%10" PRIu64 " ", sample->period);
@@ -1564,6 +1701,9 @@ static void process_event(struct perf_script *script, · 64 unmodified lines
15641701 if (PRINT_FIELD(PHYS_ADDR))
15651702 fprintf(fp, "%16" PRIx64, sample->phys_addr);
15661703 fprintf(fp, "\n");
1704+
1705+ if (PRINT_FIELD(METRIC))
1706+ perf_sample__fprint_metric(script, thread, evsel, sample, fp);
15671707 }
15681708
15691709 static struct scripting_ops *scripting_ops;
@@ -1643,8 +1783,10 @@ static int process_sample_event(struct perf_tool *tool, · 73 unmodified lines
16431783 struct perf_script *scr = container_of(tool, struct perf_script, tool);
16441784 struct addr_location al;
16451785
1646− if (perf_time__skip_sample(&scr->ptime, sample->time))
1786+ if (perf_time__ranges_skip_sample(scr->ptime_range, scr->range_num,
1787+ sample->time)) {
16471788 return 0;
1789+ }
16481790
16491791 if (debug_mode) {
16501792 if (sample->time < last_timestamp) {
@@ -1737,7 +1879,8 @@ static int process_comm_event(struct perf_tool *tool, · 86 unmodified lines
17371879 sample->tid = event->comm.tid;
17381880 sample->pid = event->comm.pid;
17391881 }
1740− perf_sample__fprintf_start(sample, thread, evsel, stdout);
1882+ perf_sample__fprintf_start(sample, thread, evsel,
1883+ PERF_RECORD_COMM, stdout);
17411884 perf_event__fprintf(event, stdout);
17421885 ret = 0;
17431886 out:
@@ -1772,7 +1915,8 @@ static int process_namespaces_event(struct perf_tool *tool, · 28 unmodified lines
17721915 sample->tid = event->namespaces.tid;
17731916 sample->pid = event->namespaces.pid;
17741917 }
1775− perf_sample__fprintf_start(sample, thread, evsel, stdout);
1918+ perf_sample__fprintf_start(sample, thread, evsel,
1919+ PERF_RECORD_NAMESPACES, stdout);
17761920 perf_event__fprintf(event, stdout);
17771921 ret = 0;
17781922 out:
@@ -1805,7 +1949,8 @@ static int process_fork_event(struct perf_tool *tool, · 26 unmodified lines
18051949 sample->tid = event->fork.tid;
18061950 sample->pid = event->fork.pid;
18071951 }
1808− perf_sample__fprintf_start(sample, thread, evsel, stdout);
1952+ perf_sample__fprintf_start(sample, thread, evsel,
1953+ PERF_RECORD_FORK, stdout);
18091954 perf_event__fprintf(event, stdout);
18101955 thread__put(thread);
18111956
@@ -1834,7 +1979,8 @@ static int process_exit_event(struct perf_tool *tool, · 22 unmodified lines
18341979 sample->tid = event->fork.tid;
18351980 sample->pid = event->fork.pid;
18361981 }
1837− perf_sample__fprintf_start(sample, thread, evsel, stdout);
1982+ perf_sample__fprintf_start(sample, thread, evsel,
1983+ PERF_RECORD_EXIT, stdout);
18381984 perf_event__fprintf(event, stdout);
18391985
18401986 if (perf_event__process_exit(tool, event, sample, machine) < 0)
@@ -1869,7 +2015,8 @@ static int process_mmap_event(struct perf_tool *tool, · 28 unmodified lines
18692015 sample->tid = event->mmap.tid;
18702016 sample->pid = event->mmap.pid;
18712017 }
1872− perf_sample__fprintf_start(sample, thread, evsel, stdout);
2018+ perf_sample__fprintf_start(sample, thread, evsel,
2019+ PERF_RECORD_MMAP, stdout);
18732020 perf_event__fprintf(event, stdout);
18742021 thread__put(thread);
18752022 return 0;
@@ -1900,7 +2047,8 @@ static int process_mmap2_event(struct perf_tool *tool, · 24 unmodified lines
19002047 sample->tid = event->mmap2.tid;
19012048 sample->pid = event->mmap2.pid;
19022049 }
1903− perf_sample__fprintf_start(sample, thread, evsel, stdout);
2050+ perf_sample__fprintf_start(sample, thread, evsel,
2051+ PERF_RECORD_MMAP2, stdout);
19042052 perf_event__fprintf(event, stdout);
19052053 thread__put(thread);
19062054 return 0;
@@ -1926,7 +2074,31 @@ static int process_switch_event(struct perf_tool *tool, · 19 unmodified lines
19262074 return -1;
19272075 }
19282076
1929− perf_sample__fprintf_start(sample, thread, evsel, stdout);
2077+ perf_sample__fprintf_start(sample, thread, evsel,
2078+ PERF_RECORD_SWITCH, stdout);
2079+ perf_event__fprintf(event, stdout);
2080+ thread__put(thread);
2081+ return 0;
2082+}
2083+
2084+static int
2085+process_lost_event(struct perf_tool *tool,
2086+ union perf_event *event,
2087+ struct perf_sample *sample,
2088+ struct machine *machine)
2089+{
2090+ struct perf_script *script = container_of(tool, struct perf_script, tool);
2091+ struct perf_session *session = script->session;
2092+ struct perf_evsel *evsel = perf_evlist__id2evsel(session->evlist, sample->id);
2093+ struct thread *thread;
2094+
2095+ thread = machine__findnew_thread(machine, sample->pid,
2096+ sample->tid);
2097+ if (thread == NULL)
2098+ return -1;
2099+
2100+ perf_sample__fprintf_start(sample, thread, evsel,
2101+ PERF_RECORD_LOST, stdout);
19302102 perf_event__fprintf(event, stdout);
19312103 thread__put(thread);
19322104 return 0;
@@ -2026,6 +2198,8 @@ static int __cmd_script(struct perf_script *script) · 93 unmodified lines
20262198 script->tool.context_switch = process_switch_event;
20272199 if (script->show_namespace_events)
20282200 script->tool.namespaces = process_namespaces_event;
2201+ if (script->show_lost_events)
2202+ script->tool.lost = process_lost_event;
20292203
20302204 if (perf_script__setup_per_event_dump(script)) {
20312205 pr_err("Couldn't create the per event dump files\n");
@@ -2311,19 +2485,6 @@ out: · 279 unmodified lines
23112485 return rc;
23122486 }
23132487
2314−/* Helper function for filesystems that return a dent->d_type DT_UNKNOWN */
2315−static int is_directory(const char *base_path, const struct dirent *dent)
2316−{
2317− char path[PATH_MAX];
2318− struct stat st;
2319−
2320− sprintf(path, "%s/%s", base_path, dent->d_name);
2321− if (stat(path, &st))
2322− return 0;
2323−
2324− return S_ISDIR(st.st_mode);
2325−}
2326−
23272488 #define for_each_lang(scripts_path, scripts_dir, lang_dirent) \
23282489 while ((lang_dirent = readdir(scripts_dir)) != NULL) \
23292490 if ((lang_dirent->d_type == DT_DIR || \
@@ -2758,9 +2919,10 @@ static void script__setup_sample_type(struct perf_script *script) · 428 unmodified lines
27582919
27592920 if (symbol_conf.use_callchain || symbol_conf.cumulate_callchain) {
27602921 if ((sample_type & PERF_SAMPLE_REGS_USER) &&
2761− (sample_type & PERF_SAMPLE_STACK_USER))
2922+ (sample_type & PERF_SAMPLE_STACK_USER)) {
27622923 callchain_param.record_mode = CALLCHAIN_DWARF;
2763− else if (sample_type & PERF_SAMPLE_BRANCH_STACK)
2924+ dwarf_callchain_users = true;
2925+ } else if (sample_type & PERF_SAMPLE_BRANCH_STACK)
27642926 callchain_param.record_mode = CALLCHAIN_LBR;
27652927 else
27662928 callchain_param.record_mode = CALLCHAIN_FP;
@@ -2975,6 +3137,8 @@ int cmd_script(int argc, const char **argv) · 208 unmodified lines
29753137 "Show context switch events (if recorded)"),
29763138 OPT_BOOLEAN('\0', "show-namespace-events", &script.show_namespace_events,
29773139 "Show namespace events (if recorded)"),
3140+ OPT_BOOLEAN('\0', "show-lost-events", &script.show_lost_events,
3141+ "Show lost events (if recorded)"),
29783142 OPT_BOOLEAN('\0', "per-event-dump", &script.per_event_dump,
29793143 "Dump trace output to files named by the monitored events"),
29803144 OPT_BOOLEAN('f', "force", &symbol_conf.force, "don't complain, do it"),
@@ -3281,18 +3445,46 @@ int cmd_script(int argc, const char **argv) · 300 unmodified lines
32813445 if (err < 0)
32823446 goto out_delete;
32833447
3284− /* needs to be parsed after looking up reference time */
3285− if (perf_time__parse_str(&script.ptime, script.time_str) != 0) {
3286− pr_err("Invalid time string\n");
3287− err = -EINVAL;
3448+ script.ptime_range = perf_time__range_alloc(script.time_str,
3449+ &script.range_size);
3450+ if (!script.ptime_range) {
3451+ err = -ENOMEM;
32883452 goto out_delete;
32893453 }
32903454
3455+ /* needs to be parsed after looking up reference time */
3456+ if (perf_time__parse_str(script.ptime_range, script.time_str) != 0) {
3457+ if (session->evlist->first_sample_time == 0 &&
3458+ session->evlist->last_sample_time == 0) {
3459+ pr_err("HINT: no first/last sample time found in perf data.\n"
3460+ "Please use latest perf binary to execute 'perf record'\n"
3461+ "(if '--buildid-all' is enabled, please set '--timestamp-boundary').\n");
3462+ err = -EINVAL;
3463+ goto out_delete;
3464+ }
3465+
3466+ script.range_num = perf_time__percent_parse_str(
3467+ script.ptime_range, script.range_size,
3468+ script.time_str,
3469+ session->evlist->first_sample_time,
3470+ session->evlist->last_sample_time);
3471+
3472+ if (script.range_num < 0) {
3473+ pr_err("Invalid time string\n");
3474+ err = -EINVAL;
3475+ goto out_delete;
3476+ }
3477+ } else {
3478+ script.range_num = 1;
3479+ }
3480+
32913481 err = __cmd_script(&script);
32923482
32933483 flush_scripting();
32943484
32953485 out_delete:
3486+ zfree(&script.ptime_range);
3487+
32963488 perf_evlist__free_stats(session->evlist);
32973489 perf_session__delete(session);
32983490
tools/perf/builtin-stat.c+147−83
@@ -63,7 +63,6 @@ · 62 unmodified lines
6363 #include "util/group.h"
6464 #include "util/session.h"
6565 #include "util/tool.h"
66−#include "util/group.h"
6766 #include "util/string2.h"
6867 #include "util/metricgroup.h"
6968 #include "asm/bug.h"
@@ -214,8 +213,13 @@ static inline void diff_timespec(struct timespec *r, struct timespec *a, · 144 unmodified lines
214213
215214 static void perf_stat__reset_stats(void)
216215 {
216+ int i;
217+
217218 perf_evlist__reset_stats(evsel_list);
218219 perf_stat__reset_shadow_stats();
220+
221+ for (i = 0; i < stat_config.stats_num; i++)
222+ perf_stat__reset_shadow_per_stat(&stat_config.stats[i]);
219223 }
220224
221225 static int create_perf_stat_counter(struct perf_evsel *evsel)
@@ -272,7 +276,7 @@ static int create_perf_stat_counter(struct perf_evsel *evsel) · 50 unmodified lines
272276 attr->enable_on_exec = 1;
273277 }
274278
275− if (target__has_cpu(&target))
279+ if (target__has_cpu(&target) && !target__has_per_thread(&target))
276280 return perf_evsel__open_per_cpu(evsel, perf_evsel__cpus(evsel));
277281
278282 return perf_evsel__open_per_thread(evsel, evsel_list->threads);
@@ -335,7 +339,7 @@ static int read_counter(struct perf_evsel *counter) · 56 unmodified lines
335339 int nthreads = thread_map__nr(evsel_list->threads);
336340 int ncpus, cpu, thread;
337341
338− if (target__has_cpu(&target))
342+ if (target__has_cpu(&target) && !target__has_per_thread(&target))
339343 ncpus = perf_evsel__nr_cpus(counter);
340344 else
341345 ncpus = 1;
@@ -458,19 +462,8 @@ static void workload_exec_failed_signal(int signo __maybe_unused, siginfo_t *inf · 116 unmodified lines
458462 workload_exec_errno = info->si_value.sival_int;
459463 }
460464
461−static bool has_unit(struct perf_evsel *counter)
462−{
463− return counter->unit && *counter->unit;
464−}
465−
466−static bool has_scale(struct perf_evsel *counter)
467−{
468− return counter->scale != 1;
469−}
470−
471465 static int perf_stat_synthesize_config(bool is_pipe)
472466 {
473− struct perf_evsel *counter;
474467 int err;
475468
476469 if (is_pipe) {
@@ -482,53 +475,10 @@ static int perf_stat_synthesize_config(bool is_pipe) · 5 unmodified lines
482475 }
483476 }
484477
485− /*
486− * Synthesize other events stuff not carried within
487− * attr event - unit, scale, name
488− */
489− evlist__for_each_entry(evsel_list, counter) {
490− if (!counter->supported)
491− continue;
492−
493− /*
494− * Synthesize unit and scale only if it's defined.
495− */
496− if (has_unit(counter)) {
497− err = perf_event__synthesize_event_update_unit(NULL, counter, process_synthesized_event);
498− if (err < 0) {
499− pr_err("Couldn't synthesize evsel unit.\n");
500− return err;
501− }
502− }
503−
504− if (has_scale(counter)) {
505− err = perf_event__synthesize_event_update_scale(NULL, counter, process_synthesized_event);
506− if (err < 0) {
507− pr_err("Couldn't synthesize evsel scale.\n");
508− return err;
509− }
510− }
511−
512− if (counter->own_cpus) {
513− err = perf_event__synthesize_event_update_cpus(NULL, counter, process_synthesized_event);
514− if (err < 0) {
515− pr_err("Couldn't synthesize evsel scale.\n");
516− return err;
517− }
518− }
519−
520− /*
521− * Name is needed only for pipe output,
522− * perf.data carries event names.
523− */
524− if (is_pipe) {
525− err = perf_event__synthesize_event_update_name(NULL, counter, process_synthesized_event);
526− if (err < 0) {
527− pr_err("Couldn't synthesize evsel name.\n");
528− return err;
529− }
530− }
531− }
478+ err = perf_event__synthesize_extra_attr(NULL,
479+ evsel_list,
480+ process_synthesized_event,
481+ is_pipe);
532482
533483 err = perf_event__synthesize_thread_map2(NULL, evsel_list->threads,
534484 process_synthesized_event,
@@ -1151,7 +1101,8 @@ static void abs_printout(int id, int nr, struct perf_evsel *evsel, double avg) · 616 unmodified lines
11511101 }
11521102
11531103 static void printout(int id, int nr, struct perf_evsel *counter, double uval,
1154− char *prefix, u64 run, u64 ena, double noise)
1104+ char *prefix, u64 run, u64 ena, double noise,
1105+ struct runtime_stat *st)
11551106 {
11561107 struct perf_stat_output_ctx out;
11571108 struct outstate os = {
@@ -1244,7 +1195,7 @@ static void printout(int id, int nr, struct perf_evsel *counter, double uval, · 86 unmodified lines
12441195
12451196 perf_stat__print_shadow_stats(counter, uval,
12461197 first_shadow_cpu(counter, id),
1247− &out, &metric_events);
1198+ &out, &metric_events, st);
12481199 if (!csv_output && !metric_only) {
12491200 print_noise(counter, noise);
12501201 print_running(run, ena);
@@ -1268,7 +1219,8 @@ static void aggr_update_shadow(void) · 17 unmodified lines
12681219 val += perf_counts(counter->counts, cpu, 0)->val;
12691220 }
12701221 perf_stat__update_shadow_stats(counter, val,
1271− first_shadow_cpu(counter, id));
1222+ first_shadow_cpu(counter, id),
1223+ &rt_stat);
12721224 }
12731225 }
12741226 }
@@ -1388,7 +1340,8 @@ static void print_aggr(char *prefix) · 113 unmodified lines
13881340 fprintf(output, "%s", prefix);
13891341
13901342 uval = val * counter->scale;
1391− printout(id, nr, counter, uval, prefix, run, ena, 1.0);
1343+ printout(id, nr, counter, uval, prefix, run, ena, 1.0,
1344+ &rt_stat);
13921345 if (!metric_only)
13931346 fputc('\n', output);
13941347 }
@@ -1397,13 +1350,24 @@ static void print_aggr(char *prefix) · 2 unmodified lines
13971350 }
13981351 }
13991352
1400−static void print_aggr_thread(struct perf_evsel *counter, char *prefix)
1353+static int cmp_val(const void *a, const void *b)
14011354 {
1402− FILE *output = stat_config.output;
1403− int nthreads = thread_map__nr(counter->threads);
1404− int ncpus = cpu_map__nr(counter->cpus);
1405− int cpu, thread;
1355+ return ((struct perf_aggr_thread_value *)b)->val -
1356+ ((struct perf_aggr_thread_value *)a)->val;
1357+}
1358+
1359+static struct perf_aggr_thread_value *sort_aggr_thread(
1360+ struct perf_evsel *counter,
1361+ int nthreads, int ncpus,
1362+ int *ret)
1363+{
1364+ int cpu, thread, i = 0;
14061365 double uval;
1366+ struct perf_aggr_thread_value *buf;
1367+
1368+ buf = calloc(nthreads, sizeof(struct perf_aggr_thread_value));
1369+ if (!buf)
1370+ return NULL;
14071371
14081372 for (thread = 0; thread < nthreads; thread++) {
14091373 u64 ena = 0, run = 0, val = 0;
@@ -1414,13 +1378,63 @@ static void print_aggr_thread(struct perf_evsel *counter, char *prefix) · 4 unmodified lines
14141378 run += perf_counts(counter->counts, cpu, thread)->run;
14151379 }
14161380
1381+ uval = val * counter->scale;
1382+
1383+ /*
1384+ * Skip value 0 when enabling --per-thread globally,
1385+ * otherwise too many 0 output.
1386+ */
1387+ if (uval == 0.0 && target__has_per_thread(&target))
1388+ continue;
1389+
1390+ buf[i].counter = counter;
1391+ buf[i].id = thread;
1392+ buf[i].uval = uval;
1393+ buf[i].val = val;
1394+ buf[i].run = run;
1395+ buf[i].ena = ena;
1396+ i++;
1397+ }
1398+
1399+ qsort(buf, i, sizeof(struct perf_aggr_thread_value), cmp_val);
1400+
1401+ if (ret)
1402+ *ret = i;
1403+
1404+ return buf;
1405+}
1406+
1407+static void print_aggr_thread(struct perf_evsel *counter, char *prefix)
1408+{
1409+ FILE *output = stat_config.output;
1410+ int nthreads = thread_map__nr(counter->threads);
1411+ int ncpus = cpu_map__nr(counter->cpus);
1412+ int thread, sorted_threads, id;
1413+ struct perf_aggr_thread_value *buf;
1414+
1415+ buf = sort_aggr_thread(counter, nthreads, ncpus, &sorted_threads);
1416+ if (!buf) {
1417+ perror("cannot sort aggr thread");
1418+ return;
1419+ }
1420+
1421+ for (thread = 0; thread < sorted_threads; thread++) {
14171422 if (prefix)
14181423 fprintf(output, "%s", prefix);
14191424
1420− uval = val * counter->scale;
1421− printout(thread, 0, counter, uval, prefix, run, ena, 1.0);
1425+ id = buf[thread].id;
1426+ if (stat_config.stats)
1427+ printout(id, 0, buf[thread].counter, buf[thread].uval,
1428+ prefix, buf[thread].run, buf[thread].ena, 1.0,
1429+ &stat_config.stats[id]);
1430+ else
1431+ printout(id, 0, buf[thread].counter, buf[thread].uval,
1432+ prefix, buf[thread].run, buf[thread].ena, 1.0,
1433+ &rt_stat);
14221434 fputc('\n', output);
14231435 }
1436+
1437+ free(buf);
14241438 }
14251439
14261440 struct caggr_data {
@@ -1455,7 +1469,8 @@ static void print_counter_aggr(struct perf_evsel *counter, char *prefix) · 28 unmodified lines
14551469 fprintf(output, "%s", prefix);
14561470
14571471 uval = cd.avg * counter->scale;
1458− printout(-1, 0, counter, uval, prefix, cd.avg_running, cd.avg_enabled, cd.avg);
1472+ printout(-1, 0, counter, uval, prefix, cd.avg_running, cd.avg_enabled,
1473+ cd.avg, &rt_stat);
14591474 if (!metric_only)
14601475 fprintf(output, "\n");
14611476 }
@@ -1494,7 +1509,8 @@ static void print_counter(struct perf_evsel *counter, char *prefix) · 32 unmodified lines
14941509 fprintf(output, "%s", prefix);
14951510
14961511 uval = val * counter->scale;
1497− printout(cpu, 0, counter, uval, prefix, run, ena, 1.0);
1512+ printout(cpu, 0, counter, uval, prefix, run, ena, 1.0,
1513+ &rt_stat);
14981514
14991515 fputc('\n', output);
15001516 }
@@ -1526,7 +1542,8 @@ static void print_no_aggr_metric(char *prefix) · 25 unmodified lines
15261542 run = perf_counts(counter->counts, cpu, 0)->run;
15271543
15281544 uval = val * counter->scale;
1529− printout(cpu, 0, counter, uval, prefix, run, ena, 1.0);
1545+ printout(cpu, 0, counter, uval, prefix, run, ena, 1.0,
1546+ &rt_stat);
15301547 }
15311548 fputc('\n', stat_config.output);
15321549 }
@@ -1582,7 +1599,8 @@ static void print_metric_headers(const char *prefix, bool no_indent) · 49 unmodified lines
15821599 perf_stat__print_shadow_stats(counter, 0,
15831600 0,
15841601 &out,
1585− &metric_events);
1602+ &metric_events,
1603+ &rt_stat);
15861604 }
15871605 fputc('\n', stat_config.output);
15881606 }
@@ -2541,6 +2559,35 @@ int process_cpu_map_event(struct perf_tool *tool, · 952 unmodified lines
25412559 return set_maps(st);
25422560 }
25432561
2562+static int runtime_stat_new(struct perf_stat_config *config, int nthreads)
2563+{
2564+ int i;
2565+
2566+ config->stats = calloc(nthreads, sizeof(struct runtime_stat));
2567+ if (!config->stats)
2568+ return -1;
2569+
2570+ config->stats_num = nthreads;
2571+
2572+ for (i = 0; i < nthreads; i++)
2573+ runtime_stat__init(&config->stats[i]);
2574+
2575+ return 0;
2576+}
2577+
2578+static void runtime_stat_delete(struct perf_stat_config *config)
2579+{
2580+ int i;
2581+
2582+ if (!config->stats)
2583+ return;
2584+
2585+ for (i = 0; i < config->stats_num; i++)
2586+ runtime_stat__exit(&config->stats[i]);
2587+
2588+ free(config->stats);
2589+}
2590+
25442591 static const char * const stat_report_usage[] = {
25452592 "perf stat report [<options>]",
25462593 NULL,
@@ -2750,12 +2797,16 @@ int cmd_stat(int argc, const char **argv) · 203 unmodified lines
27502797 run_count = 1;
27512798 }
27522799
2753− if ((stat_config.aggr_mode == AGGR_THREAD) && !target__has_task(&target)) {
2754− fprintf(stderr, "The --per-thread option is only available "
2755− "when monitoring via -p -t options.\n");
2756− parse_options_usage(NULL, stat_options, "p", 1);
2757− parse_options_usage(NULL, stat_options, "t", 1);
2758− goto out;
2800+ if ((stat_config.aggr_mode == AGGR_THREAD) &&
2801+ !target__has_task(&target)) {
2802+ if (!target.system_wide || target.cpu_list) {
2803+ fprintf(stderr, "The --per-thread option is only "
2804+ "available when monitoring via -p -t -a "
2805+ "options or only --per-thread.\n");
2806+ parse_options_usage(NULL, stat_options, "p", 1);
2807+ parse_options_usage(NULL, stat_options, "t", 1);
2808+ goto out;
2809+ }
27592810 }
27602811
27612812 /*
@@ -2779,6 +2830,9 @@ int cmd_stat(int argc, const char **argv) · 17 unmodified lines
27792830
27802831 target__validate(&target);
27812832
2833+ if ((stat_config.aggr_mode == AGGR_THREAD) && (target.system_wide))
2834+ target.per_thread = true;
2835+
27822836 if (perf_evlist__create_maps(evsel_list, &target) < 0) {
27832837 if (target__has_task(&target)) {
27842838 pr_err("Problems finding threads of monitor\n");
@@ -2796,8 +2850,15 @@ int cmd_stat(int argc, const char **argv) · 11 unmodified lines
27962850 * Initialize thread_map with comm names,
27972851 * so we could print it out on output.
27982852 */
2799− if (stat_config.aggr_mode == AGGR_THREAD)
2853+ if (stat_config.aggr_mode == AGGR_THREAD) {
28002854 thread_map__read_comms(evsel_list->threads);
2855+ if (target.system_wide) {
2856+ if (runtime_stat_new(&stat_config,
2857+ thread_map__nr(evsel_list->threads))) {
2858+ goto out;
2859+ }
2860+ }
2861+ }
28012862
28022863 if (interval && interval < 100) {
28032864 if (interval < 10) {
@@ -2887,5 +2948,8 @@ out: · 83 unmodified lines
28872948 sysfs__write_int(FREEZE_ON_SMI_PATH, 0);
28882949
28892950 perf_evlist__delete(evsel_list);
2951+
2952+ runtime_stat_delete(&stat_config);
2953+
28902954 return status;
28912955 }
tools/perf/builtin-top.c+7−3
@@ -99,6 +99,7 @@ static void perf_top__resize(struct perf_top *top) · 98 unmodified lines
9999
100100 static int perf_top__parse_source(struct perf_top *top, struct hist_entry *he)
101101 {
102+ struct perf_evsel *evsel = hists_to_evsel(he->hists);
102103 struct symbol *sym;
103104 struct annotation *notes;
104105 struct map *map;
@@ -137,7 +138,7 @@ static int perf_top__parse_source(struct perf_top *top, struct hist_entry *he) · 32 unmodified lines
137138 return err;
138139 }
139140
140− err = symbol__disassemble(sym, map, NULL, 0, NULL, NULL);
141+ err = symbol__annotate(sym, map, evsel, 0, NULL);
141142 if (err == 0) {
142143 out_assign:
143144 top->sym_filter_entry = he;
@@ -229,6 +230,7 @@ static void perf_top__record_precise_ip(struct perf_top *top, · 85 unmodified lines
229230 static void perf_top__show_details(struct perf_top *top)
230231 {
231232 struct hist_entry *he = top->sym_filter_entry;
233+ struct perf_evsel *evsel = hists_to_evsel(he->hists);
232234 struct annotation *notes;
233235 struct symbol *symbol;
234236 int more;
@@ -241,6 +243,8 @@ static void perf_top__show_details(struct perf_top *top) · 6 unmodified lines
241243
242244 pthread_mutex_lock(&notes->lock);
243245
246+ symbol__calc_percent(symbol, evsel);
247+
244248 if (notes->src == NULL)
245249 goto out_unlock;
246250
@@ -412,7 +416,7 @@ static void perf_top__print_mapped_keys(struct perf_top *top) · 165 unmodified lines
412416 fprintf(stdout, "\t[S] stop annotation.\n");
413417
414418 fprintf(stdout,
415− "\t[K] hide kernel_symbols symbols. \t(%s)\n",
419+ "\t[K] hide kernel symbols. \t(%s)\n",
416420 top->hide_kernel_symbols ? "yes" : "no");
417421 fprintf(stdout,
418422 "\t[U] hide user symbols. \t(%s)\n",
@@ -903,7 +907,7 @@ try_again: · 484 unmodified lines
903907 }
904908 }
905909
906− if (perf_evlist__mmap(evlist, opts->mmap_pages, false) < 0) {
910+ if (perf_evlist__mmap(evlist, opts->mmap_pages) < 0) {
907911 ui__error("Failed to mmap with %d (%s)\n",
908912 errno, str_error_r(errno, msg, sizeof(msg)));
909913 goto out_err;
tools/perf/builtin-trace.c+67−55
@@ -21,6 +21,7 @@ · 20 unmodified lines
2121 #include "builtin.h"
2222 #include "util/color.h"
2323 #include "util/debug.h"
24+#include "util/env.h"
2425 #include "util/event.h"
2526 #include "util/evlist.h"
2627 #include <subcmd/exec-cmd.h>
@@ -45,18 +46,17 @@ · 18 unmodified lines
4546
4647 #include <errno.h>
4748 #include <inttypes.h>
48−#include <libaudit.h> /* FIXME: Still needed for audit_errno_to_name */
4949 #include <poll.h>
5050 #include <signal.h>
5151 #include <stdlib.h>
5252 #include <string.h>
5353 #include <linux/err.h>
5454 #include <linux/filter.h>
55−#include <linux/audit.h>
5655 #include <linux/kernel.h>
5756 #include <linux/random.h>
5857 #include <linux/stringify.h>
5958 #include <linux/time64.h>
59+#include <fcntl.h>
6060
6161 #include "sane_ctype.h"
6262
@@ -111,6 +111,7 @@ struct trace { · 48 unmodified lines
111111 bool summary;
112112 bool summary_only;
113113 bool show_comm;
114+ bool print_sample;
114115 bool show_tool_stats;
115116 bool trace_syscalls;
116117 bool kernel_syscallchains;
@@ -545,9 +546,10 @@ static size_t syscall_arg__scnprintf_getrandom_flags(char *bf, size_t size, · 428 unmodified lines
545546 { .scnprintf = SCA_STRARRAY, \
546547 .parm = &strarray__##array, }
547548
549+#include "trace/beauty/arch_errno_names.c"
548550 #include "trace/beauty/eventfd.c"
549−#include "trace/beauty/flock.c"
550551 #include "trace/beauty/futex_op.c"
552+#include "trace/beauty/futex_val3.c"
551553 #include "trace/beauty/mmap.c"
552554 #include "trace/beauty/mode_t.c"
553555 #include "trace/beauty/msg_flags.c"
@@ -610,7 +612,8 @@ static struct syscall_fmt { · 56 unmodified lines
610612 { .name = "fstat", .alias = "newfstat", },
611613 { .name = "fstatat", .alias = "newfstatat", },
612614 { .name = "futex",
613− .arg = { [1] = { .scnprintf = SCA_FUTEX_OP, /* op */ }, }, },
615+ .arg = { [1] = { .scnprintf = SCA_FUTEX_OP, /* op */ },
616+ [5] = { .scnprintf = SCA_FUTEX_VAL3, /* val3 */ }, }, },
614617 { .name = "futimesat",
615618 .arg = { [0] = { .scnprintf = SCA_FDAT, /* fd */ }, }, },
616619 { .name = "getitimer",
@@ -622,6 +625,7 @@ static struct syscall_fmt { · 5 unmodified lines
622625 .arg = { [2] = { .scnprintf = SCA_GETRANDOM_FLAGS, /* flags */ }, }, },
623626 { .name = "getrlimit",
624627 .arg = { [0] = STRARRAY(resource, rlimit_resources), }, },
628+ { .name = "gettid", .errpid = true, },
625629 { .name = "ioctl",
626630 .arg = {
627631 #if defined(__i386__) || defined(__x86_64__)
@@ -819,7 +823,7 @@ static size_t fprintf_duration(unsigned long t, bool calculated, FILE *fp) · 191 unmodified lines
819823 size_t printed = fprintf(fp, "(");
820824
821825 if (!calculated)
822− printed += fprintf(fp, " ? ");
826+ printed += fprintf(fp, " ");
823827 else if (duration >= 1.0)
824828 printed += color_fprintf(fp, PERF_COLOR_RED, "%6.3f ms", duration);
825829 else if (duration >= 0.01)
@@ -1554,10 +1558,9 @@ static void thread__update_stats(struct thread_trace *ttrace, · 728 unmodified lines
15541558 update_stats(stats, duration);
15551559 }
15561560
1557−static int trace__printf_interrupted_entry(struct trace *trace, struct perf_sample *sample)
1561+static int trace__printf_interrupted_entry(struct trace *trace)
15581562 {
15591563 struct thread_trace *ttrace;
1560− u64 duration;
15611564 size_t printed;
15621565
15631566 if (trace->current == NULL)
@@ -1568,15 +1571,30 @@ static int trace__printf_interrupted_entry(struct trace *trace, struct perf_samp · 4 unmodified lines
15681571 if (!ttrace->entry_pending)
15691572 return 0;
15701573
1571− duration = sample->time - ttrace->entry_time;
1572−
1573− printed = trace__fprintf_entry_head(trace, trace->current, duration, true, ttrace->entry_time, trace->output);
1574+ printed = trace__fprintf_entry_head(trace, trace->current, 0, false, ttrace->entry_time, trace->output);
15741575 printed += fprintf(trace->output, "%-70s) ...\n", ttrace->entry_str);
15751576 ttrace->entry_pending = false;
15761577
15771578 return printed;
15781579 }
15791580
1581+static int trace__fprintf_sample(struct trace *trace, struct perf_evsel *evsel,
1582+ struct perf_sample *sample, struct thread *thread)
1583+{
1584+ int printed = 0;
1585+
1586+ if (trace->print_sample) {
1587+ double ts = (double)sample->time / NSEC_PER_MSEC;
1588+
1589+ printed += fprintf(trace->output, "%22s %10.3f %s %d/%d [%d]\n",
1590+ perf_evsel__name(evsel), ts,
1591+ thread__comm_str(thread),
1592+ sample->pid, sample->tid, sample->cpu);
1593+ }
1594+
1595+ return printed;
1596+}
1597+
15801598 static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel,
15811599 union perf_event *event __maybe_unused,
15821600 struct perf_sample *sample)
@@ -1597,6 +1615,8 @@ static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel, · 14 unmodified lines
15971615 if (ttrace == NULL)
15981616 goto out_put;
15991617
1618+ trace__fprintf_sample(trace, evsel, sample, thread);
1619+
16001620 args = perf_evsel__sc_tp_ptr(evsel, args, sample);
16011621
16021622 if (ttrace->entry_str == NULL) {
@@ -1606,7 +1626,7 @@ static int trace__sys_enter(struct trace *trace, struct perf_evsel *evsel, · 3 unmodified lines
16061626 }
16071627
16081628 if (!(trace->duration_filter || trace->summary_only || trace->min_stack))
1609− trace__printf_interrupted_entry(trace, sample);
1629+ trace__printf_interrupted_entry(trace);
16101630
16111631 ttrace->entry_time = sample->time;
16121632 msg = ttrace->entry_str;
@@ -1643,7 +1663,7 @@ static int trace__resolve_callchain(struct trace *trace, struct perf_evsel *evse · 30 unmodified lines
16431663 struct addr_location al;
16441664
16451665 if (machine__resolve(trace->host, &al, sample) < 0 ||
1646− thread__resolve_callchain(al.thread, cursor, evsel, sample, NULL, NULL, trace->max_stack))
1666+ thread__resolve_callchain(al.thread, cursor, evsel, sample, NULL, NULL, evsel->attr.sample_max_stack))
16471667 return -1;
16481668
16491669 return 0;
@@ -1659,6 +1679,14 @@ static int trace__fprintf_callchain(struct trace *trace, struct perf_sample *sam · 9 unmodified lines
16591679 return sample__fprintf_callchain(sample, 38, print_opts, &callchain_cursor, trace->output);
16601680 }
16611681
1682+static const char *errno_to_name(struct perf_evsel *evsel, int err)
1683+{
1684+ struct perf_env *env = perf_evsel__env(evsel);
1685+ const char *arch_name = perf_env__arch(env);
1686+
1687+ return arch_syscalls__strerrno(arch_name, err);
1688+}
1689+
16621690 static int trace__sys_exit(struct trace *trace, struct perf_evsel *evsel,
16631691 union perf_event *event __maybe_unused,
16641692 struct perf_sample *sample)
@@ -1679,6 +1707,8 @@ static int trace__sys_exit(struct trace *trace, struct perf_evsel *evsel, · 14 unmodified lines
16791707 if (ttrace == NULL)
16801708 goto out_put;
16811709
1710+ trace__fprintf_sample(trace, evsel, sample, thread);
1711+
16821712 if (trace->summary)
16831713 thread__update_stats(ttrace, id, sample);
16841714
@@ -1729,7 +1759,7 @@ signed_print: · 44 unmodified lines
17291759 errno_print: {
17301760 char bf[STRERR_BUFSIZE];
17311761 const char *emsg = str_error_r(-ret, bf, sizeof(bf)),
1732− *e = audit_errno_to_name(-ret);
1762+ *e = errno_to_name(evsel, -ret);
17331763
17341764 fprintf(trace->output, ") = -1 %s %s", e, emsg);
17351765 }
@@ -1910,7 +1940,7 @@ static int trace__event_handler(struct trace *trace, struct perf_evsel *evsel, · 174 unmodified lines
19101940 }
19111941 }
19121942
1913− trace__printf_interrupted_entry(trace, sample);
1943+ trace__printf_interrupted_entry(trace);
19141944 trace__fprintf_tstamp(trace, sample->time, trace->output);
19151945
19161946 if (trace->trace_syscalls)
@@ -2221,6 +2251,9 @@ static int trace__add_syscall_newtp(struct trace *trace) · 304 unmodified lines
22212251 if (perf_evsel__init_sc_tp_uint_field(sys_exit, ret))
22222252 goto out_delete_sys_exit;
22232253
2254+ perf_evsel__config_callchain(sys_enter, &trace->opts, &callchain_param);
2255+ perf_evsel__config_callchain(sys_exit, &trace->opts, &callchain_param);
2256+
22242257 perf_evlist__add(evlist, sys_enter);
22252258 perf_evlist__add(evlist, sys_exit);
22262259
@@ -2317,6 +2350,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv) · 90 unmodified lines
23172350 pgfault_maj = perf_evsel__new_pgfault(PERF_COUNT_SW_PAGE_FAULTS_MAJ);
23182351 if (pgfault_maj == NULL)
23192352 goto out_error_mem;
2353+ perf_evsel__config_callchain(pgfault_maj, &trace->opts, &callchain_param);
23202354 perf_evlist__add(evlist, pgfault_maj);
23212355 }
23222356
@@ -2324,6 +2358,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv) · 1 unmodified line
23242358 pgfault_min = perf_evsel__new_pgfault(PERF_COUNT_SW_PAGE_FAULTS_MIN);
23252359 if (pgfault_min == NULL)
23262360 goto out_error_mem;
2361+ perf_evsel__config_callchain(pgfault_min, &trace->opts, &callchain_param);
23272362 perf_evlist__add(evlist, pgfault_min);
23282363 }
23292364
@@ -2344,45 +2379,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv) · 14 unmodified lines
23442379 goto out_delete_evlist;
23452380 }
23462381
2347− perf_evlist__config(evlist, &trace->opts, NULL);
2348−
2349− if (callchain_param.enabled) {
2350− bool use_identifier = false;
2351−
2352− if (trace->syscalls.events.sys_exit) {
2353− perf_evsel__config_callchain(trace->syscalls.events.sys_exit,
2354− &trace->opts, &callchain_param);
2355− use_identifier = true;
2356− }
2357−
2358− if (pgfault_maj) {
2359− perf_evsel__config_callchain(pgfault_maj, &trace->opts, &callchain_param);
2360− use_identifier = true;
2361− }
2362−
2363− if (pgfault_min) {
2364− perf_evsel__config_callchain(pgfault_min, &trace->opts, &callchain_param);
2365− use_identifier = true;
2366− }
2367−
2368− if (use_identifier) {
2369− /*
2370− * Now we have evsels with different sample_ids, use
2371− * PERF_SAMPLE_IDENTIFIER to map from sample to evsel
2372− * from a fixed position in each ring buffer record.
2373− *
2374− * As of this the changeset introducing this comment, this
2375− * isn't strictly needed, as the fields that can come before
2376− * PERF_SAMPLE_ID are all used, but we'll probably disable
2377− * some of those for things like copying the payload of
2378− * pointer syscall arguments, and for vfs_getname we don't
2379− * need PERF_SAMPLE_ADDR and PERF_SAMPLE_IP, so do this
2380− * here as a warning we need to use PERF_SAMPLE_IDENTIFIER.
2381− */
2382− perf_evlist__set_sample_bit(evlist, IDENTIFIER);
2383− perf_evlist__reset_sample_bit(evlist, ID);
2384− }
2385− }
2382+ perf_evlist__config(evlist, &trace->opts, &callchain_param);
23862383
23872384 signal(SIGCHLD, sig_handler);
23882385 signal(SIGINT, sig_handler);
@@ -2437,7 +2434,7 @@ static int trace__run(struct trace *trace, int argc, const char **argv) · 48 unmodified lines
24372434 if (err < 0)
24382435 goto out_error_apply_filters;
24392436
2440− err = perf_evlist__mmap(evlist, trace->opts.mmap_pages, false);
2437+ err = perf_evlist__mmap(evlist, trace->opts.mmap_pages);
24412438 if (err < 0)
24422439 goto out_error_mmap;
24432440
@@ -2455,6 +2452,18 @@ static int trace__run(struct trace *trace, int argc, const char **argv) · 11 unmodified lines
24552452 trace->multiple_threads = thread_map__pid(evlist->threads, 0) == -1 ||
24562453 evlist->threads->nr > 1 ||
24572454 perf_evlist__first(evlist)->attr.inherit;
2455+
2456+ /*
2457+ * Now that we already used evsel->attr to ask the kernel to setup the
2458+ * events, lets reuse evsel->attr.sample_max_stack as the limit in
2459+ * trace__resolve_callchain(), allowing per-event max-stack settings
2460+ * to override an explicitely set --max-stack global setting.
2461+ */
2462+ evlist__for_each_entry(evlist, evsel) {
2463+ if ((evsel->attr.sample_type & PERF_SAMPLE_CALLCHAIN) &&
2464+ evsel->attr.sample_max_stack == 0)
2465+ evsel->attr.sample_max_stack = trace->max_stack;
2466+ }
24582467 again:
24592468 before = trace->nr_events;
24602469
@@ -3046,6 +3055,8 @@ int cmd_trace(int argc, const char **argv) · 585 unmodified lines
30463055 "Set the maximum stack depth when parsing the callchain, "
30473056 "anything beyond the specified depth will be ignored. "
30483057 "Default: kernel.perf_event_max_stack or " __stringify(PERF_MAX_STACK_DEPTH)),
3058+ OPT_BOOLEAN(0, "print-sample", &trace.print_sample,
3059+ "print the PERF_RECORD_SAMPLE PERF_SAMPLE_ info, for debugging"),
30493060 OPT_UINTEGER(0, "proc-map-timeout", &trace.opts.proc_map_timeout,
30503061 "per thread proc mmap processing timeout in ms"),
30513062 OPT_UINTEGER('D', "delay", &trace.opts.initial_delay,
@@ -3097,8 +3108,9 @@ int cmd_trace(int argc, const char **argv) · 45 unmodified lines
30973108 }
30983109
30993110 #ifdef HAVE_DWARF_UNWIND_SUPPORT
3100− if ((trace.min_stack || max_stack_user_set) && !callchain_param.enabled && trace.trace_syscalls)
3111+ if ((trace.min_stack || max_stack_user_set) && !callchain_param.enabled) {
31013112 record_opts__parse_callchain(&trace.opts, &callchain_param, "dwarf", false);
3113+ }
31023114 #endif
31033115
31043116 if (callchain_param.enabled) {
tools/perf/check-headers.sh+10−1
@@ -33,21 +33,30 @@ arch/s390/include/uapi/asm/kvm.h · 32 unmodified lines
3333 arch/s390/include/uapi/asm/kvm_perf.h
3434 arch/s390/include/uapi/asm/ptrace.h
3535 arch/s390/include/uapi/asm/sie.h
36+arch/s390/include/uapi/asm/unistd.h
3637 arch/arm/include/uapi/asm/kvm.h
3738 arch/arm64/include/uapi/asm/kvm.h
39+arch/alpha/include/uapi/asm/errno.h
40+arch/mips/include/asm/errno.h
41+arch/mips/include/uapi/asm/errno.h
42+arch/parisc/include/uapi/asm/errno.h
43+arch/powerpc/include/uapi/asm/errno.h
44+arch/sparc/include/uapi/asm/errno.h
45+arch/x86/include/uapi/asm/errno.h
3846 include/asm-generic/bitops/arch_hweight.h
3947 include/asm-generic/bitops/const_hweight.h
4048 include/asm-generic/bitops/__fls.h
4149 include/asm-generic/bitops/fls.h
4250 include/asm-generic/bitops/fls64.h
4351 include/linux/coresight-pmu.h
52+include/uapi/asm-generic/errno.h
53+include/uapi/asm-generic/errno-base.h
4454 include/uapi/asm-generic/ioctls.h
4555 include/uapi/asm-generic/mman-common.h
4656 '
4757
4858 check () {
4959 file=$1
50− opts="--ignore-blank-lines --ignore-space-change"
5160
5261 shift
5362 while [ -n "$*" ]; do
tools/perf/perf-completion.sh+43−4
@@ -162,8 +162,37 @@ __perf_main () · 161 unmodified lines
162162 # List possible events for -e option
163163 elif [[ $prev == @("-e"|"--event") &&
164164 $prev_skip_opts == @(record|stat|top) ]]; then
165− evts=$($cmd list --raw-dump)
166− __perfcomp_colon "$evts" "$cur"
165+
166+ local cur1=${COMP_WORDS[COMP_CWORD]}
167+ local raw_evts=$($cmd list --raw-dump)
168+ local arr s tmp result
169+
170+ if [[ "$cur1" == */* && ${cur1#*/} =~ ^[A-Z] ]]; then
171+ OLD_IFS="$IFS"
172+ IFS=" "
173+ arr=($raw_evts)
174+ IFS="$OLD_IFS"
175+
176+ for s in ${arr[@]}
177+ do
178+ if [[ "$s" == *cpu/* ]]; then
179+ tmp=${s#*cpu/}
180+ result=$result" ""cpu/"${tmp^^}
181+ else
182+ result=$result" "$s
183+ fi
184+ done
185+
186+ evts=${result}" "$(ls /sys/bus/event_source/devices/cpu/events)
187+ else
188+ evts=${raw_evts}" "$(ls /sys/bus/event_source/devices/cpu/events)
189+ fi
190+
191+ if [[ "$cur1" == , ]]; then
192+ __perfcomp_colon "$evts" ""
193+ else
194+ __perfcomp_colon "$evts" "$cur1"
195+ fi
167196 else
168197 # List subcommands for perf commands
169198 if [[ $prev_skip_opts == @(kvm|kmem|mem|lock|sched|
@@ -246,11 +275,21 @@ fi · 76 unmodified lines
246275 type perf &>/dev/null &&
247276 _perf()
248277 {
278+ if [[ "$COMP_WORDBREAKS" != *,* ]]; then
279+ COMP_WORDBREAKS="${COMP_WORDBREAKS},"
280+ export COMP_WORDBREAKS
281+ fi
282+
283+ if [[ "$COMP_WORDBREAKS" == *:* ]]; then
284+ COMP_WORDBREAKS="${COMP_WORDBREAKS/:/}"
285+ export COMP_WORDBREAKS
286+ fi
287+
249288 local cur words cword prev
250289 if [ $preload_get_comp_words_by_ref = "true" ]; then
251− _get_comp_words_by_ref -n =: cur words cword prev
290+ _get_comp_words_by_ref -n =:, cur words cword prev
252291 else
253− __perf_get_comp_words_by_ref -n =: cur words cword prev
292+ __perf_get_comp_words_by_ref -n =:, cur words cword prev
254293 fi
255294 __perf_main
256295 } &&
tools/perf/perf.c+2−2
@@ -73,7 +73,7 @@ static struct cmd_struct commands[] = { · 72 unmodified lines
7373 { "lock", cmd_lock, 0 },
7474 { "kvm", cmd_kvm, 0 },
7575 { "test", cmd_test, 0 },
76−#ifdef HAVE_LIBAUDIT_SUPPORT
76+#if defined(HAVE_LIBAUDIT_SUPPORT) || defined(HAVE_SYSCALL_TABLE)
7777 { "trace", cmd_trace, 0 },
7878 #endif
7979 { "inject", cmd_inject, 0 },
@@ -485,7 +485,7 @@ int main(int argc, const char **argv) · 405 unmodified lines
485485 argv[0] = cmd;
486486 }
487487 if (strstarts(cmd, "trace")) {
488−#ifdef HAVE_LIBAUDIT_SUPPORT
488+#if defined(HAVE_LIBAUDIT_SUPPORT) || defined(HAVE_SYSCALL_TABLE)
489489 setup_path();
490490 argv[0] = "trace";
491491 return cmd_trace(argc, argv);
tools/perf/pmu-events/arch/arm64/cavium/thunderx2-imp-def.json+62−0
@@ -0,0 +1,62 @@
1+[
2+ {
3+ "PublicDescription": "Attributable Level 1 data cache access, read",
4+ "EventCode": "0x40",
5+ "EventName": "l1d_cache_rd",
6+ "BriefDescription": "L1D cache read",
7+ },
8+ {
9+ "PublicDescription": "Attributable Level 1 data cache access, write ",
10+ "EventCode": "0x41",
11+ "EventName": "l1d_cache_wr",
12+ "BriefDescription": "L1D cache write",
13+ },
14+ {
15+ "PublicDescription": "Attributable Level 1 data cache refill, read",
16+ "EventCode": "0x42",
17+ "EventName": "l1d_cache_refill_rd",
18+ "BriefDescription": "L1D cache refill read",
19+ },
20+ {
21+ "PublicDescription": "Attributable Level 1 data cache refill, write",
22+ "EventCode": "0x43",
23+ "EventName": "l1d_cache_refill_wr",
24+ "BriefDescription": "L1D refill write",
25+ },
26+ {
27+ "PublicDescription": "Attributable Level 1 data TLB refill, read",
28+ "EventCode": "0x4C",
29+ "EventName": "l1d_tlb_refill_rd",
30+ "BriefDescription": "L1D tlb refill read",
31+ },
32+ {
33+ "PublicDescription": "Attributable Level 1 data TLB refill, write",
34+ "EventCode": "0x4D",
35+ "EventName": "l1d_tlb_refill_wr",
36+ "BriefDescription": "L1D tlb refill write",
37+ },
38+ {
39+ "PublicDescription": "Attributable Level 1 data or unified TLB access, read",
40+ "EventCode": "0x4E",
41+ "EventName": "l1d_tlb_rd",
42+ "BriefDescription": "L1D tlb read",
43+ },
44+ {
45+ "PublicDescription": "Attributable Level 1 data or unified TLB access, write",
46+ "EventCode": "0x4F",
47+ "EventName": "l1d_tlb_wr",
48+ "BriefDescription": "L1D tlb write",
49+ },
50+ {
51+ "PublicDescription": "Bus access read",
52+ "EventCode": "0x60",
53+ "EventName": "bus_access_rd",
54+ "BriefDescription": "Bus access read",
55+ },
56+ {
57+ "PublicDescription": "Bus access write",
58+ "EventCode": "0x61",
59+ "EventName": "bus_access_wr",
60+ "BriefDescription": "Bus access write",
61+ }
62+]
tools/perf/pmu-events/arch/arm64/mapfile.csv+15−0
@@ -0,0 +1,15 @@
1+# Format:
2+# MIDR,Version,JSON/file/pathname,Type
3+#
4+# where
5+# MIDR Processor version
6+# Variant[23:20] and Revision [3:0] should be zero.
7+# Version could be used to track version of of JSON file
8+# but currently unused.
9+# JSON/file/pathname is the path to JSON file, relative
10+# to tools/perf/pmu-events/arch/arm64/.
11+# Type is core, uncore etc
12+#
13+#
14+#Family-model,Version,Filename,EventType
15+0x00000000420f5160,v1,cavium,core
tools/perf/pmu-events/arch/powerpc/mapfile.csv+2−10
@@ -13,13 +13,5 @@ · 12 unmodified lines
1313 #
1414
1515 # Power8 entries
16−004b0000,1,power8,core
17−004b0201,1,power8,core
18−004c0000,1,power8,core
19−004d0000,1,power8,core
20−004d0100,1,power8,core
21−004d0200,1,power8,core
22−004c0100,1,power8,core
23−004e0100,1,power9,core
24−004e0200,1,power9,core
25−004e1200,1,power9,core
16+004[bcd][[:xdigit:]]{4},1,power8,core
17+004e[[:xdigit:]]{4},1,power9,core
tools/perf/pmu-events/arch/powerpc/power9/cache.json+0−5
@@ -124,11 +124,6 @@ · 123 unmodified lines
124124 "EventName": "PM_CMPLU_STALL_LARX",
125125 "BriefDescription": "Finish stall because the NTF instruction was a larx waiting to be satisfied"
126126 },
127− {,
128− "EventCode": "0x3006C",
129− "EventName": "PM_RUN_CYC_SMT2_MODE",
130− "BriefDescription": "Cycles in which this thread's run latch is set and the core is in SMT2 mode"
131− },
132127 {,
133128 "EventCode": "0x1C058",
134129 "EventName": "PM_DTLB_MISS_16G",
tools/perf/pmu-events/arch/powerpc/power9/frontend.json+1−6
@@ -1,9 +1,4 @@
11 [
2− {,
3− "EventCode": "0x3E15C",
4− "EventName": "PM_MRK_L2_TM_ST_ABORT_SISTER",
5− "BriefDescription": "TM marked store abort for this thread"
6− },
72 {,
83 "EventCode": "0x25044",
94 "EventName": "PM_IPTEG_FROM_L31_MOD",
@@ -369,4 +364,4 @@ · 359 unmodified lines
369364 "EventName": "PM_IPTEG_FROM_L31_ECO_MOD",
370365 "BriefDescription": "A Page Table Entry was loaded into the TLB with Modified (M) data from another core's ECO L3 on the same chip due to a instruction side request"
371366 }
372−]
367+]
tools/perf/pmu-events/arch/powerpc/power9/marked.json+6−21
@@ -1,9 +1,4 @@
11 [
2− {,
3− "EventCode": "0x3C052",
4− "EventName": "PM_DATA_SYS_PUMP_MPRED",
5− "BriefDescription": "Final Pump Scope (system) mispredicted. Either the original scope was too small (Chip/Group) or the original scope was System and it should have been smaller. Counts for a demand load"
6− },
72 {,
83 "EventCode": "0x3013E",
94 "EventName": "PM_MRK_STALL_CMPLU_CYC",
@@ -254,6 +249,11 @@ · 244 unmodified lines
254249 "EventName": "PM_RADIX_PWC_L1_PDE_FROM_L3",
255250 "BriefDescription": "A Page Directory Entry was reloaded to a level 1 page walk cache from the core's L3 data cache"
256251 },
252+ {,
253+ "EventCode": "0x3C052",
254+ "EventName": "PM_DATA_SYS_PUMP_MPRED",
255+ "BriefDescription": "Final Pump Scope (system) mispredicted. Either the original scope was too small (Chip/Group) or the original scope was System and it should have been smaller. Counts for a demand load"
256+ },
257257 {,
258258 "EventCode": "0x4D142",
259259 "EventName": "PM_MRK_DATA_FROM_L3",
@@ -434,21 +434,6 @@ · 174 unmodified lines
434434 "EventName": "PM_ITLB_MISS",
435435 "BriefDescription": "ITLB Reloaded. Counts 1 per ITLB miss for HPT but multiple for radix depending on number of levels traveresed"
436436 },
437− {,
438− "EventCode": "0x2D024",
439− "EventName": "PM_RADIX_PWC_L2_HIT",
440− "BriefDescription": "A radix translation attempt missed in the TLB but hit on both the first and second levels of page walk cache."
441− },
442− {,
443− "EventCode": "0x3F056",
444− "EventName": "PM_RADIX_PWC_L3_HIT",
445− "BriefDescription": "A radix translation attempt missed in the TLB but hit on the first, second, and third levels of page walk cache."
446− },
447− {,
448− "EventCode": "0x4E014",
449− "EventName": "PM_TM_TX_PASS_RUN_INST",
450− "BriefDescription": "Run instructions spent in successful transactions"
451− },
452437 {,
453438 "EventCode": "0x1E044",
454439 "EventName": "PM_DPTEG_FROM_L3_NO_CONFLICT",
@@ -644,4 +629,4 @@ · 189 unmodified lines
644629 "EventName": "PM_MRK_BR_MPRED_CMPL",
645630 "BriefDescription": "Marked Branch Mispredicted"
646631 }
647−]
632+]
tools/perf/pmu-events/arch/powerpc/power9/other.json+73−203
@@ -79,6 +79,11 @@ · 78 unmodified lines
7979 "EventName": "PM_RADIX_PWC_MISS",
8080 "BriefDescription": "A radix translation attempt missed in the TLB and all levels of page walk cache."
8181 },
82+ {,
83+ "EventCode": "0x26882",
84+ "EventName": "PM_L2_DC_INV",
85+ "BriefDescription": "D-cache invalidates sent over the reload bus to the core"
86+ },
8287 {,
8388 "EventCode": "0x24048",
8489 "EventName": "PM_INST_FROM_LMEM",
@@ -94,11 +99,6 @@ · 9 unmodified lines
9499 "EventName": "PM_TM_PASSED",
95100 "BriefDescription": "Number of TM transactions that passed"
96101 },
97− {,
98− "EventCode": "0xD1A0",
99− "EventName": "PM_MRK_LSU_FLUSH_LHS",
100− "BriefDescription": "Effective Address alias flush : no EA match but Real Address match. If the data has not yet been returned for this load, the instruction will just be rejected, but if it has returned data, it will be flushed"
101− },
102102 {,
103103 "EventCode": "0xF088",
104104 "EventName": "PM_LSU0_STORE_REJECT",
@@ -127,7 +127,7 @@ · 22 unmodified lines
127127 {,
128128 "EventCode": "0xD08C",
129129 "EventName": "PM_LSU2_LDMX_FIN",
130− "BriefDescription": "New P9 instruction LDMX. The definition of this new PMU event is (from the ldmx RFC02491): The thread has executed an ldmx instruction that accessed a doubleword that contains an effective address within an enabled section of the Load Monitored region. This event, therefore, should not occur if the FSCR has disabled the load monitored facility (FSCR[52]) or disabled the EBB facility (FSCR[56])"
130+ "BriefDescription": "New P9 instruction LDMX. The definition of this new PMU event is (from the ldmx RFC02491): The thread has executed an ldmx instruction that accessed a doubleword that contains an effective address within an enabled section of the Load Monitored region. This event, therefore, should not occur if the FSCR has disabled the load monitored facility (FSCR[52]) or disabled the EBB facility (FSCR[56])."
131131 },
132132 {,
133133 "EventCode": "0x300F8",
@@ -204,11 +204,6 @@ · 70 unmodified lines
204204 "EventName": "PM_MRK_DATA_FROM_L31_ECO_MOD_CYC",
205205 "BriefDescription": "Duration in cycles to reload with Modified (M) data from another core's ECO L3 on the same chip due to a marked load"
206206 },
207− {,
208− "EventCode": "0xF0B4",
209− "EventName": "PM_DC_PREF_CONS_ALLOC",
210− "BriefDescription": "Prefetch stream allocated in the conservative phase by either the hardware prefetch mechanism or software prefetch"
211− },
212207 {,
213208 "EventCode": "0xF894",
214209 "EventName": "PM_LSU3_L1_CAM_CANCEL",
@@ -219,21 +214,11 @@ · 4 unmodified lines
219214 "EventName": "PM_FLUSH_DISP_TLBIE",
220215 "BriefDescription": "Dispatch Flush: TLBIE"
221216 },
222− {,
223− "EventCode": "0xD1A4",
224− "EventName": "PM_MRK_LSU_FLUSH_SAO",
225− "BriefDescription": "A load-hit-load condition with Strong Address Ordering will have address compare disabled and flush"
226− },
227217 {,
228218 "EventCode": "0x4E11E",
229219 "EventName": "PM_MRK_DATA_FROM_DMEM_CYC",
230220 "BriefDescription": "Duration in cycles to reload from another chip's memory on the same Node or Group (Distant) due to a marked load"
231221 },
232− {,
233− "EventCode": "0x5894",
234− "EventName": "PM_LWSYNC",
235− "BriefDescription": "Lwsync instruction decoded and transferred"
236− },
237222 {,
238223 "EventCode": "0x14156",
239224 "EventName": "PM_MRK_DATA_FROM_L2_CYC",
@@ -244,11 +229,6 @@ · 4 unmodified lines
244229 "EventName": "PM_RD_CLEARING_SC",
245230 "BriefDescription": "Read clearing SC"
246231 },
247− {,
248− "EventCode": "0x50A0",
249− "EventName": "PM_HWSYNC",
250− "BriefDescription": "Hwsync instruction decoded and transferred"
251− },
252232 {,
253233 "EventCode": "0x168B0",
254234 "EventName": "PM_L3_P1_NODE_PUMP",
@@ -264,6 +244,11 @@ · 9 unmodified lines
264244 "EventName": "PM_MRK_DATA_FROM_L2_DISP_CONFLICT_LDHITST",
265245 "BriefDescription": "The processor's data cache was reloaded from local core's L2 with load hit store conflict due to a marked load"
266246 },
247+ {,
248+ "EventCode": "0x468AE",
249+ "EventName": "PM_L3_P3_CO_RTY",
250+ "BriefDescription": "L3 CO received retry port 3 (memory only), every retry counted"
251+ },
267252 {,
268253 "EventCode": "0x460A8",
269254 "EventName": "PM_SN_HIT",
@@ -279,11 +264,6 @@ · 9 unmodified lines
279264 "EventName": "PM_DC_PREF_HW_ALLOC",
280265 "BriefDescription": "Prefetch stream allocated by the hardware prefetch mechanism"
281266 },
282− {,
283− "EventCode": "0xF0BC",
284− "EventName": "PM_LS2_UNALIGNED_ST",
285− "BriefDescription": "Store instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the Store of that size. If the Store wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
286− },
287267 {,
288268 "EventCode": "0xD0AC",
289269 "EventName": "PM_SRQ_SYNC_CYC",
@@ -379,26 +359,11 @@ · 89 unmodified lines
379359 "EventName": "PM_RUN_CYC_SMT4_MODE",
380360 "BriefDescription": "Cycles in which this thread's run latch is set and the core is in SMT4 mode"
381361 },
382− {,
383− "EventCode": "0x5088",
384− "EventName": "PM_DECODE_FUSION_OP_PRESERV",
385− "BriefDescription": "Destructive op operand preservation"
386− },
387362 {,
388363 "EventCode": "0x1D14E",
389364 "EventName": "PM_MRK_DATA_FROM_OFF_CHIP_CACHE_CYC",
390365 "BriefDescription": "Duration in cycles to reload either shared or modified data from another core's L2/L3 on a different chip (remote or distant) due to a marked load"
391366 },
392− {,
393− "EventCode": "0x509C",
394− "EventName": "PM_FORCED_NOP",
395− "BriefDescription": "Instruction was forced to execute as a nop because it was found to behave like a nop (have no effect) at decode time"
396− },
397− {,
398− "EventCode": "0xC098",
399− "EventName": "PM_LS2_UNALIGNED_LD",
400− "BriefDescription": "Load instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the load of that size. If the load wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
401− },
402367 {,
403368 "EventCode": "0x20058",
404369 "EventName": "PM_DARQ1_10_12_ENTRIES",
@@ -434,11 +399,6 @@ · 29 unmodified lines
434399 "EventName": "PM_LSU1_STORE_REJECT",
435400 "BriefDescription": "All internal store rejects cause the instruction to go back to the SRQ and go to sleep until woken up to try again after the condition has been met"
436401 },
437− {,
438− "EventCode": "0x4505E",
439− "EventName": "PM_FLOP_CMPL",
440− "BriefDescription": "Floating Point Operation Finished"
441− },
442402 {,
443403 "EventCode": "0x1D144",
444404 "EventName": "PM_MRK_DATA_FROM_L3_DISP_CONFLICT",
@@ -480,14 +440,9 @@ · 35 unmodified lines
480440 "BriefDescription": "XL-form branch was mispredicted due to the predicted target address missing from EAT. The EAT forces a mispredict in this case since there is no predicated target to validate. This is a rare case that may occur when the EAT is full and a branch is issued"
481441 },
482442 {,
483− "EventCode": "0xC094",
484− "EventName": "PM_LS0_UNALIGNED_LD",
485− "BriefDescription": "Load instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the load of that size. If the load wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
486− },
487− {,
488− "EventCode": "0xF8BC",
489− "EventName": "PM_LS3_UNALIGNED_ST",
490− "BriefDescription": "Store instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the Store of that size. If the Store wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
443+ "EventCode": "0x460AE",
444+ "EventName": "PM_L3_P2_CO_RTY",
445+ "BriefDescription": "L3 CO received retry port 2 (memory only), every retry counted"
491446 },
492447 {,
493448 "EventCode": "0x58B0",
@@ -504,11 +459,6 @@ · 10 unmodified lines
504459 "EventName": "PM_TM_ST_CONF",
505460 "BriefDescription": "TM Store (fav or non-fav) ran into conflict (failed)"
506461 },
507− {,
508− "EventCode": "0xD998",
509− "EventName": "PM_MRK_LSU_FLUSH_EMSH",
510− "BriefDescription": "An ERAT miss was detected after a set-p hit. Erat tracker indicates fail due to tlbmiss and the instruction gets flushed because the instruction was working on the wrong address"
511− },
512462 {,
513463 "EventCode": "0xF8A0",
514464 "EventName": "PM_NON_DATA_STORE",
@@ -524,11 +474,6 @@ · 9 unmodified lines
524474 "EventName": "PM_BR_UNCOND",
525475 "BriefDescription": "Unconditional Branch Completed. HW branch prediction was not used for this branch. This can be an I-form branch, a B-form branch with BO-field set to branch always, or a B-form branch which was covenrted to a Resolve."
526476 },
527− {,
528− "EventCode": "0x1F056",
529− "EventName": "PM_RADIX_PWC_L1_HIT",
530− "BriefDescription": "A radix translation attempt missed in the TLB and only the first level page walk cache was a hit."
531− },
532477 {,
533478 "EventCode": "0xF8A8",
534479 "EventName": "PM_DC_PREF_FUZZY_CONF",
@@ -544,6 +489,11 @@ · 9 unmodified lines
544489 "EventName": "PM_LSU2_TM_L1_MISS",
545490 "BriefDescription": "Load tm L1 miss"
546491 },
492+ {,
493+ "EventCode": "0xC880",
494+ "EventName": "PM_LS1_LD_VECTOR_FIN",
495+ "BriefDescription": ""
496+ },
547497 {,
548498 "EventCode": "0x2894",
549499 "EventName": "PM_TM_OUTER_TEND",
@@ -564,21 +514,11 @@ · 14 unmodified lines
564514 "EventName": "PM_MRK_LSU_DERAT_MISS",
565515 "BriefDescription": "Marked derat reload (miss) for any page size"
566516 },
567− {,
568− "EventCode": "0x160A0",
569− "EventName": "PM_L3_PF_MISS_L3",
570− "BriefDescription": "L3 PF missed in L3"
571− },
572517 {,
573518 "EventCode": "0x1C04A",
574519 "EventName": "PM_DATA_FROM_RL2L3_SHR",
575520 "BriefDescription": "The processor's data cache was reloaded with Shared (S) data from another chip's L2 or L3 on the same Node or Group (Remote), as this chip due to a demand load"
576521 },
577− {,
578− "EventCode": "0xD99C",
579− "EventName": "PM_MRK_LSU_FLUSH_UE",
580− "BriefDescription": "Correctable ECC error on reload data, reported at critical data forward time"
581− },
582522 {,
583523 "EventCode": "0x268B0",
584524 "EventName": "PM_L3_P1_GRP_PUMP",
@@ -629,11 +569,6 @@ · 44 unmodified lines
629569 "EventName": "PM_TMA_REQ_L2",
630570 "BriefDescription": "addrs only req to L2 only on the first one,Indication that Load footprint is not expanding"
631571 },
632− {,
633− "EventCode": "0x5884",
634− "EventName": "PM_DECODE_LANES_NOT_AVAIL",
635− "BriefDescription": "Decode has something to transmit but dispatch lanes are not available"
636− },
637572 {,
638573 "EventCode": "0x3C042",
639574 "EventName": "PM_DATA_FROM_L3_DISP_CONFLICT",
@@ -690,9 +625,9 @@ · 50 unmodified lines
690625 "BriefDescription": "False LHS match detected"
691626 },
692627 {,
693− "EventCode": "0xD9A4",
694− "EventName": "PM_MRK_LSU_FLUSH_LARX_STCX",
695− "BriefDescription": "A larx is flushed because an older larx has an LMQ reservation for the same thread. A stcx is flushed because an older stcx is in the LMQ. The flush happens when the older larx/stcx relaunches"
628+ "EventCode": "0xF0B0",
629+ "EventName": "PM_L3_LD_PREF",
630+ "BriefDescription": "L3 load prefetch, sourced from a hardware or software stream, was sent to the nest"
696631 },
697632 {,
698633 "EventCode": "0x4D012",
@@ -715,9 +650,9 @@ · 16 unmodified lines
715650 "BriefDescription": "All successful Ld/St dispatches for this thread that were an L2 miss (excludes i_l2mru_tch_reqs)"
716651 },
717652 {,
718− "EventCode": "0xF8B8",
719− "EventName": "PM_LS1_UNALIGNED_ST",
720− "BriefDescription": "Store instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the Store of that size. If the Store wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
653+ "EventCode": "0x160A0",
654+ "EventName": "PM_L3_PF_MISS_L3",
655+ "BriefDescription": "L3 PF missed in L3"
721656 },
722657 {,
723658 "EventCode": "0x408C",
@@ -764,11 +699,6 @@ · 40 unmodified lines
764699 "EventName": "PM_TM_NESTED_TEND",
765700 "BriefDescription": "Completion time nested tend"
766701 },
767− {,
768− "EventCode": "0x36084",
769− "EventName": "PM_L2_RCST_DISP",
770− "BriefDescription": "All D-side store dispatch attempts for this thread"
771− },
772702 {,
773703 "EventCode": "0x368A0",
774704 "EventName": "PM_L3_PF_OFF_CHIP_CACHE",
@@ -829,11 +759,6 @@ · 54 unmodified lines
829759 "EventName": "PM_L3_SN_USAGE",
830760 "BriefDescription": "Rotating sample of 16 snoop valids"
831761 },
832− {,
833− "EventCode": "0x16084",
834− "EventName": "PM_L2_RCLD_DISP",
835− "BriefDescription": "All I-or-D side load dispatch attempts for this thread (excludes i_l2mru_tch_reqs)"
836− },
837762 {,
838763 "EventCode": "0x1608C",
839764 "EventName": "PM_RC0_BUSY",
@@ -842,7 +767,7 @@ · 2 unmodified lines
842767 {,
843768 "EventCode": "0x36082",
844769 "EventName": "PM_L2_LD_DISP",
845− "BriefDescription": "All successful I-or-D side load dispatches for this thread (excludes i_l2mru_tch_reqs)."
770+ "BriefDescription": "All successful I-or-D side load dispatches for this thread (excludes i_l2mru_tch_reqs)"
846771 },
847772 {,
848773 "EventCode": "0xF8B0",
@@ -904,11 +829,6 @@ · 55 unmodified lines
904829 "EventName": "PM_IC_PREF_REQ",
905830 "BriefDescription": "Instruction prefetch requests"
906831 },
907− {,
908− "EventCode": "0xC898",
909− "EventName": "PM_LS3_UNALIGNED_LD",
910− "BriefDescription": "Load instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the load of that size. If the load wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
911− },
912832 {,
913833 "EventCode": "0x488C",
914834 "EventName": "PM_IC_PREF_WRITE",
@@ -1017,7 +937,7 @@ · 102 unmodified lines
1017937 {,
1018938 "EventCode": "0x3E05E",
1019939 "EventName": "PM_L3_CO_MEPF",
1020− "BriefDescription": "L3 castouts in Mepf state for this thread"
940+ "BriefDescription": "L3 CO of line in Mep state (includes casthrough to memory). The Mepf state indicates that a line was brought in to satisfy an L3 prefetch request"
1021941 },
1022942 {,
1023943 "EventCode": "0x460A2",
@@ -1204,11 +1124,6 @@ · 180 unmodified lines
12041124 "EventName": "PM_TM_FAIL_NON_TX_CONFLICT",
12051125 "BriefDescription": "Non transactional conflict from LSU, gets reported to TEXASR"
12061126 },
1207− {,
1208− "EventCode": "0xD198",
1209− "EventName": "PM_MRK_LSU_FLUSH_ATOMIC",
1210− "BriefDescription": "Quad-word loads (lq) are considered atomic because they always span at least 2 slices. If a snoop or store from another thread changes the data the load is accessing between the 2 or 3 pieces of the lq instruction, the lq will be flushed"
1211− },
12121127 {,
12131128 "EventCode": "0x201E0",
12141129 "EventName": "PM_MRK_DATA_FROM_MEMORY",
@@ -1294,11 +1209,6 @@ · 79 unmodified lines
12941209 "EventName": "PM_ICT_NOSLOT_DISP_HELD_HB_FULL",
12951210 "BriefDescription": "Ict empty for this thread due to dispatch holds because the History Buffer was full. Could be GPR/VSR/VMR/FPR/CR/XVF; CR; XVF (XER/VSCR/FPSCR)"
12961211 },
1297− {,
1298− "EventCode": "0xC894",
1299− "EventName": "PM_LS1_UNALIGNED_LD",
1300− "BriefDescription": "Load instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the load of that size. If the load wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
1301− },
13021212 {,
13031213 "EventCode": "0x360A2",
13041214 "EventName": "PM_L3_L2_CO_HIT",
@@ -1324,11 +1234,6 @@ · 19 unmodified lines
13241234 "EventName": "PM_L2_CASTOUT_SHR",
13251235 "BriefDescription": "L2 Castouts - Shared (Tx,Sx)"
13261236 },
1327− {,
1328− "EventCode": "0xD884",
1329− "EventName": "PM_LSU3_SET_MPRED",
1330− "BriefDescription": "Set prediction(set-p) miss. The entry was not found in the Set prediction table"
1331− },
13321237 {,
13331238 "EventCode": "0x26092",
13341239 "EventName": "PM_L2_LD_MISS_64B",
@@ -1362,12 +1267,12 @@ · 27 unmodified lines
13621267 {,
13631268 "EventCode": "0xD8A8",
13641269 "EventName": "PM_ISLB_MISS",
1365− "BriefDescription": "Instruction SLB miss - Total of all segment sizes"
1270+ "BriefDescription": "Instruction SLB Miss - Total of all segment sizes"
13661271 },
13671272 {,
1368− "EventCode": "0xD19C",
1369− "EventName": "PM_MRK_LSU_FLUSH_RELAUNCH_MISS",
1370− "BriefDescription": "If a load that has already returned data and has to relaunch for any reason then gets a miss (erat, setp, data cache), it will often be flushed at relaunch time because the data might be inconsistent"
1273+ "EventCode": "0x368AE",
1274+ "EventName": "PM_L3_P1_CO_RTY",
1275+ "BriefDescription": "L3 CO received retry port 1 (memory only), every retry counted"
13711276 },
13721277 {,
13731278 "EventCode": "0x260A2",
@@ -1384,6 +1289,11 @@ · 10 unmodified lines
13841289 "EventName": "PM_CMPLU_STALL_NESTED_TBEGIN",
13851290 "BriefDescription": "Completion stall because the ISU is updating the TEXASR to keep track of the nested tbegin. This is a short delay, and it includes ROT"
13861291 },
1292+ {,
1293+ "EventCode": "0xC084",
1294+ "EventName": "PM_LS2_LD_VECTOR_FIN",
1295+ "BriefDescription": ""
1296+ },
13871297 {,
13881298 "EventCode": "0x1608E",
13891299 "EventName": "PM_ST_CAUSED_FAIL",
@@ -1409,11 +1319,6 @@ · 19 unmodified lines
14091319 "EventName": "PM_CO_USAGE",
14101320 "BriefDescription": "Continuous 16 cycle (2to1) window where this signals rotates thru sampling each CO machine busy. PMU uses this wave to then do 16 cyc count to sample total number of machs running"
14111321 },
1412− {,
1413− "EventCode": "0xD084",
1414− "EventName": "PM_LSU2_SET_MPRED",
1415− "BriefDescription": "Set prediction(set-p) miss. The entry was not found in the Set prediction table"
1416− },
14171322 {,
14181323 "EventCode": "0x48B8",
14191324 "EventName": "PM_BR_MPRED_TAKEN_TA",
@@ -1449,30 +1354,25 @@ · 29 unmodified lines
14491354 "EventName": "PM_DC_PREF_STRIDED_CONF",
14501355 "BriefDescription": "A demand load referenced a line in an active strided prefetch stream. The stream could have been allocated through the hardware prefetch mechanism or through software."
14511356 },
1357+ {,
1358+ "EventCode": "0x36084",
1359+ "EventName": "PM_L2_RCST_DISP",
1360+ "BriefDescription": "All D-side store dispatch attempts for this thread"
1361+ },
14521362 {,
14531363 "EventCode": "0x45054",
14541364 "EventName": "PM_FMA_CMPL",
14551365 "BriefDescription": "two flops operation completed (fmadd, fnmadd, fmsub, fnmsub) Scalar instructions only. "
14561366 },
1457− {,
1458− "EventCode": "0x5090",
1459− "EventName": "PM_SHL_ST_DISABLE",
1460− "BriefDescription": "Store-Hit-Load Table Read Hit with entry Disabled (entry was disabled due to the entry shown to not prevent the flush)"
1461− },
14621367 {,
14631368 "EventCode": "0x201E8",
14641369 "EventName": "PM_THRESH_EXC_512",
14651370 "BriefDescription": "Threshold counter exceeded a value of 512"
14661371 },
1467− {,
1468− "EventCode": "0x5084",
1469− "EventName": "PM_DECODE_FUSION_EXT_ADD",
1470− "BriefDescription": "32-bit extended addition"
1471− },
14721372 {,
14731373 "EventCode": "0x36080",
14741374 "EventName": "PM_L2_INST",
1475− "BriefDescription": "All successful I-side dispatches for this thread (excludes i_l2mru_tch reqs)."
1375+ "BriefDescription": "All successful I-side dispatches for this thread (excludes i_l2mru_tch reqs)"
14761376 },
14771377 {,
14781378 "EventCode": "0x3504C",
@@ -1554,21 +1454,11 @@ · 75 unmodified lines
15541454 "EventName": "PM_MEM_RWITM",
15551455 "BriefDescription": "Memory Read With Intent to Modify for this thread"
15561456 },
1557− {,
1558− "EventCode": "0x26882",
1559− "EventName": "PM_L2_DC_INV",
1560− "BriefDescription": "D-cache invalidates sent over the reload bus to the core"
1561− },
15621457 {,
15631458 "EventCode": "0xC090",
15641459 "EventName": "PM_LSU_STCX",
15651460 "BriefDescription": "STCX sent to nest, i.e. total"
15661461 },
1567− {,
1568− "EventCode": "0xD080",
1569− "EventName": "PM_LSU0_SET_MPRED",
1570− "BriefDescription": "Set prediction(set-p) miss. The entry was not found in the Set prediction table"
1571− },
15721462 {,
15731463 "EventCode": "0x2C120",
15741464 "EventName": "PM_MRK_DATA_FROM_L2_NO_CONFLICT",
@@ -1609,11 +1499,6 @@ · 34 unmodified lines
16091499 "EventName": "PM_IPTEG_FROM_L2_NO_CONFLICT",
16101500 "BriefDescription": "A Page Table Entry was loaded into the TLB from local core's L2 without conflict due to a instruction side request"
16111501 },
1612− {,
1613− "EventCode": "0xD9A0",
1614− "EventName": "PM_MRK_LSU_FLUSH_LHL_SHL",
1615− "BriefDescription": "The instruction was flushed because of a sequential load/store consistency. If a load or store hits on an older load that has either been snooped (for loads) or has stale data (for stores)."
1616− },
16171502 {,
16181503 "EventCode": "0x35042",
16191504 "EventName": "PM_IPTEG_FROM_L3_DISP_CONFLICT",
@@ -1692,7 +1577,7 @@ · 72 unmodified lines
16921577 {,
16931578 "EventCode": "0x2001A",
16941579 "EventName": "PM_NTC_ALL_FIN",
1695− "BriefDescription": "Cycles after all instructions have finished to group completed"
1580+ "BriefDescription": "Cycles after instruction finished to instruction completed."
16961581 },
16971582 {,
16981583 "EventCode": "0x3005A",
@@ -1709,6 +1594,11 @@ · 10 unmodified lines
17091594 "EventName": "PM_LSU1_L1_CAM_CANCEL",
17101595 "BriefDescription": "ls1 l1 tm cam cancel"
17111596 },
1597+ {,
1598+ "EventCode": "0x268AE",
1599+ "EventName": "PM_L3_P3_PF_RTY",
1600+ "BriefDescription": "L3 PF received retry port 3, every retry counted"
1601+ },
17121602 {,
17131603 "EventCode": "0xE884",
17141604 "EventName": "PM_LS1_ERAT_MISS_PREF",
@@ -1742,7 +1632,7 @@ · 27 unmodified lines
17421632 {,
17431633 "EventCode": "0x160B6",
17441634 "EventName": "PM_L3_WI0_BUSY",
1745− "BriefDescription": "Rotating sample of 8 WI valid"
1635+ "BriefDescription": "Rotating sample of 8 WI valid (duplicate)"
17461636 },
17471637 {,
17481638 "EventCode": "0x368AC",
@@ -1790,9 +1680,9 @@ · 41 unmodified lines
17901680 "BriefDescription": "L2 guess system (VGS or RNS) and guess was correct (ie data beyond-group)"
17911681 },
17921682 {,
1793− "EventCode": "0x589C",
1794− "EventName": "PM_PTESYNC",
1795− "BriefDescription": "ptesync instruction counted when the instruction is decoded and transmitted"
1683+ "EventCode": "0x260AE",
1684+ "EventName": "PM_L3_P2_PF_RTY",
1685+ "BriefDescription": "L3 PF received retry port 2, every retry counted"
17961686 },
17971687 {,
17981688 "EventCode": "0x26086",
@@ -1824,6 +1714,11 @@ · 25 unmodified lines
18241714 "EventName": "PM_SHL_ST_DEP_CREATED",
18251715 "BriefDescription": "Store-Hit-Load Table Read Hit with entry Enabled"
18261716 },
1717+ {,
1718+ "EventCode": "0x46882",
1719+ "EventName": "PM_L2_ST_HIT",
1720+ "BriefDescription": "All successful D-side store dispatches for this thread that were L2 hits"
1721+ },
18271722 {,
18281723 "EventCode": "0x360AC",
18291724 "EventName": "PM_L3_SN0_BUSY",
@@ -1844,11 +1739,6 @@ · 14 unmodified lines
18441739 "EventName": "PM_L2_ST_MISS",
18451740 "BriefDescription": "All successful D-Side Store dispatches that were an L2 miss for this thread"
18461741 },
1847− {,
1848− "EventCode": "0xF8B4",
1849− "EventName": "PM_DC_PREF_XCONS_ALLOC",
1850− "BriefDescription": "Prefetch stream allocated in the Ultra conservative phase by either the hardware prefetch mechanism or software prefetch"
1851− },
18521742 {,
18531743 "EventCode": "0x35048",
18541744 "EventName": "PM_IPTEG_FROM_DL2L3_SHR",
@@ -1969,11 +1859,6 @@ · 114 unmodified lines
19691859 "EventName": "PM_THRD_PRIO_2_3_CYC",
19701860 "BriefDescription": "Cycles thread running at priority level 2 or 3"
19711861 },
1972− {,
1973− "EventCode": "0x10134",
1974− "EventName": "PM_MRK_ST_DONE_L2",
1975− "BriefDescription": "marked store completed in L2 ( RC machine done)"
1976− },
19771862 {,
19781863 "EventCode": "0x368B2",
19791864 "EventName": "PM_L3_GRP_GUESS_WRONG_HIGH",
@@ -2004,11 +1889,6 @@ · 24 unmodified lines
20041889 "EventName": "PM_L2_GRP_GUESS_WRONG",
20051890 "BriefDescription": "L2 guess grp (GS or NNS) and guess was not correct (ie data on-chip OR beyond-group)"
20061891 },
2007− {,
2008− "EventCode": "0x368AE",
2009− "EventName": "PM_L3_P1_CO_RTY",
2010− "BriefDescription": "L3 CO received retry port 1 (memory only), every retry counted"
2011− },
20121892 {,
20131893 "EventCode": "0xC0AC",
20141894 "EventName": "PM_LSU_FLUSH_EMSH",
@@ -2034,11 +1914,6 @@ · 19 unmodified lines
20341914 "EventName": "PM_L2_GROUP_PUMP",
20351915 "BriefDescription": "RC requests that were on group (aka nodel) pump attempts"
20361916 },
2037− {,
2038− "EventCode": "0xF0B0",
2039− "EventName": "PM_L3_LD_PREF",
2040− "BriefDescription": "L3 load prefetch, sourced from a hardware or software stream, was sent to the nest"
2041− },
20421917 {,
20431918 "EventCode": "0x16080",
20441919 "EventName": "PM_L2_LD",
@@ -2049,6 +1924,11 @@ · 4 unmodified lines
20491924 "EventName": "PM_MATH_FLOP_CMPL",
20501925 "BriefDescription": "Math flop instruction completed"
20511926 },
1927+ {,
1928+ "EventCode": "0xC080",
1929+ "EventName": "PM_LS0_LD_VECTOR_FIN",
1930+ "BriefDescription": ""
1931+ },
20521932 {,
20531933 "EventCode": "0x368B0",
20541934 "EventName": "PM_L3_P1_SYS_PUMP",
@@ -2119,11 +1999,6 @@ · 64 unmodified lines
21191999 "EventName": "PM_BR_CORECT_PRED_TAKEN_CMPL",
21202000 "BriefDescription": "Conditional Branch Completed in which the HW correctly predicted the direction as taken. Counted at completion time"
21212001 },
2122− {,
2123− "EventCode": "0xF0B8",
2124− "EventName": "PM_LS0_UNALIGNED_ST",
2125− "BriefDescription": "Store instructions whose data crosses a double-word boundary, which causes it to require an additional slice than than what normally would be required of the Store of that size. If the Store wraps from slice 3 to slice 0, thee is an additional 3-cycle penalty"
2126− },
21272002 {,
21282003 "EventCode": "0x20132",
21292004 "EventName": "PM_MRK_DFU_FIN",
@@ -2139,6 +2014,11 @@ · 9 unmodified lines
21392014 "EventName": "PM_LSU_FLUSH_LHS",
21402015 "BriefDescription": "Effective Address alias flush : no EA match but Real Address match. If the data has not yet been returned for this load, the instruction will just be rejected, but if it has returned data, it will be flushed"
21412016 },
2017+ {,
2018+ "EventCode": "0x16084",
2019+ "EventName": "PM_L2_RCLD_DISP",
2020+ "BriefDescription": "All I-or-D side load dispatch attempts for this thread (excludes i_l2mru_tch_reqs)"
2021+ },
21422022 {,
21432023 "EventCode": "0x3F150",
21442024 "EventName": "PM_MRK_ST_DRAIN_TO_L2DISP_CYC",
@@ -2224,11 +2104,6 @@ · 79 unmodified lines
22242104 "EventName": "PM_IC_PREF_CANCEL_PAGE",
22252105 "BriefDescription": "Prefetch Canceled due to page boundary"
22262106 },
2227− {,
2228− "EventCode": "0xF09C",
2229− "EventName": "PM_SLB_TABLEWALK_CYC",
2230− "BriefDescription": "Cycles when a tablewalk is pending on this thread on the SLB table"
2231− },
22322107 {,
22332108 "EventCode": "0x460AA",
22342109 "EventName": "PM_L3_P0_CO_L31",
@@ -2247,10 +2122,10 @@ · 12 unmodified lines
22472122 {,
22482123 "EventCode": "0x46082",
22492124 "EventName": "PM_L2_ST_DISP",
2250− "BriefDescription": "All successful D-side store dispatches for this thread "
2125+ "BriefDescription": "All successful D-side store dispatches for this thread (L2 miss + L2 hits)"
22512126 },
22522127 {,
2253− "EventCode": "0x4609E",
2128+ "EventCode": "0x36880",
22542129 "EventName": "PM_L2_INST_MISS",
22552130 "BriefDescription": "All successful I-side dispatches that were an L2 miss for this thread (excludes i_l2mru_tch reqs)"
22562131 },
@@ -2340,9 +2215,9 @@ · 83 unmodified lines
23402215 "BriefDescription": "All ISU rejects"
23412216 },
23422217 {,
2343− "EventCode": "0x46882",
2344− "EventName": "PM_L2_ST_HIT",
2345− "BriefDescription": "All successful D-side store dispatches for this thread that were L2 hits"
2218+ "EventCode": "0xC884",
2219+ "EventName": "PM_LS3_LD_VECTOR_FIN",
2220+ "BriefDescription": ""
23462221 },
23472222 {,
23482223 "EventCode": "0x360A8",
@@ -2359,11 +2234,6 @@ · 10 unmodified lines
23592234 "EventName": "PM_LSU_NCST",
23602235 "BriefDescription": "Asserts when a i=1 store op is sent to the nest. No record of issue pipe (LS0/LS1) is maintained so this is for both pipes. Probably don't need separate LS0 and LS1"
23612236 },
2362− {,
2363− "EventCode": "0xD880",
2364− "EventName": "PM_LSU1_SET_MPRED",
2365− "BriefDescription": "Set prediction(set-p) miss. The entry was not found in the Set prediction table"
2366− },
23672237 {,
23682238 "EventCode": "0xD0B8",
23692239 "EventName": "PM_LSU_LMQ_FULL_CYC",
@@ -2389,4 +2259,4 @@ · 19 unmodified lines
23892259 "EventName": "PM_L3_PF_USAGE",
23902260 "BriefDescription": "Rotating sample of 32 PF actives"
23912261 }
2392−]
2262+]
tools/perf/pmu-events/arch/powerpc/power9/pipeline.json+7−7
@@ -124,6 +124,11 @@ · 123 unmodified lines
124124 "EventName": "PM_PMC5_OVERFLOW",
125125 "BriefDescription": "Overflow from counter 5"
126126 },
127+ {,
128+ "EventCode": "0x4505E",
129+ "EventName": "PM_FLOP_CMPL",
130+ "BriefDescription": "Floating Point Operation Finished"
131+ },
127132 {,
128133 "EventCode": "0x2C018",
129134 "EventName": "PM_CMPLU_STALL_DMISS_L21_L31",
@@ -389,11 +394,6 @@ · 259 unmodified lines
389394 "EventName": "PM_ICT_NOSLOT_BR_MPRED",
390395 "BriefDescription": "Ict empty for this thread due to branch mispred"
391396 },
392− {,
393− "EventCode": "0x3405E",
394− "EventName": "PM_IFETCH_THROTTLE",
395− "BriefDescription": "Cycles in which Instruction fetch throttle was active."
396− },
397397 {,
398398 "EventCode": "0x1F148",
399399 "EventName": "PM_MRK_DPTEG_FROM_ON_CHIP_CACHE",
@@ -422,7 +422,7 @@ · 22 unmodified lines
422422 {,
423423 "EventCode": "0xD0A8",
424424 "EventName": "PM_DSLB_MISS",
425− "BriefDescription": "Data SLB Miss - Total of all segment sizes"
425+ "BriefDescription": "gate_and(sd_pc_c0_comp_valid AND sd_pc_c0_comp_thread(0:1)=tid,sd_pc_c0_comp_ppc_count(0:3)) + gate_and(sd_pc_c1_comp_valid AND sd_pc_c1_comp_thread(0:1)=tid,sd_pc_c1_comp_ppc_count(0:3))"
426426 },
427427 {,
428428 "EventCode": "0x4C058",
@@ -549,4 +549,4 @@ · 120 unmodified lines
549549 "EventName": "PM_MRK_DATA_FROM_L21_SHR_CYC",
550550 "BriefDescription": "Duration in cycles to reload with Shared (S) data from another core's L2 on the same chip due to a marked load"
551551 }
552−]
552+]
tools/perf/pmu-events/arch/powerpc/power9/pmc.json+1−1
@@ -119,4 +119,4 @@ · 118 unmodified lines
119119 "EventName": "PM_1FLOP_CMPL",
120120 "BriefDescription": "one flop (fadd, fmul, fsub, fcmp, fsel, fabs, fnabs, fres, fsqrte, fneg) operation completed"
121121 }
122−]
122+]
tools/perf/pmu-events/arch/powerpc/power9/translation.json+0−5
@@ -89,11 +89,6 @@ · 88 unmodified lines
8989 "EventName": "PM_STCX_FAIL",
9090 "BriefDescription": "stcx failed"
9191 },
92− {,
93− "EventCode": "0x20112",
94− "EventName": "PM_MRK_NTF_FIN",
95− "BriefDescription": "Marked next to finish instruction finished"
96− },
9792 {,
9893 "EventCode": "0x300F0",
9994 "EventName": "PM_ST_MISS_L1",
tools/perf/pmu-events/arch/x86/broadwell/cache.json+378−177
@@ -10,13 +10,30 @@ · 9 unmodified lines
1010 "CounterHTOff": "0,1,2,3,4,5,6,7"
1111 },
1212 {
13− "PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
1413 "EventCode": "0x24",
1514 "Counter": "0,1,2,3",
16− "UMask": "0x41",
17− "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
15+ "UMask": "0x22",
16+ "EventName": "L2_RQSTS.RFO_MISS",
1817 "SampleAfterValue": "200003",
19− "BriefDescription": "Demand Data Read requests that hit L2 cache",
18+ "BriefDescription": "RFO requests that miss L2 cache.",
19+ "CounterHTOff": "0,1,2,3,4,5,6,7"
20+ },
21+ {
22+ "EventCode": "0x24",
23+ "Counter": "0,1,2,3",
24+ "UMask": "0x24",
25+ "EventName": "L2_RQSTS.CODE_RD_MISS",
26+ "SampleAfterValue": "200003",
27+ "BriefDescription": "L2 cache misses when fetching instructions.",
28+ "CounterHTOff": "0,1,2,3,4,5,6,7"
29+ },
30+ {
31+ "EventCode": "0x24",
32+ "Counter": "0,1,2,3",
33+ "UMask": "0x27",
34+ "EventName": "L2_RQSTS.ALL_DEMAND_MISS",
35+ "SampleAfterValue": "200003",
36+ "BriefDescription": "Demand requests that miss L2 cache.",
2037 "CounterHTOff": "0,1,2,3,4,5,6,7"
2138 },
2239 {
@@ -29,6 +46,43 @@ · 6 unmodified lines
2946 "BriefDescription": "L2 prefetch requests that miss L2 cache",
3047 "CounterHTOff": "0,1,2,3,4,5,6,7"
3148 },
49+ {
50+ "EventCode": "0x24",
51+ "Counter": "0,1,2,3",
52+ "UMask": "0x3f",
53+ "EventName": "L2_RQSTS.MISS",
54+ "SampleAfterValue": "200003",
55+ "BriefDescription": "All requests that miss L2 cache.",
56+ "CounterHTOff": "0,1,2,3,4,5,6,7"
57+ },
58+ {
59+ "PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
60+ "EventCode": "0x24",
61+ "Counter": "0,1,2,3",
62+ "UMask": "0x41",
63+ "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
64+ "SampleAfterValue": "200003",
65+ "BriefDescription": "Demand Data Read requests that hit L2 cache",
66+ "CounterHTOff": "0,1,2,3,4,5,6,7"
67+ },
68+ {
69+ "EventCode": "0x24",
70+ "Counter": "0,1,2,3",
71+ "UMask": "0x42",
72+ "EventName": "L2_RQSTS.RFO_HIT",
73+ "SampleAfterValue": "200003",
74+ "BriefDescription": "RFO requests that hit L2 cache.",
75+ "CounterHTOff": "0,1,2,3,4,5,6,7"
76+ },
77+ {
78+ "EventCode": "0x24",
79+ "Counter": "0,1,2,3",
80+ "UMask": "0x44",
81+ "EventName": "L2_RQSTS.CODE_RD_HIT",
82+ "SampleAfterValue": "200003",
83+ "BriefDescription": "L2 cache hits when fetching instructions, code reads.",
84+ "CounterHTOff": "0,1,2,3,4,5,6,7"
85+ },
3286 {
3387 "PublicDescription": "This event counts the number of requests from the L2 hardware prefetchers that hit L2 cache. L3 prefetch new types.",
3488 "EventCode": "0x24",
@@ -69,6 +123,15 @@ · 34 unmodified lines
69123 "BriefDescription": "L2 code requests",
70124 "CounterHTOff": "0,1,2,3,4,5,6,7"
71125 },
126+ {
127+ "EventCode": "0x24",
128+ "Counter": "0,1,2,3",
129+ "UMask": "0xe7",
130+ "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES",
131+ "SampleAfterValue": "200003",
132+ "BriefDescription": "Demand requests to L2 cache.",
133+ "CounterHTOff": "0,1,2,3,4,5,6,7"
134+ },
72135 {
73136 "PublicDescription": "This event counts the total number of requests from the L2 hardware prefetchers.",
74137 "EventCode": "0x24",
@@ -79,6 +142,15 @@ · 4 unmodified lines
79142 "BriefDescription": "Requests from L2 hardware prefetchers",
80143 "CounterHTOff": "0,1,2,3,4,5,6,7"
81144 },
145+ {
146+ "EventCode": "0x24",
147+ "Counter": "0,1,2,3",
148+ "UMask": "0xff",
149+ "EventName": "L2_RQSTS.REFERENCES",
150+ "SampleAfterValue": "200003",
151+ "BriefDescription": "All L2 requests.",
152+ "CounterHTOff": "0,1,2,3,4,5,6,7"
153+ },
82154 {
83155 "PublicDescription": "This event counts the number of WB requests that hit L2 cache.",
84156 "EventCode": "0x27",
@@ -130,6 +202,27 @@ · 45 unmodified lines
130202 "CounterMask": "1",
131203 "CounterHTOff": "2"
132204 },
205+ {
206+ "EventCode": "0x48",
207+ "Counter": "2",
208+ "UMask": "0x1",
209+ "AnyThread": "1",
210+ "EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
211+ "SampleAfterValue": "2000003",
212+ "BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
213+ "CounterMask": "1",
214+ "CounterHTOff": "2"
215+ },
216+ {
217+ "EventCode": "0x48",
218+ "Counter": "0,1,2,3",
219+ "UMask": "0x2",
220+ "EventName": "L1D_PEND_MISS.FB_FULL",
221+ "SampleAfterValue": "2000003",
222+ "BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
223+ "CounterMask": "1",
224+ "CounterHTOff": "0,1,2,3,4,5,6,7"
225+ },
133226 {
134227 "PublicDescription": "This event counts L1D data line replacements including opportunistic replacements, and replacements that require stall-for-replace or block-for-replace.",
135228 "EventCode": "0x51",
@@ -152,7 +245,30 @@ · 16 unmodified lines
152245 "CounterHTOff": "0,1,2,3,4,5,6,7"
153246 },
154247 {
155− "PublicDescription": "This event counts the number of offcore outstanding Code Reads transactions in the super queue every cycle. The \"Offcore outstanding\" state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
248+ "PublicDescription": "This event counts cycles when offcore outstanding Demand Data Read transactions are present in the super queue (SQ). A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation).",
249+ "EventCode": "0x60",
250+ "Counter": "0,1,2,3",
251+ "UMask": "0x1",
252+ "Errata": "BDM76",
253+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
254+ "SampleAfterValue": "2000003",
255+ "BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore",
256+ "CounterMask": "1",
257+ "CounterHTOff": "0,1,2,3,4,5,6,7"
258+ },
259+ {
260+ "EventCode": "0x60",
261+ "Counter": "0,1,2,3",
262+ "UMask": "0x1",
263+ "Errata": "BDM76",
264+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_GE_6",
265+ "SampleAfterValue": "2000003",
266+ "BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
267+ "CounterMask": "6",
268+ "CounterHTOff": "0,1,2,3,4,5,6,7"
269+ },
270+ {
271+ "PublicDescription": "This event counts the number of offcore outstanding Code Reads transactions in the super queue every cycle. The Offcore outstanding state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
156272 "EventCode": "0x60",
157273 "Counter": "0,1,2,3",
158274 "UMask": "0x2",
@@ -174,26 +290,26 @@ · 15 unmodified lines
174290 "CounterHTOff": "0,1,2,3,4,5,6,7"
175291 },
176292 {
177− "PublicDescription": "This event counts the number of offcore outstanding cacheable Core Data Read transactions in the super queue every cycle. A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation). See corresponding Umask under OFFCORE_REQUESTS.",
293+ "PublicDescription": "This event counts the number of offcore outstanding demand rfo Reads transactions in the super queue every cycle. The Offcore outstanding state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
178294 "EventCode": "0x60",
179295 "Counter": "0,1,2,3",
180− "UMask": "0x8",
296+ "UMask": "0x4",
181297 "Errata": "BDM76",
182− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
298+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
183299 "SampleAfterValue": "2000003",
184− "BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore",
300+ "BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle",
301+ "CounterMask": "1",
185302 "CounterHTOff": "0,1,2,3,4,5,6,7"
186303 },
187304 {
188− "PublicDescription": "This event counts cycles when offcore outstanding Demand Data Read transactions are present in the super queue (SQ). A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation).",
305+ "PublicDescription": "This event counts the number of offcore outstanding cacheable Core Data Read transactions in the super queue every cycle. A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation). See corresponding Umask under OFFCORE_REQUESTS.",
189306 "EventCode": "0x60",
190307 "Counter": "0,1,2,3",
191− "UMask": "0x1",
308+ "UMask": "0x8",
192309 "Errata": "BDM76",
193− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
310+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
194311 "SampleAfterValue": "2000003",
195− "BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore",
196− "CounterMask": "1",
312+ "BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore",
197313 "CounterHTOff": "0,1,2,3,4,5,6,7"
198314 },
199315 {
@@ -208,18 +324,6 @@ · 8 unmodified lines
208324 "CounterMask": "1",
209325 "CounterHTOff": "0,1,2,3,4,5,6,7"
210326 },
211− {
212− "PublicDescription": "This event counts the number of offcore outstanding demand rfo Reads transactions in the super queue every cycle. The \"Offcore outstanding\" state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
213− "EventCode": "0x60",
214− "Counter": "0,1,2,3",
215− "UMask": "0x4",
216− "Errata": "BDM76",
217− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
218− "SampleAfterValue": "2000003",
219− "BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle",
220− "CounterMask": "1",
221− "CounterHTOff": "0,1,2,3,4,5,6,7"
222− },
223327 {
224328 "PublicDescription": "This event counts the number of cycles when the L1D is locked. It is a superset of the 0x1 mask (BUS_LOCK_CLOCKS.BUS_LOCK_DURATION).",
225329 "EventCode": "0x63",
@@ -261,7 +365,7 @@ · 35 unmodified lines
261365 "CounterHTOff": "0,1,2,3,4,5,6,7"
262366 },
263367 {
264− "PublicDescription": "This event counts the demand and prefetch data reads. All Core Data Reads include cacheable \"Demands\" and L2 prefetchers (not L3 prefetchers). Counting also covers reads due to page walks resulted from any request type.",
368+ "PublicDescription": "This event counts the demand and prefetch data reads. All Core Data Reads include cacheable Demands and L2 prefetchers (not L3 prefetchers). Counting also covers reads due to page walks resulted from any request type.",
265369 "EventCode": "0xB0",
266370 "Counter": "0,1,2,3",
267371 "UMask": "0x8",
@@ -280,153 +384,162 @@ · 12 unmodified lines
280384 "BriefDescription": "Offcore requests buffer cannot take more entries for this thread core.",
281385 "CounterHTOff": "0,1,2,3,4,5,6,7"
282386 },
387+ {
388+ "EventCode": "0xB7, 0xBB",
389+ "Counter": "0,1,2,3",
390+ "UMask": "0x1",
391+ "EventName": "OFFCORE_RESPONSE",
392+ "SampleAfterValue": "100003",
393+ "BriefDescription": "Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
394+ "CounterHTOff": "0,1,2,3"
395+ },
283396 {
284397 "PEBS": "1",
285− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
398+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
286399 "EventCode": "0xD0",
287400 "Counter": "0,1,2,3",
288401 "UMask": "0x11",
289402 "EventName": "MEM_UOPS_RETIRED.STLB_MISS_LOADS",
290403 "SampleAfterValue": "100003",
291− "BriefDescription": "Retired load uops that miss the STLB.",
404+ "BriefDescription": "Retired load uops that miss the STLB. (Precise Event - PEBS)",
292405 "CounterHTOff": "0,1,2,3",
293406 "Data_LA": "1"
294407 },
295408 {
296409 "PEBS": "1",
297− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts store uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
410+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
298411 "EventCode": "0xD0",
299412 "Counter": "0,1,2,3",
300413 "UMask": "0x12",
301414 "EventName": "MEM_UOPS_RETIRED.STLB_MISS_STORES",
302415 "SampleAfterValue": "100003",
303− "BriefDescription": "Retired store uops that miss the STLB.",
416+ "BriefDescription": "Retired store uops that miss the STLB. (Precise Event - PEBS)",
304417 "CounterHTOff": "0,1,2,3",
305418 "Data_LA": "1",
306419 "L1_Hit_Indication": "1"
307420 },
308421 {
309422 "PEBS": "1",
310− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts load uops with locked access retired to the architected path.",
423+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with locked access retired to the architected path.",
311424 "EventCode": "0xD0",
312425 "Counter": "0,1,2,3",
313426 "UMask": "0x21",
314427 "Errata": "BDM35",
315428 "EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
316429 "SampleAfterValue": "100007",
317− "BriefDescription": "Retired load uops with locked access.",
430+ "BriefDescription": "Retired load uops with locked access. (Precise Event - PEBS)",
318431 "CounterHTOff": "0,1,2,3",
319432 "Data_LA": "1"
320433 },
321434 {
322435 "PEBS": "1",
323− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
436+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
324437 "EventCode": "0xD0",
325438 "Counter": "0,1,2,3",
326439 "UMask": "0x41",
327440 "EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
328441 "SampleAfterValue": "100003",
329− "BriefDescription": "Retired load uops that split across a cacheline boundary.",
442+ "BriefDescription": "Retired load uops that split across a cacheline boundary.(Precise Event - PEBS)",
330443 "CounterHTOff": "0,1,2,3",
331444 "Data_LA": "1"
332445 },
333446 {
334447 "PEBS": "1",
335− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
448+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
336449 "EventCode": "0xD0",
337450 "Counter": "0,1,2,3",
338451 "UMask": "0x42",
339452 "EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
340453 "SampleAfterValue": "100003",
341− "BriefDescription": "Retired store uops that split across a cacheline boundary.",
454+ "BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
342455 "CounterHTOff": "0,1,2,3",
343456 "Data_LA": "1",
344457 "L1_Hit_Indication": "1"
345458 },
346459 {
347460 "PEBS": "1",
348− "PublicDescription": "This event counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
461+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
349462 "EventCode": "0xD0",
350463 "Counter": "0,1,2,3",
351464 "UMask": "0x81",
352465 "EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
353466 "SampleAfterValue": "2000003",
354− "BriefDescription": "All retired load uops.",
467+ "BriefDescription": "All retired load uops. (Precise Event - PEBS)",
355468 "CounterHTOff": "0,1,2,3",
356469 "Data_LA": "1"
357470 },
358471 {
359472 "PEBS": "1",
360− "PublicDescription": "This event counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
473+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
361474 "EventCode": "0xD0",
362475 "Counter": "0,1,2,3",
363476 "UMask": "0x82",
364477 "EventName": "MEM_UOPS_RETIRED.ALL_STORES",
365478 "SampleAfterValue": "2000003",
366− "BriefDescription": "All retired store uops.",
479+ "BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
367480 "CounterHTOff": "0,1,2,3",
368481 "Data_LA": "1",
369482 "L1_Hit_Indication": "1"
370483 },
371484 {
372485 "PEBS": "1",
373− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
486+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data source were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
374487 "EventCode": "0xD1",
375488 "Counter": "0,1,2,3",
376489 "UMask": "0x1",
377490 "EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
378491 "SampleAfterValue": "2000003",
379− "BriefDescription": "Retired load uops with L1 cache hits as data sources.",
492+ "BriefDescription": "Retired load uops with L1 cache hits as data sources. (Precise Event - PEBS)",
380493 "CounterHTOff": "0,1,2,3",
381494 "Data_LA": "1"
382495 },
383496 {
384497 "PEBS": "1",
385− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
498+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
386499 "EventCode": "0xD1",
387500 "Counter": "0,1,2,3",
388501 "UMask": "0x2",
389502 "Errata": "BDM35",
390503 "EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
391504 "SampleAfterValue": "100003",
392− "BriefDescription": "Retired load uops with L2 cache hits as data sources.",
505+ "BriefDescription": "Retired load uops with L2 cache hits as data sources. (Precise Event - PEBS)",
393506 "CounterHTOff": "0,1,2,3",
394507 "Data_LA": "1"
395508 },
396509 {
397510 "PEBS": "1",
398− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
511+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
399512 "EventCode": "0xD1",
400513 "Counter": "0,1,2,3",
401514 "UMask": "0x4",
402515 "Errata": "BDM100",
403516 "EventName": "MEM_LOAD_UOPS_RETIRED.L3_HIT",
404517 "SampleAfterValue": "50021",
405− "BriefDescription": "Retired load uops which data sources were data hits in L3 without snoops required.",
518+ "BriefDescription": "Hit in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS)",
406519 "CounterHTOff": "0,1,2,3",
407520 "Data_LA": "1"
408521 },
409522 {
410523 "PEBS": "1",
411− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
524+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
412525 "EventCode": "0xD1",
413526 "Counter": "0,1,2,3",
414527 "UMask": "0x8",
415528 "EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
416529 "SampleAfterValue": "100003",
417− "BriefDescription": "Retired load uops misses in L1 cache as data sources.",
530+ "BriefDescription": "Retired load uops misses in L1 cache as data sources. Uses PEBS.",
418531 "CounterHTOff": "0,1,2,3",
419532 "Data_LA": "1"
420533 },
421534 {
422535 "PEBS": "1",
423− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
536+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
424537 "EventCode": "0xD1",
425538 "Counter": "0,1,2,3",
426539 "UMask": "0x10",
427540 "EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
428541 "SampleAfterValue": "50021",
429− "BriefDescription": "Miss in mid-level (L2) cache. Excludes Unknown data-source.",
542+ "BriefDescription": "Retired load uops with L2 cache misses as data sources. Uses PEBS.",
430543 "CounterHTOff": "0,1,2,3",
431544 "Data_LA": "1"
432545 },
@@ -438,84 +551,83 @@ · 5 unmodified lines
438551 "Errata": "BDM100, BDE70",
439552 "EventName": "MEM_LOAD_UOPS_RETIRED.L3_MISS",
440553 "SampleAfterValue": "100007",
441− "BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source.",
554+ "BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS).",
442555 "CounterHTOff": "0,1,2,3",
443556 "Data_LA": "1"
444557 },
445558 {
446559 "PEBS": "1",
447− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
560+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
448561 "EventCode": "0xD1",
449562 "Counter": "0,1,2,3",
450563 "UMask": "0x40",
451564 "EventName": "MEM_LOAD_UOPS_RETIRED.HIT_LFB",
452565 "SampleAfterValue": "100003",
453− "BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready.",
566+ "BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready. (Precise Event - PEBS)",
454567 "CounterHTOff": "0,1,2,3",
455568 "Data_LA": "1"
456569 },
457570 {
458571 "PEBS": "1",
459− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
572+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
460573 "EventCode": "0xD2",
461574 "Counter": "0,1,2,3",
462575 "UMask": "0x1",
463576 "Errata": "BDM100",
464577 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS",
465578 "SampleAfterValue": "20011",
466− "BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache.",
579+ "BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache. (Precise Event - PEBS)",
467580 "CounterHTOff": "0,1,2,3",
468581 "Data_LA": "1"
469582 },
470583 {
471584 "PEBS": "1",
472− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
585+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
473586 "EventCode": "0xD2",
474587 "Counter": "0,1,2,3",
475588 "UMask": "0x2",
476589 "Errata": "BDM100",
477590 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT",
478591 "SampleAfterValue": "20011",
479− "BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache.",
592+ "BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache. (Precise Event - PEBS)",
480593 "CounterHTOff": "0,1,2,3",
481594 "Data_LA": "1"
482595 },
483596 {
484597 "PEBS": "1",
485− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
598+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
486599 "EventCode": "0xD2",
487600 "Counter": "0,1,2,3",
488601 "UMask": "0x4",
489602 "Errata": "BDM100",
490603 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM",
491604 "SampleAfterValue": "20011",
492− "BriefDescription": "Retired load uops which data sources were HitM responses from shared L3.",
605+ "BriefDescription": "Retired load uops which data sources were HitM responses from shared L3. (Precise Event - PEBS)",
493606 "CounterHTOff": "0,1,2,3",
494607 "Data_LA": "1"
495608 },
496609 {
497610 "PEBS": "1",
498− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
611+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
499612 "EventCode": "0xD2",
500613 "Counter": "0,1,2,3",
501614 "UMask": "0x8",
502615 "Errata": "BDM100",
503616 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_NONE",
504617 "SampleAfterValue": "100003",
505− "BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required.",
618+ "BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required. (Precise Event - PEBS)",
506619 "CounterHTOff": "0,1,2,3",
507620 "Data_LA": "1"
508621 },
509622 {
510623 "PEBS": "1",
511− "PublicDescription": "Retired load uop whose Data Source was: local DRAM either Snoop not needed or Snoop Miss (RspI).",
624+ "PublicDescription": "This event counts retired load uops where the data came from local DRAM. This does not include hardware prefetches. This is a precise event.",
512625 "EventCode": "0xD3",
513626 "Counter": "0,1,2,3",
514627 "UMask": "0x1",
515628 "Errata": "BDE70, BDM100",
516629 "EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM",
517630 "SampleAfterValue": "100007",
518− "BriefDescription": "Data from local DRAM either Snoop not needed or Snoop Miss (RspI)",
519631 "CounterHTOff": "0,1,2,3",
520632 "Data_LA": "1"
521633 },
@@ -659,119 +771,7 @@ · 137 unmodified lines
659771 "CounterHTOff": "0,1,2,3,4,5,6,7"
660772 },
661773 {
662− "EventCode": "0x24",
663− "Counter": "0,1,2,3",
664− "UMask": "0x42",
665− "EventName": "L2_RQSTS.RFO_HIT",
666− "SampleAfterValue": "200003",
667− "BriefDescription": "RFO requests that hit L2 cache.",
668− "CounterHTOff": "0,1,2,3,4,5,6,7"
669− },
670− {
671− "EventCode": "0x24",
672− "Counter": "0,1,2,3",
673− "UMask": "0x22",
674− "EventName": "L2_RQSTS.RFO_MISS",
675− "SampleAfterValue": "200003",
676− "BriefDescription": "RFO requests that miss L2 cache.",
677− "CounterHTOff": "0,1,2,3,4,5,6,7"
678− },
679− {
680− "EventCode": "0x24",
681− "Counter": "0,1,2,3",
682− "UMask": "0x44",
683− "EventName": "L2_RQSTS.CODE_RD_HIT",
684− "SampleAfterValue": "200003",
685− "BriefDescription": "L2 cache hits when fetching instructions, code reads.",
686− "CounterHTOff": "0,1,2,3,4,5,6,7"
687− },
688− {
689− "EventCode": "0x24",
690− "Counter": "0,1,2,3",
691− "UMask": "0x24",
692− "EventName": "L2_RQSTS.CODE_RD_MISS",
693− "SampleAfterValue": "200003",
694− "BriefDescription": "L2 cache misses when fetching instructions.",
695− "CounterHTOff": "0,1,2,3,4,5,6,7"
696− },
697− {
698− "EventCode": "0x24",
699− "Counter": "0,1,2,3",
700− "UMask": "0x27",
701− "EventName": "L2_RQSTS.ALL_DEMAND_MISS",
702− "SampleAfterValue": "200003",
703− "BriefDescription": "Demand requests that miss L2 cache.",
704− "CounterHTOff": "0,1,2,3,4,5,6,7"
705− },
706− {
707− "EventCode": "0x24",
708− "Counter": "0,1,2,3",
709− "UMask": "0xe7",
710− "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES",
711− "SampleAfterValue": "200003",
712− "BriefDescription": "Demand requests to L2 cache.",
713− "CounterHTOff": "0,1,2,3,4,5,6,7"
714− },
715− {
716− "EventCode": "0x24",
717− "Counter": "0,1,2,3",
718− "UMask": "0x3f",
719− "EventName": "L2_RQSTS.MISS",
720− "SampleAfterValue": "200003",
721− "BriefDescription": "All requests that miss L2 cache.",
722− "CounterHTOff": "0,1,2,3,4,5,6,7"
723− },
724− {
725− "EventCode": "0x24",
726− "Counter": "0,1,2,3",
727− "UMask": "0xff",
728− "EventName": "L2_RQSTS.REFERENCES",
729− "SampleAfterValue": "200003",
730− "BriefDescription": "All L2 requests.",
731− "CounterHTOff": "0,1,2,3,4,5,6,7"
732− },
733− {
734− "EventCode": "0xB7, 0xBB",
735− "Counter": "0,1,2,3",
736− "UMask": "0x1",
737− "EventName": "OFFCORE_RESPONSE",
738− "SampleAfterValue": "100003",
739− "BriefDescription": "Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
740− "CounterHTOff": "0,1,2,3"
741− },
742− {
743− "EventCode": "0x60",
744− "Counter": "0,1,2,3",
745− "UMask": "0x1",
746− "Errata": "BDM76",
747− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_GE_6",
748− "SampleAfterValue": "2000003",
749− "BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
750− "CounterMask": "6",
751− "CounterHTOff": "0,1,2,3,4,5,6,7"
752− },
753− {
754− "EventCode": "0x48",
755− "Counter": "2",
756− "UMask": "0x1",
757− "AnyThread": "1",
758− "EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
759− "SampleAfterValue": "2000003",
760− "BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
761− "CounterMask": "1",
762− "CounterHTOff": "2"
763− },
764− {
765− "EventCode": "0x48",
766− "Counter": "0,1,2,3",
767− "UMask": "0x2",
768− "EventName": "L1D_PEND_MISS.FB_FULL",
769− "SampleAfterValue": "2000003",
770− "BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
771− "CounterMask": "1",
772− "CounterHTOff": "0,1,2,3,4,5,6,7"
773− },
774− {
774+ "PublicDescription": "Counts demand data reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
775775 "EventCode": "0xB7, 0xBB",
776776 "MSRValue": "0x0000010001 ",
777777 "Counter": "0,1,2,3",
@@ -784,6 +784,7 @@ · 6 unmodified lines
784784 "CounterHTOff": "0,1,2,3"
785785 },
786786 {
787+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
787788 "EventCode": "0xB7, 0xBB",
788789 "MSRValue": "0x0080020001 ",
789790 "Counter": "0,1,2,3",
@@ -796,6 +797,7 @@ · 6 unmodified lines
796797 "CounterHTOff": "0,1,2,3"
797798 },
798799 {
800+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
799801 "EventCode": "0xB7, 0xBB",
800802 "MSRValue": "0x0100020001 ",
801803 "Counter": "0,1,2,3",
@@ -808,6 +810,7 @@ · 6 unmodified lines
808810 "CounterHTOff": "0,1,2,3"
809811 },
810812 {
813+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
811814 "EventCode": "0xB7, 0xBB",
812815 "MSRValue": "0x0200020001 ",
813816 "Counter": "0,1,2,3",
@@ -820,6 +823,7 @@ · 6 unmodified lines
820823 "CounterHTOff": "0,1,2,3"
821824 },
822825 {
826+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
823827 "EventCode": "0xB7, 0xBB",
824828 "MSRValue": "0x0400020001 ",
825829 "Counter": "0,1,2,3",
@@ -832,6 +836,7 @@ · 6 unmodified lines
832836 "CounterHTOff": "0,1,2,3"
833837 },
834838 {
839+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
835840 "EventCode": "0xB7, 0xBB",
836841 "MSRValue": "0x1000020001 ",
837842 "Counter": "0,1,2,3",
@@ -844,6 +849,7 @@ · 6 unmodified lines
844849 "CounterHTOff": "0,1,2,3"
845850 },
846851 {
852+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
847853 "EventCode": "0xB7, 0xBB",
848854 "MSRValue": "0x3f80020001 ",
849855 "Counter": "0,1,2,3",
@@ -856,6 +862,7 @@ · 6 unmodified lines
856862 "CounterHTOff": "0,1,2,3"
857863 },
858864 {
865+ "PublicDescription": "Counts demand data reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
859866 "EventCode": "0xB7, 0xBB",
860867 "MSRValue": "0x00803c0001 ",
861868 "Counter": "0,1,2,3",
@@ -868,6 +875,7 @@ · 6 unmodified lines
868875 "CounterHTOff": "0,1,2,3"
869876 },
870877 {
878+ "PublicDescription": "Counts demand data reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
871879 "EventCode": "0xB7, 0xBB",
872880 "MSRValue": "0x01003c0001 ",
873881 "Counter": "0,1,2,3",
@@ -880,6 +888,7 @@ · 6 unmodified lines
880888 "CounterHTOff": "0,1,2,3"
881889 },
882890 {
891+ "PublicDescription": "Counts demand data reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
883892 "EventCode": "0xB7, 0xBB",
884893 "MSRValue": "0x02003c0001 ",
885894 "Counter": "0,1,2,3",
@@ -892,6 +901,7 @@ · 6 unmodified lines
892901 "CounterHTOff": "0,1,2,3"
893902 },
894903 {
904+ "PublicDescription": "Counts demand data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
895905 "EventCode": "0xB7, 0xBB",
896906 "MSRValue": "0x04003c0001 ",
897907 "Counter": "0,1,2,3",
@@ -904,6 +914,7 @@ · 6 unmodified lines
904914 "CounterHTOff": "0,1,2,3"
905915 },
906916 {
917+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
907918 "EventCode": "0xB7, 0xBB",
908919 "MSRValue": "0x10003c0001 ",
909920 "Counter": "0,1,2,3",
@@ -916,6 +927,7 @@ · 6 unmodified lines
916927 "CounterHTOff": "0,1,2,3"
917928 },
918929 {
930+ "PublicDescription": "Counts demand data reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
919931 "EventCode": "0xB7, 0xBB",
920932 "MSRValue": "0x3f803c0001 ",
921933 "Counter": "0,1,2,3",
@@ -928,6 +940,7 @@ · 6 unmodified lines
928940 "CounterHTOff": "0,1,2,3"
929941 },
930942 {
943+ "PublicDescription": "Counts all demand data writes (RFOs) that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
931944 "EventCode": "0xB7, 0xBB",
932945 "MSRValue": "0x0000010002 ",
933946 "Counter": "0,1,2,3",
@@ -940,6 +953,7 @@ · 6 unmodified lines
940953 "CounterHTOff": "0,1,2,3"
941954 },
942955 {
956+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
943957 "EventCode": "0xB7, 0xBB",
944958 "MSRValue": "0x00803c0002 ",
945959 "Counter": "0,1,2,3",
@@ -952,6 +966,7 @@ · 6 unmodified lines
952966 "CounterHTOff": "0,1,2,3"
953967 },
954968 {
969+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
955970 "EventCode": "0xB7, 0xBB",
956971 "MSRValue": "0x01003c0002 ",
957972 "Counter": "0,1,2,3",
@@ -964,6 +979,7 @@ · 6 unmodified lines
964979 "CounterHTOff": "0,1,2,3"
965980 },
966981 {
982+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
967983 "EventCode": "0xB7, 0xBB",
968984 "MSRValue": "0x02003c0002 ",
969985 "Counter": "0,1,2,3",
@@ -976,6 +992,7 @@ · 6 unmodified lines
976992 "CounterHTOff": "0,1,2,3"
977993 },
978994 {
995+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
979996 "EventCode": "0xB7, 0xBB",
980997 "MSRValue": "0x04003c0002 ",
981998 "Counter": "0,1,2,3",
@@ -988,6 +1005,7 @@ · 6 unmodified lines
9881005 "CounterHTOff": "0,1,2,3"
9891006 },
9901007 {
1008+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
9911009 "EventCode": "0xB7, 0xBB",
9921010 "MSRValue": "0x10003c0002 ",
9931011 "Counter": "0,1,2,3",
@@ -1000,6 +1018,7 @@ · 6 unmodified lines
10001018 "CounterHTOff": "0,1,2,3"
10011019 },
10021020 {
1021+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10031022 "EventCode": "0xB7, 0xBB",
10041023 "MSRValue": "0x3f803c0002 ",
10051024 "Counter": "0,1,2,3",
@@ -1012,6 +1031,7 @@ · 6 unmodified lines
10121031 "CounterHTOff": "0,1,2,3"
10131032 },
10141033 {
1034+ "PublicDescription": "Counts all demand code reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10151035 "EventCode": "0xB7, 0xBB",
10161036 "MSRValue": "0x0000010004 ",
10171037 "Counter": "0,1,2,3",
@@ -1024,6 +1044,7 @@ · 6 unmodified lines
10241044 "CounterHTOff": "0,1,2,3"
10251045 },
10261046 {
1047+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10271048 "EventCode": "0xB7, 0xBB",
10281049 "MSRValue": "0x0080020004 ",
10291050 "Counter": "0,1,2,3",
@@ -1036,6 +1057,7 @@ · 6 unmodified lines
10361057 "CounterHTOff": "0,1,2,3"
10371058 },
10381059 {
1060+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10391061 "EventCode": "0xB7, 0xBB",
10401062 "MSRValue": "0x0100020004 ",
10411063 "Counter": "0,1,2,3",
@@ -1048,6 +1070,7 @@ · 6 unmodified lines
10481070 "CounterHTOff": "0,1,2,3"
10491071 },
10501072 {
1073+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10511074 "EventCode": "0xB7, 0xBB",
10521075 "MSRValue": "0x0200020004 ",
10531076 "Counter": "0,1,2,3",
@@ -1060,6 +1083,7 @@ · 6 unmodified lines
10601083 "CounterHTOff": "0,1,2,3"
10611084 },
10621085 {
1086+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10631087 "EventCode": "0xB7, 0xBB",
10641088 "MSRValue": "0x0400020004 ",
10651089 "Counter": "0,1,2,3",
@@ -1072,6 +1096,7 @@ · 6 unmodified lines
10721096 "CounterHTOff": "0,1,2,3"
10731097 },
10741098 {
1099+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10751100 "EventCode": "0xB7, 0xBB",
10761101 "MSRValue": "0x1000020004 ",
10771102 "Counter": "0,1,2,3",
@@ -1084,6 +1109,7 @@ · 6 unmodified lines
10841109 "CounterHTOff": "0,1,2,3"
10851110 },
10861111 {
1112+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10871113 "EventCode": "0xB7, 0xBB",
10881114 "MSRValue": "0x3f80020004 ",
10891115 "Counter": "0,1,2,3",
@@ -1096,6 +1122,7 @@ · 6 unmodified lines
10961122 "CounterHTOff": "0,1,2,3"
10971123 },
10981124 {
1125+ "PublicDescription": "Counts all demand code reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10991126 "EventCode": "0xB7, 0xBB",
11001127 "MSRValue": "0x00803c0004 ",
11011128 "Counter": "0,1,2,3",
@@ -1108,6 +1135,7 @@ · 6 unmodified lines
11081135 "CounterHTOff": "0,1,2,3"
11091136 },
11101137 {
1138+ "PublicDescription": "Counts all demand code reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11111139 "EventCode": "0xB7, 0xBB",
11121140 "MSRValue": "0x01003c0004 ",
11131141 "Counter": "0,1,2,3",
@@ -1120,6 +1148,7 @@ · 6 unmodified lines
11201148 "CounterHTOff": "0,1,2,3"
11211149 },
11221150 {
1151+ "PublicDescription": "Counts all demand code reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11231152 "EventCode": "0xB7, 0xBB",
11241153 "MSRValue": "0x02003c0004 ",
11251154 "Counter": "0,1,2,3",
@@ -1132,6 +1161,7 @@ · 6 unmodified lines
11321161 "CounterHTOff": "0,1,2,3"
11331162 },
11341163 {
1164+ "PublicDescription": "Counts all demand code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11351165 "EventCode": "0xB7, 0xBB",
11361166 "MSRValue": "0x04003c0004 ",
11371167 "Counter": "0,1,2,3",
@@ -1144,6 +1174,7 @@ · 6 unmodified lines
11441174 "CounterHTOff": "0,1,2,3"
11451175 },
11461176 {
1177+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11471178 "EventCode": "0xB7, 0xBB",
11481179 "MSRValue": "0x10003c0004 ",
11491180 "Counter": "0,1,2,3",
@@ -1156,6 +1187,7 @@ · 6 unmodified lines
11561187 "CounterHTOff": "0,1,2,3"
11571188 },
11581189 {
1190+ "PublicDescription": "Counts all demand code reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11591191 "EventCode": "0xB7, 0xBB",
11601192 "MSRValue": "0x3f803c0004 ",
11611193 "Counter": "0,1,2,3",
@@ -1168,6 +1200,7 @@ · 6 unmodified lines
11681200 "CounterHTOff": "0,1,2,3"
11691201 },
11701202 {
1203+ "PublicDescription": "Counts writebacks (modified to exclusive) that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11711204 "EventCode": "0xB7, 0xBB",
11721205 "MSRValue": "0x0000010008 ",
11731206 "Counter": "0,1,2,3",
@@ -1180,6 +1213,7 @@ · 6 unmodified lines
11801213 "CounterHTOff": "0,1,2,3"
11811214 },
11821215 {
1216+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11831217 "EventCode": "0xB7, 0xBB",
11841218 "MSRValue": "0x0080020008 ",
11851219 "Counter": "0,1,2,3",
@@ -1192,6 +1226,7 @@ · 6 unmodified lines
11921226 "CounterHTOff": "0,1,2,3"
11931227 },
11941228 {
1229+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11951230 "EventCode": "0xB7, 0xBB",
11961231 "MSRValue": "0x0100020008 ",
11971232 "Counter": "0,1,2,3",
@@ -1204,6 +1239,7 @@ · 6 unmodified lines
12041239 "CounterHTOff": "0,1,2,3"
12051240 },
12061241 {
1242+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12071243 "EventCode": "0xB7, 0xBB",
12081244 "MSRValue": "0x0200020008 ",
12091245 "Counter": "0,1,2,3",
@@ -1216,6 +1252,7 @@ · 6 unmodified lines
12161252 "CounterHTOff": "0,1,2,3"
12171253 },
12181254 {
1255+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12191256 "EventCode": "0xB7, 0xBB",
12201257 "MSRValue": "0x0400020008 ",
12211258 "Counter": "0,1,2,3",
@@ -1228,6 +1265,7 @@ · 6 unmodified lines
12281265 "CounterHTOff": "0,1,2,3"
12291266 },
12301267 {
1268+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12311269 "EventCode": "0xB7, 0xBB",
12321270 "MSRValue": "0x1000020008 ",
12331271 "Counter": "0,1,2,3",
@@ -1240,6 +1278,7 @@ · 6 unmodified lines
12401278 "CounterHTOff": "0,1,2,3"
12411279 },
12421280 {
1281+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12431282 "EventCode": "0xB7, 0xBB",
12441283 "MSRValue": "0x3f80020008 ",
12451284 "Counter": "0,1,2,3",
@@ -1252,6 +1291,7 @@ · 6 unmodified lines
12521291 "CounterHTOff": "0,1,2,3"
12531292 },
12541293 {
1294+ "PublicDescription": "Counts writebacks (modified to exclusive) that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12551295 "EventCode": "0xB7, 0xBB",
12561296 "MSRValue": "0x00803c0008 ",
12571297 "Counter": "0,1,2,3",
@@ -1264,6 +1304,7 @@ · 6 unmodified lines
12641304 "CounterHTOff": "0,1,2,3"
12651305 },
12661306 {
1307+ "PublicDescription": "Counts writebacks (modified to exclusive) that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12671308 "EventCode": "0xB7, 0xBB",
12681309 "MSRValue": "0x01003c0008 ",
12691310 "Counter": "0,1,2,3",
@@ -1276,6 +1317,7 @@ · 6 unmodified lines
12761317 "CounterHTOff": "0,1,2,3"
12771318 },
12781319 {
1320+ "PublicDescription": "Counts writebacks (modified to exclusive) that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12791321 "EventCode": "0xB7, 0xBB",
12801322 "MSRValue": "0x02003c0008 ",
12811323 "Counter": "0,1,2,3",
@@ -1288,6 +1330,7 @@ · 6 unmodified lines
12881330 "CounterHTOff": "0,1,2,3"
12891331 },
12901332 {
1333+ "PublicDescription": "Counts writebacks (modified to exclusive) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12911334 "EventCode": "0xB7, 0xBB",
12921335 "MSRValue": "0x04003c0008 ",
12931336 "Counter": "0,1,2,3",
@@ -1300,6 +1343,7 @@ · 6 unmodified lines
13001343 "CounterHTOff": "0,1,2,3"
13011344 },
13021345 {
1346+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13031347 "EventCode": "0xB7, 0xBB",
13041348 "MSRValue": "0x10003c0008 ",
13051349 "Counter": "0,1,2,3",
@@ -1312,6 +1356,7 @@ · 6 unmodified lines
13121356 "CounterHTOff": "0,1,2,3"
13131357 },
13141358 {
1359+ "PublicDescription": "Counts writebacks (modified to exclusive) that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13151360 "EventCode": "0xB7, 0xBB",
13161361 "MSRValue": "0x3f803c0008 ",
13171362 "Counter": "0,1,2,3",
@@ -1324,6 +1369,7 @@ · 6 unmodified lines
13241369 "CounterHTOff": "0,1,2,3"
13251370 },
13261371 {
1372+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13271373 "EventCode": "0xB7, 0xBB",
13281374 "MSRValue": "0x0000010010 ",
13291375 "Counter": "0,1,2,3",
@@ -1336,6 +1382,7 @@ · 6 unmodified lines
13361382 "CounterHTOff": "0,1,2,3"
13371383 },
13381384 {
1385+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13391386 "EventCode": "0xB7, 0xBB",
13401387 "MSRValue": "0x0080020010 ",
13411388 "Counter": "0,1,2,3",
@@ -1348,6 +1395,7 @@ · 6 unmodified lines
13481395 "CounterHTOff": "0,1,2,3"
13491396 },
13501397 {
1398+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13511399 "EventCode": "0xB7, 0xBB",
13521400 "MSRValue": "0x0100020010 ",
13531401 "Counter": "0,1,2,3",
@@ -1360,6 +1408,7 @@ · 6 unmodified lines
13601408 "CounterHTOff": "0,1,2,3"
13611409 },
13621410 {
1411+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13631412 "EventCode": "0xB7, 0xBB",
13641413 "MSRValue": "0x0200020010 ",
13651414 "Counter": "0,1,2,3",
@@ -1372,6 +1421,7 @@ · 6 unmodified lines
13721421 "CounterHTOff": "0,1,2,3"
13731422 },
13741423 {
1424+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13751425 "EventCode": "0xB7, 0xBB",
13761426 "MSRValue": "0x0400020010 ",
13771427 "Counter": "0,1,2,3",
@@ -1384,6 +1434,7 @@ · 6 unmodified lines
13841434 "CounterHTOff": "0,1,2,3"
13851435 },
13861436 {
1437+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13871438 "EventCode": "0xB7, 0xBB",
13881439 "MSRValue": "0x1000020010 ",
13891440 "Counter": "0,1,2,3",
@@ -1396,6 +1447,7 @@ · 6 unmodified lines
13961447 "CounterHTOff": "0,1,2,3"
13971448 },
13981449 {
1450+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13991451 "EventCode": "0xB7, 0xBB",
14001452 "MSRValue": "0x3f80020010 ",
14011453 "Counter": "0,1,2,3",
@@ -1408,6 +1460,7 @@ · 6 unmodified lines
14081460 "CounterHTOff": "0,1,2,3"
14091461 },
14101462 {
1463+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14111464 "EventCode": "0xB7, 0xBB",
14121465 "MSRValue": "0x00803c0010 ",
14131466 "Counter": "0,1,2,3",
@@ -1420,6 +1473,7 @@ · 6 unmodified lines
14201473 "CounterHTOff": "0,1,2,3"
14211474 },
14221475 {
1476+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14231477 "EventCode": "0xB7, 0xBB",
14241478 "MSRValue": "0x01003c0010 ",
14251479 "Counter": "0,1,2,3",
@@ -1432,6 +1486,7 @@ · 6 unmodified lines
14321486 "CounterHTOff": "0,1,2,3"
14331487 },
14341488 {
1489+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14351490 "EventCode": "0xB7, 0xBB",
14361491 "MSRValue": "0x02003c0010 ",
14371492 "Counter": "0,1,2,3",
@@ -1444,6 +1499,7 @@ · 6 unmodified lines
14441499 "CounterHTOff": "0,1,2,3"
14451500 },
14461501 {
1502+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14471503 "EventCode": "0xB7, 0xBB",
14481504 "MSRValue": "0x04003c0010 ",
14491505 "Counter": "0,1,2,3",
@@ -1456,6 +1512,7 @@ · 6 unmodified lines
14561512 "CounterHTOff": "0,1,2,3"
14571513 },
14581514 {
1515+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14591516 "EventCode": "0xB7, 0xBB",
14601517 "MSRValue": "0x10003c0010 ",
14611518 "Counter": "0,1,2,3",
@@ -1468,6 +1525,7 @@ · 6 unmodified lines
14681525 "CounterHTOff": "0,1,2,3"
14691526 },
14701527 {
1528+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14711529 "EventCode": "0xB7, 0xBB",
14721530 "MSRValue": "0x3f803c0010 ",
14731531 "Counter": "0,1,2,3",
@@ -1480,6 +1538,7 @@ · 6 unmodified lines
14801538 "CounterHTOff": "0,1,2,3"
14811539 },
14821540 {
1541+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14831542 "EventCode": "0xB7, 0xBB",
14841543 "MSRValue": "0x0000010020 ",
14851544 "Counter": "0,1,2,3",
@@ -1492,6 +1551,7 @@ · 6 unmodified lines
14921551 "CounterHTOff": "0,1,2,3"
14931552 },
14941553 {
1554+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14951555 "EventCode": "0xB7, 0xBB",
14961556 "MSRValue": "0x0080020020 ",
14971557 "Counter": "0,1,2,3",
@@ -1504,6 +1564,7 @@ · 6 unmodified lines
15041564 "CounterHTOff": "0,1,2,3"
15051565 },
15061566 {
1567+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15071568 "EventCode": "0xB7, 0xBB",
15081569 "MSRValue": "0x0100020020 ",
15091570 "Counter": "0,1,2,3",
@@ -1516,6 +1577,7 @@ · 6 unmodified lines
15161577 "CounterHTOff": "0,1,2,3"
15171578 },
15181579 {
1580+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15191581 "EventCode": "0xB7, 0xBB",
15201582 "MSRValue": "0x0200020020 ",
15211583 "Counter": "0,1,2,3",
@@ -1528,6 +1590,7 @@ · 6 unmodified lines
15281590 "CounterHTOff": "0,1,2,3"
15291591 },
15301592 {
1593+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15311594 "EventCode": "0xB7, 0xBB",
15321595 "MSRValue": "0x0400020020 ",
15331596 "Counter": "0,1,2,3",
@@ -1540,6 +1603,7 @@ · 6 unmodified lines
15401603 "CounterHTOff": "0,1,2,3"
15411604 },
15421605 {
1606+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15431607 "EventCode": "0xB7, 0xBB",
15441608 "MSRValue": "0x1000020020 ",
15451609 "Counter": "0,1,2,3",
@@ -1552,6 +1616,7 @@ · 6 unmodified lines
15521616 "CounterHTOff": "0,1,2,3"
15531617 },
15541618 {
1619+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15551620 "EventCode": "0xB7, 0xBB",
15561621 "MSRValue": "0x3f80020020 ",
15571622 "Counter": "0,1,2,3",
@@ -1564,6 +1629,7 @@ · 6 unmodified lines
15641629 "CounterHTOff": "0,1,2,3"
15651630 },
15661631 {
1632+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15671633 "EventCode": "0xB7, 0xBB",
15681634 "MSRValue": "0x00803c0020 ",
15691635 "Counter": "0,1,2,3",
@@ -1576,6 +1642,7 @@ · 6 unmodified lines
15761642 "CounterHTOff": "0,1,2,3"
15771643 },
15781644 {
1645+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15791646 "EventCode": "0xB7, 0xBB",
15801647 "MSRValue": "0x01003c0020 ",
15811648 "Counter": "0,1,2,3",
@@ -1588,6 +1655,7 @@ · 6 unmodified lines
15881655 "CounterHTOff": "0,1,2,3"
15891656 },
15901657 {
1658+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15911659 "EventCode": "0xB7, 0xBB",
15921660 "MSRValue": "0x02003c0020 ",
15931661 "Counter": "0,1,2,3",
@@ -1600,6 +1668,7 @@ · 6 unmodified lines
16001668 "CounterHTOff": "0,1,2,3"
16011669 },
16021670 {
1671+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16031672 "EventCode": "0xB7, 0xBB",
16041673 "MSRValue": "0x04003c0020 ",
16051674 "Counter": "0,1,2,3",
@@ -1612,6 +1681,7 @@ · 6 unmodified lines
16121681 "CounterHTOff": "0,1,2,3"
16131682 },
16141683 {
1684+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16151685 "EventCode": "0xB7, 0xBB",
16161686 "MSRValue": "0x10003c0020 ",
16171687 "Counter": "0,1,2,3",
@@ -1624,6 +1694,7 @@ · 6 unmodified lines
16241694 "CounterHTOff": "0,1,2,3"
16251695 },
16261696 {
1697+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16271698 "EventCode": "0xB7, 0xBB",
16281699 "MSRValue": "0x3f803c0020 ",
16291700 "Counter": "0,1,2,3",
@@ -1636,6 +1707,7 @@ · 6 unmodified lines
16361707 "CounterHTOff": "0,1,2,3"
16371708 },
16381709 {
1710+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16391711 "EventCode": "0xB7, 0xBB",
16401712 "MSRValue": "0x0000010040 ",
16411713 "Counter": "0,1,2,3",
@@ -1648,6 +1720,7 @@ · 6 unmodified lines
16481720 "CounterHTOff": "0,1,2,3"
16491721 },
16501722 {
1723+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16511724 "EventCode": "0xB7, 0xBB",
16521725 "MSRValue": "0x0080020040 ",
16531726 "Counter": "0,1,2,3",
@@ -1660,6 +1733,7 @@ · 6 unmodified lines
16601733 "CounterHTOff": "0,1,2,3"
16611734 },
16621735 {
1736+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16631737 "EventCode": "0xB7, 0xBB",
16641738 "MSRValue": "0x0100020040 ",
16651739 "Counter": "0,1,2,3",
@@ -1672,6 +1746,7 @@ · 6 unmodified lines
16721746 "CounterHTOff": "0,1,2,3"
16731747 },
16741748 {
1749+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16751750 "EventCode": "0xB7, 0xBB",
16761751 "MSRValue": "0x0200020040 ",
16771752 "Counter": "0,1,2,3",
@@ -1684,6 +1759,7 @@ · 6 unmodified lines
16841759 "CounterHTOff": "0,1,2,3"
16851760 },
16861761 {
1762+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16871763 "EventCode": "0xB7, 0xBB",
16881764 "MSRValue": "0x0400020040 ",
16891765 "Counter": "0,1,2,3",
@@ -1696,6 +1772,7 @@ · 6 unmodified lines
16961772 "CounterHTOff": "0,1,2,3"
16971773 },
16981774 {
1775+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16991776 "EventCode": "0xB7, 0xBB",
17001777 "MSRValue": "0x1000020040 ",
17011778 "Counter": "0,1,2,3",
@@ -1708,6 +1785,7 @@ · 6 unmodified lines
17081785 "CounterHTOff": "0,1,2,3"
17091786 },
17101787 {
1788+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17111789 "EventCode": "0xB7, 0xBB",
17121790 "MSRValue": "0x3f80020040 ",
17131791 "Counter": "0,1,2,3",
@@ -1720,6 +1798,7 @@ · 6 unmodified lines
17201798 "CounterHTOff": "0,1,2,3"
17211799 },
17221800 {
1801+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17231802 "EventCode": "0xB7, 0xBB",
17241803 "MSRValue": "0x00803c0040 ",
17251804 "Counter": "0,1,2,3",
@@ -1732,6 +1811,7 @@ · 6 unmodified lines
17321811 "CounterHTOff": "0,1,2,3"
17331812 },
17341813 {
1814+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17351815 "EventCode": "0xB7, 0xBB",
17361816 "MSRValue": "0x01003c0040 ",
17371817 "Counter": "0,1,2,3",
@@ -1744,6 +1824,7 @@ · 6 unmodified lines
17441824 "CounterHTOff": "0,1,2,3"
17451825 },
17461826 {
1827+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17471828 "EventCode": "0xB7, 0xBB",
17481829 "MSRValue": "0x02003c0040 ",
17491830 "Counter": "0,1,2,3",
@@ -1756,6 +1837,7 @@ · 6 unmodified lines
17561837 "CounterHTOff": "0,1,2,3"
17571838 },
17581839 {
1840+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17591841 "EventCode": "0xB7, 0xBB",
17601842 "MSRValue": "0x04003c0040 ",
17611843 "Counter": "0,1,2,3",
@@ -1768,6 +1850,7 @@ · 6 unmodified lines
17681850 "CounterHTOff": "0,1,2,3"
17691851 },
17701852 {
1853+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17711854 "EventCode": "0xB7, 0xBB",
17721855 "MSRValue": "0x10003c0040 ",
17731856 "Counter": "0,1,2,3",
@@ -1780,6 +1863,7 @@ · 6 unmodified lines
17801863 "CounterHTOff": "0,1,2,3"
17811864 },
17821865 {
1866+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17831867 "EventCode": "0xB7, 0xBB",
17841868 "MSRValue": "0x3f803c0040 ",
17851869 "Counter": "0,1,2,3",
@@ -1792,6 +1876,7 @@ · 6 unmodified lines
17921876 "CounterHTOff": "0,1,2,3"
17931877 },
17941878 {
1879+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17951880 "EventCode": "0xB7, 0xBB",
17961881 "MSRValue": "0x0000010080 ",
17971882 "Counter": "0,1,2,3",
@@ -1804,6 +1889,7 @@ · 6 unmodified lines
18041889 "CounterHTOff": "0,1,2,3"
18051890 },
18061891 {
1892+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18071893 "EventCode": "0xB7, 0xBB",
18081894 "MSRValue": "0x0080020080 ",
18091895 "Counter": "0,1,2,3",
@@ -1816,6 +1902,7 @@ · 6 unmodified lines
18161902 "CounterHTOff": "0,1,2,3"
18171903 },
18181904 {
1905+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18191906 "EventCode": "0xB7, 0xBB",
18201907 "MSRValue": "0x0100020080 ",
18211908 "Counter": "0,1,2,3",
@@ -1828,6 +1915,7 @@ · 6 unmodified lines
18281915 "CounterHTOff": "0,1,2,3"
18291916 },
18301917 {
1918+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18311919 "EventCode": "0xB7, 0xBB",
18321920 "MSRValue": "0x0200020080 ",
18331921 "Counter": "0,1,2,3",
@@ -1840,6 +1928,7 @@ · 6 unmodified lines
18401928 "CounterHTOff": "0,1,2,3"
18411929 },
18421930 {
1931+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18431932 "EventCode": "0xB7, 0xBB",
18441933 "MSRValue": "0x0400020080 ",
18451934 "Counter": "0,1,2,3",
@@ -1852,6 +1941,7 @@ · 6 unmodified lines
18521941 "CounterHTOff": "0,1,2,3"
18531942 },
18541943 {
1944+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18551945 "EventCode": "0xB7, 0xBB",
18561946 "MSRValue": "0x1000020080 ",
18571947 "Counter": "0,1,2,3",
@@ -1864,6 +1954,7 @@ · 6 unmodified lines
18641954 "CounterHTOff": "0,1,2,3"
18651955 },
18661956 {
1957+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18671958 "EventCode": "0xB7, 0xBB",
18681959 "MSRValue": "0x3f80020080 ",
18691960 "Counter": "0,1,2,3",
@@ -1876,6 +1967,7 @@ · 6 unmodified lines
18761967 "CounterHTOff": "0,1,2,3"
18771968 },
18781969 {
1970+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18791971 "EventCode": "0xB7, 0xBB",
18801972 "MSRValue": "0x00803c0080 ",
18811973 "Counter": "0,1,2,3",
@@ -1888,6 +1980,7 @@ · 6 unmodified lines
18881980 "CounterHTOff": "0,1,2,3"
18891981 },
18901982 {
1983+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18911984 "EventCode": "0xB7, 0xBB",
18921985 "MSRValue": "0x01003c0080 ",
18931986 "Counter": "0,1,2,3",
@@ -1900,6 +1993,7 @@ · 6 unmodified lines
19001993 "CounterHTOff": "0,1,2,3"
19011994 },
19021995 {
1996+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19031997 "EventCode": "0xB7, 0xBB",
19041998 "MSRValue": "0x02003c0080 ",
19051999 "Counter": "0,1,2,3",
@@ -1912,6 +2006,7 @@ · 6 unmodified lines
19122006 "CounterHTOff": "0,1,2,3"
19132007 },
19142008 {
2009+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19152010 "EventCode": "0xB7, 0xBB",
19162011 "MSRValue": "0x04003c0080 ",
19172012 "Counter": "0,1,2,3",
@@ -1924,6 +2019,7 @@ · 6 unmodified lines
19242019 "CounterHTOff": "0,1,2,3"
19252020 },
19262021 {
2022+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19272023 "EventCode": "0xB7, 0xBB",
19282024 "MSRValue": "0x10003c0080 ",
19292025 "Counter": "0,1,2,3",
@@ -1936,6 +2032,7 @@ · 6 unmodified lines
19362032 "CounterHTOff": "0,1,2,3"
19372033 },
19382034 {
2035+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19392036 "EventCode": "0xB7, 0xBB",
19402037 "MSRValue": "0x3f803c0080 ",
19412038 "Counter": "0,1,2,3",
@@ -1948,6 +2045,7 @@ · 6 unmodified lines
19482045 "CounterHTOff": "0,1,2,3"
19492046 },
19502047 {
2048+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19512049 "EventCode": "0xB7, 0xBB",
19522050 "MSRValue": "0x0000010100 ",
19532051 "Counter": "0,1,2,3",
@@ -1960,6 +2058,7 @@ · 6 unmodified lines
19602058 "CounterHTOff": "0,1,2,3"
19612059 },
19622060 {
2061+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19632062 "EventCode": "0xB7, 0xBB",
19642063 "MSRValue": "0x0080020100 ",
19652064 "Counter": "0,1,2,3",
@@ -1972,6 +2071,7 @@ · 6 unmodified lines
19722071 "CounterHTOff": "0,1,2,3"
19732072 },
19742073 {
2074+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19752075 "EventCode": "0xB7, 0xBB",
19762076 "MSRValue": "0x0100020100 ",
19772077 "Counter": "0,1,2,3",
@@ -1984,6 +2084,7 @@ · 6 unmodified lines
19842084 "CounterHTOff": "0,1,2,3"
19852085 },
19862086 {
2087+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19872088 "EventCode": "0xB7, 0xBB",
19882089 "MSRValue": "0x0200020100 ",
19892090 "Counter": "0,1,2,3",
@@ -1996,6 +2097,7 @@ · 6 unmodified lines
19962097 "CounterHTOff": "0,1,2,3"
19972098 },
19982099 {
2100+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19992101 "EventCode": "0xB7, 0xBB",
20002102 "MSRValue": "0x0400020100 ",
20012103 "Counter": "0,1,2,3",
@@ -2008,6 +2110,7 @@ · 6 unmodified lines
20082110 "CounterHTOff": "0,1,2,3"
20092111 },
20102112 {
2113+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20112114 "EventCode": "0xB7, 0xBB",
20122115 "MSRValue": "0x1000020100 ",
20132116 "Counter": "0,1,2,3",
@@ -2020,6 +2123,7 @@ · 6 unmodified lines
20202123 "CounterHTOff": "0,1,2,3"
20212124 },
20222125 {
2126+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20232127 "EventCode": "0xB7, 0xBB",
20242128 "MSRValue": "0x3f80020100 ",
20252129 "Counter": "0,1,2,3",
@@ -2032,6 +2136,7 @@ · 6 unmodified lines
20322136 "CounterHTOff": "0,1,2,3"
20332137 },
20342138 {
2139+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20352140 "EventCode": "0xB7, 0xBB",
20362141 "MSRValue": "0x00803c0100 ",
20372142 "Counter": "0,1,2,3",
@@ -2044,6 +2149,7 @@ · 6 unmodified lines
20442149 "CounterHTOff": "0,1,2,3"
20452150 },
20462151 {
2152+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20472153 "EventCode": "0xB7, 0xBB",
20482154 "MSRValue": "0x01003c0100 ",
20492155 "Counter": "0,1,2,3",
@@ -2056,6 +2162,7 @@ · 6 unmodified lines
20562162 "CounterHTOff": "0,1,2,3"
20572163 },
20582164 {
2165+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20592166 "EventCode": "0xB7, 0xBB",
20602167 "MSRValue": "0x02003c0100 ",
20612168 "Counter": "0,1,2,3",
@@ -2068,6 +2175,7 @@ · 6 unmodified lines
20682175 "CounterHTOff": "0,1,2,3"
20692176 },
20702177 {
2178+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20712179 "EventCode": "0xB7, 0xBB",
20722180 "MSRValue": "0x04003c0100 ",
20732181 "Counter": "0,1,2,3",
@@ -2080,6 +2188,7 @@ · 6 unmodified lines
20802188 "CounterHTOff": "0,1,2,3"
20812189 },
20822190 {
2191+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20832192 "EventCode": "0xB7, 0xBB",
20842193 "MSRValue": "0x10003c0100 ",
20852194 "Counter": "0,1,2,3",
@@ -2092,6 +2201,7 @@ · 6 unmodified lines
20922201 "CounterHTOff": "0,1,2,3"
20932202 },
20942203 {
2204+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20952205 "EventCode": "0xB7, 0xBB",
20962206 "MSRValue": "0x3f803c0100 ",
20972207 "Counter": "0,1,2,3",
@@ -2104,6 +2214,7 @@ · 6 unmodified lines
21042214 "CounterHTOff": "0,1,2,3"
21052215 },
21062216 {
2217+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21072218 "EventCode": "0xB7, 0xBB",
21082219 "MSRValue": "0x0000010200 ",
21092220 "Counter": "0,1,2,3",
@@ -2116,6 +2227,7 @@ · 6 unmodified lines
21162227 "CounterHTOff": "0,1,2,3"
21172228 },
21182229 {
2230+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21192231 "EventCode": "0xB7, 0xBB",
21202232 "MSRValue": "0x0080020200 ",
21212233 "Counter": "0,1,2,3",
@@ -2128,6 +2240,7 @@ · 6 unmodified lines
21282240 "CounterHTOff": "0,1,2,3"
21292241 },
21302242 {
2243+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21312244 "EventCode": "0xB7, 0xBB",
21322245 "MSRValue": "0x0100020200 ",
21332246 "Counter": "0,1,2,3",
@@ -2140,6 +2253,7 @@ · 6 unmodified lines
21402253 "CounterHTOff": "0,1,2,3"
21412254 },
21422255 {
2256+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21432257 "EventCode": "0xB7, 0xBB",
21442258 "MSRValue": "0x0200020200 ",
21452259 "Counter": "0,1,2,3",
@@ -2152,6 +2266,7 @@ · 6 unmodified lines
21522266 "CounterHTOff": "0,1,2,3"
21532267 },
21542268 {
2269+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21552270 "EventCode": "0xB7, 0xBB",
21562271 "MSRValue": "0x0400020200 ",
21572272 "Counter": "0,1,2,3",
@@ -2164,6 +2279,7 @@ · 6 unmodified lines
21642279 "CounterHTOff": "0,1,2,3"
21652280 },
21662281 {
2282+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21672283 "EventCode": "0xB7, 0xBB",
21682284 "MSRValue": "0x1000020200 ",
21692285 "Counter": "0,1,2,3",
@@ -2176,6 +2292,7 @@ · 6 unmodified lines
21762292 "CounterHTOff": "0,1,2,3"
21772293 },
21782294 {
2295+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21792296 "EventCode": "0xB7, 0xBB",
21802297 "MSRValue": "0x3f80020200 ",
21812298 "Counter": "0,1,2,3",
@@ -2188,6 +2305,7 @@ · 6 unmodified lines
21882305 "CounterHTOff": "0,1,2,3"
21892306 },
21902307 {
2308+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21912309 "EventCode": "0xB7, 0xBB",
21922310 "MSRValue": "0x00803c0200 ",
21932311 "Counter": "0,1,2,3",
@@ -2200,6 +2318,7 @@ · 6 unmodified lines
22002318 "CounterHTOff": "0,1,2,3"
22012319 },
22022320 {
2321+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22032322 "EventCode": "0xB7, 0xBB",
22042323 "MSRValue": "0x01003c0200 ",
22052324 "Counter": "0,1,2,3",
@@ -2212,6 +2331,7 @@ · 6 unmodified lines
22122331 "CounterHTOff": "0,1,2,3"
22132332 },
22142333 {
2334+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22152335 "EventCode": "0xB7, 0xBB",
22162336 "MSRValue": "0x02003c0200 ",
22172337 "Counter": "0,1,2,3",
@@ -2224,6 +2344,7 @@ · 6 unmodified lines
22242344 "CounterHTOff": "0,1,2,3"
22252345 },
22262346 {
2347+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22272348 "EventCode": "0xB7, 0xBB",
22282349 "MSRValue": "0x04003c0200 ",
22292350 "Counter": "0,1,2,3",
@@ -2236,6 +2357,7 @@ · 6 unmodified lines
22362357 "CounterHTOff": "0,1,2,3"
22372358 },
22382359 {
2360+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22392361 "EventCode": "0xB7, 0xBB",
22402362 "MSRValue": "0x10003c0200 ",
22412363 "Counter": "0,1,2,3",
@@ -2248,6 +2370,7 @@ · 6 unmodified lines
22482370 "CounterHTOff": "0,1,2,3"
22492371 },
22502372 {
2373+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22512374 "EventCode": "0xB7, 0xBB",
22522375 "MSRValue": "0x3f803c0200 ",
22532376 "Counter": "0,1,2,3",
@@ -2260,6 +2383,7 @@ · 6 unmodified lines
22602383 "CounterHTOff": "0,1,2,3"
22612384 },
22622385 {
2386+ "PublicDescription": "Counts any other requests that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22632387 "EventCode": "0xB7, 0xBB",
22642388 "MSRValue": "0x0000018000 ",
22652389 "Counter": "0,1,2,3",
@@ -2272,6 +2396,7 @@ · 6 unmodified lines
22722396 "CounterHTOff": "0,1,2,3"
22732397 },
22742398 {
2399+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22752400 "EventCode": "0xB7, 0xBB",
22762401 "MSRValue": "0x0080028000 ",
22772402 "Counter": "0,1,2,3",
@@ -2284,6 +2409,7 @@ · 6 unmodified lines
22842409 "CounterHTOff": "0,1,2,3"
22852410 },
22862411 {
2412+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22872413 "EventCode": "0xB7, 0xBB",
22882414 "MSRValue": "0x0100028000 ",
22892415 "Counter": "0,1,2,3",
@@ -2296,6 +2422,7 @@ · 6 unmodified lines
22962422 "CounterHTOff": "0,1,2,3"
22972423 },
22982424 {
2425+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22992426 "EventCode": "0xB7, 0xBB",
23002427 "MSRValue": "0x0200028000 ",
23012428 "Counter": "0,1,2,3",
@@ -2308,6 +2435,7 @@ · 6 unmodified lines
23082435 "CounterHTOff": "0,1,2,3"
23092436 },
23102437 {
2438+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23112439 "EventCode": "0xB7, 0xBB",
23122440 "MSRValue": "0x0400028000 ",
23132441 "Counter": "0,1,2,3",
@@ -2320,6 +2448,7 @@ · 6 unmodified lines
23202448 "CounterHTOff": "0,1,2,3"
23212449 },
23222450 {
2451+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23232452 "EventCode": "0xB7, 0xBB",
23242453 "MSRValue": "0x1000028000 ",
23252454 "Counter": "0,1,2,3",
@@ -2332,6 +2461,7 @@ · 6 unmodified lines
23322461 "CounterHTOff": "0,1,2,3"
23332462 },
23342463 {
2464+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23352465 "EventCode": "0xB7, 0xBB",
23362466 "MSRValue": "0x3f80028000 ",
23372467 "Counter": "0,1,2,3",
@@ -2344,6 +2474,7 @@ · 6 unmodified lines
23442474 "CounterHTOff": "0,1,2,3"
23452475 },
23462476 {
2477+ "PublicDescription": "Counts any other requests that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23472478 "EventCode": "0xB7, 0xBB",
23482479 "MSRValue": "0x00803c8000 ",
23492480 "Counter": "0,1,2,3",
@@ -2356,6 +2487,7 @@ · 6 unmodified lines
23562487 "CounterHTOff": "0,1,2,3"
23572488 },
23582489 {
2490+ "PublicDescription": "Counts any other requests that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23592491 "EventCode": "0xB7, 0xBB",
23602492 "MSRValue": "0x01003c8000 ",
23612493 "Counter": "0,1,2,3",
@@ -2368,6 +2500,7 @@ · 6 unmodified lines
23682500 "CounterHTOff": "0,1,2,3"
23692501 },
23702502 {
2503+ "PublicDescription": "Counts any other requests that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23712504 "EventCode": "0xB7, 0xBB",
23722505 "MSRValue": "0x02003c8000 ",
23732506 "Counter": "0,1,2,3",
@@ -2380,6 +2513,7 @@ · 6 unmodified lines
23802513 "CounterHTOff": "0,1,2,3"
23812514 },
23822515 {
2516+ "PublicDescription": "Counts any other requests that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23832517 "EventCode": "0xB7, 0xBB",
23842518 "MSRValue": "0x04003c8000 ",
23852519 "Counter": "0,1,2,3",
@@ -2392,6 +2526,7 @@ · 6 unmodified lines
23922526 "CounterHTOff": "0,1,2,3"
23932527 },
23942528 {
2529+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23952530 "EventCode": "0xB7, 0xBB",
23962531 "MSRValue": "0x10003c8000 ",
23972532 "Counter": "0,1,2,3",
@@ -2404,6 +2539,7 @@ · 6 unmodified lines
24042539 "CounterHTOff": "0,1,2,3"
24052540 },
24062541 {
2542+ "PublicDescription": "Counts any other requests that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24072543 "EventCode": "0xB7, 0xBB",
24082544 "MSRValue": "0x3f803c8000 ",
24092545 "Counter": "0,1,2,3",
@@ -2416,6 +2552,7 @@ · 6 unmodified lines
24162552 "CounterHTOff": "0,1,2,3"
24172553 },
24182554 {
2555+ "PublicDescription": "Counts all prefetch data reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24192556 "EventCode": "0xB7, 0xBB",
24202557 "MSRValue": "0x0000010090 ",
24212558 "Counter": "0,1,2,3",
@@ -2428,6 +2565,7 @@ · 6 unmodified lines
24282565 "CounterHTOff": "0,1,2,3"
24292566 },
24302567 {
2568+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24312569 "EventCode": "0xB7, 0xBB",
24322570 "MSRValue": "0x0080020090 ",
24332571 "Counter": "0,1,2,3",
@@ -2440,6 +2578,7 @@ · 6 unmodified lines
24402578 "CounterHTOff": "0,1,2,3"
24412579 },
24422580 {
2581+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24432582 "EventCode": "0xB7, 0xBB",
24442583 "MSRValue": "0x0100020090 ",
24452584 "Counter": "0,1,2,3",
@@ -2452,6 +2591,7 @@ · 6 unmodified lines
24522591 "CounterHTOff": "0,1,2,3"
24532592 },
24542593 {
2594+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24552595 "EventCode": "0xB7, 0xBB",
24562596 "MSRValue": "0x0200020090 ",
24572597 "Counter": "0,1,2,3",
@@ -2464,6 +2604,7 @@ · 6 unmodified lines
24642604 "CounterHTOff": "0,1,2,3"
24652605 },
24662606 {
2607+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24672608 "EventCode": "0xB7, 0xBB",
24682609 "MSRValue": "0x0400020090 ",
24692610 "Counter": "0,1,2,3",
@@ -2476,6 +2617,7 @@ · 6 unmodified lines
24762617 "CounterHTOff": "0,1,2,3"
24772618 },
24782619 {
2620+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24792621 "EventCode": "0xB7, 0xBB",
24802622 "MSRValue": "0x1000020090 ",
24812623 "Counter": "0,1,2,3",
@@ -2488,6 +2630,7 @@ · 6 unmodified lines
24882630 "CounterHTOff": "0,1,2,3"
24892631 },
24902632 {
2633+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24912634 "EventCode": "0xB7, 0xBB",
24922635 "MSRValue": "0x3f80020090 ",
24932636 "Counter": "0,1,2,3",
@@ -2500,6 +2643,7 @@ · 6 unmodified lines
25002643 "CounterHTOff": "0,1,2,3"
25012644 },
25022645 {
2646+ "PublicDescription": "Counts all prefetch data reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25032647 "EventCode": "0xB7, 0xBB",
25042648 "MSRValue": "0x00803c0090 ",
25052649 "Counter": "0,1,2,3",
@@ -2512,6 +2656,7 @@ · 6 unmodified lines
25122656 "CounterHTOff": "0,1,2,3"
25132657 },
25142658 {
2659+ "PublicDescription": "Counts all prefetch data reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25152660 "EventCode": "0xB7, 0xBB",
25162661 "MSRValue": "0x01003c0090 ",
25172662 "Counter": "0,1,2,3",
@@ -2524,6 +2669,7 @@ · 6 unmodified lines
25242669 "CounterHTOff": "0,1,2,3"
25252670 },
25262671 {
2672+ "PublicDescription": "Counts all prefetch data reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25272673 "EventCode": "0xB7, 0xBB",
25282674 "MSRValue": "0x02003c0090 ",
25292675 "Counter": "0,1,2,3",
@@ -2536,6 +2682,7 @@ · 6 unmodified lines
25362682 "CounterHTOff": "0,1,2,3"
25372683 },
25382684 {
2685+ "PublicDescription": "Counts all prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25392686 "EventCode": "0xB7, 0xBB",
25402687 "MSRValue": "0x04003c0090 ",
25412688 "Counter": "0,1,2,3",
@@ -2548,6 +2695,7 @@ · 6 unmodified lines
25482695 "CounterHTOff": "0,1,2,3"
25492696 },
25502697 {
2698+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25512699 "EventCode": "0xB7, 0xBB",
25522700 "MSRValue": "0x10003c0090 ",
25532701 "Counter": "0,1,2,3",
@@ -2560,6 +2708,7 @@ · 6 unmodified lines
25602708 "CounterHTOff": "0,1,2,3"
25612709 },
25622710 {
2711+ "PublicDescription": "Counts all prefetch data reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25632712 "EventCode": "0xB7, 0xBB",
25642713 "MSRValue": "0x3f803c0090 ",
25652714 "Counter": "0,1,2,3",
@@ -2572,6 +2721,7 @@ · 6 unmodified lines
25722721 "CounterHTOff": "0,1,2,3"
25732722 },
25742723 {
2724+ "PublicDescription": "Counts prefetch RFOs that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25752725 "EventCode": "0xB7, 0xBB",
25762726 "MSRValue": "0x0000010120 ",
25772727 "Counter": "0,1,2,3",
@@ -2584,6 +2734,7 @@ · 6 unmodified lines
25842734 "CounterHTOff": "0,1,2,3"
25852735 },
25862736 {
2737+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25872738 "EventCode": "0xB7, 0xBB",
25882739 "MSRValue": "0x0080020120 ",
25892740 "Counter": "0,1,2,3",
@@ -2596,6 +2747,7 @@ · 6 unmodified lines
25962747 "CounterHTOff": "0,1,2,3"
25972748 },
25982749 {
2750+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25992751 "EventCode": "0xB7, 0xBB",
26002752 "MSRValue": "0x0100020120 ",
26012753 "Counter": "0,1,2,3",
@@ -2608,6 +2760,7 @@ · 6 unmodified lines
26082760 "CounterHTOff": "0,1,2,3"
26092761 },
26102762 {
2763+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26112764 "EventCode": "0xB7, 0xBB",
26122765 "MSRValue": "0x0200020120 ",
26132766 "Counter": "0,1,2,3",
@@ -2620,6 +2773,7 @@ · 6 unmodified lines
26202773 "CounterHTOff": "0,1,2,3"
26212774 },
26222775 {
2776+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26232777 "EventCode": "0xB7, 0xBB",
26242778 "MSRValue": "0x0400020120 ",
26252779 "Counter": "0,1,2,3",
@@ -2632,6 +2786,7 @@ · 6 unmodified lines
26322786 "CounterHTOff": "0,1,2,3"
26332787 },
26342788 {
2789+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26352790 "EventCode": "0xB7, 0xBB",
26362791 "MSRValue": "0x1000020120 ",
26372792 "Counter": "0,1,2,3",
@@ -2644,6 +2799,7 @@ · 6 unmodified lines
26442799 "CounterHTOff": "0,1,2,3"
26452800 },
26462801 {
2802+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26472803 "EventCode": "0xB7, 0xBB",
26482804 "MSRValue": "0x3f80020120 ",
26492805 "Counter": "0,1,2,3",
@@ -2656,6 +2812,7 @@ · 6 unmodified lines
26562812 "CounterHTOff": "0,1,2,3"
26572813 },
26582814 {
2815+ "PublicDescription": "Counts prefetch RFOs that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26592816 "EventCode": "0xB7, 0xBB",
26602817 "MSRValue": "0x00803c0120 ",
26612818 "Counter": "0,1,2,3",
@@ -2668,6 +2825,7 @@ · 6 unmodified lines
26682825 "CounterHTOff": "0,1,2,3"
26692826 },
26702827 {
2828+ "PublicDescription": "Counts prefetch RFOs that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26712829 "EventCode": "0xB7, 0xBB",
26722830 "MSRValue": "0x01003c0120 ",
26732831 "Counter": "0,1,2,3",
@@ -2680,6 +2838,7 @@ · 6 unmodified lines
26802838 "CounterHTOff": "0,1,2,3"
26812839 },
26822840 {
2841+ "PublicDescription": "Counts prefetch RFOs that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26832842 "EventCode": "0xB7, 0xBB",
26842843 "MSRValue": "0x02003c0120 ",
26852844 "Counter": "0,1,2,3",
@@ -2692,6 +2851,7 @@ · 6 unmodified lines
26922851 "CounterHTOff": "0,1,2,3"
26932852 },
26942853 {
2854+ "PublicDescription": "Counts prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26952855 "EventCode": "0xB7, 0xBB",
26962856 "MSRValue": "0x04003c0120 ",
26972857 "Counter": "0,1,2,3",
@@ -2704,6 +2864,7 @@ · 6 unmodified lines
27042864 "CounterHTOff": "0,1,2,3"
27052865 },
27062866 {
2867+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27072868 "EventCode": "0xB7, 0xBB",
27082869 "MSRValue": "0x10003c0120 ",
27092870 "Counter": "0,1,2,3",
@@ -2716,6 +2877,7 @@ · 6 unmodified lines
27162877 "CounterHTOff": "0,1,2,3"
27172878 },
27182879 {
2880+ "PublicDescription": "Counts prefetch RFOs that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27192881 "EventCode": "0xB7, 0xBB",
27202882 "MSRValue": "0x3f803c0120 ",
27212883 "Counter": "0,1,2,3",
@@ -2728,6 +2890,7 @@ · 6 unmodified lines
27282890 "CounterHTOff": "0,1,2,3"
27292891 },
27302892 {
2893+ "PublicDescription": "Counts all prefetch code reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27312894 "EventCode": "0xB7, 0xBB",
27322895 "MSRValue": "0x0000010240 ",
27332896 "Counter": "0,1,2,3",
@@ -2740,6 +2903,7 @@ · 6 unmodified lines
27402903 "CounterHTOff": "0,1,2,3"
27412904 },
27422905 {
2906+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27432907 "EventCode": "0xB7, 0xBB",
27442908 "MSRValue": "0x0080020240 ",
27452909 "Counter": "0,1,2,3",
@@ -2752,6 +2916,7 @@ · 6 unmodified lines
27522916 "CounterHTOff": "0,1,2,3"
27532917 },
27542918 {
2919+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27552920 "EventCode": "0xB7, 0xBB",
27562921 "MSRValue": "0x0100020240 ",
27572922 "Counter": "0,1,2,3",
@@ -2764,6 +2929,7 @@ · 6 unmodified lines
27642929 "CounterHTOff": "0,1,2,3"
27652930 },
27662931 {
2932+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27672933 "EventCode": "0xB7, 0xBB",
27682934 "MSRValue": "0x0200020240 ",
27692935 "Counter": "0,1,2,3",
@@ -2776,6 +2942,7 @@ · 6 unmodified lines
27762942 "CounterHTOff": "0,1,2,3"
27772943 },
27782944 {
2945+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27792946 "EventCode": "0xB7, 0xBB",
27802947 "MSRValue": "0x0400020240 ",
27812948 "Counter": "0,1,2,3",
@@ -2788,6 +2955,7 @@ · 6 unmodified lines
27882955 "CounterHTOff": "0,1,2,3"
27892956 },
27902957 {
2958+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27912959 "EventCode": "0xB7, 0xBB",
27922960 "MSRValue": "0x1000020240 ",
27932961 "Counter": "0,1,2,3",
@@ -2800,6 +2968,7 @@ · 6 unmodified lines
28002968 "CounterHTOff": "0,1,2,3"
28012969 },
28022970 {
2971+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28032972 "EventCode": "0xB7, 0xBB",
28042973 "MSRValue": "0x3f80020240 ",
28052974 "Counter": "0,1,2,3",
@@ -2812,6 +2981,7 @@ · 6 unmodified lines
28122981 "CounterHTOff": "0,1,2,3"
28132982 },
28142983 {
2984+ "PublicDescription": "Counts all prefetch code reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28152985 "EventCode": "0xB7, 0xBB",
28162986 "MSRValue": "0x00803c0240 ",
28172987 "Counter": "0,1,2,3",
@@ -2824,6 +2994,7 @@ · 6 unmodified lines
28242994 "CounterHTOff": "0,1,2,3"
28252995 },
28262996 {
2997+ "PublicDescription": "Counts all prefetch code reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28272998 "EventCode": "0xB7, 0xBB",
28282999 "MSRValue": "0x01003c0240 ",
28293000 "Counter": "0,1,2,3",
@@ -2836,6 +3007,7 @@ · 6 unmodified lines
28363007 "CounterHTOff": "0,1,2,3"
28373008 },
28383009 {
3010+ "PublicDescription": "Counts all prefetch code reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28393011 "EventCode": "0xB7, 0xBB",
28403012 "MSRValue": "0x02003c0240 ",
28413013 "Counter": "0,1,2,3",
@@ -2848,6 +3020,7 @@ · 6 unmodified lines
28483020 "CounterHTOff": "0,1,2,3"
28493021 },
28503022 {
3023+ "PublicDescription": "Counts all prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28513024 "EventCode": "0xB7, 0xBB",
28523025 "MSRValue": "0x04003c0240 ",
28533026 "Counter": "0,1,2,3",
@@ -2860,6 +3033,7 @@ · 6 unmodified lines
28603033 "CounterHTOff": "0,1,2,3"
28613034 },
28623035 {
3036+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28633037 "EventCode": "0xB7, 0xBB",
28643038 "MSRValue": "0x10003c0240 ",
28653039 "Counter": "0,1,2,3",
@@ -2872,6 +3046,7 @@ · 6 unmodified lines
28723046 "CounterHTOff": "0,1,2,3"
28733047 },
28743048 {
3049+ "PublicDescription": "Counts all prefetch code reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28753050 "EventCode": "0xB7, 0xBB",
28763051 "MSRValue": "0x3f803c0240 ",
28773052 "Counter": "0,1,2,3",
@@ -2884,6 +3059,7 @@ · 6 unmodified lines
28843059 "CounterHTOff": "0,1,2,3"
28853060 },
28863061 {
3062+ "PublicDescription": "Counts all demand & prefetch data reads that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28873063 "EventCode": "0xB7, 0xBB",
28883064 "MSRValue": "0x0000010091 ",
28893065 "Counter": "0,1,2,3",
@@ -2896,6 +3072,7 @@ · 6 unmodified lines
28963072 "CounterHTOff": "0,1,2,3"
28973073 },
28983074 {
3075+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28993076 "EventCode": "0xB7, 0xBB",
29003077 "MSRValue": "0x0080020091 ",
29013078 "Counter": "0,1,2,3",
@@ -2908,6 +3085,7 @@ · 6 unmodified lines
29083085 "CounterHTOff": "0,1,2,3"
29093086 },
29103087 {
3088+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29113089 "EventCode": "0xB7, 0xBB",
29123090 "MSRValue": "0x0100020091 ",
29133091 "Counter": "0,1,2,3",
@@ -2920,6 +3098,7 @@ · 6 unmodified lines
29203098 "CounterHTOff": "0,1,2,3"
29213099 },
29223100 {
3101+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29233102 "EventCode": "0xB7, 0xBB",
29243103 "MSRValue": "0x0200020091 ",
29253104 "Counter": "0,1,2,3",
@@ -2932,6 +3111,7 @@ · 6 unmodified lines
29323111 "CounterHTOff": "0,1,2,3"
29333112 },
29343113 {
3114+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29353115 "EventCode": "0xB7, 0xBB",
29363116 "MSRValue": "0x0400020091 ",
29373117 "Counter": "0,1,2,3",
@@ -2944,6 +3124,7 @@ · 6 unmodified lines
29443124 "CounterHTOff": "0,1,2,3"
29453125 },
29463126 {
3127+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29473128 "EventCode": "0xB7, 0xBB",
29483129 "MSRValue": "0x1000020091 ",
29493130 "Counter": "0,1,2,3",
@@ -2956,6 +3137,7 @@ · 6 unmodified lines
29563137 "CounterHTOff": "0,1,2,3"
29573138 },
29583139 {
3140+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29593141 "EventCode": "0xB7, 0xBB",
29603142 "MSRValue": "0x3f80020091 ",
29613143 "Counter": "0,1,2,3",
@@ -2968,6 +3150,7 @@ · 6 unmodified lines
29683150 "CounterHTOff": "0,1,2,3"
29693151 },
29703152 {
3153+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29713154 "EventCode": "0xB7, 0xBB",
29723155 "MSRValue": "0x00803c0091 ",
29733156 "Counter": "0,1,2,3",
@@ -2980,6 +3163,7 @@ · 6 unmodified lines
29803163 "CounterHTOff": "0,1,2,3"
29813164 },
29823165 {
3166+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29833167 "EventCode": "0xB7, 0xBB",
29843168 "MSRValue": "0x01003c0091 ",
29853169 "Counter": "0,1,2,3",
@@ -2992,6 +3176,7 @@ · 6 unmodified lines
29923176 "CounterHTOff": "0,1,2,3"
29933177 },
29943178 {
3179+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
29953180 "EventCode": "0xB7, 0xBB",
29963181 "MSRValue": "0x02003c0091 ",
29973182 "Counter": "0,1,2,3",
@@ -3004,6 +3189,7 @@ · 6 unmodified lines
30043189 "CounterHTOff": "0,1,2,3"
30053190 },
30063191 {
3192+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30073193 "EventCode": "0xB7, 0xBB",
30083194 "MSRValue": "0x04003c0091 ",
30093195 "Counter": "0,1,2,3",
@@ -3016,6 +3202,7 @@ · 6 unmodified lines
30163202 "CounterHTOff": "0,1,2,3"
30173203 },
30183204 {
3205+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30193206 "EventCode": "0xB7, 0xBB",
30203207 "MSRValue": "0x10003c0091 ",
30213208 "Counter": "0,1,2,3",
@@ -3028,6 +3215,7 @@ · 6 unmodified lines
30283215 "CounterHTOff": "0,1,2,3"
30293216 },
30303217 {
3218+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30313219 "EventCode": "0xB7, 0xBB",
30323220 "MSRValue": "0x3f803c0091 ",
30333221 "Counter": "0,1,2,3",
@@ -3040,6 +3228,7 @@ · 6 unmodified lines
30403228 "CounterHTOff": "0,1,2,3"
30413229 },
30423230 {
3231+ "PublicDescription": "Counts all demand & prefetch RFOs that have any response type. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30433232 "EventCode": "0xB7, 0xBB",
30443233 "MSRValue": "0x0000010122 ",
30453234 "Counter": "0,1,2,3",
@@ -3052,6 +3241,7 @@ · 6 unmodified lines
30523241 "CounterHTOff": "0,1,2,3"
30533242 },
30543243 {
3244+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30553245 "EventCode": "0xB7, 0xBB",
30563246 "MSRValue": "0x0080020122 ",
30573247 "Counter": "0,1,2,3",
@@ -3064,6 +3254,7 @@ · 6 unmodified lines
30643254 "CounterHTOff": "0,1,2,3"
30653255 },
30663256 {
3257+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30673258 "EventCode": "0xB7, 0xBB",
30683259 "MSRValue": "0x0100020122 ",
30693260 "Counter": "0,1,2,3",
@@ -3076,6 +3267,7 @@ · 6 unmodified lines
30763267 "CounterHTOff": "0,1,2,3"
30773268 },
30783269 {
3270+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30793271 "EventCode": "0xB7, 0xBB",
30803272 "MSRValue": "0x0200020122 ",
30813273 "Counter": "0,1,2,3",
@@ -3088,6 +3280,7 @@ · 6 unmodified lines
30883280 "CounterHTOff": "0,1,2,3"
30893281 },
30903282 {
3283+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
30913284 "EventCode": "0xB7, 0xBB",
30923285 "MSRValue": "0x0400020122 ",
30933286 "Counter": "0,1,2,3",
@@ -3100,6 +3293,7 @@ · 6 unmodified lines
31003293 "CounterHTOff": "0,1,2,3"
31013294 },
31023295 {
3296+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31033297 "EventCode": "0xB7, 0xBB",
31043298 "MSRValue": "0x1000020122 ",
31053299 "Counter": "0,1,2,3",
@@ -3112,6 +3306,7 @@ · 6 unmodified lines
31123306 "CounterHTOff": "0,1,2,3"
31133307 },
31143308 {
3309+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31153310 "EventCode": "0xB7, 0xBB",
31163311 "MSRValue": "0x3f80020122 ",
31173312 "Counter": "0,1,2,3",
@@ -3124,6 +3319,7 @@ · 6 unmodified lines
31243319 "CounterHTOff": "0,1,2,3"
31253320 },
31263321 {
3322+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31273323 "EventCode": "0xB7, 0xBB",
31283324 "MSRValue": "0x00803c0122 ",
31293325 "Counter": "0,1,2,3",
@@ -3136,6 +3332,7 @@ · 6 unmodified lines
31363332 "CounterHTOff": "0,1,2,3"
31373333 },
31383334 {
3335+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and sibling core snoops are not needed as either the core-valid bit is not set or the shared line is present in multiple cores. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31393336 "EventCode": "0xB7, 0xBB",
31403337 "MSRValue": "0x01003c0122 ",
31413338 "Counter": "0,1,2,3",
@@ -3148,6 +3345,7 @@ · 6 unmodified lines
31483345 "CounterHTOff": "0,1,2,3"
31493346 },
31503347 {
3348+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31513349 "EventCode": "0xB7, 0xBB",
31523350 "MSRValue": "0x02003c0122 ",
31533351 "Counter": "0,1,2,3",
@@ -3160,6 +3358,7 @@ · 6 unmodified lines
31603358 "CounterHTOff": "0,1,2,3"
31613359 },
31623360 {
3361+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31633362 "EventCode": "0xB7, 0xBB",
31643363 "MSRValue": "0x04003c0122 ",
31653364 "Counter": "0,1,2,3",
@@ -3172,6 +3371,7 @@ · 6 unmodified lines
31723371 "CounterHTOff": "0,1,2,3"
31733372 },
31743373 {
3374+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31753375 "EventCode": "0xB7, 0xBB",
31763376 "MSRValue": "0x10003c0122 ",
31773377 "Counter": "0,1,2,3",
@@ -3184,6 +3384,7 @@ · 6 unmodified lines
31843384 "CounterHTOff": "0,1,2,3"
31853385 },
31863386 {
3387+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
31873388 "EventCode": "0xB7, 0xBB",
31883389 "MSRValue": "0x3f803c0122 ",
31893390 "Counter": "0,1,2,3",
tools/perf/pmu-events/arch/x86/broadwell/floating-point.json+51−57
@@ -1,6 +1,6 @@
11 [
22 {
3− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
3+ "PublicDescription": "This event counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
44 "EventCode": "0xC1",
55 "Counter": "0,1,2,3",
66 "UMask": "0x8",
@@ -11,7 +11,7 @@ · 4 unmodified lines
1111 "CounterHTOff": "0,1,2,3,4,5,6,7"
1212 },
1313 {
14− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
14+ "PublicDescription": "This event counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
1515 "EventCode": "0xC1",
1616 "Counter": "0,1,2,3",
1717 "UMask": "0x10",
@@ -22,7 +22,6 @@ · 4 unmodified lines
2222 "CounterHTOff": "0,1,2,3,4,5,6,7"
2323 },
2424 {
25− "PEBS": "1",
2625 "EventCode": "0xC7",
2726 "Counter": "0,1,2,3",
2827 "UMask": "0x1",
@@ -32,7 +31,6 @@ · 3 unmodified lines
3231 "CounterHTOff": "0,1,2,3"
3332 },
3433 {
35− "PEBS": "1",
3634 "EventCode": "0xC7",
3735 "Counter": "0,1,2,3",
3836 "UMask": "0x2",
@@ -42,7 +40,15 @@ · 3 unmodified lines
4240 "CounterHTOff": "0,1,2,3"
4341 },
4442 {
45− "PEBS": "1",
43+ "EventCode": "0xC7",
44+ "Counter": "0,1,2,3",
45+ "UMask": "0x3",
46+ "EventName": "FP_ARITH_INST_RETIRED.SCALAR",
47+ "SampleAfterValue": "2000003",
48+ "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
49+ "CounterHTOff": "0,1,2,3"
50+ },
51+ {
4652 "EventCode": "0xC7",
4753 "Counter": "0,1,2,3",
4854 "UMask": "0x4",
@@ -52,7 +58,6 @@ · 3 unmodified lines
5258 "CounterHTOff": "0,1,2,3"
5359 },
5460 {
55− "PEBS": "1",
5661 "EventCode": "0xC7",
5762 "Counter": "0,1,2,3",
5863 "UMask": "0x8",
@@ -62,7 +67,6 @@ · 3 unmodified lines
6267 "CounterHTOff": "0,1,2,3"
6368 },
6469 {
65− "PEBS": "1",
6670 "EventCode": "0xC7",
6771 "Counter": "0,1,2,3",
6872 "UMask": "0x10",
@@ -72,7 +76,43 @@ · 3 unmodified lines
7276 "CounterHTOff": "0,1,2,3"
7377 },
7478 {
75− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
79+ "EventCode": "0xC7",
80+ "Counter": "0,1,2,3",
81+ "UMask": "0x15",
82+ "EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
83+ "SampleAfterValue": "2000006",
84+ "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
85+ "CounterHTOff": "0,1,2,3"
86+ },
87+ {
88+ "EventCode": "0xc7",
89+ "Counter": "0,1,2,3",
90+ "UMask": "0x20",
91+ "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
92+ "SampleAfterValue": "2000003",
93+ "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
94+ "CounterHTOff": "0,1,2,3"
95+ },
96+ {
97+ "EventCode": "0xC7",
98+ "Counter": "0,1,2,3",
99+ "UMask": "0x2a",
100+ "EventName": "FP_ARITH_INST_RETIRED.SINGLE",
101+ "SampleAfterValue": "2000005",
102+ "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
103+ "CounterHTOff": "0,1,2,3"
104+ },
105+ {
106+ "EventCode": "0xC7",
107+ "Counter": "0,1,2,3",
108+ "UMask": "0x3c",
109+ "EventName": "FP_ARITH_INST_RETIRED.PACKED",
110+ "SampleAfterValue": "2000004",
111+ "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
112+ "CounterHTOff": "0,1,2,3"
113+ },
114+ {
115+ "PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
76116 "EventCode": "0xCA",
77117 "Counter": "0,1,2,3",
78118 "UMask": "0x2",
@@ -82,7 +122,7 @@ · 3 unmodified lines
82122 "CounterHTOff": "0,1,2,3,4,5,6,7"
83123 },
84124 {
85− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
125+ "PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
86126 "EventCode": "0xCA",
87127 "Counter": "0,1,2,3",
88128 "UMask": "0x4",
@@ -92,7 +132,7 @@ · 3 unmodified lines
92132 "CounterHTOff": "0,1,2,3,4,5,6,7"
93133 },
94134 {
95− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
135+ "PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
96136 "EventCode": "0xCA",
97137 "Counter": "0,1,2,3",
98138 "UMask": "0x8",
@@ -102,7 +142,7 @@ · 3 unmodified lines
102142 "CounterHTOff": "0,1,2,3,4,5,6,7"
103143 },
104144 {
105− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
145+ "PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
106146 "EventCode": "0xCA",
107147 "Counter": "0,1,2,3",
108148 "UMask": "0x10",
@@ -121,51 +161,5 @@ · 12 unmodified lines
121161 "BriefDescription": "Cycles with any input/output SSE or FP assist",
122162 "CounterMask": "1",
123163 "CounterHTOff": "0,1,2,3"
124− },
125− {
126− "PEBS": "1",
127− "EventCode": "0xc7",
128− "Counter": "0,1,2,3",
129− "UMask": "0x20",
130− "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
131− "SampleAfterValue": "2000003",
132− "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
133− "CounterHTOff": "0,1,2,3"
134− },
135− {
136− "EventCode": "0xC7",
137− "Counter": "0,1,2,3",
138− "UMask": "0x3",
139− "EventName": "FP_ARITH_INST_RETIRED.SCALAR",
140− "SampleAfterValue": "2000003",
141− "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
142− "CounterHTOff": "0,1,2,3"
143− },
144− {
145− "EventCode": "0xC7",
146− "Counter": "0,1,2,3",
147− "UMask": "0x3c",
148− "EventName": "FP_ARITH_INST_RETIRED.PACKED",
149− "SampleAfterValue": "2000004",
150− "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
151− "CounterHTOff": "0,1,2,3"
152− },
153− {
154− "EventCode": "0xC7",
155− "Counter": "0,1,2,3",
156− "UMask": "0x2a",
157− "EventName": "FP_ARITH_INST_RETIRED.SINGLE",
158− "SampleAfterValue": "2000005",
159− "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
160− "CounterHTOff": "0,1,2,3"
161− },
162− {
163− "EventCode": "0xC7",
164− "Counter": "0,1,2,3",
165− "UMask": "0x15",
166− "EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
167− "SampleAfterValue": "2000006",
168− "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
169− "CounterHTOff": "0,1,2,3"
170164 }
171165 ]
tools/perf/pmu-events/arch/x86/broadwell/frontend.json+69−69
@@ -10,7 +10,7 @@ · 9 unmodified lines
1010 "CounterHTOff": "0,1,2,3"
1111 },
1212 {
13− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
13+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
1414 "EventCode": "0x79",
1515 "Counter": "0,1,2,3",
1616 "UMask": "0x4",
@@ -20,80 +20,49 @@ · 3 unmodified lines
2020 "CounterHTOff": "0,1,2,3,4,5,6,7"
2121 },
2222 {
23− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
23+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ.",
2424 "EventCode": "0x79",
2525 "Counter": "0,1,2,3",
26− "UMask": "0x8",
27− "EventName": "IDQ.DSB_UOPS",
28− "SampleAfterValue": "2000003",
29− "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
30− "CounterHTOff": "0,1,2,3,4,5,6,7"
31− },
32− {
33− "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
34− "EventCode": "0x79",
35− "Counter": "0,1,2,3",
36− "UMask": "0x10",
37− "EventName": "IDQ.MS_DSB_UOPS",
38− "SampleAfterValue": "2000003",
39− "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
40− "CounterHTOff": "0,1,2,3,4,5,6,7"
41− },
42− {
43− "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
44− "EventCode": "0x79",
45− "Counter": "0,1,2,3",
46− "UMask": "0x20",
47− "EventName": "IDQ.MS_MITE_UOPS",
48− "SampleAfterValue": "2000003",
49− "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
50− "CounterHTOff": "0,1,2,3,4,5,6,7"
51− },
52− {
53− "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
54− "EventCode": "0x79",
55− "Counter": "0,1,2,3",
56− "UMask": "0x30",
57− "EventName": "IDQ.MS_UOPS",
26+ "UMask": "0x4",
27+ "EventName": "IDQ.MITE_CYCLES",
5828 "SampleAfterValue": "2000003",
59− "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
29+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path",
30+ "CounterMask": "1",
6031 "CounterHTOff": "0,1,2,3,4,5,6,7"
6132 },
6233 {
63− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
34+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
6435 "EventCode": "0x79",
6536 "Counter": "0,1,2,3",
66− "UMask": "0x30",
67− "EventName": "IDQ.MS_CYCLES",
37+ "UMask": "0x8",
38+ "EventName": "IDQ.DSB_UOPS",
6839 "SampleAfterValue": "2000003",
69− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
70− "CounterMask": "1",
40+ "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
7141 "CounterHTOff": "0,1,2,3,4,5,6,7"
7242 },
7343 {
74− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ.",
44+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
7545 "EventCode": "0x79",
7646 "Counter": "0,1,2,3",
77− "UMask": "0x4",
78− "EventName": "IDQ.MITE_CYCLES",
47+ "UMask": "0x8",
48+ "EventName": "IDQ.DSB_CYCLES",
7949 "SampleAfterValue": "2000003",
80− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path",
50+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path",
8151 "CounterMask": "1",
8252 "CounterHTOff": "0,1,2,3,4,5,6,7"
8353 },
8454 {
85− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
55+ "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
8656 "EventCode": "0x79",
8757 "Counter": "0,1,2,3",
88− "UMask": "0x8",
89− "EventName": "IDQ.DSB_CYCLES",
58+ "UMask": "0x10",
59+ "EventName": "IDQ.MS_DSB_UOPS",
9060 "SampleAfterValue": "2000003",
91− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path",
92− "CounterMask": "1",
61+ "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
9362 "CounterHTOff": "0,1,2,3,4,5,6,7"
9463 },
9564 {
96− "PublicDescription": "This event counts cycles during which uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
65+ "PublicDescription": "This event counts cycles during which uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
9766 "EventCode": "0x79",
9867 "Counter": "0,1,2,3",
9968 "UMask": "0x10",
@@ -104,7 +73,7 @@ · 4 unmodified lines
10473 "CounterHTOff": "0,1,2,3,4,5,6,7"
10574 },
10675 {
107− "PublicDescription": "This event counts the number of deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
76+ "PublicDescription": "This event counts the number of deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
10877 "EventCode": "0x79",
10978 "Counter": "0,1,2,3",
11079 "UMask": "0x10",
@@ -116,7 +85,7 @@ · 5 unmodified lines
11685 "CounterHTOff": "0,1,2,3,4,5,6,7"
11786 },
11887 {
119− "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
88+ "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
12089 "EventCode": "0x79",
12190 "Counter": "0,1,2,3",
12291 "UMask": "0x18",
@@ -127,7 +96,7 @@ · 4 unmodified lines
12796 "CounterHTOff": "0,1,2,3,4,5,6,7"
12897 },
12998 {
130− "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
99+ "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
131100 "EventCode": "0x79",
132101 "Counter": "0,1,2,3",
133102 "UMask": "0x18",
@@ -138,7 +107,17 @@ · 4 unmodified lines
138107 "CounterHTOff": "0,1,2,3,4,5,6,7"
139108 },
140109 {
141− "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
110+ "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ.",
111+ "EventCode": "0x79",
112+ "Counter": "0,1,2,3",
113+ "UMask": "0x20",
114+ "EventName": "IDQ.MS_MITE_UOPS",
115+ "SampleAfterValue": "2000003",
116+ "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
117+ "CounterHTOff": "0,1,2,3,4,5,6,7"
118+ },
119+ {
120+ "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
142121 "EventCode": "0x79",
143122 "Counter": "0,1,2,3",
144123 "UMask": "0x24",
@@ -149,7 +128,7 @@ · 4 unmodified lines
149128 "CounterHTOff": "0,1,2,3,4,5,6,7"
150129 },
151130 {
152− "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
131+ "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
153132 "EventCode": "0x79",
154133 "Counter": "0,1,2,3",
155134 "UMask": "0x24",
@@ -160,7 +139,39 @@ · 4 unmodified lines
160139 "CounterHTOff": "0,1,2,3,4,5,6,7"
161140 },
162141 {
163− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
142+ "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
143+ "EventCode": "0x79",
144+ "Counter": "0,1,2,3",
145+ "UMask": "0x30",
146+ "EventName": "IDQ.MS_UOPS",
147+ "SampleAfterValue": "2000003",
148+ "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
149+ "CounterHTOff": "0,1,2,3,4,5,6,7"
150+ },
151+ {
152+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
153+ "EventCode": "0x79",
154+ "Counter": "0,1,2,3",
155+ "UMask": "0x30",
156+ "EventName": "IDQ.MS_CYCLES",
157+ "SampleAfterValue": "2000003",
158+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
159+ "CounterMask": "1",
160+ "CounterHTOff": "0,1,2,3,4,5,6,7"
161+ },
162+ {
163+ "EventCode": "0x79",
164+ "Counter": "0,1,2,3",
165+ "UMask": "0x30",
166+ "EdgeDetect": "1",
167+ "EventName": "IDQ.MS_SWITCHES",
168+ "SampleAfterValue": "2000003",
169+ "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
170+ "CounterMask": "1",
171+ "CounterHTOff": "0,1,2,3,4,5,6,7"
172+ },
173+ {
174+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
164175 "EventCode": "0x79",
165176 "Counter": "0,1,2,3",
166177 "UMask": "0x3c",
@@ -200,7 +211,7 @@ · 33 unmodified lines
200211 "CounterHTOff": "0,1,2,3,4,5,6,7"
201212 },
202213 {
203− "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding ?4 ? x? when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
214+ "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
204215 "EventCode": "0x9C",
205216 "Counter": "0,1,2,3",
206217 "UMask": "0x1",
@@ -263,7 +274,7 @@ · 56 unmodified lines
263274 "CounterHTOff": "0,1,2,3"
264275 },
265276 {
266− "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 0?2 cycles.",
277+ "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
267278 "EventCode": "0xAB",
268279 "Counter": "0,1,2,3",
269280 "UMask": "0x2",
@@ -271,16 +282,5 @@ · 1 unmodified line
271282 "SampleAfterValue": "2000003",
272283 "BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.",
273284 "CounterHTOff": "0,1,2,3,4,5,6,7"
274− },
275− {
276− "EventCode": "0x79",
277− "Counter": "0,1,2,3",
278− "UMask": "0x30",
279− "EdgeDetect": "1",
280− "EventName": "IDQ.MS_SWITCHES",
281− "SampleAfterValue": "2000003",
282− "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
283− "CounterMask": "1",
284− "CounterHTOff": "0,1,2,3,4,5,6,7"
285285 }
286286 ]
tools/perf/pmu-events/arch/x86/broadwell/memory.json+205−5
@@ -90,7 +90,6 @@ · 89 unmodified lines
9090 "CounterHTOff": "0,1,2,3,4,5,6,7"
9191 },
9292 {
93− "PublicDescription": "Unfriendly TSX abort triggered by a flowmarker.",
9493 "EventCode": "0x5d",
9594 "Counter": "0,1,2,3",
9695 "UMask": "0x1",
@@ -170,13 +169,13 @@ · 73 unmodified lines
170169 },
171170 {
172171 "PEBS": "1",
173− "PublicDescription": "Number of times HLE abort was triggered.",
172+ "PublicDescription": "Number of times HLE abort was triggered (PEBS).",
174173 "EventCode": "0xc8",
175174 "Counter": "0,1,2,3",
176175 "UMask": "0x4",
177176 "EventName": "HLE_RETIRED.ABORTED",
178177 "SampleAfterValue": "2000003",
179− "BriefDescription": "Number of times HLE abort was triggered",
178+ "BriefDescription": "Number of times HLE abort was triggered (PEBS)",
180179 "CounterHTOff": "0,1,2,3,4,5,6,7"
181180 },
182181 {
@@ -251,13 +250,13 @@ · 68 unmodified lines
251250 },
252251 {
253252 "PEBS": "1",
254− "PublicDescription": "Number of times RTM abort was triggered .",
253+ "PublicDescription": "Number of times RTM abort was triggered (PEBS).",
255254 "EventCode": "0xc9",
256255 "Counter": "0,1,2,3",
257256 "UMask": "0x4",
258257 "EventName": "RTM_RETIRED.ABORTED",
259258 "SampleAfterValue": "2000003",
260− "BriefDescription": "Number of times RTM abort was triggered",
259+ "BriefDescription": "Number of times RTM abort was triggered (PEBS)",
261260 "CounterHTOff": "0,1,2,3"
262261 },
263262 {
@@ -431,6 +430,7 @@ · 167 unmodified lines
431430 "CounterHTOff": "3"
432431 },
433432 {
433+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
434434 "EventCode": "0xB7, 0xBB",
435435 "MSRValue": "0x2000020001 ",
436436 "Counter": "0,1,2,3",
@@ -443,6 +443,7 @@ · 6 unmodified lines
443443 "CounterHTOff": "0,1,2,3"
444444 },
445445 {
446+ "PublicDescription": "Counts demand data reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
446447 "EventCode": "0xB7, 0xBB",
447448 "MSRValue": "0x20003c0001 ",
448449 "Counter": "0,1,2,3",
@@ -455,6 +456,7 @@ · 6 unmodified lines
455456 "CounterHTOff": "0,1,2,3"
456457 },
457458 {
459+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
458460 "EventCode": "0xB7, 0xBB",
459461 "MSRValue": "0x0084000001 ",
460462 "Counter": "0,1,2,3",
@@ -467,6 +469,7 @@ · 6 unmodified lines
467469 "CounterHTOff": "0,1,2,3"
468470 },
469471 {
472+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
470473 "EventCode": "0xB7, 0xBB",
471474 "MSRValue": "0x0104000001 ",
472475 "Counter": "0,1,2,3",
@@ -479,6 +482,7 @@ · 6 unmodified lines
479482 "CounterHTOff": "0,1,2,3"
480483 },
481484 {
485+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
482486 "EventCode": "0xB7, 0xBB",
483487 "MSRValue": "0x0204000001 ",
484488 "Counter": "0,1,2,3",
@@ -491,6 +495,7 @@ · 6 unmodified lines
491495 "CounterHTOff": "0,1,2,3"
492496 },
493497 {
498+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
494499 "EventCode": "0xB7, 0xBB",
495500 "MSRValue": "0x0404000001 ",
496501 "Counter": "0,1,2,3",
@@ -503,6 +508,7 @@ · 6 unmodified lines
503508 "CounterHTOff": "0,1,2,3"
504509 },
505510 {
511+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
506512 "EventCode": "0xB7, 0xBB",
507513 "MSRValue": "0x1004000001 ",
508514 "Counter": "0,1,2,3",
@@ -515,6 +521,7 @@ · 6 unmodified lines
515521 "CounterHTOff": "0,1,2,3"
516522 },
517523 {
524+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
518525 "EventCode": "0xB7, 0xBB",
519526 "MSRValue": "0x2004000001 ",
520527 "Counter": "0,1,2,3",
@@ -527,6 +534,7 @@ · 6 unmodified lines
527534 "CounterHTOff": "0,1,2,3"
528535 },
529536 {
537+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
530538 "EventCode": "0xB7, 0xBB",
531539 "MSRValue": "0x3f84000001 ",
532540 "Counter": "0,1,2,3",
@@ -539,6 +547,7 @@ · 6 unmodified lines
539547 "CounterHTOff": "0,1,2,3"
540548 },
541549 {
550+ "PublicDescription": "Counts demand data reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
542551 "EventCode": "0xB7, 0xBB",
543552 "MSRValue": "0x00bc000001 ",
544553 "Counter": "0,1,2,3",
@@ -551,6 +560,7 @@ · 6 unmodified lines
551560 "CounterHTOff": "0,1,2,3"
552561 },
553562 {
563+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
554564 "EventCode": "0xB7, 0xBB",
555565 "MSRValue": "0x013c000001 ",
556566 "Counter": "0,1,2,3",
@@ -563,6 +573,7 @@ · 6 unmodified lines
563573 "CounterHTOff": "0,1,2,3"
564574 },
565575 {
576+ "PublicDescription": "Counts demand data reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
566577 "EventCode": "0xB7, 0xBB",
567578 "MSRValue": "0x023c000001 ",
568579 "Counter": "0,1,2,3",
@@ -575,6 +586,7 @@ · 6 unmodified lines
575586 "CounterHTOff": "0,1,2,3"
576587 },
577588 {
589+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
578590 "EventCode": "0xB7, 0xBB",
579591 "MSRValue": "0x043c000001 ",
580592 "Counter": "0,1,2,3",
@@ -587,6 +599,7 @@ · 6 unmodified lines
587599 "CounterHTOff": "0,1,2,3"
588600 },
589601 {
602+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
590603 "EventCode": "0xB7, 0xBB",
591604 "MSRValue": "0x20003c0002 ",
592605 "Counter": "0,1,2,3",
@@ -599,6 +612,7 @@ · 6 unmodified lines
599612 "CounterHTOff": "0,1,2,3"
600613 },
601614 {
615+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
602616 "EventCode": "0xB7, 0xBB",
603617 "MSRValue": "0x3f84000002 ",
604618 "Counter": "0,1,2,3",
@@ -611,6 +625,7 @@ · 6 unmodified lines
611625 "CounterHTOff": "0,1,2,3"
612626 },
613627 {
628+ "PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
614629 "EventCode": "0xB7, 0xBB",
615630 "MSRValue": "0x00bc000002 ",
616631 "Counter": "0,1,2,3",
@@ -623,6 +638,7 @@ · 6 unmodified lines
623638 "CounterHTOff": "0,1,2,3"
624639 },
625640 {
641+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
626642 "EventCode": "0xB7, 0xBB",
627643 "MSRValue": "0x013c000002 ",
628644 "Counter": "0,1,2,3",
@@ -635,6 +651,7 @@ · 6 unmodified lines
635651 "CounterHTOff": "0,1,2,3"
636652 },
637653 {
654+ "PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
638655 "EventCode": "0xB7, 0xBB",
639656 "MSRValue": "0x023c000002 ",
640657 "Counter": "0,1,2,3",
@@ -647,6 +664,7 @@ · 6 unmodified lines
647664 "CounterHTOff": "0,1,2,3"
648665 },
649666 {
667+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
650668 "EventCode": "0xB7, 0xBB",
651669 "MSRValue": "0x043c000002 ",
652670 "Counter": "0,1,2,3",
@@ -659,6 +677,7 @@ · 6 unmodified lines
659677 "CounterHTOff": "0,1,2,3"
660678 },
661679 {
680+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
662681 "EventCode": "0xB7, 0xBB",
663682 "MSRValue": "0x2000020004 ",
664683 "Counter": "0,1,2,3",
@@ -671,6 +690,7 @@ · 6 unmodified lines
671690 "CounterHTOff": "0,1,2,3"
672691 },
673692 {
693+ "PublicDescription": "Counts all demand code reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
674694 "EventCode": "0xB7, 0xBB",
675695 "MSRValue": "0x20003c0004 ",
676696 "Counter": "0,1,2,3",
@@ -683,6 +703,7 @@ · 6 unmodified lines
683703 "CounterHTOff": "0,1,2,3"
684704 },
685705 {
706+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
686707 "EventCode": "0xB7, 0xBB",
687708 "MSRValue": "0x0084000004 ",
688709 "Counter": "0,1,2,3",
@@ -695,6 +716,7 @@ · 6 unmodified lines
695716 "CounterHTOff": "0,1,2,3"
696717 },
697718 {
719+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
698720 "EventCode": "0xB7, 0xBB",
699721 "MSRValue": "0x0104000004 ",
700722 "Counter": "0,1,2,3",
@@ -707,6 +729,7 @@ · 6 unmodified lines
707729 "CounterHTOff": "0,1,2,3"
708730 },
709731 {
732+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
710733 "EventCode": "0xB7, 0xBB",
711734 "MSRValue": "0x0204000004 ",
712735 "Counter": "0,1,2,3",
@@ -719,6 +742,7 @@ · 6 unmodified lines
719742 "CounterHTOff": "0,1,2,3"
720743 },
721744 {
745+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
722746 "EventCode": "0xB7, 0xBB",
723747 "MSRValue": "0x0404000004 ",
724748 "Counter": "0,1,2,3",
@@ -731,6 +755,7 @@ · 6 unmodified lines
731755 "CounterHTOff": "0,1,2,3"
732756 },
733757 {
758+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
734759 "EventCode": "0xB7, 0xBB",
735760 "MSRValue": "0x1004000004 ",
736761 "Counter": "0,1,2,3",
@@ -743,6 +768,7 @@ · 6 unmodified lines
743768 "CounterHTOff": "0,1,2,3"
744769 },
745770 {
771+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
746772 "EventCode": "0xB7, 0xBB",
747773 "MSRValue": "0x2004000004 ",
748774 "Counter": "0,1,2,3",
@@ -755,6 +781,7 @@ · 6 unmodified lines
755781 "CounterHTOff": "0,1,2,3"
756782 },
757783 {
784+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
758785 "EventCode": "0xB7, 0xBB",
759786 "MSRValue": "0x3f84000004 ",
760787 "Counter": "0,1,2,3",
@@ -767,6 +794,7 @@ · 6 unmodified lines
767794 "CounterHTOff": "0,1,2,3"
768795 },
769796 {
797+ "PublicDescription": "Counts all demand code reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
770798 "EventCode": "0xB7, 0xBB",
771799 "MSRValue": "0x00bc000004 ",
772800 "Counter": "0,1,2,3",
@@ -779,6 +807,7 @@ · 6 unmodified lines
779807 "CounterHTOff": "0,1,2,3"
780808 },
781809 {
810+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
782811 "EventCode": "0xB7, 0xBB",
783812 "MSRValue": "0x013c000004 ",
784813 "Counter": "0,1,2,3",
@@ -791,6 +820,7 @@ · 6 unmodified lines
791820 "CounterHTOff": "0,1,2,3"
792821 },
793822 {
823+ "PublicDescription": "Counts all demand code reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
794824 "EventCode": "0xB7, 0xBB",
795825 "MSRValue": "0x023c000004 ",
796826 "Counter": "0,1,2,3",
@@ -803,6 +833,7 @@ · 6 unmodified lines
803833 "CounterHTOff": "0,1,2,3"
804834 },
805835 {
836+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
806837 "EventCode": "0xB7, 0xBB",
807838 "MSRValue": "0x043c000004 ",
808839 "Counter": "0,1,2,3",
@@ -815,6 +846,7 @@ · 6 unmodified lines
815846 "CounterHTOff": "0,1,2,3"
816847 },
817848 {
849+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
818850 "EventCode": "0xB7, 0xBB",
819851 "MSRValue": "0x2000020008 ",
820852 "Counter": "0,1,2,3",
@@ -827,6 +859,7 @@ · 6 unmodified lines
827859 "CounterHTOff": "0,1,2,3"
828860 },
829861 {
862+ "PublicDescription": "Counts writebacks (modified to exclusive) that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
830863 "EventCode": "0xB7, 0xBB",
831864 "MSRValue": "0x20003c0008 ",
832865 "Counter": "0,1,2,3",
@@ -839,6 +872,7 @@ · 6 unmodified lines
839872 "CounterHTOff": "0,1,2,3"
840873 },
841874 {
875+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
842876 "EventCode": "0xB7, 0xBB",
843877 "MSRValue": "0x0084000008 ",
844878 "Counter": "0,1,2,3",
@@ -851,6 +885,7 @@ · 6 unmodified lines
851885 "CounterHTOff": "0,1,2,3"
852886 },
853887 {
888+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
854889 "EventCode": "0xB7, 0xBB",
855890 "MSRValue": "0x0104000008 ",
856891 "Counter": "0,1,2,3",
@@ -863,6 +898,7 @@ · 6 unmodified lines
863898 "CounterHTOff": "0,1,2,3"
864899 },
865900 {
901+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
866902 "EventCode": "0xB7, 0xBB",
867903 "MSRValue": "0x0204000008 ",
868904 "Counter": "0,1,2,3",
@@ -875,6 +911,7 @@ · 6 unmodified lines
875911 "CounterHTOff": "0,1,2,3"
876912 },
877913 {
914+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
878915 "EventCode": "0xB7, 0xBB",
879916 "MSRValue": "0x0404000008 ",
880917 "Counter": "0,1,2,3",
@@ -887,6 +924,7 @@ · 6 unmodified lines
887924 "CounterHTOff": "0,1,2,3"
888925 },
889926 {
927+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
890928 "EventCode": "0xB7, 0xBB",
891929 "MSRValue": "0x1004000008 ",
892930 "Counter": "0,1,2,3",
@@ -899,6 +937,7 @@ · 6 unmodified lines
899937 "CounterHTOff": "0,1,2,3"
900938 },
901939 {
940+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
902941 "EventCode": "0xB7, 0xBB",
903942 "MSRValue": "0x2004000008 ",
904943 "Counter": "0,1,2,3",
@@ -911,6 +950,7 @@ · 6 unmodified lines
911950 "CounterHTOff": "0,1,2,3"
912951 },
913952 {
953+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
914954 "EventCode": "0xB7, 0xBB",
915955 "MSRValue": "0x3f84000008 ",
916956 "Counter": "0,1,2,3",
@@ -923,6 +963,7 @@ · 6 unmodified lines
923963 "CounterHTOff": "0,1,2,3"
924964 },
925965 {
966+ "PublicDescription": "Counts writebacks (modified to exclusive) that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
926967 "EventCode": "0xB7, 0xBB",
927968 "MSRValue": "0x00bc000008 ",
928969 "Counter": "0,1,2,3",
@@ -935,6 +976,7 @@ · 6 unmodified lines
935976 "CounterHTOff": "0,1,2,3"
936977 },
937978 {
979+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
938980 "EventCode": "0xB7, 0xBB",
939981 "MSRValue": "0x013c000008 ",
940982 "Counter": "0,1,2,3",
@@ -947,6 +989,7 @@ · 6 unmodified lines
947989 "CounterHTOff": "0,1,2,3"
948990 },
949991 {
992+ "PublicDescription": "Counts writebacks (modified to exclusive) that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
950993 "EventCode": "0xB7, 0xBB",
951994 "MSRValue": "0x023c000008 ",
952995 "Counter": "0,1,2,3",
@@ -959,6 +1002,7 @@ · 6 unmodified lines
9591002 "CounterHTOff": "0,1,2,3"
9601003 },
9611004 {
1005+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
9621006 "EventCode": "0xB7, 0xBB",
9631007 "MSRValue": "0x043c000008 ",
9641008 "Counter": "0,1,2,3",
@@ -971,6 +1015,7 @@ · 6 unmodified lines
9711015 "CounterHTOff": "0,1,2,3"
9721016 },
9731017 {
1018+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
9741019 "EventCode": "0xB7, 0xBB",
9751020 "MSRValue": "0x2000020010 ",
9761021 "Counter": "0,1,2,3",
@@ -983,6 +1028,7 @@ · 6 unmodified lines
9831028 "CounterHTOff": "0,1,2,3"
9841029 },
9851030 {
1031+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
9861032 "EventCode": "0xB7, 0xBB",
9871033 "MSRValue": "0x20003c0010 ",
9881034 "Counter": "0,1,2,3",
@@ -995,6 +1041,7 @@ · 6 unmodified lines
9951041 "CounterHTOff": "0,1,2,3"
9961042 },
9971043 {
1044+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
9981045 "EventCode": "0xB7, 0xBB",
9991046 "MSRValue": "0x0084000010 ",
10001047 "Counter": "0,1,2,3",
@@ -1007,6 +1054,7 @@ · 6 unmodified lines
10071054 "CounterHTOff": "0,1,2,3"
10081055 },
10091056 {
1057+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10101058 "EventCode": "0xB7, 0xBB",
10111059 "MSRValue": "0x0104000010 ",
10121060 "Counter": "0,1,2,3",
@@ -1019,6 +1067,7 @@ · 6 unmodified lines
10191067 "CounterHTOff": "0,1,2,3"
10201068 },
10211069 {
1070+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10221071 "EventCode": "0xB7, 0xBB",
10231072 "MSRValue": "0x0204000010 ",
10241073 "Counter": "0,1,2,3",
@@ -1031,6 +1080,7 @@ · 6 unmodified lines
10311080 "CounterHTOff": "0,1,2,3"
10321081 },
10331082 {
1083+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10341084 "EventCode": "0xB7, 0xBB",
10351085 "MSRValue": "0x0404000010 ",
10361086 "Counter": "0,1,2,3",
@@ -1043,6 +1093,7 @@ · 6 unmodified lines
10431093 "CounterHTOff": "0,1,2,3"
10441094 },
10451095 {
1096+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10461097 "EventCode": "0xB7, 0xBB",
10471098 "MSRValue": "0x1004000010 ",
10481099 "Counter": "0,1,2,3",
@@ -1055,6 +1106,7 @@ · 6 unmodified lines
10551106 "CounterHTOff": "0,1,2,3"
10561107 },
10571108 {
1109+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10581110 "EventCode": "0xB7, 0xBB",
10591111 "MSRValue": "0x2004000010 ",
10601112 "Counter": "0,1,2,3",
@@ -1067,6 +1119,7 @@ · 6 unmodified lines
10671119 "CounterHTOff": "0,1,2,3"
10681120 },
10691121 {
1122+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10701123 "EventCode": "0xB7, 0xBB",
10711124 "MSRValue": "0x3f84000010 ",
10721125 "Counter": "0,1,2,3",
@@ -1079,6 +1132,7 @@ · 6 unmodified lines
10791132 "CounterHTOff": "0,1,2,3"
10801133 },
10811134 {
1135+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10821136 "EventCode": "0xB7, 0xBB",
10831137 "MSRValue": "0x00bc000010 ",
10841138 "Counter": "0,1,2,3",
@@ -1091,6 +1145,7 @@ · 6 unmodified lines
10911145 "CounterHTOff": "0,1,2,3"
10921146 },
10931147 {
1148+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
10941149 "EventCode": "0xB7, 0xBB",
10951150 "MSRValue": "0x013c000010 ",
10961151 "Counter": "0,1,2,3",
@@ -1103,6 +1158,7 @@ · 6 unmodified lines
11031158 "CounterHTOff": "0,1,2,3"
11041159 },
11051160 {
1161+ "PublicDescription": "Counts prefetch (that bring data to L2) data reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11061162 "EventCode": "0xB7, 0xBB",
11071163 "MSRValue": "0x023c000010 ",
11081164 "Counter": "0,1,2,3",
@@ -1115,6 +1171,7 @@ · 6 unmodified lines
11151171 "CounterHTOff": "0,1,2,3"
11161172 },
11171173 {
1174+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11181175 "EventCode": "0xB7, 0xBB",
11191176 "MSRValue": "0x043c000010 ",
11201177 "Counter": "0,1,2,3",
@@ -1127,6 +1184,7 @@ · 6 unmodified lines
11271184 "CounterHTOff": "0,1,2,3"
11281185 },
11291186 {
1187+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11301188 "EventCode": "0xB7, 0xBB",
11311189 "MSRValue": "0x2000020020 ",
11321190 "Counter": "0,1,2,3",
@@ -1139,6 +1197,7 @@ · 6 unmodified lines
11391197 "CounterHTOff": "0,1,2,3"
11401198 },
11411199 {
1200+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11421201 "EventCode": "0xB7, 0xBB",
11431202 "MSRValue": "0x20003c0020 ",
11441203 "Counter": "0,1,2,3",
@@ -1151,6 +1210,7 @@ · 6 unmodified lines
11511210 "CounterHTOff": "0,1,2,3"
11521211 },
11531212 {
1213+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11541214 "EventCode": "0xB7, 0xBB",
11551215 "MSRValue": "0x0084000020 ",
11561216 "Counter": "0,1,2,3",
@@ -1163,6 +1223,7 @@ · 6 unmodified lines
11631223 "CounterHTOff": "0,1,2,3"
11641224 },
11651225 {
1226+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11661227 "EventCode": "0xB7, 0xBB",
11671228 "MSRValue": "0x0104000020 ",
11681229 "Counter": "0,1,2,3",
@@ -1175,6 +1236,7 @@ · 6 unmodified lines
11751236 "CounterHTOff": "0,1,2,3"
11761237 },
11771238 {
1239+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11781240 "EventCode": "0xB7, 0xBB",
11791241 "MSRValue": "0x0204000020 ",
11801242 "Counter": "0,1,2,3",
@@ -1187,6 +1249,7 @@ · 6 unmodified lines
11871249 "CounterHTOff": "0,1,2,3"
11881250 },
11891251 {
1252+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
11901253 "EventCode": "0xB7, 0xBB",
11911254 "MSRValue": "0x0404000020 ",
11921255 "Counter": "0,1,2,3",
@@ -1199,6 +1262,7 @@ · 6 unmodified lines
11991262 "CounterHTOff": "0,1,2,3"
12001263 },
12011264 {
1265+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12021266 "EventCode": "0xB7, 0xBB",
12031267 "MSRValue": "0x1004000020 ",
12041268 "Counter": "0,1,2,3",
@@ -1211,6 +1275,7 @@ · 6 unmodified lines
12111275 "CounterHTOff": "0,1,2,3"
12121276 },
12131277 {
1278+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12141279 "EventCode": "0xB7, 0xBB",
12151280 "MSRValue": "0x2004000020 ",
12161281 "Counter": "0,1,2,3",
@@ -1223,6 +1288,7 @@ · 6 unmodified lines
12231288 "CounterHTOff": "0,1,2,3"
12241289 },
12251290 {
1291+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12261292 "EventCode": "0xB7, 0xBB",
12271293 "MSRValue": "0x3f84000020 ",
12281294 "Counter": "0,1,2,3",
@@ -1235,6 +1301,7 @@ · 6 unmodified lines
12351301 "CounterHTOff": "0,1,2,3"
12361302 },
12371303 {
1304+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12381305 "EventCode": "0xB7, 0xBB",
12391306 "MSRValue": "0x00bc000020 ",
12401307 "Counter": "0,1,2,3",
@@ -1247,6 +1314,7 @@ · 6 unmodified lines
12471314 "CounterHTOff": "0,1,2,3"
12481315 },
12491316 {
1317+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12501318 "EventCode": "0xB7, 0xBB",
12511319 "MSRValue": "0x013c000020 ",
12521320 "Counter": "0,1,2,3",
@@ -1259,6 +1327,7 @@ · 6 unmodified lines
12591327 "CounterHTOff": "0,1,2,3"
12601328 },
12611329 {
1330+ "PublicDescription": "Counts all prefetch (that bring data to L2) RFOs that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12621331 "EventCode": "0xB7, 0xBB",
12631332 "MSRValue": "0x023c000020 ",
12641333 "Counter": "0,1,2,3",
@@ -1271,6 +1340,7 @@ · 6 unmodified lines
12711340 "CounterHTOff": "0,1,2,3"
12721341 },
12731342 {
1343+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12741344 "EventCode": "0xB7, 0xBB",
12751345 "MSRValue": "0x043c000020 ",
12761346 "Counter": "0,1,2,3",
@@ -1283,6 +1353,7 @@ · 6 unmodified lines
12831353 "CounterHTOff": "0,1,2,3"
12841354 },
12851355 {
1356+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12861357 "EventCode": "0xB7, 0xBB",
12871358 "MSRValue": "0x2000020040 ",
12881359 "Counter": "0,1,2,3",
@@ -1295,6 +1366,7 @@ · 6 unmodified lines
12951366 "CounterHTOff": "0,1,2,3"
12961367 },
12971368 {
1369+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
12981370 "EventCode": "0xB7, 0xBB",
12991371 "MSRValue": "0x20003c0040 ",
13001372 "Counter": "0,1,2,3",
@@ -1307,6 +1379,7 @@ · 6 unmodified lines
13071379 "CounterHTOff": "0,1,2,3"
13081380 },
13091381 {
1382+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13101383 "EventCode": "0xB7, 0xBB",
13111384 "MSRValue": "0x0084000040 ",
13121385 "Counter": "0,1,2,3",
@@ -1319,6 +1392,7 @@ · 6 unmodified lines
13191392 "CounterHTOff": "0,1,2,3"
13201393 },
13211394 {
1395+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13221396 "EventCode": "0xB7, 0xBB",
13231397 "MSRValue": "0x0104000040 ",
13241398 "Counter": "0,1,2,3",
@@ -1331,6 +1405,7 @@ · 6 unmodified lines
13311405 "CounterHTOff": "0,1,2,3"
13321406 },
13331407 {
1408+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13341409 "EventCode": "0xB7, 0xBB",
13351410 "MSRValue": "0x0204000040 ",
13361411 "Counter": "0,1,2,3",
@@ -1343,6 +1418,7 @@ · 6 unmodified lines
13431418 "CounterHTOff": "0,1,2,3"
13441419 },
13451420 {
1421+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13461422 "EventCode": "0xB7, 0xBB",
13471423 "MSRValue": "0x0404000040 ",
13481424 "Counter": "0,1,2,3",
@@ -1355,6 +1431,7 @@ · 6 unmodified lines
13551431 "CounterHTOff": "0,1,2,3"
13561432 },
13571433 {
1434+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13581435 "EventCode": "0xB7, 0xBB",
13591436 "MSRValue": "0x1004000040 ",
13601437 "Counter": "0,1,2,3",
@@ -1367,6 +1444,7 @@ · 6 unmodified lines
13671444 "CounterHTOff": "0,1,2,3"
13681445 },
13691446 {
1447+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13701448 "EventCode": "0xB7, 0xBB",
13711449 "MSRValue": "0x2004000040 ",
13721450 "Counter": "0,1,2,3",
@@ -1379,6 +1457,7 @@ · 6 unmodified lines
13791457 "CounterHTOff": "0,1,2,3"
13801458 },
13811459 {
1460+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13821461 "EventCode": "0xB7, 0xBB",
13831462 "MSRValue": "0x3f84000040 ",
13841463 "Counter": "0,1,2,3",
@@ -1391,6 +1470,7 @@ · 6 unmodified lines
13911470 "CounterHTOff": "0,1,2,3"
13921471 },
13931472 {
1473+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
13941474 "EventCode": "0xB7, 0xBB",
13951475 "MSRValue": "0x00bc000040 ",
13961476 "Counter": "0,1,2,3",
@@ -1403,6 +1483,7 @@ · 6 unmodified lines
14031483 "CounterHTOff": "0,1,2,3"
14041484 },
14051485 {
1486+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14061487 "EventCode": "0xB7, 0xBB",
14071488 "MSRValue": "0x013c000040 ",
14081489 "Counter": "0,1,2,3",
@@ -1415,6 +1496,7 @@ · 6 unmodified lines
14151496 "CounterHTOff": "0,1,2,3"
14161497 },
14171498 {
1499+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) code reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14181500 "EventCode": "0xB7, 0xBB",
14191501 "MSRValue": "0x023c000040 ",
14201502 "Counter": "0,1,2,3",
@@ -1427,6 +1509,7 @@ · 6 unmodified lines
14271509 "CounterHTOff": "0,1,2,3"
14281510 },
14291511 {
1512+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14301513 "EventCode": "0xB7, 0xBB",
14311514 "MSRValue": "0x043c000040 ",
14321515 "Counter": "0,1,2,3",
@@ -1439,6 +1522,7 @@ · 6 unmodified lines
14391522 "CounterHTOff": "0,1,2,3"
14401523 },
14411524 {
1525+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14421526 "EventCode": "0xB7, 0xBB",
14431527 "MSRValue": "0x2000020080 ",
14441528 "Counter": "0,1,2,3",
@@ -1451,6 +1535,7 @@ · 6 unmodified lines
14511535 "CounterHTOff": "0,1,2,3"
14521536 },
14531537 {
1538+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14541539 "EventCode": "0xB7, 0xBB",
14551540 "MSRValue": "0x20003c0080 ",
14561541 "Counter": "0,1,2,3",
@@ -1463,6 +1548,7 @@ · 6 unmodified lines
14631548 "CounterHTOff": "0,1,2,3"
14641549 },
14651550 {
1551+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14661552 "EventCode": "0xB7, 0xBB",
14671553 "MSRValue": "0x0084000080 ",
14681554 "Counter": "0,1,2,3",
@@ -1475,6 +1561,7 @@ · 6 unmodified lines
14751561 "CounterHTOff": "0,1,2,3"
14761562 },
14771563 {
1564+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14781565 "EventCode": "0xB7, 0xBB",
14791566 "MSRValue": "0x0104000080 ",
14801567 "Counter": "0,1,2,3",
@@ -1487,6 +1574,7 @@ · 6 unmodified lines
14871574 "CounterHTOff": "0,1,2,3"
14881575 },
14891576 {
1577+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
14901578 "EventCode": "0xB7, 0xBB",
14911579 "MSRValue": "0x0204000080 ",
14921580 "Counter": "0,1,2,3",
@@ -1499,6 +1587,7 @@ · 6 unmodified lines
14991587 "CounterHTOff": "0,1,2,3"
15001588 },
15011589 {
1590+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15021591 "EventCode": "0xB7, 0xBB",
15031592 "MSRValue": "0x0404000080 ",
15041593 "Counter": "0,1,2,3",
@@ -1511,6 +1600,7 @@ · 6 unmodified lines
15111600 "CounterHTOff": "0,1,2,3"
15121601 },
15131602 {
1603+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15141604 "EventCode": "0xB7, 0xBB",
15151605 "MSRValue": "0x1004000080 ",
15161606 "Counter": "0,1,2,3",
@@ -1523,6 +1613,7 @@ · 6 unmodified lines
15231613 "CounterHTOff": "0,1,2,3"
15241614 },
15251615 {
1616+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15261617 "EventCode": "0xB7, 0xBB",
15271618 "MSRValue": "0x2004000080 ",
15281619 "Counter": "0,1,2,3",
@@ -1535,6 +1626,7 @@ · 6 unmodified lines
15351626 "CounterHTOff": "0,1,2,3"
15361627 },
15371628 {
1629+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15381630 "EventCode": "0xB7, 0xBB",
15391631 "MSRValue": "0x3f84000080 ",
15401632 "Counter": "0,1,2,3",
@@ -1547,6 +1639,7 @@ · 6 unmodified lines
15471639 "CounterHTOff": "0,1,2,3"
15481640 },
15491641 {
1642+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15501643 "EventCode": "0xB7, 0xBB",
15511644 "MSRValue": "0x00bc000080 ",
15521645 "Counter": "0,1,2,3",
@@ -1559,6 +1652,7 @@ · 6 unmodified lines
15591652 "CounterHTOff": "0,1,2,3"
15601653 },
15611654 {
1655+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15621656 "EventCode": "0xB7, 0xBB",
15631657 "MSRValue": "0x013c000080 ",
15641658 "Counter": "0,1,2,3",
@@ -1571,6 +1665,7 @@ · 6 unmodified lines
15711665 "CounterHTOff": "0,1,2,3"
15721666 },
15731667 {
1668+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) data reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15741669 "EventCode": "0xB7, 0xBB",
15751670 "MSRValue": "0x023c000080 ",
15761671 "Counter": "0,1,2,3",
@@ -1583,6 +1678,7 @@ · 6 unmodified lines
15831678 "CounterHTOff": "0,1,2,3"
15841679 },
15851680 {
1681+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15861682 "EventCode": "0xB7, 0xBB",
15871683 "MSRValue": "0x043c000080 ",
15881684 "Counter": "0,1,2,3",
@@ -1595,6 +1691,7 @@ · 6 unmodified lines
15951691 "CounterHTOff": "0,1,2,3"
15961692 },
15971693 {
1694+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
15981695 "EventCode": "0xB7, 0xBB",
15991696 "MSRValue": "0x2000020100 ",
16001697 "Counter": "0,1,2,3",
@@ -1607,6 +1704,7 @@ · 6 unmodified lines
16071704 "CounterHTOff": "0,1,2,3"
16081705 },
16091706 {
1707+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16101708 "EventCode": "0xB7, 0xBB",
16111709 "MSRValue": "0x20003c0100 ",
16121710 "Counter": "0,1,2,3",
@@ -1619,6 +1717,7 @@ · 6 unmodified lines
16191717 "CounterHTOff": "0,1,2,3"
16201718 },
16211719 {
1720+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16221721 "EventCode": "0xB7, 0xBB",
16231722 "MSRValue": "0x0084000100 ",
16241723 "Counter": "0,1,2,3",
@@ -1631,6 +1730,7 @@ · 6 unmodified lines
16311730 "CounterHTOff": "0,1,2,3"
16321731 },
16331732 {
1733+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16341734 "EventCode": "0xB7, 0xBB",
16351735 "MSRValue": "0x0104000100 ",
16361736 "Counter": "0,1,2,3",
@@ -1643,6 +1743,7 @@ · 6 unmodified lines
16431743 "CounterHTOff": "0,1,2,3"
16441744 },
16451745 {
1746+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16461747 "EventCode": "0xB7, 0xBB",
16471748 "MSRValue": "0x0204000100 ",
16481749 "Counter": "0,1,2,3",
@@ -1655,6 +1756,7 @@ · 6 unmodified lines
16551756 "CounterHTOff": "0,1,2,3"
16561757 },
16571758 {
1759+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16581760 "EventCode": "0xB7, 0xBB",
16591761 "MSRValue": "0x0404000100 ",
16601762 "Counter": "0,1,2,3",
@@ -1667,6 +1769,7 @@ · 6 unmodified lines
16671769 "CounterHTOff": "0,1,2,3"
16681770 },
16691771 {
1772+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16701773 "EventCode": "0xB7, 0xBB",
16711774 "MSRValue": "0x1004000100 ",
16721775 "Counter": "0,1,2,3",
@@ -1679,6 +1782,7 @@ · 6 unmodified lines
16791782 "CounterHTOff": "0,1,2,3"
16801783 },
16811784 {
1785+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16821786 "EventCode": "0xB7, 0xBB",
16831787 "MSRValue": "0x2004000100 ",
16841788 "Counter": "0,1,2,3",
@@ -1691,6 +1795,7 @@ · 6 unmodified lines
16911795 "CounterHTOff": "0,1,2,3"
16921796 },
16931797 {
1798+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
16941799 "EventCode": "0xB7, 0xBB",
16951800 "MSRValue": "0x3f84000100 ",
16961801 "Counter": "0,1,2,3",
@@ -1703,6 +1808,7 @@ · 6 unmodified lines
17031808 "CounterHTOff": "0,1,2,3"
17041809 },
17051810 {
1811+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17061812 "EventCode": "0xB7, 0xBB",
17071813 "MSRValue": "0x00bc000100 ",
17081814 "Counter": "0,1,2,3",
@@ -1715,6 +1821,7 @@ · 6 unmodified lines
17151821 "CounterHTOff": "0,1,2,3"
17161822 },
17171823 {
1824+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17181825 "EventCode": "0xB7, 0xBB",
17191826 "MSRValue": "0x013c000100 ",
17201827 "Counter": "0,1,2,3",
@@ -1727,6 +1834,7 @@ · 6 unmodified lines
17271834 "CounterHTOff": "0,1,2,3"
17281835 },
17291836 {
1837+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17301838 "EventCode": "0xB7, 0xBB",
17311839 "MSRValue": "0x023c000100 ",
17321840 "Counter": "0,1,2,3",
@@ -1739,6 +1847,7 @@ · 6 unmodified lines
17391847 "CounterHTOff": "0,1,2,3"
17401848 },
17411849 {
1850+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17421851 "EventCode": "0xB7, 0xBB",
17431852 "MSRValue": "0x043c000100 ",
17441853 "Counter": "0,1,2,3",
@@ -1751,6 +1860,7 @@ · 6 unmodified lines
17511860 "CounterHTOff": "0,1,2,3"
17521861 },
17531862 {
1863+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17541864 "EventCode": "0xB7, 0xBB",
17551865 "MSRValue": "0x2000020200 ",
17561866 "Counter": "0,1,2,3",
@@ -1763,6 +1873,7 @@ · 6 unmodified lines
17631873 "CounterHTOff": "0,1,2,3"
17641874 },
17651875 {
1876+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17661877 "EventCode": "0xB7, 0xBB",
17671878 "MSRValue": "0x20003c0200 ",
17681879 "Counter": "0,1,2,3",
@@ -1775,6 +1886,7 @@ · 6 unmodified lines
17751886 "CounterHTOff": "0,1,2,3"
17761887 },
17771888 {
1889+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17781890 "EventCode": "0xB7, 0xBB",
17791891 "MSRValue": "0x0084000200 ",
17801892 "Counter": "0,1,2,3",
@@ -1787,6 +1899,7 @@ · 6 unmodified lines
17871899 "CounterHTOff": "0,1,2,3"
17881900 },
17891901 {
1902+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
17901903 "EventCode": "0xB7, 0xBB",
17911904 "MSRValue": "0x0104000200 ",
17921905 "Counter": "0,1,2,3",
@@ -1799,6 +1912,7 @@ · 6 unmodified lines
17991912 "CounterHTOff": "0,1,2,3"
18001913 },
18011914 {
1915+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18021916 "EventCode": "0xB7, 0xBB",
18031917 "MSRValue": "0x0204000200 ",
18041918 "Counter": "0,1,2,3",
@@ -1811,6 +1925,7 @@ · 6 unmodified lines
18111925 "CounterHTOff": "0,1,2,3"
18121926 },
18131927 {
1928+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18141929 "EventCode": "0xB7, 0xBB",
18151930 "MSRValue": "0x0404000200 ",
18161931 "Counter": "0,1,2,3",
@@ -1823,6 +1938,7 @@ · 6 unmodified lines
18231938 "CounterHTOff": "0,1,2,3"
18241939 },
18251940 {
1941+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18261942 "EventCode": "0xB7, 0xBB",
18271943 "MSRValue": "0x1004000200 ",
18281944 "Counter": "0,1,2,3",
@@ -1835,6 +1951,7 @@ · 6 unmodified lines
18351951 "CounterHTOff": "0,1,2,3"
18361952 },
18371953 {
1954+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18381955 "EventCode": "0xB7, 0xBB",
18391956 "MSRValue": "0x2004000200 ",
18401957 "Counter": "0,1,2,3",
@@ -1847,6 +1964,7 @@ · 6 unmodified lines
18471964 "CounterHTOff": "0,1,2,3"
18481965 },
18491966 {
1967+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18501968 "EventCode": "0xB7, 0xBB",
18511969 "MSRValue": "0x3f84000200 ",
18521970 "Counter": "0,1,2,3",
@@ -1859,6 +1977,7 @@ · 6 unmodified lines
18591977 "CounterHTOff": "0,1,2,3"
18601978 },
18611979 {
1980+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18621981 "EventCode": "0xB7, 0xBB",
18631982 "MSRValue": "0x00bc000200 ",
18641983 "Counter": "0,1,2,3",
@@ -1871,6 +1990,7 @@ · 6 unmodified lines
18711990 "CounterHTOff": "0,1,2,3"
18721991 },
18731992 {
1993+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18741994 "EventCode": "0xB7, 0xBB",
18751995 "MSRValue": "0x013c000200 ",
18761996 "Counter": "0,1,2,3",
@@ -1883,6 +2003,7 @@ · 6 unmodified lines
18832003 "CounterHTOff": "0,1,2,3"
18842004 },
18852005 {
2006+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18862007 "EventCode": "0xB7, 0xBB",
18872008 "MSRValue": "0x023c000200 ",
18882009 "Counter": "0,1,2,3",
@@ -1895,6 +2016,7 @@ · 6 unmodified lines
18952016 "CounterHTOff": "0,1,2,3"
18962017 },
18972018 {
2019+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
18982020 "EventCode": "0xB7, 0xBB",
18992021 "MSRValue": "0x043c000200 ",
19002022 "Counter": "0,1,2,3",
@@ -1907,6 +2029,7 @@ · 6 unmodified lines
19072029 "CounterHTOff": "0,1,2,3"
19082030 },
19092031 {
2032+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19102033 "EventCode": "0xB7, 0xBB",
19112034 "MSRValue": "0x2000028000 ",
19122035 "Counter": "0,1,2,3",
@@ -1919,6 +2042,7 @@ · 6 unmodified lines
19192042 "CounterHTOff": "0,1,2,3"
19202043 },
19212044 {
2045+ "PublicDescription": "Counts any other requests that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19222046 "EventCode": "0xB7, 0xBB",
19232047 "MSRValue": "0x20003c8000 ",
19242048 "Counter": "0,1,2,3",
@@ -1931,6 +2055,7 @@ · 6 unmodified lines
19312055 "CounterHTOff": "0,1,2,3"
19322056 },
19332057 {
2058+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19342059 "EventCode": "0xB7, 0xBB",
19352060 "MSRValue": "0x0084008000 ",
19362061 "Counter": "0,1,2,3",
@@ -1943,6 +2068,7 @@ · 6 unmodified lines
19432068 "CounterHTOff": "0,1,2,3"
19442069 },
19452070 {
2071+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19462072 "EventCode": "0xB7, 0xBB",
19472073 "MSRValue": "0x0104008000 ",
19482074 "Counter": "0,1,2,3",
@@ -1955,6 +2081,7 @@ · 6 unmodified lines
19552081 "CounterHTOff": "0,1,2,3"
19562082 },
19572083 {
2084+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19582085 "EventCode": "0xB7, 0xBB",
19592086 "MSRValue": "0x0204008000 ",
19602087 "Counter": "0,1,2,3",
@@ -1967,6 +2094,7 @@ · 6 unmodified lines
19672094 "CounterHTOff": "0,1,2,3"
19682095 },
19692096 {
2097+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19702098 "EventCode": "0xB7, 0xBB",
19712099 "MSRValue": "0x0404008000 ",
19722100 "Counter": "0,1,2,3",
@@ -1979,6 +2107,7 @@ · 6 unmodified lines
19792107 "CounterHTOff": "0,1,2,3"
19802108 },
19812109 {
2110+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19822111 "EventCode": "0xB7, 0xBB",
19832112 "MSRValue": "0x1004008000 ",
19842113 "Counter": "0,1,2,3",
@@ -1991,6 +2120,7 @@ · 6 unmodified lines
19912120 "CounterHTOff": "0,1,2,3"
19922121 },
19932122 {
2123+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
19942124 "EventCode": "0xB7, 0xBB",
19952125 "MSRValue": "0x2004008000 ",
19962126 "Counter": "0,1,2,3",
@@ -2003,6 +2133,7 @@ · 6 unmodified lines
20032133 "CounterHTOff": "0,1,2,3"
20042134 },
20052135 {
2136+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20062137 "EventCode": "0xB7, 0xBB",
20072138 "MSRValue": "0x3f84008000 ",
20082139 "Counter": "0,1,2,3",
@@ -2015,6 +2146,7 @@ · 6 unmodified lines
20152146 "CounterHTOff": "0,1,2,3"
20162147 },
20172148 {
2149+ "PublicDescription": "Counts any other requests that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20182150 "EventCode": "0xB7, 0xBB",
20192151 "MSRValue": "0x00bc008000 ",
20202152 "Counter": "0,1,2,3",
@@ -2027,6 +2159,7 @@ · 6 unmodified lines
20272159 "CounterHTOff": "0,1,2,3"
20282160 },
20292161 {
2162+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20302163 "EventCode": "0xB7, 0xBB",
20312164 "MSRValue": "0x013c008000 ",
20322165 "Counter": "0,1,2,3",
@@ -2039,6 +2172,7 @@ · 6 unmodified lines
20392172 "CounterHTOff": "0,1,2,3"
20402173 },
20412174 {
2175+ "PublicDescription": "Counts any other requests that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20422176 "EventCode": "0xB7, 0xBB",
20432177 "MSRValue": "0x023c008000 ",
20442178 "Counter": "0,1,2,3",
@@ -2051,6 +2185,7 @@ · 6 unmodified lines
20512185 "CounterHTOff": "0,1,2,3"
20522186 },
20532187 {
2188+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20542189 "EventCode": "0xB7, 0xBB",
20552190 "MSRValue": "0x043c008000 ",
20562191 "Counter": "0,1,2,3",
@@ -2063,6 +2198,7 @@ · 6 unmodified lines
20632198 "CounterHTOff": "0,1,2,3"
20642199 },
20652200 {
2201+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20662202 "EventCode": "0xB7, 0xBB",
20672203 "MSRValue": "0x2000020090 ",
20682204 "Counter": "0,1,2,3",
@@ -2075,6 +2211,7 @@ · 6 unmodified lines
20752211 "CounterHTOff": "0,1,2,3"
20762212 },
20772213 {
2214+ "PublicDescription": "Counts all prefetch data reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20782215 "EventCode": "0xB7, 0xBB",
20792216 "MSRValue": "0x20003c0090 ",
20802217 "Counter": "0,1,2,3",
@@ -2087,6 +2224,7 @@ · 6 unmodified lines
20872224 "CounterHTOff": "0,1,2,3"
20882225 },
20892226 {
2227+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
20902228 "EventCode": "0xB7, 0xBB",
20912229 "MSRValue": "0x0084000090 ",
20922230 "Counter": "0,1,2,3",
@@ -2099,6 +2237,7 @@ · 6 unmodified lines
20992237 "CounterHTOff": "0,1,2,3"
21002238 },
21012239 {
2240+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21022241 "EventCode": "0xB7, 0xBB",
21032242 "MSRValue": "0x0104000090 ",
21042243 "Counter": "0,1,2,3",
@@ -2111,6 +2250,7 @@ · 6 unmodified lines
21112250 "CounterHTOff": "0,1,2,3"
21122251 },
21132252 {
2253+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21142254 "EventCode": "0xB7, 0xBB",
21152255 "MSRValue": "0x0204000090 ",
21162256 "Counter": "0,1,2,3",
@@ -2123,6 +2263,7 @@ · 6 unmodified lines
21232263 "CounterHTOff": "0,1,2,3"
21242264 },
21252265 {
2266+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21262267 "EventCode": "0xB7, 0xBB",
21272268 "MSRValue": "0x0404000090 ",
21282269 "Counter": "0,1,2,3",
@@ -2135,6 +2276,7 @@ · 6 unmodified lines
21352276 "CounterHTOff": "0,1,2,3"
21362277 },
21372278 {
2279+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21382280 "EventCode": "0xB7, 0xBB",
21392281 "MSRValue": "0x1004000090 ",
21402282 "Counter": "0,1,2,3",
@@ -2147,6 +2289,7 @@ · 6 unmodified lines
21472289 "CounterHTOff": "0,1,2,3"
21482290 },
21492291 {
2292+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21502293 "EventCode": "0xB7, 0xBB",
21512294 "MSRValue": "0x2004000090 ",
21522295 "Counter": "0,1,2,3",
@@ -2159,6 +2302,7 @@ · 6 unmodified lines
21592302 "CounterHTOff": "0,1,2,3"
21602303 },
21612304 {
2305+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21622306 "EventCode": "0xB7, 0xBB",
21632307 "MSRValue": "0x3f84000090 ",
21642308 "Counter": "0,1,2,3",
@@ -2171,6 +2315,7 @@ · 6 unmodified lines
21712315 "CounterHTOff": "0,1,2,3"
21722316 },
21732317 {
2318+ "PublicDescription": "Counts all prefetch data reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21742319 "EventCode": "0xB7, 0xBB",
21752320 "MSRValue": "0x00bc000090 ",
21762321 "Counter": "0,1,2,3",
@@ -2183,6 +2328,7 @@ · 6 unmodified lines
21832328 "CounterHTOff": "0,1,2,3"
21842329 },
21852330 {
2331+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21862332 "EventCode": "0xB7, 0xBB",
21872333 "MSRValue": "0x013c000090 ",
21882334 "Counter": "0,1,2,3",
@@ -2195,6 +2341,7 @@ · 6 unmodified lines
21952341 "CounterHTOff": "0,1,2,3"
21962342 },
21972343 {
2344+ "PublicDescription": "Counts all prefetch data reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
21982345 "EventCode": "0xB7, 0xBB",
21992346 "MSRValue": "0x023c000090 ",
22002347 "Counter": "0,1,2,3",
@@ -2207,6 +2354,7 @@ · 6 unmodified lines
22072354 "CounterHTOff": "0,1,2,3"
22082355 },
22092356 {
2357+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22102358 "EventCode": "0xB7, 0xBB",
22112359 "MSRValue": "0x043c000090 ",
22122360 "Counter": "0,1,2,3",
@@ -2219,6 +2367,7 @@ · 6 unmodified lines
22192367 "CounterHTOff": "0,1,2,3"
22202368 },
22212369 {
2370+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22222371 "EventCode": "0xB7, 0xBB",
22232372 "MSRValue": "0x2000020120 ",
22242373 "Counter": "0,1,2,3",
@@ -2231,6 +2380,7 @@ · 6 unmodified lines
22312380 "CounterHTOff": "0,1,2,3"
22322381 },
22332382 {
2383+ "PublicDescription": "Counts prefetch RFOs that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22342384 "EventCode": "0xB7, 0xBB",
22352385 "MSRValue": "0x20003c0120 ",
22362386 "Counter": "0,1,2,3",
@@ -2243,6 +2393,7 @@ · 6 unmodified lines
22432393 "CounterHTOff": "0,1,2,3"
22442394 },
22452395 {
2396+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22462397 "EventCode": "0xB7, 0xBB",
22472398 "MSRValue": "0x0084000120 ",
22482399 "Counter": "0,1,2,3",
@@ -2255,6 +2406,7 @@ · 6 unmodified lines
22552406 "CounterHTOff": "0,1,2,3"
22562407 },
22572408 {
2409+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22582410 "EventCode": "0xB7, 0xBB",
22592411 "MSRValue": "0x0104000120 ",
22602412 "Counter": "0,1,2,3",
@@ -2267,6 +2419,7 @@ · 6 unmodified lines
22672419 "CounterHTOff": "0,1,2,3"
22682420 },
22692421 {
2422+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22702423 "EventCode": "0xB7, 0xBB",
22712424 "MSRValue": "0x0204000120 ",
22722425 "Counter": "0,1,2,3",
@@ -2279,6 +2432,7 @@ · 6 unmodified lines
22792432 "CounterHTOff": "0,1,2,3"
22802433 },
22812434 {
2435+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22822436 "EventCode": "0xB7, 0xBB",
22832437 "MSRValue": "0x0404000120 ",
22842438 "Counter": "0,1,2,3",
@@ -2291,6 +2445,7 @@ · 6 unmodified lines
22912445 "CounterHTOff": "0,1,2,3"
22922446 },
22932447 {
2448+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
22942449 "EventCode": "0xB7, 0xBB",
22952450 "MSRValue": "0x1004000120 ",
22962451 "Counter": "0,1,2,3",
@@ -2303,6 +2458,7 @@ · 6 unmodified lines
23032458 "CounterHTOff": "0,1,2,3"
23042459 },
23052460 {
2461+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23062462 "EventCode": "0xB7, 0xBB",
23072463 "MSRValue": "0x2004000120 ",
23082464 "Counter": "0,1,2,3",
@@ -2315,6 +2471,7 @@ · 6 unmodified lines
23152471 "CounterHTOff": "0,1,2,3"
23162472 },
23172473 {
2474+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23182475 "EventCode": "0xB7, 0xBB",
23192476 "MSRValue": "0x3f84000120 ",
23202477 "Counter": "0,1,2,3",
@@ -2327,6 +2484,7 @@ · 6 unmodified lines
23272484 "CounterHTOff": "0,1,2,3"
23282485 },
23292486 {
2487+ "PublicDescription": "Counts prefetch RFOs that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23302488 "EventCode": "0xB7, 0xBB",
23312489 "MSRValue": "0x00bc000120 ",
23322490 "Counter": "0,1,2,3",
@@ -2339,6 +2497,7 @@ · 6 unmodified lines
23392497 "CounterHTOff": "0,1,2,3"
23402498 },
23412499 {
2500+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23422501 "EventCode": "0xB7, 0xBB",
23432502 "MSRValue": "0x013c000120 ",
23442503 "Counter": "0,1,2,3",
@@ -2351,6 +2510,7 @@ · 6 unmodified lines
23512510 "CounterHTOff": "0,1,2,3"
23522511 },
23532512 {
2513+ "PublicDescription": "Counts prefetch RFOs that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23542514 "EventCode": "0xB7, 0xBB",
23552515 "MSRValue": "0x023c000120 ",
23562516 "Counter": "0,1,2,3",
@@ -2363,6 +2523,7 @@ · 6 unmodified lines
23632523 "CounterHTOff": "0,1,2,3"
23642524 },
23652525 {
2526+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23662527 "EventCode": "0xB7, 0xBB",
23672528 "MSRValue": "0x043c000120 ",
23682529 "Counter": "0,1,2,3",
@@ -2375,6 +2536,7 @@ · 6 unmodified lines
23752536 "CounterHTOff": "0,1,2,3"
23762537 },
23772538 {
2539+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23782540 "EventCode": "0xB7, 0xBB",
23792541 "MSRValue": "0x2000020240 ",
23802542 "Counter": "0,1,2,3",
@@ -2387,6 +2549,7 @@ · 6 unmodified lines
23872549 "CounterHTOff": "0,1,2,3"
23882550 },
23892551 {
2552+ "PublicDescription": "Counts all prefetch code reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
23902553 "EventCode": "0xB7, 0xBB",
23912554 "MSRValue": "0x20003c0240 ",
23922555 "Counter": "0,1,2,3",
@@ -2399,6 +2562,7 @@ · 6 unmodified lines
23992562 "CounterHTOff": "0,1,2,3"
24002563 },
24012564 {
2565+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24022566 "EventCode": "0xB7, 0xBB",
24032567 "MSRValue": "0x0084000240 ",
24042568 "Counter": "0,1,2,3",
@@ -2411,6 +2575,7 @@ · 6 unmodified lines
24112575 "CounterHTOff": "0,1,2,3"
24122576 },
24132577 {
2578+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24142579 "EventCode": "0xB7, 0xBB",
24152580 "MSRValue": "0x0104000240 ",
24162581 "Counter": "0,1,2,3",
@@ -2423,6 +2588,7 @@ · 6 unmodified lines
24232588 "CounterHTOff": "0,1,2,3"
24242589 },
24252590 {
2591+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24262592 "EventCode": "0xB7, 0xBB",
24272593 "MSRValue": "0x0204000240 ",
24282594 "Counter": "0,1,2,3",
@@ -2435,6 +2601,7 @@ · 6 unmodified lines
24352601 "CounterHTOff": "0,1,2,3"
24362602 },
24372603 {
2604+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24382605 "EventCode": "0xB7, 0xBB",
24392606 "MSRValue": "0x0404000240 ",
24402607 "Counter": "0,1,2,3",
@@ -2447,6 +2614,7 @@ · 6 unmodified lines
24472614 "CounterHTOff": "0,1,2,3"
24482615 },
24492616 {
2617+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24502618 "EventCode": "0xB7, 0xBB",
24512619 "MSRValue": "0x1004000240 ",
24522620 "Counter": "0,1,2,3",
@@ -2459,6 +2627,7 @@ · 6 unmodified lines
24592627 "CounterHTOff": "0,1,2,3"
24602628 },
24612629 {
2630+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24622631 "EventCode": "0xB7, 0xBB",
24632632 "MSRValue": "0x2004000240 ",
24642633 "Counter": "0,1,2,3",
@@ -2471,6 +2640,7 @@ · 6 unmodified lines
24712640 "CounterHTOff": "0,1,2,3"
24722641 },
24732642 {
2643+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24742644 "EventCode": "0xB7, 0xBB",
24752645 "MSRValue": "0x3f84000240 ",
24762646 "Counter": "0,1,2,3",
@@ -2483,6 +2653,7 @@ · 6 unmodified lines
24832653 "CounterHTOff": "0,1,2,3"
24842654 },
24852655 {
2656+ "PublicDescription": "Counts all prefetch code reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24862657 "EventCode": "0xB7, 0xBB",
24872658 "MSRValue": "0x00bc000240 ",
24882659 "Counter": "0,1,2,3",
@@ -2495,6 +2666,7 @@ · 6 unmodified lines
24952666 "CounterHTOff": "0,1,2,3"
24962667 },
24972668 {
2669+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
24982670 "EventCode": "0xB7, 0xBB",
24992671 "MSRValue": "0x013c000240 ",
25002672 "Counter": "0,1,2,3",
@@ -2507,6 +2679,7 @@ · 6 unmodified lines
25072679 "CounterHTOff": "0,1,2,3"
25082680 },
25092681 {
2682+ "PublicDescription": "Counts all prefetch code reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25102683 "EventCode": "0xB7, 0xBB",
25112684 "MSRValue": "0x023c000240 ",
25122685 "Counter": "0,1,2,3",
@@ -2519,6 +2692,7 @@ · 6 unmodified lines
25192692 "CounterHTOff": "0,1,2,3"
25202693 },
25212694 {
2695+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25222696 "EventCode": "0xB7, 0xBB",
25232697 "MSRValue": "0x043c000240 ",
25242698 "Counter": "0,1,2,3",
@@ -2531,6 +2705,7 @@ · 6 unmodified lines
25312705 "CounterHTOff": "0,1,2,3"
25322706 },
25332707 {
2708+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25342709 "EventCode": "0xB7, 0xBB",
25352710 "MSRValue": "0x2000020091 ",
25362711 "Counter": "0,1,2,3",
@@ -2543,6 +2718,7 @@ · 6 unmodified lines
25432718 "CounterHTOff": "0,1,2,3"
25442719 },
25452720 {
2721+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25462722 "EventCode": "0xB7, 0xBB",
25472723 "MSRValue": "0x20003c0091 ",
25482724 "Counter": "0,1,2,3",
@@ -2555,6 +2731,7 @@ · 6 unmodified lines
25552731 "CounterHTOff": "0,1,2,3"
25562732 },
25572733 {
2734+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25582735 "EventCode": "0xB7, 0xBB",
25592736 "MSRValue": "0x0084000091 ",
25602737 "Counter": "0,1,2,3",
@@ -2567,6 +2744,7 @@ · 6 unmodified lines
25672744 "CounterHTOff": "0,1,2,3"
25682745 },
25692746 {
2747+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25702748 "EventCode": "0xB7, 0xBB",
25712749 "MSRValue": "0x0104000091 ",
25722750 "Counter": "0,1,2,3",
@@ -2579,6 +2757,7 @@ · 6 unmodified lines
25792757 "CounterHTOff": "0,1,2,3"
25802758 },
25812759 {
2760+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25822761 "EventCode": "0xB7, 0xBB",
25832762 "MSRValue": "0x0204000091 ",
25842763 "Counter": "0,1,2,3",
@@ -2591,6 +2770,7 @@ · 6 unmodified lines
25912770 "CounterHTOff": "0,1,2,3"
25922771 },
25932772 {
2773+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
25942774 "EventCode": "0xB7, 0xBB",
25952775 "MSRValue": "0x0404000091 ",
25962776 "Counter": "0,1,2,3",
@@ -2603,6 +2783,7 @@ · 6 unmodified lines
26032783 "CounterHTOff": "0,1,2,3"
26042784 },
26052785 {
2786+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26062787 "EventCode": "0xB7, 0xBB",
26072788 "MSRValue": "0x1004000091 ",
26082789 "Counter": "0,1,2,3",
@@ -2615,6 +2796,7 @@ · 6 unmodified lines
26152796 "CounterHTOff": "0,1,2,3"
26162797 },
26172798 {
2799+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26182800 "EventCode": "0xB7, 0xBB",
26192801 "MSRValue": "0x2004000091 ",
26202802 "Counter": "0,1,2,3",
@@ -2627,6 +2809,7 @@ · 6 unmodified lines
26272809 "CounterHTOff": "0,1,2,3"
26282810 },
26292811 {
2812+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26302813 "EventCode": "0xB7, 0xBB",
26312814 "MSRValue": "0x3f84000091 ",
26322815 "Counter": "0,1,2,3",
@@ -2639,6 +2822,7 @@ · 6 unmodified lines
26392822 "CounterHTOff": "0,1,2,3"
26402823 },
26412824 {
2825+ "PublicDescription": "Counts all demand & prefetch data reads that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26422826 "EventCode": "0xB7, 0xBB",
26432827 "MSRValue": "0x00bc000091 ",
26442828 "Counter": "0,1,2,3",
@@ -2651,6 +2835,7 @@ · 6 unmodified lines
26512835 "CounterHTOff": "0,1,2,3"
26522836 },
26532837 {
2838+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26542839 "EventCode": "0xB7, 0xBB",
26552840 "MSRValue": "0x013c000091 ",
26562841 "Counter": "0,1,2,3",
@@ -2663,6 +2848,7 @@ · 6 unmodified lines
26632848 "CounterHTOff": "0,1,2,3"
26642849 },
26652850 {
2851+ "PublicDescription": "Counts all demand & prefetch data reads that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26662852 "EventCode": "0xB7, 0xBB",
26672853 "MSRValue": "0x023c000091 ",
26682854 "Counter": "0,1,2,3",
@@ -2675,6 +2861,7 @@ · 6 unmodified lines
26752861 "CounterHTOff": "0,1,2,3"
26762862 },
26772863 {
2864+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26782865 "EventCode": "0xB7, 0xBB",
26792866 "MSRValue": "0x043c000091 ",
26802867 "Counter": "0,1,2,3",
@@ -2687,6 +2874,7 @@ · 6 unmodified lines
26872874 "CounterHTOff": "0,1,2,3"
26882875 },
26892876 {
2877+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
26902878 "EventCode": "0xB7, 0xBB",
26912879 "MSRValue": "0x2000020122 ",
26922880 "Counter": "0,1,2,3",
@@ -2699,6 +2887,7 @@ · 6 unmodified lines
26992887 "CounterHTOff": "0,1,2,3"
27002888 },
27012889 {
2890+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the target was non-DRAM system address. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27022891 "EventCode": "0xB7, 0xBB",
27032892 "MSRValue": "0x20003c0122 ",
27042893 "Counter": "0,1,2,3",
@@ -2711,6 +2900,7 @@ · 6 unmodified lines
27112900 "CounterHTOff": "0,1,2,3"
27122901 },
27132902 {
2903+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27142904 "EventCode": "0xB7, 0xBB",
27152905 "MSRValue": "0x0084000122 ",
27162906 "Counter": "0,1,2,3",
@@ -2723,6 +2913,7 @@ · 6 unmodified lines
27232913 "CounterHTOff": "0,1,2,3"
27242914 },
27252915 {
2916+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27262917 "EventCode": "0xB7, 0xBB",
27272918 "MSRValue": "0x0104000122 ",
27282919 "Counter": "0,1,2,3",
@@ -2735,6 +2926,7 @@ · 6 unmodified lines
27352926 "CounterHTOff": "0,1,2,3"
27362927 },
27372928 {
2929+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27382930 "EventCode": "0xB7, 0xBB",
27392931 "MSRValue": "0x0204000122 ",
27402932 "Counter": "0,1,2,3",
@@ -2747,6 +2939,7 @@ · 6 unmodified lines
27472939 "CounterHTOff": "0,1,2,3"
27482940 },
27492941 {
2942+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27502943 "EventCode": "0xB7, 0xBB",
27512944 "MSRValue": "0x0404000122 ",
27522945 "Counter": "0,1,2,3",
@@ -2759,6 +2952,7 @@ · 6 unmodified lines
27592952 "CounterHTOff": "0,1,2,3"
27602953 },
27612954 {
2955+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27622956 "EventCode": "0xB7, 0xBB",
27632957 "MSRValue": "0x1004000122 ",
27642958 "Counter": "0,1,2,3",
@@ -2771,6 +2965,7 @@ · 6 unmodified lines
27712965 "CounterHTOff": "0,1,2,3"
27722966 },
27732967 {
2968+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27742969 "EventCode": "0xB7, 0xBB",
27752970 "MSRValue": "0x2004000122 ",
27762971 "Counter": "0,1,2,3",
@@ -2783,6 +2978,7 @@ · 6 unmodified lines
27832978 "CounterHTOff": "0,1,2,3"
27842979 },
27852980 {
2981+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27862982 "EventCode": "0xB7, 0xBB",
27872983 "MSRValue": "0x3f84000122 ",
27882984 "Counter": "0,1,2,3",
@@ -2795,6 +2991,7 @@ · 6 unmodified lines
27952991 "CounterHTOff": "0,1,2,3"
27962992 },
27972993 {
2994+ "PublicDescription": "Counts all demand & prefetch RFOs that miss the L3 with no details on snoop-related information. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
27982995 "EventCode": "0xB7, 0xBB",
27992996 "MSRValue": "0x00bc000122 ",
28002997 "Counter": "0,1,2,3",
@@ -2807,6 +3004,7 @@ · 6 unmodified lines
28073004 "CounterHTOff": "0,1,2,3"
28083005 },
28093006 {
3007+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28103008 "EventCode": "0xB7, 0xBB",
28113009 "MSRValue": "0x013c000122 ",
28123010 "Counter": "0,1,2,3",
@@ -2819,6 +3017,7 @@ · 6 unmodified lines
28193017 "CounterHTOff": "0,1,2,3"
28203018 },
28213019 {
3020+ "PublicDescription": "Counts all demand & prefetch RFOs that miss the L3 with a snoop miss response. Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28223021 "EventCode": "0xB7, 0xBB",
28233022 "MSRValue": "0x023c000122 ",
28243023 "Counter": "0,1,2,3",
@@ -2831,6 +3030,7 @@ · 6 unmodified lines
28313030 "CounterHTOff": "0,1,2,3"
28323031 },
28333032 {
3033+ "PublicDescription": "tbd Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
28343034 "EventCode": "0xB7, 0xBB",
28353035 "MSRValue": "0x043c000122 ",
28363036 "Counter": "0,1,2,3",
tools/perf/pmu-events/arch/x86/broadwell/other.json+10−10
@@ -9,16 +9,6 @@ · 8 unmodified lines
99 "BriefDescription": "Unhalted core cycles when the thread is in ring 0",
1010 "CounterHTOff": "0,1,2,3,4,5,6,7"
1111 },
12− {
13− "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
14− "EventCode": "0x5C",
15− "Counter": "0,1,2,3",
16− "UMask": "0x2",
17− "EventName": "CPL_CYCLES.RING123",
18− "SampleAfterValue": "2000003",
19− "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3",
20− "CounterHTOff": "0,1,2,3,4,5,6,7"
21− },
2212 {
2313 "PublicDescription": "This event counts when there is a transition from ring 1,2 or 3 to ring0.",
2414 "EventCode": "0x5C",
@@ -31,6 +21,16 @@ · 6 unmodified lines
3121 "CounterMask": "1",
3222 "CounterHTOff": "0,1,2,3,4,5,6,7"
3323 },
24+ {
25+ "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
26+ "EventCode": "0x5C",
27+ "Counter": "0,1,2,3",
28+ "UMask": "0x2",
29+ "EventName": "CPL_CYCLES.RING123",
30+ "SampleAfterValue": "2000003",
31+ "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3",
32+ "CounterHTOff": "0,1,2,3,4,5,6,7"
33+ },
3434 {
3535 "PublicDescription": "This event counts cycles in which the L1 and L2 are locked due to a UC lock or split lock. A lock is asserted in case of locked memory access, due to noncacheable memory, locked operation that spans two cache lines, or a page walk from the noncacheable page table. L1D and L2 locks have a very high performance penalty and it is highly recommended to avoid such access.",
3636 "EventCode": "0x63",
tools/perf/pmu-events/arch/x86/broadwell/pipeline.json+613−603
@@ -2,32 +2,42 @@ · 1 unmodified line
22 {
33 "PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. \nNotes: INST_RETIRED.ANY is counted by a designated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. INST_RETIRED.ANY_P is counted by a programmable counter and it is an architectural performance event. \nCounting: Faulting executions of GETSEC/VM entry/VM Exit/MWait will not count as retired instructions.",
44 "EventCode": "0x00",
5− "Counter": "Fixed counter 1",
5+ "Counter": "Fixed counter 0",
66 "UMask": "0x1",
77 "EventName": "INST_RETIRED.ANY",
88 "SampleAfterValue": "2000003",
99 "BriefDescription": "Instructions retired from execution.",
10− "CounterHTOff": "Fixed counter 1"
10+ "CounterHTOff": "Fixed counter 0"
1111 },
1212 {
1313 "PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
1414 "EventCode": "0x00",
15− "Counter": "Fixed counter 2",
15+ "Counter": "Fixed counter 1",
1616 "UMask": "0x2",
1717 "EventName": "CPU_CLK_UNHALTED.THREAD",
1818 "SampleAfterValue": "2000003",
1919 "BriefDescription": "Core cycles when the thread is not in halt state",
20− "CounterHTOff": "Fixed counter 2"
20+ "CounterHTOff": "Fixed counter 1"
21+ },
22+ {
23+ "EventCode": "0x00",
24+ "Counter": "Fixed counter 1",
25+ "UMask": "0x2",
26+ "AnyThread": "1",
27+ "EventName": "CPU_CLK_UNHALTED.THREAD_ANY",
28+ "SampleAfterValue": "2000003",
29+ "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
30+ "CounterHTOff": "Fixed counter 1"
2131 },
2232 {
2333 "PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. \nNote: On all current platforms this event stops counting during 'throttling (TM)' states duty off periods the processor is 'halted'. This event is clocked by base clock (100 Mhz) on Sandy Bridge. The counter update is done at a lower clock rate then the core clock the overflow status bit for this counter may appear 'sticky'. After the counter has overflowed and software clears the overflow status bit and resets the counter to less than MAX. The reset value to the counter is not clocked immediately so the overflow status bit will flip 'high (1)' and generate another PMI (if enabled) after which the reset value gets clocked into the counter. Therefore, software will get the interrupt, read the overflow status bit '1 for bit 34 while the counter value is less than MAX. Software should ignore this case.",
2434 "EventCode": "0x00",
25− "Counter": "Fixed counter 3",
35+ "Counter": "Fixed counter 2",
2636 "UMask": "0x3",
2737 "EventName": "CPU_CLK_UNHALTED.REF_TSC",
2838 "SampleAfterValue": "2000003",
2939 "BriefDescription": "Reference cycles when the core is not in halt state.",
30− "CounterHTOff": "Fixed counter 3"
40+ "CounterHTOff": "Fixed counter 2"
3141 },
3242 {
3343 "PublicDescription": "This event counts how many times the load operation got the true Block-on-Store blocking code preventing store forwarding. This includes cases when:\n - preceding store conflicts with the load (incomplete overlap);\n - store forwarding is impossible due to u-arch limitations;\n - preceding lock RMW operations are not forwarded;\n - store has the no-forward bit set (uncacheable/page-split/masked stores);\n - all-blocking stores are used (mostly, fences and port I/O);\nand others.\nThe most common case is a load blocked due to its address range overlapping with a preceding smaller uncompleted store. Note: This event does not take into account cases of out-of-SW-control (for example, SbTailHit), unknown physical STA, and cases of blocking loads on store due to being non-WB memory type or a lock. These cases are covered by other events.\nSee the table of not supported store forwards in the Optimization Guide.",
@@ -59,26 +69,37 @@ · 25 unmodified lines
5969 "CounterHTOff": "0,1,2,3,4,5,6,7"
6070 },
6171 {
62− "PublicDescription": "This event counts the number of cycles during which Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the current thread. This also includes the cycles during which the Allocator is serving another thread.",
72+ "PublicDescription": "Cycles checkpoints in Resource Allocation Table (RAT) are recovering from JEClear or machine clear.",
6373 "EventCode": "0x0D",
6474 "Counter": "0,1,2,3",
65− "UMask": "0x8",
66− "EventName": "INT_MISC.RAT_STALL_CYCLES",
75+ "UMask": "0x3",
76+ "EventName": "INT_MISC.RECOVERY_CYCLES",
6777 "SampleAfterValue": "2000003",
68− "BriefDescription": "Cycles when Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the thread",
78+ "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for this thread (e.g. misprediction or memory nuke)",
79+ "CounterMask": "1",
6980 "CounterHTOff": "0,1,2,3,4,5,6,7"
7081 },
7182 {
72− "PublicDescription": "Cycles checkpoints in Resource Allocation Table (RAT) are recovering from JEClear or machine clear.",
7383 "EventCode": "0x0D",
7484 "Counter": "0,1,2,3",
7585 "UMask": "0x3",
76− "EventName": "INT_MISC.RECOVERY_CYCLES",
86+ "AnyThread": "1",
87+ "EventName": "INT_MISC.RECOVERY_CYCLES_ANY",
7788 "SampleAfterValue": "2000003",
78− "BriefDescription": "Number of cycles waiting for the checkpoints in Resource Allocation Table (RAT) to be recovered after Nuke due to all other cases except JEClear (e.g. whenever a ucode assist is needed like SSE exception, memory disambiguation, etc...)",
89+ "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for any thread running on the physical core (e.g. misprediction or memory nuke).",
7990 "CounterMask": "1",
8091 "CounterHTOff": "0,1,2,3,4,5,6,7"
8192 },
93+ {
94+ "PublicDescription": "This event counts the number of cycles during which Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the current thread. This also includes the cycles during which the Allocator is serving another thread.",
95+ "EventCode": "0x0D",
96+ "Counter": "0,1,2,3",
97+ "UMask": "0x8",
98+ "EventName": "INT_MISC.RAT_STALL_CYCLES",
99+ "SampleAfterValue": "2000003",
100+ "BriefDescription": "Cycles when Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the thread",
101+ "CounterHTOff": "0,1,2,3,4,5,6,7"
102+ },
82103 {
83104 "PublicDescription": "This event counts the number of Uops issued by the Resource Allocation Table (RAT) to the reservation station (RS).",
84105 "EventCode": "0x0E",
@@ -89,6 +110,18 @@ · 4 unmodified lines
89110 "BriefDescription": "Uops that Resource Allocation Table (RAT) issues to Reservation Station (RS)",
90111 "CounterHTOff": "0,1,2,3,4,5,6,7"
91112 },
113+ {
114+ "PublicDescription": "This event counts cycles during which the Resource Allocation Table (RAT) does not issue any Uops to the reservation station (RS) for the current thread.",
115+ "EventCode": "0x0E",
116+ "Invert": "1",
117+ "Counter": "0,1,2,3",
118+ "UMask": "0x1",
119+ "EventName": "UOPS_ISSUED.STALL_CYCLES",
120+ "SampleAfterValue": "2000003",
121+ "BriefDescription": "Cycles when Resource Allocation Table (RAT) does not issue Uops to Reservation Station (RS) for the thread",
122+ "CounterMask": "1",
123+ "CounterHTOff": "0,1,2,3"
124+ },
92125 {
93126 "PublicDescription": "Number of flags-merge uops being allocated. Such uops considered perf sensitive\n added by GSR u-arch.",
94127 "EventCode": "0x0E",
@@ -117,18 +150,6 @@ · 22 unmodified lines
117150 "BriefDescription": "Number of Multiply packed/scalar single precision uops allocated.",
118151 "CounterHTOff": "0,1,2,3,4,5,6,7"
119152 },
120− {
121− "PublicDescription": "This event counts cycles during which the Resource Allocation Table (RAT) does not issue any Uops to the reservation station (RS) for the current thread.",
122− "EventCode": "0x0E",
123− "Invert": "1",
124− "Counter": "0,1,2,3",
125− "UMask": "0x1",
126− "EventName": "UOPS_ISSUED.STALL_CYCLES",
127− "SampleAfterValue": "2000003",
128− "BriefDescription": "Cycles when Resource Allocation Table (RAT) does not issue Uops to Reservation Station (RS) for the thread",
129− "CounterMask": "1",
130− "CounterHTOff": "0,1,2,3"
131− },
132153 {
133154 "PublicDescription": "This event counts the number of the divide operations executed. Uses edge-detect and a cmask value of 1 on ARITH.FPU_DIV_ACTIVE to get the number of the divide operations executed.",
134155 "EventCode": "0x14",
@@ -139,6 +160,26 @@ · 4 unmodified lines
139160 "BriefDescription": "Cycles when divider is busy executing divide operations",
140161 "CounterHTOff": "0,1,2,3,4,5,6,7"
141162 },
163+ {
164+ "PublicDescription": "This is an architectural event that counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling. For this reason, this event may have a changing ratio with regards to wall clock time.",
165+ "EventCode": "0x3C",
166+ "Counter": "0,1,2,3",
167+ "UMask": "0x0",
168+ "EventName": "CPU_CLK_UNHALTED.THREAD_P",
169+ "SampleAfterValue": "2000003",
170+ "BriefDescription": "Thread cycles when thread is not in halt state",
171+ "CounterHTOff": "0,1,2,3,4,5,6,7"
172+ },
173+ {
174+ "EventCode": "0x3C",
175+ "Counter": "0,1,2,3",
176+ "UMask": "0x0",
177+ "AnyThread": "1",
178+ "EventName": "CPU_CLK_UNHALTED.THREAD_P_ANY",
179+ "SampleAfterValue": "2000003",
180+ "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
181+ "CounterHTOff": "0,1,2,3,4,5,6,7"
182+ },
142183 {
143184 "PublicDescription": "This is a fixed-frequency event programmed to general counters. It counts when the core is unhalted at 100 Mhz.",
144185 "EventCode": "0x3C",
@@ -149,6 +190,36 @@ · 4 unmodified lines
149190 "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
150191 "CounterHTOff": "0,1,2,3,4,5,6,7"
151192 },
193+ {
194+ "EventCode": "0x3C",
195+ "Counter": "0,1,2,3",
196+ "UMask": "0x1",
197+ "AnyThread": "1",
198+ "EventName": "CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY",
199+ "SampleAfterValue": "2000003",
200+ "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
201+ "CounterHTOff": "0,1,2,3,4,5,6,7"
202+ },
203+ {
204+ "PublicDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate).",
205+ "EventCode": "0x3C",
206+ "Counter": "0,1,2,3",
207+ "UMask": "0x1",
208+ "EventName": "CPU_CLK_UNHALTED.REF_XCLK",
209+ "SampleAfterValue": "2000003",
210+ "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
211+ "CounterHTOff": "0,1,2,3,4,5,6,7"
212+ },
213+ {
214+ "EventCode": "0x3C",
215+ "Counter": "0,1,2,3",
216+ "UMask": "0x1",
217+ "AnyThread": "1",
218+ "EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
219+ "SampleAfterValue": "2000003",
220+ "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
221+ "CounterHTOff": "0,1,2,3,4,5,6,7"
222+ },
152223 {
153224 "EventCode": "0x3c",
154225 "Counter": "0,1,2,3",
@@ -158,6 +229,15 @@ · 3 unmodified lines
158229 "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
159230 "CounterHTOff": "0,1,2,3"
160231 },
232+ {
233+ "EventCode": "0x3C",
234+ "Counter": "0,1,2,3",
235+ "UMask": "0x2",
236+ "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE",
237+ "SampleAfterValue": "2000003",
238+ "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
239+ "CounterHTOff": "0,1,2,3,4,5,6,7"
240+ },
161241 {
162242 "PublicDescription": "This event counts all not software-prefetch load dispatches that hit the fill buffer (FB) allocated for the software prefetch. It can also be incremented by some lock instructions. So it should only be used with profiling so that the locks can be excluded by asm inspection of the nearby instructions.",
163243 "EventCode": "0x4c",
@@ -224,6 +304,18 @@ · 60 unmodified lines
224304 "BriefDescription": "Cycles when Reservation Station (RS) is empty for the thread",
225305 "CounterHTOff": "0,1,2,3,4,5,6,7"
226306 },
307+ {
308+ "EventCode": "0x5E",
309+ "Invert": "1",
310+ "Counter": "0,1,2,3",
311+ "UMask": "0x1",
312+ "EdgeDetect": "1",
313+ "EventName": "RS_EVENTS.EMPTY_END",
314+ "SampleAfterValue": "200003",
315+ "BriefDescription": "Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues.",
316+ "CounterMask": "1",
317+ "CounterHTOff": "0,1,2,3,4,5,6,7"
318+ },
227319 {
228320 "PublicDescription": "This event counts stalls occured due to changing prefix length (66, 67 or REX.W when they change the length of the decoded instruction). Occurrences counting is proportional to the number of prefixes in a 16B-line. This may result in the following penalties: three-cycle penalty for each LCP in a 16-byte chunk.",
229321 "EventCode": "0x87",
@@ -404,6 +496,15 @@ · 174 unmodified lines
404496 "BriefDescription": "Taken speculative and retired mispredicted indirect branches with return mnemonic",
405497 "CounterHTOff": "0,1,2,3,4,5,6,7"
406498 },
499+ {
500+ "EventCode": "0x89",
501+ "Counter": "0,1,2,3",
502+ "UMask": "0xa0",
503+ "EventName": "BR_MISP_EXEC.TAKEN_INDIRECT_NEAR_CALL",
504+ "SampleAfterValue": "200003",
505+ "BriefDescription": "Taken speculative and retired mispredicted indirect calls.",
506+ "CounterHTOff": "0,1,2,3,4,5,6,7"
507+ },
407508 {
408509 "PublicDescription": "This event counts both taken and not taken speculative and retired mispredicted macro conditional branch instructions.",
409510 "EventCode": "0x89",
@@ -434,6 +535,16 @@ · 24 unmodified lines
434535 "BriefDescription": "Speculative and retired mispredicted macro conditional branches",
435536 "CounterHTOff": "0,1,2,3,4,5,6,7"
436537 },
538+ {
539+ "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
540+ "EventCode": "0xA0",
541+ "Counter": "0,1,2,3",
542+ "UMask": "0x3",
543+ "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF",
544+ "SampleAfterValue": "2000003",
545+ "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports",
546+ "CounterHTOff": "0,1,2,3"
547+ },
437548 {
438549 "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
439550 "EventCode": "0xA1",
@@ -445,602 +556,472 @@ · 5 unmodified lines
445556 "CounterHTOff": "0,1,2,3,4,5,6,7"
446557 },
447558 {
448− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
449559 "EventCode": "0xA1",
450560 "Counter": "0,1,2,3",
451− "UMask": "0x2",
452− "EventName": "UOPS_DISPATCHED_PORT.PORT_1",
561+ "UMask": "0x1",
562+ "AnyThread": "1",
563+ "EventName": "UOPS_EXECUTED_PORT.PORT_0_CORE",
453564 "SampleAfterValue": "2000003",
454− "BriefDescription": "Cycles per thread when uops are executed in port 1",
565+ "BriefDescription": "Cycles per core when uops are exectuted in port 0.",
455566 "CounterHTOff": "0,1,2,3,4,5,6,7"
456567 },
457568 {
458− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
569+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
459570 "EventCode": "0xA1",
460571 "Counter": "0,1,2,3",
461− "UMask": "0x4",
462− "EventName": "UOPS_DISPATCHED_PORT.PORT_2",
572+ "UMask": "0x1",
573+ "EventName": "UOPS_EXECUTED_PORT.PORT_0",
463574 "SampleAfterValue": "2000003",
464− "BriefDescription": "Cycles per thread when uops are executed in port 2",
575+ "BriefDescription": "Cycles per thread when uops are executed in port 0",
465576 "CounterHTOff": "0,1,2,3,4,5,6,7"
466577 },
467578 {
468− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
579+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
469580 "EventCode": "0xA1",
470581 "Counter": "0,1,2,3",
471− "UMask": "0x8",
472− "EventName": "UOPS_DISPATCHED_PORT.PORT_3",
582+ "UMask": "0x2",
583+ "EventName": "UOPS_DISPATCHED_PORT.PORT_1",
473584 "SampleAfterValue": "2000003",
474− "BriefDescription": "Cycles per thread when uops are executed in port 3",
585+ "BriefDescription": "Cycles per thread when uops are executed in port 1",
475586 "CounterHTOff": "0,1,2,3,4,5,6,7"
476587 },
477588 {
478− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
479589 "EventCode": "0xA1",
480590 "Counter": "0,1,2,3",
481− "UMask": "0x10",
482− "EventName": "UOPS_DISPATCHED_PORT.PORT_4",
591+ "UMask": "0x2",
592+ "AnyThread": "1",
593+ "EventName": "UOPS_EXECUTED_PORT.PORT_1_CORE",
483594 "SampleAfterValue": "2000003",
484− "BriefDescription": "Cycles per thread when uops are executed in port 4",
595+ "BriefDescription": "Cycles per core when uops are exectuted in port 1.",
485596 "CounterHTOff": "0,1,2,3,4,5,6,7"
486597 },
487598 {
488− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
599+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
489600 "EventCode": "0xA1",
490601 "Counter": "0,1,2,3",
491− "UMask": "0x20",
492− "EventName": "UOPS_DISPATCHED_PORT.PORT_5",
602+ "UMask": "0x2",
603+ "EventName": "UOPS_EXECUTED_PORT.PORT_1",
493604 "SampleAfterValue": "2000003",
494− "BriefDescription": "Cycles per thread when uops are executed in port 5",
605+ "BriefDescription": "Cycles per thread when uops are executed in port 1",
495606 "CounterHTOff": "0,1,2,3,4,5,6,7"
496607 },
497608 {
498− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
609+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
499610 "EventCode": "0xA1",
500611 "Counter": "0,1,2,3",
501− "UMask": "0x40",
502− "EventName": "UOPS_DISPATCHED_PORT.PORT_6",
612+ "UMask": "0x4",
613+ "EventName": "UOPS_DISPATCHED_PORT.PORT_2",
503614 "SampleAfterValue": "2000003",
504− "BriefDescription": "Cycles per thread when uops are executed in port 6",
615+ "BriefDescription": "Cycles per thread when uops are executed in port 2",
505616 "CounterHTOff": "0,1,2,3,4,5,6,7"
506617 },
507618 {
508− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
509619 "EventCode": "0xA1",
510620 "Counter": "0,1,2,3",
511− "UMask": "0x80",
512− "EventName": "UOPS_DISPATCHED_PORT.PORT_7",
621+ "UMask": "0x4",
622+ "AnyThread": "1",
623+ "EventName": "UOPS_EXECUTED_PORT.PORT_2_CORE",
513624 "SampleAfterValue": "2000003",
514− "BriefDescription": "Cycles per thread when uops are executed in port 7",
625+ "BriefDescription": "Cycles per core when uops are dispatched to port 2.",
515626 "CounterHTOff": "0,1,2,3,4,5,6,7"
516627 },
517628 {
518− "PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
519− "EventCode": "0xA2",
629+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
630+ "EventCode": "0xA1",
520631 "Counter": "0,1,2,3",
521− "UMask": "0x1",
522− "EventName": "RESOURCE_STALLS.ANY",
632+ "UMask": "0x4",
633+ "EventName": "UOPS_EXECUTED_PORT.PORT_2",
523634 "SampleAfterValue": "2000003",
524− "BriefDescription": "Resource-related stall cycles",
635+ "BriefDescription": "Cycles per thread when uops are executed in port 2",
525636 "CounterHTOff": "0,1,2,3,4,5,6,7"
526637 },
527638 {
528− "PublicDescription": "This event counts stall cycles caused by absence of eligible entries in the reservation station (RS). This may result from RS overflow, or from RS deallocation because of the RS array Write Port allocation scheme (each RS entry has two write ports instead of four. As a result, empty entries could not be used, although RS is not really full). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
529− "EventCode": "0xA2",
639+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
640+ "EventCode": "0xA1",
530641 "Counter": "0,1,2,3",
531− "UMask": "0x4",
532− "EventName": "RESOURCE_STALLS.RS",
642+ "UMask": "0x8",
643+ "EventName": "UOPS_DISPATCHED_PORT.PORT_3",
533644 "SampleAfterValue": "2000003",
534− "BriefDescription": "Cycles stalled due to no eligible RS entry available.",
645+ "BriefDescription": "Cycles per thread when uops are executed in port 3",
535646 "CounterHTOff": "0,1,2,3,4,5,6,7"
536647 },
537648 {
538− "PublicDescription": "This event counts stall cycles caused by the store buffer (SB) overflow (excluding draining from synch). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
539− "EventCode": "0xA2",
649+ "EventCode": "0xA1",
540650 "Counter": "0,1,2,3",
541651 "UMask": "0x8",
542− "EventName": "RESOURCE_STALLS.SB",
652+ "AnyThread": "1",
653+ "EventName": "UOPS_EXECUTED_PORT.PORT_3_CORE",
543654 "SampleAfterValue": "2000003",
544− "BriefDescription": "Cycles stalled due to no store buffers available. (not including draining form sync).",
655+ "BriefDescription": "Cycles per core when uops are dispatched to port 3.",
545656 "CounterHTOff": "0,1,2,3,4,5,6,7"
546657 },
547658 {
548− "PublicDescription": "This event counts ROB full stall cycles. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
549− "EventCode": "0xA2",
659+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
660+ "EventCode": "0xA1",
550661 "Counter": "0,1,2,3",
551− "UMask": "0x10",
552− "EventName": "RESOURCE_STALLS.ROB",
662+ "UMask": "0x8",
663+ "EventName": "UOPS_EXECUTED_PORT.PORT_3",
553664 "SampleAfterValue": "2000003",
554− "BriefDescription": "Cycles stalled due to re-order buffer full.",
665+ "BriefDescription": "Cycles per thread when uops are executed in port 3",
555666 "CounterHTOff": "0,1,2,3,4,5,6,7"
556667 },
557668 {
558− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand* load request missing the L2 cache.",
559− "EventCode": "0xA3",
669+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
670+ "EventCode": "0xA1",
560671 "Counter": "0,1,2,3",
561− "UMask": "0x1",
562− "EventName": "CYCLE_ACTIVITY.CYCLES_L2_PENDING",
563− "SampleAfterValue": "2000003",
564− "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
565− "CounterMask": "1",
566− "CounterHTOff": "0,1,2,3,4,5,6,7"
567− },
568− {
569− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request missing the L1 data cache.",
570− "EventCode": "0xA3",
571− "Counter": "2",
572− "UMask": "0x8",
573− "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_PENDING",
672+ "UMask": "0x10",
673+ "EventName": "UOPS_DISPATCHED_PORT.PORT_4",
574674 "SampleAfterValue": "2000003",
575− "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
576− "CounterMask": "8",
577− "CounterHTOff": "2"
675+ "BriefDescription": "Cycles per thread when uops are executed in port 4",
676+ "CounterHTOff": "0,1,2,3,4,5,6,7"
578677 },
579678 {
580− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request (that is cycles with non-completed load waiting for its data from memory subsystem).",
581− "EventCode": "0xA3",
679+ "EventCode": "0xA1",
582680 "Counter": "0,1,2,3",
583− "UMask": "0x2",
584− "EventName": "CYCLE_ACTIVITY.CYCLES_LDM_PENDING",
681+ "UMask": "0x10",
682+ "AnyThread": "1",
683+ "EventName": "UOPS_EXECUTED_PORT.PORT_4_CORE",
585684 "SampleAfterValue": "2000003",
586− "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
587− "CounterMask": "2",
685+ "BriefDescription": "Cycles per core when uops are exectuted in port 4.",
588686 "CounterHTOff": "0,1,2,3,4,5,6,7"
589687 },
590688 {
591− "PublicDescription": "Counts number of cycles nothing is executed on any execution port.",
592− "EventCode": "0xA3",
689+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
690+ "EventCode": "0xA1",
593691 "Counter": "0,1,2,3",
594− "UMask": "0x4",
595− "EventName": "CYCLE_ACTIVITY.CYCLES_NO_EXECUTE",
692+ "UMask": "0x10",
693+ "EventName": "UOPS_EXECUTED_PORT.PORT_4",
596694 "SampleAfterValue": "2000003",
597− "BriefDescription": "Total execution stalls",
598− "CounterMask": "4",
599− "CounterHTOff": "0,1,2,3"
695+ "BriefDescription": "Cycles per thread when uops are executed in port 4",
696+ "CounterHTOff": "0,1,2,3,4,5,6,7"
600697 },
601698 {
602− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand* load request missing the L2 cache.(as a footprint) * includes also L1 HW prefetch requests that may or may not be required by demands.",
603− "EventCode": "0xA3",
699+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
700+ "EventCode": "0xA1",
604701 "Counter": "0,1,2,3",
605− "UMask": "0x5",
606− "EventName": "CYCLE_ACTIVITY.STALLS_L2_PENDING",
702+ "UMask": "0x20",
703+ "EventName": "UOPS_DISPATCHED_PORT.PORT_5",
607704 "SampleAfterValue": "2000003",
608− "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
609− "CounterMask": "5",
610− "CounterHTOff": "0,1,2,3"
705+ "BriefDescription": "Cycles per thread when uops are executed in port 5",
706+ "CounterHTOff": "0,1,2,3,4,5,6,7"
611707 },
612708 {
613− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request.",
614− "EventCode": "0xA3",
709+ "EventCode": "0xA1",
615710 "Counter": "0,1,2,3",
616− "UMask": "0x6",
617− "EventName": "CYCLE_ACTIVITY.STALLS_LDM_PENDING",
711+ "UMask": "0x20",
712+ "AnyThread": "1",
713+ "EventName": "UOPS_EXECUTED_PORT.PORT_5_CORE",
618714 "SampleAfterValue": "2000003",
619− "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
620− "CounterMask": "6",
621− "CounterHTOff": "0,1,2,3"
715+ "BriefDescription": "Cycles per core when uops are exectuted in port 5.",
716+ "CounterHTOff": "0,1,2,3,4,5,6,7"
622717 },
623718 {
624− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request missing the L1 data cache.",
625− "EventCode": "0xA3",
626− "Counter": "2",
627− "UMask": "0xc",
628− "EventName": "CYCLE_ACTIVITY.STALLS_L1D_PENDING",
719+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
720+ "EventCode": "0xA1",
721+ "Counter": "0,1,2,3",
722+ "UMask": "0x20",
723+ "EventName": "UOPS_EXECUTED_PORT.PORT_5",
629724 "SampleAfterValue": "2000003",
630− "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
631− "CounterMask": "12",
632− "CounterHTOff": "2"
725+ "BriefDescription": "Cycles per thread when uops are executed in port 5",
726+ "CounterHTOff": "0,1,2,3,4,5,6,7"
633727 },
634728 {
635− "PublicDescription": "Number of Uops delivered by the LSD. ",
636− "EventCode": "0xA8",
729+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
730+ "EventCode": "0xA1",
637731 "Counter": "0,1,2,3",
638− "UMask": "0x1",
639− "EventName": "LSD.UOPS",
732+ "UMask": "0x40",
733+ "EventName": "UOPS_DISPATCHED_PORT.PORT_6",
640734 "SampleAfterValue": "2000003",
641− "BriefDescription": "Number of Uops delivered by the LSD.",
735+ "BriefDescription": "Cycles per thread when uops are executed in port 6",
642736 "CounterHTOff": "0,1,2,3,4,5,6,7"
643737 },
644738 {
645− "PublicDescription": "Number of uops to be executed per-thread each cycle.",
646− "EventCode": "0xB1",
739+ "EventCode": "0xA1",
647740 "Counter": "0,1,2,3",
648− "UMask": "0x1",
649− "EventName": "UOPS_EXECUTED.THREAD",
741+ "UMask": "0x40",
742+ "AnyThread": "1",
743+ "EventName": "UOPS_EXECUTED_PORT.PORT_6_CORE",
650744 "SampleAfterValue": "2000003",
651− "BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
745+ "BriefDescription": "Cycles per core when uops are exectuted in port 6.",
652746 "CounterHTOff": "0,1,2,3,4,5,6,7"
653747 },
654748 {
655− "PublicDescription": "Number of uops executed from any thread.",
656− "EventCode": "0xB1",
749+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
750+ "EventCode": "0xA1",
657751 "Counter": "0,1,2,3",
658− "UMask": "0x2",
659− "EventName": "UOPS_EXECUTED.CORE",
752+ "UMask": "0x40",
753+ "EventName": "UOPS_EXECUTED_PORT.PORT_6",
660754 "SampleAfterValue": "2000003",
661− "BriefDescription": "Number of uops executed on the core.",
755+ "BriefDescription": "Cycles per thread when uops are executed in port 6",
662756 "CounterHTOff": "0,1,2,3,4,5,6,7"
663757 },
664758 {
665− "PublicDescription": "This event counts cycles during which no uops were dispatched from the Reservation Station (RS) per thread.",
666− "EventCode": "0xB1",
667− "Invert": "1",
759+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
760+ "EventCode": "0xA1",
668761 "Counter": "0,1,2,3",
669− "UMask": "0x1",
670− "EventName": "UOPS_EXECUTED.STALL_CYCLES",
762+ "UMask": "0x80",
763+ "EventName": "UOPS_DISPATCHED_PORT.PORT_7",
671764 "SampleAfterValue": "2000003",
672− "BriefDescription": "Counts number of cycles no uops were dispatched to be executed on this thread.",
673− "CounterMask": "1",
674− "CounterHTOff": "0,1,2,3"
765+ "BriefDescription": "Cycles per thread when uops are executed in port 7",
766+ "CounterHTOff": "0,1,2,3,4,5,6,7"
675767 },
676768 {
677− "PublicDescription": "This event counts the number of instructions (EOMs) retired. Counting covers macro-fused instructions individually (that is, increments by two).",
678− "EventCode": "0xC0",
769+ "EventCode": "0xA1",
679770 "Counter": "0,1,2,3",
680− "UMask": "0x0",
681− "Errata": "BDM61",
682− "EventName": "INST_RETIRED.ANY_P",
771+ "UMask": "0x80",
772+ "AnyThread": "1",
773+ "EventName": "UOPS_EXECUTED_PORT.PORT_7_CORE",
683774 "SampleAfterValue": "2000003",
684− "BriefDescription": "Number of instructions retired. General Counter - architectural event",
775+ "BriefDescription": "Cycles per core when uops are dispatched to port 7.",
685776 "CounterHTOff": "0,1,2,3,4,5,6,7"
686777 },
687778 {
688− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
689− "EventCode": "0xC0",
779+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
780+ "EventCode": "0xA1",
690781 "Counter": "0,1,2,3",
691− "UMask": "0x2",
692− "EventName": "INST_RETIRED.X87",
782+ "UMask": "0x80",
783+ "EventName": "UOPS_EXECUTED_PORT.PORT_7",
693784 "SampleAfterValue": "2000003",
694− "BriefDescription": "FP operations retired. X87 FP operations that have no exceptions:",
785+ "BriefDescription": "Cycles per thread when uops are executed in port 7",
695786 "CounterHTOff": "0,1,2,3,4,5,6,7"
696787 },
697788 {
698− "PEBS": "2",
699− "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts instructions retired.",
700− "EventCode": "0xC0",
701− "Counter": "1",
789+ "PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
790+ "EventCode": "0xA2",
791+ "Counter": "0,1,2,3",
702792 "UMask": "0x1",
703− "Errata": "BDM11, BDM55",
704− "EventName": "INST_RETIRED.PREC_DIST",
793+ "EventName": "RESOURCE_STALLS.ANY",
705794 "SampleAfterValue": "2000003",
706− "BriefDescription": "Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution",
707− "CounterHTOff": "1"
795+ "BriefDescription": "Resource-related stall cycles",
796+ "CounterHTOff": "0,1,2,3,4,5,6,7"
708797 },
709798 {
710− "EventCode": "0xC1",
799+ "PublicDescription": "This event counts stall cycles caused by absence of eligible entries in the reservation station (RS). This may result from RS overflow, or from RS deallocation because of the RS array Write Port allocation scheme (each RS entry has two write ports instead of four. As a result, empty entries could not be used, although RS is not really full). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
800+ "EventCode": "0xA2",
711801 "Counter": "0,1,2,3",
712− "UMask": "0x40",
713− "EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
714− "SampleAfterValue": "100003",
715− "BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
802+ "UMask": "0x4",
803+ "EventName": "RESOURCE_STALLS.RS",
804+ "SampleAfterValue": "2000003",
805+ "BriefDescription": "Cycles stalled due to no eligible RS entry available.",
716806 "CounterHTOff": "0,1,2,3,4,5,6,7"
717807 },
718808 {
719− "PEBS": "1",
720− "PublicDescription": "This event counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
721− "EventCode": "0xC2",
809+ "PublicDescription": "This event counts stall cycles caused by the store buffer (SB) overflow (excluding draining from synch). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
810+ "EventCode": "0xA2",
722811 "Counter": "0,1,2,3",
723− "UMask": "0x1",
724− "EventName": "UOPS_RETIRED.ALL",
812+ "UMask": "0x8",
813+ "EventName": "RESOURCE_STALLS.SB",
725814 "SampleAfterValue": "2000003",
726− "BriefDescription": "Actually retired uops.",
727− "CounterHTOff": "0,1,2,3,4,5,6,7",
728− "Data_LA": "1"
815+ "BriefDescription": "Cycles stalled due to no store buffers available. (not including draining form sync).",
816+ "CounterHTOff": "0,1,2,3,4,5,6,7"
729817 },
730818 {
731− "PEBS": "1",
732− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of retirement slots used.",
733− "EventCode": "0xC2",
819+ "PublicDescription": "This event counts ROB full stall cycles. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
820+ "EventCode": "0xA2",
734821 "Counter": "0,1,2,3",
735− "UMask": "0x2",
736− "EventName": "UOPS_RETIRED.RETIRE_SLOTS",
822+ "UMask": "0x10",
823+ "EventName": "RESOURCE_STALLS.ROB",
737824 "SampleAfterValue": "2000003",
738− "BriefDescription": "Retirement slots used.",
825+ "BriefDescription": "Cycles stalled due to re-order buffer full.",
739826 "CounterHTOff": "0,1,2,3,4,5,6,7"
740827 },
741828 {
742− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts cycles without actually retired uops.",
743− "EventCode": "0xC2",
744− "Invert": "1",
829+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand* load request missing the L2 cache.",
830+ "EventCode": "0xA3",
745831 "Counter": "0,1,2,3",
746832 "UMask": "0x1",
747− "EventName": "UOPS_RETIRED.STALL_CYCLES",
833+ "EventName": "CYCLE_ACTIVITY.CYCLES_L2_PENDING",
748834 "SampleAfterValue": "2000003",
749− "BriefDescription": "Cycles without actually retired uops.",
835+ "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
750836 "CounterMask": "1",
751− "CounterHTOff": "0,1,2,3"
837+ "CounterHTOff": "0,1,2,3,4,5,6,7"
752838 },
753839 {
754− "PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
755− "EventCode": "0xC2",
756− "Invert": "1",
840+ "EventCode": "0xA3",
757841 "Counter": "0,1,2,3",
758842 "UMask": "0x1",
759− "EventName": "UOPS_RETIRED.TOTAL_CYCLES",
843+ "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS",
760844 "SampleAfterValue": "2000003",
761− "BriefDescription": "Cycles with less than 10 actually retired uops.",
762− "CounterMask": "10",
763− "CounterHTOff": "0,1,2,3"
845+ "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
846+ "CounterMask": "1",
847+ "CounterHTOff": "0,1,2,3,4,5,6,7"
764848 },
765849 {
766− "PublicDescription": "This event counts both thread-specific (TS) and all-thread (AT) nukes.",
767− "EventCode": "0xC3",
850+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request (that is cycles with non-completed load waiting for its data from memory subsystem).",
851+ "EventCode": "0xA3",
768852 "Counter": "0,1,2,3",
769− "UMask": "0x1",
770− "EventName": "MACHINE_CLEARS.CYCLES",
853+ "UMask": "0x2",
854+ "EventName": "CYCLE_ACTIVITY.CYCLES_LDM_PENDING",
771855 "SampleAfterValue": "2000003",
772− "BriefDescription": "Cycles there was a Nuke. Account for both thread-specific and All Thread Nukes.",
856+ "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
857+ "CounterMask": "2",
773858 "CounterHTOff": "0,1,2,3,4,5,6,7"
774859 },
775860 {
776− "PublicDescription": "This event counts self-modifying code (SMC) detected, which causes a machine clear.",
777− "EventCode": "0xC3",
861+ "EventCode": "0xA3",
778862 "Counter": "0,1,2,3",
779− "UMask": "0x4",
780− "EventName": "MACHINE_CLEARS.SMC",
781− "SampleAfterValue": "100003",
782− "BriefDescription": "Self-modifying code (SMC) detected.",
783− "CounterHTOff": "0,1,2,3,4,5,6,7"
863+ "UMask": "0x2",
864+ "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY",
865+ "SampleAfterValue": "2000003",
866+ "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
867+ "CounterMask": "2",
868+ "CounterHTOff": "0,1,2,3"
784869 },
785870 {
786− "PublicDescription": "Maskmov false fault - counts number of time ucode passes through Maskmov flow due to instruction's mask being 0 while the flow was completed without raising a fault.",
787− "EventCode": "0xC3",
871+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port.",
872+ "EventCode": "0xA3",
788873 "Counter": "0,1,2,3",
789− "UMask": "0x20",
790− "EventName": "MACHINE_CLEARS.MASKMOV",
791− "SampleAfterValue": "100003",
792− "BriefDescription": "This event counts the number of executed Intel AVX masked load operations that refer to an illegal address range with the mask bits set to 0.",
793− "CounterHTOff": "0,1,2,3,4,5,6,7"
874+ "UMask": "0x4",
875+ "EventName": "CYCLE_ACTIVITY.CYCLES_NO_EXECUTE",
876+ "SampleAfterValue": "2000003",
877+ "BriefDescription": "This event increments by 1 for every cycle where there was no execute for this thread.",
878+ "CounterMask": "4",
879+ "CounterHTOff": "0,1,2,3"
794880 },
795881 {
796− "PEBS": "1",
797− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts conditional branch instructions retired.",
798− "EventCode": "0xC4",
799− "Counter": "0,1,2,3",
800− "UMask": "0x1",
801− "EventName": "BR_INST_RETIRED.CONDITIONAL",
802− "SampleAfterValue": "400009",
803− "BriefDescription": "Conditional branch instructions retired.",
804− "CounterHTOff": "0,1,2,3,4,5,6,7"
805− },
806− {
807− "PEBS": "1",
808− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts both direct and indirect near call instructions retired.",
809− "EventCode": "0xC4",
810− "Counter": "0,1,2,3",
811− "UMask": "0x2",
812− "EventName": "BR_INST_RETIRED.NEAR_CALL",
813− "SampleAfterValue": "100007",
814− "BriefDescription": "Direct and indirect near call instructions retired.",
815− "CounterHTOff": "0,1,2,3,4,5,6,7"
816− },
817− {
818− "PublicDescription": "This event counts all (macro) branch instructions retired.",
819− "EventCode": "0xC4",
820− "Counter": "0,1,2,3",
821− "UMask": "0x0",
822− "EventName": "BR_INST_RETIRED.ALL_BRANCHES",
823− "SampleAfterValue": "400009",
824− "BriefDescription": "All (macro) branch instructions retired.",
825− "CounterHTOff": "0,1,2,3,4,5,6,7"
826− },
827− {
828− "PEBS": "1",
829− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts return instructions retired.",
830− "EventCode": "0xC4",
831− "Counter": "0,1,2,3",
832− "UMask": "0x8",
833− "EventName": "BR_INST_RETIRED.NEAR_RETURN",
834− "SampleAfterValue": "100007",
835− "BriefDescription": "Return instructions retired.",
836− "CounterHTOff": "0,1,2,3,4,5,6,7"
837− },
838− {
839− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts not taken branch instructions retired.",
840− "EventCode": "0xC4",
841− "Counter": "0,1,2,3",
842− "UMask": "0x10",
843− "EventName": "BR_INST_RETIRED.NOT_TAKEN",
844− "SampleAfterValue": "400009",
845− "BriefDescription": "Not taken branch instructions retired.",
846− "CounterHTOff": "0,1,2,3,4,5,6,7"
847− },
848− {
849− "PEBS": "1",
850− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts taken branch instructions retired.",
851− "EventCode": "0xC4",
852− "Counter": "0,1,2,3",
853− "UMask": "0x20",
854− "EventName": "BR_INST_RETIRED.NEAR_TAKEN",
855− "SampleAfterValue": "400009",
856− "BriefDescription": "Taken branch instructions retired.",
857− "CounterHTOff": "0,1,2,3,4,5,6,7"
858− },
859− {
860− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts far branch instructions retired.",
861− "EventCode": "0xC4",
862− "Counter": "0,1,2,3",
863− "UMask": "0x40",
864− "Errata": "BDW98",
865− "EventName": "BR_INST_RETIRED.FAR_BRANCH",
866− "SampleAfterValue": "100007",
867− "BriefDescription": "Far branch instructions retired.",
868− "CounterHTOff": "0,1,2,3,4,5,6,7"
869− },
870− {
871− "PEBS": "2",
872− "PublicDescription": "This is a precise version of BR_INST_RETIRED.ALL_BRANCHES that counts all (macro) branch instructions retired.",
873− "EventCode": "0xC4",
882+ "EventCode": "0xA3",
874883 "Counter": "0,1,2,3",
875884 "UMask": "0x4",
876− "Errata": "BDW98",
877− "EventName": "BR_INST_RETIRED.ALL_BRANCHES_PEBS",
878− "SampleAfterValue": "400009",
879− "BriefDescription": "All (macro) branch instructions retired. (Precise Event - PEBS)",
880− "CounterHTOff": "0,1,2,3"
881− },
882− {
883− "PEBS": "1",
884− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted conditional branch instructions retired.",
885− "EventCode": "0xC5",
886− "Counter": "0,1,2,3",
887− "UMask": "0x1",
888− "EventName": "BR_MISP_RETIRED.CONDITIONAL",
889− "SampleAfterValue": "400009",
890− "BriefDescription": "Mispredicted conditional branch instructions retired.",
891− "CounterHTOff": "0,1,2,3,4,5,6,7"
892− },
893− {
894− "PublicDescription": "This event counts all mispredicted macro branch instructions retired.",
895− "EventCode": "0xC5",
896− "Counter": "0,1,2,3",
897− "UMask": "0x0",
898− "EventName": "BR_MISP_RETIRED.ALL_BRANCHES",
899− "SampleAfterValue": "400009",
900− "BriefDescription": "All mispredicted macro branch instructions retired.",
901− "CounterHTOff": "0,1,2,3,4,5,6,7"
902− },
903− {
904− "PEBS": "1",
905− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted return instructions retired.",
906− "EventCode": "0xC5",
907− "Counter": "0,1,2,3",
908− "UMask": "0x8",
909− "EventName": "BR_MISP_RETIRED.RET",
910− "SampleAfterValue": "100007",
911− "BriefDescription": "This event counts the number of mispredicted ret instructions retired. Non PEBS",
885+ "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL",
886+ "SampleAfterValue": "2000003",
887+ "BriefDescription": "Total execution stalls.",
888+ "CounterMask": "4",
912889 "CounterHTOff": "0,1,2,3,4,5,6,7"
913890 },
914891 {
915− "PEBS": "2",
916− "PublicDescription": "This is a precise version of BR_MISP_RETIRED.ALL_BRANCHES that counts all mispredicted macro branch instructions retired.",
917− "EventCode": "0xC5",
892+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand* load request missing the L2 cache.(as a footprint) * includes also L1 HW prefetch requests that may or may not be required by demands.",
893+ "EventCode": "0xA3",
918894 "Counter": "0,1,2,3",
919− "UMask": "0x4",
920− "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_PEBS",
921− "SampleAfterValue": "400009",
922− "BriefDescription": "Mispredicted macro branch instructions retired. (Precise Event - PEBS)",
895+ "UMask": "0x5",
896+ "EventName": "CYCLE_ACTIVITY.STALLS_L2_PENDING",
897+ "SampleAfterValue": "2000003",
898+ "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
899+ "CounterMask": "5",
923900 "CounterHTOff": "0,1,2,3"
924901 },
925902 {
926− "PublicDescription": "This event counts cases of saving new LBR records by hardware. This assumes proper enabling of LBRs and takes into account LBR filtering done by the LBR_SELECT register.",
927− "EventCode": "0xCC",
903+ "EventCode": "0xA3",
928904 "Counter": "0,1,2,3",
929− "UMask": "0x20",
930− "EventName": "ROB_MISC_EVENTS.LBR_INSERTS",
905+ "UMask": "0x5",
906+ "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS",
931907 "SampleAfterValue": "2000003",
932− "BriefDescription": "Count cases of saving new LBR",
908+ "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
909+ "CounterMask": "5",
933910 "CounterHTOff": "0,1,2,3,4,5,6,7"
934911 },
935912 {
936− "PublicDescription": "This is an architectural event that counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling. For this reason, this event may have a changing ratio with regards to wall clock time.",
937− "EventCode": "0x3C",
913+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request.",
914+ "EventCode": "0xA3",
938915 "Counter": "0,1,2,3",
939− "UMask": "0x0",
940− "EventName": "CPU_CLK_UNHALTED.THREAD_P",
916+ "UMask": "0x6",
917+ "EventName": "CYCLE_ACTIVITY.STALLS_LDM_PENDING",
941918 "SampleAfterValue": "2000003",
942− "BriefDescription": "Thread cycles when thread is not in halt state",
943− "CounterHTOff": "0,1,2,3,4,5,6,7"
919+ "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
920+ "CounterMask": "6",
921+ "CounterHTOff": "0,1,2,3"
944922 },
945923 {
946− "EventCode": "0x89",
924+ "EventCode": "0xA3",
947925 "Counter": "0,1,2,3",
948− "UMask": "0xa0",
949− "EventName": "BR_MISP_EXEC.TAKEN_INDIRECT_NEAR_CALL",
950− "SampleAfterValue": "200003",
951− "BriefDescription": "Taken speculative and retired mispredicted indirect calls.",
926+ "UMask": "0x6",
927+ "EventName": "CYCLE_ACTIVITY.STALLS_MEM_ANY",
928+ "SampleAfterValue": "2000003",
929+ "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
930+ "CounterMask": "6",
952931 "CounterHTOff": "0,1,2,3,4,5,6,7"
953932 },
954933 {
955− "EventCode": "0xA1",
956− "Counter": "0,1,2,3",
957− "UMask": "0x1",
958− "AnyThread": "1",
959− "EventName": "UOPS_EXECUTED_PORT.PORT_0_CORE",
934+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request missing the L1 data cache.",
935+ "EventCode": "0xA3",
936+ "Counter": "2",
937+ "UMask": "0x8",
938+ "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_PENDING",
960939 "SampleAfterValue": "2000003",
961− "BriefDescription": "Cycles per core when uops are exectuted in port 0.",
962− "CounterHTOff": "0,1,2,3,4,5,6,7"
940+ "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
941+ "CounterMask": "8",
942+ "CounterHTOff": "2"
963943 },
964944 {
965− "EventCode": "0xA1",
966− "Counter": "0,1,2,3",
967− "UMask": "0x2",
968− "AnyThread": "1",
969− "EventName": "UOPS_EXECUTED_PORT.PORT_1_CORE",
945+ "EventCode": "0xA3",
946+ "Counter": "2",
947+ "UMask": "0x8",
948+ "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS",
970949 "SampleAfterValue": "2000003",
971− "BriefDescription": "Cycles per core when uops are exectuted in port 1.",
972− "CounterHTOff": "0,1,2,3,4,5,6,7"
950+ "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
951+ "CounterMask": "8",
952+ "CounterHTOff": "2"
973953 },
974954 {
975− "EventCode": "0xA1",
976− "Counter": "0,1,2,3",
977− "UMask": "0x4",
978− "AnyThread": "1",
979− "EventName": "UOPS_EXECUTED_PORT.PORT_2_CORE",
955+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request missing the L1 data cache.",
956+ "EventCode": "0xA3",
957+ "Counter": "2",
958+ "UMask": "0xc",
959+ "EventName": "CYCLE_ACTIVITY.STALLS_L1D_PENDING",
980960 "SampleAfterValue": "2000003",
981− "BriefDescription": "Cycles per core when uops are dispatched to port 2.",
982− "CounterHTOff": "0,1,2,3,4,5,6,7"
961+ "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
962+ "CounterMask": "12",
963+ "CounterHTOff": "2"
983964 },
984965 {
985− "EventCode": "0xA1",
986− "Counter": "0,1,2,3",
987− "UMask": "0x8",
988− "AnyThread": "1",
989− "EventName": "UOPS_EXECUTED_PORT.PORT_3_CORE",
966+ "EventCode": "0xA3",
967+ "Counter": "2",
968+ "UMask": "0xc",
969+ "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS",
990970 "SampleAfterValue": "2000003",
991− "BriefDescription": "Cycles per core when uops are dispatched to port 3.",
992− "CounterHTOff": "0,1,2,3,4,5,6,7"
971+ "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
972+ "CounterMask": "12",
973+ "CounterHTOff": "2"
993974 },
994975 {
995− "EventCode": "0xA1",
976+ "EventCode": "0xA8",
996977 "Counter": "0,1,2,3",
997− "UMask": "0x10",
998− "AnyThread": "1",
999− "EventName": "UOPS_EXECUTED_PORT.PORT_4_CORE",
978+ "UMask": "0x1",
979+ "EventName": "LSD.UOPS",
1000980 "SampleAfterValue": "2000003",
1001− "BriefDescription": "Cycles per core when uops are exectuted in port 4.",
981+ "BriefDescription": "Number of Uops delivered by the LSD.",
1002982 "CounterHTOff": "0,1,2,3,4,5,6,7"
1003983 },
1004984 {
1005− "EventCode": "0xA1",
985+ "EventCode": "0xA8",
1006986 "Counter": "0,1,2,3",
1007− "UMask": "0x20",
1008− "AnyThread": "1",
1009− "EventName": "UOPS_EXECUTED_PORT.PORT_5_CORE",
987+ "UMask": "0x1",
988+ "EventName": "LSD.CYCLES_4_UOPS",
1010989 "SampleAfterValue": "2000003",
1011− "BriefDescription": "Cycles per core when uops are exectuted in port 5.",
990+ "BriefDescription": "Cycles 4 Uops delivered by the LSD, but didn't come from the decoder.",
991+ "CounterMask": "4",
1012992 "CounterHTOff": "0,1,2,3,4,5,6,7"
1013993 },
1014994 {
1015− "EventCode": "0xA1",
995+ "EventCode": "0xA8",
1016996 "Counter": "0,1,2,3",
1017− "UMask": "0x40",
1018− "AnyThread": "1",
1019− "EventName": "UOPS_EXECUTED_PORT.PORT_6_CORE",
997+ "UMask": "0x1",
998+ "EventName": "LSD.CYCLES_ACTIVE",
1020999 "SampleAfterValue": "2000003",
1021− "BriefDescription": "Cycles per core when uops are exectuted in port 6.",
1000+ "BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
1001+ "CounterMask": "1",
10221002 "CounterHTOff": "0,1,2,3,4,5,6,7"
10231003 },
10241004 {
1025− "EventCode": "0xA1",
1005+ "PublicDescription": "Number of uops to be executed per-thread each cycle.",
1006+ "EventCode": "0xB1",
10261007 "Counter": "0,1,2,3",
1027− "UMask": "0x80",
1028− "AnyThread": "1",
1029− "EventName": "UOPS_EXECUTED_PORT.PORT_7_CORE",
1008+ "UMask": "0x1",
1009+ "EventName": "UOPS_EXECUTED.THREAD",
10301010 "SampleAfterValue": "2000003",
1031− "BriefDescription": "Cycles per core when uops are dispatched to port 7.",
1011+ "BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
10321012 "CounterHTOff": "0,1,2,3,4,5,6,7"
10331013 },
10341014 {
1035− "PEBS": "1",
1036− "PublicDescription": "Number of near branch instructions retired that were mispredicted and taken.",
1037− "EventCode": "0xC5",
1015+ "PublicDescription": "This event counts cycles during which no uops were dispatched from the Reservation Station (RS) per thread.",
1016+ "EventCode": "0xB1",
1017+ "Invert": "1",
10381018 "Counter": "0,1,2,3",
1039− "UMask": "0x20",
1040− "EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
1041− "SampleAfterValue": "400009",
1042− "BriefDescription": "number of near branch instructions retired that were mispredicted and taken.",
1043− "CounterHTOff": "0,1,2,3,4,5,6,7"
1019+ "UMask": "0x1",
1020+ "EventName": "UOPS_EXECUTED.STALL_CYCLES",
1021+ "SampleAfterValue": "2000003",
1022+ "BriefDescription": "Counts number of cycles no uops were dispatched to be executed on this thread.",
1023+ "CounterMask": "1",
1024+ "CounterHTOff": "0,1,2,3"
10441025 },
10451026 {
10461027 "EventCode": "0xB1",
@@ -1083,335 +1064,364 @@ · 36 unmodified lines
10831064 "CounterHTOff": "0,1,2,3"
10841065 },
10851066 {
1086− "EventCode": "0xe6",
1067+ "PublicDescription": "Number of uops executed from any thread.",
1068+ "EventCode": "0xB1",
10871069 "Counter": "0,1,2,3",
1088− "UMask": "0x1f",
1089− "EventName": "BACLEARS.ANY",
1090− "SampleAfterValue": "100003",
1091− "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.",
1070+ "UMask": "0x2",
1071+ "EventName": "UOPS_EXECUTED.CORE",
1072+ "SampleAfterValue": "2000003",
1073+ "BriefDescription": "Number of uops executed on the core.",
10921074 "CounterHTOff": "0,1,2,3,4,5,6,7"
10931075 },
10941076 {
1095− "EventCode": "0xA3",
1096− "Counter": "2",
1097− "UMask": "0x8",
1098− "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS",
1077+ "EventCode": "0xb1",
1078+ "Counter": "0,1,2,3",
1079+ "UMask": "0x2",
1080+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
10991081 "SampleAfterValue": "2000003",
1100− "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
1101− "CounterMask": "8",
1102− "CounterHTOff": "2"
1082+ "BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
1083+ "CounterMask": "1",
1084+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11031085 },
11041086 {
1105− "EventCode": "0xA3",
1087+ "EventCode": "0xb1",
11061088 "Counter": "0,1,2,3",
1107− "UMask": "0x1",
1108− "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS",
1089+ "UMask": "0x2",
1090+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
11091091 "SampleAfterValue": "2000003",
1110− "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
1111− "CounterMask": "1",
1092+ "BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
1093+ "CounterMask": "2",
11121094 "CounterHTOff": "0,1,2,3,4,5,6,7"
11131095 },
11141096 {
1115− "EventCode": "0xA3",
1097+ "EventCode": "0xb1",
11161098 "Counter": "0,1,2,3",
11171099 "UMask": "0x2",
1118− "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY",
1100+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3",
11191101 "SampleAfterValue": "2000003",
1120− "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
1121− "CounterMask": "2",
1122− "CounterHTOff": "0,1,2,3"
1102+ "BriefDescription": "Cycles at least 3 micro-op is executed from any thread on physical core.",
1103+ "CounterMask": "3",
1104+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11231105 },
11241106 {
1125− "EventCode": "0xA3",
1107+ "EventCode": "0xb1",
11261108 "Counter": "0,1,2,3",
1127− "UMask": "0x4",
1128− "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL",
1109+ "UMask": "0x2",
1110+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
11291111 "SampleAfterValue": "2000003",
1130− "BriefDescription": "Total execution stalls.",
1112+ "BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
11311113 "CounterMask": "4",
11321114 "CounterHTOff": "0,1,2,3,4,5,6,7"
11331115 },
11341116 {
1135− "EventCode": "0xA3",
1136− "Counter": "2",
1137− "UMask": "0xc",
1138− "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS",
1117+ "EventCode": "0xb1",
1118+ "Invert": "1",
1119+ "Counter": "0,1,2,3",
1120+ "UMask": "0x2",
1121+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
11391122 "SampleAfterValue": "2000003",
1140− "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
1141− "CounterMask": "12",
1142− "CounterHTOff": "2"
1123+ "BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
1124+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11431125 },
11441126 {
1145− "EventCode": "0xA3",
1127+ "PublicDescription": "This event counts the number of instructions (EOMs) retired. Counting covers macro-fused instructions individually (that is, increments by two).",
1128+ "EventCode": "0xC0",
11461129 "Counter": "0,1,2,3",
1147− "UMask": "0x5",
1148− "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS",
1130+ "UMask": "0x0",
1131+ "Errata": "BDM61",
1132+ "EventName": "INST_RETIRED.ANY_P",
11491133 "SampleAfterValue": "2000003",
1150− "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
1151− "CounterMask": "5",
1134+ "BriefDescription": "Number of instructions retired. General Counter - architectural event",
11521135 "CounterHTOff": "0,1,2,3,4,5,6,7"
11531136 },
11541137 {
1155− "EventCode": "0xA3",
1138+ "PEBS": "2",
1139+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts instructions retired.",
1140+ "EventCode": "0xC0",
1141+ "Counter": "1",
1142+ "UMask": "0x1",
1143+ "Errata": "BDM11, BDM55",
1144+ "EventName": "INST_RETIRED.PREC_DIST",
1145+ "SampleAfterValue": "2000003",
1146+ "BriefDescription": "Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution",
1147+ "CounterHTOff": "1"
1148+ },
1149+ {
1150+ "PublicDescription": "This event counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
1151+ "EventCode": "0xC0",
11561152 "Counter": "0,1,2,3",
1157− "UMask": "0x6",
1158− "EventName": "CYCLE_ACTIVITY.STALLS_MEM_ANY",
1153+ "UMask": "0x2",
1154+ "EventName": "INST_RETIRED.X87",
11591155 "SampleAfterValue": "2000003",
1160− "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
1161− "CounterMask": "6",
1156+ "BriefDescription": "FP operations retired. X87 FP operations that have no exceptions:",
11621157 "CounterHTOff": "0,1,2,3,4,5,6,7"
11631158 },
11641159 {
1165− "EventCode": "0xC3",
1160+ "EventCode": "0xC1",
11661161 "Counter": "0,1,2,3",
1167− "UMask": "0x1",
1168− "EdgeDetect": "1",
1169− "EventName": "MACHINE_CLEARS.COUNT",
1162+ "UMask": "0x40",
1163+ "EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
11701164 "SampleAfterValue": "100003",
1171− "BriefDescription": "Number of machine clears (nukes) of any type.",
1172− "CounterMask": "1",
1165+ "BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
11731166 "CounterHTOff": "0,1,2,3,4,5,6,7"
11741167 },
11751168 {
1176− "EventCode": "0xA8",
1169+ "PEBS": "1",
1170+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
1171+ "EventCode": "0xC2",
11771172 "Counter": "0,1,2,3",
11781173 "UMask": "0x1",
1179− "EventName": "LSD.CYCLES_4_UOPS",
1174+ "EventName": "UOPS_RETIRED.ALL",
11801175 "SampleAfterValue": "2000003",
1181− "BriefDescription": "Cycles 4 Uops delivered by the LSD, but didn't come from the decoder.",
1182− "CounterMask": "4",
1183− "CounterHTOff": "0,1,2,3,4,5,6,7"
1176+ "BriefDescription": "Actually retired uops. (Precise Event - PEBS)",
1177+ "CounterHTOff": "0,1,2,3,4,5,6,7",
1178+ "Data_LA": "1"
11841179 },
11851180 {
1186− "EventCode": "0x5E",
1181+ "PublicDescription": "This event counts cycles without actually retired uops.",
1182+ "EventCode": "0xC2",
11871183 "Invert": "1",
11881184 "Counter": "0,1,2,3",
11891185 "UMask": "0x1",
1190− "EdgeDetect": "1",
1191− "EventName": "RS_EVENTS.EMPTY_END",
1192− "SampleAfterValue": "200003",
1193− "BriefDescription": "Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues.",
1194− "CounterMask": "1",
1195− "CounterHTOff": "0,1,2,3,4,5,6,7"
1196− },
1197− {
1198− "EventCode": "0xA8",
1199− "Counter": "0,1,2,3",
1200− "UMask": "0x1",
1201− "EventName": "LSD.CYCLES_ACTIVE",
1186+ "EventName": "UOPS_RETIRED.STALL_CYCLES",
12021187 "SampleAfterValue": "2000003",
1203− "BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
1188+ "BriefDescription": "Cycles without actually retired uops.",
12041189 "CounterMask": "1",
1205− "CounterHTOff": "0,1,2,3,4,5,6,7"
1190+ "CounterHTOff": "0,1,2,3"
12061191 },
12071192 {
1208− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
1209− "EventCode": "0xA1",
1193+ "PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
1194+ "EventCode": "0xC2",
1195+ "Invert": "1",
12101196 "Counter": "0,1,2,3",
12111197 "UMask": "0x1",
1212− "EventName": "UOPS_EXECUTED_PORT.PORT_0",
1198+ "EventName": "UOPS_RETIRED.TOTAL_CYCLES",
12131199 "SampleAfterValue": "2000003",
1214− "BriefDescription": "Cycles per thread when uops are executed in port 0",
1215− "CounterHTOff": "0,1,2,3,4,5,6,7"
1200+ "BriefDescription": "Cycles with less than 10 actually retired uops.",
1201+ "CounterMask": "10",
1202+ "CounterHTOff": "0,1,2,3"
12161203 },
12171204 {
1218− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
1219− "EventCode": "0xA1",
1205+ "PEBS": "1",
1206+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of retirement slots used.",
1207+ "EventCode": "0xC2",
12201208 "Counter": "0,1,2,3",
12211209 "UMask": "0x2",
1222− "EventName": "UOPS_EXECUTED_PORT.PORT_1",
1210+ "EventName": "UOPS_RETIRED.RETIRE_SLOTS",
12231211 "SampleAfterValue": "2000003",
1224− "BriefDescription": "Cycles per thread when uops are executed in port 1",
1212+ "BriefDescription": "Retirement slots used. (Precise Event - PEBS)",
12251213 "CounterHTOff": "0,1,2,3,4,5,6,7"
12261214 },
12271215 {
1228− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
1229− "EventCode": "0xA1",
1216+ "PublicDescription": "This event counts both thread-specific (TS) and all-thread (AT) nukes.",
1217+ "EventCode": "0xC3",
12301218 "Counter": "0,1,2,3",
1231− "UMask": "0x4",
1232− "EventName": "UOPS_EXECUTED_PORT.PORT_2",
1219+ "UMask": "0x1",
1220+ "EventName": "MACHINE_CLEARS.CYCLES",
12331221 "SampleAfterValue": "2000003",
1234− "BriefDescription": "Cycles per thread when uops are executed in port 2",
1222+ "BriefDescription": "Cycles there was a Nuke. Account for both thread-specific and All Thread Nukes.",
12351223 "CounterHTOff": "0,1,2,3,4,5,6,7"
12361224 },
12371225 {
1238− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
1239− "EventCode": "0xA1",
1226+ "EventCode": "0xC3",
12401227 "Counter": "0,1,2,3",
1241− "UMask": "0x8",
1242− "EventName": "UOPS_EXECUTED_PORT.PORT_3",
1243− "SampleAfterValue": "2000003",
1244− "BriefDescription": "Cycles per thread when uops are executed in port 3",
1228+ "UMask": "0x1",
1229+ "EdgeDetect": "1",
1230+ "EventName": "MACHINE_CLEARS.COUNT",
1231+ "SampleAfterValue": "100003",
1232+ "BriefDescription": "Number of machine clears (nukes) of any type.",
1233+ "CounterMask": "1",
12451234 "CounterHTOff": "0,1,2,3,4,5,6,7"
12461235 },
12471236 {
1248− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
1249− "EventCode": "0xA1",
1237+ "PublicDescription": "This event counts self-modifying code (SMC) detected, which causes a machine clear.",
1238+ "EventCode": "0xC3",
12501239 "Counter": "0,1,2,3",
1251− "UMask": "0x10",
1252− "EventName": "UOPS_EXECUTED_PORT.PORT_4",
1253− "SampleAfterValue": "2000003",
1254− "BriefDescription": "Cycles per thread when uops are executed in port 4",
1240+ "UMask": "0x4",
1241+ "EventName": "MACHINE_CLEARS.SMC",
1242+ "SampleAfterValue": "100003",
1243+ "BriefDescription": "Self-modifying code (SMC) detected.",
12551244 "CounterHTOff": "0,1,2,3,4,5,6,7"
12561245 },
12571246 {
1258− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
1259− "EventCode": "0xA1",
1247+ "PublicDescription": "Maskmov false fault - counts number of time ucode passes through Maskmov flow due to instruction's mask being 0 while the flow was completed without raising a fault.",
1248+ "EventCode": "0xC3",
12601249 "Counter": "0,1,2,3",
12611250 "UMask": "0x20",
1262− "EventName": "UOPS_EXECUTED_PORT.PORT_5",
1263− "SampleAfterValue": "2000003",
1264− "BriefDescription": "Cycles per thread when uops are executed in port 5",
1251+ "EventName": "MACHINE_CLEARS.MASKMOV",
1252+ "SampleAfterValue": "100003",
1253+ "BriefDescription": "This event counts the number of executed Intel AVX masked load operations that refer to an illegal address range with the mask bits set to 0.",
12651254 "CounterHTOff": "0,1,2,3,4,5,6,7"
12661255 },
12671256 {
1268− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
1269− "EventCode": "0xA1",
1257+ "PublicDescription": "This event counts all (macro) branch instructions retired.",
1258+ "EventCode": "0xC4",
12701259 "Counter": "0,1,2,3",
1271− "UMask": "0x40",
1272− "EventName": "UOPS_EXECUTED_PORT.PORT_6",
1273− "SampleAfterValue": "2000003",
1274− "BriefDescription": "Cycles per thread when uops are executed in port 6",
1260+ "UMask": "0x0",
1261+ "EventName": "BR_INST_RETIRED.ALL_BRANCHES",
1262+ "SampleAfterValue": "400009",
1263+ "BriefDescription": "All (macro) branch instructions retired.",
12751264 "CounterHTOff": "0,1,2,3,4,5,6,7"
12761265 },
12771266 {
1278− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
1279− "EventCode": "0xA1",
1267+ "PEBS": "1",
1268+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts conditional branch instructions retired.",
1269+ "EventCode": "0xC4",
12801270 "Counter": "0,1,2,3",
1281− "UMask": "0x80",
1282− "EventName": "UOPS_EXECUTED_PORT.PORT_7",
1283− "SampleAfterValue": "2000003",
1284− "BriefDescription": "Cycles per thread when uops are executed in port 7",
1271+ "UMask": "0x1",
1272+ "EventName": "BR_INST_RETIRED.CONDITIONAL",
1273+ "SampleAfterValue": "400009",
1274+ "BriefDescription": "Conditional branch instructions retired. (Precise Event - PEBS)",
12851275 "CounterHTOff": "0,1,2,3,4,5,6,7"
12861276 },
12871277 {
1288− "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
1289− "EventCode": "0xA0",
1278+ "PEBS": "1",
1279+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect near call instructions retired.",
1280+ "EventCode": "0xC4",
12901281 "Counter": "0,1,2,3",
1291− "UMask": "0x3",
1292− "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF",
1293− "SampleAfterValue": "2000003",
1294− "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports",
1295− "CounterHTOff": "0,1,2,3"
1282+ "UMask": "0x2",
1283+ "EventName": "BR_INST_RETIRED.NEAR_CALL",
1284+ "SampleAfterValue": "100007",
1285+ "BriefDescription": "Direct and indirect near call instructions retired. (Precise Event - PEBS)",
1286+ "CounterHTOff": "0,1,2,3,4,5,6,7"
12961287 },
12971288 {
1298− "EventCode": "0x00",
1299− "Counter": "Fixed counter 2",
1289+ "PEBS": "1",
1290+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect macro near call instructions retired (captured in ring 3).",
1291+ "EventCode": "0xC4",
1292+ "Counter": "0,1,2,3",
13001293 "UMask": "0x2",
1301− "AnyThread": "1",
1302− "EventName": "CPU_CLK_UNHALTED.THREAD_ANY",
1303− "SampleAfterValue": "2000003",
1304− "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
1305− "CounterHTOff": "Fixed counter 2"
1294+ "EventName": "BR_INST_RETIRED.NEAR_CALL_R3",
1295+ "SampleAfterValue": "100007",
1296+ "BriefDescription": "Direct and indirect macro near call instructions retired (captured in ring 3). (Precise Event - PEBS)",
1297+ "CounterHTOff": "0,1,2,3,4,5,6,7"
13061298 },
13071299 {
1308− "EventCode": "0x3C",
1300+ "PEBS": "2",
1301+ "PublicDescription": "This is a precise version of BR_INST_RETIRED.ALL_BRANCHES that counts all (macro) branch instructions retired.",
1302+ "EventCode": "0xC4",
13091303 "Counter": "0,1,2,3",
1310− "UMask": "0x0",
1311− "AnyThread": "1",
1312− "EventName": "CPU_CLK_UNHALTED.THREAD_P_ANY",
1313− "SampleAfterValue": "2000003",
1314− "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
1315− "CounterHTOff": "0,1,2,3,4,5,6,7"
1304+ "UMask": "0x4",
1305+ "Errata": "BDW98",
1306+ "EventName": "BR_INST_RETIRED.ALL_BRANCHES_PEBS",
1307+ "SampleAfterValue": "400009",
1308+ "BriefDescription": "All (macro) branch instructions retired. (Precise Event - PEBS)",
1309+ "CounterHTOff": "0,1,2,3"
13161310 },
13171311 {
1318− "EventCode": "0x3C",
1312+ "PEBS": "1",
1313+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts return instructions retired.",
1314+ "EventCode": "0xC4",
13191315 "Counter": "0,1,2,3",
1320− "UMask": "0x1",
1321− "AnyThread": "1",
1322− "EventName": "CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY",
1323− "SampleAfterValue": "2000003",
1324− "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
1316+ "UMask": "0x8",
1317+ "EventName": "BR_INST_RETIRED.NEAR_RETURN",
1318+ "SampleAfterValue": "100007",
1319+ "BriefDescription": "Return instructions retired. (Precise Event - PEBS)",
13251320 "CounterHTOff": "0,1,2,3,4,5,6,7"
13261321 },
13271322 {
1328− "EventCode": "0x0D",
1323+ "PublicDescription": "This event counts not taken branch instructions retired.",
1324+ "EventCode": "0xC4",
13291325 "Counter": "0,1,2,3",
1330− "UMask": "0x3",
1331− "AnyThread": "1",
1332− "EventName": "INT_MISC.RECOVERY_CYCLES_ANY",
1333− "SampleAfterValue": "2000003",
1334− "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for any thread running on the physical core (e.g. misprediction or memory nuke).",
1335− "CounterMask": "1",
1326+ "UMask": "0x10",
1327+ "EventName": "BR_INST_RETIRED.NOT_TAKEN",
1328+ "SampleAfterValue": "400009",
1329+ "BriefDescription": "Not taken branch instructions retired.",
13361330 "CounterHTOff": "0,1,2,3,4,5,6,7"
13371331 },
13381332 {
1339− "EventCode": "0xb1",
1333+ "PEBS": "1",
1334+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts taken branch instructions retired.",
1335+ "EventCode": "0xC4",
13401336 "Counter": "0,1,2,3",
1341− "UMask": "0x2",
1342− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
1343− "SampleAfterValue": "2000003",
1344− "BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
1345− "CounterMask": "1",
1337+ "UMask": "0x20",
1338+ "EventName": "BR_INST_RETIRED.NEAR_TAKEN",
1339+ "SampleAfterValue": "400009",
1340+ "BriefDescription": "Taken branch instructions retired. (Precise Event - PEBS)",
13461341 "CounterHTOff": "0,1,2,3,4,5,6,7"
13471342 },
13481343 {
1349− "EventCode": "0xb1",
1344+ "PublicDescription": "This event counts far branch instructions retired.",
1345+ "EventCode": "0xC4",
13501346 "Counter": "0,1,2,3",
1351− "UMask": "0x2",
1352− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
1353− "SampleAfterValue": "2000003",
1354− "BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
1355− "CounterMask": "2",
1347+ "UMask": "0x40",
1348+ "Errata": "BDW98",
1349+ "EventName": "BR_INST_RETIRED.FAR_BRANCH",
1350+ "SampleAfterValue": "100007",
1351+ "BriefDescription": "Far branch instructions retired.",
13561352 "CounterHTOff": "0,1,2,3,4,5,6,7"
13571353 },
13581354 {
1359− "EventCode": "0xb1",
1355+ "PublicDescription": "This event counts all mispredicted macro branch instructions retired.",
1356+ "EventCode": "0xC5",
13601357 "Counter": "0,1,2,3",
1361− "UMask": "0x2",
1362− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3",
1363− "SampleAfterValue": "2000003",
1364− "BriefDescription": "Cycles at least 3 micro-op is executed from any thread on physical core.",
1365− "CounterMask": "3",
1358+ "UMask": "0x0",
1359+ "EventName": "BR_MISP_RETIRED.ALL_BRANCHES",
1360+ "SampleAfterValue": "400009",
1361+ "BriefDescription": "All mispredicted macro branch instructions retired.",
13661362 "CounterHTOff": "0,1,2,3,4,5,6,7"
13671363 },
13681364 {
1369− "EventCode": "0xb1",
1365+ "PEBS": "1",
1366+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted conditional branch instructions retired.",
1367+ "EventCode": "0xC5",
13701368 "Counter": "0,1,2,3",
1371− "UMask": "0x2",
1372− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
1373− "SampleAfterValue": "2000003",
1374− "BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
1375− "CounterMask": "4",
1369+ "UMask": "0x1",
1370+ "EventName": "BR_MISP_RETIRED.CONDITIONAL",
1371+ "SampleAfterValue": "400009",
1372+ "BriefDescription": "Mispredicted conditional branch instructions retired. (Precise Event - PEBS)",
13761373 "CounterHTOff": "0,1,2,3,4,5,6,7"
13771374 },
13781375 {
1379− "EventCode": "0xb1",
1380− "Invert": "1",
1376+ "PEBS": "2",
1377+ "PublicDescription": "This is a precise version of BR_MISP_RETIRED.ALL_BRANCHES that counts all mispredicted macro branch instructions retired.",
1378+ "EventCode": "0xC5",
13811379 "Counter": "0,1,2,3",
1382− "UMask": "0x2",
1383− "EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
1384− "SampleAfterValue": "2000003",
1385− "BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
1380+ "UMask": "0x4",
1381+ "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_PEBS",
1382+ "SampleAfterValue": "400009",
1383+ "BriefDescription": "Mispredicted macro branch instructions retired. (Precise Event - PEBS)",
1384+ "CounterHTOff": "0,1,2,3"
1385+ },
1386+ {
1387+ "PEBS": "1",
1388+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted return instructions retired.",
1389+ "EventCode": "0xC5",
1390+ "Counter": "0,1,2,3",
1391+ "UMask": "0x8",
1392+ "EventName": "BR_MISP_RETIRED.RET",
1393+ "SampleAfterValue": "100007",
1394+ "BriefDescription": "This event counts the number of mispredicted ret instructions retired.(Precise Event)",
13861395 "CounterHTOff": "0,1,2,3,4,5,6,7"
13871396 },
13881397 {
1389− "PublicDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate).",
1390− "EventCode": "0x3C",
1398+ "PEBS": "1",
1399+ "PublicDescription": "Number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
1400+ "EventCode": "0xC5",
13911401 "Counter": "0,1,2,3",
1392− "UMask": "0x1",
1393− "EventName": "CPU_CLK_UNHALTED.REF_XCLK",
1394− "SampleAfterValue": "2000003",
1395− "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
1402+ "UMask": "0x20",
1403+ "EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
1404+ "SampleAfterValue": "400009",
1405+ "BriefDescription": "number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
13961406 "CounterHTOff": "0,1,2,3,4,5,6,7"
13971407 },
13981408 {
1399− "EventCode": "0x3C",
1409+ "PublicDescription": "This event counts cases of saving new LBR records by hardware. This assumes proper enabling of LBRs and takes into account LBR filtering done by the LBR_SELECT register.",
1410+ "EventCode": "0xCC",
14001411 "Counter": "0,1,2,3",
1401− "UMask": "0x1",
1402− "AnyThread": "1",
1403− "EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
1412+ "UMask": "0x20",
1413+ "EventName": "ROB_MISC_EVENTS.LBR_INSERTS",
14041414 "SampleAfterValue": "2000003",
1405− "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
1415+ "BriefDescription": "Count cases of saving new LBR",
14061416 "CounterHTOff": "0,1,2,3,4,5,6,7"
14071417 },
14081418 {
1409− "EventCode": "0x3C",
1419+ "EventCode": "0xe6",
14101420 "Counter": "0,1,2,3",
1411− "UMask": "0x2",
1412− "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE",
1413− "SampleAfterValue": "2000003",
1414− "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
1421+ "UMask": "0x1f",
1422+ "EventName": "BACLEARS.ANY",
1423+ "SampleAfterValue": "100003",
1424+ "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.",
14151425 "CounterHTOff": "0,1,2,3,4,5,6,7"
14161426 }
14171427 ]
tools/perf/pmu-events/arch/x86/broadwell/virtual-memory.json+75−75
@@ -43,6 +43,16 @@ · 42 unmodified lines
4343 "BriefDescription": "Load miss in all TLB levels causes a page walk that completes. (1G)",
4444 "CounterHTOff": "0,1,2,3,4,5,6,7"
4545 },
46+ {
47+ "EventCode": "0x08",
48+ "Counter": "0,1,2,3",
49+ "UMask": "0xe",
50+ "Errata": "BDM69",
51+ "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED",
52+ "SampleAfterValue": "100003",
53+ "BriefDescription": "Demand load Miss in all translation lookaside buffer (TLB) levels causes a page walk that completes of any page size.",
54+ "CounterHTOff": "0,1,2,3,4,5,6,7"
55+ },
4656 {
4757 "PublicDescription": "This event counts the number of cycles while PMH is busy with the page walk.",
4858 "EventCode": "0x08",
@@ -72,6 +82,15 @@ · 23 unmodified lines
7282 "BriefDescription": "Load misses that miss the DTLB and hit the STLB (2M).",
7383 "CounterHTOff": "0,1,2,3,4,5,6,7"
7484 },
85+ {
86+ "EventCode": "0x08",
87+ "Counter": "0,1,2,3",
88+ "UMask": "0x60",
89+ "EventName": "DTLB_LOAD_MISSES.STLB_HIT",
90+ "SampleAfterValue": "2000003",
91+ "BriefDescription": "Load operations that miss the first DTLB level but hit the second and do not cause page walks.",
92+ "CounterHTOff": "0,1,2,3,4,5,6,7"
93+ },
7594 {
7695 "PublicDescription": "This event counts store misses in all DTLB levels that cause page walks of any page size (4K/2M/4M/1G).",
7796 "EventCode": "0x49",
@@ -116,6 +135,16 @@ · 38 unmodified lines
116135 "BriefDescription": "Store misses in all DTLB levels that cause completed page walks (1G)",
117136 "CounterHTOff": "0,1,2,3,4,5,6,7"
118137 },
138+ {
139+ "EventCode": "0x49",
140+ "Counter": "0,1,2,3",
141+ "UMask": "0xe",
142+ "Errata": "BDM69",
143+ "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED",
144+ "SampleAfterValue": "100003",
145+ "BriefDescription": "Store misses in all DTLB levels that cause completed page walks.",
146+ "CounterHTOff": "0,1,2,3,4,5,6,7"
147+ },
119148 {
120149 "PublicDescription": "This event counts the number of cycles while PMH is busy with the page walk.",
121150 "EventCode": "0x49",
@@ -145,6 +174,15 @@ · 23 unmodified lines
145174 "BriefDescription": "Store misses that miss the DTLB and hit the STLB (2M).",
146175 "CounterHTOff": "0,1,2,3,4,5,6,7"
147176 },
177+ {
178+ "EventCode": "0x49",
179+ "Counter": "0,1,2,3",
180+ "UMask": "0x60",
181+ "EventName": "DTLB_STORE_MISSES.STLB_HIT",
182+ "SampleAfterValue": "100003",
183+ "BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
184+ "CounterHTOff": "0,1,2,3,4,5,6,7"
185+ },
148186 {
149187 "PublicDescription": "This event counts cycles for an extended page table walk. The Extended Page directory cache differs from standard TLB caches by the operating system that use it. Virtual machine operating systems use the extended page directory cache, while guest operating systems use the standard TLB caches.",
150188 "EventCode": "0x4F",
@@ -199,6 +237,16 @@ · 48 unmodified lines
199237 "BriefDescription": "Store miss in all TLB levels causes a page walk that completes. (1G)",
200238 "CounterHTOff": "0,1,2,3,4,5,6,7"
201239 },
240+ {
241+ "EventCode": "0x85",
242+ "Counter": "0,1,2,3",
243+ "UMask": "0xe",
244+ "Errata": "BDM69",
245+ "EventName": "ITLB_MISSES.WALK_COMPLETED",
246+ "SampleAfterValue": "100003",
247+ "BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
248+ "CounterHTOff": "0,1,2,3,4,5,6,7"
249+ },
202250 {
203251 "PublicDescription": "This event counts the number of cycles while PMH is busy with the page walk.",
204252 "EventCode": "0x85",
@@ -228,6 +276,15 @@ · 23 unmodified lines
228276 "BriefDescription": "Code misses that miss the DTLB and hit the STLB (2M).",
229277 "CounterHTOff": "0,1,2,3,4,5,6,7"
230278 },
279+ {
280+ "EventCode": "0x85",
281+ "Counter": "0,1,2,3",
282+ "UMask": "0x60",
283+ "EventName": "ITLB_MISSES.STLB_HIT",
284+ "SampleAfterValue": "100003",
285+ "BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
286+ "CounterHTOff": "0,1,2,3,4,5,6,7"
287+ },
231288 {
232289 "PublicDescription": "This event counts the number of flushes of the big or small ITLB pages. Counting include both TLB Flush (covering all sets) and TLB Set Clear (set-specific).",
233290 "EventCode": "0xAE",
@@ -251,61 +308,61 @@ · 17 unmodified lines
251308 {
252309 "EventCode": "0xBC",
253310 "Counter": "0,1,2,3",
254− "UMask": "0x21",
311+ "UMask": "0x12",
255312 "Errata": "BDM69, BDM98",
256− "EventName": "PAGE_WALKER_LOADS.ITLB_L1",
313+ "EventName": "PAGE_WALKER_LOADS.DTLB_L2",
257314 "SampleAfterValue": "2000003",
258− "BriefDescription": "Number of ITLB page walker hits in the L1+FB.",
315+ "BriefDescription": "Number of DTLB page walker hits in the L2.",
259316 "CounterHTOff": "0,1,2,3"
260317 },
261318 {
262319 "EventCode": "0xBC",
263320 "Counter": "0,1,2,3",
264− "UMask": "0x12",
321+ "UMask": "0x14",
265322 "Errata": "BDM69, BDM98",
266− "EventName": "PAGE_WALKER_LOADS.DTLB_L2",
323+ "EventName": "PAGE_WALKER_LOADS.DTLB_L3",
267324 "SampleAfterValue": "2000003",
268− "BriefDescription": "Number of DTLB page walker hits in the L2.",
325+ "BriefDescription": "Number of DTLB page walker hits in the L3 + XSNP.",
269326 "CounterHTOff": "0,1,2,3"
270327 },
271328 {
272329 "EventCode": "0xBC",
273330 "Counter": "0,1,2,3",
274− "UMask": "0x22",
331+ "UMask": "0x18",
275332 "Errata": "BDM69, BDM98",
276− "EventName": "PAGE_WALKER_LOADS.ITLB_L2",
333+ "EventName": "PAGE_WALKER_LOADS.DTLB_MEMORY",
277334 "SampleAfterValue": "2000003",
278− "BriefDescription": "Number of ITLB page walker hits in the L2.",
335+ "BriefDescription": "Number of DTLB page walker hits in Memory.",
279336 "CounterHTOff": "0,1,2,3"
280337 },
281338 {
282339 "EventCode": "0xBC",
283340 "Counter": "0,1,2,3",
284− "UMask": "0x14",
341+ "UMask": "0x21",
285342 "Errata": "BDM69, BDM98",
286− "EventName": "PAGE_WALKER_LOADS.DTLB_L3",
343+ "EventName": "PAGE_WALKER_LOADS.ITLB_L1",
287344 "SampleAfterValue": "2000003",
288− "BriefDescription": "Number of DTLB page walker hits in the L3 + XSNP.",
345+ "BriefDescription": "Number of ITLB page walker hits in the L1+FB.",
289346 "CounterHTOff": "0,1,2,3"
290347 },
291348 {
292349 "EventCode": "0xBC",
293350 "Counter": "0,1,2,3",
294− "UMask": "0x24",
351+ "UMask": "0x22",
295352 "Errata": "BDM69, BDM98",
296− "EventName": "PAGE_WALKER_LOADS.ITLB_L3",
353+ "EventName": "PAGE_WALKER_LOADS.ITLB_L2",
297354 "SampleAfterValue": "2000003",
298− "BriefDescription": "Number of ITLB page walker hits in the L3 + XSNP.",
355+ "BriefDescription": "Number of ITLB page walker hits in the L2.",
299356 "CounterHTOff": "0,1,2,3"
300357 },
301358 {
302359 "EventCode": "0xBC",
303360 "Counter": "0,1,2,3",
304− "UMask": "0x18",
361+ "UMask": "0x24",
305362 "Errata": "BDM69, BDM98",
306− "EventName": "PAGE_WALKER_LOADS.DTLB_MEMORY",
363+ "EventName": "PAGE_WALKER_LOADS.ITLB_L3",
307364 "SampleAfterValue": "2000003",
308− "BriefDescription": "Number of DTLB page walker hits in Memory.",
365+ "BriefDescription": "Number of ITLB page walker hits in the L3 + XSNP.",
309366 "CounterHTOff": "0,1,2,3"
310367 },
311368 {
@@ -327,62 +384,5 @@ · 15 unmodified lines
327384 "SampleAfterValue": "100007",
328385 "BriefDescription": "STLB flush attempts",
329386 "CounterHTOff": "0,1,2,3,4,5,6,7"
330− },
331− {
332− "EventCode": "0x08",
333− "Counter": "0,1,2,3",
334− "UMask": "0xe",
335− "Errata": "BDM69",
336− "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED",
337− "SampleAfterValue": "100003",
338− "BriefDescription": "Demand load Miss in all translation lookaside buffer (TLB) levels causes a page walk that completes of any page size.",
339− "CounterHTOff": "0,1,2,3,4,5,6,7"
340− },
341− {
342− "EventCode": "0x08",
343− "Counter": "0,1,2,3",
344− "UMask": "0x60",
345− "EventName": "DTLB_LOAD_MISSES.STLB_HIT",
346− "SampleAfterValue": "2000003",
347− "BriefDescription": "Load operations that miss the first DTLB level but hit the second and do not cause page walks.",
348− "CounterHTOff": "0,1,2,3,4,5,6,7"
349− },
350− {
351− "EventCode": "0x49",
352− "Counter": "0,1,2,3",
353− "UMask": "0xe",
354− "Errata": "BDM69",
355− "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED",
356− "SampleAfterValue": "100003",
357− "BriefDescription": "Store misses in all DTLB levels that cause completed page walks.",
358− "CounterHTOff": "0,1,2,3,4,5,6,7"
359− },
360− {
361− "EventCode": "0x49",
362− "Counter": "0,1,2,3",
363− "UMask": "0x60",
364− "EventName": "DTLB_STORE_MISSES.STLB_HIT",
365− "SampleAfterValue": "100003",
366− "BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
367− "CounterHTOff": "0,1,2,3,4,5,6,7"
368− },
369− {
370− "EventCode": "0x85",
371− "Counter": "0,1,2,3",
372− "UMask": "0xe",
373− "Errata": "BDM69",
374− "EventName": "ITLB_MISSES.WALK_COMPLETED",
375− "SampleAfterValue": "100003",
376− "BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
377− "CounterHTOff": "0,1,2,3,4,5,6,7"
378− },
379− {
380− "EventCode": "0x85",
381− "Counter": "0,1,2,3",
382− "UMask": "0x60",
383− "EventName": "ITLB_MISSES.STLB_HIT",
384− "SampleAfterValue": "100003",
385− "BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
386− "CounterHTOff": "0,1,2,3,4,5,6,7"
387387 }
388388 ]
tools/perf/pmu-events/arch/x86/broadwellde/cache.json+212−177
@@ -11,11 +11,28 @@ · 10 unmodified lines
1111 },
1212 {
1313 "EventCode": "0x24",
14− "UMask": "0x41",
15− "BriefDescription": "Demand Data Read requests that hit L2 cache",
14+ "UMask": "0x22",
15+ "BriefDescription": "RFO requests that miss L2 cache.",
1616 "Counter": "0,1,2,3",
17− "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
18− "PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
17+ "EventName": "L2_RQSTS.RFO_MISS",
18+ "SampleAfterValue": "200003",
19+ "CounterHTOff": "0,1,2,3,4,5,6,7"
20+ },
21+ {
22+ "EventCode": "0x24",
23+ "UMask": "0x24",
24+ "BriefDescription": "L2 cache misses when fetching instructions.",
25+ "Counter": "0,1,2,3",
26+ "EventName": "L2_RQSTS.CODE_RD_MISS",
27+ "SampleAfterValue": "200003",
28+ "CounterHTOff": "0,1,2,3,4,5,6,7"
29+ },
30+ {
31+ "EventCode": "0x24",
32+ "UMask": "0x27",
33+ "BriefDescription": "Demand requests that miss L2 cache.",
34+ "Counter": "0,1,2,3",
35+ "EventName": "L2_RQSTS.ALL_DEMAND_MISS",
1936 "SampleAfterValue": "200003",
2037 "CounterHTOff": "0,1,2,3,4,5,6,7"
2138 },
@@ -29,6 +46,43 @@ · 7 unmodified lines
2946 "SampleAfterValue": "200003",
3047 "CounterHTOff": "0,1,2,3,4,5,6,7"
3148 },
49+ {
50+ "EventCode": "0x24",
51+ "UMask": "0x3f",
52+ "BriefDescription": "All requests that miss L2 cache.",
53+ "Counter": "0,1,2,3",
54+ "EventName": "L2_RQSTS.MISS",
55+ "SampleAfterValue": "200003",
56+ "CounterHTOff": "0,1,2,3,4,5,6,7"
57+ },
58+ {
59+ "EventCode": "0x24",
60+ "UMask": "0x41",
61+ "BriefDescription": "Demand Data Read requests that hit L2 cache",
62+ "Counter": "0,1,2,3",
63+ "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
64+ "PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
65+ "SampleAfterValue": "200003",
66+ "CounterHTOff": "0,1,2,3,4,5,6,7"
67+ },
68+ {
69+ "EventCode": "0x24",
70+ "UMask": "0x42",
71+ "BriefDescription": "RFO requests that hit L2 cache.",
72+ "Counter": "0,1,2,3",
73+ "EventName": "L2_RQSTS.RFO_HIT",
74+ "SampleAfterValue": "200003",
75+ "CounterHTOff": "0,1,2,3,4,5,6,7"
76+ },
77+ {
78+ "EventCode": "0x24",
79+ "UMask": "0x44",
80+ "BriefDescription": "L2 cache hits when fetching instructions, code reads.",
81+ "Counter": "0,1,2,3",
82+ "EventName": "L2_RQSTS.CODE_RD_HIT",
83+ "SampleAfterValue": "200003",
84+ "CounterHTOff": "0,1,2,3,4,5,6,7"
85+ },
3286 {
3387 "EventCode": "0x24",
3488 "UMask": "0x50",
@@ -69,6 +123,15 @@ · 34 unmodified lines
69123 "SampleAfterValue": "200003",
70124 "CounterHTOff": "0,1,2,3,4,5,6,7"
71125 },
126+ {
127+ "EventCode": "0x24",
128+ "UMask": "0xe7",
129+ "BriefDescription": "Demand requests to L2 cache.",
130+ "Counter": "0,1,2,3",
131+ "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES",
132+ "SampleAfterValue": "200003",
133+ "CounterHTOff": "0,1,2,3,4,5,6,7"
134+ },
72135 {
73136 "EventCode": "0x24",
74137 "UMask": "0xf8",
@@ -79,6 +142,15 @@ · 4 unmodified lines
79142 "SampleAfterValue": "200003",
80143 "CounterHTOff": "0,1,2,3,4,5,6,7"
81144 },
145+ {
146+ "EventCode": "0x24",
147+ "UMask": "0xff",
148+ "BriefDescription": "All L2 requests.",
149+ "Counter": "0,1,2,3",
150+ "EventName": "L2_RQSTS.REFERENCES",
151+ "SampleAfterValue": "200003",
152+ "CounterHTOff": "0,1,2,3,4,5,6,7"
153+ },
82154 {
83155 "EventCode": "0x27",
84156 "UMask": "0x50",
@@ -130,6 +202,27 @@ · 45 unmodified lines
130202 "SampleAfterValue": "2000003",
131203 "CounterHTOff": "2"
132204 },
205+ {
206+ "EventCode": "0x48",
207+ "UMask": "0x1",
208+ "BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
209+ "Counter": "2",
210+ "EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
211+ "AnyThread": "1",
212+ "CounterMask": "1",
213+ "SampleAfterValue": "2000003",
214+ "CounterHTOff": "2"
215+ },
216+ {
217+ "EventCode": "0x48",
218+ "UMask": "0x2",
219+ "BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
220+ "Counter": "0,1,2,3",
221+ "EventName": "L1D_PEND_MISS.FB_FULL",
222+ "CounterMask": "1",
223+ "SampleAfterValue": "2000003",
224+ "CounterHTOff": "0,1,2,3,4,5,6,7"
225+ },
133226 {
134227 "EventCode": "0x51",
135228 "UMask": "0x1",
@@ -151,6 +244,29 @@ · 15 unmodified lines
151244 "SampleAfterValue": "2000003",
152245 "CounterHTOff": "0,1,2,3,4,5,6,7"
153246 },
247+ {
248+ "EventCode": "0x60",
249+ "UMask": "0x1",
250+ "BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore",
251+ "Counter": "0,1,2,3",
252+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
253+ "CounterMask": "1",
254+ "Errata": "BDM76",
255+ "PublicDescription": "This event counts cycles when offcore outstanding Demand Data Read transactions are present in the super queue (SQ). A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation).",
256+ "SampleAfterValue": "2000003",
257+ "CounterHTOff": "0,1,2,3,4,5,6,7"
258+ },
259+ {
260+ "EventCode": "0x60",
261+ "UMask": "0x1",
262+ "BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
263+ "Counter": "0,1,2,3",
264+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_GE_6",
265+ "CounterMask": "6",
266+ "Errata": "BDM76",
267+ "SampleAfterValue": "2000003",
268+ "CounterHTOff": "0,1,2,3,4,5,6,7"
269+ },
154270 {
155271 "EventCode": "0x60",
156272 "UMask": "0x2",
@@ -158,7 +274,7 @@ · 1 unmodified line
158274 "Counter": "0,1,2,3",
159275 "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD",
160276 "Errata": "BDM76",
161− "PublicDescription": "This event counts the number of offcore outstanding Code Reads transactions in the super queue every cycle. The \"Offcore outstanding\" state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
277+ "PublicDescription": "This event counts the number of offcore outstanding Code Reads transactions in the super queue every cycle. The Offcore outstanding state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
162278 "SampleAfterValue": "2000003",
163279 "CounterHTOff": "0,1,2,3,4,5,6,7"
164280 },
@@ -175,24 +291,24 @@ · 10 unmodified lines
175291 },
176292 {
177293 "EventCode": "0x60",
178− "UMask": "0x8",
179− "BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore",
294+ "UMask": "0x4",
295+ "BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle",
180296 "Counter": "0,1,2,3",
181− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
297+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
298+ "CounterMask": "1",
182299 "Errata": "BDM76",
183− "PublicDescription": "This event counts the number of offcore outstanding cacheable Core Data Read transactions in the super queue every cycle. A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation). See corresponding Umask under OFFCORE_REQUESTS.",
300+ "PublicDescription": "This event counts the number of offcore outstanding demand rfo Reads transactions in the super queue every cycle. The Offcore outstanding state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
184301 "SampleAfterValue": "2000003",
185302 "CounterHTOff": "0,1,2,3,4,5,6,7"
186303 },
187304 {
188305 "EventCode": "0x60",
189− "UMask": "0x1",
190− "BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore",
306+ "UMask": "0x8",
307+ "BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore",
191308 "Counter": "0,1,2,3",
192− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
193− "CounterMask": "1",
309+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
194310 "Errata": "BDM76",
195− "PublicDescription": "This event counts cycles when offcore outstanding Demand Data Read transactions are present in the super queue (SQ). A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation).",
311+ "PublicDescription": "This event counts the number of offcore outstanding cacheable Core Data Read transactions in the super queue every cycle. A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation). See corresponding Umask under OFFCORE_REQUESTS.",
196312 "SampleAfterValue": "2000003",
197313 "CounterHTOff": "0,1,2,3,4,5,6,7"
198314 },
@@ -208,18 +324,6 @@ · 9 unmodified lines
208324 "SampleAfterValue": "2000003",
209325 "CounterHTOff": "0,1,2,3,4,5,6,7"
210326 },
211− {
212− "EventCode": "0x60",
213− "UMask": "0x4",
214− "BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle",
215− "Counter": "0,1,2,3",
216− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
217− "CounterMask": "1",
218− "Errata": "BDM76",
219− "PublicDescription": "This event counts the number of offcore outstanding demand rfo Reads transactions in the super queue every cycle. The \"Offcore outstanding\" state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
220− "SampleAfterValue": "2000003",
221− "CounterHTOff": "0,1,2,3,4,5,6,7"
222− },
223327 {
224328 "EventCode": "0x63",
225329 "UMask": "0x2",
@@ -266,7 +370,7 @@ · 40 unmodified lines
266370 "BriefDescription": "Demand and prefetch data reads",
267371 "Counter": "0,1,2,3",
268372 "EventName": "OFFCORE_REQUESTS.ALL_DATA_RD",
269− "PublicDescription": "This event counts the demand and prefetch data reads. All Core Data Reads include cacheable \"Demands\" and L2 prefetchers (not L3 prefetchers). Counting also covers reads due to page walks resulted from any request type.",
373+ "PublicDescription": "This event counts the demand and prefetch data reads. All Core Data Reads include cacheable Demands and L2 prefetchers (not L3 prefetchers). Counting also covers reads due to page walks resulted from any request type.",
270374 "SampleAfterValue": "100003",
271375 "CounterHTOff": "0,1,2,3,4,5,6,7"
272376 },
@@ -280,27 +384,36 @@ · 7 unmodified lines
280384 "SampleAfterValue": "2000003",
281385 "CounterHTOff": "0,1,2,3,4,5,6,7"
282386 },
387+ {
388+ "EventCode": "0xB7, 0xBB",
389+ "UMask": "0x1",
390+ "BriefDescription": "Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
391+ "Counter": "0,1,2,3",
392+ "EventName": "OFFCORE_RESPONSE",
393+ "SampleAfterValue": "100003",
394+ "CounterHTOff": "0,1,2,3"
395+ },
283396 {
284397 "EventCode": "0xD0",
285398 "UMask": "0x11",
286− "BriefDescription": "Retired load uops that miss the STLB.",
399+ "BriefDescription": "Retired load uops that miss the STLB. (Precise Event - PEBS)",
287400 "Data_LA": "1",
288401 "PEBS": "1",
289402 "Counter": "0,1,2,3",
290403 "EventName": "MEM_UOPS_RETIRED.STLB_MISS_LOADS",
291− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
404+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
292405 "SampleAfterValue": "100003",
293406 "CounterHTOff": "0,1,2,3"
294407 },
295408 {
296409 "EventCode": "0xD0",
297410 "UMask": "0x12",
298− "BriefDescription": "Retired store uops that miss the STLB.",
411+ "BriefDescription": "Retired store uops that miss the STLB. (Precise Event - PEBS)",
299412 "Data_LA": "1",
300413 "PEBS": "1",
301414 "Counter": "0,1,2,3",
302415 "EventName": "MEM_UOPS_RETIRED.STLB_MISS_STORES",
303− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts store uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
416+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
304417 "SampleAfterValue": "100003",
305418 "L1_Hit_Indication": "1",
306419 "CounterHTOff": "0,1,2,3"
@@ -308,37 +421,37 @@ · 1 unmodified line
308421 {
309422 "EventCode": "0xD0",
310423 "UMask": "0x21",
311− "BriefDescription": "Retired load uops with locked access.",
424+ "BriefDescription": "Retired load uops with locked access. (Precise Event - PEBS)",
312425 "Data_LA": "1",
313426 "PEBS": "1",
314427 "Counter": "0,1,2,3",
315428 "EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
316429 "Errata": "BDM35",
317− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts load uops with locked access retired to the architected path.",
430+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with locked access retired to the architected path.",
318431 "SampleAfterValue": "100007",
319432 "CounterHTOff": "0,1,2,3"
320433 },
321434 {
322435 "EventCode": "0xD0",
323436 "UMask": "0x41",
324− "BriefDescription": "Retired load uops that split across a cacheline boundary.",
437+ "BriefDescription": "Retired load uops that split across a cacheline boundary.(Precise Event - PEBS)",
325438 "Data_LA": "1",
326439 "PEBS": "1",
327440 "Counter": "0,1,2,3",
328441 "EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
329− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
442+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
330443 "SampleAfterValue": "100003",
331444 "CounterHTOff": "0,1,2,3"
332445 },
333446 {
334447 "EventCode": "0xD0",
335448 "UMask": "0x42",
336− "BriefDescription": "Retired store uops that split across a cacheline boundary.",
449+ "BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
337450 "Data_LA": "1",
338451 "PEBS": "1",
339452 "Counter": "0,1,2,3",
340453 "EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
341− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
454+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
342455 "SampleAfterValue": "100003",
343456 "L1_Hit_Indication": "1",
344457 "CounterHTOff": "0,1,2,3"
@@ -346,24 +459,24 @@ · 1 unmodified line
346459 {
347460 "EventCode": "0xD0",
348461 "UMask": "0x81",
349− "BriefDescription": "All retired load uops.",
462+ "BriefDescription": "All retired load uops. (Precise Event - PEBS)",
350463 "Data_LA": "1",
351464 "PEBS": "1",
352465 "Counter": "0,1,2,3",
353466 "EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
354− "PublicDescription": "This event counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
467+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
355468 "SampleAfterValue": "2000003",
356469 "CounterHTOff": "0,1,2,3"
357470 },
358471 {
359472 "EventCode": "0xD0",
360473 "UMask": "0x82",
361− "BriefDescription": "All retired store uops.",
474+ "BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
362475 "Data_LA": "1",
363476 "PEBS": "1",
364477 "Counter": "0,1,2,3",
365478 "EventName": "MEM_UOPS_RETIRED.ALL_STORES",
366− "PublicDescription": "This event counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
479+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
367480 "SampleAfterValue": "2000003",
368481 "L1_Hit_Indication": "1",
369482 "CounterHTOff": "0,1,2,3"
@@ -371,69 +484,69 @@ · 1 unmodified line
371484 {
372485 "EventCode": "0xD1",
373486 "UMask": "0x1",
374− "BriefDescription": "Retired load uops with L1 cache hits as data sources.",
487+ "BriefDescription": "Retired load uops with L1 cache hits as data sources. (Precise Event - PEBS)",
375488 "Data_LA": "1",
376489 "PEBS": "1",
377490 "Counter": "0,1,2,3",
378491 "EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
379− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
492+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data source were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
380493 "SampleAfterValue": "2000003",
381494 "CounterHTOff": "0,1,2,3"
382495 },
383496 {
384497 "EventCode": "0xD1",
385498 "UMask": "0x2",
386− "BriefDescription": "Retired load uops with L2 cache hits as data sources.",
499+ "BriefDescription": "Retired load uops with L2 cache hits as data sources. (Precise Event - PEBS)",
387500 "Data_LA": "1",
388501 "PEBS": "1",
389502 "Counter": "0,1,2,3",
390503 "EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
391504 "Errata": "BDM35",
392− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
505+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
393506 "SampleAfterValue": "100003",
394507 "CounterHTOff": "0,1,2,3"
395508 },
396509 {
397510 "EventCode": "0xD1",
398511 "UMask": "0x4",
399− "BriefDescription": "Retired load uops which data sources were data hits in L3 without snoops required.",
512+ "BriefDescription": "Hit in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS)",
400513 "Data_LA": "1",
401514 "PEBS": "1",
402515 "Counter": "0,1,2,3",
403516 "EventName": "MEM_LOAD_UOPS_RETIRED.L3_HIT",
404517 "Errata": "BDM100",
405− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
518+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
406519 "SampleAfterValue": "50021",
407520 "CounterHTOff": "0,1,2,3"
408521 },
409522 {
410523 "EventCode": "0xD1",
411524 "UMask": "0x8",
412− "BriefDescription": "Retired load uops misses in L1 cache as data sources.",
525+ "BriefDescription": "Retired load uops misses in L1 cache as data sources. Uses PEBS.",
413526 "Data_LA": "1",
414527 "PEBS": "1",
415528 "Counter": "0,1,2,3",
416529 "EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
417− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
530+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
418531 "SampleAfterValue": "100003",
419532 "CounterHTOff": "0,1,2,3"
420533 },
421534 {
422535 "EventCode": "0xD1",
423536 "UMask": "0x10",
424− "BriefDescription": "Miss in mid-level (L2) cache. Excludes Unknown data-source.",
537+ "BriefDescription": "Retired load uops with L2 cache misses as data sources. Uses PEBS.",
425538 "Data_LA": "1",
426539 "PEBS": "1",
427540 "Counter": "0,1,2,3",
428541 "EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
429− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
542+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
430543 "SampleAfterValue": "50021",
431544 "CounterHTOff": "0,1,2,3"
432545 },
433546 {
434547 "EventCode": "0xD1",
435548 "UMask": "0x20",
436− "BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source.",
549+ "BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS).",
437550 "Data_LA": "1",
438551 "PEBS": "1",
439552 "Counter": "0,1,2,3",
@@ -445,77 +558,112 @@ · 5 unmodified lines
445558 {
446559 "EventCode": "0xD1",
447560 "UMask": "0x40",
448− "BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready.",
561+ "BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready. (Precise Event - PEBS)",
449562 "Data_LA": "1",
450563 "PEBS": "1",
451564 "Counter": "0,1,2,3",
452565 "EventName": "MEM_LOAD_UOPS_RETIRED.HIT_LFB",
453− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
566+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
454567 "SampleAfterValue": "100003",
455568 "CounterHTOff": "0,1,2,3"
456569 },
457570 {
458571 "EventCode": "0xD2",
459572 "UMask": "0x1",
460− "BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache.",
573+ "BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache. (Precise Event - PEBS)",
461574 "Data_LA": "1",
462575 "PEBS": "1",
463576 "Counter": "0,1,2,3",
464577 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS",
465578 "Errata": "BDM100",
466− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
579+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
467580 "SampleAfterValue": "20011",
468581 "CounterHTOff": "0,1,2,3"
469582 },
470583 {
471584 "EventCode": "0xD2",
472585 "UMask": "0x2",
473− "BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache.",
586+ "BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache. (Precise Event - PEBS)",
474587 "Data_LA": "1",
475588 "PEBS": "1",
476589 "Counter": "0,1,2,3",
477590 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT",
478591 "Errata": "BDM100",
479− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
592+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
480593 "SampleAfterValue": "20011",
481594 "CounterHTOff": "0,1,2,3"
482595 },
483596 {
484597 "EventCode": "0xD2",
485598 "UMask": "0x4",
486− "BriefDescription": "Retired load uops which data sources were HitM responses from shared L3.",
599+ "BriefDescription": "Retired load uops which data sources were HitM responses from shared L3. (Precise Event - PEBS)",
487600 "Data_LA": "1",
488601 "PEBS": "1",
489602 "Counter": "0,1,2,3",
490603 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM",
491604 "Errata": "BDM100",
492− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
605+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
493606 "SampleAfterValue": "20011",
494607 "CounterHTOff": "0,1,2,3"
495608 },
496609 {
497610 "EventCode": "0xD2",
498611 "UMask": "0x8",
499− "BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required.",
612+ "BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required. (Precise Event - PEBS)",
500613 "Data_LA": "1",
501614 "PEBS": "1",
502615 "Counter": "0,1,2,3",
503616 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_NONE",
504617 "Errata": "BDM100",
505− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
618+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
506619 "SampleAfterValue": "100003",
507620 "CounterHTOff": "0,1,2,3"
508621 },
509622 {
510623 "EventCode": "0xD3",
511624 "UMask": "0x1",
512− "BriefDescription": "Data from local DRAM either Snoop not needed or Snoop Miss (RspI)",
513625 "Data_LA": "1",
514626 "PEBS": "1",
515627 "Counter": "0,1,2,3",
516628 "EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM",
517629 "Errata": "BDE70, BDM100",
518− "PublicDescription": "Retired load uop whose Data Source was: local DRAM either Snoop not needed or Snoop Miss (RspI).",
630+ "PublicDescription": "This event counts retired load uops where the data came from local DRAM. This does not include hardware prefetches. This is a precise event.",
631+ "SampleAfterValue": "100007",
632+ "CounterHTOff": "0,1,2,3"
633+ },
634+ {
635+ "EventCode": "0xD3",
636+ "UMask": "0x4",
637+ "BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI) (Precise Event)",
638+ "Data_LA": "1",
639+ "PEBS": "1",
640+ "Counter": "0,1,2,3",
641+ "EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_DRAM",
642+ "Errata": "BDE70",
643+ "SampleAfterValue": "100007",
644+ "CounterHTOff": "0,1,2,3"
645+ },
646+ {
647+ "EventCode": "0xD3",
648+ "UMask": "0x10",
649+ "BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM (Precise Event)",
650+ "Data_LA": "1",
651+ "PEBS": "1",
652+ "Counter": "0,1,2,3",
653+ "EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_HITM",
654+ "Errata": "BDE70",
655+ "SampleAfterValue": "100007",
656+ "CounterHTOff": "0,1,2,3"
657+ },
658+ {
659+ "EventCode": "0xD3",
660+ "UMask": "0x20",
661+ "BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache (Precise Event)",
662+ "Data_LA": "1",
663+ "PEBS": "1",
664+ "Counter": "0,1,2,3",
665+ "EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.REMOTE_FWD",
666+ "Errata": "BDE70",
519667 "SampleAfterValue": "100007",
520668 "CounterHTOff": "0,1,2,3"
521669 },
@@ -657,118 +805,5 @@ · 135 unmodified lines
657805 "PublicDescription": "This event counts the number of split locks in the super queue.",
658806 "SampleAfterValue": "100003",
659807 "CounterHTOff": "0,1,2,3,4,5,6,7"
660− },
661− {
662− "EventCode": "0x24",
663− "UMask": "0x42",
664− "BriefDescription": "RFO requests that hit L2 cache.",
665− "Counter": "0,1,2,3",
666− "EventName": "L2_RQSTS.RFO_HIT",
667− "SampleAfterValue": "200003",
668− "CounterHTOff": "0,1,2,3,4,5,6,7"
669− },
670− {
671− "EventCode": "0x24",
672− "UMask": "0x22",
673− "BriefDescription": "RFO requests that miss L2 cache.",
674− "Counter": "0,1,2,3",
675− "EventName": "L2_RQSTS.RFO_MISS",
676− "SampleAfterValue": "200003",
677− "CounterHTOff": "0,1,2,3,4,5,6,7"
678− },
679− {
680− "EventCode": "0x24",
681− "UMask": "0x44",
682− "BriefDescription": "L2 cache hits when fetching instructions, code reads.",
683− "Counter": "0,1,2,3",
684− "EventName": "L2_RQSTS.CODE_RD_HIT",
685− "SampleAfterValue": "200003",
686− "CounterHTOff": "0,1,2,3,4,5,6,7"
687− },
688− {
689− "EventCode": "0x24",
690− "UMask": "0x24",
691− "BriefDescription": "L2 cache misses when fetching instructions.",
692− "Counter": "0,1,2,3",
693− "EventName": "L2_RQSTS.CODE_RD_MISS",
694− "SampleAfterValue": "200003",
695− "CounterHTOff": "0,1,2,3,4,5,6,7"
696− },
697− {
698− "EventCode": "0x24",
699− "UMask": "0x27",
700− "BriefDescription": "Demand requests that miss L2 cache.",
701− "Counter": "0,1,2,3",
702− "EventName": "L2_RQSTS.ALL_DEMAND_MISS",
703− "SampleAfterValue": "200003",
704− "CounterHTOff": "0,1,2,3,4,5,6,7"
705− },
706− {
707− "EventCode": "0x24",
708− "UMask": "0xe7",
709− "BriefDescription": "Demand requests to L2 cache.",
710− "Counter": "0,1,2,3",
711− "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES",
712− "SampleAfterValue": "200003",
713− "CounterHTOff": "0,1,2,3,4,5,6,7"
714− },
715− {
716− "EventCode": "0x24",
717− "UMask": "0x3f",
718− "BriefDescription": "All requests that miss L2 cache.",
719− "Counter": "0,1,2,3",
720− "EventName": "L2_RQSTS.MISS",
721− "SampleAfterValue": "200003",
722− "CounterHTOff": "0,1,2,3,4,5,6,7"
723− },
724− {
725− "EventCode": "0x24",
726− "UMask": "0xff",
727− "BriefDescription": "All L2 requests.",
728− "Counter": "0,1,2,3",
729− "EventName": "L2_RQSTS.REFERENCES",
730− "SampleAfterValue": "200003",
731− "CounterHTOff": "0,1,2,3,4,5,6,7"
732− },
733− {
734− "EventCode": "0xB7, 0xBB",
735− "UMask": "0x1",
736− "BriefDescription": "Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
737− "Counter": "0,1,2,3",
738− "EventName": "OFFCORE_RESPONSE",
739− "SampleAfterValue": "100003",
740− "CounterHTOff": "0,1,2,3"
741− },
742− {
743− "EventCode": "0x60",
744− "UMask": "0x1",
745− "BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
746− "Counter": "0,1,2,3",
747− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_GE_6",
748− "CounterMask": "6",
749− "Errata": "BDM76",
750− "SampleAfterValue": "2000003",
751− "CounterHTOff": "0,1,2,3,4,5,6,7"
752− },
753− {
754− "EventCode": "0x48",
755− "UMask": "0x1",
756− "BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
757− "Counter": "2",
758− "EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
759− "AnyThread": "1",
760− "CounterMask": "1",
761− "SampleAfterValue": "2000003",
762− "CounterHTOff": "2"
763− },
764− {
765− "EventCode": "0x48",
766− "UMask": "0x2",
767− "BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
768− "Counter": "0,1,2,3",
769− "EventName": "L1D_PEND_MISS.FB_FULL",
770− "CounterMask": "1",
771− "SampleAfterValue": "2000003",
772− "CounterHTOff": "0,1,2,3,4,5,6,7"
773808 }
774809 ]
tools/perf/pmu-events/arch/x86/broadwellde/floating-point.json+51−57
@@ -6,7 +6,7 @@ · 5 unmodified lines
66 "Counter": "0,1,2,3",
77 "EventName": "OTHER_ASSISTS.AVX_TO_SSE",
88 "Errata": "BDM30",
9− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
9+ "PublicDescription": "This event counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
1010 "SampleAfterValue": "100003",
1111 "CounterHTOff": "0,1,2,3,4,5,6,7"
1212 },
@@ -17,7 +17,7 @@ · 4 unmodified lines
1717 "Counter": "0,1,2,3",
1818 "EventName": "OTHER_ASSISTS.SSE_TO_AVX",
1919 "Errata": "BDM30",
20− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
20+ "PublicDescription": "This event counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
2121 "SampleAfterValue": "100003",
2222 "CounterHTOff": "0,1,2,3,4,5,6,7"
2323 },
@@ -25,7 +25,6 @@ · 1 unmodified line
2525 "EventCode": "0xC7",
2626 "UMask": "0x1",
2727 "BriefDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired. Each count represents 1 computation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
28− "PEBS": "1",
2928 "Counter": "0,1,2,3",
3029 "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE",
3130 "SampleAfterValue": "2000003",
@@ -35,17 +34,24 @@ · 3 unmodified lines
3534 "EventCode": "0xC7",
3635 "UMask": "0x2",
3736 "BriefDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired. Each count represents 1 computation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
38− "PEBS": "1",
3937 "Counter": "0,1,2,3",
4038 "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE",
4139 "SampleAfterValue": "2000003",
4240 "CounterHTOff": "0,1,2,3"
4341 },
42+ {
43+ "EventCode": "0xC7",
44+ "UMask": "0x3",
45+ "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
46+ "Counter": "0,1,2,3",
47+ "EventName": "FP_ARITH_INST_RETIRED.SCALAR",
48+ "SampleAfterValue": "2000003",
49+ "CounterHTOff": "0,1,2,3"
50+ },
4451 {
4552 "EventCode": "0xC7",
4653 "UMask": "0x4",
4754 "BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired. Each count represents 2 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
48− "PEBS": "1",
4955 "Counter": "0,1,2,3",
5056 "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE",
5157 "SampleAfterValue": "2000003",
@@ -55,7 +61,6 @@ · 3 unmodified lines
5561 "EventCode": "0xC7",
5662 "UMask": "0x8",
5763 "BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
58− "PEBS": "1",
5964 "Counter": "0,1,2,3",
6065 "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE",
6166 "SampleAfterValue": "2000003",
@@ -65,19 +70,54 @@ · 3 unmodified lines
6570 "EventCode": "0xC7",
6671 "UMask": "0x10",
6772 "BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
68− "PEBS": "1",
6973 "Counter": "0,1,2,3",
7074 "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE",
7175 "SampleAfterValue": "2000003",
7276 "CounterHTOff": "0,1,2,3"
7377 },
78+ {
79+ "EventCode": "0xC7",
80+ "UMask": "0x15",
81+ "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
82+ "Counter": "0,1,2,3",
83+ "EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
84+ "SampleAfterValue": "2000006",
85+ "CounterHTOff": "0,1,2,3"
86+ },
87+ {
88+ "EventCode": "0xc7",
89+ "UMask": "0x20",
90+ "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
91+ "Counter": "0,1,2,3",
92+ "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
93+ "SampleAfterValue": "2000003",
94+ "CounterHTOff": "0,1,2,3"
95+ },
96+ {
97+ "EventCode": "0xC7",
98+ "UMask": "0x2a",
99+ "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
100+ "Counter": "0,1,2,3",
101+ "EventName": "FP_ARITH_INST_RETIRED.SINGLE",
102+ "SampleAfterValue": "2000005",
103+ "CounterHTOff": "0,1,2,3"
104+ },
105+ {
106+ "EventCode": "0xC7",
107+ "UMask": "0x3c",
108+ "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
109+ "Counter": "0,1,2,3",
110+ "EventName": "FP_ARITH_INST_RETIRED.PACKED",
111+ "SampleAfterValue": "2000004",
112+ "CounterHTOff": "0,1,2,3"
113+ },
74114 {
75115 "EventCode": "0xCA",
76116 "UMask": "0x2",
77117 "BriefDescription": "Number of X87 assists due to output value.",
78118 "Counter": "0,1,2,3",
79119 "EventName": "FP_ASSIST.X87_OUTPUT",
80− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
120+ "PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
81121 "SampleAfterValue": "100003",
82122 "CounterHTOff": "0,1,2,3,4,5,6,7"
83123 },
@@ -87,7 +127,7 @@ · 3 unmodified lines
87127 "BriefDescription": "Number of X87 assists due to input value.",
88128 "Counter": "0,1,2,3",
89129 "EventName": "FP_ASSIST.X87_INPUT",
90− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
130+ "PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
91131 "SampleAfterValue": "100003",
92132 "CounterHTOff": "0,1,2,3,4,5,6,7"
93133 },
@@ -97,7 +137,7 @@ · 3 unmodified lines
97137 "BriefDescription": "Number of SIMD FP assists due to Output values",
98138 "Counter": "0,1,2,3",
99139 "EventName": "FP_ASSIST.SIMD_OUTPUT",
100− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
140+ "PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
101141 "SampleAfterValue": "100003",
102142 "CounterHTOff": "0,1,2,3,4,5,6,7"
103143 },
@@ -107,7 +147,7 @@ · 3 unmodified lines
107147 "BriefDescription": "Number of SIMD FP assists due to input values",
108148 "Counter": "0,1,2,3",
109149 "EventName": "FP_ASSIST.SIMD_INPUT",
110− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
150+ "PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
111151 "SampleAfterValue": "100003",
112152 "CounterHTOff": "0,1,2,3,4,5,6,7"
113153 },
@@ -121,51 +161,5 @@ · 7 unmodified lines
121161 "PublicDescription": "This event counts cycles with any input and output SSE or x87 FP assist. If an input and output assist are detected on the same cycle the event increments by 1.",
122162 "SampleAfterValue": "100003",
123163 "CounterHTOff": "0,1,2,3"
124− },
125− {
126− "EventCode": "0xc7",
127− "UMask": "0x20",
128− "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
129− "PEBS": "1",
130− "Counter": "0,1,2,3",
131− "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
132− "SampleAfterValue": "2000003",
133− "CounterHTOff": "0,1,2,3"
134− },
135− {
136− "EventCode": "0xC7",
137− "UMask": "0x3",
138− "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
139− "Counter": "0,1,2,3",
140− "EventName": "FP_ARITH_INST_RETIRED.SCALAR",
141− "SampleAfterValue": "2000003",
142− "CounterHTOff": "0,1,2,3"
143− },
144− {
145− "EventCode": "0xC7",
146− "UMask": "0x3c",
147− "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
148− "Counter": "0,1,2,3",
149− "EventName": "FP_ARITH_INST_RETIRED.PACKED",
150− "SampleAfterValue": "2000004",
151− "CounterHTOff": "0,1,2,3"
152− },
153− {
154− "EventCode": "0xC7",
155− "UMask": "0x2a",
156− "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
157− "Counter": "0,1,2,3",
158− "EventName": "FP_ARITH_INST_RETIRED.SINGLE",
159− "SampleAfterValue": "2000005",
160− "CounterHTOff": "0,1,2,3"
161− },
162− {
163− "EventCode": "0xC7",
164− "UMask": "0x15",
165− "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
166− "Counter": "0,1,2,3",
167− "EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
168− "SampleAfterValue": "2000006",
169− "CounterHTOff": "0,1,2,3"
170164 }
171165 ]
tools/perf/pmu-events/arch/x86/broadwellde/frontend.json+69−69
@@ -15,80 +15,49 @@ · 14 unmodified lines
1515 "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path",
1616 "Counter": "0,1,2,3",
1717 "EventName": "IDQ.MITE_UOPS",
18− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
18+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
1919 "SampleAfterValue": "2000003",
2020 "CounterHTOff": "0,1,2,3,4,5,6,7"
2121 },
2222 {
2323 "EventCode": "0x79",
24− "UMask": "0x8",
25− "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
26− "Counter": "0,1,2,3",
27− "EventName": "IDQ.DSB_UOPS",
28− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
29− "SampleAfterValue": "2000003",
30− "CounterHTOff": "0,1,2,3,4,5,6,7"
31− },
32− {
33− "EventCode": "0x79",
34− "UMask": "0x10",
35− "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
36− "Counter": "0,1,2,3",
37− "EventName": "IDQ.MS_DSB_UOPS",
38− "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
39− "SampleAfterValue": "2000003",
40− "CounterHTOff": "0,1,2,3,4,5,6,7"
41− },
42− {
43− "EventCode": "0x79",
44− "UMask": "0x20",
45− "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
46− "Counter": "0,1,2,3",
47− "EventName": "IDQ.MS_MITE_UOPS",
48− "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
49− "SampleAfterValue": "2000003",
50− "CounterHTOff": "0,1,2,3,4,5,6,7"
51− },
52− {
53− "EventCode": "0x79",
54− "UMask": "0x30",
55− "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
24+ "UMask": "0x4",
25+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path",
5626 "Counter": "0,1,2,3",
57− "EventName": "IDQ.MS_UOPS",
58− "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
27+ "EventName": "IDQ.MITE_CYCLES",
28+ "CounterMask": "1",
29+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ.",
5930 "SampleAfterValue": "2000003",
6031 "CounterHTOff": "0,1,2,3,4,5,6,7"
6132 },
6233 {
6334 "EventCode": "0x79",
64− "UMask": "0x30",
65− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
35+ "UMask": "0x8",
36+ "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
6637 "Counter": "0,1,2,3",
67− "EventName": "IDQ.MS_CYCLES",
68− "CounterMask": "1",
69− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
38+ "EventName": "IDQ.DSB_UOPS",
39+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
7040 "SampleAfterValue": "2000003",
7141 "CounterHTOff": "0,1,2,3,4,5,6,7"
7242 },
7343 {
7444 "EventCode": "0x79",
75− "UMask": "0x4",
76− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path",
45+ "UMask": "0x8",
46+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path",
7747 "Counter": "0,1,2,3",
78− "EventName": "IDQ.MITE_CYCLES",
48+ "EventName": "IDQ.DSB_CYCLES",
7949 "CounterMask": "1",
80− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ.",
50+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
8151 "SampleAfterValue": "2000003",
8252 "CounterHTOff": "0,1,2,3,4,5,6,7"
8353 },
8454 {
8555 "EventCode": "0x79",
86− "UMask": "0x8",
87− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path",
56+ "UMask": "0x10",
57+ "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
8858 "Counter": "0,1,2,3",
89− "EventName": "IDQ.DSB_CYCLES",
90− "CounterMask": "1",
91− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
59+ "EventName": "IDQ.MS_DSB_UOPS",
60+ "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
9261 "SampleAfterValue": "2000003",
9362 "CounterHTOff": "0,1,2,3,4,5,6,7"
9463 },
@@ -99,7 +68,7 @@ · 4 unmodified lines
9968 "Counter": "0,1,2,3",
10069 "EventName": "IDQ.MS_DSB_CYCLES",
10170 "CounterMask": "1",
102− "PublicDescription": "This event counts cycles during which uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
71+ "PublicDescription": "This event counts cycles during which uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
10372 "SampleAfterValue": "2000003",
10473 "CounterHTOff": "0,1,2,3,4,5,6,7"
10574 },
@@ -111,7 +80,7 @@ · 5 unmodified lines
11180 "Counter": "0,1,2,3",
11281 "EventName": "IDQ.MS_DSB_OCCUR",
11382 "CounterMask": "1",
114− "PublicDescription": "This event counts the number of deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
83+ "PublicDescription": "This event counts the number of deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
11584 "SampleAfterValue": "2000003",
11685 "CounterHTOff": "0,1,2,3,4,5,6,7"
11786 },
@@ -122,7 +91,7 @@ · 4 unmodified lines
12291 "Counter": "0,1,2,3",
12392 "EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
12493 "CounterMask": "4",
125− "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
94+ "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
12695 "SampleAfterValue": "2000003",
12796 "CounterHTOff": "0,1,2,3,4,5,6,7"
12897 },
@@ -133,7 +102,17 @@ · 4 unmodified lines
133102 "Counter": "0,1,2,3",
134103 "EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
135104 "CounterMask": "1",
136− "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
105+ "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
106+ "SampleAfterValue": "2000003",
107+ "CounterHTOff": "0,1,2,3,4,5,6,7"
108+ },
109+ {
110+ "EventCode": "0x79",
111+ "UMask": "0x20",
112+ "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
113+ "Counter": "0,1,2,3",
114+ "EventName": "IDQ.MS_MITE_UOPS",
115+ "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ.",
137116 "SampleAfterValue": "2000003",
138117 "CounterHTOff": "0,1,2,3,4,5,6,7"
139118 },
@@ -144,7 +123,7 @@ · 4 unmodified lines
144123 "Counter": "0,1,2,3",
145124 "EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
146125 "CounterMask": "4",
147− "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
126+ "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
148127 "SampleAfterValue": "2000003",
149128 "CounterHTOff": "0,1,2,3,4,5,6,7"
150129 },
@@ -155,7 +134,39 @@ · 4 unmodified lines
155134 "Counter": "0,1,2,3",
156135 "EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
157136 "CounterMask": "1",
158− "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
137+ "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
138+ "SampleAfterValue": "2000003",
139+ "CounterHTOff": "0,1,2,3,4,5,6,7"
140+ },
141+ {
142+ "EventCode": "0x79",
143+ "UMask": "0x30",
144+ "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
145+ "Counter": "0,1,2,3",
146+ "EventName": "IDQ.MS_UOPS",
147+ "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
148+ "SampleAfterValue": "2000003",
149+ "CounterHTOff": "0,1,2,3,4,5,6,7"
150+ },
151+ {
152+ "EventCode": "0x79",
153+ "UMask": "0x30",
154+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
155+ "Counter": "0,1,2,3",
156+ "EventName": "IDQ.MS_CYCLES",
157+ "CounterMask": "1",
158+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
159+ "SampleAfterValue": "2000003",
160+ "CounterHTOff": "0,1,2,3,4,5,6,7"
161+ },
162+ {
163+ "EdgeDetect": "1",
164+ "EventCode": "0x79",
165+ "UMask": "0x30",
166+ "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
167+ "Counter": "0,1,2,3",
168+ "EventName": "IDQ.MS_SWITCHES",
169+ "CounterMask": "1",
159170 "SampleAfterValue": "2000003",
160171 "CounterHTOff": "0,1,2,3,4,5,6,7"
161172 },
@@ -165,7 +176,7 @@ · 3 unmodified lines
165176 "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path",
166177 "Counter": "0,1,2,3",
167178 "EventName": "IDQ.MITE_ALL_UOPS",
168− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
179+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
169180 "SampleAfterValue": "2000003",
170181 "CounterHTOff": "0,1,2,3,4,5,6,7"
171182 },
@@ -205,7 +216,7 @@ · 33 unmodified lines
205216 "BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled",
206217 "Counter": "0,1,2,3",
207218 "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE",
208− "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding ?4 ? x? when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
219+ "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
209220 "SampleAfterValue": "2000003",
210221 "CounterHTOff": "0,1,2,3"
211222 },
@@ -268,18 +279,7 @@ · 56 unmodified lines
268279 "BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.",
269280 "Counter": "0,1,2,3",
270281 "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES",
271− "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 0?2 cycles.",
272− "SampleAfterValue": "2000003",
273− "CounterHTOff": "0,1,2,3,4,5,6,7"
274− },
275− {
276− "EdgeDetect": "1",
277− "EventCode": "0x79",
278− "UMask": "0x30",
279− "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
280− "Counter": "0,1,2,3",
281− "EventName": "IDQ.MS_SWITCHES",
282− "CounterMask": "1",
282+ "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
283283 "SampleAfterValue": "2000003",
284284 "CounterHTOff": "0,1,2,3,4,5,6,7"
285285 }
tools/perf/pmu-events/arch/x86/broadwellde/memory.json+4−5
@@ -95,7 +95,6 @@ · 94 unmodified lines
9595 "BriefDescription": "Counts the number of times a class of instructions that may cause a transactional abort was executed. Since this is the count of execution, it may not always cause a transactional abort.",
9696 "Counter": "0,1,2,3",
9797 "EventName": "TX_EXEC.MISC1",
98− "PublicDescription": "Unfriendly TSX abort triggered by a flowmarker.",
9998 "SampleAfterValue": "2000003",
10099 "CounterHTOff": "0,1,2,3,4,5,6,7"
101100 },
@@ -171,11 +170,11 @@ · 69 unmodified lines
171170 {
172171 "EventCode": "0xc8",
173172 "UMask": "0x4",
174− "BriefDescription": "Number of times HLE abort was triggered",
173+ "BriefDescription": "Number of times HLE abort was triggered (PEBS)",
175174 "PEBS": "1",
176175 "Counter": "0,1,2,3",
177176 "EventName": "HLE_RETIRED.ABORTED",
178− "PublicDescription": "Number of times HLE abort was triggered.",
177+ "PublicDescription": "Number of times HLE abort was triggered (PEBS).",
179178 "SampleAfterValue": "2000003",
180179 "CounterHTOff": "0,1,2,3,4,5,6,7"
181180 },
@@ -252,11 +251,11 @@ · 70 unmodified lines
252251 {
253252 "EventCode": "0xc9",
254253 "UMask": "0x4",
255− "BriefDescription": "Number of times RTM abort was triggered",
254+ "BriefDescription": "Number of times RTM abort was triggered (PEBS)",
256255 "PEBS": "1",
257256 "Counter": "0,1,2,3",
258257 "EventName": "RTM_RETIRED.ABORTED",
259− "PublicDescription": "Number of times RTM abort was triggered .",
258+ "PublicDescription": "Number of times RTM abort was triggered (PEBS).",
260259 "SampleAfterValue": "2000003",
261260 "CounterHTOff": "0,1,2,3"
262261 },
tools/perf/pmu-events/arch/x86/broadwellde/other.json+10−10
@@ -9,16 +9,6 @@ · 8 unmodified lines
99 "SampleAfterValue": "2000003",
1010 "CounterHTOff": "0,1,2,3,4,5,6,7"
1111 },
12− {
13− "EventCode": "0x5C",
14− "UMask": "0x2",
15− "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3",
16− "Counter": "0,1,2,3",
17− "EventName": "CPL_CYCLES.RING123",
18− "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
19− "SampleAfterValue": "2000003",
20− "CounterHTOff": "0,1,2,3,4,5,6,7"
21− },
2212 {
2313 "EdgeDetect": "1",
2414 "EventCode": "0x5C",
@@ -31,6 +21,16 @@ · 6 unmodified lines
3121 "SampleAfterValue": "100007",
3222 "CounterHTOff": "0,1,2,3,4,5,6,7"
3323 },
24+ {
25+ "EventCode": "0x5C",
26+ "UMask": "0x2",
27+ "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3",
28+ "Counter": "0,1,2,3",
29+ "EventName": "CPL_CYCLES.RING123",
30+ "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
31+ "SampleAfterValue": "2000003",
32+ "CounterHTOff": "0,1,2,3,4,5,6,7"
33+ },
3434 {
3535 "EventCode": "0x63",
3636 "UMask": "0x1",
tools/perf/pmu-events/arch/x86/broadwellde/pipeline.json+612−602
@@ -3,31 +3,41 @@ · 2 unmodified lines
33 "EventCode": "0x00",
44 "UMask": "0x1",
55 "BriefDescription": "Instructions retired from execution.",
6− "Counter": "Fixed counter 1",
6+ "Counter": "Fixed counter 0",
77 "EventName": "INST_RETIRED.ANY",
88 "PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. \nNotes: INST_RETIRED.ANY is counted by a designated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. INST_RETIRED.ANY_P is counted by a programmable counter and it is an architectural performance event. \nCounting: Faulting executions of GETSEC/VM entry/VM Exit/MWait will not count as retired instructions.",
99 "SampleAfterValue": "2000003",
10− "CounterHTOff": "Fixed counter 1"
10+ "CounterHTOff": "Fixed counter 0"
1111 },
1212 {
1313 "EventCode": "0x00",
1414 "UMask": "0x2",
1515 "BriefDescription": "Core cycles when the thread is not in halt state",
16− "Counter": "Fixed counter 2",
16+ "Counter": "Fixed counter 1",
1717 "EventName": "CPU_CLK_UNHALTED.THREAD",
1818 "PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
1919 "SampleAfterValue": "2000003",
20− "CounterHTOff": "Fixed counter 2"
20+ "CounterHTOff": "Fixed counter 1"
21+ },
22+ {
23+ "EventCode": "0x00",
24+ "UMask": "0x2",
25+ "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
26+ "Counter": "Fixed counter 1",
27+ "EventName": "CPU_CLK_UNHALTED.THREAD_ANY",
28+ "AnyThread": "1",
29+ "SampleAfterValue": "2000003",
30+ "CounterHTOff": "Fixed counter 1"
2131 },
2232 {
2333 "EventCode": "0x00",
2434 "UMask": "0x3",
2535 "BriefDescription": "Reference cycles when the core is not in halt state.",
26− "Counter": "Fixed counter 3",
36+ "Counter": "Fixed counter 2",
2737 "EventName": "CPU_CLK_UNHALTED.REF_TSC",
2838 "PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. \nNote: On all current platforms this event stops counting during 'throttling (TM)' states duty off periods the processor is 'halted'. This event is clocked by base clock (100 Mhz) on Sandy Bridge. The counter update is done at a lower clock rate then the core clock the overflow status bit for this counter may appear 'sticky'. After the counter has overflowed and software clears the overflow status bit and resets the counter to less than MAX. The reset value to the counter is not clocked immediately so the overflow status bit will flip 'high (1)' and generate another PMI (if enabled) after which the reset value gets clocked into the counter. Therefore, software will get the interrupt, read the overflow status bit '1 for bit 34 while the counter value is less than MAX. Software should ignore this case.",
2939 "SampleAfterValue": "2000003",
30− "CounterHTOff": "Fixed counter 3"
40+ "CounterHTOff": "Fixed counter 2"
3141 },
3242 {
3343 "EventCode": "0x03",
@@ -60,22 +70,33 @@ · 26 unmodified lines
6070 },
6171 {
6272 "EventCode": "0x0D",
63− "UMask": "0x8",
64− "BriefDescription": "Cycles when Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the thread",
73+ "UMask": "0x3",
74+ "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for this thread (e.g. misprediction or memory nuke)",
6575 "Counter": "0,1,2,3",
66− "EventName": "INT_MISC.RAT_STALL_CYCLES",
67− "PublicDescription": "This event counts the number of cycles during which Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the current thread. This also includes the cycles during which the Allocator is serving another thread.",
76+ "EventName": "INT_MISC.RECOVERY_CYCLES",
77+ "CounterMask": "1",
78+ "PublicDescription": "Cycles checkpoints in Resource Allocation Table (RAT) are recovering from JEClear or machine clear.",
6879 "SampleAfterValue": "2000003",
6980 "CounterHTOff": "0,1,2,3,4,5,6,7"
7081 },
7182 {
7283 "EventCode": "0x0D",
7384 "UMask": "0x3",
74− "BriefDescription": "Number of cycles waiting for the checkpoints in Resource Allocation Table (RAT) to be recovered after Nuke due to all other cases except JEClear (e.g. whenever a ucode assist is needed like SSE exception, memory disambiguation, etc...)",
85+ "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for any thread running on the physical core (e.g. misprediction or memory nuke).",
7586 "Counter": "0,1,2,3",
76− "EventName": "INT_MISC.RECOVERY_CYCLES",
87+ "EventName": "INT_MISC.RECOVERY_CYCLES_ANY",
88+ "AnyThread": "1",
7789 "CounterMask": "1",
78− "PublicDescription": "Cycles checkpoints in Resource Allocation Table (RAT) are recovering from JEClear or machine clear.",
90+ "SampleAfterValue": "2000003",
91+ "CounterHTOff": "0,1,2,3,4,5,6,7"
92+ },
93+ {
94+ "EventCode": "0x0D",
95+ "UMask": "0x8",
96+ "BriefDescription": "Cycles when Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the thread",
97+ "Counter": "0,1,2,3",
98+ "EventName": "INT_MISC.RAT_STALL_CYCLES",
99+ "PublicDescription": "This event counts the number of cycles during which Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the current thread. This also includes the cycles during which the Allocator is serving another thread.",
79100 "SampleAfterValue": "2000003",
80101 "CounterHTOff": "0,1,2,3,4,5,6,7"
81102 },
@@ -89,6 +110,18 @@ · 7 unmodified lines
89110 "SampleAfterValue": "2000003",
90111 "CounterHTOff": "0,1,2,3,4,5,6,7"
91112 },
113+ {
114+ "Invert": "1",
115+ "EventCode": "0x0E",
116+ "UMask": "0x1",
117+ "BriefDescription": "Cycles when Resource Allocation Table (RAT) does not issue Uops to Reservation Station (RS) for the thread",
118+ "Counter": "0,1,2,3",
119+ "EventName": "UOPS_ISSUED.STALL_CYCLES",
120+ "CounterMask": "1",
121+ "PublicDescription": "This event counts cycles during which the Resource Allocation Table (RAT) does not issue any Uops to the reservation station (RS) for the current thread.",
122+ "SampleAfterValue": "2000003",
123+ "CounterHTOff": "0,1,2,3"
124+ },
92125 {
93126 "EventCode": "0x0E",
94127 "UMask": "0x10",
@@ -117,18 +150,6 @@ · 22 unmodified lines
117150 "SampleAfterValue": "2000003",
118151 "CounterHTOff": "0,1,2,3,4,5,6,7"
119152 },
120− {
121− "Invert": "1",
122− "EventCode": "0x0E",
123− "UMask": "0x1",
124− "BriefDescription": "Cycles when Resource Allocation Table (RAT) does not issue Uops to Reservation Station (RS) for the thread",
125− "Counter": "0,1,2,3",
126− "EventName": "UOPS_ISSUED.STALL_CYCLES",
127− "CounterMask": "1",
128− "PublicDescription": "This event counts cycles during which the Resource Allocation Table (RAT) does not issue any Uops to the reservation station (RS) for the current thread.",
129− "SampleAfterValue": "2000003",
130− "CounterHTOff": "0,1,2,3"
131− },
132153 {
133154 "EventCode": "0x14",
134155 "UMask": "0x1",
@@ -139,6 +160,26 @@ · 4 unmodified lines
139160 "SampleAfterValue": "2000003",
140161 "CounterHTOff": "0,1,2,3,4,5,6,7"
141162 },
163+ {
164+ "EventCode": "0x3C",
165+ "UMask": "0x0",
166+ "BriefDescription": "Thread cycles when thread is not in halt state",
167+ "Counter": "0,1,2,3",
168+ "EventName": "CPU_CLK_UNHALTED.THREAD_P",
169+ "PublicDescription": "This is an architectural event that counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling. For this reason, this event may have a changing ratio with regards to wall clock time.",
170+ "SampleAfterValue": "2000003",
171+ "CounterHTOff": "0,1,2,3,4,5,6,7"
172+ },
173+ {
174+ "EventCode": "0x3C",
175+ "UMask": "0x0",
176+ "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
177+ "Counter": "0,1,2,3",
178+ "EventName": "CPU_CLK_UNHALTED.THREAD_P_ANY",
179+ "AnyThread": "1",
180+ "SampleAfterValue": "2000003",
181+ "CounterHTOff": "0,1,2,3,4,5,6,7"
182+ },
142183 {
143184 "EventCode": "0x3C",
144185 "UMask": "0x1",
@@ -149,6 +190,36 @@ · 4 unmodified lines
149190 "SampleAfterValue": "2000003",
150191 "CounterHTOff": "0,1,2,3,4,5,6,7"
151192 },
193+ {
194+ "EventCode": "0x3C",
195+ "UMask": "0x1",
196+ "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
197+ "Counter": "0,1,2,3",
198+ "EventName": "CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY",
199+ "AnyThread": "1",
200+ "SampleAfterValue": "2000003",
201+ "CounterHTOff": "0,1,2,3,4,5,6,7"
202+ },
203+ {
204+ "EventCode": "0x3C",
205+ "UMask": "0x1",
206+ "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
207+ "Counter": "0,1,2,3",
208+ "EventName": "CPU_CLK_UNHALTED.REF_XCLK",
209+ "PublicDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate).",
210+ "SampleAfterValue": "2000003",
211+ "CounterHTOff": "0,1,2,3,4,5,6,7"
212+ },
213+ {
214+ "EventCode": "0x3C",
215+ "UMask": "0x1",
216+ "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
217+ "Counter": "0,1,2,3",
218+ "EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
219+ "AnyThread": "1",
220+ "SampleAfterValue": "2000003",
221+ "CounterHTOff": "0,1,2,3,4,5,6,7"
222+ },
152223 {
153224 "EventCode": "0x3c",
154225 "UMask": "0x2",
@@ -158,6 +229,15 @@ · 3 unmodified lines
158229 "SampleAfterValue": "2000003",
159230 "CounterHTOff": "0,1,2,3"
160231 },
232+ {
233+ "EventCode": "0x3C",
234+ "UMask": "0x2",
235+ "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
236+ "Counter": "0,1,2,3",
237+ "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE",
238+ "SampleAfterValue": "2000003",
239+ "CounterHTOff": "0,1,2,3,4,5,6,7"
240+ },
161241 {
162242 "EventCode": "0x4c",
163243 "UMask": "0x1",
@@ -224,6 +304,18 @@ · 60 unmodified lines
224304 "SampleAfterValue": "2000003",
225305 "CounterHTOff": "0,1,2,3,4,5,6,7"
226306 },
307+ {
308+ "EdgeDetect": "1",
309+ "Invert": "1",
310+ "EventCode": "0x5E",
311+ "UMask": "0x1",
312+ "BriefDescription": "Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues.",
313+ "Counter": "0,1,2,3",
314+ "EventName": "RS_EVENTS.EMPTY_END",
315+ "CounterMask": "1",
316+ "SampleAfterValue": "200003",
317+ "CounterHTOff": "0,1,2,3,4,5,6,7"
318+ },
227319 {
228320 "EventCode": "0x87",
229321 "UMask": "0x1",
@@ -404,6 +496,15 @@ · 174 unmodified lines
404496 "SampleAfterValue": "200003",
405497 "CounterHTOff": "0,1,2,3,4,5,6,7"
406498 },
499+ {
500+ "EventCode": "0x89",
501+ "UMask": "0xa0",
502+ "BriefDescription": "Taken speculative and retired mispredicted indirect calls.",
503+ "Counter": "0,1,2,3",
504+ "EventName": "BR_MISP_EXEC.TAKEN_INDIRECT_NEAR_CALL",
505+ "SampleAfterValue": "200003",
506+ "CounterHTOff": "0,1,2,3,4,5,6,7"
507+ },
407508 {
408509 "EventCode": "0x89",
409510 "UMask": "0xc1",
@@ -434,6 +535,16 @@ · 24 unmodified lines
434535 "SampleAfterValue": "200003",
435536 "CounterHTOff": "0,1,2,3,4,5,6,7"
436537 },
538+ {
539+ "EventCode": "0xA0",
540+ "UMask": "0x3",
541+ "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports",
542+ "Counter": "0,1,2,3",
543+ "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF",
544+ "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
545+ "SampleAfterValue": "2000003",
546+ "CounterHTOff": "0,1,2,3"
547+ },
437548 {
438549 "EventCode": "0xA1",
439550 "UMask": "0x1",
@@ -446,601 +557,471 @@ · 6 unmodified lines
446557 },
447558 {
448559 "EventCode": "0xA1",
449− "UMask": "0x2",
450− "BriefDescription": "Cycles per thread when uops are executed in port 1",
560+ "UMask": "0x1",
561+ "BriefDescription": "Cycles per core when uops are exectuted in port 0.",
451562 "Counter": "0,1,2,3",
452− "EventName": "UOPS_DISPATCHED_PORT.PORT_1",
453− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
563+ "EventName": "UOPS_EXECUTED_PORT.PORT_0_CORE",
564+ "AnyThread": "1",
454565 "SampleAfterValue": "2000003",
455566 "CounterHTOff": "0,1,2,3,4,5,6,7"
456567 },
457568 {
458569 "EventCode": "0xA1",
459− "UMask": "0x4",
460− "BriefDescription": "Cycles per thread when uops are executed in port 2",
570+ "UMask": "0x1",
571+ "BriefDescription": "Cycles per thread when uops are executed in port 0",
461572 "Counter": "0,1,2,3",
462− "EventName": "UOPS_DISPATCHED_PORT.PORT_2",
463− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
573+ "EventName": "UOPS_EXECUTED_PORT.PORT_0",
574+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
464575 "SampleAfterValue": "2000003",
465576 "CounterHTOff": "0,1,2,3,4,5,6,7"
466577 },
467578 {
468579 "EventCode": "0xA1",
469− "UMask": "0x8",
470− "BriefDescription": "Cycles per thread when uops are executed in port 3",
580+ "UMask": "0x2",
581+ "BriefDescription": "Cycles per thread when uops are executed in port 1",
471582 "Counter": "0,1,2,3",
472− "EventName": "UOPS_DISPATCHED_PORT.PORT_3",
473− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
583+ "EventName": "UOPS_DISPATCHED_PORT.PORT_1",
584+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
474585 "SampleAfterValue": "2000003",
475586 "CounterHTOff": "0,1,2,3,4,5,6,7"
476587 },
477588 {
478589 "EventCode": "0xA1",
479− "UMask": "0x10",
480− "BriefDescription": "Cycles per thread when uops are executed in port 4",
590+ "UMask": "0x2",
591+ "BriefDescription": "Cycles per core when uops are exectuted in port 1.",
481592 "Counter": "0,1,2,3",
482− "EventName": "UOPS_DISPATCHED_PORT.PORT_4",
483− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
593+ "EventName": "UOPS_EXECUTED_PORT.PORT_1_CORE",
594+ "AnyThread": "1",
484595 "SampleAfterValue": "2000003",
485596 "CounterHTOff": "0,1,2,3,4,5,6,7"
486597 },
487598 {
488599 "EventCode": "0xA1",
489− "UMask": "0x20",
490− "BriefDescription": "Cycles per thread when uops are executed in port 5",
600+ "UMask": "0x2",
601+ "BriefDescription": "Cycles per thread when uops are executed in port 1",
491602 "Counter": "0,1,2,3",
492− "EventName": "UOPS_DISPATCHED_PORT.PORT_5",
493− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
603+ "EventName": "UOPS_EXECUTED_PORT.PORT_1",
604+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
494605 "SampleAfterValue": "2000003",
495606 "CounterHTOff": "0,1,2,3,4,5,6,7"
496607 },
497608 {
498609 "EventCode": "0xA1",
499− "UMask": "0x40",
500− "BriefDescription": "Cycles per thread when uops are executed in port 6",
610+ "UMask": "0x4",
611+ "BriefDescription": "Cycles per thread when uops are executed in port 2",
501612 "Counter": "0,1,2,3",
502− "EventName": "UOPS_DISPATCHED_PORT.PORT_6",
503− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
613+ "EventName": "UOPS_DISPATCHED_PORT.PORT_2",
614+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
504615 "SampleAfterValue": "2000003",
505616 "CounterHTOff": "0,1,2,3,4,5,6,7"
506617 },
507618 {
508619 "EventCode": "0xA1",
509− "UMask": "0x80",
510− "BriefDescription": "Cycles per thread when uops are executed in port 7",
620+ "UMask": "0x4",
621+ "BriefDescription": "Cycles per core when uops are dispatched to port 2.",
511622 "Counter": "0,1,2,3",
512− "EventName": "UOPS_DISPATCHED_PORT.PORT_7",
513− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
623+ "EventName": "UOPS_EXECUTED_PORT.PORT_2_CORE",
624+ "AnyThread": "1",
514625 "SampleAfterValue": "2000003",
515626 "CounterHTOff": "0,1,2,3,4,5,6,7"
516627 },
517628 {
518− "EventCode": "0xA2",
519− "UMask": "0x1",
520− "BriefDescription": "Resource-related stall cycles",
629+ "EventCode": "0xA1",
630+ "UMask": "0x4",
631+ "BriefDescription": "Cycles per thread when uops are executed in port 2",
521632 "Counter": "0,1,2,3",
522− "EventName": "RESOURCE_STALLS.ANY",
523− "PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
633+ "EventName": "UOPS_EXECUTED_PORT.PORT_2",
634+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
524635 "SampleAfterValue": "2000003",
525636 "CounterHTOff": "0,1,2,3,4,5,6,7"
526637 },
527638 {
528− "EventCode": "0xA2",
529− "UMask": "0x4",
530− "BriefDescription": "Cycles stalled due to no eligible RS entry available.",
639+ "EventCode": "0xA1",
640+ "UMask": "0x8",
641+ "BriefDescription": "Cycles per thread when uops are executed in port 3",
531642 "Counter": "0,1,2,3",
532− "EventName": "RESOURCE_STALLS.RS",
533− "PublicDescription": "This event counts stall cycles caused by absence of eligible entries in the reservation station (RS). This may result from RS overflow, or from RS deallocation because of the RS array Write Port allocation scheme (each RS entry has two write ports instead of four. As a result, empty entries could not be used, although RS is not really full). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
643+ "EventName": "UOPS_DISPATCHED_PORT.PORT_3",
644+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
534645 "SampleAfterValue": "2000003",
535646 "CounterHTOff": "0,1,2,3,4,5,6,7"
536647 },
537648 {
538− "EventCode": "0xA2",
649+ "EventCode": "0xA1",
539650 "UMask": "0x8",
540− "BriefDescription": "Cycles stalled due to no store buffers available. (not including draining form sync).",
651+ "BriefDescription": "Cycles per core when uops are dispatched to port 3.",
541652 "Counter": "0,1,2,3",
542− "EventName": "RESOURCE_STALLS.SB",
543− "PublicDescription": "This event counts stall cycles caused by the store buffer (SB) overflow (excluding draining from synch). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
653+ "EventName": "UOPS_EXECUTED_PORT.PORT_3_CORE",
654+ "AnyThread": "1",
544655 "SampleAfterValue": "2000003",
545656 "CounterHTOff": "0,1,2,3,4,5,6,7"
546657 },
547658 {
548− "EventCode": "0xA2",
549− "UMask": "0x10",
550− "BriefDescription": "Cycles stalled due to re-order buffer full.",
659+ "EventCode": "0xA1",
660+ "UMask": "0x8",
661+ "BriefDescription": "Cycles per thread when uops are executed in port 3",
551662 "Counter": "0,1,2,3",
552− "EventName": "RESOURCE_STALLS.ROB",
553− "PublicDescription": "This event counts ROB full stall cycles. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
663+ "EventName": "UOPS_EXECUTED_PORT.PORT_3",
664+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
554665 "SampleAfterValue": "2000003",
555666 "CounterHTOff": "0,1,2,3,4,5,6,7"
556667 },
557668 {
558− "EventCode": "0xA3",
559− "UMask": "0x1",
560− "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
669+ "EventCode": "0xA1",
670+ "UMask": "0x10",
671+ "BriefDescription": "Cycles per thread when uops are executed in port 4",
561672 "Counter": "0,1,2,3",
562− "EventName": "CYCLE_ACTIVITY.CYCLES_L2_PENDING",
563− "CounterMask": "1",
564− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand* load request missing the L2 cache.",
673+ "EventName": "UOPS_DISPATCHED_PORT.PORT_4",
674+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
565675 "SampleAfterValue": "2000003",
566676 "CounterHTOff": "0,1,2,3,4,5,6,7"
567677 },
568678 {
569− "EventCode": "0xA3",
570− "UMask": "0x8",
571− "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
572− "Counter": "2",
573− "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_PENDING",
574− "CounterMask": "8",
575− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request missing the L1 data cache.",
576− "SampleAfterValue": "2000003",
577− "CounterHTOff": "2"
578− },
579− {
580− "EventCode": "0xA3",
581− "UMask": "0x2",
582− "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
679+ "EventCode": "0xA1",
680+ "UMask": "0x10",
681+ "BriefDescription": "Cycles per core when uops are exectuted in port 4.",
583682 "Counter": "0,1,2,3",
584− "EventName": "CYCLE_ACTIVITY.CYCLES_LDM_PENDING",
585− "CounterMask": "2",
586− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request (that is cycles with non-completed load waiting for its data from memory subsystem).",
683+ "EventName": "UOPS_EXECUTED_PORT.PORT_4_CORE",
684+ "AnyThread": "1",
587685 "SampleAfterValue": "2000003",
588686 "CounterHTOff": "0,1,2,3,4,5,6,7"
589687 },
590688 {
591− "EventCode": "0xA3",
592− "UMask": "0x4",
593− "BriefDescription": "Total execution stalls",
689+ "EventCode": "0xA1",
690+ "UMask": "0x10",
691+ "BriefDescription": "Cycles per thread when uops are executed in port 4",
594692 "Counter": "0,1,2,3",
595− "EventName": "CYCLE_ACTIVITY.CYCLES_NO_EXECUTE",
596− "CounterMask": "4",
597− "PublicDescription": "Counts number of cycles nothing is executed on any execution port.",
693+ "EventName": "UOPS_EXECUTED_PORT.PORT_4",
694+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
598695 "SampleAfterValue": "2000003",
599− "CounterHTOff": "0,1,2,3"
696+ "CounterHTOff": "0,1,2,3,4,5,6,7"
600697 },
601698 {
602− "EventCode": "0xA3",
603− "UMask": "0x5",
604− "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
699+ "EventCode": "0xA1",
700+ "UMask": "0x20",
701+ "BriefDescription": "Cycles per thread when uops are executed in port 5",
605702 "Counter": "0,1,2,3",
606− "EventName": "CYCLE_ACTIVITY.STALLS_L2_PENDING",
607− "CounterMask": "5",
608− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand* load request missing the L2 cache.(as a footprint) * includes also L1 HW prefetch requests that may or may not be required by demands.",
703+ "EventName": "UOPS_DISPATCHED_PORT.PORT_5",
704+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
609705 "SampleAfterValue": "2000003",
610− "CounterHTOff": "0,1,2,3"
706+ "CounterHTOff": "0,1,2,3,4,5,6,7"
611707 },
612708 {
613− "EventCode": "0xA3",
614− "UMask": "0x6",
615− "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
709+ "EventCode": "0xA1",
710+ "UMask": "0x20",
711+ "BriefDescription": "Cycles per core when uops are exectuted in port 5.",
616712 "Counter": "0,1,2,3",
617− "EventName": "CYCLE_ACTIVITY.STALLS_LDM_PENDING",
618− "CounterMask": "6",
619− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request.",
713+ "EventName": "UOPS_EXECUTED_PORT.PORT_5_CORE",
714+ "AnyThread": "1",
620715 "SampleAfterValue": "2000003",
621− "CounterHTOff": "0,1,2,3"
716+ "CounterHTOff": "0,1,2,3,4,5,6,7"
622717 },
623718 {
624− "EventCode": "0xA3",
625− "UMask": "0xc",
626− "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
627− "Counter": "2",
628− "EventName": "CYCLE_ACTIVITY.STALLS_L1D_PENDING",
629− "CounterMask": "12",
630− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request missing the L1 data cache.",
719+ "EventCode": "0xA1",
720+ "UMask": "0x20",
721+ "BriefDescription": "Cycles per thread when uops are executed in port 5",
722+ "Counter": "0,1,2,3",
723+ "EventName": "UOPS_EXECUTED_PORT.PORT_5",
724+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
631725 "SampleAfterValue": "2000003",
632− "CounterHTOff": "2"
726+ "CounterHTOff": "0,1,2,3,4,5,6,7"
633727 },
634728 {
635− "EventCode": "0xA8",
636− "UMask": "0x1",
637− "BriefDescription": "Number of Uops delivered by the LSD.",
729+ "EventCode": "0xA1",
730+ "UMask": "0x40",
731+ "BriefDescription": "Cycles per thread when uops are executed in port 6",
638732 "Counter": "0,1,2,3",
639− "EventName": "LSD.UOPS",
640− "PublicDescription": "Number of Uops delivered by the LSD. ",
733+ "EventName": "UOPS_DISPATCHED_PORT.PORT_6",
734+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
641735 "SampleAfterValue": "2000003",
642736 "CounterHTOff": "0,1,2,3,4,5,6,7"
643737 },
644738 {
645− "EventCode": "0xB1",
646− "UMask": "0x1",
647− "BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
739+ "EventCode": "0xA1",
740+ "UMask": "0x40",
741+ "BriefDescription": "Cycles per core when uops are exectuted in port 6.",
648742 "Counter": "0,1,2,3",
649− "EventName": "UOPS_EXECUTED.THREAD",
650− "PublicDescription": "Number of uops to be executed per-thread each cycle.",
743+ "EventName": "UOPS_EXECUTED_PORT.PORT_6_CORE",
744+ "AnyThread": "1",
651745 "SampleAfterValue": "2000003",
652746 "CounterHTOff": "0,1,2,3,4,5,6,7"
653747 },
654748 {
655− "EventCode": "0xB1",
656− "UMask": "0x2",
657− "BriefDescription": "Number of uops executed on the core.",
749+ "EventCode": "0xA1",
750+ "UMask": "0x40",
751+ "BriefDescription": "Cycles per thread when uops are executed in port 6",
658752 "Counter": "0,1,2,3",
659− "EventName": "UOPS_EXECUTED.CORE",
660− "PublicDescription": "Number of uops executed from any thread.",
753+ "EventName": "UOPS_EXECUTED_PORT.PORT_6",
754+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
661755 "SampleAfterValue": "2000003",
662756 "CounterHTOff": "0,1,2,3,4,5,6,7"
663757 },
664758 {
665− "Invert": "1",
666− "EventCode": "0xB1",
667− "UMask": "0x1",
668− "BriefDescription": "Counts number of cycles no uops were dispatched to be executed on this thread.",
759+ "EventCode": "0xA1",
760+ "UMask": "0x80",
761+ "BriefDescription": "Cycles per thread when uops are executed in port 7",
669762 "Counter": "0,1,2,3",
670− "EventName": "UOPS_EXECUTED.STALL_CYCLES",
671− "CounterMask": "1",
672− "PublicDescription": "This event counts cycles during which no uops were dispatched from the Reservation Station (RS) per thread.",
763+ "EventName": "UOPS_DISPATCHED_PORT.PORT_7",
764+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
673765 "SampleAfterValue": "2000003",
674− "CounterHTOff": "0,1,2,3"
766+ "CounterHTOff": "0,1,2,3,4,5,6,7"
675767 },
676768 {
677− "EventCode": "0xC0",
678− "UMask": "0x0",
679− "BriefDescription": "Number of instructions retired. General Counter - architectural event",
769+ "EventCode": "0xA1",
770+ "UMask": "0x80",
771+ "BriefDescription": "Cycles per core when uops are dispatched to port 7.",
680772 "Counter": "0,1,2,3",
681− "EventName": "INST_RETIRED.ANY_P",
682− "Errata": "BDM61",
683− "PublicDescription": "This event counts the number of instructions (EOMs) retired. Counting covers macro-fused instructions individually (that is, increments by two).",
773+ "EventName": "UOPS_EXECUTED_PORT.PORT_7_CORE",
774+ "AnyThread": "1",
684775 "SampleAfterValue": "2000003",
685776 "CounterHTOff": "0,1,2,3,4,5,6,7"
686777 },
687778 {
688− "EventCode": "0xC0",
689− "UMask": "0x2",
690− "BriefDescription": "FP operations retired. X87 FP operations that have no exceptions:",
779+ "EventCode": "0xA1",
780+ "UMask": "0x80",
781+ "BriefDescription": "Cycles per thread when uops are executed in port 7",
691782 "Counter": "0,1,2,3",
692− "EventName": "INST_RETIRED.X87",
693− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
783+ "EventName": "UOPS_EXECUTED_PORT.PORT_7",
784+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
694785 "SampleAfterValue": "2000003",
695786 "CounterHTOff": "0,1,2,3,4,5,6,7"
696787 },
697788 {
698− "EventCode": "0xC0",
789+ "EventCode": "0xA2",
699790 "UMask": "0x1",
700− "BriefDescription": "Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution",
701− "PEBS": "2",
702− "Counter": "1",
703− "EventName": "INST_RETIRED.PREC_DIST",
704− "Errata": "BDM11, BDM55",
705− "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts instructions retired.",
791+ "BriefDescription": "Resource-related stall cycles",
792+ "Counter": "0,1,2,3",
793+ "EventName": "RESOURCE_STALLS.ANY",
794+ "PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
706795 "SampleAfterValue": "2000003",
707− "CounterHTOff": "1"
796+ "CounterHTOff": "0,1,2,3,4,5,6,7"
708797 },
709798 {
710− "EventCode": "0xC1",
711− "UMask": "0x40",
712− "BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
799+ "EventCode": "0xA2",
800+ "UMask": "0x4",
801+ "BriefDescription": "Cycles stalled due to no eligible RS entry available.",
713802 "Counter": "0,1,2,3",
714− "EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
715− "SampleAfterValue": "100003",
803+ "EventName": "RESOURCE_STALLS.RS",
804+ "PublicDescription": "This event counts stall cycles caused by absence of eligible entries in the reservation station (RS). This may result from RS overflow, or from RS deallocation because of the RS array Write Port allocation scheme (each RS entry has two write ports instead of four. As a result, empty entries could not be used, although RS is not really full). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
805+ "SampleAfterValue": "2000003",
716806 "CounterHTOff": "0,1,2,3,4,5,6,7"
717807 },
718808 {
719− "EventCode": "0xC2",
720− "UMask": "0x1",
721− "BriefDescription": "Actually retired uops.",
722− "Data_LA": "1",
723− "PEBS": "1",
809+ "EventCode": "0xA2",
810+ "UMask": "0x8",
811+ "BriefDescription": "Cycles stalled due to no store buffers available. (not including draining form sync).",
724812 "Counter": "0,1,2,3",
725− "EventName": "UOPS_RETIRED.ALL",
726− "PublicDescription": "This event counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
813+ "EventName": "RESOURCE_STALLS.SB",
814+ "PublicDescription": "This event counts stall cycles caused by the store buffer (SB) overflow (excluding draining from synch). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
727815 "SampleAfterValue": "2000003",
728816 "CounterHTOff": "0,1,2,3,4,5,6,7"
729817 },
730818 {
731− "EventCode": "0xC2",
732− "UMask": "0x2",
733− "BriefDescription": "Retirement slots used.",
734− "PEBS": "1",
819+ "EventCode": "0xA2",
820+ "UMask": "0x10",
821+ "BriefDescription": "Cycles stalled due to re-order buffer full.",
735822 "Counter": "0,1,2,3",
736− "EventName": "UOPS_RETIRED.RETIRE_SLOTS",
737− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of retirement slots used.",
823+ "EventName": "RESOURCE_STALLS.ROB",
824+ "PublicDescription": "This event counts ROB full stall cycles. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
738825 "SampleAfterValue": "2000003",
739826 "CounterHTOff": "0,1,2,3,4,5,6,7"
740827 },
741828 {
742− "Invert": "1",
743− "EventCode": "0xC2",
829+ "EventCode": "0xA3",
744830 "UMask": "0x1",
745− "BriefDescription": "Cycles without actually retired uops.",
831+ "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
746832 "Counter": "0,1,2,3",
747− "EventName": "UOPS_RETIRED.STALL_CYCLES",
833+ "EventName": "CYCLE_ACTIVITY.CYCLES_L2_PENDING",
748834 "CounterMask": "1",
749− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts cycles without actually retired uops.",
835+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand* load request missing the L2 cache.",
750836 "SampleAfterValue": "2000003",
751− "CounterHTOff": "0,1,2,3"
837+ "CounterHTOff": "0,1,2,3,4,5,6,7"
752838 },
753839 {
754− "Invert": "1",
755− "EventCode": "0xC2",
840+ "EventCode": "0xA3",
756841 "UMask": "0x1",
757− "BriefDescription": "Cycles with less than 10 actually retired uops.",
842+ "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
758843 "Counter": "0,1,2,3",
759− "EventName": "UOPS_RETIRED.TOTAL_CYCLES",
760− "CounterMask": "10",
761− "PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
844+ "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS",
845+ "CounterMask": "1",
762846 "SampleAfterValue": "2000003",
763− "CounterHTOff": "0,1,2,3"
847+ "CounterHTOff": "0,1,2,3,4,5,6,7"
764848 },
765849 {
766− "EventCode": "0xC3",
767− "UMask": "0x1",
768− "BriefDescription": "Cycles there was a Nuke. Account for both thread-specific and All Thread Nukes.",
850+ "EventCode": "0xA3",
851+ "UMask": "0x2",
852+ "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
769853 "Counter": "0,1,2,3",
770− "EventName": "MACHINE_CLEARS.CYCLES",
771− "PublicDescription": "This event counts both thread-specific (TS) and all-thread (AT) nukes.",
854+ "EventName": "CYCLE_ACTIVITY.CYCLES_LDM_PENDING",
855+ "CounterMask": "2",
856+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request (that is cycles with non-completed load waiting for its data from memory subsystem).",
772857 "SampleAfterValue": "2000003",
773858 "CounterHTOff": "0,1,2,3,4,5,6,7"
774859 },
775860 {
776− "EventCode": "0xC3",
861+ "EventCode": "0xA3",
862+ "UMask": "0x2",
863+ "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
864+ "Counter": "0,1,2,3",
865+ "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY",
866+ "CounterMask": "2",
867+ "SampleAfterValue": "2000003",
868+ "CounterHTOff": "0,1,2,3"
869+ },
870+ {
871+ "EventCode": "0xA3",
777872 "UMask": "0x4",
778− "BriefDescription": "Self-modifying code (SMC) detected.",
873+ "BriefDescription": "This event increments by 1 for every cycle where there was no execute for this thread.",
779874 "Counter": "0,1,2,3",
780− "EventName": "MACHINE_CLEARS.SMC",
781− "PublicDescription": "This event counts self-modifying code (SMC) detected, which causes a machine clear.",
782− "SampleAfterValue": "100003",
783− "CounterHTOff": "0,1,2,3,4,5,6,7"
875+ "EventName": "CYCLE_ACTIVITY.CYCLES_NO_EXECUTE",
876+ "CounterMask": "4",
877+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port.",
878+ "SampleAfterValue": "2000003",
879+ "CounterHTOff": "0,1,2,3"
784880 },
785881 {
786− "EventCode": "0xC3",
787− "UMask": "0x20",
788− "BriefDescription": "This event counts the number of executed Intel AVX masked load operations that refer to an illegal address range with the mask bits set to 0.",
882+ "EventCode": "0xA3",
883+ "UMask": "0x4",
884+ "BriefDescription": "Total execution stalls.",
789885 "Counter": "0,1,2,3",
790− "EventName": "MACHINE_CLEARS.MASKMOV",
791− "PublicDescription": "Maskmov false fault - counts number of time ucode passes through Maskmov flow due to instruction's mask being 0 while the flow was completed without raising a fault.",
792− "SampleAfterValue": "100003",
886+ "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL",
887+ "CounterMask": "4",
888+ "SampleAfterValue": "2000003",
793889 "CounterHTOff": "0,1,2,3,4,5,6,7"
794890 },
795891 {
796− "EventCode": "0xC4",
797− "UMask": "0x1",
798− "BriefDescription": "Conditional branch instructions retired.",
799− "PEBS": "1",
800− "Counter": "0,1,2,3",
801− "EventName": "BR_INST_RETIRED.CONDITIONAL",
802− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts conditional branch instructions retired.",
803− "SampleAfterValue": "400009",
804− "CounterHTOff": "0,1,2,3,4,5,6,7"
805− },
806− {
807− "EventCode": "0xC4",
808− "UMask": "0x2",
809− "BriefDescription": "Direct and indirect near call instructions retired.",
810− "PEBS": "1",
811− "Counter": "0,1,2,3",
812− "EventName": "BR_INST_RETIRED.NEAR_CALL",
813− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts both direct and indirect near call instructions retired.",
814− "SampleAfterValue": "100007",
815− "CounterHTOff": "0,1,2,3,4,5,6,7"
816− },
817− {
818− "EventCode": "0xC4",
819− "UMask": "0x0",
820− "BriefDescription": "All (macro) branch instructions retired.",
821− "Counter": "0,1,2,3",
822− "EventName": "BR_INST_RETIRED.ALL_BRANCHES",
823− "PublicDescription": "This event counts all (macro) branch instructions retired.",
824− "SampleAfterValue": "400009",
825− "CounterHTOff": "0,1,2,3,4,5,6,7"
826− },
827− {
828− "EventCode": "0xC4",
829− "UMask": "0x8",
830− "BriefDescription": "Return instructions retired.",
831− "PEBS": "1",
832− "Counter": "0,1,2,3",
833− "EventName": "BR_INST_RETIRED.NEAR_RETURN",
834− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts return instructions retired.",
835− "SampleAfterValue": "100007",
836− "CounterHTOff": "0,1,2,3,4,5,6,7"
837− },
838− {
839− "EventCode": "0xC4",
840− "UMask": "0x10",
841− "BriefDescription": "Not taken branch instructions retired.",
842− "Counter": "0,1,2,3",
843− "EventName": "BR_INST_RETIRED.NOT_TAKEN",
844− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts not taken branch instructions retired.",
845− "SampleAfterValue": "400009",
846− "CounterHTOff": "0,1,2,3,4,5,6,7"
847− },
848− {
849− "EventCode": "0xC4",
850− "UMask": "0x20",
851− "BriefDescription": "Taken branch instructions retired.",
852− "PEBS": "1",
853− "Counter": "0,1,2,3",
854− "EventName": "BR_INST_RETIRED.NEAR_TAKEN",
855− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts taken branch instructions retired.",
856− "SampleAfterValue": "400009",
857− "CounterHTOff": "0,1,2,3,4,5,6,7"
858− },
859− {
860− "EventCode": "0xC4",
861− "UMask": "0x40",
862− "BriefDescription": "Far branch instructions retired.",
863− "Counter": "0,1,2,3",
864− "EventName": "BR_INST_RETIRED.FAR_BRANCH",
865− "Errata": "BDW98",
866− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts far branch instructions retired.",
867− "SampleAfterValue": "100007",
868− "CounterHTOff": "0,1,2,3,4,5,6,7"
869− },
870− {
871− "EventCode": "0xC4",
872− "UMask": "0x4",
873− "BriefDescription": "All (macro) branch instructions retired. (Precise Event - PEBS)",
874− "PEBS": "2",
892+ "EventCode": "0xA3",
893+ "UMask": "0x5",
894+ "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
875895 "Counter": "0,1,2,3",
876− "EventName": "BR_INST_RETIRED.ALL_BRANCHES_PEBS",
877− "Errata": "BDW98",
878− "PublicDescription": "This is a precise version of BR_INST_RETIRED.ALL_BRANCHES that counts all (macro) branch instructions retired.",
879− "SampleAfterValue": "400009",
896+ "EventName": "CYCLE_ACTIVITY.STALLS_L2_PENDING",
897+ "CounterMask": "5",
898+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand* load request missing the L2 cache.(as a footprint) * includes also L1 HW prefetch requests that may or may not be required by demands.",
899+ "SampleAfterValue": "2000003",
880900 "CounterHTOff": "0,1,2,3"
881901 },
882902 {
883− "EventCode": "0xC5",
884− "UMask": "0x1",
885− "BriefDescription": "Mispredicted conditional branch instructions retired.",
886− "PEBS": "1",
887− "Counter": "0,1,2,3",
888− "EventName": "BR_MISP_RETIRED.CONDITIONAL",
889− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted conditional branch instructions retired.",
890− "SampleAfterValue": "400009",
891− "CounterHTOff": "0,1,2,3,4,5,6,7"
892− },
893− {
894− "EventCode": "0xC5",
895− "UMask": "0x0",
896− "BriefDescription": "All mispredicted macro branch instructions retired.",
897− "Counter": "0,1,2,3",
898− "EventName": "BR_MISP_RETIRED.ALL_BRANCHES",
899− "PublicDescription": "This event counts all mispredicted macro branch instructions retired.",
900− "SampleAfterValue": "400009",
901− "CounterHTOff": "0,1,2,3,4,5,6,7"
902− },
903− {
904− "EventCode": "0xC5",
905− "UMask": "0x8",
906− "BriefDescription": "This event counts the number of mispredicted ret instructions retired. Non PEBS",
907− "PEBS": "1",
903+ "EventCode": "0xA3",
904+ "UMask": "0x5",
905+ "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
908906 "Counter": "0,1,2,3",
909− "EventName": "BR_MISP_RETIRED.RET",
910− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted return instructions retired.",
911− "SampleAfterValue": "100007",
907+ "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS",
908+ "CounterMask": "5",
909+ "SampleAfterValue": "2000003",
912910 "CounterHTOff": "0,1,2,3,4,5,6,7"
913911 },
914912 {
915− "EventCode": "0xC5",
916− "UMask": "0x4",
917− "BriefDescription": "Mispredicted macro branch instructions retired. (Precise Event - PEBS)",
918− "PEBS": "2",
913+ "EventCode": "0xA3",
914+ "UMask": "0x6",
915+ "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
919916 "Counter": "0,1,2,3",
920− "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_PEBS",
921− "PublicDescription": "This is a precise version of BR_MISP_RETIRED.ALL_BRANCHES that counts all mispredicted macro branch instructions retired.",
922− "SampleAfterValue": "400009",
917+ "EventName": "CYCLE_ACTIVITY.STALLS_LDM_PENDING",
918+ "CounterMask": "6",
919+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request.",
920+ "SampleAfterValue": "2000003",
923921 "CounterHTOff": "0,1,2,3"
924922 },
925923 {
926− "EventCode": "0xCC",
927− "UMask": "0x20",
928− "BriefDescription": "Count cases of saving new LBR",
924+ "EventCode": "0xA3",
925+ "UMask": "0x6",
926+ "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
929927 "Counter": "0,1,2,3",
930− "EventName": "ROB_MISC_EVENTS.LBR_INSERTS",
931− "PublicDescription": "This event counts cases of saving new LBR records by hardware. This assumes proper enabling of LBRs and takes into account LBR filtering done by the LBR_SELECT register.",
928+ "EventName": "CYCLE_ACTIVITY.STALLS_MEM_ANY",
929+ "CounterMask": "6",
932930 "SampleAfterValue": "2000003",
933931 "CounterHTOff": "0,1,2,3,4,5,6,7"
934932 },
935933 {
936− "EventCode": "0x3C",
937− "UMask": "0x0",
938− "BriefDescription": "Thread cycles when thread is not in halt state",
939− "Counter": "0,1,2,3",
940− "EventName": "CPU_CLK_UNHALTED.THREAD_P",
941− "PublicDescription": "This is an architectural event that counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling. For this reason, this event may have a changing ratio with regards to wall clock time.",
934+ "EventCode": "0xA3",
935+ "UMask": "0x8",
936+ "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
937+ "Counter": "2",
938+ "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_PENDING",
939+ "CounterMask": "8",
940+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request missing the L1 data cache.",
942941 "SampleAfterValue": "2000003",
943− "CounterHTOff": "0,1,2,3,4,5,6,7"
944− },
945− {
946− "EventCode": "0x89",
947− "UMask": "0xa0",
948− "BriefDescription": "Taken speculative and retired mispredicted indirect calls.",
949− "Counter": "0,1,2,3",
950− "EventName": "BR_MISP_EXEC.TAKEN_INDIRECT_NEAR_CALL",
951− "SampleAfterValue": "200003",
952− "CounterHTOff": "0,1,2,3,4,5,6,7"
942+ "CounterHTOff": "2"
953943 },
954944 {
955− "EventCode": "0xA1",
956− "UMask": "0x1",
957− "BriefDescription": "Cycles per core when uops are exectuted in port 0.",
958− "Counter": "0,1,2,3",
959− "EventName": "UOPS_EXECUTED_PORT.PORT_0_CORE",
960− "AnyThread": "1",
945+ "EventCode": "0xA3",
946+ "UMask": "0x8",
947+ "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
948+ "Counter": "2",
949+ "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS",
950+ "CounterMask": "8",
961951 "SampleAfterValue": "2000003",
962− "CounterHTOff": "0,1,2,3,4,5,6,7"
952+ "CounterHTOff": "2"
963953 },
964954 {
965− "EventCode": "0xA1",
966− "UMask": "0x2",
967− "BriefDescription": "Cycles per core when uops are exectuted in port 1.",
968− "Counter": "0,1,2,3",
969− "EventName": "UOPS_EXECUTED_PORT.PORT_1_CORE",
970− "AnyThread": "1",
955+ "EventCode": "0xA3",
956+ "UMask": "0xc",
957+ "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
958+ "Counter": "2",
959+ "EventName": "CYCLE_ACTIVITY.STALLS_L1D_PENDING",
960+ "CounterMask": "12",
961+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request missing the L1 data cache.",
971962 "SampleAfterValue": "2000003",
972− "CounterHTOff": "0,1,2,3,4,5,6,7"
963+ "CounterHTOff": "2"
973964 },
974965 {
975− "EventCode": "0xA1",
976− "UMask": "0x4",
977− "BriefDescription": "Cycles per core when uops are dispatched to port 2.",
978− "Counter": "0,1,2,3",
979− "EventName": "UOPS_EXECUTED_PORT.PORT_2_CORE",
980− "AnyThread": "1",
966+ "EventCode": "0xA3",
967+ "UMask": "0xc",
968+ "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
969+ "Counter": "2",
970+ "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS",
971+ "CounterMask": "12",
981972 "SampleAfterValue": "2000003",
982− "CounterHTOff": "0,1,2,3,4,5,6,7"
973+ "CounterHTOff": "2"
983974 },
984975 {
985− "EventCode": "0xA1",
986− "UMask": "0x8",
987− "BriefDescription": "Cycles per core when uops are dispatched to port 3.",
976+ "EventCode": "0xA8",
977+ "UMask": "0x1",
978+ "BriefDescription": "Number of Uops delivered by the LSD.",
988979 "Counter": "0,1,2,3",
989− "EventName": "UOPS_EXECUTED_PORT.PORT_3_CORE",
990− "AnyThread": "1",
980+ "EventName": "LSD.UOPS",
991981 "SampleAfterValue": "2000003",
992982 "CounterHTOff": "0,1,2,3,4,5,6,7"
993983 },
994984 {
995− "EventCode": "0xA1",
996− "UMask": "0x10",
997− "BriefDescription": "Cycles per core when uops are exectuted in port 4.",
985+ "EventCode": "0xA8",
986+ "UMask": "0x1",
987+ "BriefDescription": "Cycles 4 Uops delivered by the LSD, but didn't come from the decoder.",
998988 "Counter": "0,1,2,3",
999− "EventName": "UOPS_EXECUTED_PORT.PORT_4_CORE",
1000− "AnyThread": "1",
989+ "EventName": "LSD.CYCLES_4_UOPS",
990+ "CounterMask": "4",
1001991 "SampleAfterValue": "2000003",
1002992 "CounterHTOff": "0,1,2,3,4,5,6,7"
1003993 },
1004994 {
1005− "EventCode": "0xA1",
1006− "UMask": "0x20",
1007− "BriefDescription": "Cycles per core when uops are exectuted in port 5.",
995+ "EventCode": "0xA8",
996+ "UMask": "0x1",
997+ "BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
1008998 "Counter": "0,1,2,3",
1009− "EventName": "UOPS_EXECUTED_PORT.PORT_5_CORE",
1010− "AnyThread": "1",
999+ "EventName": "LSD.CYCLES_ACTIVE",
1000+ "CounterMask": "1",
10111001 "SampleAfterValue": "2000003",
10121002 "CounterHTOff": "0,1,2,3,4,5,6,7"
10131003 },
10141004 {
1015− "EventCode": "0xA1",
1016− "UMask": "0x40",
1017− "BriefDescription": "Cycles per core when uops are exectuted in port 6.",
1005+ "EventCode": "0xB1",
1006+ "UMask": "0x1",
1007+ "BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
10181008 "Counter": "0,1,2,3",
1019− "EventName": "UOPS_EXECUTED_PORT.PORT_6_CORE",
1020− "AnyThread": "1",
1009+ "EventName": "UOPS_EXECUTED.THREAD",
1010+ "PublicDescription": "Number of uops to be executed per-thread each cycle.",
10211011 "SampleAfterValue": "2000003",
10221012 "CounterHTOff": "0,1,2,3,4,5,6,7"
10231013 },
10241014 {
1025− "EventCode": "0xA1",
1026− "UMask": "0x80",
1027− "BriefDescription": "Cycles per core when uops are dispatched to port 7.",
1015+ "Invert": "1",
1016+ "EventCode": "0xB1",
1017+ "UMask": "0x1",
1018+ "BriefDescription": "Counts number of cycles no uops were dispatched to be executed on this thread.",
10281019 "Counter": "0,1,2,3",
1029− "EventName": "UOPS_EXECUTED_PORT.PORT_7_CORE",
1030− "AnyThread": "1",
1020+ "EventName": "UOPS_EXECUTED.STALL_CYCLES",
1021+ "CounterMask": "1",
1022+ "PublicDescription": "This event counts cycles during which no uops were dispatched from the Reservation Station (RS) per thread.",
10311023 "SampleAfterValue": "2000003",
1032− "CounterHTOff": "0,1,2,3,4,5,6,7"
1033− },
1034− {
1035− "EventCode": "0xC5",
1036− "UMask": "0x20",
1037− "BriefDescription": "number of near branch instructions retired that were mispredicted and taken.",
1038− "PEBS": "1",
1039− "Counter": "0,1,2,3",
1040− "EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
1041− "PublicDescription": "Number of near branch instructions retired that were mispredicted and taken.",
1042− "SampleAfterValue": "400009",
1043− "CounterHTOff": "0,1,2,3,4,5,6,7"
1024+ "CounterHTOff": "0,1,2,3"
10441025 },
10451026 {
10461027 "EventCode": "0xB1",
@@ -1083,335 +1064,364 @@ · 36 unmodified lines
10831064 "CounterHTOff": "0,1,2,3"
10841065 },
10851066 {
1086− "EventCode": "0xe6",
1087− "UMask": "0x1f",
1088− "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.",
1067+ "EventCode": "0xB1",
1068+ "UMask": "0x2",
1069+ "BriefDescription": "Number of uops executed on the core.",
10891070 "Counter": "0,1,2,3",
1090− "EventName": "BACLEARS.ANY",
1091− "SampleAfterValue": "100003",
1092− "CounterHTOff": "0,1,2,3,4,5,6,7"
1093− },
1094− {
1095− "EventCode": "0xA3",
1096− "UMask": "0x8",
1097− "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
1098− "Counter": "2",
1099− "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS",
1100− "CounterMask": "8",
1071+ "EventName": "UOPS_EXECUTED.CORE",
1072+ "PublicDescription": "Number of uops executed from any thread.",
11011073 "SampleAfterValue": "2000003",
1102− "CounterHTOff": "2"
1074+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11031075 },
11041076 {
1105− "EventCode": "0xA3",
1106− "UMask": "0x1",
1107− "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
1077+ "EventCode": "0xb1",
1078+ "UMask": "0x2",
1079+ "BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
11081080 "Counter": "0,1,2,3",
1109− "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS",
1081+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
11101082 "CounterMask": "1",
11111083 "SampleAfterValue": "2000003",
11121084 "CounterHTOff": "0,1,2,3,4,5,6,7"
11131085 },
11141086 {
1115− "EventCode": "0xA3",
1087+ "EventCode": "0xb1",
11161088 "UMask": "0x2",
1117− "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
1089+ "BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
11181090 "Counter": "0,1,2,3",
1119− "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY",
1091+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
11201092 "CounterMask": "2",
11211093 "SampleAfterValue": "2000003",
1122− "CounterHTOff": "0,1,2,3"
1094+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11231095 },
11241096 {
1125− "EventCode": "0xA3",
1126− "UMask": "0x4",
1127− "BriefDescription": "Total execution stalls.",
1097+ "EventCode": "0xb1",
1098+ "UMask": "0x2",
1099+ "BriefDescription": "Cycles at least 3 micro-op is executed from any thread on physical core.",
11281100 "Counter": "0,1,2,3",
1129− "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL",
1130− "CounterMask": "4",
1101+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3",
1102+ "CounterMask": "3",
11311103 "SampleAfterValue": "2000003",
11321104 "CounterHTOff": "0,1,2,3,4,5,6,7"
11331105 },
11341106 {
1135− "EventCode": "0xA3",
1136− "UMask": "0xc",
1137− "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
1138− "Counter": "2",
1139− "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS",
1140− "CounterMask": "12",
1107+ "EventCode": "0xb1",
1108+ "UMask": "0x2",
1109+ "BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
1110+ "Counter": "0,1,2,3",
1111+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
1112+ "CounterMask": "4",
11411113 "SampleAfterValue": "2000003",
1142− "CounterHTOff": "2"
1114+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11431115 },
11441116 {
1145− "EventCode": "0xA3",
1146− "UMask": "0x5",
1147− "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
1117+ "Invert": "1",
1118+ "EventCode": "0xb1",
1119+ "UMask": "0x2",
1120+ "BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
11481121 "Counter": "0,1,2,3",
1149− "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS",
1150− "CounterMask": "5",
1122+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
11511123 "SampleAfterValue": "2000003",
11521124 "CounterHTOff": "0,1,2,3,4,5,6,7"
11531125 },
11541126 {
1155− "EventCode": "0xA3",
1156− "UMask": "0x6",
1157− "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
1127+ "EventCode": "0xC0",
1128+ "UMask": "0x0",
1129+ "BriefDescription": "Number of instructions retired. General Counter - architectural event",
11581130 "Counter": "0,1,2,3",
1159− "EventName": "CYCLE_ACTIVITY.STALLS_MEM_ANY",
1160− "CounterMask": "6",
1131+ "EventName": "INST_RETIRED.ANY_P",
1132+ "Errata": "BDM61",
1133+ "PublicDescription": "This event counts the number of instructions (EOMs) retired. Counting covers macro-fused instructions individually (that is, increments by two).",
11611134 "SampleAfterValue": "2000003",
11621135 "CounterHTOff": "0,1,2,3,4,5,6,7"
11631136 },
11641137 {
1165− "EdgeDetect": "1",
1166− "EventCode": "0xC3",
1138+ "EventCode": "0xC0",
11671139 "UMask": "0x1",
1168− "BriefDescription": "Number of machine clears (nukes) of any type.",
1140+ "BriefDescription": "Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution",
1141+ "PEBS": "2",
1142+ "Counter": "1",
1143+ "EventName": "INST_RETIRED.PREC_DIST",
1144+ "Errata": "BDM11, BDM55",
1145+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts instructions retired.",
1146+ "SampleAfterValue": "2000003",
1147+ "CounterHTOff": "1"
1148+ },
1149+ {
1150+ "EventCode": "0xC0",
1151+ "UMask": "0x2",
1152+ "BriefDescription": "FP operations retired. X87 FP operations that have no exceptions:",
11691153 "Counter": "0,1,2,3",
1170− "EventName": "MACHINE_CLEARS.COUNT",
1171− "CounterMask": "1",
1172− "SampleAfterValue": "100003",
1154+ "EventName": "INST_RETIRED.X87",
1155+ "PublicDescription": "This event counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
1156+ "SampleAfterValue": "2000003",
11731157 "CounterHTOff": "0,1,2,3,4,5,6,7"
11741158 },
11751159 {
1176− "EventCode": "0xA8",
1177− "UMask": "0x1",
1178− "BriefDescription": "Cycles 4 Uops delivered by the LSD, but didn't come from the decoder.",
1160+ "EventCode": "0xC1",
1161+ "UMask": "0x40",
1162+ "BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
11791163 "Counter": "0,1,2,3",
1180− "EventName": "LSD.CYCLES_4_UOPS",
1181− "CounterMask": "4",
1182− "SampleAfterValue": "2000003",
1164+ "EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
1165+ "SampleAfterValue": "100003",
11831166 "CounterHTOff": "0,1,2,3,4,5,6,7"
11841167 },
11851168 {
1186− "EdgeDetect": "1",
1187− "Invert": "1",
1188− "EventCode": "0x5E",
1169+ "EventCode": "0xC2",
11891170 "UMask": "0x1",
1190− "BriefDescription": "Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues.",
1171+ "BriefDescription": "Actually retired uops. (Precise Event - PEBS)",
1172+ "Data_LA": "1",
1173+ "PEBS": "1",
11911174 "Counter": "0,1,2,3",
1192− "EventName": "RS_EVENTS.EMPTY_END",
1193− "CounterMask": "1",
1194− "SampleAfterValue": "200003",
1175+ "EventName": "UOPS_RETIRED.ALL",
1176+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
1177+ "SampleAfterValue": "2000003",
11951178 "CounterHTOff": "0,1,2,3,4,5,6,7"
11961179 },
11971180 {
1198− "EventCode": "0xA8",
1181+ "Invert": "1",
1182+ "EventCode": "0xC2",
11991183 "UMask": "0x1",
1200− "BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
1184+ "BriefDescription": "Cycles without actually retired uops.",
12011185 "Counter": "0,1,2,3",
1202− "EventName": "LSD.CYCLES_ACTIVE",
1186+ "EventName": "UOPS_RETIRED.STALL_CYCLES",
12031187 "CounterMask": "1",
1188+ "PublicDescription": "This event counts cycles without actually retired uops.",
12041189 "SampleAfterValue": "2000003",
1205− "CounterHTOff": "0,1,2,3,4,5,6,7"
1190+ "CounterHTOff": "0,1,2,3"
12061191 },
12071192 {
1208− "EventCode": "0xA1",
1193+ "Invert": "1",
1194+ "EventCode": "0xC2",
12091195 "UMask": "0x1",
1210− "BriefDescription": "Cycles per thread when uops are executed in port 0",
1196+ "BriefDescription": "Cycles with less than 10 actually retired uops.",
12111197 "Counter": "0,1,2,3",
1212− "EventName": "UOPS_EXECUTED_PORT.PORT_0",
1213− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
1198+ "EventName": "UOPS_RETIRED.TOTAL_CYCLES",
1199+ "CounterMask": "10",
1200+ "PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
12141201 "SampleAfterValue": "2000003",
1215− "CounterHTOff": "0,1,2,3,4,5,6,7"
1202+ "CounterHTOff": "0,1,2,3"
12161203 },
12171204 {
1218− "EventCode": "0xA1",
1205+ "EventCode": "0xC2",
12191206 "UMask": "0x2",
1220− "BriefDescription": "Cycles per thread when uops are executed in port 1",
1207+ "BriefDescription": "Retirement slots used. (Precise Event - PEBS)",
1208+ "PEBS": "1",
12211209 "Counter": "0,1,2,3",
1222− "EventName": "UOPS_EXECUTED_PORT.PORT_1",
1223− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
1210+ "EventName": "UOPS_RETIRED.RETIRE_SLOTS",
1211+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of retirement slots used.",
12241212 "SampleAfterValue": "2000003",
12251213 "CounterHTOff": "0,1,2,3,4,5,6,7"
12261214 },
12271215 {
1228− "EventCode": "0xA1",
1229− "UMask": "0x4",
1230− "BriefDescription": "Cycles per thread when uops are executed in port 2",
1216+ "EventCode": "0xC3",
1217+ "UMask": "0x1",
1218+ "BriefDescription": "Cycles there was a Nuke. Account for both thread-specific and All Thread Nukes.",
12311219 "Counter": "0,1,2,3",
1232− "EventName": "UOPS_EXECUTED_PORT.PORT_2",
1233− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
1220+ "EventName": "MACHINE_CLEARS.CYCLES",
1221+ "PublicDescription": "This event counts both thread-specific (TS) and all-thread (AT) nukes.",
12341222 "SampleAfterValue": "2000003",
12351223 "CounterHTOff": "0,1,2,3,4,5,6,7"
12361224 },
12371225 {
1238− "EventCode": "0xA1",
1239− "UMask": "0x8",
1240− "BriefDescription": "Cycles per thread when uops are executed in port 3",
1226+ "EdgeDetect": "1",
1227+ "EventCode": "0xC3",
1228+ "UMask": "0x1",
1229+ "BriefDescription": "Number of machine clears (nukes) of any type.",
12411230 "Counter": "0,1,2,3",
1242− "EventName": "UOPS_EXECUTED_PORT.PORT_3",
1243− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
1244− "SampleAfterValue": "2000003",
1231+ "EventName": "MACHINE_CLEARS.COUNT",
1232+ "CounterMask": "1",
1233+ "SampleAfterValue": "100003",
12451234 "CounterHTOff": "0,1,2,3,4,5,6,7"
12461235 },
12471236 {
1248− "EventCode": "0xA1",
1249− "UMask": "0x10",
1250− "BriefDescription": "Cycles per thread when uops are executed in port 4",
1237+ "EventCode": "0xC3",
1238+ "UMask": "0x4",
1239+ "BriefDescription": "Self-modifying code (SMC) detected.",
12511240 "Counter": "0,1,2,3",
1252− "EventName": "UOPS_EXECUTED_PORT.PORT_4",
1253− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
1254− "SampleAfterValue": "2000003",
1241+ "EventName": "MACHINE_CLEARS.SMC",
1242+ "PublicDescription": "This event counts self-modifying code (SMC) detected, which causes a machine clear.",
1243+ "SampleAfterValue": "100003",
12551244 "CounterHTOff": "0,1,2,3,4,5,6,7"
12561245 },
12571246 {
1258− "EventCode": "0xA1",
1247+ "EventCode": "0xC3",
12591248 "UMask": "0x20",
1260− "BriefDescription": "Cycles per thread when uops are executed in port 5",
1249+ "BriefDescription": "This event counts the number of executed Intel AVX masked load operations that refer to an illegal address range with the mask bits set to 0.",
12611250 "Counter": "0,1,2,3",
1262− "EventName": "UOPS_EXECUTED_PORT.PORT_5",
1263− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
1264− "SampleAfterValue": "2000003",
1251+ "EventName": "MACHINE_CLEARS.MASKMOV",
1252+ "PublicDescription": "Maskmov false fault - counts number of time ucode passes through Maskmov flow due to instruction's mask being 0 while the flow was completed without raising a fault.",
1253+ "SampleAfterValue": "100003",
12651254 "CounterHTOff": "0,1,2,3,4,5,6,7"
12661255 },
12671256 {
1268− "EventCode": "0xA1",
1269− "UMask": "0x40",
1270− "BriefDescription": "Cycles per thread when uops are executed in port 6",
1257+ "EventCode": "0xC4",
1258+ "UMask": "0x0",
1259+ "BriefDescription": "All (macro) branch instructions retired.",
12711260 "Counter": "0,1,2,3",
1272− "EventName": "UOPS_EXECUTED_PORT.PORT_6",
1273− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
1274− "SampleAfterValue": "2000003",
1261+ "EventName": "BR_INST_RETIRED.ALL_BRANCHES",
1262+ "PublicDescription": "This event counts all (macro) branch instructions retired.",
1263+ "SampleAfterValue": "400009",
12751264 "CounterHTOff": "0,1,2,3,4,5,6,7"
12761265 },
12771266 {
1278− "EventCode": "0xA1",
1279− "UMask": "0x80",
1280− "BriefDescription": "Cycles per thread when uops are executed in port 7",
1267+ "EventCode": "0xC4",
1268+ "UMask": "0x1",
1269+ "BriefDescription": "Conditional branch instructions retired. (Precise Event - PEBS)",
1270+ "PEBS": "1",
12811271 "Counter": "0,1,2,3",
1282− "EventName": "UOPS_EXECUTED_PORT.PORT_7",
1283− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
1284− "SampleAfterValue": "2000003",
1272+ "EventName": "BR_INST_RETIRED.CONDITIONAL",
1273+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts conditional branch instructions retired.",
1274+ "SampleAfterValue": "400009",
12851275 "CounterHTOff": "0,1,2,3,4,5,6,7"
12861276 },
12871277 {
1288− "EventCode": "0xA0",
1289− "UMask": "0x3",
1290− "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports",
1278+ "EventCode": "0xC4",
1279+ "UMask": "0x2",
1280+ "BriefDescription": "Direct and indirect near call instructions retired. (Precise Event - PEBS)",
1281+ "PEBS": "1",
12911282 "Counter": "0,1,2,3",
1292− "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF",
1293− "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
1294− "SampleAfterValue": "2000003",
1295− "CounterHTOff": "0,1,2,3"
1283+ "EventName": "BR_INST_RETIRED.NEAR_CALL",
1284+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect near call instructions retired.",
1285+ "SampleAfterValue": "100007",
1286+ "CounterHTOff": "0,1,2,3,4,5,6,7"
12961287 },
12971288 {
1298− "EventCode": "0x00",
1289+ "EventCode": "0xC4",
12991290 "UMask": "0x2",
1300− "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
1301− "Counter": "Fixed counter 2",
1302− "EventName": "CPU_CLK_UNHALTED.THREAD_ANY",
1303− "AnyThread": "1",
1304− "SampleAfterValue": "2000003",
1305− "CounterHTOff": "Fixed counter 2"
1291+ "BriefDescription": "Direct and indirect macro near call instructions retired (captured in ring 3). (Precise Event - PEBS)",
1292+ "PEBS": "1",
1293+ "Counter": "0,1,2,3",
1294+ "EventName": "BR_INST_RETIRED.NEAR_CALL_R3",
1295+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect macro near call instructions retired (captured in ring 3).",
1296+ "SampleAfterValue": "100007",
1297+ "CounterHTOff": "0,1,2,3,4,5,6,7"
13061298 },
13071299 {
1308− "EventCode": "0x3C",
1309− "UMask": "0x0",
1310− "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
1300+ "EventCode": "0xC4",
1301+ "UMask": "0x4",
1302+ "BriefDescription": "All (macro) branch instructions retired. (Precise Event - PEBS)",
1303+ "PEBS": "2",
13111304 "Counter": "0,1,2,3",
1312− "EventName": "CPU_CLK_UNHALTED.THREAD_P_ANY",
1313− "AnyThread": "1",
1314− "SampleAfterValue": "2000003",
1315− "CounterHTOff": "0,1,2,3,4,5,6,7"
1305+ "EventName": "BR_INST_RETIRED.ALL_BRANCHES_PEBS",
1306+ "Errata": "BDW98",
1307+ "PublicDescription": "This is a precise version of BR_INST_RETIRED.ALL_BRANCHES that counts all (macro) branch instructions retired.",
1308+ "SampleAfterValue": "400009",
1309+ "CounterHTOff": "0,1,2,3"
13161310 },
13171311 {
1318− "EventCode": "0x3C",
1319− "UMask": "0x1",
1320− "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
1312+ "EventCode": "0xC4",
1313+ "UMask": "0x8",
1314+ "BriefDescription": "Return instructions retired. (Precise Event - PEBS)",
1315+ "PEBS": "1",
13211316 "Counter": "0,1,2,3",
1322− "EventName": "CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY",
1323− "AnyThread": "1",
1324− "SampleAfterValue": "2000003",
1317+ "EventName": "BR_INST_RETIRED.NEAR_RETURN",
1318+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts return instructions retired.",
1319+ "SampleAfterValue": "100007",
13251320 "CounterHTOff": "0,1,2,3,4,5,6,7"
13261321 },
13271322 {
1328− "EventCode": "0x0D",
1329− "UMask": "0x3",
1330− "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for any thread running on the physical core (e.g. misprediction or memory nuke).",
1323+ "EventCode": "0xC4",
1324+ "UMask": "0x10",
1325+ "BriefDescription": "Not taken branch instructions retired.",
13311326 "Counter": "0,1,2,3",
1332− "EventName": "INT_MISC.RECOVERY_CYCLES_ANY",
1333− "AnyThread": "1",
1334− "CounterMask": "1",
1335− "SampleAfterValue": "2000003",
1327+ "EventName": "BR_INST_RETIRED.NOT_TAKEN",
1328+ "PublicDescription": "This event counts not taken branch instructions retired.",
1329+ "SampleAfterValue": "400009",
13361330 "CounterHTOff": "0,1,2,3,4,5,6,7"
13371331 },
13381332 {
1339− "EventCode": "0xb1",
1340− "UMask": "0x2",
1341− "BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
1333+ "EventCode": "0xC4",
1334+ "UMask": "0x20",
1335+ "BriefDescription": "Taken branch instructions retired. (Precise Event - PEBS)",
1336+ "PEBS": "1",
13421337 "Counter": "0,1,2,3",
1343− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
1344− "CounterMask": "1",
1345− "SampleAfterValue": "2000003",
1338+ "EventName": "BR_INST_RETIRED.NEAR_TAKEN",
1339+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts taken branch instructions retired.",
1340+ "SampleAfterValue": "400009",
13461341 "CounterHTOff": "0,1,2,3,4,5,6,7"
13471342 },
13481343 {
1349− "EventCode": "0xb1",
1350− "UMask": "0x2",
1351− "BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
1344+ "EventCode": "0xC4",
1345+ "UMask": "0x40",
1346+ "BriefDescription": "Far branch instructions retired.",
13521347 "Counter": "0,1,2,3",
1353− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
1354− "CounterMask": "2",
1355− "SampleAfterValue": "2000003",
1348+ "EventName": "BR_INST_RETIRED.FAR_BRANCH",
1349+ "Errata": "BDW98",
1350+ "PublicDescription": "This event counts far branch instructions retired.",
1351+ "SampleAfterValue": "100007",
13561352 "CounterHTOff": "0,1,2,3,4,5,6,7"
13571353 },
13581354 {
1359− "EventCode": "0xb1",
1360− "UMask": "0x2",
1361− "BriefDescription": "Cycles at least 3 micro-op is executed from any thread on physical core.",
1355+ "EventCode": "0xC5",
1356+ "UMask": "0x0",
1357+ "BriefDescription": "All mispredicted macro branch instructions retired.",
13621358 "Counter": "0,1,2,3",
1363− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3",
1364− "CounterMask": "3",
1365− "SampleAfterValue": "2000003",
1359+ "EventName": "BR_MISP_RETIRED.ALL_BRANCHES",
1360+ "PublicDescription": "This event counts all mispredicted macro branch instructions retired.",
1361+ "SampleAfterValue": "400009",
13661362 "CounterHTOff": "0,1,2,3,4,5,6,7"
13671363 },
13681364 {
1369− "EventCode": "0xb1",
1370− "UMask": "0x2",
1371− "BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
1365+ "EventCode": "0xC5",
1366+ "UMask": "0x1",
1367+ "BriefDescription": "Mispredicted conditional branch instructions retired. (Precise Event - PEBS)",
1368+ "PEBS": "1",
13721369 "Counter": "0,1,2,3",
1373− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
1374− "CounterMask": "4",
1375− "SampleAfterValue": "2000003",
1370+ "EventName": "BR_MISP_RETIRED.CONDITIONAL",
1371+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted conditional branch instructions retired.",
1372+ "SampleAfterValue": "400009",
13761373 "CounterHTOff": "0,1,2,3,4,5,6,7"
13771374 },
13781375 {
1379− "Invert": "1",
1380− "EventCode": "0xb1",
1381− "UMask": "0x2",
1382− "BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
1376+ "EventCode": "0xC5",
1377+ "UMask": "0x4",
1378+ "BriefDescription": "Mispredicted macro branch instructions retired. (Precise Event - PEBS)",
1379+ "PEBS": "2",
13831380 "Counter": "0,1,2,3",
1384− "EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
1385− "SampleAfterValue": "2000003",
1381+ "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_PEBS",
1382+ "PublicDescription": "This is a precise version of BR_MISP_RETIRED.ALL_BRANCHES that counts all mispredicted macro branch instructions retired.",
1383+ "SampleAfterValue": "400009",
1384+ "CounterHTOff": "0,1,2,3"
1385+ },
1386+ {
1387+ "EventCode": "0xC5",
1388+ "UMask": "0x8",
1389+ "BriefDescription": "This event counts the number of mispredicted ret instructions retired.(Precise Event)",
1390+ "PEBS": "1",
1391+ "Counter": "0,1,2,3",
1392+ "EventName": "BR_MISP_RETIRED.RET",
1393+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted return instructions retired.",
1394+ "SampleAfterValue": "100007",
13861395 "CounterHTOff": "0,1,2,3,4,5,6,7"
13871396 },
13881397 {
1389− "EventCode": "0x3C",
1390− "UMask": "0x1",
1391− "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
1398+ "EventCode": "0xC5",
1399+ "UMask": "0x20",
1400+ "BriefDescription": "number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
1401+ "PEBS": "1",
13921402 "Counter": "0,1,2,3",
1393− "EventName": "CPU_CLK_UNHALTED.REF_XCLK",
1394− "PublicDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate).",
1395− "SampleAfterValue": "2000003",
1403+ "EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
1404+ "PublicDescription": "Number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
1405+ "SampleAfterValue": "400009",
13961406 "CounterHTOff": "0,1,2,3,4,5,6,7"
13971407 },
13981408 {
1399− "EventCode": "0x3C",
1400− "UMask": "0x1",
1401− "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
1409+ "EventCode": "0xCC",
1410+ "UMask": "0x20",
1411+ "BriefDescription": "Count cases of saving new LBR",
14021412 "Counter": "0,1,2,3",
1403− "EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
1404− "AnyThread": "1",
1413+ "EventName": "ROB_MISC_EVENTS.LBR_INSERTS",
1414+ "PublicDescription": "This event counts cases of saving new LBR records by hardware. This assumes proper enabling of LBRs and takes into account LBR filtering done by the LBR_SELECT register.",
14051415 "SampleAfterValue": "2000003",
14061416 "CounterHTOff": "0,1,2,3,4,5,6,7"
14071417 },
14081418 {
1409− "EventCode": "0x3C",
1410− "UMask": "0x2",
1411− "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
1419+ "EventCode": "0xe6",
1420+ "UMask": "0x1f",
1421+ "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.",
14121422 "Counter": "0,1,2,3",
1413− "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE",
1414− "SampleAfterValue": "2000003",
1423+ "EventName": "BACLEARS.ANY",
1424+ "SampleAfterValue": "100003",
14151425 "CounterHTOff": "0,1,2,3,4,5,6,7"
14161426 }
14171427 ]
tools/perf/pmu-events/arch/x86/broadwellde/virtual-memory.json+75−75
@@ -43,6 +43,16 @@ · 42 unmodified lines
4343 "SampleAfterValue": "2000003",
4444 "CounterHTOff": "0,1,2,3,4,5,6,7"
4545 },
46+ {
47+ "EventCode": "0x08",
48+ "UMask": "0xe",
49+ "BriefDescription": "Demand load Miss in all translation lookaside buffer (TLB) levels causes a page walk that completes of any page size.",
50+ "Counter": "0,1,2,3",
51+ "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED",
52+ "Errata": "BDM69",
53+ "SampleAfterValue": "100003",
54+ "CounterHTOff": "0,1,2,3,4,5,6,7"
55+ },
4656 {
4757 "EventCode": "0x08",
4858 "UMask": "0x10",
@@ -72,6 +82,15 @@ · 23 unmodified lines
7282 "SampleAfterValue": "2000003",
7383 "CounterHTOff": "0,1,2,3,4,5,6,7"
7484 },
85+ {
86+ "EventCode": "0x08",
87+ "UMask": "0x60",
88+ "BriefDescription": "Load operations that miss the first DTLB level but hit the second and do not cause page walks.",
89+ "Counter": "0,1,2,3",
90+ "EventName": "DTLB_LOAD_MISSES.STLB_HIT",
91+ "SampleAfterValue": "2000003",
92+ "CounterHTOff": "0,1,2,3,4,5,6,7"
93+ },
7594 {
7695 "EventCode": "0x49",
7796 "UMask": "0x1",
@@ -116,6 +135,16 @@ · 38 unmodified lines
116135 "SampleAfterValue": "100003",
117136 "CounterHTOff": "0,1,2,3,4,5,6,7"
118137 },
138+ {
139+ "EventCode": "0x49",
140+ "UMask": "0xe",
141+ "BriefDescription": "Store misses in all DTLB levels that cause completed page walks.",
142+ "Counter": "0,1,2,3",
143+ "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED",
144+ "Errata": "BDM69",
145+ "SampleAfterValue": "100003",
146+ "CounterHTOff": "0,1,2,3,4,5,6,7"
147+ },
119148 {
120149 "EventCode": "0x49",
121150 "UMask": "0x10",
@@ -145,6 +174,15 @@ · 23 unmodified lines
145174 "SampleAfterValue": "100003",
146175 "CounterHTOff": "0,1,2,3,4,5,6,7"
147176 },
177+ {
178+ "EventCode": "0x49",
179+ "UMask": "0x60",
180+ "BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
181+ "Counter": "0,1,2,3",
182+ "EventName": "DTLB_STORE_MISSES.STLB_HIT",
183+ "SampleAfterValue": "100003",
184+ "CounterHTOff": "0,1,2,3,4,5,6,7"
185+ },
148186 {
149187 "EventCode": "0x4F",
150188 "UMask": "0x10",
@@ -199,6 +237,16 @@ · 48 unmodified lines
199237 "SampleAfterValue": "100003",
200238 "CounterHTOff": "0,1,2,3,4,5,6,7"
201239 },
240+ {
241+ "EventCode": "0x85",
242+ "UMask": "0xe",
243+ "BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
244+ "Counter": "0,1,2,3",
245+ "EventName": "ITLB_MISSES.WALK_COMPLETED",
246+ "Errata": "BDM69",
247+ "SampleAfterValue": "100003",
248+ "CounterHTOff": "0,1,2,3,4,5,6,7"
249+ },
202250 {
203251 "EventCode": "0x85",
204252 "UMask": "0x10",
@@ -228,6 +276,15 @@ · 23 unmodified lines
228276 "SampleAfterValue": "100003",
229277 "CounterHTOff": "0,1,2,3,4,5,6,7"
230278 },
279+ {
280+ "EventCode": "0x85",
281+ "UMask": "0x60",
282+ "BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
283+ "Counter": "0,1,2,3",
284+ "EventName": "ITLB_MISSES.STLB_HIT",
285+ "SampleAfterValue": "100003",
286+ "CounterHTOff": "0,1,2,3,4,5,6,7"
287+ },
231288 {
232289 "EventCode": "0xAE",
233290 "UMask": "0x1",
@@ -250,60 +307,60 @@ · 16 unmodified lines
250307 },
251308 {
252309 "EventCode": "0xBC",
253− "UMask": "0x21",
254− "BriefDescription": "Number of ITLB page walker hits in the L1+FB.",
310+ "UMask": "0x12",
311+ "BriefDescription": "Number of DTLB page walker hits in the L2.",
255312 "Counter": "0,1,2,3",
256− "EventName": "PAGE_WALKER_LOADS.ITLB_L1",
313+ "EventName": "PAGE_WALKER_LOADS.DTLB_L2",
257314 "Errata": "BDM69, BDM98",
258315 "SampleAfterValue": "2000003",
259316 "CounterHTOff": "0,1,2,3"
260317 },
261318 {
262319 "EventCode": "0xBC",
263− "UMask": "0x12",
264− "BriefDescription": "Number of DTLB page walker hits in the L2.",
320+ "UMask": "0x14",
321+ "BriefDescription": "Number of DTLB page walker hits in the L3 + XSNP.",
265322 "Counter": "0,1,2,3",
266− "EventName": "PAGE_WALKER_LOADS.DTLB_L2",
323+ "EventName": "PAGE_WALKER_LOADS.DTLB_L3",
267324 "Errata": "BDM69, BDM98",
268325 "SampleAfterValue": "2000003",
269326 "CounterHTOff": "0,1,2,3"
270327 },
271328 {
272329 "EventCode": "0xBC",
273− "UMask": "0x22",
274− "BriefDescription": "Number of ITLB page walker hits in the L2.",
330+ "UMask": "0x18",
331+ "BriefDescription": "Number of DTLB page walker hits in Memory.",
275332 "Counter": "0,1,2,3",
276− "EventName": "PAGE_WALKER_LOADS.ITLB_L2",
333+ "EventName": "PAGE_WALKER_LOADS.DTLB_MEMORY",
277334 "Errata": "BDM69, BDM98",
278335 "SampleAfterValue": "2000003",
279336 "CounterHTOff": "0,1,2,3"
280337 },
281338 {
282339 "EventCode": "0xBC",
283− "UMask": "0x14",
284− "BriefDescription": "Number of DTLB page walker hits in the L3 + XSNP.",
340+ "UMask": "0x21",
341+ "BriefDescription": "Number of ITLB page walker hits in the L1+FB.",
285342 "Counter": "0,1,2,3",
286− "EventName": "PAGE_WALKER_LOADS.DTLB_L3",
343+ "EventName": "PAGE_WALKER_LOADS.ITLB_L1",
287344 "Errata": "BDM69, BDM98",
288345 "SampleAfterValue": "2000003",
289346 "CounterHTOff": "0,1,2,3"
290347 },
291348 {
292349 "EventCode": "0xBC",
293− "UMask": "0x24",
294− "BriefDescription": "Number of ITLB page walker hits in the L3 + XSNP.",
350+ "UMask": "0x22",
351+ "BriefDescription": "Number of ITLB page walker hits in the L2.",
295352 "Counter": "0,1,2,3",
296− "EventName": "PAGE_WALKER_LOADS.ITLB_L3",
353+ "EventName": "PAGE_WALKER_LOADS.ITLB_L2",
297354 "Errata": "BDM69, BDM98",
298355 "SampleAfterValue": "2000003",
299356 "CounterHTOff": "0,1,2,3"
300357 },
301358 {
302359 "EventCode": "0xBC",
303− "UMask": "0x18",
304− "BriefDescription": "Number of DTLB page walker hits in Memory.",
360+ "UMask": "0x24",
361+ "BriefDescription": "Number of ITLB page walker hits in the L3 + XSNP.",
305362 "Counter": "0,1,2,3",
306− "EventName": "PAGE_WALKER_LOADS.DTLB_MEMORY",
363+ "EventName": "PAGE_WALKER_LOADS.ITLB_L3",
307364 "Errata": "BDM69, BDM98",
308365 "SampleAfterValue": "2000003",
309366 "CounterHTOff": "0,1,2,3"
@@ -327,62 +384,5 @@ · 17 unmodified lines
327384 "PublicDescription": "This event counts the number of any STLB flush attempts (such as entire, VPID, PCID, InvPage, CR3 write, and so on).",
328385 "SampleAfterValue": "100007",
329386 "CounterHTOff": "0,1,2,3,4,5,6,7"
330− },
331− {
332− "EventCode": "0x08",
333− "UMask": "0xe",
334− "BriefDescription": "Demand load Miss in all translation lookaside buffer (TLB) levels causes a page walk that completes of any page size.",
335− "Counter": "0,1,2,3",
336− "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED",
337− "Errata": "BDM69",
338− "SampleAfterValue": "100003",
339− "CounterHTOff": "0,1,2,3,4,5,6,7"
340− },
341− {
342− "EventCode": "0x08",
343− "UMask": "0x60",
344− "BriefDescription": "Load operations that miss the first DTLB level but hit the second and do not cause page walks.",
345− "Counter": "0,1,2,3",
346− "EventName": "DTLB_LOAD_MISSES.STLB_HIT",
347− "SampleAfterValue": "2000003",
348− "CounterHTOff": "0,1,2,3,4,5,6,7"
349− },
350− {
351− "EventCode": "0x49",
352− "UMask": "0xe",
353− "BriefDescription": "Store misses in all DTLB levels that cause completed page walks.",
354− "Counter": "0,1,2,3",
355− "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED",
356− "Errata": "BDM69",
357− "SampleAfterValue": "100003",
358− "CounterHTOff": "0,1,2,3,4,5,6,7"
359− },
360− {
361− "EventCode": "0x49",
362− "UMask": "0x60",
363− "BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
364− "Counter": "0,1,2,3",
365− "EventName": "DTLB_STORE_MISSES.STLB_HIT",
366− "SampleAfterValue": "100003",
367− "CounterHTOff": "0,1,2,3,4,5,6,7"
368− },
369− {
370− "EventCode": "0x85",
371− "UMask": "0xe",
372− "BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
373− "Counter": "0,1,2,3",
374− "EventName": "ITLB_MISSES.WALK_COMPLETED",
375− "Errata": "BDM69",
376− "SampleAfterValue": "100003",
377− "CounterHTOff": "0,1,2,3,4,5,6,7"
378− },
379− {
380− "EventCode": "0x85",
381− "UMask": "0x60",
382− "BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
383− "Counter": "0,1,2,3",
384− "EventName": "ITLB_MISSES.STLB_HIT",
385− "SampleAfterValue": "100003",
386− "CounterHTOff": "0,1,2,3,4,5,6,7"
387387 }
388388 ]
tools/perf/pmu-events/arch/x86/broadwellx/cache.json+203−180
@@ -11,11 +11,28 @@ · 10 unmodified lines
1111 },
1212 {
1313 "EventCode": "0x24",
14− "UMask": "0x41",
15− "BriefDescription": "Demand Data Read requests that hit L2 cache",
14+ "UMask": "0x22",
15+ "BriefDescription": "RFO requests that miss L2 cache.",
1616 "Counter": "0,1,2,3",
17− "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
18− "PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
17+ "EventName": "L2_RQSTS.RFO_MISS",
18+ "SampleAfterValue": "200003",
19+ "CounterHTOff": "0,1,2,3,4,5,6,7"
20+ },
21+ {
22+ "EventCode": "0x24",
23+ "UMask": "0x24",
24+ "BriefDescription": "L2 cache misses when fetching instructions.",
25+ "Counter": "0,1,2,3",
26+ "EventName": "L2_RQSTS.CODE_RD_MISS",
27+ "SampleAfterValue": "200003",
28+ "CounterHTOff": "0,1,2,3,4,5,6,7"
29+ },
30+ {
31+ "EventCode": "0x24",
32+ "UMask": "0x27",
33+ "BriefDescription": "Demand requests that miss L2 cache.",
34+ "Counter": "0,1,2,3",
35+ "EventName": "L2_RQSTS.ALL_DEMAND_MISS",
1936 "SampleAfterValue": "200003",
2037 "CounterHTOff": "0,1,2,3,4,5,6,7"
2138 },
@@ -29,6 +46,43 @@ · 7 unmodified lines
2946 "SampleAfterValue": "200003",
3047 "CounterHTOff": "0,1,2,3,4,5,6,7"
3148 },
49+ {
50+ "EventCode": "0x24",
51+ "UMask": "0x3f",
52+ "BriefDescription": "All requests that miss L2 cache.",
53+ "Counter": "0,1,2,3",
54+ "EventName": "L2_RQSTS.MISS",
55+ "SampleAfterValue": "200003",
56+ "CounterHTOff": "0,1,2,3,4,5,6,7"
57+ },
58+ {
59+ "EventCode": "0x24",
60+ "UMask": "0x41",
61+ "BriefDescription": "Demand Data Read requests that hit L2 cache",
62+ "Counter": "0,1,2,3",
63+ "EventName": "L2_RQSTS.DEMAND_DATA_RD_HIT",
64+ "PublicDescription": "This event counts the number of demand Data Read requests that hit L2 cache. Only not rejected loads are counted.",
65+ "SampleAfterValue": "200003",
66+ "CounterHTOff": "0,1,2,3,4,5,6,7"
67+ },
68+ {
69+ "EventCode": "0x24",
70+ "UMask": "0x42",
71+ "BriefDescription": "RFO requests that hit L2 cache.",
72+ "Counter": "0,1,2,3",
73+ "EventName": "L2_RQSTS.RFO_HIT",
74+ "SampleAfterValue": "200003",
75+ "CounterHTOff": "0,1,2,3,4,5,6,7"
76+ },
77+ {
78+ "EventCode": "0x24",
79+ "UMask": "0x44",
80+ "BriefDescription": "L2 cache hits when fetching instructions, code reads.",
81+ "Counter": "0,1,2,3",
82+ "EventName": "L2_RQSTS.CODE_RD_HIT",
83+ "SampleAfterValue": "200003",
84+ "CounterHTOff": "0,1,2,3,4,5,6,7"
85+ },
3286 {
3387 "EventCode": "0x24",
3488 "UMask": "0x50",
@@ -69,6 +123,15 @@ · 34 unmodified lines
69123 "SampleAfterValue": "200003",
70124 "CounterHTOff": "0,1,2,3,4,5,6,7"
71125 },
126+ {
127+ "EventCode": "0x24",
128+ "UMask": "0xe7",
129+ "BriefDescription": "Demand requests to L2 cache.",
130+ "Counter": "0,1,2,3",
131+ "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES",
132+ "SampleAfterValue": "200003",
133+ "CounterHTOff": "0,1,2,3,4,5,6,7"
134+ },
72135 {
73136 "EventCode": "0x24",
74137 "UMask": "0xf8",
@@ -79,6 +142,15 @@ · 4 unmodified lines
79142 "SampleAfterValue": "200003",
80143 "CounterHTOff": "0,1,2,3,4,5,6,7"
81144 },
145+ {
146+ "EventCode": "0x24",
147+ "UMask": "0xff",
148+ "BriefDescription": "All L2 requests.",
149+ "Counter": "0,1,2,3",
150+ "EventName": "L2_RQSTS.REFERENCES",
151+ "SampleAfterValue": "200003",
152+ "CounterHTOff": "0,1,2,3,4,5,6,7"
153+ },
82154 {
83155 "EventCode": "0x27",
84156 "UMask": "0x50",
@@ -130,6 +202,27 @@ · 45 unmodified lines
130202 "SampleAfterValue": "2000003",
131203 "CounterHTOff": "2"
132204 },
205+ {
206+ "EventCode": "0x48",
207+ "UMask": "0x1",
208+ "BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
209+ "Counter": "2",
210+ "EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
211+ "AnyThread": "1",
212+ "CounterMask": "1",
213+ "SampleAfterValue": "2000003",
214+ "CounterHTOff": "2"
215+ },
216+ {
217+ "EventCode": "0x48",
218+ "UMask": "0x2",
219+ "BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
220+ "Counter": "0,1,2,3",
221+ "EventName": "L1D_PEND_MISS.FB_FULL",
222+ "CounterMask": "1",
223+ "SampleAfterValue": "2000003",
224+ "CounterHTOff": "0,1,2,3,4,5,6,7"
225+ },
133226 {
134227 "EventCode": "0x51",
135228 "UMask": "0x1",
@@ -151,6 +244,29 @@ · 15 unmodified lines
151244 "SampleAfterValue": "2000003",
152245 "CounterHTOff": "0,1,2,3,4,5,6,7"
153246 },
247+ {
248+ "EventCode": "0x60",
249+ "UMask": "0x1",
250+ "BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore",
251+ "Counter": "0,1,2,3",
252+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
253+ "CounterMask": "1",
254+ "Errata": "BDM76",
255+ "PublicDescription": "This event counts cycles when offcore outstanding Demand Data Read transactions are present in the super queue (SQ). A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation).",
256+ "SampleAfterValue": "2000003",
257+ "CounterHTOff": "0,1,2,3,4,5,6,7"
258+ },
259+ {
260+ "EventCode": "0x60",
261+ "UMask": "0x1",
262+ "BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
263+ "Counter": "0,1,2,3",
264+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_GE_6",
265+ "CounterMask": "6",
266+ "Errata": "BDM76",
267+ "SampleAfterValue": "2000003",
268+ "CounterHTOff": "0,1,2,3,4,5,6,7"
269+ },
154270 {
155271 "EventCode": "0x60",
156272 "UMask": "0x2",
@@ -158,7 +274,7 @@ · 1 unmodified line
158274 "Counter": "0,1,2,3",
159275 "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_CODE_RD",
160276 "Errata": "BDM76",
161− "PublicDescription": "This event counts the number of offcore outstanding Code Reads transactions in the super queue every cycle. The \"Offcore outstanding\" state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
277+ "PublicDescription": "This event counts the number of offcore outstanding Code Reads transactions in the super queue every cycle. The Offcore outstanding state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
162278 "SampleAfterValue": "2000003",
163279 "CounterHTOff": "0,1,2,3,4,5,6,7"
164280 },
@@ -175,24 +291,24 @@ · 10 unmodified lines
175291 },
176292 {
177293 "EventCode": "0x60",
178− "UMask": "0x8",
179− "BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore",
294+ "UMask": "0x4",
295+ "BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle",
180296 "Counter": "0,1,2,3",
181− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
297+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
298+ "CounterMask": "1",
182299 "Errata": "BDM76",
183− "PublicDescription": "This event counts the number of offcore outstanding cacheable Core Data Read transactions in the super queue every cycle. A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation). See corresponding Umask under OFFCORE_REQUESTS.",
300+ "PublicDescription": "This event counts the number of offcore outstanding demand rfo Reads transactions in the super queue every cycle. The Offcore outstanding state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
184301 "SampleAfterValue": "2000003",
185302 "CounterHTOff": "0,1,2,3,4,5,6,7"
186303 },
187304 {
188305 "EventCode": "0x60",
189− "UMask": "0x1",
190− "BriefDescription": "Cycles when offcore outstanding Demand Data Read transactions are present in SuperQueue (SQ), queue to uncore",
306+ "UMask": "0x8",
307+ "BriefDescription": "Offcore outstanding cacheable Core Data Read transactions in SuperQueue (SQ), queue to uncore",
191308 "Counter": "0,1,2,3",
192− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_DATA_RD",
193− "CounterMask": "1",
309+ "EventName": "OFFCORE_REQUESTS_OUTSTANDING.ALL_DATA_RD",
194310 "Errata": "BDM76",
195− "PublicDescription": "This event counts cycles when offcore outstanding Demand Data Read transactions are present in the super queue (SQ). A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation).",
311+ "PublicDescription": "This event counts the number of offcore outstanding cacheable Core Data Read transactions in the super queue every cycle. A transaction is considered to be in the Offcore outstanding state between L2 miss and transaction completion sent to requestor (SQ de-allocation). See corresponding Umask under OFFCORE_REQUESTS.",
196312 "SampleAfterValue": "2000003",
197313 "CounterHTOff": "0,1,2,3,4,5,6,7"
198314 },
@@ -208,18 +324,6 @@ · 9 unmodified lines
208324 "SampleAfterValue": "2000003",
209325 "CounterHTOff": "0,1,2,3,4,5,6,7"
210326 },
211− {
212− "EventCode": "0x60",
213− "UMask": "0x4",
214− "BriefDescription": "Offcore outstanding demand rfo reads transactions in SuperQueue (SQ), queue to uncore, every cycle",
215− "Counter": "0,1,2,3",
216− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.CYCLES_WITH_DEMAND_RFO",
217− "CounterMask": "1",
218− "Errata": "BDM76",
219− "PublicDescription": "This event counts the number of offcore outstanding demand rfo Reads transactions in the super queue every cycle. The \"Offcore outstanding\" state of the transaction lasts from the L2 miss until the sending transaction completion to requestor (SQ deallocation). See the corresponding Umask under OFFCORE_REQUESTS.",
220− "SampleAfterValue": "2000003",
221− "CounterHTOff": "0,1,2,3,4,5,6,7"
222− },
223327 {
224328 "EventCode": "0x63",
225329 "UMask": "0x2",
@@ -266,7 +370,7 @@ · 40 unmodified lines
266370 "BriefDescription": "Demand and prefetch data reads",
267371 "Counter": "0,1,2,3",
268372 "EventName": "OFFCORE_REQUESTS.ALL_DATA_RD",
269− "PublicDescription": "This event counts the demand and prefetch data reads. All Core Data Reads include cacheable \"Demands\" and L2 prefetchers (not L3 prefetchers). Counting also covers reads due to page walks resulted from any request type.",
373+ "PublicDescription": "This event counts the demand and prefetch data reads. All Core Data Reads include cacheable Demands and L2 prefetchers (not L3 prefetchers). Counting also covers reads due to page walks resulted from any request type.",
270374 "SampleAfterValue": "100003",
271375 "CounterHTOff": "0,1,2,3,4,5,6,7"
272376 },
@@ -280,27 +384,36 @@ · 7 unmodified lines
280384 "SampleAfterValue": "2000003",
281385 "CounterHTOff": "0,1,2,3,4,5,6,7"
282386 },
387+ {
388+ "EventCode": "0xB7, 0xBB",
389+ "UMask": "0x1",
390+ "BriefDescription": "Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
391+ "Counter": "0,1,2,3",
392+ "EventName": "OFFCORE_RESPONSE",
393+ "SampleAfterValue": "100003",
394+ "CounterHTOff": "0,1,2,3"
395+ },
283396 {
284397 "EventCode": "0xD0",
285398 "UMask": "0x11",
286− "BriefDescription": "Retired load uops that miss the STLB.",
399+ "BriefDescription": "Retired load uops that miss the STLB. (Precise Event - PEBS)",
287400 "Data_LA": "1",
288401 "PEBS": "1",
289402 "Counter": "0,1,2,3",
290403 "EventName": "MEM_UOPS_RETIRED.STLB_MISS_LOADS",
291− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
404+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
292405 "SampleAfterValue": "100003",
293406 "CounterHTOff": "0,1,2,3"
294407 },
295408 {
296409 "EventCode": "0xD0",
297410 "UMask": "0x12",
298− "BriefDescription": "Retired store uops that miss the STLB.",
411+ "BriefDescription": "Retired store uops that miss the STLB. (Precise Event - PEBS)",
299412 "Data_LA": "1",
300413 "PEBS": "1",
301414 "Counter": "0,1,2,3",
302415 "EventName": "MEM_UOPS_RETIRED.STLB_MISS_STORES",
303− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts store uops with true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
416+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops true STLB miss retired to the architected path. True STLB miss is an uop triggering page walk that gets completed without blocks, and later gets retired. This page walk can end up with or without a fault.",
304417 "SampleAfterValue": "100003",
305418 "L1_Hit_Indication": "1",
306419 "CounterHTOff": "0,1,2,3"
@@ -308,37 +421,37 @@ · 1 unmodified line
308421 {
309422 "EventCode": "0xD0",
310423 "UMask": "0x21",
311− "BriefDescription": "Retired load uops with locked access.",
424+ "BriefDescription": "Retired load uops with locked access. (Precise Event - PEBS)",
312425 "Data_LA": "1",
313426 "PEBS": "1",
314427 "Counter": "0,1,2,3",
315428 "EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
316429 "Errata": "BDM35",
317− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts load uops with locked access retired to the architected path.",
430+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops with locked access retired to the architected path.",
318431 "SampleAfterValue": "100007",
319432 "CounterHTOff": "0,1,2,3"
320433 },
321434 {
322435 "EventCode": "0xD0",
323436 "UMask": "0x41",
324− "BriefDescription": "Retired load uops that split across a cacheline boundary.",
437+ "BriefDescription": "Retired load uops that split across a cacheline boundary.(Precise Event - PEBS)",
325438 "Data_LA": "1",
326439 "PEBS": "1",
327440 "Counter": "0,1,2,3",
328441 "EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
329− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
442+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted load uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
330443 "SampleAfterValue": "100003",
331444 "CounterHTOff": "0,1,2,3"
332445 },
333446 {
334447 "EventCode": "0xD0",
335448 "UMask": "0x42",
336− "BriefDescription": "Retired store uops that split across a cacheline boundary.",
449+ "BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
337450 "Data_LA": "1",
338451 "PEBS": "1",
339452 "Counter": "0,1,2,3",
340453 "EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
341− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
454+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts line-splitted store uops retired to the architected path. A line split is across 64B cache-line which includes a page split (4K).",
342455 "SampleAfterValue": "100003",
343456 "L1_Hit_Indication": "1",
344457 "CounterHTOff": "0,1,2,3"
@@ -346,24 +459,24 @@ · 1 unmodified line
346459 {
347460 "EventCode": "0xD0",
348461 "UMask": "0x81",
349− "BriefDescription": "All retired load uops.",
462+ "BriefDescription": "All retired load uops. (Precise Event - PEBS)",
350463 "Data_LA": "1",
351464 "PEBS": "1",
352465 "Counter": "0,1,2,3",
353466 "EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
354− "PublicDescription": "This event counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
467+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts load uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement. This event also counts SW prefetches.",
355468 "SampleAfterValue": "2000003",
356469 "CounterHTOff": "0,1,2,3"
357470 },
358471 {
359472 "EventCode": "0xD0",
360473 "UMask": "0x82",
361− "BriefDescription": "All retired store uops.",
474+ "BriefDescription": "Retired store uops that split across a cacheline boundary. (Precise Event - PEBS)",
362475 "Data_LA": "1",
363476 "PEBS": "1",
364477 "Counter": "0,1,2,3",
365478 "EventName": "MEM_UOPS_RETIRED.ALL_STORES",
366− "PublicDescription": "This event counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event counts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
479+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts store uops retired to the architected path with a filter on bits 0 and 1 applied.\nNote: This event ?ounts AVX-256bit load/store double-pump memory uops as a single uop at retirement.",
367480 "SampleAfterValue": "2000003",
368481 "L1_Hit_Indication": "1",
369482 "CounterHTOff": "0,1,2,3"
@@ -371,69 +484,69 @@ · 1 unmodified line
371484 {
372485 "EventCode": "0xD1",
373486 "UMask": "0x1",
374− "BriefDescription": "Retired load uops with L1 cache hits as data sources.",
487+ "BriefDescription": "Retired load uops with L1 cache hits as data sources. (Precise Event - PEBS)",
375488 "Data_LA": "1",
376489 "PEBS": "1",
377490 "Counter": "0,1,2,3",
378491 "EventName": "MEM_LOAD_UOPS_RETIRED.L1_HIT",
379− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
492+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data source were hits in the nearest-level (L1) cache.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load. This event also counts SW prefetches independent of the actual data source.",
380493 "SampleAfterValue": "2000003",
381494 "CounterHTOff": "0,1,2,3"
382495 },
383496 {
384497 "EventCode": "0xD1",
385498 "UMask": "0x2",
386− "BriefDescription": "Retired load uops with L2 cache hits as data sources.",
499+ "BriefDescription": "Retired load uops with L2 cache hits as data sources. (Precise Event - PEBS)",
387500 "Data_LA": "1",
388501 "PEBS": "1",
389502 "Counter": "0,1,2,3",
390503 "EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
391504 "Errata": "BDM35",
392− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
505+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the mid-level (L2) cache.",
393506 "SampleAfterValue": "100003",
394507 "CounterHTOff": "0,1,2,3"
395508 },
396509 {
397510 "EventCode": "0xD1",
398511 "UMask": "0x4",
399− "BriefDescription": "Retired load uops which data sources were data hits in L3 without snoops required.",
512+ "BriefDescription": "Hit in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS)",
400513 "Data_LA": "1",
401514 "PEBS": "1",
402515 "Counter": "0,1,2,3",
403516 "EventName": "MEM_LOAD_UOPS_RETIRED.L3_HIT",
404517 "Errata": "BDM100",
405− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
518+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were data hits in the last-level (L3) cache without snoops required.",
406519 "SampleAfterValue": "50021",
407520 "CounterHTOff": "0,1,2,3"
408521 },
409522 {
410523 "EventCode": "0xD1",
411524 "UMask": "0x8",
412− "BriefDescription": "Retired load uops misses in L1 cache as data sources.",
525+ "BriefDescription": "Retired load uops misses in L1 cache as data sources. Uses PEBS.",
413526 "Data_LA": "1",
414527 "PEBS": "1",
415528 "Counter": "0,1,2,3",
416529 "EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
417− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
530+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the nearest-level (L1) cache. Counting excludes unknown and UC data source.",
418531 "SampleAfterValue": "100003",
419532 "CounterHTOff": "0,1,2,3"
420533 },
421534 {
422535 "EventCode": "0xD1",
423536 "UMask": "0x10",
424− "BriefDescription": "Miss in mid-level (L2) cache. Excludes Unknown data-source.",
537+ "BriefDescription": "Retired load uops with L2 cache misses as data sources. Uses PEBS.",
425538 "Data_LA": "1",
426539 "PEBS": "1",
427540 "Counter": "0,1,2,3",
428541 "EventName": "MEM_LOAD_UOPS_RETIRED.L2_MISS",
429− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
542+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were misses in the mid-level (L2) cache. Counting excludes unknown and UC data source.",
430543 "SampleAfterValue": "50021",
431544 "CounterHTOff": "0,1,2,3"
432545 },
433546 {
434547 "EventCode": "0xD1",
435548 "UMask": "0x20",
436− "BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source.",
549+ "BriefDescription": "Miss in last-level (L3) cache. Excludes Unknown data-source. (Precise Event - PEBS).",
437550 "Data_LA": "1",
438551 "PEBS": "1",
439552 "Counter": "0,1,2,3",
@@ -445,84 +558,83 @@ · 5 unmodified lines
445558 {
446559 "EventCode": "0xD1",
447560 "UMask": "0x40",
448− "BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready.",
561+ "BriefDescription": "Retired load uops which data sources were load uops missed L1 but hit FB due to preceding miss to the same cache line with data not ready. (Precise Event - PEBS)",
449562 "Data_LA": "1",
450563 "PEBS": "1",
451564 "Counter": "0,1,2,3",
452565 "EventName": "MEM_LOAD_UOPS_RETIRED.HIT_LFB",
453− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
566+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were load uops missed L1 but hit a fill buffer due to a preceding miss to the same cache line with the data not ready.\nNote: Only two data-sources of L1/FB are applicable for AVX-256bit even though the corresponding AVX load could be serviced by a deeper level in the memory hierarchy. Data source is reported for the Low-half load.",
454567 "SampleAfterValue": "100003",
455568 "CounterHTOff": "0,1,2,3"
456569 },
457570 {
458571 "EventCode": "0xD2",
459572 "UMask": "0x1",
460− "BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache.",
573+ "BriefDescription": "Retired load uops which data sources were L3 hit and cross-core snoop missed in on-pkg core cache. (Precise Event - PEBS)",
461574 "Data_LA": "1",
462575 "PEBS": "1",
463576 "Counter": "0,1,2,3",
464577 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_MISS",
465578 "Errata": "BDM100",
466− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
579+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 Hit and a cross-core snoop missed in the on-pkg core cache.",
467580 "SampleAfterValue": "20011",
468581 "CounterHTOff": "0,1,2,3"
469582 },
470583 {
471584 "EventCode": "0xD2",
472585 "UMask": "0x2",
473− "BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache.",
586+ "BriefDescription": "Retired load uops which data sources were L3 and cross-core snoop hits in on-pkg core cache. (Precise Event - PEBS)",
474587 "Data_LA": "1",
475588 "PEBS": "1",
476589 "Counter": "0,1,2,3",
477590 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HIT",
478591 "Errata": "BDM100",
479− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
592+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were L3 hit and a cross-core snoop hit in the on-pkg core cache.",
480593 "SampleAfterValue": "20011",
481594 "CounterHTOff": "0,1,2,3"
482595 },
483596 {
484597 "EventCode": "0xD2",
485598 "UMask": "0x4",
486− "BriefDescription": "Retired load uops which data sources were HitM responses from shared L3.",
599+ "BriefDescription": "Retired load uops which data sources were HitM responses from shared L3. (Precise Event - PEBS)",
487600 "Data_LA": "1",
488601 "PEBS": "1",
489602 "Counter": "0,1,2,3",
490603 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_HITM",
491604 "Errata": "BDM100",
492− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
605+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were HitM responses from a core on same socket (shared L3).",
493606 "SampleAfterValue": "20011",
494607 "CounterHTOff": "0,1,2,3"
495608 },
496609 {
497610 "EventCode": "0xD2",
498611 "UMask": "0x8",
499− "BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required.",
612+ "BriefDescription": "Retired load uops which data sources were hits in L3 without snoops required. (Precise Event - PEBS)",
500613 "Data_LA": "1",
501614 "PEBS": "1",
502615 "Counter": "0,1,2,3",
503616 "EventName": "MEM_LOAD_UOPS_L3_HIT_RETIRED.XSNP_NONE",
504617 "Errata": "BDM100",
505− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
618+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts retired load uops which data sources were hits in the last-level (L3) cache without snoops required.",
506619 "SampleAfterValue": "100003",
507620 "CounterHTOff": "0,1,2,3"
508621 },
509622 {
510623 "EventCode": "0xD3",
511624 "UMask": "0x1",
512− "BriefDescription": "Data from local DRAM either Snoop not needed or Snoop Miss (RspI)",
513625 "Data_LA": "1",
514626 "PEBS": "1",
515627 "Counter": "0,1,2,3",
516628 "EventName": "MEM_LOAD_UOPS_L3_MISS_RETIRED.LOCAL_DRAM",
517629 "Errata": "BDE70, BDM100",
518− "PublicDescription": "Retired load uop whose Data Source was: local DRAM either Snoop not needed or Snoop Miss (RspI).",
630+ "PublicDescription": "This event counts retired load uops where the data came from local DRAM. This does not include hardware prefetches. This is a precise event.",
519631 "SampleAfterValue": "100007",
520632 "CounterHTOff": "0,1,2,3"
521633 },
522634 {
523635 "EventCode": "0xD3",
524636 "UMask": "0x4",
525− "BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI)",
637+ "BriefDescription": "Retired load uop whose Data Source was: remote DRAM either Snoop not needed or Snoop Miss (RspI) (Precise Event)",
526638 "Data_LA": "1",
527639 "PEBS": "1",
528640 "Counter": "0,1,2,3",
@@ -534,7 +646,7 @@ · 5 unmodified lines
534646 {
535647 "EventCode": "0xD3",
536648 "UMask": "0x10",
537− "BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM",
649+ "BriefDescription": "Retired load uop whose Data Source was: Remote cache HITM (Precise Event)",
538650 "Data_LA": "1",
539651 "PEBS": "1",
540652 "Counter": "0,1,2,3",
@@ -546,7 +658,7 @@ · 5 unmodified lines
546658 {
547659 "EventCode": "0xD3",
548660 "UMask": "0x20",
549− "BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache",
661+ "BriefDescription": "Retired load uop whose Data Source was: forwarded from remote cache (Precise Event)",
550662 "Data_LA": "1",
551663 "PEBS": "1",
552664 "Counter": "0,1,2,3",
@@ -694,119 +806,6 @@ · 141 unmodified lines
694806 "SampleAfterValue": "100003",
695807 "CounterHTOff": "0,1,2,3,4,5,6,7"
696808 },
697− {
698− "EventCode": "0x24",
699− "UMask": "0x42",
700− "BriefDescription": "RFO requests that hit L2 cache.",
701− "Counter": "0,1,2,3",
702− "EventName": "L2_RQSTS.RFO_HIT",
703− "SampleAfterValue": "200003",
704− "CounterHTOff": "0,1,2,3,4,5,6,7"
705− },
706− {
707− "EventCode": "0x24",
708− "UMask": "0x22",
709− "BriefDescription": "RFO requests that miss L2 cache.",
710− "Counter": "0,1,2,3",
711− "EventName": "L2_RQSTS.RFO_MISS",
712− "SampleAfterValue": "200003",
713− "CounterHTOff": "0,1,2,3,4,5,6,7"
714− },
715− {
716− "EventCode": "0x24",
717− "UMask": "0x44",
718− "BriefDescription": "L2 cache hits when fetching instructions, code reads.",
719− "Counter": "0,1,2,3",
720− "EventName": "L2_RQSTS.CODE_RD_HIT",
721− "SampleAfterValue": "200003",
722− "CounterHTOff": "0,1,2,3,4,5,6,7"
723− },
724− {
725− "EventCode": "0x24",
726− "UMask": "0x24",
727− "BriefDescription": "L2 cache misses when fetching instructions.",
728− "Counter": "0,1,2,3",
729− "EventName": "L2_RQSTS.CODE_RD_MISS",
730− "SampleAfterValue": "200003",
731− "CounterHTOff": "0,1,2,3,4,5,6,7"
732− },
733− {
734− "EventCode": "0x24",
735− "UMask": "0x27",
736− "BriefDescription": "Demand requests that miss L2 cache.",
737− "Counter": "0,1,2,3",
738− "EventName": "L2_RQSTS.ALL_DEMAND_MISS",
739− "SampleAfterValue": "200003",
740− "CounterHTOff": "0,1,2,3,4,5,6,7"
741− },
742− {
743− "EventCode": "0x24",
744− "UMask": "0xe7",
745− "BriefDescription": "Demand requests to L2 cache.",
746− "Counter": "0,1,2,3",
747− "EventName": "L2_RQSTS.ALL_DEMAND_REFERENCES",
748− "SampleAfterValue": "200003",
749− "CounterHTOff": "0,1,2,3,4,5,6,7"
750− },
751− {
752− "EventCode": "0x24",
753− "UMask": "0x3f",
754− "BriefDescription": "All requests that miss L2 cache.",
755− "Counter": "0,1,2,3",
756− "EventName": "L2_RQSTS.MISS",
757− "SampleAfterValue": "200003",
758− "CounterHTOff": "0,1,2,3,4,5,6,7"
759− },
760− {
761− "EventCode": "0x24",
762− "UMask": "0xff",
763− "BriefDescription": "All L2 requests.",
764− "Counter": "0,1,2,3",
765− "EventName": "L2_RQSTS.REFERENCES",
766− "SampleAfterValue": "200003",
767− "CounterHTOff": "0,1,2,3,4,5,6,7"
768− },
769− {
770− "EventCode": "0xB7, 0xBB",
771− "UMask": "0x1",
772− "BriefDescription": "Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
773− "Counter": "0,1,2,3",
774− "EventName": "OFFCORE_RESPONSE",
775− "SampleAfterValue": "100003",
776− "CounterHTOff": "0,1,2,3"
777− },
778− {
779− "EventCode": "0x60",
780− "UMask": "0x1",
781− "BriefDescription": "Cycles with at least 6 offcore outstanding Demand Data Read transactions in uncore queue.",
782− "Counter": "0,1,2,3",
783− "EventName": "OFFCORE_REQUESTS_OUTSTANDING.DEMAND_DATA_RD_GE_6",
784− "CounterMask": "6",
785− "Errata": "BDM76",
786− "SampleAfterValue": "2000003",
787− "CounterHTOff": "0,1,2,3,4,5,6,7"
788− },
789− {
790− "EventCode": "0x48",
791− "UMask": "0x1",
792− "BriefDescription": "Cycles with L1D load Misses outstanding from any thread on physical core.",
793− "Counter": "2",
794− "EventName": "L1D_PEND_MISS.PENDING_CYCLES_ANY",
795− "AnyThread": "1",
796− "CounterMask": "1",
797− "SampleAfterValue": "2000003",
798− "CounterHTOff": "2"
799− },
800− {
801− "EventCode": "0x48",
802− "UMask": "0x2",
803− "BriefDescription": "Cycles a demand request was blocked due to Fill Buffers inavailability.",
804− "Counter": "0,1,2,3",
805− "EventName": "L1D_PEND_MISS.FB_FULL",
806− "CounterMask": "1",
807− "SampleAfterValue": "2000003",
808− "CounterHTOff": "0,1,2,3,4,5,6,7"
809− },
810809 {
811810 "Offcore": "1",
812811 "EventCode": "0xB7, 0xBB",
@@ -816,6 +815,7 @@ · 3 unmodified lines
816815 "Counter": "0,1,2,3",
817816 "EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.LLC_HIT.ANY_RESPONSE",
818817 "MSRIndex": "0x1a6,0x1a7",
818+ "PublicDescription": "Counts all requests that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
819819 "SampleAfterValue": "100003",
820820 "CounterHTOff": "0,1,2,3"
821821 },
@@ -828,6 +828,7 @@ · 6 unmodified lines
828828 "Counter": "0,1,2,3",
829829 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_HIT.HITM_OTHER_CORE",
830830 "MSRIndex": "0x1a6,0x1a7",
831+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
831832 "SampleAfterValue": "100003",
832833 "CounterHTOff": "0,1,2,3"
833834 },
@@ -840,6 +841,7 @@ · 6 unmodified lines
840841 "Counter": "0,1,2,3",
841842 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
842843 "MSRIndex": "0x1a6,0x1a7",
844+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
843845 "SampleAfterValue": "100003",
844846 "CounterHTOff": "0,1,2,3"
845847 },
@@ -852,6 +854,7 @@ · 6 unmodified lines
852854 "Counter": "0,1,2,3",
853855 "EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
854856 "MSRIndex": "0x1a6,0x1a7",
857+ "PublicDescription": "Counts all demand & prefetch code reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
855858 "SampleAfterValue": "100003",
856859 "CounterHTOff": "0,1,2,3"
857860 },
@@ -864,6 +867,7 @@ · 6 unmodified lines
864867 "Counter": "0,1,2,3",
865868 "EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_HIT.HITM_OTHER_CORE",
866869 "MSRIndex": "0x1a6,0x1a7",
870+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
867871 "SampleAfterValue": "100003",
868872 "CounterHTOff": "0,1,2,3"
869873 },
@@ -876,6 +880,7 @@ · 6 unmodified lines
876880 "Counter": "0,1,2,3",
877881 "EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
878882 "MSRIndex": "0x1a6,0x1a7",
883+ "PublicDescription": "Counts all demand & prefetch RFOs that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
879884 "SampleAfterValue": "100003",
880885 "CounterHTOff": "0,1,2,3"
881886 },
@@ -888,6 +893,7 @@ · 6 unmodified lines
888893 "Counter": "0,1,2,3",
889894 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_HIT.HITM_OTHER_CORE",
890895 "MSRIndex": "0x1a6,0x1a7",
896+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
891897 "SampleAfterValue": "100003",
892898 "CounterHTOff": "0,1,2,3"
893899 },
@@ -900,6 +906,7 @@ · 6 unmodified lines
900906 "Counter": "0,1,2,3",
901907 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_HIT.HIT_OTHER_CORE_NO_FWD",
902908 "MSRIndex": "0x1a6,0x1a7",
909+ "PublicDescription": "Counts all demand & prefetch data reads that hit in the L3 and the snoops to sibling cores hit in either E/S state and the line is not forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
903910 "SampleAfterValue": "100003",
904911 "CounterHTOff": "0,1,2,3"
905912 },
@@ -912,6 +919,7 @@ · 6 unmodified lines
912919 "Counter": "0,1,2,3",
913920 "EventName": "OFFCORE_RESPONSE.PF_LLC_CODE_RD.LLC_HIT.ANY_RESPONSE",
914921 "MSRIndex": "0x1a6,0x1a7",
922+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
915923 "SampleAfterValue": "100003",
916924 "CounterHTOff": "0,1,2,3"
917925 },
@@ -924,6 +932,7 @@ · 6 unmodified lines
924932 "Counter": "0,1,2,3",
925933 "EventName": "OFFCORE_RESPONSE.PF_LLC_RFO.LLC_HIT.ANY_RESPONSE",
926934 "MSRIndex": "0x1a6,0x1a7",
935+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
927936 "SampleAfterValue": "100003",
928937 "CounterHTOff": "0,1,2,3"
929938 },
@@ -936,6 +945,20 @@ · 6 unmodified lines
936945 "Counter": "0,1,2,3",
937946 "EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.HITM_OTHER_CORE",
938947 "MSRIndex": "0x1a6,0x1a7",
948+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 and the snoop to one of the sibling cores hits the line in M state and the line is forwarded Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
949+ "SampleAfterValue": "100003",
950+ "CounterHTOff": "0,1,2,3"
951+ },
952+ {
953+ "Offcore": "1",
954+ "EventCode": "0xB7, 0xBB",
955+ "UMask": "0x1",
956+ "BriefDescription": "Counts all demand data writes (RFOs) that hit in the L3",
957+ "MSRValue": "0x3f803c0002",
958+ "Counter": "0,1,2,3",
959+ "EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_HIT.ANY_RESPONSE",
960+ "MSRIndex": "0x1a6,0x1a7",
961+ "PublicDescription": "Counts all demand data writes (RFOs) that hit in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
939962 "SampleAfterValue": "100003",
940963 "CounterHTOff": "0,1,2,3"
941964 }
tools/perf/pmu-events/arch/x86/broadwellx/floating-point.json+51−57
@@ -6,7 +6,7 @@ · 5 unmodified lines
66 "Counter": "0,1,2,3",
77 "EventName": "OTHER_ASSISTS.AVX_TO_SSE",
88 "Errata": "BDM30",
9− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
9+ "PublicDescription": "This event counts the number of transitions from AVX-256 to legacy SSE when penalty is applicable.",
1010 "SampleAfterValue": "100003",
1111 "CounterHTOff": "0,1,2,3,4,5,6,7"
1212 },
@@ -17,7 +17,7 @@ · 4 unmodified lines
1717 "Counter": "0,1,2,3",
1818 "EventName": "OTHER_ASSISTS.SSE_TO_AVX",
1919 "Errata": "BDM30",
20− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
20+ "PublicDescription": "This event counts the number of transitions from legacy SSE to AVX-256 when penalty is applicable.",
2121 "SampleAfterValue": "100003",
2222 "CounterHTOff": "0,1,2,3,4,5,6,7"
2323 },
@@ -25,7 +25,6 @@ · 1 unmodified line
2525 "EventCode": "0xC7",
2626 "UMask": "0x1",
2727 "BriefDescription": "Number of SSE/AVX computational scalar double precision floating-point instructions retired. Each count represents 1 computation. Applies to SSE* and AVX* scalar double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
28− "PEBS": "1",
2928 "Counter": "0,1,2,3",
3029 "EventName": "FP_ARITH_INST_RETIRED.SCALAR_DOUBLE",
3130 "SampleAfterValue": "2000003",
@@ -35,17 +34,24 @@ · 3 unmodified lines
3534 "EventCode": "0xC7",
3635 "UMask": "0x2",
3736 "BriefDescription": "Number of SSE/AVX computational scalar single precision floating-point instructions retired. Each count represents 1 computation. Applies to SSE* and AVX* scalar single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
38− "PEBS": "1",
3937 "Counter": "0,1,2,3",
4038 "EventName": "FP_ARITH_INST_RETIRED.SCALAR_SINGLE",
4139 "SampleAfterValue": "2000003",
4240 "CounterHTOff": "0,1,2,3"
4341 },
42+ {
43+ "EventCode": "0xC7",
44+ "UMask": "0x3",
45+ "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
46+ "Counter": "0,1,2,3",
47+ "EventName": "FP_ARITH_INST_RETIRED.SCALAR",
48+ "SampleAfterValue": "2000003",
49+ "CounterHTOff": "0,1,2,3"
50+ },
4451 {
4552 "EventCode": "0xC7",
4653 "UMask": "0x4",
4754 "BriefDescription": "Number of SSE/AVX computational 128-bit packed double precision floating-point instructions retired. Each count represents 2 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
48− "PEBS": "1",
4955 "Counter": "0,1,2,3",
5056 "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_DOUBLE",
5157 "SampleAfterValue": "2000003",
@@ -55,7 +61,6 @@ · 3 unmodified lines
5561 "EventCode": "0xC7",
5662 "UMask": "0x8",
5763 "BriefDescription": "Number of SSE/AVX computational 128-bit packed single precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
58− "PEBS": "1",
5964 "Counter": "0,1,2,3",
6065 "EventName": "FP_ARITH_INST_RETIRED.128B_PACKED_SINGLE",
6166 "SampleAfterValue": "2000003",
@@ -65,19 +70,54 @@ · 3 unmodified lines
6570 "EventCode": "0xC7",
6671 "UMask": "0x10",
6772 "BriefDescription": "Number of SSE/AVX computational 256-bit packed double precision floating-point instructions retired. Each count represents 4 computations. Applies to SSE* and AVX* packed double precision floating-point instructions: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
68− "PEBS": "1",
6973 "Counter": "0,1,2,3",
7074 "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_DOUBLE",
7175 "SampleAfterValue": "2000003",
7276 "CounterHTOff": "0,1,2,3"
7377 },
78+ {
79+ "EventCode": "0xC7",
80+ "UMask": "0x15",
81+ "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
82+ "Counter": "0,1,2,3",
83+ "EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
84+ "SampleAfterValue": "2000006",
85+ "CounterHTOff": "0,1,2,3"
86+ },
87+ {
88+ "EventCode": "0xc7",
89+ "UMask": "0x20",
90+ "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
91+ "Counter": "0,1,2,3",
92+ "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
93+ "SampleAfterValue": "2000003",
94+ "CounterHTOff": "0,1,2,3"
95+ },
96+ {
97+ "EventCode": "0xC7",
98+ "UMask": "0x2a",
99+ "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
100+ "Counter": "0,1,2,3",
101+ "EventName": "FP_ARITH_INST_RETIRED.SINGLE",
102+ "SampleAfterValue": "2000005",
103+ "CounterHTOff": "0,1,2,3"
104+ },
105+ {
106+ "EventCode": "0xC7",
107+ "UMask": "0x3c",
108+ "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
109+ "Counter": "0,1,2,3",
110+ "EventName": "FP_ARITH_INST_RETIRED.PACKED",
111+ "SampleAfterValue": "2000004",
112+ "CounterHTOff": "0,1,2,3"
113+ },
74114 {
75115 "EventCode": "0xCA",
76116 "UMask": "0x2",
77117 "BriefDescription": "Number of X87 assists due to output value.",
78118 "Counter": "0,1,2,3",
79119 "EventName": "FP_ASSIST.X87_OUTPUT",
80− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
120+ "PublicDescription": "This event counts the number of x87 floating point (FP) micro-code assist (numeric overflow/underflow, inexact result) when the output value (destination register) is invalid.",
81121 "SampleAfterValue": "100003",
82122 "CounterHTOff": "0,1,2,3,4,5,6,7"
83123 },
@@ -87,7 +127,7 @@ · 3 unmodified lines
87127 "BriefDescription": "Number of X87 assists due to input value.",
88128 "Counter": "0,1,2,3",
89129 "EventName": "FP_ASSIST.X87_INPUT",
90− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
130+ "PublicDescription": "This event counts x87 floating point (FP) micro-code assist (invalid operation, denormal operand, SNaN operand) when the input value (one of the source operands to an FP instruction) is invalid.",
91131 "SampleAfterValue": "100003",
92132 "CounterHTOff": "0,1,2,3,4,5,6,7"
93133 },
@@ -97,7 +137,7 @@ · 3 unmodified lines
97137 "BriefDescription": "Number of SIMD FP assists due to Output values",
98138 "Counter": "0,1,2,3",
99139 "EventName": "FP_ASSIST.SIMD_OUTPUT",
100− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
140+ "PublicDescription": "This event counts the number of SSE* floating point (FP) micro-code assist (numeric overflow/underflow) when the output value (destination register) is invalid. Counting covers only cases involving penalties that require micro-code assist intervention.",
101141 "SampleAfterValue": "100003",
102142 "CounterHTOff": "0,1,2,3,4,5,6,7"
103143 },
@@ -107,7 +147,7 @@ · 3 unmodified lines
107147 "BriefDescription": "Number of SIMD FP assists due to input values",
108148 "Counter": "0,1,2,3",
109149 "EventName": "FP_ASSIST.SIMD_INPUT",
110− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
150+ "PublicDescription": "This event counts any input SSE* FP assist - invalid operation, denormal operand, dividing by zero, SNaN operand. Counting includes only cases involving penalties that required micro-code assist intervention.",
111151 "SampleAfterValue": "100003",
112152 "CounterHTOff": "0,1,2,3,4,5,6,7"
113153 },
@@ -121,51 +161,5 @@ · 7 unmodified lines
121161 "PublicDescription": "This event counts cycles with any input and output SSE or x87 FP assist. If an input and output assist are detected on the same cycle the event increments by 1.",
122162 "SampleAfterValue": "100003",
123163 "CounterHTOff": "0,1,2,3"
124− },
125− {
126− "EventCode": "0xc7",
127− "UMask": "0x20",
128− "BriefDescription": "Number of SSE/AVX computational 256-bit packed single precision floating-point instructions retired. Each count represents 8 computations. Applies to SSE* and AVX* packed single precision floating-point instructions: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
129− "PEBS": "1",
130− "Counter": "0,1,2,3",
131− "EventName": "FP_ARITH_INST_RETIRED.256B_PACKED_SINGLE",
132− "SampleAfterValue": "2000003",
133− "CounterHTOff": "0,1,2,3"
134− },
135− {
136− "EventCode": "0xC7",
137− "UMask": "0x3",
138− "BriefDescription": "Number of SSE/AVX computational scalar floating-point instructions retired. Applies to SSE* and AVX* scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT FM(N)ADD/SUB. FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
139− "Counter": "0,1,2,3",
140− "EventName": "FP_ARITH_INST_RETIRED.SCALAR",
141− "SampleAfterValue": "2000003",
142− "CounterHTOff": "0,1,2,3"
143− },
144− {
145− "EventCode": "0xC7",
146− "UMask": "0x3c",
147− "BriefDescription": "Number of SSE/AVX computational packed floating-point instructions retired. Applies to SSE* and AVX*, packed, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RSQRT RCP SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element.",
148− "Counter": "0,1,2,3",
149− "EventName": "FP_ARITH_INST_RETIRED.PACKED",
150− "SampleAfterValue": "2000004",
151− "CounterHTOff": "0,1,2,3"
152− },
153− {
154− "EventCode": "0xC7",
155− "UMask": "0x2a",
156− "BriefDescription": "Number of SSE/AVX computational single precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX RCP RSQRT SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
157− "Counter": "0,1,2,3",
158− "EventName": "FP_ARITH_INST_RETIRED.SINGLE",
159− "SampleAfterValue": "2000005",
160− "CounterHTOff": "0,1,2,3"
161− },
162− {
163− "EventCode": "0xC7",
164− "UMask": "0x15",
165− "BriefDescription": "Number of SSE/AVX computational double precision floating-point instructions retired. Applies to SSE* and AVX*scalar, double and single precision floating-point: ADD SUB MUL DIV MIN MAX SQRT DPP FM(N)ADD/SUB. DPP and FM(N)ADD/SUB instructions count twice as they perform multiple calculations per element. ?.",
166− "Counter": "0,1,2,3",
167− "EventName": "FP_ARITH_INST_RETIRED.DOUBLE",
168− "SampleAfterValue": "2000006",
169− "CounterHTOff": "0,1,2,3"
170164 }
171165 ]
tools/perf/pmu-events/arch/x86/broadwellx/frontend.json+69−69
@@ -15,80 +15,49 @@ · 14 unmodified lines
1515 "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path",
1616 "Counter": "0,1,2,3",
1717 "EventName": "IDQ.MITE_UOPS",
18− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
18+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
1919 "SampleAfterValue": "2000003",
2020 "CounterHTOff": "0,1,2,3,4,5,6,7"
2121 },
2222 {
2323 "EventCode": "0x79",
24− "UMask": "0x8",
25− "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
26− "Counter": "0,1,2,3",
27− "EventName": "IDQ.DSB_UOPS",
28− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
29− "SampleAfterValue": "2000003",
30− "CounterHTOff": "0,1,2,3,4,5,6,7"
31− },
32− {
33− "EventCode": "0x79",
34− "UMask": "0x10",
35− "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
36− "Counter": "0,1,2,3",
37− "EventName": "IDQ.MS_DSB_UOPS",
38− "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
39− "SampleAfterValue": "2000003",
40− "CounterHTOff": "0,1,2,3,4,5,6,7"
41− },
42− {
43− "EventCode": "0x79",
44− "UMask": "0x20",
45− "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
46− "Counter": "0,1,2,3",
47− "EventName": "IDQ.MS_MITE_UOPS",
48− "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
49− "SampleAfterValue": "2000003",
50− "CounterHTOff": "0,1,2,3,4,5,6,7"
51− },
52− {
53− "EventCode": "0x79",
54− "UMask": "0x30",
55− "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
24+ "UMask": "0x4",
25+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path",
5626 "Counter": "0,1,2,3",
57− "EventName": "IDQ.MS_UOPS",
58− "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
27+ "EventName": "IDQ.MITE_CYCLES",
28+ "CounterMask": "1",
29+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ.",
5930 "SampleAfterValue": "2000003",
6031 "CounterHTOff": "0,1,2,3,4,5,6,7"
6132 },
6233 {
6334 "EventCode": "0x79",
64− "UMask": "0x30",
65− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
35+ "UMask": "0x8",
36+ "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path",
6637 "Counter": "0,1,2,3",
67− "EventName": "IDQ.MS_CYCLES",
68− "CounterMask": "1",
69− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may \"bypass\" the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
38+ "EventName": "IDQ.DSB_UOPS",
39+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
7040 "SampleAfterValue": "2000003",
7141 "CounterHTOff": "0,1,2,3,4,5,6,7"
7242 },
7343 {
7444 "EventCode": "0x79",
75− "UMask": "0x4",
76− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from MITE path",
45+ "UMask": "0x8",
46+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path",
7747 "Counter": "0,1,2,3",
78− "EventName": "IDQ.MITE_CYCLES",
48+ "EventName": "IDQ.DSB_CYCLES",
7949 "CounterMask": "1",
80− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ.",
50+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
8151 "SampleAfterValue": "2000003",
8252 "CounterHTOff": "0,1,2,3,4,5,6,7"
8353 },
8454 {
8555 "EventCode": "0x79",
86− "UMask": "0x8",
87− "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) from Decode Stream Buffer (DSB) path",
56+ "UMask": "0x10",
57+ "BriefDescription": "Uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
8858 "Counter": "0,1,2,3",
89− "EventName": "IDQ.DSB_CYCLES",
90− "CounterMask": "1",
91− "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
59+ "EventName": "IDQ.MS_DSB_UOPS",
60+ "PublicDescription": "This event counts the number of uops initiated by Decode Stream Buffer (DSB) that are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
9261 "SampleAfterValue": "2000003",
9362 "CounterHTOff": "0,1,2,3,4,5,6,7"
9463 },
@@ -99,7 +68,7 @@ · 4 unmodified lines
9968 "Counter": "0,1,2,3",
10069 "EventName": "IDQ.MS_DSB_CYCLES",
10170 "CounterMask": "1",
102− "PublicDescription": "This event counts cycles during which uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
71+ "PublicDescription": "This event counts cycles during which uops initiated by Decode Stream Buffer (DSB) are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
10372 "SampleAfterValue": "2000003",
10473 "CounterHTOff": "0,1,2,3,4,5,6,7"
10574 },
@@ -111,7 +80,7 @@ · 5 unmodified lines
11180 "Counter": "0,1,2,3",
11281 "EventName": "IDQ.MS_DSB_OCCUR",
11382 "CounterMask": "1",
114− "PublicDescription": "This event counts the number of deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while the Microcode Sequencer (MS) is busy. Counting includes uops that may \"bypass\" the IDQ.",
83+ "PublicDescription": "This event counts the number of deliveries to Instruction Decode Queue (IDQ) initiated by Decode Stream Buffer (DSB) while the Microcode Sequencer (MS) is busy. Counting includes uops that may bypass the IDQ.",
11584 "SampleAfterValue": "2000003",
11685 "CounterHTOff": "0,1,2,3,4,5,6,7"
11786 },
@@ -122,7 +91,7 @@ · 4 unmodified lines
12291 "Counter": "0,1,2,3",
12392 "EventName": "IDQ.ALL_DSB_CYCLES_4_UOPS",
12493 "CounterMask": "4",
125− "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
94+ "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
12695 "SampleAfterValue": "2000003",
12796 "CounterHTOff": "0,1,2,3,4,5,6,7"
12897 },
@@ -133,7 +102,17 @@ · 4 unmodified lines
133102 "Counter": "0,1,2,3",
134103 "EventName": "IDQ.ALL_DSB_CYCLES_ANY_UOPS",
135104 "CounterMask": "1",
136− "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may \"bypass\" the IDQ.",
105+ "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the Decode Stream Buffer (DSB) path. Counting includes uops that may bypass the IDQ.",
106+ "SampleAfterValue": "2000003",
107+ "CounterHTOff": "0,1,2,3,4,5,6,7"
108+ },
109+ {
110+ "EventCode": "0x79",
111+ "UMask": "0x20",
112+ "BriefDescription": "Uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
113+ "Counter": "0,1,2,3",
114+ "EventName": "IDQ.MS_MITE_UOPS",
115+ "PublicDescription": "This event counts the number of uops initiated by MITE and delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ.",
137116 "SampleAfterValue": "2000003",
138117 "CounterHTOff": "0,1,2,3,4,5,6,7"
139118 },
@@ -144,7 +123,7 @@ · 4 unmodified lines
144123 "Counter": "0,1,2,3",
145124 "EventName": "IDQ.ALL_MITE_CYCLES_4_UOPS",
146125 "CounterMask": "4",
147− "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
126+ "PublicDescription": "This event counts the number of cycles 4 uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
148127 "SampleAfterValue": "2000003",
149128 "CounterHTOff": "0,1,2,3,4,5,6,7"
150129 },
@@ -155,7 +134,39 @@ · 4 unmodified lines
155134 "Counter": "0,1,2,3",
156135 "EventName": "IDQ.ALL_MITE_CYCLES_ANY_UOPS",
157136 "CounterMask": "1",
158− "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
137+ "PublicDescription": "This event counts the number of cycles uops were delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
138+ "SampleAfterValue": "2000003",
139+ "CounterHTOff": "0,1,2,3,4,5,6,7"
140+ },
141+ {
142+ "EventCode": "0x79",
143+ "UMask": "0x30",
144+ "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
145+ "Counter": "0,1,2,3",
146+ "EventName": "IDQ.MS_UOPS",
147+ "PublicDescription": "This event counts the total number of uops delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
148+ "SampleAfterValue": "2000003",
149+ "CounterHTOff": "0,1,2,3,4,5,6,7"
150+ },
151+ {
152+ "EventCode": "0x79",
153+ "UMask": "0x30",
154+ "BriefDescription": "Cycles when uops are being delivered to Instruction Decode Queue (IDQ) while Microcode Sequenser (MS) is busy",
155+ "Counter": "0,1,2,3",
156+ "EventName": "IDQ.MS_CYCLES",
157+ "CounterMask": "1",
158+ "PublicDescription": "This event counts cycles during which uops are being delivered to Instruction Decode Queue (IDQ) while the Microcode Sequenser (MS) is busy. Counting includes uops that may bypass the IDQ. Uops maybe initiated by Decode Stream Buffer (DSB) or MITE.",
159+ "SampleAfterValue": "2000003",
160+ "CounterHTOff": "0,1,2,3,4,5,6,7"
161+ },
162+ {
163+ "EdgeDetect": "1",
164+ "EventCode": "0x79",
165+ "UMask": "0x30",
166+ "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
167+ "Counter": "0,1,2,3",
168+ "EventName": "IDQ.MS_SWITCHES",
169+ "CounterMask": "1",
159170 "SampleAfterValue": "2000003",
160171 "CounterHTOff": "0,1,2,3,4,5,6,7"
161172 },
@@ -165,7 +176,7 @@ · 3 unmodified lines
165176 "BriefDescription": "Uops delivered to Instruction Decode Queue (IDQ) from MITE path",
166177 "Counter": "0,1,2,3",
167178 "EventName": "IDQ.MITE_ALL_UOPS",
168− "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may \"bypass\" the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
179+ "PublicDescription": "This event counts the number of uops delivered to Instruction Decode Queue (IDQ) from the MITE path. Counting includes uops that may bypass the IDQ. This also means that uops are not being delivered from the Decode Stream Buffer (DSB).",
169180 "SampleAfterValue": "2000003",
170181 "CounterHTOff": "0,1,2,3,4,5,6,7"
171182 },
@@ -205,7 +216,7 @@ · 33 unmodified lines
205216 "BriefDescription": "Uops not delivered to Resource Allocation Table (RAT) per thread when backend of the machine is not stalled",
206217 "Counter": "0,1,2,3",
207218 "EventName": "IDQ_UOPS_NOT_DELIVERED.CORE",
208− "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding ?4 ? x? when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
219+ "PublicDescription": "This event counts the number of uops not delivered to Resource Allocation Table (RAT) per thread adding 4 x when Resource Allocation Table (RAT) is not stalled and Instruction Decode Queue (IDQ) delivers x uops to Resource Allocation Table (RAT) (where x belongs to {0,1,2,3}). Counting does not cover cases when:\n a. IDQ-Resource Allocation Table (RAT) pipe serves the other thread;\n b. Resource Allocation Table (RAT) is stalled for the thread (including uop drops and clear BE conditions); \n c. Instruction Decode Queue (IDQ) delivers four uops.",
209220 "SampleAfterValue": "2000003",
210221 "CounterHTOff": "0,1,2,3"
211222 },
@@ -268,18 +279,7 @@ · 56 unmodified lines
268279 "BriefDescription": "Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles.",
269280 "Counter": "0,1,2,3",
270281 "EventName": "DSB2MITE_SWITCHES.PENALTY_CYCLES",
271− "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 0?2 cycles.",
272− "SampleAfterValue": "2000003",
273− "CounterHTOff": "0,1,2,3,4,5,6,7"
274− },
275− {
276− "EdgeDetect": "1",
277− "EventCode": "0x79",
278− "UMask": "0x30",
279− "BriefDescription": "Number of switches from DSB (Decode Stream Buffer) or MITE (legacy decode pipeline) to the Microcode Sequencer.",
280− "Counter": "0,1,2,3",
281− "EventName": "IDQ.MS_SWITCHES",
282− "CounterMask": "1",
282+ "PublicDescription": "This event counts Decode Stream Buffer (DSB)-to-MITE switch true penalty cycles. These cycles do not include uops routed through because of the switch itself, for example, when Instruction Decode Queue (IDQ) pre-allocation is unavailable, or Instruction Decode Queue (IDQ) is full. SBD-to-MITE switch true penalty cycles happen after the merge mux (MM) receives Decode Stream Buffer (DSB) Sync-indication until receiving the first MITE uop. \nMM is placed before Instruction Decode Queue (IDQ) to merge uops being fed from the MITE and Decode Stream Buffer (DSB) paths. Decode Stream Buffer (DSB) inserts the Sync-indication whenever a Decode Stream Buffer (DSB)-to-MITE switch occurs.\nPenalty: A Decode Stream Buffer (DSB) hit followed by a Decode Stream Buffer (DSB) miss can cost up to six cycles in which no uops are delivered to the IDQ. Most often, such switches from the Decode Stream Buffer (DSB) to the legacy pipeline cost 02 cycles.",
283283 "SampleAfterValue": "2000003",
284284 "CounterHTOff": "0,1,2,3,4,5,6,7"
285285 }
tools/perf/pmu-events/arch/x86/broadwellx/memory.json+35−5
@@ -95,7 +95,6 @@ · 94 unmodified lines
9595 "BriefDescription": "Counts the number of times a class of instructions that may cause a transactional abort was executed. Since this is the count of execution, it may not always cause a transactional abort.",
9696 "Counter": "0,1,2,3",
9797 "EventName": "TX_EXEC.MISC1",
98− "PublicDescription": "Unfriendly TSX abort triggered by a flowmarker.",
9998 "SampleAfterValue": "2000003",
10099 "CounterHTOff": "0,1,2,3,4,5,6,7"
101100 },
@@ -171,11 +170,11 @@ · 69 unmodified lines
171170 {
172171 "EventCode": "0xc8",
173172 "UMask": "0x4",
174− "BriefDescription": "Number of times HLE abort was triggered",
173+ "BriefDescription": "Number of times HLE abort was triggered (PEBS)",
175174 "PEBS": "1",
176175 "Counter": "0,1,2,3",
177176 "EventName": "HLE_RETIRED.ABORTED",
178− "PublicDescription": "Number of times HLE abort was triggered.",
177+ "PublicDescription": "Number of times HLE abort was triggered (PEBS).",
179178 "SampleAfterValue": "2000003",
180179 "CounterHTOff": "0,1,2,3,4,5,6,7"
181180 },
@@ -252,11 +251,11 @@ · 70 unmodified lines
252251 {
253252 "EventCode": "0xc9",
254253 "UMask": "0x4",
255− "BriefDescription": "Number of times RTM abort was triggered",
254+ "BriefDescription": "Number of times RTM abort was triggered (PEBS)",
256255 "PEBS": "1",
257256 "Counter": "0,1,2,3",
258257 "EventName": "RTM_RETIRED.ABORTED",
259− "PublicDescription": "Number of times RTM abort was triggered .",
258+ "PublicDescription": "Number of times RTM abort was triggered (PEBS).",
260259 "SampleAfterValue": "2000003",
261260 "CounterHTOff": "0,1,2,3"
262261 },
@@ -439,6 +438,7 @@ · 176 unmodified lines
439438 "Counter": "0,1,2,3",
440439 "EventName": "OFFCORE_RESPONSE.ALL_REQUESTS.LLC_MISS.ANY_RESPONSE",
441440 "MSRIndex": "0x1a6,0x1a7",
441+ "PublicDescription": "Counts all requests that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
442442 "SampleAfterValue": "100003",
443443 "CounterHTOff": "0,1,2,3"
444444 },
@@ -451,6 +451,7 @@ · 6 unmodified lines
451451 "Counter": "0,1,2,3",
452452 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_HIT_FORWARD",
453453 "MSRIndex": "0x1a6,0x1a7",
454+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and clean or shared data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
454455 "SampleAfterValue": "100003",
455456 "CounterHTOff": "0,1,2,3"
456457 },
@@ -463,6 +464,7 @@ · 6 unmodified lines
463464 "Counter": "0,1,2,3",
464465 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_HITM",
465466 "MSRIndex": "0x1a6,0x1a7",
467+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
466468 "SampleAfterValue": "100003",
467469 "CounterHTOff": "0,1,2,3"
468470 },
@@ -475,6 +477,7 @@ · 6 unmodified lines
475477 "Counter": "0,1,2,3",
476478 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.REMOTE_DRAM",
477479 "MSRIndex": "0x1a6,0x1a7",
480+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from remote dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
478481 "SampleAfterValue": "100003",
479482 "CounterHTOff": "0,1,2,3"
480483 },
@@ -487,6 +490,7 @@ · 6 unmodified lines
487490 "Counter": "0,1,2,3",
488491 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.LOCAL_DRAM",
489492 "MSRIndex": "0x1a6,0x1a7",
493+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
490494 "SampleAfterValue": "100003",
491495 "CounterHTOff": "0,1,2,3"
492496 },
@@ -499,6 +503,7 @@ · 6 unmodified lines
499503 "Counter": "0,1,2,3",
500504 "EventName": "OFFCORE_RESPONSE.ALL_READS.LLC_MISS.ANY_RESPONSE",
501505 "MSRIndex": "0x1a6,0x1a7",
506+ "PublicDescription": "Counts all data/code/rfo reads (demand & prefetch) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
502507 "SampleAfterValue": "100003",
503508 "CounterHTOff": "0,1,2,3"
504509 },
@@ -511,6 +516,7 @@ · 6 unmodified lines
511516 "Counter": "0,1,2,3",
512517 "EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_MISS.LOCAL_DRAM",
513518 "MSRIndex": "0x1a6,0x1a7",
519+ "PublicDescription": "Counts all demand & prefetch code reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
514520 "SampleAfterValue": "100003",
515521 "CounterHTOff": "0,1,2,3"
516522 },
@@ -523,6 +529,7 @@ · 6 unmodified lines
523529 "Counter": "0,1,2,3",
524530 "EventName": "OFFCORE_RESPONSE.ALL_CODE_RD.LLC_MISS.ANY_RESPONSE",
525531 "MSRIndex": "0x1a6,0x1a7",
532+ "PublicDescription": "Counts all demand & prefetch code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
526533 "SampleAfterValue": "100003",
527534 "CounterHTOff": "0,1,2,3"
528535 },
@@ -535,6 +542,7 @@ · 6 unmodified lines
535542 "Counter": "0,1,2,3",
536543 "EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_MISS.LOCAL_DRAM",
537544 "MSRIndex": "0x1a6,0x1a7",
545+ "PublicDescription": "Counts all demand & prefetch RFOs that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
538546 "SampleAfterValue": "100003",
539547 "CounterHTOff": "0,1,2,3"
540548 },
@@ -547,6 +555,7 @@ · 6 unmodified lines
547555 "Counter": "0,1,2,3",
548556 "EventName": "OFFCORE_RESPONSE.ALL_RFO.LLC_MISS.ANY_RESPONSE",
549557 "MSRIndex": "0x1a6,0x1a7",
558+ "PublicDescription": "Counts all demand & prefetch RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
550559 "SampleAfterValue": "100003",
551560 "CounterHTOff": "0,1,2,3"
552561 },
@@ -559,6 +568,7 @@ · 6 unmodified lines
559568 "Counter": "0,1,2,3",
560569 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_HIT_FORWARD",
561570 "MSRIndex": "0x1a6,0x1a7",
571+ "PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and clean or shared data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
562572 "SampleAfterValue": "100003",
563573 "CounterHTOff": "0,1,2,3"
564574 },
@@ -571,6 +581,7 @@ · 6 unmodified lines
571581 "Counter": "0,1,2,3",
572582 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_HITM",
573583 "MSRIndex": "0x1a6,0x1a7",
584+ "PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
574585 "SampleAfterValue": "100003",
575586 "CounterHTOff": "0,1,2,3"
576587 },
@@ -583,6 +594,7 @@ · 6 unmodified lines
583594 "Counter": "0,1,2,3",
584595 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.REMOTE_DRAM",
585596 "MSRIndex": "0x1a6,0x1a7",
597+ "PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from remote dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
586598 "SampleAfterValue": "100003",
587599 "CounterHTOff": "0,1,2,3"
588600 },
@@ -595,6 +607,7 @@ · 6 unmodified lines
595607 "Counter": "0,1,2,3",
596608 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.LOCAL_DRAM",
597609 "MSRIndex": "0x1a6,0x1a7",
610+ "PublicDescription": "Counts all demand & prefetch data reads that miss the L3 and the data is returned from local dram Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
598611 "SampleAfterValue": "100003",
599612 "CounterHTOff": "0,1,2,3"
600613 },
@@ -607,6 +620,7 @@ · 6 unmodified lines
607620 "Counter": "0,1,2,3",
608621 "EventName": "OFFCORE_RESPONSE.ALL_DATA_RD.LLC_MISS.ANY_RESPONSE",
609622 "MSRIndex": "0x1a6,0x1a7",
623+ "PublicDescription": "Counts all demand & prefetch data reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
610624 "SampleAfterValue": "100003",
611625 "CounterHTOff": "0,1,2,3"
612626 },
@@ -619,6 +633,7 @@ · 6 unmodified lines
619633 "Counter": "0,1,2,3",
620634 "EventName": "OFFCORE_RESPONSE.PF_LLC_CODE_RD.LLC_MISS.ANY_RESPONSE",
621635 "MSRIndex": "0x1a6,0x1a7",
636+ "PublicDescription": "Counts prefetch (that bring data to LLC only) code reads that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
622637 "SampleAfterValue": "100003",
623638 "CounterHTOff": "0,1,2,3"
624639 },
@@ -631,6 +646,7 @@ · 6 unmodified lines
631646 "Counter": "0,1,2,3",
632647 "EventName": "OFFCORE_RESPONSE.PF_LLC_RFO.LLC_MISS.ANY_RESPONSE",
633648 "MSRIndex": "0x1a6,0x1a7",
649+ "PublicDescription": "Counts all prefetch (that bring data to LLC only) RFOs that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
634650 "SampleAfterValue": "100003",
635651 "CounterHTOff": "0,1,2,3"
636652 },
@@ -643,6 +659,20 @@ · 6 unmodified lines
643659 "Counter": "0,1,2,3",
644660 "EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.REMOTE_HITM",
645661 "MSRIndex": "0x1a6,0x1a7",
662+ "PublicDescription": "Counts all demand data writes (RFOs) that miss the L3 and the modified data is transferred from remote cache Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
663+ "SampleAfterValue": "100003",
664+ "CounterHTOff": "0,1,2,3"
665+ },
666+ {
667+ "Offcore": "1",
668+ "EventCode": "0xB7, 0xBB",
669+ "UMask": "0x1",
670+ "BriefDescription": "Counts all demand data writes (RFOs) that miss in the L3",
671+ "MSRValue": "0x3fbfc00002",
672+ "Counter": "0,1,2,3",
673+ "EventName": "OFFCORE_RESPONSE.DEMAND_RFO.LLC_MISS.ANY_RESPONSE",
674+ "MSRIndex": "0x1a6,0x1a7",
675+ "PublicDescription": "Counts all demand data writes (RFOs) that miss in the L3 Offcore response can be programmed only with a specific pair of event select and counter MSR, and with specific event codes and predefine mask bit value in a dedicated MSR to specify attributes of the offcore transaction.",
646676 "SampleAfterValue": "100003",
647677 "CounterHTOff": "0,1,2,3"
648678 }
tools/perf/pmu-events/arch/x86/broadwellx/other.json+10−10
@@ -9,16 +9,6 @@ · 8 unmodified lines
99 "SampleAfterValue": "2000003",
1010 "CounterHTOff": "0,1,2,3,4,5,6,7"
1111 },
12− {
13− "EventCode": "0x5C",
14− "UMask": "0x2",
15− "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3",
16− "Counter": "0,1,2,3",
17− "EventName": "CPL_CYCLES.RING123",
18− "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
19− "SampleAfterValue": "2000003",
20− "CounterHTOff": "0,1,2,3,4,5,6,7"
21− },
2212 {
2313 "EdgeDetect": "1",
2414 "EventCode": "0x5C",
@@ -31,6 +21,16 @@ · 6 unmodified lines
3121 "SampleAfterValue": "100007",
3222 "CounterHTOff": "0,1,2,3,4,5,6,7"
3323 },
24+ {
25+ "EventCode": "0x5C",
26+ "UMask": "0x2",
27+ "BriefDescription": "Unhalted core cycles when thread is in rings 1, 2, or 3",
28+ "Counter": "0,1,2,3",
29+ "EventName": "CPL_CYCLES.RING123",
30+ "PublicDescription": "This event counts unhalted core cycles during which the thread is in rings 1, 2, or 3.",
31+ "SampleAfterValue": "2000003",
32+ "CounterHTOff": "0,1,2,3,4,5,6,7"
33+ },
3434 {
3535 "EventCode": "0x63",
3636 "UMask": "0x1",
tools/perf/pmu-events/arch/x86/broadwellx/pipeline.json+612−602
@@ -3,31 +3,41 @@ · 2 unmodified lines
33 "EventCode": "0x00",
44 "UMask": "0x1",
55 "BriefDescription": "Instructions retired from execution.",
6− "Counter": "Fixed counter 1",
6+ "Counter": "Fixed counter 0",
77 "EventName": "INST_RETIRED.ANY",
88 "PublicDescription": "This event counts the number of instructions retired from execution. For instructions that consist of multiple micro-ops, this event counts the retirement of the last micro-op of the instruction. Counting continues during hardware interrupts, traps, and inside interrupt handlers. \nNotes: INST_RETIRED.ANY is counted by a designated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. INST_RETIRED.ANY_P is counted by a programmable counter and it is an architectural performance event. \nCounting: Faulting executions of GETSEC/VM entry/VM Exit/MWait will not count as retired instructions.",
99 "SampleAfterValue": "2000003",
10− "CounterHTOff": "Fixed counter 1"
10+ "CounterHTOff": "Fixed counter 0"
1111 },
1212 {
1313 "EventCode": "0x00",
1414 "UMask": "0x2",
1515 "BriefDescription": "Core cycles when the thread is not in halt state",
16− "Counter": "Fixed counter 2",
16+ "Counter": "Fixed counter 1",
1717 "EventName": "CPU_CLK_UNHALTED.THREAD",
1818 "PublicDescription": "This event counts the number of core cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. This event is a component in many key event ratios. The core frequency may change from time to time due to transitions associated with Enhanced Intel SpeedStep Technology or TM2. For this reason this event may have a changing ratio with regards to time. When the core frequency is constant, this event can approximate elapsed time while the core was not in the halt state. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events.",
1919 "SampleAfterValue": "2000003",
20− "CounterHTOff": "Fixed counter 2"
20+ "CounterHTOff": "Fixed counter 1"
21+ },
22+ {
23+ "EventCode": "0x00",
24+ "UMask": "0x2",
25+ "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
26+ "Counter": "Fixed counter 1",
27+ "EventName": "CPU_CLK_UNHALTED.THREAD_ANY",
28+ "AnyThread": "1",
29+ "SampleAfterValue": "2000003",
30+ "CounterHTOff": "Fixed counter 1"
2131 },
2232 {
2333 "EventCode": "0x00",
2434 "UMask": "0x3",
2535 "BriefDescription": "Reference cycles when the core is not in halt state.",
26− "Counter": "Fixed counter 3",
36+ "Counter": "Fixed counter 2",
2737 "EventName": "CPU_CLK_UNHALTED.REF_TSC",
2838 "PublicDescription": "This event counts the number of reference cycles when the core is not in a halt state. The core enters the halt state when it is running the HLT instruction or the MWAIT instruction. This event is not affected by core frequency changes (for example, P states, TM2 transitions) but has the same incrementing frequency as the time stamp counter. This event can approximate elapsed time while the core was not in a halt state. This event has a constant ratio with the CPU_CLK_UNHALTED.REF_XCLK event. It is counted on a dedicated fixed counter, leaving the four (eight when Hyperthreading is disabled) programmable counters available for other events. \nNote: On all current platforms this event stops counting during 'throttling (TM)' states duty off periods the processor is 'halted'. This event is clocked by base clock (100 Mhz) on Sandy Bridge. The counter update is done at a lower clock rate then the core clock the overflow status bit for this counter may appear 'sticky'. After the counter has overflowed and software clears the overflow status bit and resets the counter to less than MAX. The reset value to the counter is not clocked immediately so the overflow status bit will flip 'high (1)' and generate another PMI (if enabled) after which the reset value gets clocked into the counter. Therefore, software will get the interrupt, read the overflow status bit '1 for bit 34 while the counter value is less than MAX. Software should ignore this case.",
2939 "SampleAfterValue": "2000003",
30− "CounterHTOff": "Fixed counter 3"
40+ "CounterHTOff": "Fixed counter 2"
3141 },
3242 {
3343 "EventCode": "0x03",
@@ -60,22 +70,33 @@ · 26 unmodified lines
6070 },
6171 {
6272 "EventCode": "0x0D",
63− "UMask": "0x8",
64− "BriefDescription": "Cycles when Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the thread",
73+ "UMask": "0x3",
74+ "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for this thread (e.g. misprediction or memory nuke)",
6575 "Counter": "0,1,2,3",
66− "EventName": "INT_MISC.RAT_STALL_CYCLES",
67− "PublicDescription": "This event counts the number of cycles during which Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the current thread. This also includes the cycles during which the Allocator is serving another thread.",
76+ "EventName": "INT_MISC.RECOVERY_CYCLES",
77+ "CounterMask": "1",
78+ "PublicDescription": "Cycles checkpoints in Resource Allocation Table (RAT) are recovering from JEClear or machine clear.",
6879 "SampleAfterValue": "2000003",
6980 "CounterHTOff": "0,1,2,3,4,5,6,7"
7081 },
7182 {
7283 "EventCode": "0x0D",
7384 "UMask": "0x3",
74− "BriefDescription": "Number of cycles waiting for the checkpoints in Resource Allocation Table (RAT) to be recovered after Nuke due to all other cases except JEClear (e.g. whenever a ucode assist is needed like SSE exception, memory disambiguation, etc...)",
85+ "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for any thread running on the physical core (e.g. misprediction or memory nuke).",
7586 "Counter": "0,1,2,3",
76− "EventName": "INT_MISC.RECOVERY_CYCLES",
87+ "EventName": "INT_MISC.RECOVERY_CYCLES_ANY",
88+ "AnyThread": "1",
7789 "CounterMask": "1",
78− "PublicDescription": "Cycles checkpoints in Resource Allocation Table (RAT) are recovering from JEClear or machine clear.",
90+ "SampleAfterValue": "2000003",
91+ "CounterHTOff": "0,1,2,3,4,5,6,7"
92+ },
93+ {
94+ "EventCode": "0x0D",
95+ "UMask": "0x8",
96+ "BriefDescription": "Cycles when Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the thread",
97+ "Counter": "0,1,2,3",
98+ "EventName": "INT_MISC.RAT_STALL_CYCLES",
99+ "PublicDescription": "This event counts the number of cycles during which Resource Allocation Table (RAT) external stall is sent to Instruction Decode Queue (IDQ) for the current thread. This also includes the cycles during which the Allocator is serving another thread.",
79100 "SampleAfterValue": "2000003",
80101 "CounterHTOff": "0,1,2,3,4,5,6,7"
81102 },
@@ -89,6 +110,18 @@ · 7 unmodified lines
89110 "SampleAfterValue": "2000003",
90111 "CounterHTOff": "0,1,2,3,4,5,6,7"
91112 },
113+ {
114+ "Invert": "1",
115+ "EventCode": "0x0E",
116+ "UMask": "0x1",
117+ "BriefDescription": "Cycles when Resource Allocation Table (RAT) does not issue Uops to Reservation Station (RS) for the thread",
118+ "Counter": "0,1,2,3",
119+ "EventName": "UOPS_ISSUED.STALL_CYCLES",
120+ "CounterMask": "1",
121+ "PublicDescription": "This event counts cycles during which the Resource Allocation Table (RAT) does not issue any Uops to the reservation station (RS) for the current thread.",
122+ "SampleAfterValue": "2000003",
123+ "CounterHTOff": "0,1,2,3"
124+ },
92125 {
93126 "EventCode": "0x0E",
94127 "UMask": "0x10",
@@ -117,18 +150,6 @@ · 22 unmodified lines
117150 "SampleAfterValue": "2000003",
118151 "CounterHTOff": "0,1,2,3,4,5,6,7"
119152 },
120− {
121− "Invert": "1",
122− "EventCode": "0x0E",
123− "UMask": "0x1",
124− "BriefDescription": "Cycles when Resource Allocation Table (RAT) does not issue Uops to Reservation Station (RS) for the thread",
125− "Counter": "0,1,2,3",
126− "EventName": "UOPS_ISSUED.STALL_CYCLES",
127− "CounterMask": "1",
128− "PublicDescription": "This event counts cycles during which the Resource Allocation Table (RAT) does not issue any Uops to the reservation station (RS) for the current thread.",
129− "SampleAfterValue": "2000003",
130− "CounterHTOff": "0,1,2,3"
131− },
132153 {
133154 "EventCode": "0x14",
134155 "UMask": "0x1",
@@ -139,6 +160,26 @@ · 4 unmodified lines
139160 "SampleAfterValue": "2000003",
140161 "CounterHTOff": "0,1,2,3,4,5,6,7"
141162 },
163+ {
164+ "EventCode": "0x3C",
165+ "UMask": "0x0",
166+ "BriefDescription": "Thread cycles when thread is not in halt state",
167+ "Counter": "0,1,2,3",
168+ "EventName": "CPU_CLK_UNHALTED.THREAD_P",
169+ "PublicDescription": "This is an architectural event that counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling. For this reason, this event may have a changing ratio with regards to wall clock time.",
170+ "SampleAfterValue": "2000003",
171+ "CounterHTOff": "0,1,2,3,4,5,6,7"
172+ },
173+ {
174+ "EventCode": "0x3C",
175+ "UMask": "0x0",
176+ "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
177+ "Counter": "0,1,2,3",
178+ "EventName": "CPU_CLK_UNHALTED.THREAD_P_ANY",
179+ "AnyThread": "1",
180+ "SampleAfterValue": "2000003",
181+ "CounterHTOff": "0,1,2,3,4,5,6,7"
182+ },
142183 {
143184 "EventCode": "0x3C",
144185 "UMask": "0x1",
@@ -149,6 +190,36 @@ · 4 unmodified lines
149190 "SampleAfterValue": "2000003",
150191 "CounterHTOff": "0,1,2,3,4,5,6,7"
151192 },
193+ {
194+ "EventCode": "0x3C",
195+ "UMask": "0x1",
196+ "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
197+ "Counter": "0,1,2,3",
198+ "EventName": "CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY",
199+ "AnyThread": "1",
200+ "SampleAfterValue": "2000003",
201+ "CounterHTOff": "0,1,2,3,4,5,6,7"
202+ },
203+ {
204+ "EventCode": "0x3C",
205+ "UMask": "0x1",
206+ "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
207+ "Counter": "0,1,2,3",
208+ "EventName": "CPU_CLK_UNHALTED.REF_XCLK",
209+ "PublicDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate).",
210+ "SampleAfterValue": "2000003",
211+ "CounterHTOff": "0,1,2,3,4,5,6,7"
212+ },
213+ {
214+ "EventCode": "0x3C",
215+ "UMask": "0x1",
216+ "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
217+ "Counter": "0,1,2,3",
218+ "EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
219+ "AnyThread": "1",
220+ "SampleAfterValue": "2000003",
221+ "CounterHTOff": "0,1,2,3,4,5,6,7"
222+ },
152223 {
153224 "EventCode": "0x3c",
154225 "UMask": "0x2",
@@ -158,6 +229,15 @@ · 3 unmodified lines
158229 "SampleAfterValue": "2000003",
159230 "CounterHTOff": "0,1,2,3"
160231 },
232+ {
233+ "EventCode": "0x3C",
234+ "UMask": "0x2",
235+ "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
236+ "Counter": "0,1,2,3",
237+ "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE",
238+ "SampleAfterValue": "2000003",
239+ "CounterHTOff": "0,1,2,3,4,5,6,7"
240+ },
161241 {
162242 "EventCode": "0x4c",
163243 "UMask": "0x1",
@@ -224,6 +304,18 @@ · 60 unmodified lines
224304 "SampleAfterValue": "2000003",
225305 "CounterHTOff": "0,1,2,3,4,5,6,7"
226306 },
307+ {
308+ "EdgeDetect": "1",
309+ "Invert": "1",
310+ "EventCode": "0x5E",
311+ "UMask": "0x1",
312+ "BriefDescription": "Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues.",
313+ "Counter": "0,1,2,3",
314+ "EventName": "RS_EVENTS.EMPTY_END",
315+ "CounterMask": "1",
316+ "SampleAfterValue": "200003",
317+ "CounterHTOff": "0,1,2,3,4,5,6,7"
318+ },
227319 {
228320 "EventCode": "0x87",
229321 "UMask": "0x1",
@@ -404,6 +496,15 @@ · 174 unmodified lines
404496 "SampleAfterValue": "200003",
405497 "CounterHTOff": "0,1,2,3,4,5,6,7"
406498 },
499+ {
500+ "EventCode": "0x89",
501+ "UMask": "0xa0",
502+ "BriefDescription": "Taken speculative and retired mispredicted indirect calls.",
503+ "Counter": "0,1,2,3",
504+ "EventName": "BR_MISP_EXEC.TAKEN_INDIRECT_NEAR_CALL",
505+ "SampleAfterValue": "200003",
506+ "CounterHTOff": "0,1,2,3,4,5,6,7"
507+ },
407508 {
408509 "EventCode": "0x89",
409510 "UMask": "0xc1",
@@ -434,6 +535,16 @@ · 24 unmodified lines
434535 "SampleAfterValue": "200003",
435536 "CounterHTOff": "0,1,2,3,4,5,6,7"
436537 },
538+ {
539+ "EventCode": "0xA0",
540+ "UMask": "0x3",
541+ "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports",
542+ "Counter": "0,1,2,3",
543+ "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF",
544+ "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
545+ "SampleAfterValue": "2000003",
546+ "CounterHTOff": "0,1,2,3"
547+ },
437548 {
438549 "EventCode": "0xA1",
439550 "UMask": "0x1",
@@ -446,601 +557,471 @@ · 6 unmodified lines
446557 },
447558 {
448559 "EventCode": "0xA1",
449− "UMask": "0x2",
450− "BriefDescription": "Cycles per thread when uops are executed in port 1",
560+ "UMask": "0x1",
561+ "BriefDescription": "Cycles per core when uops are exectuted in port 0.",
451562 "Counter": "0,1,2,3",
452− "EventName": "UOPS_DISPATCHED_PORT.PORT_1",
453− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
563+ "EventName": "UOPS_EXECUTED_PORT.PORT_0_CORE",
564+ "AnyThread": "1",
454565 "SampleAfterValue": "2000003",
455566 "CounterHTOff": "0,1,2,3,4,5,6,7"
456567 },
457568 {
458569 "EventCode": "0xA1",
459− "UMask": "0x4",
460− "BriefDescription": "Cycles per thread when uops are executed in port 2",
570+ "UMask": "0x1",
571+ "BriefDescription": "Cycles per thread when uops are executed in port 0",
461572 "Counter": "0,1,2,3",
462− "EventName": "UOPS_DISPATCHED_PORT.PORT_2",
463− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
573+ "EventName": "UOPS_EXECUTED_PORT.PORT_0",
574+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
464575 "SampleAfterValue": "2000003",
465576 "CounterHTOff": "0,1,2,3,4,5,6,7"
466577 },
467578 {
468579 "EventCode": "0xA1",
469− "UMask": "0x8",
470− "BriefDescription": "Cycles per thread when uops are executed in port 3",
580+ "UMask": "0x2",
581+ "BriefDescription": "Cycles per thread when uops are executed in port 1",
471582 "Counter": "0,1,2,3",
472− "EventName": "UOPS_DISPATCHED_PORT.PORT_3",
473− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
583+ "EventName": "UOPS_DISPATCHED_PORT.PORT_1",
584+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
474585 "SampleAfterValue": "2000003",
475586 "CounterHTOff": "0,1,2,3,4,5,6,7"
476587 },
477588 {
478589 "EventCode": "0xA1",
479− "UMask": "0x10",
480− "BriefDescription": "Cycles per thread when uops are executed in port 4",
590+ "UMask": "0x2",
591+ "BriefDescription": "Cycles per core when uops are exectuted in port 1.",
481592 "Counter": "0,1,2,3",
482− "EventName": "UOPS_DISPATCHED_PORT.PORT_4",
483− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
593+ "EventName": "UOPS_EXECUTED_PORT.PORT_1_CORE",
594+ "AnyThread": "1",
484595 "SampleAfterValue": "2000003",
485596 "CounterHTOff": "0,1,2,3,4,5,6,7"
486597 },
487598 {
488599 "EventCode": "0xA1",
489− "UMask": "0x20",
490− "BriefDescription": "Cycles per thread when uops are executed in port 5",
600+ "UMask": "0x2",
601+ "BriefDescription": "Cycles per thread when uops are executed in port 1",
491602 "Counter": "0,1,2,3",
492− "EventName": "UOPS_DISPATCHED_PORT.PORT_5",
493− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
603+ "EventName": "UOPS_EXECUTED_PORT.PORT_1",
604+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
494605 "SampleAfterValue": "2000003",
495606 "CounterHTOff": "0,1,2,3,4,5,6,7"
496607 },
497608 {
498609 "EventCode": "0xA1",
499− "UMask": "0x40",
500− "BriefDescription": "Cycles per thread when uops are executed in port 6",
610+ "UMask": "0x4",
611+ "BriefDescription": "Cycles per thread when uops are executed in port 2",
501612 "Counter": "0,1,2,3",
502− "EventName": "UOPS_DISPATCHED_PORT.PORT_6",
503− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
613+ "EventName": "UOPS_DISPATCHED_PORT.PORT_2",
614+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
504615 "SampleAfterValue": "2000003",
505616 "CounterHTOff": "0,1,2,3,4,5,6,7"
506617 },
507618 {
508619 "EventCode": "0xA1",
509− "UMask": "0x80",
510− "BriefDescription": "Cycles per thread when uops are executed in port 7",
620+ "UMask": "0x4",
621+ "BriefDescription": "Cycles per core when uops are dispatched to port 2.",
511622 "Counter": "0,1,2,3",
512− "EventName": "UOPS_DISPATCHED_PORT.PORT_7",
513− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
623+ "EventName": "UOPS_EXECUTED_PORT.PORT_2_CORE",
624+ "AnyThread": "1",
514625 "SampleAfterValue": "2000003",
515626 "CounterHTOff": "0,1,2,3,4,5,6,7"
516627 },
517628 {
518− "EventCode": "0xA2",
519− "UMask": "0x1",
520− "BriefDescription": "Resource-related stall cycles",
629+ "EventCode": "0xA1",
630+ "UMask": "0x4",
631+ "BriefDescription": "Cycles per thread when uops are executed in port 2",
521632 "Counter": "0,1,2,3",
522− "EventName": "RESOURCE_STALLS.ANY",
523− "PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
633+ "EventName": "UOPS_EXECUTED_PORT.PORT_2",
634+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
524635 "SampleAfterValue": "2000003",
525636 "CounterHTOff": "0,1,2,3,4,5,6,7"
526637 },
527638 {
528− "EventCode": "0xA2",
529− "UMask": "0x4",
530− "BriefDescription": "Cycles stalled due to no eligible RS entry available.",
639+ "EventCode": "0xA1",
640+ "UMask": "0x8",
641+ "BriefDescription": "Cycles per thread when uops are executed in port 3",
531642 "Counter": "0,1,2,3",
532− "EventName": "RESOURCE_STALLS.RS",
533− "PublicDescription": "This event counts stall cycles caused by absence of eligible entries in the reservation station (RS). This may result from RS overflow, or from RS deallocation because of the RS array Write Port allocation scheme (each RS entry has two write ports instead of four. As a result, empty entries could not be used, although RS is not really full). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
643+ "EventName": "UOPS_DISPATCHED_PORT.PORT_3",
644+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
534645 "SampleAfterValue": "2000003",
535646 "CounterHTOff": "0,1,2,3,4,5,6,7"
536647 },
537648 {
538− "EventCode": "0xA2",
649+ "EventCode": "0xA1",
539650 "UMask": "0x8",
540− "BriefDescription": "Cycles stalled due to no store buffers available. (not including draining form sync).",
651+ "BriefDescription": "Cycles per core when uops are dispatched to port 3.",
541652 "Counter": "0,1,2,3",
542− "EventName": "RESOURCE_STALLS.SB",
543− "PublicDescription": "This event counts stall cycles caused by the store buffer (SB) overflow (excluding draining from synch). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
653+ "EventName": "UOPS_EXECUTED_PORT.PORT_3_CORE",
654+ "AnyThread": "1",
544655 "SampleAfterValue": "2000003",
545656 "CounterHTOff": "0,1,2,3,4,5,6,7"
546657 },
547658 {
548− "EventCode": "0xA2",
549− "UMask": "0x10",
550− "BriefDescription": "Cycles stalled due to re-order buffer full.",
659+ "EventCode": "0xA1",
660+ "UMask": "0x8",
661+ "BriefDescription": "Cycles per thread when uops are executed in port 3",
551662 "Counter": "0,1,2,3",
552− "EventName": "RESOURCE_STALLS.ROB",
553− "PublicDescription": "This event counts ROB full stall cycles. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
663+ "EventName": "UOPS_EXECUTED_PORT.PORT_3",
664+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
554665 "SampleAfterValue": "2000003",
555666 "CounterHTOff": "0,1,2,3,4,5,6,7"
556667 },
557668 {
558− "EventCode": "0xA3",
559− "UMask": "0x1",
560− "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
669+ "EventCode": "0xA1",
670+ "UMask": "0x10",
671+ "BriefDescription": "Cycles per thread when uops are executed in port 4",
561672 "Counter": "0,1,2,3",
562− "EventName": "CYCLE_ACTIVITY.CYCLES_L2_PENDING",
563− "CounterMask": "1",
564− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand* load request missing the L2 cache.",
673+ "EventName": "UOPS_DISPATCHED_PORT.PORT_4",
674+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
565675 "SampleAfterValue": "2000003",
566676 "CounterHTOff": "0,1,2,3,4,5,6,7"
567677 },
568678 {
569− "EventCode": "0xA3",
570− "UMask": "0x8",
571− "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
572− "Counter": "2",
573− "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_PENDING",
574− "CounterMask": "8",
575− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request missing the L1 data cache.",
576− "SampleAfterValue": "2000003",
577− "CounterHTOff": "2"
578− },
579− {
580− "EventCode": "0xA3",
581− "UMask": "0x2",
582− "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
679+ "EventCode": "0xA1",
680+ "UMask": "0x10",
681+ "BriefDescription": "Cycles per core when uops are exectuted in port 4.",
583682 "Counter": "0,1,2,3",
584− "EventName": "CYCLE_ACTIVITY.CYCLES_LDM_PENDING",
585− "CounterMask": "2",
586− "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request (that is cycles with non-completed load waiting for its data from memory subsystem).",
683+ "EventName": "UOPS_EXECUTED_PORT.PORT_4_CORE",
684+ "AnyThread": "1",
587685 "SampleAfterValue": "2000003",
588686 "CounterHTOff": "0,1,2,3,4,5,6,7"
589687 },
590688 {
591− "EventCode": "0xA3",
592− "UMask": "0x4",
593− "BriefDescription": "Total execution stalls",
689+ "EventCode": "0xA1",
690+ "UMask": "0x10",
691+ "BriefDescription": "Cycles per thread when uops are executed in port 4",
594692 "Counter": "0,1,2,3",
595− "EventName": "CYCLE_ACTIVITY.CYCLES_NO_EXECUTE",
596− "CounterMask": "4",
597− "PublicDescription": "Counts number of cycles nothing is executed on any execution port.",
693+ "EventName": "UOPS_EXECUTED_PORT.PORT_4",
694+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
598695 "SampleAfterValue": "2000003",
599− "CounterHTOff": "0,1,2,3"
696+ "CounterHTOff": "0,1,2,3,4,5,6,7"
600697 },
601698 {
602− "EventCode": "0xA3",
603− "UMask": "0x5",
604− "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
699+ "EventCode": "0xA1",
700+ "UMask": "0x20",
701+ "BriefDescription": "Cycles per thread when uops are executed in port 5",
605702 "Counter": "0,1,2,3",
606− "EventName": "CYCLE_ACTIVITY.STALLS_L2_PENDING",
607− "CounterMask": "5",
608− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand* load request missing the L2 cache.(as a footprint) * includes also L1 HW prefetch requests that may or may not be required by demands.",
703+ "EventName": "UOPS_DISPATCHED_PORT.PORT_5",
704+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
609705 "SampleAfterValue": "2000003",
610− "CounterHTOff": "0,1,2,3"
706+ "CounterHTOff": "0,1,2,3,4,5,6,7"
611707 },
612708 {
613− "EventCode": "0xA3",
614− "UMask": "0x6",
615− "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
709+ "EventCode": "0xA1",
710+ "UMask": "0x20",
711+ "BriefDescription": "Cycles per core when uops are exectuted in port 5.",
616712 "Counter": "0,1,2,3",
617− "EventName": "CYCLE_ACTIVITY.STALLS_LDM_PENDING",
618− "CounterMask": "6",
619− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request.",
713+ "EventName": "UOPS_EXECUTED_PORT.PORT_5_CORE",
714+ "AnyThread": "1",
620715 "SampleAfterValue": "2000003",
621− "CounterHTOff": "0,1,2,3"
716+ "CounterHTOff": "0,1,2,3,4,5,6,7"
622717 },
623718 {
624− "EventCode": "0xA3",
625− "UMask": "0xc",
626− "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
627− "Counter": "2",
628− "EventName": "CYCLE_ACTIVITY.STALLS_L1D_PENDING",
629− "CounterMask": "12",
630− "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request missing the L1 data cache.",
719+ "EventCode": "0xA1",
720+ "UMask": "0x20",
721+ "BriefDescription": "Cycles per thread when uops are executed in port 5",
722+ "Counter": "0,1,2,3",
723+ "EventName": "UOPS_EXECUTED_PORT.PORT_5",
724+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
631725 "SampleAfterValue": "2000003",
632− "CounterHTOff": "2"
726+ "CounterHTOff": "0,1,2,3,4,5,6,7"
633727 },
634728 {
635− "EventCode": "0xA8",
636− "UMask": "0x1",
637− "BriefDescription": "Number of Uops delivered by the LSD.",
729+ "EventCode": "0xA1",
730+ "UMask": "0x40",
731+ "BriefDescription": "Cycles per thread when uops are executed in port 6",
638732 "Counter": "0,1,2,3",
639− "EventName": "LSD.UOPS",
640− "PublicDescription": "Number of Uops delivered by the LSD. ",
733+ "EventName": "UOPS_DISPATCHED_PORT.PORT_6",
734+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
641735 "SampleAfterValue": "2000003",
642736 "CounterHTOff": "0,1,2,3,4,5,6,7"
643737 },
644738 {
645− "EventCode": "0xB1",
646− "UMask": "0x1",
647− "BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
739+ "EventCode": "0xA1",
740+ "UMask": "0x40",
741+ "BriefDescription": "Cycles per core when uops are exectuted in port 6.",
648742 "Counter": "0,1,2,3",
649− "EventName": "UOPS_EXECUTED.THREAD",
650− "PublicDescription": "Number of uops to be executed per-thread each cycle.",
743+ "EventName": "UOPS_EXECUTED_PORT.PORT_6_CORE",
744+ "AnyThread": "1",
651745 "SampleAfterValue": "2000003",
652746 "CounterHTOff": "0,1,2,3,4,5,6,7"
653747 },
654748 {
655− "EventCode": "0xB1",
656− "UMask": "0x2",
657− "BriefDescription": "Number of uops executed on the core.",
749+ "EventCode": "0xA1",
750+ "UMask": "0x40",
751+ "BriefDescription": "Cycles per thread when uops are executed in port 6",
658752 "Counter": "0,1,2,3",
659− "EventName": "UOPS_EXECUTED.CORE",
660− "PublicDescription": "Number of uops executed from any thread.",
753+ "EventName": "UOPS_EXECUTED_PORT.PORT_6",
754+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
661755 "SampleAfterValue": "2000003",
662756 "CounterHTOff": "0,1,2,3,4,5,6,7"
663757 },
664758 {
665− "Invert": "1",
666− "EventCode": "0xB1",
667− "UMask": "0x1",
668− "BriefDescription": "Counts number of cycles no uops were dispatched to be executed on this thread.",
759+ "EventCode": "0xA1",
760+ "UMask": "0x80",
761+ "BriefDescription": "Cycles per thread when uops are executed in port 7",
669762 "Counter": "0,1,2,3",
670− "EventName": "UOPS_EXECUTED.STALL_CYCLES",
671− "CounterMask": "1",
672− "PublicDescription": "This event counts cycles during which no uops were dispatched from the Reservation Station (RS) per thread.",
763+ "EventName": "UOPS_DISPATCHED_PORT.PORT_7",
764+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
673765 "SampleAfterValue": "2000003",
674− "CounterHTOff": "0,1,2,3"
766+ "CounterHTOff": "0,1,2,3,4,5,6,7"
675767 },
676768 {
677− "EventCode": "0xC0",
678− "UMask": "0x0",
679− "BriefDescription": "Number of instructions retired. General Counter - architectural event",
769+ "EventCode": "0xA1",
770+ "UMask": "0x80",
771+ "BriefDescription": "Cycles per core when uops are dispatched to port 7.",
680772 "Counter": "0,1,2,3",
681− "EventName": "INST_RETIRED.ANY_P",
682− "Errata": "BDM61",
683− "PublicDescription": "This event counts the number of instructions (EOMs) retired. Counting covers macro-fused instructions individually (that is, increments by two).",
773+ "EventName": "UOPS_EXECUTED_PORT.PORT_7_CORE",
774+ "AnyThread": "1",
684775 "SampleAfterValue": "2000003",
685776 "CounterHTOff": "0,1,2,3,4,5,6,7"
686777 },
687778 {
688− "EventCode": "0xC0",
689− "UMask": "0x2",
690− "BriefDescription": "FP operations retired. X87 FP operations that have no exceptions:",
779+ "EventCode": "0xA1",
780+ "UMask": "0x80",
781+ "BriefDescription": "Cycles per thread when uops are executed in port 7",
691782 "Counter": "0,1,2,3",
692− "EventName": "INST_RETIRED.X87",
693− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
783+ "EventName": "UOPS_EXECUTED_PORT.PORT_7",
784+ "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
694785 "SampleAfterValue": "2000003",
695786 "CounterHTOff": "0,1,2,3,4,5,6,7"
696787 },
697788 {
698− "EventCode": "0xC0",
789+ "EventCode": "0xA2",
699790 "UMask": "0x1",
700− "BriefDescription": "Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution",
701− "PEBS": "2",
702− "Counter": "1",
703− "EventName": "INST_RETIRED.PREC_DIST",
704− "Errata": "BDM11, BDM55",
705− "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts instructions retired.",
791+ "BriefDescription": "Resource-related stall cycles",
792+ "Counter": "0,1,2,3",
793+ "EventName": "RESOURCE_STALLS.ANY",
794+ "PublicDescription": "This event counts resource-related stall cycles. Reasons for stalls can be as follows:\n - *any* u-arch structure got full (LB, SB, RS, ROB, BOB, LM, Physical Register Reclaim Table (PRRT), or Physical History Table (PHT) slots)\n - *any* u-arch structure got empty (like INT/SIMD FreeLists)\n - FPU control word (FPCW), MXCSR\nand others. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
706795 "SampleAfterValue": "2000003",
707− "CounterHTOff": "1"
796+ "CounterHTOff": "0,1,2,3,4,5,6,7"
708797 },
709798 {
710− "EventCode": "0xC1",
711− "UMask": "0x40",
712− "BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
799+ "EventCode": "0xA2",
800+ "UMask": "0x4",
801+ "BriefDescription": "Cycles stalled due to no eligible RS entry available.",
713802 "Counter": "0,1,2,3",
714− "EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
715− "SampleAfterValue": "100003",
803+ "EventName": "RESOURCE_STALLS.RS",
804+ "PublicDescription": "This event counts stall cycles caused by absence of eligible entries in the reservation station (RS). This may result from RS overflow, or from RS deallocation because of the RS array Write Port allocation scheme (each RS entry has two write ports instead of four. As a result, empty entries could not be used, although RS is not really full). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
805+ "SampleAfterValue": "2000003",
716806 "CounterHTOff": "0,1,2,3,4,5,6,7"
717807 },
718808 {
719− "EventCode": "0xC2",
720− "UMask": "0x1",
721− "BriefDescription": "Actually retired uops.",
722− "Data_LA": "1",
723− "PEBS": "1",
809+ "EventCode": "0xA2",
810+ "UMask": "0x8",
811+ "BriefDescription": "Cycles stalled due to no store buffers available. (not including draining form sync).",
724812 "Counter": "0,1,2,3",
725− "EventName": "UOPS_RETIRED.ALL",
726− "PublicDescription": "This event counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
813+ "EventName": "RESOURCE_STALLS.SB",
814+ "PublicDescription": "This event counts stall cycles caused by the store buffer (SB) overflow (excluding draining from synch). This counts cycles that the pipeline backend blocked uop delivery from the front end.",
727815 "SampleAfterValue": "2000003",
728816 "CounterHTOff": "0,1,2,3,4,5,6,7"
729817 },
730818 {
731− "EventCode": "0xC2",
732− "UMask": "0x2",
733− "BriefDescription": "Retirement slots used.",
734− "PEBS": "1",
819+ "EventCode": "0xA2",
820+ "UMask": "0x10",
821+ "BriefDescription": "Cycles stalled due to re-order buffer full.",
735822 "Counter": "0,1,2,3",
736− "EventName": "UOPS_RETIRED.RETIRE_SLOTS",
737− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts the number of retirement slots used.",
823+ "EventName": "RESOURCE_STALLS.ROB",
824+ "PublicDescription": "This event counts ROB full stall cycles. This counts cycles that the pipeline backend blocked uop delivery from the front end.",
738825 "SampleAfterValue": "2000003",
739826 "CounterHTOff": "0,1,2,3,4,5,6,7"
740827 },
741828 {
742− "Invert": "1",
743− "EventCode": "0xC2",
829+ "EventCode": "0xA3",
744830 "UMask": "0x1",
745− "BriefDescription": "Cycles without actually retired uops.",
831+ "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
746832 "Counter": "0,1,2,3",
747− "EventName": "UOPS_RETIRED.STALL_CYCLES",
833+ "EventName": "CYCLE_ACTIVITY.CYCLES_L2_PENDING",
748834 "CounterMask": "1",
749− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts cycles without actually retired uops.",
835+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand* load request missing the L2 cache.",
750836 "SampleAfterValue": "2000003",
751− "CounterHTOff": "0,1,2,3"
837+ "CounterHTOff": "0,1,2,3,4,5,6,7"
752838 },
753839 {
754− "Invert": "1",
755− "EventCode": "0xC2",
840+ "EventCode": "0xA3",
756841 "UMask": "0x1",
757− "BriefDescription": "Cycles with less than 10 actually retired uops.",
842+ "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
758843 "Counter": "0,1,2,3",
759− "EventName": "UOPS_RETIRED.TOTAL_CYCLES",
760− "CounterMask": "10",
761− "PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
844+ "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS",
845+ "CounterMask": "1",
762846 "SampleAfterValue": "2000003",
763− "CounterHTOff": "0,1,2,3"
847+ "CounterHTOff": "0,1,2,3,4,5,6,7"
764848 },
765849 {
766− "EventCode": "0xC3",
767− "UMask": "0x1",
768− "BriefDescription": "Cycles there was a Nuke. Account for both thread-specific and All Thread Nukes.",
850+ "EventCode": "0xA3",
851+ "UMask": "0x2",
852+ "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
769853 "Counter": "0,1,2,3",
770− "EventName": "MACHINE_CLEARS.CYCLES",
771− "PublicDescription": "This event counts both thread-specific (TS) and all-thread (AT) nukes.",
854+ "EventName": "CYCLE_ACTIVITY.CYCLES_LDM_PENDING",
855+ "CounterMask": "2",
856+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request (that is cycles with non-completed load waiting for its data from memory subsystem).",
772857 "SampleAfterValue": "2000003",
773858 "CounterHTOff": "0,1,2,3,4,5,6,7"
774859 },
775860 {
776− "EventCode": "0xC3",
861+ "EventCode": "0xA3",
862+ "UMask": "0x2",
863+ "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
864+ "Counter": "0,1,2,3",
865+ "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY",
866+ "CounterMask": "2",
867+ "SampleAfterValue": "2000003",
868+ "CounterHTOff": "0,1,2,3"
869+ },
870+ {
871+ "EventCode": "0xA3",
777872 "UMask": "0x4",
778− "BriefDescription": "Self-modifying code (SMC) detected.",
873+ "BriefDescription": "This event increments by 1 for every cycle where there was no execute for this thread.",
779874 "Counter": "0,1,2,3",
780− "EventName": "MACHINE_CLEARS.SMC",
781− "PublicDescription": "This event counts self-modifying code (SMC) detected, which causes a machine clear.",
782− "SampleAfterValue": "100003",
783− "CounterHTOff": "0,1,2,3,4,5,6,7"
875+ "EventName": "CYCLE_ACTIVITY.CYCLES_NO_EXECUTE",
876+ "CounterMask": "4",
877+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port.",
878+ "SampleAfterValue": "2000003",
879+ "CounterHTOff": "0,1,2,3"
784880 },
785881 {
786− "EventCode": "0xC3",
787− "UMask": "0x20",
788− "BriefDescription": "This event counts the number of executed Intel AVX masked load operations that refer to an illegal address range with the mask bits set to 0.",
882+ "EventCode": "0xA3",
883+ "UMask": "0x4",
884+ "BriefDescription": "Total execution stalls.",
789885 "Counter": "0,1,2,3",
790− "EventName": "MACHINE_CLEARS.MASKMOV",
791− "PublicDescription": "Maskmov false fault - counts number of time ucode passes through Maskmov flow due to instruction's mask being 0 while the flow was completed without raising a fault.",
792− "SampleAfterValue": "100003",
886+ "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL",
887+ "CounterMask": "4",
888+ "SampleAfterValue": "2000003",
793889 "CounterHTOff": "0,1,2,3,4,5,6,7"
794890 },
795891 {
796− "EventCode": "0xC4",
797− "UMask": "0x1",
798− "BriefDescription": "Conditional branch instructions retired.",
799− "PEBS": "1",
800− "Counter": "0,1,2,3",
801− "EventName": "BR_INST_RETIRED.CONDITIONAL",
802− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts conditional branch instructions retired.",
803− "SampleAfterValue": "400009",
804− "CounterHTOff": "0,1,2,3,4,5,6,7"
805− },
806− {
807− "EventCode": "0xC4",
808− "UMask": "0x2",
809− "BriefDescription": "Direct and indirect near call instructions retired.",
810− "PEBS": "1",
811− "Counter": "0,1,2,3",
812− "EventName": "BR_INST_RETIRED.NEAR_CALL",
813− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts both direct and indirect near call instructions retired.",
814− "SampleAfterValue": "100007",
815− "CounterHTOff": "0,1,2,3,4,5,6,7"
816− },
817− {
818− "EventCode": "0xC4",
819− "UMask": "0x0",
820− "BriefDescription": "All (macro) branch instructions retired.",
821− "Counter": "0,1,2,3",
822− "EventName": "BR_INST_RETIRED.ALL_BRANCHES",
823− "PublicDescription": "This event counts all (macro) branch instructions retired.",
824− "SampleAfterValue": "400009",
825− "CounterHTOff": "0,1,2,3,4,5,6,7"
826− },
827− {
828− "EventCode": "0xC4",
829− "UMask": "0x8",
830− "BriefDescription": "Return instructions retired.",
831− "PEBS": "1",
832− "Counter": "0,1,2,3",
833− "EventName": "BR_INST_RETIRED.NEAR_RETURN",
834− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts return instructions retired.",
835− "SampleAfterValue": "100007",
836− "CounterHTOff": "0,1,2,3,4,5,6,7"
837− },
838− {
839− "EventCode": "0xC4",
840− "UMask": "0x10",
841− "BriefDescription": "Not taken branch instructions retired.",
842− "Counter": "0,1,2,3",
843− "EventName": "BR_INST_RETIRED.NOT_TAKEN",
844− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts not taken branch instructions retired.",
845− "SampleAfterValue": "400009",
846− "CounterHTOff": "0,1,2,3,4,5,6,7"
847− },
848− {
849− "EventCode": "0xC4",
850− "UMask": "0x20",
851− "BriefDescription": "Taken branch instructions retired.",
852− "PEBS": "1",
853− "Counter": "0,1,2,3",
854− "EventName": "BR_INST_RETIRED.NEAR_TAKEN",
855− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts taken branch instructions retired.",
856− "SampleAfterValue": "400009",
857− "CounterHTOff": "0,1,2,3,4,5,6,7"
858− },
859− {
860− "EventCode": "0xC4",
861− "UMask": "0x40",
862− "BriefDescription": "Far branch instructions retired.",
863− "Counter": "0,1,2,3",
864− "EventName": "BR_INST_RETIRED.FAR_BRANCH",
865− "Errata": "BDW98",
866− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts far branch instructions retired.",
867− "SampleAfterValue": "100007",
868− "CounterHTOff": "0,1,2,3,4,5,6,7"
869− },
870− {
871− "EventCode": "0xC4",
872− "UMask": "0x4",
873− "BriefDescription": "All (macro) branch instructions retired. (Precise Event - PEBS)",
874− "PEBS": "2",
892+ "EventCode": "0xA3",
893+ "UMask": "0x5",
894+ "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
875895 "Counter": "0,1,2,3",
876− "EventName": "BR_INST_RETIRED.ALL_BRANCHES_PEBS",
877− "Errata": "BDW98",
878− "PublicDescription": "This is a precise version of BR_INST_RETIRED.ALL_BRANCHES that counts all (macro) branch instructions retired.",
879− "SampleAfterValue": "400009",
896+ "EventName": "CYCLE_ACTIVITY.STALLS_L2_PENDING",
897+ "CounterMask": "5",
898+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand* load request missing the L2 cache.(as a footprint) * includes also L1 HW prefetch requests that may or may not be required by demands.",
899+ "SampleAfterValue": "2000003",
880900 "CounterHTOff": "0,1,2,3"
881901 },
882902 {
883− "EventCode": "0xC5",
884− "UMask": "0x1",
885− "BriefDescription": "Mispredicted conditional branch instructions retired.",
886− "PEBS": "1",
887− "Counter": "0,1,2,3",
888− "EventName": "BR_MISP_RETIRED.CONDITIONAL",
889− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted conditional branch instructions retired.",
890− "SampleAfterValue": "400009",
891− "CounterHTOff": "0,1,2,3,4,5,6,7"
892− },
893− {
894− "EventCode": "0xC5",
895− "UMask": "0x0",
896− "BriefDescription": "All mispredicted macro branch instructions retired.",
897− "Counter": "0,1,2,3",
898− "EventName": "BR_MISP_RETIRED.ALL_BRANCHES",
899− "PublicDescription": "This event counts all mispredicted macro branch instructions retired.",
900− "SampleAfterValue": "400009",
901− "CounterHTOff": "0,1,2,3,4,5,6,7"
902− },
903− {
904− "EventCode": "0xC5",
905− "UMask": "0x8",
906− "BriefDescription": "This event counts the number of mispredicted ret instructions retired. Non PEBS",
907− "PEBS": "1",
903+ "EventCode": "0xA3",
904+ "UMask": "0x5",
905+ "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
908906 "Counter": "0,1,2,3",
909− "EventName": "BR_MISP_RETIRED.RET",
910− "PublicDescription": "This is a non-precise version (that is, does not use PEBS) of the event that counts mispredicted return instructions retired.",
911− "SampleAfterValue": "100007",
907+ "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS",
908+ "CounterMask": "5",
909+ "SampleAfterValue": "2000003",
912910 "CounterHTOff": "0,1,2,3,4,5,6,7"
913911 },
914912 {
915− "EventCode": "0xC5",
916− "UMask": "0x4",
917− "BriefDescription": "Mispredicted macro branch instructions retired. (Precise Event - PEBS)",
918− "PEBS": "2",
913+ "EventCode": "0xA3",
914+ "UMask": "0x6",
915+ "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
919916 "Counter": "0,1,2,3",
920− "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_PEBS",
921− "PublicDescription": "This is a precise version of BR_MISP_RETIRED.ALL_BRANCHES that counts all mispredicted macro branch instructions retired.",
922− "SampleAfterValue": "400009",
917+ "EventName": "CYCLE_ACTIVITY.STALLS_LDM_PENDING",
918+ "CounterMask": "6",
919+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request.",
920+ "SampleAfterValue": "2000003",
923921 "CounterHTOff": "0,1,2,3"
924922 },
925923 {
926− "EventCode": "0xCC",
927− "UMask": "0x20",
928− "BriefDescription": "Count cases of saving new LBR",
924+ "EventCode": "0xA3",
925+ "UMask": "0x6",
926+ "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
929927 "Counter": "0,1,2,3",
930− "EventName": "ROB_MISC_EVENTS.LBR_INSERTS",
931− "PublicDescription": "This event counts cases of saving new LBR records by hardware. This assumes proper enabling of LBRs and takes into account LBR filtering done by the LBR_SELECT register.",
928+ "EventName": "CYCLE_ACTIVITY.STALLS_MEM_ANY",
929+ "CounterMask": "6",
932930 "SampleAfterValue": "2000003",
933931 "CounterHTOff": "0,1,2,3,4,5,6,7"
934932 },
935933 {
936− "EventCode": "0x3C",
937− "UMask": "0x0",
938− "BriefDescription": "Thread cycles when thread is not in halt state",
939− "Counter": "0,1,2,3",
940− "EventName": "CPU_CLK_UNHALTED.THREAD_P",
941− "PublicDescription": "This is an architectural event that counts the number of thread cycles while the thread is not in a halt state. The thread enters the halt state when it is running the HLT instruction. The core frequency may change from time to time due to power or thermal throttling. For this reason, this event may have a changing ratio with regards to wall clock time.",
934+ "EventCode": "0xA3",
935+ "UMask": "0x8",
936+ "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
937+ "Counter": "2",
938+ "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_PENDING",
939+ "CounterMask": "8",
940+ "PublicDescription": "Counts number of cycles the CPU has at least one pending demand load request missing the L1 data cache.",
942941 "SampleAfterValue": "2000003",
943− "CounterHTOff": "0,1,2,3,4,5,6,7"
944− },
945− {
946− "EventCode": "0x89",
947− "UMask": "0xa0",
948− "BriefDescription": "Taken speculative and retired mispredicted indirect calls.",
949− "Counter": "0,1,2,3",
950− "EventName": "BR_MISP_EXEC.TAKEN_INDIRECT_NEAR_CALL",
951− "SampleAfterValue": "200003",
952− "CounterHTOff": "0,1,2,3,4,5,6,7"
942+ "CounterHTOff": "2"
953943 },
954944 {
955− "EventCode": "0xA1",
956− "UMask": "0x1",
957− "BriefDescription": "Cycles per core when uops are exectuted in port 0.",
958− "Counter": "0,1,2,3",
959− "EventName": "UOPS_EXECUTED_PORT.PORT_0_CORE",
960− "AnyThread": "1",
945+ "EventCode": "0xA3",
946+ "UMask": "0x8",
947+ "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
948+ "Counter": "2",
949+ "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS",
950+ "CounterMask": "8",
961951 "SampleAfterValue": "2000003",
962− "CounterHTOff": "0,1,2,3,4,5,6,7"
952+ "CounterHTOff": "2"
963953 },
964954 {
965− "EventCode": "0xA1",
966− "UMask": "0x2",
967− "BriefDescription": "Cycles per core when uops are exectuted in port 1.",
968− "Counter": "0,1,2,3",
969− "EventName": "UOPS_EXECUTED_PORT.PORT_1_CORE",
970− "AnyThread": "1",
955+ "EventCode": "0xA3",
956+ "UMask": "0xc",
957+ "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
958+ "Counter": "2",
959+ "EventName": "CYCLE_ACTIVITY.STALLS_L1D_PENDING",
960+ "CounterMask": "12",
961+ "PublicDescription": "Counts number of cycles nothing is executed on any execution port, while there was at least one pending demand load request missing the L1 data cache.",
971962 "SampleAfterValue": "2000003",
972− "CounterHTOff": "0,1,2,3,4,5,6,7"
963+ "CounterHTOff": "2"
973964 },
974965 {
975− "EventCode": "0xA1",
976− "UMask": "0x4",
977− "BriefDescription": "Cycles per core when uops are dispatched to port 2.",
978− "Counter": "0,1,2,3",
979− "EventName": "UOPS_EXECUTED_PORT.PORT_2_CORE",
980− "AnyThread": "1",
966+ "EventCode": "0xA3",
967+ "UMask": "0xc",
968+ "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
969+ "Counter": "2",
970+ "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS",
971+ "CounterMask": "12",
981972 "SampleAfterValue": "2000003",
982− "CounterHTOff": "0,1,2,3,4,5,6,7"
973+ "CounterHTOff": "2"
983974 },
984975 {
985− "EventCode": "0xA1",
986− "UMask": "0x8",
987− "BriefDescription": "Cycles per core when uops are dispatched to port 3.",
976+ "EventCode": "0xA8",
977+ "UMask": "0x1",
978+ "BriefDescription": "Number of Uops delivered by the LSD.",
988979 "Counter": "0,1,2,3",
989− "EventName": "UOPS_EXECUTED_PORT.PORT_3_CORE",
990− "AnyThread": "1",
980+ "EventName": "LSD.UOPS",
991981 "SampleAfterValue": "2000003",
992982 "CounterHTOff": "0,1,2,3,4,5,6,7"
993983 },
994984 {
995− "EventCode": "0xA1",
996− "UMask": "0x10",
997− "BriefDescription": "Cycles per core when uops are exectuted in port 4.",
985+ "EventCode": "0xA8",
986+ "UMask": "0x1",
987+ "BriefDescription": "Cycles 4 Uops delivered by the LSD, but didn't come from the decoder.",
998988 "Counter": "0,1,2,3",
999− "EventName": "UOPS_EXECUTED_PORT.PORT_4_CORE",
1000− "AnyThread": "1",
989+ "EventName": "LSD.CYCLES_4_UOPS",
990+ "CounterMask": "4",
1001991 "SampleAfterValue": "2000003",
1002992 "CounterHTOff": "0,1,2,3,4,5,6,7"
1003993 },
1004994 {
1005− "EventCode": "0xA1",
1006− "UMask": "0x20",
1007− "BriefDescription": "Cycles per core when uops are exectuted in port 5.",
995+ "EventCode": "0xA8",
996+ "UMask": "0x1",
997+ "BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
1008998 "Counter": "0,1,2,3",
1009− "EventName": "UOPS_EXECUTED_PORT.PORT_5_CORE",
1010− "AnyThread": "1",
999+ "EventName": "LSD.CYCLES_ACTIVE",
1000+ "CounterMask": "1",
10111001 "SampleAfterValue": "2000003",
10121002 "CounterHTOff": "0,1,2,3,4,5,6,7"
10131003 },
10141004 {
1015− "EventCode": "0xA1",
1016− "UMask": "0x40",
1017− "BriefDescription": "Cycles per core when uops are exectuted in port 6.",
1005+ "EventCode": "0xB1",
1006+ "UMask": "0x1",
1007+ "BriefDescription": "Counts the number of uops to be executed per-thread each cycle.",
10181008 "Counter": "0,1,2,3",
1019− "EventName": "UOPS_EXECUTED_PORT.PORT_6_CORE",
1020− "AnyThread": "1",
1009+ "EventName": "UOPS_EXECUTED.THREAD",
1010+ "PublicDescription": "Number of uops to be executed per-thread each cycle.",
10211011 "SampleAfterValue": "2000003",
10221012 "CounterHTOff": "0,1,2,3,4,5,6,7"
10231013 },
10241014 {
1025− "EventCode": "0xA1",
1026− "UMask": "0x80",
1027− "BriefDescription": "Cycles per core when uops are dispatched to port 7.",
1015+ "Invert": "1",
1016+ "EventCode": "0xB1",
1017+ "UMask": "0x1",
1018+ "BriefDescription": "Counts number of cycles no uops were dispatched to be executed on this thread.",
10281019 "Counter": "0,1,2,3",
1029− "EventName": "UOPS_EXECUTED_PORT.PORT_7_CORE",
1030− "AnyThread": "1",
1020+ "EventName": "UOPS_EXECUTED.STALL_CYCLES",
1021+ "CounterMask": "1",
1022+ "PublicDescription": "This event counts cycles during which no uops were dispatched from the Reservation Station (RS) per thread.",
10311023 "SampleAfterValue": "2000003",
1032− "CounterHTOff": "0,1,2,3,4,5,6,7"
1033− },
1034− {
1035− "EventCode": "0xC5",
1036− "UMask": "0x20",
1037− "BriefDescription": "number of near branch instructions retired that were mispredicted and taken.",
1038− "PEBS": "1",
1039− "Counter": "0,1,2,3",
1040− "EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
1041− "PublicDescription": "Number of near branch instructions retired that were mispredicted and taken.",
1042− "SampleAfterValue": "400009",
1043− "CounterHTOff": "0,1,2,3,4,5,6,7"
1024+ "CounterHTOff": "0,1,2,3"
10441025 },
10451026 {
10461027 "EventCode": "0xB1",
@@ -1083,335 +1064,364 @@ · 36 unmodified lines
10831064 "CounterHTOff": "0,1,2,3"
10841065 },
10851066 {
1086− "EventCode": "0xe6",
1087− "UMask": "0x1f",
1088− "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.",
1067+ "EventCode": "0xB1",
1068+ "UMask": "0x2",
1069+ "BriefDescription": "Number of uops executed on the core.",
10891070 "Counter": "0,1,2,3",
1090− "EventName": "BACLEARS.ANY",
1091− "SampleAfterValue": "100003",
1092− "CounterHTOff": "0,1,2,3,4,5,6,7"
1093− },
1094− {
1095− "EventCode": "0xA3",
1096− "UMask": "0x8",
1097− "BriefDescription": "Cycles while L1 cache miss demand load is outstanding.",
1098− "Counter": "2",
1099− "EventName": "CYCLE_ACTIVITY.CYCLES_L1D_MISS",
1100− "CounterMask": "8",
1071+ "EventName": "UOPS_EXECUTED.CORE",
1072+ "PublicDescription": "Number of uops executed from any thread.",
11011073 "SampleAfterValue": "2000003",
1102− "CounterHTOff": "2"
1074+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11031075 },
11041076 {
1105− "EventCode": "0xA3",
1106− "UMask": "0x1",
1107− "BriefDescription": "Cycles while L2 cache miss demand load is outstanding.",
1077+ "EventCode": "0xb1",
1078+ "UMask": "0x2",
1079+ "BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
11081080 "Counter": "0,1,2,3",
1109− "EventName": "CYCLE_ACTIVITY.CYCLES_L2_MISS",
1081+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
11101082 "CounterMask": "1",
11111083 "SampleAfterValue": "2000003",
11121084 "CounterHTOff": "0,1,2,3,4,5,6,7"
11131085 },
11141086 {
1115− "EventCode": "0xA3",
1087+ "EventCode": "0xb1",
11161088 "UMask": "0x2",
1117− "BriefDescription": "Cycles while memory subsystem has an outstanding load.",
1089+ "BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
11181090 "Counter": "0,1,2,3",
1119− "EventName": "CYCLE_ACTIVITY.CYCLES_MEM_ANY",
1091+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
11201092 "CounterMask": "2",
11211093 "SampleAfterValue": "2000003",
1122− "CounterHTOff": "0,1,2,3"
1094+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11231095 },
11241096 {
1125− "EventCode": "0xA3",
1126− "UMask": "0x4",
1127− "BriefDescription": "Total execution stalls.",
1097+ "EventCode": "0xb1",
1098+ "UMask": "0x2",
1099+ "BriefDescription": "Cycles at least 3 micro-op is executed from any thread on physical core.",
11281100 "Counter": "0,1,2,3",
1129− "EventName": "CYCLE_ACTIVITY.STALLS_TOTAL",
1130− "CounterMask": "4",
1101+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3",
1102+ "CounterMask": "3",
11311103 "SampleAfterValue": "2000003",
11321104 "CounterHTOff": "0,1,2,3,4,5,6,7"
11331105 },
11341106 {
1135− "EventCode": "0xA3",
1136− "UMask": "0xc",
1137− "BriefDescription": "Execution stalls while L1 cache miss demand load is outstanding.",
1138− "Counter": "2",
1139− "EventName": "CYCLE_ACTIVITY.STALLS_L1D_MISS",
1140− "CounterMask": "12",
1107+ "EventCode": "0xb1",
1108+ "UMask": "0x2",
1109+ "BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
1110+ "Counter": "0,1,2,3",
1111+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
1112+ "CounterMask": "4",
11411113 "SampleAfterValue": "2000003",
1142− "CounterHTOff": "2"
1114+ "CounterHTOff": "0,1,2,3,4,5,6,7"
11431115 },
11441116 {
1145− "EventCode": "0xA3",
1146− "UMask": "0x5",
1147− "BriefDescription": "Execution stalls while L2 cache miss demand load is outstanding.",
1117+ "Invert": "1",
1118+ "EventCode": "0xb1",
1119+ "UMask": "0x2",
1120+ "BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
11481121 "Counter": "0,1,2,3",
1149− "EventName": "CYCLE_ACTIVITY.STALLS_L2_MISS",
1150− "CounterMask": "5",
1122+ "EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
11511123 "SampleAfterValue": "2000003",
11521124 "CounterHTOff": "0,1,2,3,4,5,6,7"
11531125 },
11541126 {
1155− "EventCode": "0xA3",
1156− "UMask": "0x6",
1157− "BriefDescription": "Execution stalls while memory subsystem has an outstanding load.",
1127+ "EventCode": "0xC0",
1128+ "UMask": "0x0",
1129+ "BriefDescription": "Number of instructions retired. General Counter - architectural event",
11581130 "Counter": "0,1,2,3",
1159− "EventName": "CYCLE_ACTIVITY.STALLS_MEM_ANY",
1160− "CounterMask": "6",
1131+ "EventName": "INST_RETIRED.ANY_P",
1132+ "Errata": "BDM61",
1133+ "PublicDescription": "This event counts the number of instructions (EOMs) retired. Counting covers macro-fused instructions individually (that is, increments by two).",
11611134 "SampleAfterValue": "2000003",
11621135 "CounterHTOff": "0,1,2,3,4,5,6,7"
11631136 },
11641137 {
1165− "EdgeDetect": "1",
1166− "EventCode": "0xC3",
1138+ "EventCode": "0xC0",
11671139 "UMask": "0x1",
1168− "BriefDescription": "Number of machine clears (nukes) of any type.",
1140+ "BriefDescription": "Precise instruction retired event with HW to reduce effect of PEBS shadow in IP distribution",
1141+ "PEBS": "2",
1142+ "Counter": "1",
1143+ "EventName": "INST_RETIRED.PREC_DIST",
1144+ "Errata": "BDM11, BDM55",
1145+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts instructions retired.",
1146+ "SampleAfterValue": "2000003",
1147+ "CounterHTOff": "1"
1148+ },
1149+ {
1150+ "EventCode": "0xC0",
1151+ "UMask": "0x2",
1152+ "BriefDescription": "FP operations retired. X87 FP operations that have no exceptions:",
11691153 "Counter": "0,1,2,3",
1170− "EventName": "MACHINE_CLEARS.COUNT",
1171− "CounterMask": "1",
1172− "SampleAfterValue": "100003",
1154+ "EventName": "INST_RETIRED.X87",
1155+ "PublicDescription": "This event counts FP operations retired. For X87 FP operations that have no exceptions counting also includes flows that have several X87, or flows that use X87 uops in the exception handling.",
1156+ "SampleAfterValue": "2000003",
11731157 "CounterHTOff": "0,1,2,3,4,5,6,7"
11741158 },
11751159 {
1176− "EventCode": "0xA8",
1177− "UMask": "0x1",
1178− "BriefDescription": "Cycles 4 Uops delivered by the LSD, but didn't come from the decoder.",
1160+ "EventCode": "0xC1",
1161+ "UMask": "0x40",
1162+ "BriefDescription": "Number of times any microcode assist is invoked by HW upon uop writeback.",
11791163 "Counter": "0,1,2,3",
1180− "EventName": "LSD.CYCLES_4_UOPS",
1181− "CounterMask": "4",
1182− "SampleAfterValue": "2000003",
1164+ "EventName": "OTHER_ASSISTS.ANY_WB_ASSIST",
1165+ "SampleAfterValue": "100003",
11831166 "CounterHTOff": "0,1,2,3,4,5,6,7"
11841167 },
11851168 {
1186− "EdgeDetect": "1",
1187− "Invert": "1",
1188− "EventCode": "0x5E",
1169+ "EventCode": "0xC2",
11891170 "UMask": "0x1",
1190− "BriefDescription": "Counts end of periods where the Reservation Station (RS) was empty. Could be useful to precisely locate Frontend Latency Bound issues.",
1171+ "BriefDescription": "Actually retired uops. (Precise Event - PEBS)",
1172+ "Data_LA": "1",
1173+ "PEBS": "1",
11911174 "Counter": "0,1,2,3",
1192− "EventName": "RS_EVENTS.EMPTY_END",
1193− "CounterMask": "1",
1194− "SampleAfterValue": "200003",
1175+ "EventName": "UOPS_RETIRED.ALL",
1176+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts all actually retired uops. Counting increments by two for micro-fused uops, and by one for macro-fused and other uops. Maximal increment value for one cycle is eight.",
1177+ "SampleAfterValue": "2000003",
11951178 "CounterHTOff": "0,1,2,3,4,5,6,7"
11961179 },
11971180 {
1198− "EventCode": "0xA8",
1181+ "Invert": "1",
1182+ "EventCode": "0xC2",
11991183 "UMask": "0x1",
1200− "BriefDescription": "Cycles Uops delivered by the LSD, but didn't come from the decoder.",
1184+ "BriefDescription": "Cycles without actually retired uops.",
12011185 "Counter": "0,1,2,3",
1202− "EventName": "LSD.CYCLES_ACTIVE",
1186+ "EventName": "UOPS_RETIRED.STALL_CYCLES",
12031187 "CounterMask": "1",
1188+ "PublicDescription": "This event counts cycles without actually retired uops.",
12041189 "SampleAfterValue": "2000003",
1205− "CounterHTOff": "0,1,2,3,4,5,6,7"
1190+ "CounterHTOff": "0,1,2,3"
12061191 },
12071192 {
1208− "EventCode": "0xA1",
1193+ "Invert": "1",
1194+ "EventCode": "0xC2",
12091195 "UMask": "0x1",
1210− "BriefDescription": "Cycles per thread when uops are executed in port 0",
1196+ "BriefDescription": "Cycles with less than 10 actually retired uops.",
12111197 "Counter": "0,1,2,3",
1212− "EventName": "UOPS_EXECUTED_PORT.PORT_0",
1213− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 0.",
1198+ "EventName": "UOPS_RETIRED.TOTAL_CYCLES",
1199+ "CounterMask": "10",
1200+ "PublicDescription": "Number of cycles using always true condition (uops_ret < 16) applied to non PEBS uops retired event.",
12141201 "SampleAfterValue": "2000003",
1215− "CounterHTOff": "0,1,2,3,4,5,6,7"
1202+ "CounterHTOff": "0,1,2,3"
12161203 },
12171204 {
1218− "EventCode": "0xA1",
1205+ "EventCode": "0xC2",
12191206 "UMask": "0x2",
1220− "BriefDescription": "Cycles per thread when uops are executed in port 1",
1207+ "BriefDescription": "Retirement slots used. (Precise Event - PEBS)",
1208+ "PEBS": "1",
12211209 "Counter": "0,1,2,3",
1222− "EventName": "UOPS_EXECUTED_PORT.PORT_1",
1223− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 1.",
1210+ "EventName": "UOPS_RETIRED.RETIRE_SLOTS",
1211+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts the number of retirement slots used.",
12241212 "SampleAfterValue": "2000003",
12251213 "CounterHTOff": "0,1,2,3,4,5,6,7"
12261214 },
12271215 {
1228− "EventCode": "0xA1",
1229− "UMask": "0x4",
1230− "BriefDescription": "Cycles per thread when uops are executed in port 2",
1216+ "EventCode": "0xC3",
1217+ "UMask": "0x1",
1218+ "BriefDescription": "Cycles there was a Nuke. Account for both thread-specific and All Thread Nukes.",
12311219 "Counter": "0,1,2,3",
1232− "EventName": "UOPS_EXECUTED_PORT.PORT_2",
1233− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 2.",
1220+ "EventName": "MACHINE_CLEARS.CYCLES",
1221+ "PublicDescription": "This event counts both thread-specific (TS) and all-thread (AT) nukes.",
12341222 "SampleAfterValue": "2000003",
12351223 "CounterHTOff": "0,1,2,3,4,5,6,7"
12361224 },
12371225 {
1238− "EventCode": "0xA1",
1239− "UMask": "0x8",
1240− "BriefDescription": "Cycles per thread when uops are executed in port 3",
1226+ "EdgeDetect": "1",
1227+ "EventCode": "0xC3",
1228+ "UMask": "0x1",
1229+ "BriefDescription": "Number of machine clears (nukes) of any type.",
12411230 "Counter": "0,1,2,3",
1242− "EventName": "UOPS_EXECUTED_PORT.PORT_3",
1243− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 3.",
1244− "SampleAfterValue": "2000003",
1231+ "EventName": "MACHINE_CLEARS.COUNT",
1232+ "CounterMask": "1",
1233+ "SampleAfterValue": "100003",
12451234 "CounterHTOff": "0,1,2,3,4,5,6,7"
12461235 },
12471236 {
1248− "EventCode": "0xA1",
1249− "UMask": "0x10",
1250− "BriefDescription": "Cycles per thread when uops are executed in port 4",
1237+ "EventCode": "0xC3",
1238+ "UMask": "0x4",
1239+ "BriefDescription": "Self-modifying code (SMC) detected.",
12511240 "Counter": "0,1,2,3",
1252− "EventName": "UOPS_EXECUTED_PORT.PORT_4",
1253− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 4.",
1254− "SampleAfterValue": "2000003",
1241+ "EventName": "MACHINE_CLEARS.SMC",
1242+ "PublicDescription": "This event counts self-modifying code (SMC) detected, which causes a machine clear.",
1243+ "SampleAfterValue": "100003",
12551244 "CounterHTOff": "0,1,2,3,4,5,6,7"
12561245 },
12571246 {
1258− "EventCode": "0xA1",
1247+ "EventCode": "0xC3",
12591248 "UMask": "0x20",
1260− "BriefDescription": "Cycles per thread when uops are executed in port 5",
1249+ "BriefDescription": "This event counts the number of executed Intel AVX masked load operations that refer to an illegal address range with the mask bits set to 0.",
12611250 "Counter": "0,1,2,3",
1262− "EventName": "UOPS_EXECUTED_PORT.PORT_5",
1263− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 5.",
1264− "SampleAfterValue": "2000003",
1251+ "EventName": "MACHINE_CLEARS.MASKMOV",
1252+ "PublicDescription": "Maskmov false fault - counts number of time ucode passes through Maskmov flow due to instruction's mask being 0 while the flow was completed without raising a fault.",
1253+ "SampleAfterValue": "100003",
12651254 "CounterHTOff": "0,1,2,3,4,5,6,7"
12661255 },
12671256 {
1268− "EventCode": "0xA1",
1269− "UMask": "0x40",
1270− "BriefDescription": "Cycles per thread when uops are executed in port 6",
1257+ "EventCode": "0xC4",
1258+ "UMask": "0x0",
1259+ "BriefDescription": "All (macro) branch instructions retired.",
12711260 "Counter": "0,1,2,3",
1272− "EventName": "UOPS_EXECUTED_PORT.PORT_6",
1273− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 6.",
1274− "SampleAfterValue": "2000003",
1261+ "EventName": "BR_INST_RETIRED.ALL_BRANCHES",
1262+ "PublicDescription": "This event counts all (macro) branch instructions retired.",
1263+ "SampleAfterValue": "400009",
12751264 "CounterHTOff": "0,1,2,3,4,5,6,7"
12761265 },
12771266 {
1278− "EventCode": "0xA1",
1279− "UMask": "0x80",
1280− "BriefDescription": "Cycles per thread when uops are executed in port 7",
1267+ "EventCode": "0xC4",
1268+ "UMask": "0x1",
1269+ "BriefDescription": "Conditional branch instructions retired. (Precise Event - PEBS)",
1270+ "PEBS": "1",
12811271 "Counter": "0,1,2,3",
1282− "EventName": "UOPS_EXECUTED_PORT.PORT_7",
1283− "PublicDescription": "This event counts, on the per-thread basis, cycles during which uops are dispatched from the Reservation Station (RS) to port 7.",
1284− "SampleAfterValue": "2000003",
1272+ "EventName": "BR_INST_RETIRED.CONDITIONAL",
1273+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts conditional branch instructions retired.",
1274+ "SampleAfterValue": "400009",
12851275 "CounterHTOff": "0,1,2,3,4,5,6,7"
12861276 },
12871277 {
1288− "EventCode": "0xA0",
1289− "UMask": "0x3",
1290− "BriefDescription": "Micro-op dispatches cancelled due to insufficient SIMD physical register file read ports",
1278+ "EventCode": "0xC4",
1279+ "UMask": "0x2",
1280+ "BriefDescription": "Direct and indirect near call instructions retired. (Precise Event - PEBS)",
1281+ "PEBS": "1",
12911282 "Counter": "0,1,2,3",
1292− "EventName": "UOP_DISPATCHES_CANCELLED.SIMD_PRF",
1293− "PublicDescription": "This event counts the number of micro-operations cancelled after they were dispatched from the scheduler to the execution units when the total number of physical register read ports across all dispatch ports exceeds the read bandwidth of the physical register file. The SIMD_PRF subevent applies to the following instructions: VDPPS, DPPS, VPCMPESTRI, PCMPESTRI, VPCMPESTRM, PCMPESTRM, VFMADD*, VFMADDSUB*, VFMSUB*, VMSUBADD*, VFNMADD*, VFNMSUB*. See the Broadwell Optimization Guide for more information.",
1294− "SampleAfterValue": "2000003",
1295− "CounterHTOff": "0,1,2,3"
1283+ "EventName": "BR_INST_RETIRED.NEAR_CALL",
1284+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect near call instructions retired.",
1285+ "SampleAfterValue": "100007",
1286+ "CounterHTOff": "0,1,2,3,4,5,6,7"
12961287 },
12971288 {
1298− "EventCode": "0x00",
1289+ "EventCode": "0xC4",
12991290 "UMask": "0x2",
1300− "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
1301− "Counter": "Fixed counter 2",
1302− "EventName": "CPU_CLK_UNHALTED.THREAD_ANY",
1303− "AnyThread": "1",
1304− "SampleAfterValue": "2000003",
1305− "CounterHTOff": "Fixed counter 2"
1291+ "BriefDescription": "Direct and indirect macro near call instructions retired (captured in ring 3). (Precise Event - PEBS)",
1292+ "PEBS": "1",
1293+ "Counter": "0,1,2,3",
1294+ "EventName": "BR_INST_RETIRED.NEAR_CALL_R3",
1295+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts both direct and indirect macro near call instructions retired (captured in ring 3).",
1296+ "SampleAfterValue": "100007",
1297+ "CounterHTOff": "0,1,2,3,4,5,6,7"
13061298 },
13071299 {
1308− "EventCode": "0x3C",
1309− "UMask": "0x0",
1310− "BriefDescription": "Core cycles when at least one thread on the physical core is not in halt state.",
1300+ "EventCode": "0xC4",
1301+ "UMask": "0x4",
1302+ "BriefDescription": "All (macro) branch instructions retired. (Precise Event - PEBS)",
1303+ "PEBS": "2",
13111304 "Counter": "0,1,2,3",
1312− "EventName": "CPU_CLK_UNHALTED.THREAD_P_ANY",
1313− "AnyThread": "1",
1314− "SampleAfterValue": "2000003",
1315− "CounterHTOff": "0,1,2,3,4,5,6,7"
1305+ "EventName": "BR_INST_RETIRED.ALL_BRANCHES_PEBS",
1306+ "Errata": "BDW98",
1307+ "PublicDescription": "This is a precise version of BR_INST_RETIRED.ALL_BRANCHES that counts all (macro) branch instructions retired.",
1308+ "SampleAfterValue": "400009",
1309+ "CounterHTOff": "0,1,2,3"
13161310 },
13171311 {
1318− "EventCode": "0x3C",
1319− "UMask": "0x1",
1320− "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
1312+ "EventCode": "0xC4",
1313+ "UMask": "0x8",
1314+ "BriefDescription": "Return instructions retired. (Precise Event - PEBS)",
1315+ "PEBS": "1",
13211316 "Counter": "0,1,2,3",
1322− "EventName": "CPU_CLK_THREAD_UNHALTED.REF_XCLK_ANY",
1323− "AnyThread": "1",
1324− "SampleAfterValue": "2000003",
1317+ "EventName": "BR_INST_RETIRED.NEAR_RETURN",
1318+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts return instructions retired.",
1319+ "SampleAfterValue": "100007",
13251320 "CounterHTOff": "0,1,2,3,4,5,6,7"
13261321 },
13271322 {
1328− "EventCode": "0x0D",
1329− "UMask": "0x3",
1330− "BriefDescription": "Core cycles the allocator was stalled due to recovery from earlier clear event for any thread running on the physical core (e.g. misprediction or memory nuke).",
1323+ "EventCode": "0xC4",
1324+ "UMask": "0x10",
1325+ "BriefDescription": "Not taken branch instructions retired.",
13311326 "Counter": "0,1,2,3",
1332− "EventName": "INT_MISC.RECOVERY_CYCLES_ANY",
1333− "AnyThread": "1",
1334− "CounterMask": "1",
1335− "SampleAfterValue": "2000003",
1327+ "EventName": "BR_INST_RETIRED.NOT_TAKEN",
1328+ "PublicDescription": "This event counts not taken branch instructions retired.",
1329+ "SampleAfterValue": "400009",
13361330 "CounterHTOff": "0,1,2,3,4,5,6,7"
13371331 },
13381332 {
1339− "EventCode": "0xb1",
1340− "UMask": "0x2",
1341− "BriefDescription": "Cycles at least 1 micro-op is executed from any thread on physical core.",
1333+ "EventCode": "0xC4",
1334+ "UMask": "0x20",
1335+ "BriefDescription": "Taken branch instructions retired. (Precise Event - PEBS)",
1336+ "PEBS": "1",
13421337 "Counter": "0,1,2,3",
1343− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_1",
1344− "CounterMask": "1",
1345− "SampleAfterValue": "2000003",
1338+ "EventName": "BR_INST_RETIRED.NEAR_TAKEN",
1339+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts taken branch instructions retired.",
1340+ "SampleAfterValue": "400009",
13461341 "CounterHTOff": "0,1,2,3,4,5,6,7"
13471342 },
13481343 {
1349− "EventCode": "0xb1",
1350− "UMask": "0x2",
1351− "BriefDescription": "Cycles at least 2 micro-op is executed from any thread on physical core.",
1344+ "EventCode": "0xC4",
1345+ "UMask": "0x40",
1346+ "BriefDescription": "Far branch instructions retired.",
13521347 "Counter": "0,1,2,3",
1353− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_2",
1354− "CounterMask": "2",
1355− "SampleAfterValue": "2000003",
1348+ "EventName": "BR_INST_RETIRED.FAR_BRANCH",
1349+ "Errata": "BDW98",
1350+ "PublicDescription": "This event counts far branch instructions retired.",
1351+ "SampleAfterValue": "100007",
13561352 "CounterHTOff": "0,1,2,3,4,5,6,7"
13571353 },
13581354 {
1359− "EventCode": "0xb1",
1360− "UMask": "0x2",
1361− "BriefDescription": "Cycles at least 3 micro-op is executed from any thread on physical core.",
1355+ "EventCode": "0xC5",
1356+ "UMask": "0x0",
1357+ "BriefDescription": "All mispredicted macro branch instructions retired.",
13621358 "Counter": "0,1,2,3",
1363− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_3",
1364− "CounterMask": "3",
1365− "SampleAfterValue": "2000003",
1359+ "EventName": "BR_MISP_RETIRED.ALL_BRANCHES",
1360+ "PublicDescription": "This event counts all mispredicted macro branch instructions retired.",
1361+ "SampleAfterValue": "400009",
13661362 "CounterHTOff": "0,1,2,3,4,5,6,7"
13671363 },
13681364 {
1369− "EventCode": "0xb1",
1370− "UMask": "0x2",
1371− "BriefDescription": "Cycles at least 4 micro-op is executed from any thread on physical core.",
1365+ "EventCode": "0xC5",
1366+ "UMask": "0x1",
1367+ "BriefDescription": "Mispredicted conditional branch instructions retired. (Precise Event - PEBS)",
1368+ "PEBS": "1",
13721369 "Counter": "0,1,2,3",
1373− "EventName": "UOPS_EXECUTED.CORE_CYCLES_GE_4",
1374− "CounterMask": "4",
1375− "SampleAfterValue": "2000003",
1370+ "EventName": "BR_MISP_RETIRED.CONDITIONAL",
1371+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted conditional branch instructions retired.",
1372+ "SampleAfterValue": "400009",
13761373 "CounterHTOff": "0,1,2,3,4,5,6,7"
13771374 },
13781375 {
1379− "Invert": "1",
1380− "EventCode": "0xb1",
1381− "UMask": "0x2",
1382− "BriefDescription": "Cycles with no micro-ops executed from any thread on physical core.",
1376+ "EventCode": "0xC5",
1377+ "UMask": "0x4",
1378+ "BriefDescription": "Mispredicted macro branch instructions retired. (Precise Event - PEBS)",
1379+ "PEBS": "2",
13831380 "Counter": "0,1,2,3",
1384− "EventName": "UOPS_EXECUTED.CORE_CYCLES_NONE",
1385− "SampleAfterValue": "2000003",
1381+ "EventName": "BR_MISP_RETIRED.ALL_BRANCHES_PEBS",
1382+ "PublicDescription": "This is a precise version of BR_MISP_RETIRED.ALL_BRANCHES that counts all mispredicted macro branch instructions retired.",
1383+ "SampleAfterValue": "400009",
1384+ "CounterHTOff": "0,1,2,3"
1385+ },
1386+ {
1387+ "EventCode": "0xC5",
1388+ "UMask": "0x8",
1389+ "BriefDescription": "This event counts the number of mispredicted ret instructions retired.(Precise Event)",
1390+ "PEBS": "1",
1391+ "Counter": "0,1,2,3",
1392+ "EventName": "BR_MISP_RETIRED.RET",
1393+ "PublicDescription": "This is a precise version (that is, uses PEBS) of the event that counts mispredicted return instructions retired.",
1394+ "SampleAfterValue": "100007",
13861395 "CounterHTOff": "0,1,2,3,4,5,6,7"
13871396 },
13881397 {
1389− "EventCode": "0x3C",
1390− "UMask": "0x1",
1391− "BriefDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate)",
1398+ "EventCode": "0xC5",
1399+ "UMask": "0x20",
1400+ "BriefDescription": "number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
1401+ "PEBS": "1",
13921402 "Counter": "0,1,2,3",
1393− "EventName": "CPU_CLK_UNHALTED.REF_XCLK",
1394− "PublicDescription": "Reference cycles when the thread is unhalted (counts at 100 MHz rate).",
1395− "SampleAfterValue": "2000003",
1403+ "EventName": "BR_MISP_RETIRED.NEAR_TAKEN",
1404+ "PublicDescription": "Number of near branch instructions retired that were mispredicted and taken. (Precise Event - PEBS).",
1405+ "SampleAfterValue": "400009",
13961406 "CounterHTOff": "0,1,2,3,4,5,6,7"
13971407 },
13981408 {
1399− "EventCode": "0x3C",
1400− "UMask": "0x1",
1401− "BriefDescription": "Reference cycles when the at least one thread on the physical core is unhalted (counts at 100 MHz rate).",
1409+ "EventCode": "0xCC",
1410+ "UMask": "0x20",
1411+ "BriefDescription": "Count cases of saving new LBR",
14021412 "Counter": "0,1,2,3",
1403− "EventName": "CPU_CLK_UNHALTED.REF_XCLK_ANY",
1404− "AnyThread": "1",
1413+ "EventName": "ROB_MISC_EVENTS.LBR_INSERTS",
1414+ "PublicDescription": "This event counts cases of saving new LBR records by hardware. This assumes proper enabling of LBRs and takes into account LBR filtering done by the LBR_SELECT register.",
14051415 "SampleAfterValue": "2000003",
14061416 "CounterHTOff": "0,1,2,3,4,5,6,7"
14071417 },
14081418 {
1409− "EventCode": "0x3C",
1410− "UMask": "0x2",
1411− "BriefDescription": "Count XClk pulses when this thread is unhalted and the other thread is halted.",
1419+ "EventCode": "0xe6",
1420+ "UMask": "0x1f",
1421+ "BriefDescription": "Counts the total number when the front end is resteered, mainly when the BPU cannot provide a correct prediction and this is corrected by other branch handling mechanisms at the front end.",
14121422 "Counter": "0,1,2,3",
1413− "EventName": "CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE",
1414− "SampleAfterValue": "2000003",
1423+ "EventName": "BACLEARS.ANY",
1424+ "SampleAfterValue": "100003",
14151425 "CounterHTOff": "0,1,2,3,4,5,6,7"
14161426 }
14171427 ]
tools/perf/pmu-events/arch/x86/broadwellx/virtual-memory.json+75−75
@@ -43,6 +43,16 @@ · 42 unmodified lines
4343 "SampleAfterValue": "2000003",
4444 "CounterHTOff": "0,1,2,3,4,5,6,7"
4545 },
46+ {
47+ "EventCode": "0x08",
48+ "UMask": "0xe",
49+ "BriefDescription": "Demand load Miss in all translation lookaside buffer (TLB) levels causes a page walk that completes of any page size.",
50+ "Counter": "0,1,2,3",
51+ "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED",
52+ "Errata": "BDM69",
53+ "SampleAfterValue": "100003",
54+ "CounterHTOff": "0,1,2,3,4,5,6,7"
55+ },
4656 {
4757 "EventCode": "0x08",
4858 "UMask": "0x10",
@@ -72,6 +82,15 @@ · 23 unmodified lines
7282 "SampleAfterValue": "2000003",
7383 "CounterHTOff": "0,1,2,3,4,5,6,7"
7484 },
85+ {
86+ "EventCode": "0x08",
87+ "UMask": "0x60",
88+ "BriefDescription": "Load operations that miss the first DTLB level but hit the second and do not cause page walks.",
89+ "Counter": "0,1,2,3",
90+ "EventName": "DTLB_LOAD_MISSES.STLB_HIT",
91+ "SampleAfterValue": "2000003",
92+ "CounterHTOff": "0,1,2,3,4,5,6,7"
93+ },
7594 {
7695 "EventCode": "0x49",
7796 "UMask": "0x1",
@@ -116,6 +135,16 @@ · 38 unmodified lines
116135 "SampleAfterValue": "100003",
117136 "CounterHTOff": "0,1,2,3,4,5,6,7"
118137 },
138+ {
139+ "EventCode": "0x49",
140+ "UMask": "0xe",
141+ "BriefDescription": "Store misses in all DTLB levels that cause completed page walks.",
142+ "Counter": "0,1,2,3",
143+ "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED",
144+ "Errata": "BDM69",
145+ "SampleAfterValue": "100003",
146+ "CounterHTOff": "0,1,2,3,4,5,6,7"
147+ },
119148 {
120149 "EventCode": "0x49",
121150 "UMask": "0x10",
@@ -145,6 +174,15 @@ · 23 unmodified lines
145174 "SampleAfterValue": "100003",
146175 "CounterHTOff": "0,1,2,3,4,5,6,7"
147176 },
177+ {
178+ "EventCode": "0x49",
179+ "UMask": "0x60",
180+ "BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
181+ "Counter": "0,1,2,3",
182+ "EventName": "DTLB_STORE_MISSES.STLB_HIT",
183+ "SampleAfterValue": "100003",
184+ "CounterHTOff": "0,1,2,3,4,5,6,7"
185+ },
148186 {
149187 "EventCode": "0x4F",
150188 "UMask": "0x10",
@@ -199,6 +237,16 @@ · 48 unmodified lines
199237 "SampleAfterValue": "100003",
200238 "CounterHTOff": "0,1,2,3,4,5,6,7"
201239 },
240+ {
241+ "EventCode": "0x85",
242+ "UMask": "0xe",
243+ "BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
244+ "Counter": "0,1,2,3",
245+ "EventName": "ITLB_MISSES.WALK_COMPLETED",
246+ "Errata": "BDM69",
247+ "SampleAfterValue": "100003",
248+ "CounterHTOff": "0,1,2,3,4,5,6,7"
249+ },
202250 {
203251 "EventCode": "0x85",
204252 "UMask": "0x10",
@@ -228,6 +276,15 @@ · 23 unmodified lines
228276 "SampleAfterValue": "100003",
229277 "CounterHTOff": "0,1,2,3,4,5,6,7"
230278 },
279+ {
280+ "EventCode": "0x85",
281+ "UMask": "0x60",
282+ "BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
283+ "Counter": "0,1,2,3",
284+ "EventName": "ITLB_MISSES.STLB_HIT",
285+ "SampleAfterValue": "100003",
286+ "CounterHTOff": "0,1,2,3,4,5,6,7"
287+ },
231288 {
232289 "EventCode": "0xAE",
233290 "UMask": "0x1",
@@ -250,60 +307,60 @@ · 16 unmodified lines
250307 },
251308 {
252309 "EventCode": "0xBC",
253− "UMask": "0x21",
254− "BriefDescription": "Number of ITLB page walker hits in the L1+FB.",
310+ "UMask": "0x12",
311+ "BriefDescription": "Number of DTLB page walker hits in the L2.",
255312 "Counter": "0,1,2,3",
256− "EventName": "PAGE_WALKER_LOADS.ITLB_L1",
313+ "EventName": "PAGE_WALKER_LOADS.DTLB_L2",
257314 "Errata": "BDM69, BDM98",
258315 "SampleAfterValue": "2000003",
259316 "CounterHTOff": "0,1,2,3"
260317 },
261318 {
262319 "EventCode": "0xBC",
263− "UMask": "0x12",
264− "BriefDescription": "Number of DTLB page walker hits in the L2.",
320+ "UMask": "0x14",
321+ "BriefDescription": "Number of DTLB page walker hits in the L3 + XSNP.",
265322 "Counter": "0,1,2,3",
266− "EventName": "PAGE_WALKER_LOADS.DTLB_L2",
323+ "EventName": "PAGE_WALKER_LOADS.DTLB_L3",
267324 "Errata": "BDM69, BDM98",
268325 "SampleAfterValue": "2000003",
269326 "CounterHTOff": "0,1,2,3"
270327 },
271328 {
272329 "EventCode": "0xBC",
273− "UMask": "0x22",
274− "BriefDescription": "Number of ITLB page walker hits in the L2.",
330+ "UMask": "0x18",
331+ "BriefDescription": "Number of DTLB page walker hits in Memory.",
275332 "Counter": "0,1,2,3",
276− "EventName": "PAGE_WALKER_LOADS.ITLB_L2",
333+ "EventName": "PAGE_WALKER_LOADS.DTLB_MEMORY",
277334 "Errata": "BDM69, BDM98",
278335 "SampleAfterValue": "2000003",
279336 "CounterHTOff": "0,1,2,3"
280337 },
281338 {
282339 "EventCode": "0xBC",
283− "UMask": "0x14",
284− "BriefDescription": "Number of DTLB page walker hits in the L3 + XSNP.",
340+ "UMask": "0x21",
341+ "BriefDescription": "Number of ITLB page walker hits in the L1+FB.",
285342 "Counter": "0,1,2,3",
286− "EventName": "PAGE_WALKER_LOADS.DTLB_L3",
343+ "EventName": "PAGE_WALKER_LOADS.ITLB_L1",
287344 "Errata": "BDM69, BDM98",
288345 "SampleAfterValue": "2000003",
289346 "CounterHTOff": "0,1,2,3"
290347 },
291348 {
292349 "EventCode": "0xBC",
293− "UMask": "0x24",
294− "BriefDescription": "Number of ITLB page walker hits in the L3 + XSNP.",
350+ "UMask": "0x22",
351+ "BriefDescription": "Number of ITLB page walker hits in the L2.",
295352 "Counter": "0,1,2,3",
296− "EventName": "PAGE_WALKER_LOADS.ITLB_L3",
353+ "EventName": "PAGE_WALKER_LOADS.ITLB_L2",
297354 "Errata": "BDM69, BDM98",
298355 "SampleAfterValue": "2000003",
299356 "CounterHTOff": "0,1,2,3"
300357 },
301358 {
302359 "EventCode": "0xBC",
303− "UMask": "0x18",
304− "BriefDescription": "Number of DTLB page walker hits in Memory.",
360+ "UMask": "0x24",
361+ "BriefDescription": "Number of ITLB page walker hits in the L3 + XSNP.",
305362 "Counter": "0,1,2,3",
306− "EventName": "PAGE_WALKER_LOADS.DTLB_MEMORY",
363+ "EventName": "PAGE_WALKER_LOADS.ITLB_L3",
307364 "Errata": "BDM69, BDM98",
308365 "SampleAfterValue": "2000003",
309366 "CounterHTOff": "0,1,2,3"
@@ -327,62 +384,5 @@ · 17 unmodified lines
327384 "PublicDescription": "This event counts the number of any STLB flush attempts (such as entire, VPID, PCID, InvPage, CR3 write, and so on).",
328385 "SampleAfterValue": "100007",
329386 "CounterHTOff": "0,1,2,3,4,5,6,7"
330− },
331− {
332− "EventCode": "0x08",
333− "UMask": "0xe",
334− "BriefDescription": "Demand load Miss in all translation lookaside buffer (TLB) levels causes a page walk that completes of any page size.",
335− "Counter": "0,1,2,3",
336− "EventName": "DTLB_LOAD_MISSES.WALK_COMPLETED",
337− "Errata": "BDM69",
338− "SampleAfterValue": "100003",
339− "CounterHTOff": "0,1,2,3,4,5,6,7"
340− },
341− {
342− "EventCode": "0x08",
343− "UMask": "0x60",
344− "BriefDescription": "Load operations that miss the first DTLB level but hit the second and do not cause page walks.",
345− "Counter": "0,1,2,3",
346− "EventName": "DTLB_LOAD_MISSES.STLB_HIT",
347− "SampleAfterValue": "2000003",
348− "CounterHTOff": "0,1,2,3,4,5,6,7"
349− },
350− {
351− "EventCode": "0x49",
352− "UMask": "0xe",
353− "BriefDescription": "Store misses in all DTLB levels that cause completed page walks.",
354− "Counter": "0,1,2,3",
355− "EventName": "DTLB_STORE_MISSES.WALK_COMPLETED",
356− "Errata": "BDM69",
357− "SampleAfterValue": "100003",
358− "CounterHTOff": "0,1,2,3,4,5,6,7"
359− },
360− {
361− "EventCode": "0x49",
362− "UMask": "0x60",
363− "BriefDescription": "Store operations that miss the first TLB level but hit the second and do not cause page walks.",
364− "Counter": "0,1,2,3",
365− "EventName": "DTLB_STORE_MISSES.STLB_HIT",
366− "SampleAfterValue": "100003",
367− "CounterHTOff": "0,1,2,3,4,5,6,7"
368− },
369− {
370− "EventCode": "0x85",
371− "UMask": "0xe",
372− "BriefDescription": "Misses in all ITLB levels that cause completed page walks.",
373− "Counter": "0,1,2,3",
374− "EventName": "ITLB_MISSES.WALK_COMPLETED",
375− "Errata": "BDM69",
376− "SampleAfterValue": "100003",
377− "CounterHTOff": "0,1,2,3,4,5,6,7"
378− },
379− {
380− "EventCode": "0x85",
381− "UMask": "0x60",
382− "BriefDescription": "Operations that miss the first ITLB level but hit the second and do not cause any page walks.",
383− "Counter": "0,1,2,3",
384− "EventName": "ITLB_MISSES.STLB_HIT",
385− "SampleAfterValue": "100003",
386− "CounterHTOff": "0,1,2,3,4,5,6,7"
387387 }
388388 ]
tools/perf/pmu-events/arch/x86/goldmont/cache.json+1081−163
@@ -1,4 +1,24 @@
11 [
2+ {
3+ "CollectPEBSRecord": "1",
4+ "PublicDescription": "Counts memory requests originating from the core that miss in the L2 cache.",
5+ "EventCode": "0x2E",
6+ "Counter": "0,1,2,3",
7+ "UMask": "0x41",
8+ "EventName": "LONGEST_LAT_CACHE.MISS",
9+ "SampleAfterValue": "200003",
10+ "BriefDescription": "L2 cache request misses"
11+ },
12+ {
13+ "CollectPEBSRecord": "1",
14+ "PublicDescription": "Counts memory requests originating from the core that reference a cache line in the L2 cache.",
15+ "EventCode": "0x2E",
16+ "Counter": "0,1,2,3",
17+ "UMask": "0x4f",
18+ "EventName": "LONGEST_LAT_CACHE.REFERENCE",
19+ "SampleAfterValue": "200003",
20+ "BriefDescription": "L2 cache requests"
21+ },
222 {
323 "CollectPEBSRecord": "1",
424 "PublicDescription": "Counts the number of demand and prefetch transactions that the L2 XQ rejects due to a full or near full condition which likely indicates back pressure from the intra-die interconnect (IDI) fabric. The XQ may reject transactions from the L2Q (non-cacheable requests), L2 misses and L2 write-back victims.",
@@ -11,120 +31,119 @@ · 6 unmodified lines
1131 },
1232 {
1333 "CollectPEBSRecord": "1",
14− "PublicDescription": "Counts the number of demand and L1 prefetcher requests rejected by the L2Q due to a full or nearly full condition which likely indicates back pressure from L2Q. It also counts requests that would have gone directly to the XQ, but are rejected due to a full or nearly full condition, indicating back pressure from the IDI link. The L2Q may also reject transactions from a core to insure fairness between cores, or to delay a core's dirty eviction when the address conflicts with incoming external snoops.",
34+ "PublicDescription": "Counts the number of demand and L1 prefetcher requests rejected by the L2Q due to a full or nearly full condition which likely indicates back pressure from L2Q. It also counts requests that would have gone directly to the XQ, but are rejected due to a full or nearly full condition, indicating back pressure from the IDI link. The L2Q may also reject transactions from a core to ensure fairness between cores, or to delay a core's dirty eviction when the address conflicts with incoming external snoops.",
1535 "EventCode": "0x31",
1636 "Counter": "0,1,2,3",
1737 "UMask": "0x0",
1838 "EventName": "CORE_REJECT_L2Q.ALL",
1939 "SampleAfterValue": "200003",
20− "BriefDescription": "Requests rejected by the L2Q "
40+ "BriefDescription": "Requests rejected by the L2Q"
2141 },
2242 {
2343 "CollectPEBSRecord": "1",
24− "PublicDescription": "Counts memory requests originating from the core that reference a cache line in the L2 cache.",
25− "EventCode": "0x2E",
44+ "PublicDescription": "Counts when a modified (dirty) cache line is evicted from the data L1 cache and needs to be written back to memory. No count will occur if the evicted line is clean, and hence does not require a writeback.",
45+ "EventCode": "0x51",
2646 "Counter": "0,1,2,3",
27− "UMask": "0x4f",
28− "EventName": "LONGEST_LAT_CACHE.REFERENCE",
47+ "UMask": "0x1",
48+ "EventName": "DL1.DIRTY_EVICTION",
2949 "SampleAfterValue": "200003",
30− "BriefDescription": "L2 cache requests"
50+ "BriefDescription": "L1 Cache evictions for dirty data"
3151 },
3252 {
3353 "CollectPEBSRecord": "1",
34− "PublicDescription": "Counts memory requests originating from the core that miss in the L2 cache.",
35− "EventCode": "0x2E",
54+ "PublicDescription": "Counts cycles that fetch is stalled due to an outstanding ICache miss. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes due to an ICache miss. Note: this event is not the same as the total number of cycles spent retrieving instruction cache lines from the memory hierarchy.",
55+ "EventCode": "0x86",
3656 "Counter": "0,1,2,3",
37− "UMask": "0x41",
38− "EventName": "LONGEST_LAT_CACHE.MISS",
57+ "UMask": "0x2",
58+ "EventName": "FETCH_STALL.ICACHE_FILL_PENDING_CYCLES",
3959 "SampleAfterValue": "200003",
40− "BriefDescription": "L2 cache request misses"
60+ "BriefDescription": "Cycles code-fetch stalled due to an outstanding ICache miss."
4161 },
4262 {
4363 "CollectPEBSRecord": "1",
44− "PublicDescription": "Counts cycles that an ICache miss is outstanding, and instruction fetch is stalled. That is, the decoder queue is able to accept bytes, but the fetch unit is unable to provide bytes, while an Icache miss outstanding. Note this event is not the same as cycles to retrieve an instruction due to an Icache miss. Rather, it is the part of the Instruction Cache (ICache) miss time where no bytes are available for the decoder.",
45− "EventCode": "0x86",
64+ "EventCode": "0xB7",
4665 "Counter": "0,1,2,3",
47− "UMask": "0x2",
48− "EventName": "FETCH_STALL.ICACHE_FILL_PENDING_CYCLES",
49− "SampleAfterValue": "200003",
50− "BriefDescription": "Cycles where code-fetch is stalled and an ICache miss is outstanding. This is not the same as an ICache Miss."
66+ "UMask": "0x1",
67+ "EventName": "OFFCORE_RESPONSE",
68+ "SampleAfterValue": "100007",
69+ "BriefDescription": "Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)"
5170 },
5271 {
5372 "PEBS": "2",
5473 "CollectPEBSRecord": "2",
55− "PublicDescription": "Counts the number of load uops retired.",
74+ "PublicDescription": "Counts locked memory uops retired. This includes regular locks and bus locks. (To specifically count bus locks only, see the Offcore response event.) A locked access is one with a lock prefix, or an exchange to memory. See the SDM for a complete description of which memory load accesses are locks.",
5675 "EventCode": "0xD0",
5776 "Counter": "0,1,2,3",
58− "UMask": "0x81",
59− "EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
77+ "UMask": "0x21",
78+ "EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
6079 "SampleAfterValue": "200003",
61− "BriefDescription": "Load uops retired (Precise event capable)"
80+ "BriefDescription": "Locked load uops retired (Precise event capable)"
6281 },
6382 {
6483 "PEBS": "2",
6584 "CollectPEBSRecord": "2",
66− "PublicDescription": "Counts the number of store uops retired.",
85+ "PublicDescription": "Counts load uops retired where the data requested spans a 64 byte cache line boundary.",
6786 "EventCode": "0xD0",
6887 "Counter": "0,1,2,3",
69− "UMask": "0x82",
70− "EventName": "MEM_UOPS_RETIRED.ALL_STORES",
88+ "UMask": "0x41",
89+ "EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
7190 "SampleAfterValue": "200003",
72− "BriefDescription": "Store uops retired (Precise event capable)"
91+ "BriefDescription": "Load uops retired that split a cache-line (Precise event capable)"
7392 },
7493 {
7594 "PEBS": "2",
7695 "CollectPEBSRecord": "2",
77− "PublicDescription": "Counts the number of memory uops retired that is either a loads or a store or both.",
96+ "PublicDescription": "Counts store uops retired where the data requested spans a 64 byte cache line boundary.",
7897 "EventCode": "0xD0",
7998 "Counter": "0,1,2,3",
80− "UMask": "0x83",
81− "EventName": "MEM_UOPS_RETIRED.ALL",
99+ "UMask": "0x42",
100+ "EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
82101 "SampleAfterValue": "200003",
83− "BriefDescription": "Memory uops retired (Precise event capable)"
102+ "BriefDescription": "Stores uops retired that split a cache-line (Precise event capable)"
84103 },
85104 {
86105 "PEBS": "2",
87106 "CollectPEBSRecord": "2",
88− "PublicDescription": "Counts locked memory uops retired. This includes \"regular\" locks and bus locks. (To specifically count bus locks only, see the Offcore response event.) A locked access is one with a lock prefix, or an exchange to memory. See the SDM for a complete description of which memory load accesses are locks.",
107+ "PublicDescription": "Counts memory uops retired where the data requested spans a 64 byte cache line boundary.",
89108 "EventCode": "0xD0",
90109 "Counter": "0,1,2,3",
91− "UMask": "0x21",
92− "EventName": "MEM_UOPS_RETIRED.LOCK_LOADS",
110+ "UMask": "0x43",
111+ "EventName": "MEM_UOPS_RETIRED.SPLIT",
93112 "SampleAfterValue": "200003",
94− "BriefDescription": "Locked load uops retired (Precise event capable)"
113+ "BriefDescription": "Memory uops retired that split a cache-line (Precise event capable)"
95114 },
96115 {
97116 "PEBS": "2",
98117 "CollectPEBSRecord": "2",
99− "PublicDescription": "Counts load uops retired where the data requested spans a 64 byte cache line boundary.",
118+ "PublicDescription": "Counts the number of load uops retired.",
100119 "EventCode": "0xD0",
101120 "Counter": "0,1,2,3",
102− "UMask": "0x41",
103− "EventName": "MEM_UOPS_RETIRED.SPLIT_LOADS",
121+ "UMask": "0x81",
122+ "EventName": "MEM_UOPS_RETIRED.ALL_LOADS",
104123 "SampleAfterValue": "200003",
105− "BriefDescription": "Load uops retired that split a cache-line (Precise event capable)"
124+ "BriefDescription": "Load uops retired (Precise event capable)"
106125 },
107126 {
108127 "PEBS": "2",
109128 "CollectPEBSRecord": "2",
110− "PublicDescription": "Counts store uops retired where the data requested spans a 64 byte cache line boundary.",
129+ "PublicDescription": "Counts the number of store uops retired.",
111130 "EventCode": "0xD0",
112131 "Counter": "0,1,2,3",
113− "UMask": "0x42",
114− "EventName": "MEM_UOPS_RETIRED.SPLIT_STORES",
132+ "UMask": "0x82",
133+ "EventName": "MEM_UOPS_RETIRED.ALL_STORES",
115134 "SampleAfterValue": "200003",
116− "BriefDescription": "Stores uops retired that split a cache-line (Precise event capable)"
135+ "BriefDescription": "Store uops retired (Precise event capable)"
117136 },
118137 {
119138 "PEBS": "2",
120139 "CollectPEBSRecord": "2",
121− "PublicDescription": "Counts memory uops retired where the data requested spans a 64 byte cache line boundary.",
140+ "PublicDescription": "Counts the number of memory uops retired that is either a loads or a store or both.",
122141 "EventCode": "0xD0",
123142 "Counter": "0,1,2,3",
124− "UMask": "0x43",
125− "EventName": "MEM_UOPS_RETIRED.SPLIT",
143+ "UMask": "0x83",
144+ "EventName": "MEM_UOPS_RETIRED.ALL",
126145 "SampleAfterValue": "200003",
127− "BriefDescription": "Memory uops retired that split a cache-line (Precise event capable)"
146+ "BriefDescription": "Memory uops retired (Precise event capable)"
128147 },
129148 {
130149 "PEBS": "2",
@@ -140,24 +159,24 @@ · 9 unmodified lines
140159 {
141160 "PEBS": "2",
142161 "CollectPEBSRecord": "2",
143− "PublicDescription": "Counts load uops retired that miss the L1 data cache.",
162+ "PublicDescription": "Counts load uops retired that hit in the L2 cache.",
144163 "EventCode": "0xD1",
145164 "Counter": "0,1,2,3",
146− "UMask": "0x8",
147− "EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
165+ "UMask": "0x2",
166+ "EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
148167 "SampleAfterValue": "200003",
149− "BriefDescription": "Load uops retired that missed L1 data cache (Precise event capable)"
168+ "BriefDescription": "Load uops retired that hit L2 (Precise event capable)"
150169 },
151170 {
152171 "PEBS": "2",
153172 "CollectPEBSRecord": "2",
154− "PublicDescription": "Counts load uops retired that hit in the L2 cache.",
173+ "PublicDescription": "Counts load uops retired that miss the L1 data cache.",
155174 "EventCode": "0xD1",
156175 "Counter": "0,1,2,3",
157− "UMask": "0x2",
158− "EventName": "MEM_LOAD_UOPS_RETIRED.L2_HIT",
176+ "UMask": "0x8",
177+ "EventName": "MEM_LOAD_UOPS_RETIRED.L1_MISS",
159178 "SampleAfterValue": "200003",
160− "BriefDescription": "Load uops retired that hit L2 (Precise event capable)"
179+ "BriefDescription": "Load uops retired that missed L1 data cache (Precise event capable)"
161180 },
162181 {
163182 "PEBS": "2",
@@ -205,24 +224,20 @@ · 41 unmodified lines
205224 },
206225 {
207226 "CollectPEBSRecord": "1",
208− "PublicDescription": "Counts when a modified (dirty) cache line is evicted from the data L1 cache and needs to be written back to memory. No count will occur if the evicted line is clean, and hence does not require a writeback.",
209− "EventCode": "0x51",
210− "Counter": "0,1,2,3",
211− "UMask": "0x1",
212− "EventName": "DL1.DIRTY_EVICTION",
213− "SampleAfterValue": "200003",
214− "BriefDescription": "L1 Cache evictions for dirty data"
215− },
216− {
217− "CollectPEBSRecord": "1",
227+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that are outstanding, per cycle, from the time of the L2 miss to when any response is received. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
218228 "EventCode": "0xB7",
229+ "MSRValue": "0x40000032b7 ",
219230 "Counter": "0,1,2,3",
220231 "UMask": "0x1",
221− "EventName": "OFFCORE_RESPONSE",
232+ "EventName": "OFFCORE_RESPONSE.ANY_READ.OUTSTANDING",
233+ "MSRIndex": "0x1a6",
222234 "SampleAfterValue": "100007",
223− "BriefDescription": "Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)"
235+ "BriefDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that are outstanding, per cycle, from the time of the L2 miss to when any response is received.",
236+ "Offcore": "1"
224237 },
225238 {
239+ "CollectPEBSRecord": "1",
240+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that miss the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
226241 "EventCode": "0xB7",
227242 "MSRValue": "0x36000032b7 ",
228243 "Counter": "0,1,2,3",
@@ -234,6 +249,8 @@ · 5 unmodified lines
234249 "Offcore": "1"
235250 },
236251 {
252+ "CollectPEBSRecord": "1",
253+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that miss the L2 cache with a snoop hit in the other processor module, data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
237254 "EventCode": "0xB7",
238255 "MSRValue": "0x10000032b7 ",
239256 "Counter": "0,1,2,3",
@@ -245,6 +262,8 @@ · 5 unmodified lines
245262 "Offcore": "1"
246263 },
247264 {
265+ "CollectPEBSRecord": "1",
266+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that miss the L2 cache with a snoop hit in the other processor module, no data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
248267 "EventCode": "0xB7",
249268 "MSRValue": "0x04000032b7 ",
250269 "Counter": "0,1,2,3",
@@ -256,6 +275,8 @@ · 5 unmodified lines
256275 "Offcore": "1"
257276 },
258277 {
278+ "CollectPEBSRecord": "1",
279+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that true miss for the L2 cache with a snoop miss in the other processor module. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
259280 "EventCode": "0xB7",
260281 "MSRValue": "0x02000032b7 ",
261282 "Counter": "0,1,2,3",
@@ -267,6 +288,8 @@ · 5 unmodified lines
267288 "Offcore": "1"
268289 },
269290 {
291+ "CollectPEBSRecord": "1",
292+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that hit the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
270293 "EventCode": "0xB7",
271294 "MSRValue": "0x00000432b7 ",
272295 "Counter": "0,1,2,3",
@@ -278,6 +301,34 @@ · 5 unmodified lines
278301 "Offcore": "1"
279302 },
280303 {
304+ "CollectPEBSRecord": "1",
305+ "PublicDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that have any transaction responses from the uncore subsystem. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
306+ "EventCode": "0xB7",
307+ "MSRValue": "0x00000132b7 ",
308+ "Counter": "0,1,2,3",
309+ "UMask": "0x1",
310+ "EventName": "OFFCORE_RESPONSE.ANY_READ.ANY_RESPONSE",
311+ "MSRIndex": "0x1a6,0x1a7",
312+ "SampleAfterValue": "100007",
313+ "BriefDescription": "Counts data read, code read, and read for ownership (RFO) requests (demand & prefetch) that have any transaction responses from the uncore subsystem.",
314+ "Offcore": "1"
315+ },
316+ {
317+ "CollectPEBSRecord": "1",
318+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that are outstanding, per cycle, from the time of the L2 miss to when any response is received. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
319+ "EventCode": "0xB7",
320+ "MSRValue": "0x4000000022 ",
321+ "Counter": "0,1,2,3",
322+ "UMask": "0x1",
323+ "EventName": "OFFCORE_RESPONSE.ANY_RFO.OUTSTANDING",
324+ "MSRIndex": "0x1a6",
325+ "SampleAfterValue": "100007",
326+ "BriefDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that are outstanding, per cycle, from the time of the L2 miss to when any response is received.",
327+ "Offcore": "1"
328+ },
329+ {
330+ "CollectPEBSRecord": "1",
331+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that miss the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
281332 "EventCode": "0xB7",
282333 "MSRValue": "0x3600000022 ",
283334 "Counter": "0,1,2,3",
@@ -289,6 +340,8 @@ · 5 unmodified lines
289340 "Offcore": "1"
290341 },
291342 {
343+ "CollectPEBSRecord": "1",
344+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that miss the L2 cache with a snoop hit in the other processor module, data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
292345 "EventCode": "0xB7",
293346 "MSRValue": "0x1000000022 ",
294347 "Counter": "0,1,2,3",
@@ -300,6 +353,8 @@ · 5 unmodified lines
300353 "Offcore": "1"
301354 },
302355 {
356+ "CollectPEBSRecord": "1",
357+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that miss the L2 cache with a snoop hit in the other processor module, no data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
303358 "EventCode": "0xB7",
304359 "MSRValue": "0x0400000022 ",
305360 "Counter": "0,1,2,3",
@@ -311,6 +366,8 @@ · 5 unmodified lines
311366 "Offcore": "1"
312367 },
313368 {
369+ "CollectPEBSRecord": "1",
370+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that true miss for the L2 cache with a snoop miss in the other processor module. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
314371 "EventCode": "0xB7",
315372 "MSRValue": "0x0200000022 ",
316373 "Counter": "0,1,2,3",
@@ -322,6 +379,8 @@ · 5 unmodified lines
322379 "Offcore": "1"
323380 },
324381 {
382+ "CollectPEBSRecord": "1",
383+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that hit the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
325384 "EventCode": "0xB7",
326385 "MSRValue": "0x0000040022 ",
327386 "Counter": "0,1,2,3",
@@ -333,6 +392,34 @@ · 5 unmodified lines
333392 "Offcore": "1"
334393 },
335394 {
395+ "CollectPEBSRecord": "1",
396+ "PublicDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that have any transaction responses from the uncore subsystem. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
397+ "EventCode": "0xB7",
398+ "MSRValue": "0x0000010022 ",
399+ "Counter": "0,1,2,3",
400+ "UMask": "0x1",
401+ "EventName": "OFFCORE_RESPONSE.ANY_RFO.ANY_RESPONSE",
402+ "MSRIndex": "0x1a6,0x1a7",
403+ "SampleAfterValue": "100007",
404+ "BriefDescription": "Counts reads for ownership (RFO) requests (demand & prefetch) that have any transaction responses from the uncore subsystem.",
405+ "Offcore": "1"
406+ },
407+ {
408+ "CollectPEBSRecord": "1",
409+ "PublicDescription": "Counts data reads (demand & prefetch) that are outstanding, per cycle, from the time of the L2 miss to when any response is received. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
410+ "EventCode": "0xB7",
411+ "MSRValue": "0x4000003091",
412+ "Counter": "0,1,2,3",
413+ "UMask": "0x1",
414+ "EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.OUTSTANDING",
415+ "MSRIndex": "0x1a6",
416+ "SampleAfterValue": "100007",
417+ "BriefDescription": "Counts data reads (demand & prefetch) that are outstanding, per cycle, from the time of the L2 miss to when any response is received.",
418+ "Offcore": "1"
419+ },
420+ {
421+ "CollectPEBSRecord": "1",
422+ "PublicDescription": "Counts data reads (demand & prefetch) that miss the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
336423 "EventCode": "0xB7",
337424 "MSRValue": "0x3600003091",
338425 "Counter": "0,1,2,3",
@@ -344,6 +431,8 @@ · 5 unmodified lines
344431 "Offcore": "1"
345432 },
346433 {
434+ "CollectPEBSRecord": "1",
435+ "PublicDescription": "Counts data reads (demand & prefetch) that miss the L2 cache with a snoop hit in the other processor module, data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
347436 "EventCode": "0xB7",
348437 "MSRValue": "0x1000003091",
349438 "Counter": "0,1,2,3",
@@ -355,6 +444,8 @@ · 5 unmodified lines
355444 "Offcore": "1"
356445 },
357446 {
447+ "CollectPEBSRecord": "1",
448+ "PublicDescription": "Counts data reads (demand & prefetch) that miss the L2 cache with a snoop hit in the other processor module, no data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
358449 "EventCode": "0xB7",
359450 "MSRValue": "0x0400003091",
360451 "Counter": "0,1,2,3",
@@ -366,6 +457,8 @@ · 5 unmodified lines
366457 "Offcore": "1"
367458 },
368459 {
460+ "CollectPEBSRecord": "1",
461+ "PublicDescription": "Counts data reads (demand & prefetch) that true miss for the L2 cache with a snoop miss in the other processor module. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
369462 "EventCode": "0xB7",
370463 "MSRValue": "0x0200003091",
371464 "Counter": "0,1,2,3",
@@ -377,6 +470,8 @@ · 5 unmodified lines
377470 "Offcore": "1"
378471 },
379472 {
473+ "CollectPEBSRecord": "1",
474+ "PublicDescription": "Counts data reads (demand & prefetch) that hit the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
380475 "EventCode": "0xB7",
381476 "MSRValue": "0x0000043091",
382477 "Counter": "0,1,2,3",
@@ -388,6 +483,34 @@ · 5 unmodified lines
388483 "Offcore": "1"
389484 },
390485 {
486+ "CollectPEBSRecord": "1",
487+ "PublicDescription": "Counts data reads (demand & prefetch) that have any transaction responses from the uncore subsystem. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
488+ "EventCode": "0xB7",
489+ "MSRValue": "0x0000013091",
490+ "Counter": "0,1,2,3",
491+ "UMask": "0x1",
492+ "EventName": "OFFCORE_RESPONSE.ANY_DATA_RD.ANY_RESPONSE",
493+ "MSRIndex": "0x1a6,0x1a7",
494+ "SampleAfterValue": "100007",
495+ "BriefDescription": "Counts data reads (demand & prefetch) that have any transaction responses from the uncore subsystem.",
496+ "Offcore": "1"
497+ },
498+ {
499+ "CollectPEBSRecord": "1",
500+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that are outstanding, per cycle, from the time of the L2 miss to when any response is received. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
501+ "EventCode": "0xB7",
502+ "MSRValue": "0x4000003010 ",
503+ "Counter": "0,1,2,3",
504+ "UMask": "0x1",
505+ "EventName": "OFFCORE_RESPONSE.ANY_PF_DATA_RD.OUTSTANDING",
506+ "MSRIndex": "0x1a6",
507+ "SampleAfterValue": "100007",
508+ "BriefDescription": "Counts data reads generated by L1 or L2 prefetchers that are outstanding, per cycle, from the time of the L2 miss to when any response is received.",
509+ "Offcore": "1"
510+ },
511+ {
512+ "CollectPEBSRecord": "1",
513+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that miss the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
391514 "EventCode": "0xB7",
392515 "MSRValue": "0x3600003010 ",
393516 "Counter": "0,1,2,3",
@@ -399,6 +522,8 @@ · 5 unmodified lines
399522 "Offcore": "1"
400523 },
401524 {
525+ "CollectPEBSRecord": "1",
526+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that miss the L2 cache with a snoop hit in the other processor module, data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
402527 "EventCode": "0xB7",
403528 "MSRValue": "0x1000003010 ",
404529 "Counter": "0,1,2,3",
@@ -410,6 +535,8 @@ · 5 unmodified lines
410535 "Offcore": "1"
411536 },
412537 {
538+ "CollectPEBSRecord": "1",
539+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that miss the L2 cache with a snoop hit in the other processor module, no data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
413540 "EventCode": "0xB7",
414541 "MSRValue": "0x0400003010 ",
415542 "Counter": "0,1,2,3",
@@ -421,6 +548,8 @@ · 5 unmodified lines
421548 "Offcore": "1"
422549 },
423550 {
551+ "CollectPEBSRecord": "1",
552+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that true miss for the L2 cache with a snoop miss in the other processor module. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
424553 "EventCode": "0xB7",
425554 "MSRValue": "0x0200003010 ",
426555 "Counter": "0,1,2,3",
@@ -432,347 +561,957 @@ · 5 unmodified lines
432561 "Offcore": "1"
433562 },
434563 {
564+ "CollectPEBSRecord": "1",
565+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that hit the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
435566 "EventCode": "0xB7",
436567 "MSRValue": "0x0000043010 ",
437568 "Counter": "0,1,2,3",
438569 "UMask": "0x1",
439− "EventName": "OFFCORE_RESPONSE.ANY_PF_DATA_RD.L2_HIT",
570+ "EventName": "OFFCORE_RESPONSE.ANY_PF_DATA_RD.L2_HIT",
571+ "MSRIndex": "0x1a6,0x1a7",
572+ "SampleAfterValue": "100007",
573+ "BriefDescription": "Counts data reads generated by L1 or L2 prefetchers that hit the L2 cache.",
574+ "Offcore": "1"
575+ },
576+ {
577+ "CollectPEBSRecord": "1",
578+ "PublicDescription": "Counts data reads generated by L1 or L2 prefetchers that have any transaction responses from the uncore subsystem. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
579+ "EventCode": "0xB7",
580+ "MSRValue": "0x0000013010 ",
581+ "Counter": "0,1,2,3",
582+ "UMask": "0x1",
583+ "EventName": "OFFCORE_RESPONSE.ANY_PF_DATA_RD.ANY_RESPONSE",
584+ "MSRIndex": "0x1a6,0x1a7",
585+ "SampleAfterValue": "100007",
586+ "BriefDescription": "Counts data reads generated by L1 or L2 prefetchers that have any transaction responses from the uncore subsystem.",
587+ "Offcore": "1"
588+ },
589+ {
590+ "CollectPEBSRecord": "1",
591+ "PublicDescription": "Counts requests to the uncore subsystem that are outstanding, per cycle, from the time of the L2 miss to when any response is received. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
592+ "EventCode": "0xB7",
593+ "MSRValue": "0x4000008000 ",
594+ "Counter": "0,1,2,3",
595+ "UMask": "0x1",
596+ "EventName": "OFFCORE_RESPONSE.ANY_REQUEST.OUTSTANDING",
597+ "MSRIndex": "0x1a6",
598+ "SampleAfterValue": "100007",
599+ "BriefDescription": "Counts requests to the uncore subsystem that are outstanding, per cycle, from the time of the L2 miss to when any response is received.",
600+ "Offcore": "1"
601+ },
602+ {
603+ "CollectPEBSRecord": "1",
604+ "PublicDescription": "Counts requests to the uncore subsystem that miss the L2 cache. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
605+ "EventCode": "0xB7",
606+ "MSRValue": "0x3600008000 ",
607+ "Counter": "0,1,2,3",
608+ "UMask": "0x1",
609+ "EventName": "OFFCORE_RESPONSE.ANY_REQUEST.L2_MISS.ANY",
610+ "MSRIndex": "0x1a6,0x1a7",
611+ "SampleAfterValue": "100007",
612+ "BriefDescription": "Counts requests to the uncore subsystem that miss the L2 cache.",
613+ "Offcore": "1"
614+ },
615+ {
616+ "CollectPEBSRecord": "1",
617+ "PublicDescription": "Counts requests to the uncore subsystem that miss the L2 cache with a snoop hit in the other processor module, data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
618+ "EventCode": "0xB7",
619+ "MSRValue": "0x1000008000 ",
620+ "Counter": "0,1,2,3",
621+ "UMask": "0x1",
622+ "EventName": "OFFCORE_RESPONSE.ANY_REQUEST.L2_MISS.HITM_OTHER_CORE",
623+ "MSRIndex": "0x1a6,0x1a7",
624+ "SampleAfterValue": "100007",
625+ "BriefDescription": "Counts requests to the uncore subsystem that miss the L2 cache with a snoop hit in the other processor module, data forwarding is required.",
626+ "Offcore": "1"
627+ },
628+ {
629+ "CollectPEBSRecord": "1",
630+ "PublicDescription": "Counts requests to the uncore subsystem that miss the L2 cache with a snoop hit in the other processor module, no data forwarding is required. Requires MSR_OFFCORE_RESP[0,1] to specify request type and response. (duplicated for both MSRs)",
631+ "EventCode": "0xB7",
632+ "MSRValue": "0x0400008000 ",

1383 more lines aren’t shown here.

tools/perf/pmu-events/arch/x86/goldmont/memory.json+270−10

This diff is large: its 291 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/goldmont/other.json+42−12

This diff is large: its 89 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/goldmont/pipeline.json+264−242

This diff is large: its 661 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/goldmont/virtual-memory.json+30−30

This diff is large: its 71 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/cache.json+194−171

This diff is large: its 735 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/floating-point.json+10−10

This diff is large: its 33 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/frontend.json+66−66

This diff is large: its 188 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/memory.json+21−0

This diff is large: its 168 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/other.json+10−10

This diff is large: its 34 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/pipeline.json+570−561

This diff is large: its 1433 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswell/virtual-memory.json+106−106

This diff is large: its 342 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/cache.json+200−177

This diff is large: its 764 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/floating-point.json+10−10

This diff is large: its 33 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/frontend.json+66−66

This diff is large: its 186 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/memory.json+28−0

This diff is large: its 224 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/other.json+10−10

This diff is large: its 34 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/pipeline.json+571−562

This diff is large: its 1442 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/haswellx/virtual-memory.json+106−106

This diff is large: its 338 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivybridge/cache.json+111−132

This diff is large: its 482 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivybridge/frontend.json+61−61

This diff is large: its 178 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivybridge/memory.json+12−12

This diff is large: its 38 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivybridge/other.json+10−10

This diff is large: its 34 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivybridge/pipeline.json+412−410

This diff is large: its 1153 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivybridge/virtual-memory.json+30−30

This diff is large: its 71 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivytown/cache.json+112−124

This diff is large: its 479 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivytown/frontend.json+61−61

This diff is large: its 178 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivytown/memory.json+12−12

This diff is large: its 38 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivytown/other.json+10−10

This diff is large: its 34 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivytown/pipeline.json+412−410

This diff is large: its 1153 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/ivytown/virtual-memory.json+30−30

This diff is large: its 85 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/mapfile.csv+1−4

This diff is large: its 12 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/silvermont/cache.json+2−1

This diff is large: its 15 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/cache.json+515−3875

This diff is large: its 4776 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/floating-point.json+2−3

This diff is large: its 24 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/frontend.json+121−111

This diff is large: its 507 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/memory.json+212−1906

This diff is large: its 2245 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/other.json+38−2

This diff is large: its 50 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/pipeline.json+492−481

This diff is large: its 1394 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylake/virtual-memory.json+137−125

This diff is large: its 401 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/cache.json+167−90

This diff is large: its 893 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/floating-point.json+1−2

This diff is large: its 12 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/frontend.json+24−24

This diff is large: its 205 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/memory.json+154−77

This diff is large: its 758 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/other.json+93−1

This diff is large: its 112 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/pipeline.json+22−22

This diff is large: its 169 lines aren’t shown here.

tools/perf/pmu-events/arch/x86/skylakex/virtual-memory.json+21−21

This diff is large: its 138 lines aren’t shown here.

tools/perf/pmu-events/jevents.c+38−1

This diff is large: its 53 lines aren’t shown here.

tools/perf/scripts/python/bin/mem-phys-addr-record+19−0

This diff is large: its 20 lines aren’t shown here.

tools/perf/scripts/python/bin/mem-phys-addr-report+3−0

This diff is large: its 4 lines aren’t shown here.

tools/perf/scripts/python/mem-phys-addr.py+95−0

This diff is large: its 96 lines aren’t shown here.

tools/perf/tests/attr.c+6−0

This diff is large: its 13 lines aren’t shown here.

tools/perf/tests/backward-ring-buffer.c+3−3

This diff is large: its 20 lines aren’t shown here.

tools/perf/tests/bp_signal.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/bpf-script-example.c+2−2

This diff is large: its 11 lines aren’t shown here.

tools/perf/tests/bpf.c+32−36

This diff is large: its 116 lines aren’t shown here.

tools/perf/tests/builtin-test.c+5−5

This diff is large: its 39 lines aren’t shown here.

tools/perf/tests/code-reading.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/dwarf-unwind.c+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/tests/keep-tracking.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/mmap-basic.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/openat-syscall-tp-fields.c+4−1

This diff is large: its 18 lines aren’t shown here.

tools/perf/tests/parse-events.c+0−1

This diff is large: its 8 lines aren’t shown here.

tools/perf/tests/perf-record.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/sample-parsing.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/shell/trace+probe_vfs_getname.sh+3−4

This diff is large: its 14 lines aren’t shown here.

tools/perf/tests/sw-clock.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/switch-tracking.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/task-exit.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/tests/thread-map.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/trace/beauty/Build+1−0

This diff is large: its 7 lines aren’t shown here.

tools/perf/trace/beauty/arch_errno_names.c+1−0

This diff is large: its 2 lines aren’t shown here.

tools/perf/trace/beauty/arch_errno_names.sh+100−0

This diff is large: its 101 lines aren’t shown here.

tools/perf/trace/beauty/beauty.h+5−0

This diff is large: its 17 lines aren’t shown here.

tools/perf/trace/beauty/flock.c+5−5

This diff is large: its 26 lines aren’t shown here.

tools/perf/trace/beauty/futex_val3.c+18−0

This diff is large: its 19 lines aren’t shown here.

tools/perf/ui/browsers/annotate.c+197−202

This diff is large: its 823 lines aren’t shown here.

tools/perf/ui/gtk/annotate.c+13−12

This diff is large: its 86 lines aren’t shown here.

tools/perf/util/Build+9−1

This diff is large: its 24 lines aren’t shown here.

tools/perf/util/annotate.c+340−312

This diff is large: its 1003 lines aren’t shown here.

tools/perf/util/annotate.h+42−34

This diff is large: its 122 lines aren’t shown here.

tools/perf/util/arm-spe-pkt-decoder.c+462−0

This diff is large: its 463 lines aren’t shown here.

tools/perf/util/arm-spe-pkt-decoder.h+43−0

This diff is large: its 44 lines aren’t shown here.

tools/perf/util/arm-spe.c+231−0

This diff is large: its 232 lines aren’t shown here.

tools/perf/util/arm-spe.h+31−0

This diff is large: its 32 lines aren’t shown here.

tools/perf/util/auxtrace.c+5−3

This diff is large: its 32 lines aren’t shown here.

tools/perf/util/auxtrace.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/bpf-loader.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/util/callchain.c+10−0

This diff is large: its 24 lines aren’t shown here.

tools/perf/util/callchain.h+2−0

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/cgroup.c+3−0

This diff is large: its 10 lines aren’t shown here.

tools/perf/util/cs-etm-decoder/Build+1−0

This diff is large: its 2 lines aren’t shown here.

tools/perf/util/cs-etm-decoder/cs-etm-decoder.c+513−0

This diff is large: its 514 lines aren’t shown here.

tools/perf/util/cs-etm-decoder/cs-etm-decoder.h+105−0

This diff is large: its 106 lines aren’t shown here.

tools/perf/util/cs-etm.c+1023−0

This diff is large: its 1024 lines aren’t shown here.

tools/perf/util/cs-etm.h+18−0

This diff is large: its 37 lines aren’t shown here.

tools/perf/util/data.c+0−10

This diff is large: its 17 lines aren’t shown here.

tools/perf/util/dso.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/env.c+47−0

This diff is large: its 60 lines aren’t shown here.

tools/perf/util/env.h+2−0

This diff is large: its 7 lines aren’t shown here.

tools/perf/util/event.c+8−0

This diff is large: its 22 lines aren’t shown here.

tools/perf/util/event.h+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/util/evlist.c+43−27

This diff is large: its 223 lines aren’t shown here.

tools/perf/util/evlist.h+9−6

This diff is large: its 56 lines aren’t shown here.

tools/perf/util/evsel.c+163−64

This diff is large: its 439 lines aren’t shown here.

tools/perf/util/evsel.h+7−5

This diff is large: its 43 lines aren’t shown here.

tools/perf/util/generate-cmdlist.sh+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/header.c+128−2

This diff is large: its 180 lines aren’t shown here.

tools/perf/util/header.h+8−1

This diff is large: its 35 lines aren’t shown here.

tools/perf/util/intel-bts.c+1−5

This diff is large: its 27 lines aren’t shown here.

tools/perf/util/intel-pt-decoder/Build+15−9

This diff is large: its 31 lines aren’t shown here.

tools/perf/util/intel-pt.c+3−8

This diff is large: its 41 lines aren’t shown here.

tools/perf/util/machine.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/util/map.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/metricgroup.c+6−4

This diff is large: its 42 lines aren’t shown here.

tools/perf/util/mmap.c+25−48

This diff is large: its 158 lines aren’t shown here.

tools/perf/util/mmap.h+2−2

This diff is large: its 13 lines aren’t shown here.

tools/perf/util/ordered-events.c+1−2

This diff is large: its 11 lines aren’t shown here.

tools/perf/util/ordered-events.h+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/parse-events.c+3−0

This diff is large: its 10 lines aren’t shown here.

tools/perf/util/path.c+14−0

This diff is large: its 25 lines aren’t shown here.

tools/perf/util/path.h+3−0

This diff is large: its 13 lines aren’t shown here.

tools/perf/util/pmu.c+71−16

This diff is large: its 182 lines aren’t shown here.

tools/perf/util/pmu.h+1−1

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/probe-event.c+61−24

This diff is large: its 211 lines aren’t shown here.

tools/perf/util/python-ext-sources+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/python.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/rblist.c+12−7

This diff is large: its 29 lines aren’t shown here.

tools/perf/util/rblist.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/scripting-engines/trace-event-python.c+2−1

This diff is large: its 17 lines aren’t shown here.

tools/perf/util/session.c+26−28

This diff is large: its 140 lines aren’t shown here.

tools/perf/util/session.h+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/sort.c+12−8

This diff is large: its 71 lines aren’t shown here.

tools/perf/util/srcline.c+5−4

This diff is large: its 31 lines aren’t shown here.

tools/perf/util/srcline.h+3−2

This diff is large: its 13 lines aren’t shown here.

tools/perf/util/stat-shadow.c+258−168

This diff is large: its 856 lines aren’t shown here.

tools/perf/util/stat.c+11−4

This diff is large: its 29 lines aren’t shown here.

tools/perf/util/stat.h+60−3

This diff is large: its 115 lines aren’t shown here.

tools/perf/util/string.c+46−0

This diff is large: its 50 lines aren’t shown here.

tools/perf/util/string2.h+2−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/symbol.c+5−0

This diff is large: its 12 lines aren’t shown here.

tools/perf/util/symbol.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/syscalltbl.c+4−0

This diff is large: its 11 lines aren’t shown here.

tools/perf/util/target.h+7−0

This diff is large: its 21 lines aren’t shown here.

tools/perf/util/thread_map.c+20−7

This diff is large: its 70 lines aren’t shown here.

tools/perf/util/thread_map.h+2−1

This diff is large: its 17 lines aren’t shown here.

tools/perf/util/time-utils.c+289−12

This diff is large: its 348 lines aren’t shown here.

tools/perf/util/time-utils.h+8−0

This diff is large: its 17 lines aren’t shown here.

tools/perf/util/tool.h+1−0

This diff is large: its 8 lines aren’t shown here.

tools/perf/util/unwind-libunwind-local.c+3−6

This diff is large: its 31 lines aren’t shown here.

tools/perf/util/unwind-libunwind.c+2−2

This diff is large: its 18 lines aren’t shown here.

tools/perf/util/util.c+1−1

This diff is large: its 9 lines aren’t shown here.

tools/perf/util/util.h+10−0

This diff is large: its 15 lines aren’t shown here.

Keyboard shortcuts

General

Show keyboard shortcuts
?
Focus the current search
/

Issues and pull requests

Open filters
f

Issues, pull requests and Inbox

Select next item
j↓
Select previous item
k↑
Open selected item
Enter
Dismiss filters or clear selection
Esc

Repository code

Go to file
t