History

arm64: enable perf events based hard lockup detector

With the recent feature added to enable perf events to use pseudo NMIs as interrupts on platforms which support GICv3 or later, its now been possible to enable hard lockup detector (or NMI watchdog) on arm64 platforms. So enable corresponding support. One thing to note here is that normally lockup detector is initialized just after the early initcalls but PMU on arm64 comes up much later as device_initcall(). To cope with that, override arch_perf_nmi_is_available() to let the watchdog framework know PMU not ready, and inform the framework to re-initialize lockup detection once PMU has been initialized. [dianders@chromium.org: only HAVE_HARDLOCKUP_DETECTOR_PERF if the PMU config is enabled] Link: https://lkml.kernel.org/r/20230523073952.1.I60217a63acc35621e13f10be16c0cd7c363caf8c@changeid Link: https://lkml.kernel.org/r/20230519101840.v5.18.Ia44852044cdcb074f387e80df6b45e892965d4a1@changeid Co-developed-by: Sumit Garg <sumit.garg@linaro.org> Signed-off-by: Sumit Garg <sumit.garg@linaro.org> Co-developed-by: Pingfan Liu <kernelfans@gmail.com> Signed-off-by: Pingfan Liu <kernelfans@gmail.com> Signed-off-by: Lecopzer Chen <lecopzer.chen@mediatek.com> Signed-off-by: Douglas Anderson <dianders@chromium.org> Cc: Andi Kleen <ak@linux.intel.com> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: Chen-Yu Tsai <wens@csie.org> Cc: Christophe Leroy <christophe.leroy@csgroup.eu> Cc: Colin Cross <ccross@android.com> Cc: Daniel Thompson <daniel.thompson@linaro.org> Cc: "David S. Miller" <davem@davemloft.net> Cc: Guenter Roeck <groeck@chromium.org> Cc: Ian Rogers <irogers@google.com> Cc: Marc Zyngier <maz@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Masayoshi Mizuma <msys.mizuma@gmail.com> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Nicholas Piggin <npiggin@gmail.com> Cc: Petr Mladek <pmladek@suse.com> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: "Ravi V. Shankar" <ravi.v.shankar@intel.com> Cc: Ricardo Neri <ricardo.neri@intel.com> Cc: Stephane Eranian <eranian@google.com> Cc: Stephen Boyd <swboyd@chromium.org> Cc: Tzung-Bi Shih <tzungbi@chromium.org> Cc: Will Deacon <will@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

Douglas Anderson committed 3 years ago

Commit d7a0fe91 parent 94946f95 files · +32 −2Browse files

Unified · Split

5 files
  1. arch/arm64/Kconfig +3 −0
  2. arch/arm64/kernel/watchdog_hld.c +12 −0
  3. drivers/perf/arm_pmu.c +5 −0
  4. drivers/perf/arm_pmuv3.c +10 −2
  5. include/linux/perf/arm_pmu.h +2 −0
arch/arm64/Kconfig+3−0
@@ -203,12 +203,15 @@ config ARM64 · 202 unmodified lines
203203 select HAVE_FUNCTION_ERROR_INJECTION
204204 select HAVE_FUNCTION_GRAPH_TRACER
205205 select HAVE_GCC_PLUGINS
206+ select HAVE_HARDLOCKUP_DETECTOR_PERF if PERF_EVENTS && \
207+ HW_PERF_EVENTS && HAVE_PERF_EVENTS_NMI
206208 select HAVE_HW_BREAKPOINT if PERF_EVENTS
207209 select HAVE_IOREMAP_PROT
208210 select HAVE_IRQ_TIME_ACCOUNTING
209211 select HAVE_KVM
210212 select HAVE_NMI
211213 select HAVE_PERF_EVENTS
214+ select HAVE_PERF_EVENTS_NMI if ARM64_PSEUDO_NMI
212215 select HAVE_PERF_REGS
213216 select HAVE_PERF_USER_STACK_DUMP
214217 select HAVE_PREEMPT_DYNAMIC_KEY
arch/arm64/kernel/watchdog_hld.c+12−0
@@ -1,5 +1,7 @@
11 // SPDX-License-Identifier: GPL-2.0
2+#include <linux/nmi.h>
23 #include <linux/cpufreq.h>
4+#include <linux/perf/arm_pmu.h>
35
46 /*
57 * Safe maximum CPU frequency in case a particular platform doesn't implement
@@ -22,3 +24,13 @@ u64 hw_nmi_get_sample_period(int watchdog_thresh) · 16 unmodified lines
2224
2325 return (u64)max_cpu_freq * watchdog_thresh;
2426 }
27+
28+bool __init arch_perf_nmi_is_available(void)
29+{
30+ /*
31+ * hardlockup_detector_perf_init() will success even if Pseudo-NMI turns off,
32+ * however, the pmu interrupts will act like a normal interrupt instead of
33+ * NMI and the hardlockup detector would be broken.
34+ */
35+ return arm_pmu_irq_is_nmi();
36+}
drivers/perf/arm_pmu.c+5−0
@@ -687,6 +687,11 @@ static int armpmu_get_cpu_irq(struct arm_pmu *pmu, int cpu) · 686 unmodified lines
687687 return per_cpu(hw_events->irq, cpu);
688688 }
689689
690+bool arm_pmu_irq_is_nmi(void)
691+{
692+ return has_nmi;
693+}
694+
690695 /*
691696 * PMU hardware loses all context when a CPU goes offline.
692697 * When a CPU is hotplugged back in, since some hardware registers are
drivers/perf/arm_pmuv3.c+10−2
@@ -22,6 +22,7 @@ · 21 unmodified lines
2222 #include <linux/platform_device.h>
2323 #include <linux/sched_clock.h>
2424 #include <linux/smp.h>
25+#include <linux/nmi.h>
2526
2627 #include <asm/arm_pmuv3.h>
2728
@@ -1348,10 +1349,17 @@ static struct platform_driver armv8_pmu_driver = { · 1320 unmodified lines
13481349
13491350 static int __init armv8_pmu_driver_init(void)
13501351 {
1352+ int ret;
1353+
13511354 if (acpi_disabled)
1352− return platform_driver_register(&armv8_pmu_driver);
1355+ ret = platform_driver_register(&armv8_pmu_driver);
13531356 else
1354− return arm_pmu_acpi_probe(armv8_pmuv3_pmu_init);
1357+ ret = arm_pmu_acpi_probe(armv8_pmuv3_pmu_init);
1358+
1359+ if (!ret)
1360+ lockup_detector_retry_init();
1361+
1362+ return ret;
13551363 }
13561364 device_initcall(armv8_pmu_driver_init)
13571365
include/linux/perf/arm_pmu.h+2−0
@@ -171,6 +171,8 @@ void kvm_host_pmu_init(struct arm_pmu *pmu); · 170 unmodified lines
171171 #define kvm_host_pmu_init(x) do { } while(0)
172172 #endif
173173
174+bool arm_pmu_irq_is_nmi(void);
175+
174176 /* Internal functions only for core arm_pmu code */
175177 struct arm_pmu *armpmu_alloc(void);
176178 void armpmu_free(struct arm_pmu *pmu);