System wide sampling like 'perf top' or 'perf record -a' read all threads /proc/xxx/maps before sampling. If there are any threads which generating a keeping growing huge maps, perf will do infinite loop during synthesizing. Nothing will be sampled. This patch fixes this issue by adding per-thread timeout to force stop this kind of endless proc map processing. PERF_RECORD_MISC_PROC_MAP_PARSE_TIME_OUT is introduced to indicate that the mmap record are truncated by time out. User will get warning notification when truncated mmap records are detected.
Showing the changes to include/uapi (1 of 4 files). Show all
Unified · Split
| @@ -565,6 +565,10 @@ struct perf_event_mmap_page { · 564 unmodified lines | |||
| 565 | #define PERF_RECORD_MISC_GUEST_KERNEL (4 << 0) | 565 | #define PERF_RECORD_MISC_GUEST_KERNEL (4 << 0) |
| 566 | #define PERF_RECORD_MISC_GUEST_USER (5 << 0) | 566 | #define PERF_RECORD_MISC_GUEST_USER (5 << 0) |
| 567 | 567 | ||
| 568 | /* | ||
| 569 | * Indicates that /proc/PID/maps parsing are truncated by time out. | ||
| 570 | */ | ||
| 571 | #define PERF_RECORD_MISC_PROC_MAP_PARSE_TIMEOUT (1 << 12) | ||
| 568 | /* | 572 | /* |
| 569 | * PERF_RECORD_MISC_MMAP_DATA and PERF_RECORD_MISC_COMM_EXEC are used on | 573 | * PERF_RECORD_MISC_MMAP_DATA and PERF_RECORD_MISC_COMM_EXEC are used on |
| 570 | * different events so can reuse the same bit position. | 574 | * different events so can reuse the same bit position. |