kernel/tools/perf/util/evsel.c, branch linux-rolling-stable

perf record: Add --call-graph fp,defer option for deferred callchains

2025-12-03T05:59:13Z

Add a new callchain record mode option for deferred callchains. For now it only works with FP (frame-pointer) mode. And add the missing feature detection logic to clear the flag on old kernels. $ perf record --call-graph fp,defer -vv true ... ------------------------------------------------------------ perf_event_attr: type 0 (PERF_TYPE_HARDWARE) size 136 config 0 (PERF_COUNT_HW_CPU_CYCLES) { sample_period, sample_freq } 4000 sample_type IP|TID|TIME|CALLCHAIN|PERIOD read_format ID|LOST disabled 1 inherit 1 mmap 1 comm 1 freq 1 enable_on_exec 1 task 1 sample_id_all 1 mmap2 1 comm_exec 1 ksymbol 1 bpf_event 1 defer_callchain 1 defer_output 1 ------------------------------------------------------------ sys_perf_event_open: pid 162755 cpu 0 group_fd -1 flags 0x8 sys_perf_event_open failed, error -22 switching off deferred callchain support Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim

perf tools: Minimal DEFERRED_CALLCHAIN support

2025-12-03T00:13:32Z

Add a new event type for deferred callchains and a new callback for the struct perf_tool. For now it doesn't actually handle the deferred callchains but it just marks the sample if it has the PERF_CONTEXT_ USER_DEFFERED in the callchain array. At least, perf report can dump the raw data with this change. Actually this requires the next commit to enable attr.defer_callchain, but if you already have a data file, it'll show the following result. $ perf report -D ... 0x2158@perf.data [0x40]: event: 22 . . ... raw event: size 64 bytes . 0000: 16 00 00 00 02 00 40 00 06 00 00 00 0b 00 00 00 ......@......... . 0010: 03 00 00 00 00 00 00 00 a7 7f 33 fe 18 7f 00 00 ..........3..... . 0020: 0f 0e 33 fe 18 7f 00 00 48 14 33 fe 18 7f 00 00 ..3.....H.3..... . 0030: 08 09 00 00 08 09 00 00 e6 7a e7 35 1c 00 00 00 .........z.5.... 121163447014 0x2158 [0x40]: PERF_RECORD_CALLCHAIN_DEFERRED(IP, 0x2): 2312/2312: 0xb00000006 ... FP chain: nr:3 ..... 0: 00007f18fe337fa7 ..... 1: 00007f18fe330e0f ..... 2: 00007f18fe331448 : unhandled! Reviewed-by: Ian Rogers Signed-off-by: Namhyung Kim

perf python: Correct copying of metric_leader in an evsel

2025-12-03T00:12:49Z

Ensure the metric_leader is copied and set up correctly. In compute_metric determine the correct metric_leader event to match the requested CPU. Fixes the handling of metrics particularly on hybrid machines. Signed-off-by: Ian Rogers Tested-by: Thomas Falcon Signed-off-by: Namhyung Kim

perf evsel: Skip store_evsel_ids for non-perf-event PMUs

2025-11-20T01:49:45Z

The IDs are associated with perf events and not applicable to non-perf event PMUs. The failure to generate the ids was causing perf stat record to fail. ``` $ perf stat record -a sleep 1 Performance counter stats for 'system wide': 47,941 context-switches # nan cs/sec cs_per_second 0.00 msec cpu-clock # 0.0 CPUs CPUs_utilized 3,261 cpu-migrations # nan migrations/sec migrations_per_second 516 page-faults # nan faults/sec page_faults_per_second 7,525,483 cpu_core/branch-misses/ # 2.3 % branch_miss_rate 322,069,004 cpu_core/branches/ # nan M/sec branch_frequency 1,895,684,291 cpu_core/cpu-cycles/ # nan GHz cycles_frequency 2,789,777,426 cpu_core/instructions/ # 1.5 instructions insn_per_cycle 7,074,765 cpu_atom/branch-misses/ # 3.2 % branch_miss_rate (49.89%) 224,225,412 cpu_atom/branches/ # nan M/sec branch_frequency (50.29%) 2,061,679,981 cpu_atom/cpu-cycles/ # nan GHz cycles_frequency (50.33%) 2,011,242,533 cpu_atom/instructions/ # 1.0 instructions insn_per_cycle (50.33%) TopdownL1 (cpu_core) # 9.0 % tma_bad_speculation # 28.3 % tma_frontend_bound # 35.2 % tma_backend_bound # 27.5 % tma_retiring TopdownL1 (cpu_atom) # 36.8 % tma_backend_bound (59.65%) # 22.8 % tma_frontend_bound (59.60%) # 11.6 % tma_bad_speculation # 28.8 % tma_retiring (59.59%) 1.006777519 seconds time elapsed $ perf stat report Performance counter stats for 'perf': 1,013,376,154 duration_time duration_time duration_time duration_time duration_time duration_time 47,941 context-switches 0.00 msec cpu-clock 3,261 cpu-migrations 516 page-faults 7,525,483 cpu_core/branch-misses/ 322,069,814 cpu_core/branches/ 322,069,004 cpu_core/branches/ 1,895,684,291 cpu_core/cpu-cycles/ 1,895,679,209 cpu_core/cpu-cycles/ 2,789,777,426 cpu_core/instructions/ cpu_core/cpu-cycles/ cpu_core/stalled-cycles-frontend/ cpu_core/cpu-cycles/ cpu_core/stalled-cycles-backend/ cpu_core/stalled-cycles-backend/ cpu_core/instructions/ cpu_core/stalled-cycles-frontend/ 7,074,765 cpu_atom/branch-misses/ (49.89%) 221,679,088 cpu_atom/branches/ (49.89%) 224,225,412 cpu_atom/branches/ (50.29%) 2,061,679,981 cpu_atom/cpu-cycles/ (50.33%) 2,016,259,567 cpu_atom/cpu-cycles/ (50.33%) 2,011,242,533 cpu_atom/instructions/ (50.33%) cpu_atom/cpu-cycles/ cpu_atom/stalled-cycles-frontend/ cpu_atom/cpu-cycles/ cpu_atom/stalled-cycles-backend/ cpu_atom/stalled-cycles-backend/ cpu_atom/instructions/ cpu_atom/stalled-cycles-frontend/ 17,145,113 cpu_core/INT_MISC.UOP_DROPPING/ 10,594,226,100 cpu_core/TOPDOWN.SLOTS/ 2,919,021,401 cpu_core/topdown-retiring/ 943,101,838 cpu_core/topdown-bad-spec/ 3,031,152,533 cpu_core/topdown-fe-bound/ 3,739,756,791 cpu_core/topdown-be-bound/ 1,909,501,648 cpu_atom/CPU_CLK_UNHALTED.CORE/ (60.04%) 3,516,608,359 cpu_atom/TOPDOWN_BE_BOUND.ALL/ (59.65%) 2,179,403,876 cpu_atom/TOPDOWN_FE_BOUND.ALL/ (59.60%) 2,745,732,458 cpu_atom/TOPDOWN_RETIRING.ALL/ (59.59%) 1.006777519 seconds time elapsed Some events weren't counted. Try disabling the NMI watchdog: echo 0 > /proc/sys/kernel/nmi_watchdog perf stat ... echo 1 > /proc/sys/kernel/nmi_watchdog ``` Reported-by: James Clark Closes: https://lore.kernel.org/lkml/ca0f0cd3-7335-48f9-8737-2f70a75b019a@linaro.org/ Signed-off-by: Ian Rogers Tested-by: Thomas Falcon Signed-off-by: Namhyung Kim

perf tools: Fix missing feature check for inherit + SAMPLE_READ

2025-11-12T00:43:37Z

It should also have PERF_SAMPLE_TID to enable inherit and PERF_SAMPLE_READ on recent kernels. Not having _TID makes the feature check wrongly detect the inherit and _READ support. It was reported that the following command failed due to the error in the missing feature check on Intel SPR machines. $ perf record -e '{cpu/mem-loads-aux/S,cpu/mem-loads,ldlat=3/PS}' -- ls Error: Failure to open event 'cpu/mem-loads,ldlat=3/PS' on PMU 'cpu' which will be removed. Invalid event (cpu/mem-loads,ldlat=3/PS) in per-thread mode, enable system wide with '-a'. Reviewed-by: Ian Rogers Fixes: 3b193a57baf15c468 ("perf tools: Detect missing kernel features properly") Reported-and-tested-by: Chen, Zide Closes: https://lore.kernel.org/lkml/20251022220802.1335131-1-zide.chen@intel.com/ Signed-off-by: Namhyung Kim

perf evsel: Remove unused metric_events variable

2025-11-04T04:57:21Z

The metric_events exist in the metric_expr list and so this variable has been unused for a while. Signed-off-by: Ian Rogers Signed-off-by: Namhyung Kim

perf evsel: Improvements to evselmatch

2025-10-15T14:59:11Z

Ensure both the perf_event_attr and alternate_hw_config are checked in the match. Don't mask the config if the perf_event_attr isn't a HARDWARE or HW_CACHE event. Add common early exit cases. Signed-off-by: Ian Rogers Tested-by: James Clark Signed-off-by: Namhyung Kim

perf record: Add ratio-to-prev term

2025-10-03T19:49:51Z

Provide ratio-to-prev term which allows the user to set the event sample period of two events corresponding to a desired ratio. If using on an Intel x86 platform with Auto Counter Reload support, also set corresponding event's config2 attribute with a bitmask which counters to reset and which counters to sample if the desired ratio is met or exceeded. On other platforms, only the sample period is affected by the ratio-to-prev term. Reviewed-by: Ian Rogers Signed-off-by: Thomas Falcon Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Andi Kleen Cc: Dapeng Mi Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Refactor retry/skip/fatal error handling

2025-10-03T19:49:51Z

For the sake of Intel topdown events commit 9eac5612da1c9102 ("perf stat: Don't skip failing group events") changed 'perf stat' error handling making it so that more errors were fatal and didn't report "" events. The change outside of topdown events was unintentional. The notion of "fatal" error handling was introduced in commit e0e6a6ca3ac211cc ("perf stat: Factor out open error handling") and refined in commits like commit cb5ef60067c11cc8 ("perf stat: Error out unsupported group leader immediately") to be an approach for avoiding later assertion failures in the code base. This change fixes those issues and removes the notion of a fatal error on an event. If all events fail to open then a fatal error occurs with the previous fatal error message. This seems to best match the notion of supported events and allowing some errors not to stop 'perf stat', while allowing the truly fatal no event case to terminate the tool early. The evsel->errored flag is only used in the stat code but always just meaning !evsel->supported although there is a comment about it being sticky. Force all evsels to be supported in evsel__init and then clear this when evsel__open fails. When an event is tried the supported is set to true again. This simplifies the notion of whether an evsel is broken. In the get_group_fd code, fail to get a group fd when the evsel isn't supported. If the leader isn't supported then it is also expected that there is no group_fd as the leader will have been skipped. Therefore change the BUG_ON test to be on supported rather than skippable. This corrects the assertion errors that were the reason for the previous fatal error handling. Fixes: 9eac5612da1c9102 ("perf stat: Don't skip failing group events") Reviewed-by: James Clark Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Chun-Tse Shao Cc: Dapeng Mi Cc: Howard Chu Cc: Ian Rogers Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Thomas Falcon Link: https://lore.kernel.org/r/20251002220727.1889799-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Ensure the fallback message is always written to

2025-09-19T19:52:24Z

The fallback message is unconditionally printed in places like record__open(). If no fallback is attempted this can lead to printing uninitialized data, crashes, etc. Fixes: c0a54341c0e89333 ("perf evsel: Introduce event fallback method") Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Howard Chu Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Signed-off-by: Arnaldo Carvalho de Melo

kernel/tools/perf/util/evsel.c, branch linux-rolling-stable

perf record: Add --call-graph fp,defer option for deferred callchains

perf tools: Minimal DEFERRED_CALLCHAIN support

perf python: Correct copying of metric_leader in an evsel

perf evsel: Skip store_evsel_ids for non-perf-event PMUs

perf tools: Fix missing feature check for inherit + SAMPLE_READ

perf evsel: Remove unused metric_events variable

perf evsel: Improvements to __evsel__match

perf record: Add ratio-to-prev term

perf stat: Refactor retry/skip/fatal error handling

perf evsel: Ensure the fallback message is always written to

perf evsel: Improvements to evselmatch