kernel/linux.git/tools/perf/util/evsel.c, branch v6.4

perf evsel: Separate bpf_counter_list and bpf_filters, can be used at the same time

2023-05-26T19:03:43+00:00

'struct evsel' uses a union for the two lists. This turned out to be error prone. For example: [root@quaco ~]# perf stat --bpf-prog 5246 Error: cpu-clock event does not have sample flags 66c660 failed to set filter "(null)" on event cpu-clock with 2 (No such file or directory) [root@quaco ~]# perf stat --bpf-prog 5246 Fix this issue by separating the two lists. Fixes: 56ec9457a4a20c5e ("perf bpf filter: Implement event sample filtering") Signed-off-by: Song Liu Acked-by: Namhyung Kim Cc: Jiri Olsa Cc: Namhyung Kim Cc: kernel-team@meta.com Link: https://lore.kernel.org/r/20230519235757.3613719-1-song@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf parse-events: Do not break up AUX event group

2023-05-10T17:19:20+00:00

Do not assume which events may have a PMU name, allowing the logic to keep an AUX event group together. Example: Before: $ perf record --no-bpf-event -c 10 -e '{intel_pt//,tlb_flush.stlb_any/aux-sample-size=8192/pp}:u' -- sleep 0.1 WARNING: events were regrouped to match PMUs Cannot add AUX area sampling to a group leader $ After: $ perf record --no-bpf-event -c 10 -e '{intel_pt//,tlb_flush.stlb_any/aux-sample-size=8192/pp}:u' -- sleep 0.1 [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.078 MB perf.data ] $ perf script -F-dso,+addr | grep -C5 tlb_flush.stlb_any | head -11 sleep 20444 [003] 7939.510243: 1 branches:uH: 7f5350cc82a2 dl_main+0x9a2 => 7f5350cb38f0 _dl_add_to_namespace_list+0x0 sleep 20444 [003] 7939.510243: 1 branches:uH: 7f5350cb3908 _dl_add_to_namespace_list+0x18 => 7f5350cbb080 rtld_mutex_dummy+0x0 sleep 20444 [003] 7939.510243: 1 branches:uH: 7f5350cc8350 dl_main+0xa50 => 0 [unknown] sleep 20444 [003] 7939.510244: 1 branches:uH: 7f5350cc83ca dl_main+0xaca => 7f5350caeb60 _dl_process_pt_gnu_property+0x0 sleep 20444 [003] 7939.510245: 1 branches:uH: 7f5350caeb60 _dl_process_pt_gnu_property+0x0 => 0 [unknown] sleep 20444 7939.510245: 10 tlb_flush.stlb_any/aux-sample-size=8192/pp: 0 7f5350caeb60 _dl_process_pt_gnu_property+0x0 sleep 20444 [003] 7939.510254: 1 branches:uH: 7f5350cc87fe dl_main+0xefe => 7f5350ccd240 strcmp+0x0 sleep 20444 [003] 7939.510254: 1 branches:uH: 7f5350cc8862 dl_main+0xf62 => 0 [unknown] sleep 20444 [003] 7939.510255: 1 branches:uH: 7f5350cc9cdc dl_main+0x23dc => 0 [unknown] sleep 20444 [003] 7939.510257: 1 branches:uH: 7f5350cc89f6 dl_main+0x10f6 => 7f5350cb9530 _dl_setup_hash+0x0 sleep 20444 [003] 7939.510257: 1 branches:uH: 7f5350cc8a2d dl_main+0x112d => 7f5350cb3990 _dl_new_object+0x0 $ Fixes: 347c2f0a0988c59c ("perf parse-events: Sort and group parsed events") Suggested-by: Ian Rogers Signed-off-by: Adrian Hunter Tested-by: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lore.kernel.org/r/20230508093952.27482-3-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Modify group pmu name for software events

2023-05-10T17:19:20+00:00

If we have a group of {cycles,faults} then we need the faults software event to appear to be on the same PMU as cycles so that we don't split the group in parse_events__sort_events_and_fix_groups. This case is relatively easy as cycles is the leader and will have a PMU name. In the reverse case, {faults,cycles} we still need faults to appear to have the PMU name of cycles but the old behavior is just to return "cpu". For hybrid this fails as cycles will be on "cpu_core" or "cpu_atom", causing faults to be split into a different group. Change the behavior for software events so that the whole group is searched for the named PMU. Signed-off-by: Ian Rogers Tested-by: Adrian Hunter Cc: Ahmad Yasin Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Caleb Biggers Cc: Edward Baker Cc: Florian Fischer Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: John Garry Cc: Kajol Jain Cc: Kan Liang Cc: Kang Minchul Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Perry Taylor Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Rob Herring Cc: Samantha Alt Cc: Stephane Eranian Cc: Sumanth Korikkar Cc: Suzuki Poulouse Cc: Thomas Richter Cc: Tiezhu Yang Cc: Weilin Wang Cc: Xing Zhengjun Cc: Yang Jihong Link: https://lore.kernel.org/r/20230502223851.2234828-20-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo

perf stat: Introduce skippable evsels

2023-05-10T15:35:02+00:00

'perf stat' with no arguments will use default events and metrics. These events may fail to open even with kernel and hypervisor disabled. When these fail then the permissions error appears even though they were implicitly selected. This is particularly a problem with the automatic selection of the TopdownL1 metric group on certain architectures like Skylake: $ perf stat true Error: Access to performance monitoring and observability operations is limited. Consider adjusting /proc/sys/kernel/perf_event_paranoid setting to open access to performance monitoring and observability operations for processes without CAP_PERFMON, CAP_SYS_PTRACE or CAP_SYS_ADMIN Linux capability. More information can be found at 'Perf events and tool security' document: https://www.kernel.org/doc/html/latest/admin-guide/perf-security.html perf_event_paranoid setting is 2: -1: Allow use of (almost) all events by all users Ignore mlock limit after perf_event_mlock_kb without CAP_IPC_LOCK >= 0: Disallow raw and ftrace function tracepoint access >= 1: Disallow CPU event access >= 2: Disallow kernel profiling To make the adjusted perf_event_paranoid setting permanent preserve it in /etc/sysctl.conf (e.g. kernel.perf_event_paranoid = ) $ This patch adds skippable evsels that when they fail to open won't cause termination and will appear as "" in output. The TopdownL1 events, from the metric group, are marked as skippable. This turns the failure above to: $ perf stat perf bench internals synthesize Computing performance of single threaded perf event synthesis by synthesizing events on the perf process itself: Average synthesis took: 49.287 usec (+- 0.083 usec) Average num. events: 3.000 (+- 0.000) Average time per event 16.429 usec Average data synthesis took: 49.641 usec (+- 0.085 usec) Average num. events: 11.000 (+- 0.000) Average time per event 4.513 usec Performance counter stats for 'perf bench internals synthesize': 1,222.38 msec task-clock:u # 0.993 CPUs utilized 0 context-switches:u # 0.000 /sec 0 cpu-migrations:u # 0.000 /sec 162 page-faults:u # 132.529 /sec 774,445,184 cycles:u # 0.634 GHz (49.61%) 1,640,969,811 instructions:u # 2.12 insn per cycle (59.67%) 302,052,148 branches:u # 247.102 M/sec (59.69%) 1,807,718 branch-misses:u # 0.60% of all branches (59.68%) 5,218,927 CPU_CLK_UNHALTED.REF_XCLK:u # 4.269 M/sec # 17.3 % tma_frontend_bound # 56.4 % tma_retiring # nan % tma_backend_bound # nan % tma_bad_speculation (60.01%) 536,580,469 IDQ_UOPS_NOT_DELIVERED.CORE:u # 438.965 M/sec (60.33%) INT_MISC.RECOVERY_CYCLES_ANY:u 5,223,936 CPU_CLK_UNHALTED.ONE_THREAD_ACTIVE:u # 4.274 M/sec (40.31%) 774,127,250 CPU_CLK_UNHALTED.THREAD:u # 633.297 M/sec (50.34%) 1,746,579,518 UOPS_RETIRED.RETIRE_SLOTS:u # 1.429 G/sec (50.12%) 1,940,625,702 UOPS_ISSUED.ANY:u # 1.588 G/sec (49.70%) 1.231055525 seconds time elapsed 0.258327000 seconds user 0.965749000 seconds sys $ The event INT_MISC.RECOVERY_CYCLES_ANY:u is skipped as it can't be opened with paranoia 2 on Skylake. With a lower paranoia, or as root, all events/metrics are computed. Signed-off-by: Ian Rogers Tested-by: Kan Liang Cc: Adrian Hunter Cc: Ahmad Yasin Cc: Alexander Shishkin Cc: Andi Kleen Cc: Athira Rajeev Cc: Caleb Biggers Cc: Edward Baker Cc: Florian Fischer Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: John Garry Cc: Kajol Jain Cc: Kang Minchul Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Perry Taylor Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Rob Herring Cc: Samantha Alt Cc: Stephane Eranian Cc: Sumanth Korikkar Cc: Suzuki Poulouse Cc: Thomas Richter Cc: Tiezhu Yang Cc: Weilin Wang Cc: Xing Zhengjun Cc: Yang Jihong Link: https://lore.kernel.org/r/20230502223851.2234828-3-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Introduce evsel__name_is() method to check if the evsel name is equal to a given string

2023-04-24T17:28:11+00:00

This makes the logic a bit clear by avoiding the !strcmp() pattern and also a way to intercept the pointer if we need to do extra validation on it or to do lazy setting of evsel->name via evsel__name(evsel). Reviewed-by: "Liang, Kan" Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lore.kernel.org/lkml/ZEGLM8VehJbS0gP2@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Use zfree() to reduce chances of use after free

2023-04-12T13:03:37+00:00

Do defensive programming by using zfree() to initialize freed pointers to NULL, so that eventual use after free result in a NULL pointer deref instead of more subtle behaviour. Also remove one NULL test before free(), as it accepts a NULL arg and we get one line shaved not doing it explicitely. Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Avoid SEGV if delete is called on NULL

2023-04-11T00:59:29+00:00

Seen in "perf stat --bpf-counters --for-each-cgroup test" running in a container: libbpf: Failed to bump RLIMIT_MEMLOCK (err = -1), you might need to do it explicitly! libbpf: Error in bpf_object__probe_loading():Operation not permitted(1). Couldn't load trivial BPF program. Make sure your kernel supports BPF (CONFIG_BPF_SYSCALL=y) and/or that RLIMIT_MEMLOCK is set to big enough value. libbpf: failed to load object 'bperf_cgroup_bpf' libbpf: failed to load BPF skeleton 'bperf_cgroup_bpf': -1 Failed to load cgroup skeleton #0 0x55f28a650981 in list_empty tools/include/linux/list.h:189 #1 0x55f28a6593b4 in evsel__exit util/evsel.c:1518 #2 0x55f28a6596af in evsel__delete util/evsel.c:1544 #3 0x55f28a89d166 in bperf_cgrp__destroy util/bpf_counter_cgroup.c:283 #4 0x55f28a899e9a in bpf_counter__destroy util/bpf_counter.c:816 #5 0x55f28a659455 in evsel__exit util/evsel.c:1520 #6 0x55f28a6596af in evsel__delete util/evsel.c:1544 #7 0x55f28a640d4d in evlist__purge util/evlist.c:148 #8 0x55f28a640ea6 in evlist__delete util/evlist.c:169 #9 0x55f28a4efbf2 in cmd_stat tools/perf/builtin-stat.c:2598 #10 0x55f28a6050c2 in run_builtin tools/perf/perf.c:330 #11 0x55f28a605633 in handle_internal_command tools/perf/perf.c:384 #12 0x55f28a6059fb in run_argv tools/perf/perf.c:428 #13 0x55f28a6061d3 in main tools/perf/perf.c:562 Signed-off-by: Ian Rogers Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Florian Fischer Cc: Ingo Molnar Cc: Jiri Olsa Cc: Kan Liang Cc: Mark Rutland Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20230410205659.3131608-1-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Add BPF event filter support

2023-03-15T14:08:34+00:00

Use --filter option to set BPF filter for generic events other than the tracepoints or Intel PT. The BPF program will check the sample data and filter according to the expression. For example, the below is the typical perf record for frequency mode. The sample period started from 1 and increased gradually. $ sudo ./perf record -e cycles true $ sudo ./perf script perf-exec 2272336 546683.916875: 1 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916892: 1 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916899: 3 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916905: 17 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916911: 100 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916917: 589 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916924: 3470 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) perf-exec 2272336 546683.916930: 20465 cycles: ffffffff828499b8 perf_event_exec+0x298 ([kernel.kallsyms]) true 2272336 546683.916940: 119873 cycles: ffffffff8283afdd perf_iterate_ctx+0x2d ([kernel.kallsyms]) true 2272336 546683.917003: 461349 cycles: ffffffff82892517 vma_interval_tree_insert+0x37 ([kernel.kallsyms]) true 2272336 546683.917237: 635778 cycles: ffffffff82a11400 security_mmap_file+0x20 ([kernel.kallsyms]) When you add a BPF filter to get samples having periods greater than 1000, the output would look like below: $ sudo ./perf record -e cycles --filter 'period > 1000' true $ sudo ./perf script perf-exec 2273949 546850.708501: 5029 cycles: ffffffff826f9e25 finish_wait+0x5 ([kernel.kallsyms]) perf-exec 2273949 546850.708508: 32409 cycles: ffffffff826f9e25 finish_wait+0x5 ([kernel.kallsyms]) perf-exec 2273949 546850.708526: 143369 cycles: ffffffff82b4cdbf xas_start+0x5f ([kernel.kallsyms]) perf-exec 2273949 546850.708600: 372650 cycles: ffffffff8286b8f7 __pagevec_lru_add+0x117 ([kernel.kallsyms]) perf-exec 2273949 546850.708791: 482953 cycles: ffffffff829190de __mod_memcg_lruvec_state+0x4e ([kernel.kallsyms]) true 2273949 546850.709036: 501985 cycles: ffffffff828add7c tlb_gather_mmu+0x4c ([kernel.kallsyms]) true 2273949 546850.709292: 503065 cycles: 7f2446d97c03 _dl_map_object_deps+0x973 (/usr/lib/x86_64-linux-gnu/ld-linux-x86-64.so.2) Committer notes: Add stubs for perf_bpf_filter__prepare() and perf_bpf_filter__destroy() to tools/perf/util/python.c to keep it building. Signed-off-by: Namhyung Kim Acked-by: Jiri Olsa Cc: Adrian Hunter Cc: Andi Kleen Cc: Hao Luo Cc: Ian Rogers Cc: Ingo Molnar Cc: James Clark Cc: Kan Liang Cc: Leo Yan Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Song Liu Cc: Stephane Eranian Cc: bpf@vger.kernel.org Link: https://lore.kernel.org/r/20230314234237.3008956-2-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Remove use_uncore_alias

2023-03-13T20:42:27+00:00

This flag used to be used when regrouping uncore events in particular due to wildcard matches. This is now handled by sorting evlist and so the flag is redundant. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Florian Fischer Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: John Garry Cc: Kajol Jain Cc: Kan Liang Cc: Kim Phillips Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sean Christopherson Cc: Steinar H. Gunderson Cc: Stephane Eranian Cc: Suzuki Poulouse Cc: Xing Zhengjun Link: https://lore.kernel.org/r/20230312021543.3060328-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo

perf evsel: Add function to compute group PMU name

2023-03-13T20:42:02+00:00

The computed name respects software events and aux event groups, such that the pmu_name is changed to be that of the aux event leader or group leader for software events. This is done as a later change will split events that are in different PMUs into different groups. Committer notes: Added a stub for this new function so that 'perf test python' passes. Signed-off-by: Ian Rogers Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Florian Fischer Cc: Ingo Molnar Cc: James Clark Cc: Jiri Olsa Cc: John Garry Cc: Kajol Jain Cc: Kan Liang Cc: Kim Phillips Cc: Leo Yan Cc: Mark Rutland Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Ravi Bangoria Cc: Sean Christopherson Cc: Steinar H. Gunderson Cc: Stephane Eranian Cc: Suzuki Poulouse Cc: Xing Zhengjun Link: https://lore.kernel.org/r/20230312021543.3060328-2-irogers@google.com Signed-off-by: Arnaldo Carvalho de Melo