- 15 Feb, 2022 5 commits
-
-
Ian Rogers authored
Based on TMA_metrics-full.csv version 4.3 at 01.org: https://download.01.org/perfmon/ Events are updated to version 22: https://download.01.org/perfmon/HSX Json files generated by: https://github.com/intel/event-converter-for-linux-perf Tested: ... 6: Parse event definition strings : Ok 7: Simple expression parser : Ok ... 9: Parse perf pmu format : Ok 10: PMU events : 10.1: PMU event table sanity : Ok 10.2: PMU event map aliases : Ok 10.3: Parsing of PMU event table metrics : Ok 10.4: Parsing of PMU event table metrics with fake PMUs : Ok ... 68: Parse and process metrics : Ok ... 88: perf stat metrics (shadow stat) test : Ok 89: perf all metricgroups test : Ok 90: perf all metrics test : FAILED! 91: perf all PMU test : Ok ... Test 90 failed for Load_Miss_Real_Latency with <not counted> events: Performance counter stats for 'system wide': <not counted> mem_load_uops_retired.hit_lfb (0.00%) <not counted> MEM_LOAD_UOPS_RETIRED.L1_MISS (0.00%) <not counted> L1D_PEND_MISS.PENDING (0.00%) 1002638743 ns duration_time This is exposing a somewhat known issue with weak groups that can be worked around with: $ perf stat --metric-no-group -M Load_Miss_Real_Latency -a sleep 1 Performance counter stats for 'system wide': 9539883 mem_load_uops_retired.hit_lfb # 25.87 Load_Miss_Real_Latency (83.24%) 10876212 MEM_LOAD_UOPS_RETIRED.L1_MISS (66.68%) 528172960 L1D_PEND_MISS.PENDING (83.26%) 1001964165 ns duration_time Reviewed-by:Kan Liang <kan.liang@linux.intel.com> Signed-off-by:
Ian Rogers <irogers@google.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Alexandre Torgue <alexandre.torgue@foss.st.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: James Clark <james.clark@arm.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: John Garry <john.garry@huawei.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Cc: Zhengjun Xing <zhengjun.xing@linux.intel.com> Link: https://lore.kernel.org/r/20220201015858.1226914-6-irogers@google.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Ian Rogers authored
Based on TMA_metrics-full.csv version 4.3 at 01.org: https://download.01.org/perfmon/ Events are updated to version 1.14: https://download.01.org/perfmon/CLX Json files generated by: https://github.com/intel/event-converter-for-linux-perf Tested: ... 6: Parse event definition strings : Ok 7: Simple expression parser : Ok ... 9: Parse perf pmu format : Ok 10: PMU events : 10.1: PMU event table sanity : Ok 10.2: PMU event map aliases : Ok 10.3: Parsing of PMU event table metrics : Ok 10.4: Parsing of PMU event table metrics with fake PMUs : Ok ... 68: Parse and process metrics : Ok ... 88: perf stat metrics (shadow stat) test : Ok 89: perf all metricgroups test : Ok 90: perf all metrics test : FAILED! 91: perf all PMU test : Ok ... Test 90 failed due to MEM_PMM_Read_Latency as the test machine lacks optane memory, and the divide by 0 causes the metric not to print - which is intended behavior. Reviewed-by: -
Ian Rogers authored
Based on TMA_metrics-full.csv version 4.3 at 01.org: https://download.01.org/perfmon/ Events are still at version 7: https://download.01.org/perfmon/BDW-DE Json files generated by: https://github.com/intel/event-converter-for-linux-perf This adds TopdownL1_SMT metrics to bdwde-metrics.json. A discussed in: https://lore.kernel.org/all/20220129080929.837293-4-irogers@google.com/ The TMA_Metrics-full.csv was modified so that BDW-DE is in the server column with BDX, the Page_Walks_Utilization and Page_Walks_Utilization_SMT metrics are then copied from BDW. Tested: ... 6: Parse event definition strings : Ok 7: Simple expression parser : Ok ... 9: Parse perf pmu format : Ok 10: PMU events : 10.1: PMU event table sanity : Ok 10.2: PMU event map aliases : Ok 10.3: Parsing of PMU event table metrics : Ok 10.4: Parsing of PMU event table metrics with fake PMUs : Ok ... 68: Parse and process metrics : Ok ... 88: perf stat metrics (shadow stat) test : Ok 89: perf all metricgroups test : Ok 90: perf all metrics test : Skip 91: perf all PMU test : Ok ... 90 skips due to a lack of floating point samples, which is understandable. Suggested-by: -
Ian Rogers authored
Based on TMA_metrics-full.csv version 4.3 at 01.org: https://download.01.org/perfmon/ Events are updated to version 1.26: https://download.01.org/perfmon/SKX Json files generated by: https://github.com/intel/event-converter-for-linux-perf Fixes were made that allow the skx-metrics.json to successfully generate, bringing back TopdownL1 metrics. Tested: $ perf test ... 6: Parse event definition strings : Ok 7: Simple expression parser : Ok ... 9: Parse perf pmu format : Ok 10: PMU events : 10.1: PMU event table sanity : Ok 10.2: PMU event map aliases : Ok 10.3: Parsing of PMU event table metrics : Ok 10.4: Parsing of PMU event table metrics with fake PMUs : Ok ... 68: Parse and process metrics : Ok ... 88: perf stat metrics (shadow stat) test : Ok 89: perf all metricgroups test : Ok 90: perf all metrics test : Skip 91: perf all PMU test : Ok ... 90 skips due to a lack of floating point samples, which is understandable. Fixes: c4ad8fab ("perf vendor events: Update metrics for SkyLake Server") Reviewed-by: -
Ian Rogers authored
Some Intel TMA metrics compute a ratio that may divide by 0, which causes the metric not to print. This happens for metrics with FP_ARITH events. If we see these events in the result and would otherwise fail, then switch to a skip. Also, don't early exit when processing metrics. Reviewed-by:
John Garry <john.garry@huawei.com> Reviewed-by:
-
- 14 Feb, 2022 7 commits
-
-
Ian Rogers authored
Now unmap_ip is const, make contains symbol const. Signed-off-by:
-
Ian Rogers authored
The maps code has its own header, move the corresponding C function definitions to their own C file. In the process tidy and minimize includes. Committer notes: Add back the 'static' for maps__init() and maps__exit(). Signed-off-by:
-
Ian Rogers authored
Now purely accessed through new and delete, so reduce to file scope. Signed-off-by:
-
Ian Rogers authored
struct maps is reference counted, using a pointer is more idiomatic. Committer notes: Check maps__new() return. Signed-off-by:
-
Ian Rogers authored
struct maps is reference counted, using a pointer is more idiomatic. Committer notes: Delay: maps = machine__kernel_maps(&vmlinux); To after: machine__init(&vmlinux, "", HOST_KERNEL_ID); To avoid this on f34: In file included from /var/home/acme/git/perf/tools/perf/util/build-id.h:10, from /var/home/acme/git/perf/tools/perf/util/dso.h:13, from tests/vmlinux-kallsyms.c:8: In function ‘machine__kernel_maps’, inlined from ‘test__vmlinux_matches_kallsyms’ at tests/vmlinux-kallsyms.c:122:22: /var/home/acme/git/perf/tools/perf/util/machine.h:86:23: error: ‘vmlinux.kmaps’ is used uninitialized [-Werror=uninitialized] 86 | return machine->kmaps; | ~~~~~~~^~~~~~~ tests/vmlinux-kallsyms.c: In function ‘test__vmlinux_matches_kallsyms’: tests/vmlinux-kallsyms.c:121:34: note: ‘vmlinux’ declared here 121 | struct machine kallsyms, vmlinux; | ^~~~~~~ cc1: all warnings being treated as errors Signed-off-by: -
Christy Lee authored
Libbpf has deprecated the ability to keep track of object list inside libbpf, it now requires applications to track usage multiple bpf objects directly. Remove usage of bpf_object__next() API and hoist the tracking logic to perf. Signed-off-by:
Christy Lee <christylee@fb.com> Acked-by:
Song Liu <songliubraving@fb.com> Cc: Alexei Starovoitov <ast@kernel.org> Cc: Daniel Borkmann <daniel@iogearbox.net> Cc: bpf@vger.kernel.org Cc: kernel-team@fb.com Link: https://lore.kernel.org/bpf/20220212073054.1052880-3-andrii@kernel.orgSigned-off-by:
Andrii Nakryiko <andrii@kernel.org> Signed-off-by:
Jiri Olsa <jolsa@kernel.org> Signed-off-by:
-
Christy Lee authored
bpf_load_program() API is deprecated, remove perf's usage of the deprecated function. Add a __weak function declaration for libbpf version compatibility. Signed-off-by:
-
- 11 Feb, 2022 4 commits
-
-
Ian Rogers authored
Having functions to access nsinfo reduces the places where reference counting checking needs to be added. Signed-off-by:
-
Ian Rogers authored
Functions purely determine a value from the map and don't need to modify it. Move functions to C file as they are most commonly used via a function pointer. Signed-off-by:
-
Ian Rogers authored
Improve readability in perf_pmu__cpus_match() by using perf_cpu_map__for_each_cpu(). Signed-off-by:
-
Ian Rogers authored
Switch from directly accessing the perf_cpu_map to using the appropriate libperf API when possible. Using the API simplifies the job of refactoring use of perf_cpu_map. Signed-off-by:
-
- 10 Feb, 2022 24 commits
-
-
Alexey Bayduraev authored
Print path and name of a data file into raw dump (-D) <file_offset>@<path/file>: 0x2226a@perf.data [0x30]: event: 9 or 0x15cc36@perf.data/data.7 [0x30]: event: 9 Reviewed-by:
Riccardo Mancini <rickyman7@gmail.com> Signed-off-by:
Alexey Bayduraev <alexey.v.bayduraev@linux.intel.com> Tested-by:
Andi Kleen <ak@linux.intel.com> Acked-by:
Namhyung Kim <namhyung@kernel.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Antonov <alexander.antonov@linux.intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Alexei Budankov <abudankov@huawei.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: https://lore.kernel.org/r/e8378fd4910c10751b001be880705653989283c2.1642440724.git.alexey.v.bayduraev@linux.intel.comSigned-off-by:
-
Alexey Bayduraev authored
Load data directory files and provide basic raw dump and aggregated analysis support of data directories in report mode, still with no memory consumption optimizations. READER_MAX_SIZE is chosen based on the results of measurements on different machines on perf.data directory sizes >1GB. On machines with big core count (192 cores) the difference between 1MB and 2MB is about 4%. Other sizes (>2MB) are quite equal to 2MB. On machines with small core count (4-24) there is no differences between 1-16 MB sizes. So this constant is 2MB. Suggested-by:
Namhyung Kim <namhyung@gmail.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Antonov <alexander.antonov@linux.intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Alexei Budankov <abudankov@huawei.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: https://lore.kernel.org/r/3f10c13a226c0ceb53e88a082f847b91c1ae2c25.1642440724.git.alexey.v.bayduraev@linux.intel.comSigned-off-by:
-
Alexey Bayduraev authored
Implement compatibility checks for other modes and related command line options: asynchronous (--aio) trace streaming and affinity (--affinity) modes, pipe mode, AUX area tracing --snapshot and --aux-sample options, --switch-output, --switch-output-event, --switch-max-files and --timestamp-filename options. Parallel data streaming is compatible with Zstd compression (--compression-level) and external control commands (--control). CPU mask provided via -C option filters --threads specification masks. Reviewed-by:
-
Alexey Bayduraev authored
Extend --threads option in perf record command line interface. The option can have a value in the form of masks that specify CPUs to be monitored with data streaming threads and its layout in system topology. The masks can be filtered using CPU mask provided via -C option. The specification value can be user defined list of masks. Masks separated by colon define CPUs to be monitored by one thread and affinity mask of that thread is separated by slash. For example: <cpus mask 1>/<affinity mask 1>:<cpu mask 2>/<affinity mask 2> specifies parallel threads layout that consists of two threads with corresponding assigned CPUs to be monitored. The specification value can be a string e.g. "cpu", "core" or "package" meaning creation of data streaming thread for every CPU or core or package to monitor distinct CPUs or CPUs grouped by core or package. The option provided with no or empty value defaults to per-cpu parallel threads layout creating data streaming thread for every CPU being monitored. Document --threads option syntax and parallel data streaming modes in Documentation/perf-record.txt. Suggested-by:
-
Alexey Bayduraev authored
Provide --threads option in perf record command line interface. The option creates a data streaming thread for each CPU in the system. Document --threads option in Documentation/perf-record.txt. Reviewed-by:
-
Alexey Bayduraev authored
Introduce bytes_transferred and bytes_compressed stats so they would capture statistics for the related data buffer transfers. Reviewed-by:
-
Alexey Bayduraev authored
Introduce compressor object into mmap object so it could be used to pack the data stream from the corresponding kernel data buffer. Initialize and make use of the introduced per mmap compressor. Signed-off-by:
-
Alexey Bayduraev authored
Introduce a function to calculate the total amount of data written and use it to support the --max-size option. Reviewed-by:
-
Alexey Bayduraev authored
Introduce data file objects into mmap object so it could be used to process and store data stream from the corresponding kernel data buffer. Initialize data files located at mmap buffer objects so trace data can be written into several data file located at data directory. Signed-off-by:
-
Alexey Bayduraev authored
Start thread in detached state because its management is implemented via messaging to avoid any scaling issues. Block signals prior thread start so only main tool thread would be notified on external async signals during data collection. Thread affinity mask is used to assign eligible CPUs for the thread to run. Wait and sync on thread start using thread ack pipe. Reviewed-by:
-
Alexey Bayduraev authored
Signal thread to terminate by closing write fd of msg pipe. Receive THREAD_MSG__READY message as the confirmation of the thread's termination. Stop threads created for parallel trace streaming prior their stats processing. Reviewed-by:
-
Alexey Bayduraev authored
Introduce thread local variable and use it for threaded trace streaming. Use thread affinity mask instead of record affinity mask in affinity modes. Use evlist__ctlfd_update() to propagate control commands from thread object to global evlist object to enable evlist__ctlfd_* functionality. Move waking and sample statistic to struct record_thread and introduce record__waking function to calculate the total number of wakes. Reviewed-by:
-
Alexey Bayduraev authored
Introduce evlist__ctlfd_update() function to propagate external control commands to global evlist object. Reviewed-by:
-
Alexey Bayduraev authored
Introduce thread specific data object and array of such objects to store and manage thread local data. Implement functions to allocate, initialize, finalize and release thread specific data. Thread local maps and overwrite_maps arrays keep pointers to mmap buffer objects to serve according to maps thread mask. Thread local pollfd array keeps event fds connected to mmaps buffers according to maps thread mask. Thread control commands are delivered via thread local comm pipes and ctlfd_pos fd. External control commands (--control option) are delivered via evlist ctlfd_pos fd and handled by the main tool thread. Reviewed-by:
-
Alexey Bayduraev authored
Introduce a function to duplicate an existing file descriptor in the fdarray structure. The function returns the position of the duplicated file descriptor. Reviewed-by:
-
Alexey Bayduraev authored
Introduce affinity and mmap thread masks. Thread affinity mask defines CPUs that a thread is allowed to run on. Thread maps mask defines mmap data buffers the thread serves to stream profiling data from. Reviewed-by:
-
Namhyung Kim authored
Stats from discarded entries should be omitted. But a lock class may have both good and bad entries. If the first entry was bad, we can zero-fill the stats and only add good stats if any. The entry can remove the discard state if it finds a good entry later. Signed-off-by:
Jiri Olsa <jolsa@redhat.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: https://lore.kernel.org/r/20220127000050.3011493-7-namhyung@kernel.orgSigned-off-by:
-
Namhyung Kim authored
The -c or --combine-locks option is to merge lock instances in the same class into a single entry. It compares the name of the locks and marks duplicated entries using lock_stat->combined. # perf lock report Name acquired contended avg wait (ns) total wait (ns) max wait (ns) min wait (ns) rcu_read_lock 251225 0 0 0 0 0 &(ei->i_block_re... 8731 0 0 0 0 0 &sb->s_type->i_l... 8731 0 0 0 0 0 hrtimer_bases.lock 5261 0 0 0 0 0 hrtimer_bases.lock 2626 0 0 0 0 0 hrtimer_bases.lock 1953 0 0 0 0 0 hrtimer_bases.lock 1382 0 0 0 0 0 cpu_hotplug_lock 1350 0 0 0 0 0 hrtimer_bases.lock 1273 0 0 0 0 0 hrtimer_bases.lock 1269 0 0 0 0 0 # perf lock report -c Name acquired contended avg wait (ns) total wait (ns) max wait (ns) min wait (ns) rcu_read_lock 251225 0 0 0 0 0 hrtimer_bases.lock 39450 0 0 0 0 0 &sb->s_type->i_l... 10301 1 662 662 662 662 ptlock_ptr(page) 10173 2 701 1402 760 642 &(ei->i_block_re... 8732 0 0 0 0 0 &xa->xa_lock 8088 0 0 0 0 0 &base->lock 6705 0 0 0 0 0 &p->pi_lock 5549 0 0 0 0 0 &dentry->d_lockr... 5010 4 1274 5097 1844 789 &ep->lock 3958 0 0 0 0 0 Signed-off-by: -
Namhyung Kim authored
It has 20 character spaces for name so lock names shorter than 20 should be printed without ellipsis. Before: # perf lock report Name acquired contended avg wait (ns) total wait (ns) max wait (ns) min wait (ns) rcu_read_lock 251225 0 0 0 0 0 &(ei->i_block_re... 8731 0 0 0 0 0 &sb->s_type->i_l... 8731 0 0 0 0 0 hrtimer_bases.lo... 5261 0 0 0 0 0 hrtimer_bases.lo... 2626 0 0 0 0 0 hrtimer_bases.lo... 1953 0 0 0 0 0 hrtimer_bases.lo... 1382 0 0 0 0 0 cpu_hotplug_lock... 1350 0 0 0 0 0 After: # perf lock report Name acquired contended avg wait (ns) total wait (ns) max wait (ns) min wait (ns) rcu_read_lock 251225 0 0 0 0 0 &(ei->i_block_re... 8731 0 0 0 0 0 &sb->s_type->i_l... 8731 0 0 0 0 0 hrtimer_bases.lock 5261 0 0 0 0 0 hrtimer_bases.lock 2626 0 0 0 0 0 hrtimer_bases.lock 1953 0 0 0 0 0 hrtimer_bases.lock 1382 0 0 0 0 0 cpu_hotplug_lock 1350 0 0 0 0 0 Signed-off-by: -
Namhyung Kim authored
Instead of the random order, sort it by lock class name. Before: # perf lock info -m Address of instance: name of class 0xffffa0d940ac5310: &dentry->d_lockref.lock 0xffffa0c20b0e1cb0: &dentry->d_lockref.lock 0xffffa0d8e051cc48: &base->lock 0xffffa0d94f992110: &anon_vma->rwsem 0xffffa0d947a4f278: (null) 0xffffa0c208f6e108: &map->lock 0xffffa0c213ad32c8: &cfs_rq->removed.lock 0xffffa0c20d695888: &parent->list_lock 0xffffa0c278775278: (null) 0xffffa0c212ad4690: &dentry->d_lockref.lock After: # perf lock info -m Address of instance: name of class 0xffffa0c20d538800: &(&sig->stats_lock)->lock 0xffffa0c216d4ec40: &(&sig->stats_lock)->lock 0xffffa1fe4cb04610: &(__futex_data.queues)[i].lock 0xffffa1fe4cb07750: &(__futex_data.queues)[i].lock 0xffffa1fe4cb07b50: &(__futex_data.queues)[i].lock 0xffffa1fe4cb0b850: &(__futex_data.queues)[i].lock 0xffffa1fe4cb0bcd0: &(__futex_data.queues)[i].lock 0xffffa1fe4cb0e5d0: &(__futex_data.queues)[i].lock 0xffffa1fe4cb11ad0: &(__futex_data.queues)[i].lock Signed-off-by:
-
Namhyung Kim authored
As evsel__intval() returns u64, we can just use it as is. Signed-off-by:
-
Namhyung Kim authored
The hlist_head has a single entry so we can save some memory. Signed-off-by:
-
Namhyung Kim authored
Likewise, it should use a proper name in case the task runs under chroot. Signed-off-by:
-
Namhyung Kim authored
When reading build-id from a DSO, it should consider if it's from a chroot task. In that case, the path is different so it needs to prepend the root directory to access the file correctly. Signed-off-by:
-
