- 23 May, 2023 7 commits
-
-
Ian Rogers authored
Metrics are updated to make TMA info metric names synchronized. Metrics were generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
Kan Liang <kan.liang@linux.intel.com> Signed-off-by:
Ian Rogers <irogers@google.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: John Garry <john.g.garry@oracle.com> Cc: Kajol Jain <kjain@linux.ibm.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Richter <tmricht@linux.ibm.com> Cc: Xing Zhengjun <zhengjun.xing@linux.intel.com> Link: https://lore.kernel.org/r/20230517173805.602113-8-irogers@google.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Ian Rogers authored
Update icelake events to v1.18 including the new events MEM_LOAD_MISC_RETIRED.UC and SQ_MISC.BUS_LOCK. Metrics are updated to make TMA info metric names synchronized. Events and metrics were generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
-
Ian Rogers authored
Metrics are updated to make TMA info metric names synchronized. Metrics were generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
-
Ian Rogers authored
Update elkhartlake to v1.04 that marks deprecated a number of events and adds additional description to MEM_BOUND_STALLS.IFETCH. The events data was generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
-
Ian Rogers authored
Update cascadelakex to v1.18 including the new events INT_MISC.CLEARS_COUNT, FP_ARITH_INST_RETIRED.VECTOR, FP_ARITH_INST_RETIRED.SCALAR, FP_ARITH_INST_RETIRED.8_FLOPS and FP_ARITH_INST_RETIRED.4_FLOPS. Metrics are updated to make TMA info metric names synchronized. Events and metrics were generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
-
Ian Rogers authored
Update broadwell events to v28, broadwellde to v10, broadwellx to v21. Including the new events FP_ARITH_INST_RETIRED.VECTOR, and FP_ARITH_INST_RETIRED.4_FLOPS. Metrics are updated to make TMA info metric names synchronized. Events and metrics were generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
-
Ian Rogers authored
Update events to v21 including the new event SQ_MISC.BUS_LOCK and improved comments. Metrics are updated to make TMA info metric names synchronized. Events and metrics were generated by: https://github.com/intel/perfmon/blob/main/scripts/create_perf_json.pyReviewed-by:
-
- 22 May, 2023 2 commits
-
-
Anup Sharma authored
This commit adds support for testing the JSON output generated by the 'perf data' command's conversion to JSON functionality. The test script now includes a step to ensure that the resulting JSON file contains valid data. Changes: V1 -> V2: Added a check for the existence of the result output file. Replaced the usage of jq with json.load for validating the JSON format. Checks using ShellCheck and checkpatch, addressing and resolving warnings. Removed the unnecessary root permission check. Modified the 'perf record' command to avoid requiring root permissions. Committer testing: $ perf test to-json 115: 'perf data convert --to-json' command test : Ok $ perf test -v to-json Couldn't bump rlimit(MEMLOCK), failures may take place when creating BPF maps, etc 115: 'perf data convert --to-json' command test : --- start --- test child forked, pid 1746867 Testing Perf Data Convertion Command to JSON Perf Data Converter Command to JSON [SUCCESS] Validating Perf Data Converted JSON file The file contains valid JSON format [SUCCESS] test child finished with 0 ---- end ---- 'perf data convert --to-json' command test: Ok $ Signed-off-by:
Anup Sharma <anupnewsmail@gmail.com> Acked-by:
-
Arnaldo Carvalho de Melo authored
To pick up fixes that were already merged upstream. Signed-off-by:
-
- 19 May, 2023 3 commits
-
-
Tiezhu Yang authored
The __NR_execve definition for i386 was deleted by mistake in the commit ece7f7c0 ("perf bench syscall: Add fork syscall benchmark"), add it to fix the build error on i386. Fixes: ece7f7c0 ("perf bench syscall: Add fork syscall benchmark") Reported-by:
Linux Kernel Functional Testing <lkft@linaro.org> Signed-off-by:
Tiezhu Yang <yangtiezhu@loongson.cn> Tested-by:
Naresh Kamboju <naresh.kamboju@linaro.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Tiezhu Yang <yangtiezhu@loongson.cn> Cc: loongson-kernel@lists.loongnix.cn Closes: https://lore.kernel.org/all/CA+G9fYvgBR1iB0CorM8OC4AM_w_tFzyQKHc+rF6qPzJL=TbfDQ@mail.gmail.com/ Link: https://lore.kernel.org/r/1684480657-2375-1-git-send-email-yangtiezhu@loongson.cnSigned-off-by:
-
Ian Rogers authored
Address the warning: ``` tests/attr.py:155: DeprecationWarning: The SafeConfigParser class has been renamed to ConfigParser in Python 3.2. This alias will be removed in Python 3.12. Use ConfigParser directly instead. parser = configparser.SafeConfigParser() ``` by removing the word 'Safe'. Reviewed-by:
James Clark <james.clark@arm.com> Signed-off-by:
-
Ian Rogers authored
Previously hard coded events/metrics were used, update for the use of the TopdownL1 json metric group. Reported-by:
-
- 17 May, 2023 4 commits
-
-
Arnaldo Carvalho de Melo authored
To pick the changes from: e0bddc19 ("x86/mm: Reduce untagged_addr() overhead for systems without LAM") This only causes these perf files to be rebuilt: CC /tmp/build/perf/bench/mem-memcpy-x86-64-asm.o CC /tmp/build/perf/bench/mem-memset-x86-64-asm.o And addresses this perf build warning: Warning: Kernel ABI header at 'tools/arch/x86/include/asm/disabled-features.h' differs from latest version at 'arch/x86/include/asm/disabled-features.h' diff -u tools/arch/x86/include/asm/disabled-features.h arch/x86/include/asm/disabled-features.h Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: https://lore.kernel.org/lkml/ZGTpdlzrlRjjnY6K@kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
To pick the changes in this cset: a03c376e ("x86/arch_prctl: Add AMX feature numbers as ABI constants") 23e5d9ec ("x86/mm/iommu/sva: Make LAM and SVA mutually exclusive") 2f8794bd ("x86/mm: Provide arch_prctl() interface for LAM") This picks these new prctls in a third range, that was also added to the tools/perf/trace/beauty/arch_prctl.c beautifier. $ tools/perf/trace/beauty/x86_arch_prctl.sh > /tmp/before $ cp arch/x86/include/uapi/asm/prctl.h tools/arch/x86/include/uapi/asm/prctl.h $ tools/perf/trace/beauty/x86_arch_prctl.sh > /tmp/after $ diff -u /tmp/before /tmp/after @@ -20,3 +20,11 @@ [0x2003 - 0x2001]= "MAP_VDSO_64", }; +#define x86_arch_prctl_codes_3_offset 0x4001 +static const char *x86_arch_prctl_codes_3[] = { + [0x4001 - 0x4001]= "GET_UNTAG_MASK", + [0x4002 - 0x4001]= "ENABLE_TAGGED_ADDR", + [0x4003 - 0x4001]= "GET_MAX_TAG_BITS", + [0x4004 - 0x4001]= "FORCE_TAGGED_SVA", +}; + $ With this 'perf trace' can translate those numbers into strings and use the strings in filter expressions: # perf trace -e prctl 0.000 ( 0.011 ms): DOM Worker/3722622 prctl(option: SET_NAME, arg2: 0x7f9c014b7df5) = 0 0.032 ( 0.002 ms): DOM Worker/3722622 prctl(option: SET_NAME, arg2: 0x7f9bb6b51580) = 0 5.452 ( 0.003 ms): StreamT~ns #30/3722623 prctl(option: SET_NAME, arg2: 0x7f9bdbdfeb70) = 0 5.468 ( 0.002 ms): StreamT~ns #30/3722623 prctl(option: SET_NAME, arg2: 0x7f9bdbdfea70) = 0 24.494 ( 0.009 ms): IndexedDB #556/3722624 prctl(option: SET_NAME, arg2: 0x7f562a32ae28) = 0 24.540 ( 0.002 ms): IndexedDB #556/3722624 prctl(option: SET_NAME, arg2: 0x7f563c6d4b30) = 0 670.281 ( 0.008 ms): systemd-userwo/3722339 prctl(option: SET_NAME, arg2: 0x564be30805c8) = 0 670.293 ( 0.002 ms): systemd-userwo/3722339 prctl(option: SET_NAME, arg2: 0x564be30800f0) = 0 ^C# This addresses this perf build warning: Warning: Kernel ABI header at 'tools/arch/x86/include/uapi/asm/prctl.h' differs from latest version at 'arch/x86/include/uapi/asm/prctl.h' diff -u tools/arch/x86/include/uapi/asm/prctl.h arch/x86/include/uapi/asm/prctl.h Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Chang S. Bae <chang.seok.bae@intel.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: https://lore.kernel.org/lkml/ZGTjNPpD3FOWfetM@kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
This is to get the changes from: 68674f94 ("x86: don't use REP_GOOD or ERMS for small memory copies") 20f3337d ("x86: don't use REP_GOOD or ERMS for small memory clearing") This also make the 'perf bench mem' files stop referring to the erms versions that gone away with the above patches. That addresses these perf tools build warning: Warning: Kernel ABI header at 'tools/arch/x86/lib/memcpy_64.S' differs from latest version at 'arch/x86/lib/memcpy_64.S' diff -u tools/arch/x86/lib/memcpy_64.S arch/x86/lib/memcpy_64.S Warning: Kernel ABI header at 'tools/arch/x86/lib/memset_64.S' differs from latest version at 'arch/x86/lib/memset_64.S' diff -u tools/arch/x86/lib/memset_64.S arch/x86/lib/memset_64.S Signed-off-by:
-
Arnaldo Carvalho de Melo authored
To pick the changes from: 3d8f61bf ("x86: KVM: Add common feature flag for AMD's PSFD") 3763bf58 ("x86/cpufeatures: Redefine synthetic virtual NMI bit as AMD's "real" vNMI") 6449dcb0 ("x86: CPUID and CR3/CR4 flags for Linear Address Masking") be8de49b ("x86/speculation: Identify processors vulnerable to SMT RSB predictions") e7862eda ("x86/cpu: Support AMD Automatic IBRS") faabfcb1 ("x86/cpu, kvm: Add the SMM_CTL MSR not present feature") 5b909d4a ("x86/cpu, kvm: Add the Null Selector Clears Base feature") 84168ae7 ("x86/cpu, kvm: Move X86_FEATURE_LFENCE_RDTSC to its native leaf") a9dc9ec5 ("x86/cpu, kvm: Add the NO_NESTED_DATA_BP feature") f8df91e7 ("x86/cpufeatures: Add macros for Intel's new fast rep string features") 78335aac ("x86/cpufeatures: Add Bandwidth Monitoring Event Configuration feature flag") f334f723 ("x86/cpufeatures: Add Slow Memory Bandwidth Allocation feature flag") a018d2e3 ("x86/cpufeatures: Add Architectural PerfMon Extension bit") This causes these perf files to be rebuilt and brings some X86_FEATURE that will be used when updating the copies of tools/arch/x86/lib/mem{cpy,set}_64.S with the kernel sources: CC /tmp/build/perf/bench/mem-memcpy-x86-64-asm.o CC /tmp/build/perf/bench/mem-memset-x86-64-asm.o And addresses this perf build warning: Warning: Kernel ABI header at 'tools/arch/x86/include/asm/cpufeatures.h' differs from latest version at 'arch/x86/include/asm/cpufeatures.h' diff -u tools/arch/x86/include/asm/cpufeatures.h arch/x86/include/asm/cpufeatures.h Cc: Borislav Petkov (AMD) <bp@alien8.de> Cc: Kim Phillips <kim.phillips@amd.com> Cc: Jim Mattson <jmattson@google.com> Cc: Babu Moger <babu.moger@amd.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Kan Liang <kan.liang@linux.intel.com> Cc: Sean Christopherson <seanjc@google.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Paolo Bonzini <pbonzini@redhat.com> Cc: Tom Lendacky <thomas.lendacky@amd.com> Link: https://lore.kernel.org/lkml/ZGTTw642q8mWgv2Y@kernel.orgSigned-off-by:
-
- 16 May, 2023 3 commits
-
-
Arnaldo Carvalho de Melo authored
To pick the changes in these csets: 7608f70a ("s390: wire up memfd_secret system call") That add support for this new syscall in tools such as 'perf trace'. For instance, this is now possible (adapted from the x86_64 test output): # perf trace -v -e memfd_secret event qualifier tracepoint filter: (common_pid != 13375 && common_pid != 3713) && (id == 447) ^C# That is the filter expression attached to the raw_syscalls:sys_{enter,exit} tracepoints. $ grep memfd_secret tools/perf/arch/x86/entry/syscalls/syscall_64.tbl 447 common memfd_secret sys_memfd_secret $ This addresses this perf build warnings: Warning: Kernel ABI header at 'tools/perf/arch/s390/entry/syscalls/syscall.tbl' differs from latest version at 'arch/s390/kernel/syscalls/syscall.tbl' diff -u tools/perf/arch/s390/entry/syscalls/syscall.tbl arch/s390/kernel/syscalls/syscall.tbl Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Heiko Carstens <hca@linux.ibm.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Vasily Gorbik <gor@linux.ibm.com> Link: https://lore.kernel.org/lkml/ZGPMW0p++D1Jdvf6@kernel.orgSigned-off-by:
-
Arnaldo Carvalho de Melo authored
ddc65971 ("prctl: add PR_GET_AUXV to copy auxv to userspace") To pick the changes in: That don't result in any changes in tooling: $ tools/perf/trace/beauty/prctl_option.sh > before $ cp include/uapi/linux/prctl.h tools/include/uapi/linux/prctl.h $ tools/perf/trace/beauty/prctl_option.sh > after $ diff -u before after $ This actually adds a new prctl arg, but it has to be dealt with differently, as it is not in sequence with the other arguments. Just silences this perf tools build warning: Warning: Kernel ABI header at 'tools/include/uapi/linux/prctl.h' differs from latest version at 'include/uapi/linux/prctl.h' diff -u tools/include/uapi/linux/prctl.h include/uapi/linux/prctl.h Signed-off-by:
-
Ian Rogers authored
Some metrics may not have a metric_group which can result in segvs with "perf stat --topdown". Add a condition for the no metric_group case. Fixes: 1647cd5b ("perf stat: Implement --topdown using json metrics") Reported-by:
-
- 15 May, 2023 21 commits
-
-
Ilkka Koskinen authored
Add JSON files for AmpereOne core PMU events. Reviewed-by:
John Garry <john.g.garry@oracle.com> Signed-off-by:
Doug Rady <dcrady@os.amperecomputing.com> Signed-off-by:
Ilkka Koskinen <ilkka@os.amperecomputing.com> Acked-by:
-
Changbin Du authored
The pager will result stdout in full buffering mode instead of line buffering. We need to make the trace visible timely. Signed-off-by:
Changbin Du <changbin.du@huawei.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: https://lore.kernel.org/r/20230513074000.733550-1-changbin.du@huawei.comSigned-off-by:
-
Namhyung Kim authored
hists__find_annotations() allows to move to next or previous symbols for annotation using the arrow keys. But TUI annotate_browser__run() uses the RIGHT key as ENTER to handle jump/call instructions. That makes the navigation to the next function impossible. I'd like to change it back to move the next symbol but I'm afraid if some users get confused. So I added a new pair of keys to handle that. Signed-off-by:
Namhyung Kim <namhyung@kernel.org> Acked-by:
-
Namhyung Kim authored
When the source argument of the "mov" instruction looks like below, it didn't parse the whole operand and just stopped at the first comma. mov (%rbx,%rax,1),%rcx Fix it by checking the parentheses and move it to the closing one. Signed-off-by:
-
Namhyung Kim authored
I found that the "decq", "incq", "testq", "tzcnt" instructions didn't parse the operands properly. Add them to the "x86__instructions" table to fix the issue. Signed-off-by:
-
Ben Hutchings authored
When building man pages from a Git checkout, we consistently set the man page date based on when the input was last changed. Otherwise, it defaults to the build time, which is not reproducible. Allow the date to be set through the KBUILD_BUILD_TIMESTAMP variable, as for timestamps in the kernel itself. Signed-off-by:
Ben Hutchings <benh@debian.org> Acked-by: Ian Rogers<irogers@google.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Salvatore Bonaccorso <carnil@debian.org> Link: https://lore.kernel.org/r/ZF/1F1P+b9qZ/vVH@decadent.org.ukSigned-off-by:
-
Ben Hutchings authored
When building perf documentation with asciidoc, we use "git log" to find the last commit date of each doc source and pass that to asciidoc to use as the man page date. When using asciidoctor, however, the current date is always used instead. Defining perf_date like we do for asciidoc also doesn't work because we're not using DocBook as an intermediate format. The asciidoctor man page backend looks for the variable "docdate", so set that instead. Signed-off-by:
-
Ian Rogers authored
Check a bogus PMU fails and that a known PMU succeeds. Limit to PMUs known cpu, cpu_atom and armv8_pmuv3_0 ones. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
perf build: Don't use -ftree-loop-distribute-patterns and -gno-variable-location-views in the python feature test when building with clang-13 Using -ftree-loop-distribute-patterns and -gno-variable-location-views in the python feature test when building with clang-16 results in: 16 80.04 clearlinux:latest : FAIL clang version 16.0.1 clang-16: error: unknown argument: '-gno-variable-location-views' clang-16: error: unknown argument: '-gno-variable-location-views' clang-16: error: optimization flag '-ftree-loop-distribute-patterns' is not supported [-Werror,-Wignored-optimization-argument] clang-16: error: optimization flag '-ftree-loop-distribute-patterns' is not supported [-Werror,-Wignored-optimization-argument] error: command '/usr/sbin/clang' failed with exit code 1 Noticed when building on a docker.io/library/clearlinux:latest container. Signed-off-by: -
Ian Rogers authored
Move to print-events.c and make static. Signed-off-by:
-
Ian Rogers authored
Now that hybrid bugs are fixed sufficient to run TopdownL1 metrics, don't implicitly disable them for hybrid. Signed-off-by:
-
Ian Rogers authored
Ids/events from a metric are turned into an event string and parsed; setup_metric_events matches the id back to the parsed evsel. With hybrid the same event may exist on both PMUs with the same name and be being used by metrics at the same time. A metric on cpu_core therefore shouldn't match against evsels on cpu_atom, or the metric will compute the wrong value. Make the matching sensitive to the PMU being parsed. Signed-off-by:
-
Ian Rogers authored
Don't rewrite metrics across PMUs as the result events likely won't be found. Identify metrics with a pair of PMU name and metric name. Signed-off-by:
-
Ian Rogers authored
Fix the metrics tma_memory_bound on alderlake cpu_core and tma_microcode_sequencer on alderlake cpu_atom, where metrics had be rewritten across PMUs. Fix MEM_BOUND_STALLS_AT_RET_CORRECTION which is an aux metric but lacks a hash prefix. Add PMU prefixes for cpu_core/cpu_atom events to avoid wildcard opening the events. Signed-off-by:
-
Ian Rogers authored
Wire up the --cputype value to limit which metrics are parsed. Signed-off-by:
-
Ian Rogers authored
Hybrid systems may define the same metric for different PMUs, this can cause confusion of events. To avoid this make the referenced metric searches PMU specific, matching that in the table. Signed-off-by:
-
Ian Rogers authored
On hybrid systems the topdown events don't share a fixed counter on the atom core, so they don't require the sorting the perf metric supporting PMUs do. Signed-off-by:
-
Ian Rogers authored
Bring back the behavior of not auto-merging hybrid events by delegating to a test in pmu. Signed-off-by:
-
Ian Rogers authored
Avoid the parser error: ''' $ perf stat -e 'cycles/name=l1d/' true event syntax error: 'cycles/name=l1d/' \___ parser error ''' by combining the name and legacy cache cases in the parser. Signed-off-by: -
Ian Rogers authored
Avoid the parser error: ''' $ perf stat -e 'cycles/name=name/' true event syntax error: 'cycles/name=name/' \___ parser error ''' by turning the term back to a string if it is on the right. Add PMU and generic parsing tests. Signed-off-by: -
Ian Rogers authored
An event like "cpu/instructions/" typically parses due to there being a sysfs event called instructions. On hybrid recursive parsing means that the hardware event is encoded in the attribute, with the PMU being placed in the high bits of the config: ''' $ perf stat -vv -e 'cpu_core/cycles/' true ... ------------------------------------------------------------ perf_event_attr: size 136 config 0x400000000 sample_type IDENTIFIER read_format TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING disabled 1 inherit 1 enable_on_exec 1 exclude_guest 1 ------------------------------------------------------------ ''' Make this behavior the default by adding a new term type and token for hardware events. The token gathers both the numeric config and the parsed name, so that if the token appears like "cycles/name=cycles/" then the token can be handled like a name. The numeric value isn't sufficient to distinguish say "cpu-cycles" from "cycles". Extend the parse-events test so that all current non-PMU hardware parsing tests, also test with the PMU cpu - more than half the change. Signed-off-by:
-
