- 03 May, 2024 2 commits
-
-
Ian Rogers authored
To avoid directory scans in perf it is going to be assumed that sysfs event names are either lower or upper case. 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: Bjorn Helgaas <bhelgaas@google.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: James Clark <james.clark@arm.com> Cc: Jing Zhang <renyu.zj@linux.alibaba.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Jonathan Corbet <corbet@lwn.net> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Randy Dunlap <rdunlap@infradead.org> Cc: Ravi Bangoria <ravi.bangoria@amd.com> Cc: Thomas Richter <tmricht@linux.ibm.com> Link: https://lore.kernel.org/r/20240502213507.2339733-3-irogers@google.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Ian Rogers authored
Add JSON to the test name. Reviewed-by:
-
- 02 May, 2024 10 commits
-
-
Namhyung Kim authored
I found that the debug build was a slowed down a lot by the maps lock code since it checks the invariants whenever it gets the pointer to the lock. This means it checks twice the invariants before and after the access. Instead, let's move the checking code within the lock area but after any modification and remove it from the read paths. This would remove (more than) half of the maps lock overhead. The time for perf report with a huge data file (200k+ of MMAP2 events). Non-debug Before After --------- -------- -------- 2m 43s 6m 45s 4m 21s Reviewed-by:Namhyung Kim <namhyung@kernel.org> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kan Liang <kan.liang@linux.intel.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: https://lore.kernel.org/r/20240429225738.1491791-1-namhyung@kernel.orgSigned-off-by:
-
James Clark authored
When the config validation functions are warning about ETMv3, they do it based on "not ETMv4". If the drivers aren't all loaded or the hardware doesn't support Coresight it will appear as "not ETMv4" and then Perf will print the error message "... not supported in ETMv3 ..." which is wrong and confusing. cs_etm_is_etmv4() is also misnamed because it also returns true for ETE because ETE has a superset of the ETMv4 metadata files. Although this was always done in the correct order so it wasn't a bug. Improve all this by making a single get version function which also handles not present as a separate case. Change the ETMv3 error message to only print when ETMv3 is detected, and add a new error message for the not present case. Reviewed-by:
Leo Yan <leo.yan@linux.dev> Signed-off-by:
James Clark <james.clark@arm.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: Kan Liang <kan.liang@linux.intel.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Mike Leach <mike.leach@linaro.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Suzuki Poulouse <suzuki.poulose@arm.com> Cc: Will Deacon <will@kernel.org> Link: https://lore.kernel.org/r/20240501135753.508022-4-james.clark@arm.comSigned-off-by:
-
James Clark authored
Most functions already have cs_etm_pmu, so it's a bit neater to pass it through rather than itr only to convert it again. Signed-off-by:
-
James Clark authored
The perf_cpu struct makes some iterators simpler and avoids some mistakes with interchanging CPU IDs with indexes etc. At the moment in this file the conversion to an integer is done somewhere in the middle of the call tree. Change it to delay the conversion to an int until the leaf functions. Some of the usage patterns are duplicated, so instead of changing them all, make cs_etm_get_ro() more reusable and use that everywhere. cs_etm_get_ro() didn't return an error before, but return one now so that it can also be used where an error is needed. Continue to ignore the error where it was already ignored. Use cs_etm_pmu_path_exists() instead of cs_etm_get_ro() in cs_etm_is_etmv4() because cs_etm_get_ro() prints a warning, but path exists is sufficient for this use case. Signed-off-by:
-
Namhyung Kim authored
I sometimes see ("unknown type") in the result and it was because it didn't check the type of stack variables properly during the instruction tracking. The stack can carry constant values (without type info) and if the target instruction is accessing the stack location, it resulted in the "unknown type". Maybe we could pick one of integer types for the constant, but it doesn't really mean anything useful. Let's just drop the stack slot if it doesn't have a valid type info. Here's an example how it got the unknown type. Note that 0xffffff48 = -0xb8. ----------------------------------------------------------- find data type for 0xffffff48(reg6) at ... CU for ... frame base: cfa=0 fbreg=6 scope: [2/2] (die:11cb97f) bb: [37 - 3a] var [37] reg15 type='int' size=0x4 (die:0x1180633) bb: [40 - 4b] mov [40] imm=0x1 -> reg13 var [45] reg8 type='sigset_t*' size=0x8 (die:0x11a39ee) mov [45] imm=0x1 -> reg2 <--- here reg2 has a constant bb: [215 - 237] mov [218] reg2 -> -0xb8(stack) constant <--- and save it to the stack mov [225] reg13 -> -0xc4(stack) constant call [22f] find_task_by_vgpid call [22f] return -> reg0 type='struct task_struct*' size=0x8 (die:0x11881e8) bb: [5c8 - 5cf] bb: [2fb - 302] mov [2fb] -0xc4(stack) -> reg13 constant bb: [13b - 14d] mov [143] 0xd50(reg3) -> reg5 type='struct task_struct*' size=0x8 (die:0xa31f3c) bb: [153 - 153] chk [153] reg6 offset=0xffffff48 ok=0 kind=0 fbreg <--- access here found by insn track: 0xffffff48(reg6) type-offset=0 type='G<EF>^K<F6><AF>U' size=0 (die:0xffffffffffffffff) Signed-off-by: -
Namhyung Kim authored
The instruction tracking should be the same for the both registers. Just do it once and compare the result with multi regs as with the previous patches. Then we don't need to call find_data_type_block() separately for each reg. Let's remove the 'reg' argument from the relevant functions. Signed-off-by:
-
Namhyung Kim authored
The following instruction pattern is used to access a global variable. mov $0x231c0, %rax movsql %edi, %rcx mov -0x7dc94ae0(,%rcx,8), %rcx cmpl $0x0, 0xa60(%rcx,%rax,1) <<<--- here The first instruction set the address of the per-cpu variable (here, it is 'runqueues' of type 'struct rq'). The second instruction seems like a cpu number of the per-cpu base. The third instruction get the base offset of per-cpu area for that cpu. The last instruction compares the value of the per-cpu variable at the offset of 0xa60. Signed-off-by:
-
Namhyung Kim authored
Like per-cpu base offset array, sometimes it accesses the global variable directly using the offset. Allow this type of instructions as long as it finds a global variable for the address. movslq %edi, %rcx mov -0x7dc94ae0(,%rcx,8), %rcx <<<--- here As %rcx has a valid type (i.e. array index) from the first instruction, it will be checked by the first case in check_matching_type(). But as it's not a pointer type, the match will fail. But in this case, it should check if it accesses the kernel global array variable. Signed-off-by:
-
Namhyung Kim authored
Currently it looks up global variables from the current CU using address and name. But it sometimes fails to find a variable as the variable can come from a different CU - but it's still strange it failed to find a declaration for some reason. Anyway, it can collect all global variables from all CU once and then lookup them later on. This slightly improves the success rate of my test data set. Signed-off-by:
-
Namhyung Kim authored
This function is to search all global variables in the CU. We want to have the list of global variables at once and match them later. Signed-off-by:
-
- 27 Apr, 2024 28 commits
-
-
Arnaldo Carvalho de Melo authored
We can't default to doing parallel tests as there are tests that compete for the same resources and thus clash, for instance tests that put in place 'perf probe' probes, that clean the probes without regard to other tests needs, ARM64 coresight tests, Intel PT ones, etc. So reintroduce --p/--parallel and make -S/--sequential the default. We need to come up with infrastructure that state which tests can't run in parallel because they need exclusive access to some resource, something as simple as "probes" that would then avoid 'perf probe' tests from running while other such test is running, or make the tests more resilient, till then we can't use parallel mode as default. While at it, document all these options in the 'perf test' man page. Reported-by:
Adrian Hunter <adrian.hunter@intel.com> Reported-by:
-
Arnaldo Carvalho de Melo authored
To pick up the changes in this cset: 3c7a8e19 ("uapi: introduce uapi-friendly macros for GENMASK") That just causes perf to rebuild. Its just some macros going to an uapi header that we now have to grab a copy into tools/ as well. This addresses this perf build warning: Warning: Kernel ABI header differences: diff -u tools/include/linux/bits.h include/linux/bits.h Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kan Liang <kan.liang@linux.intel.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Paolo Bonzini <pbonzini@redhat.com> Link: https://lore.kernel.org/lkml/ZiwJsFOBez0MS4r9@x1Signed-off-by:
-
Arnaldo Carvalho de Melo authored
To pick the changes from: 95a6ccbd ("x86/bhi: Mitigate KVM by default") ec9404e4 ("x86/bhi: Add BHI mitigation knob") be482ff9 ("x86/bhi: Enumerate Branch History Injection (BHI) bug") 0f4a8376 ("x86/bhi: Define SPEC_CTRL_BHI_DIS_S") 7390db8a ("x86/bhi: Add support for clearing branch history at syscall entry") 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 differences: diff -u tools/arch/x86/include/asm/cpufeatures.h arch/x86/include/asm/cpufeatures.h Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Daniel Sneddon <daniel.sneddon@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Pawan Gupta <pawan.kumar.gupta@linux.intel.com> Cc: Thomas Gleixner <tglx@linutronix.de> Link: https://lore.kernel.org/lkml/ZirIx4kPtJwGFZS0@x1Signed-off-by:
-
Namhyung Kim authored
The loop in hists__find_annotations() never set the 'nd' pointer to NULL and it makes stdio output repeating the last element forever. I think it doesn't set to NULL for TUI to prevent it from exiting unexpectedly. But it should just set on stdio mode. Fixes: d001c7a7 ("perf annotate-data: Add hist_entry__annotate_data_tui()") Signed-off-by:
-
Ian Rogers authored
Memory sanitizer lacks an interceptor for scandirat, reporting all memory it allocates as uninitialized. Memory sanitizer has a scandir interceptor so use the fallback function in this case. This allows 'perf test' to run under memory sanitizer. Additional notes from Ian on running in this mode: Note, as msan needs to instrument memory allocations libraries need to be compiled with it. I lacked the msan built libraries and so built with: ``` $ make -C tools/perf O=/tmp/perf DEBUG=1 EXTRA_CFLAGS="-O0 -g -fno-omit-frame-pointer -fsanitize=memory -fsanitize-memory-track-origins" CC=clang CXX=clang++ HOSTCC=clang NO_LIBTRACEEVENT=1 NO_LIBELF=1 BUILD_BPF_SKEL=0 NO_LIBPFM=1 ``` oh, I disabled libbpf here as the bpf system call also lacks msan interceptors. Signed-off-by:
-
Adrian Hunter authored
MemorySanitizer discovered instances where the instruction op value was not assigned.: WARNING: MemorySanitizer: use-of-uninitialized-value #0 0x5581c00a76b3 in intel_pt_sample_flags tools/perf/util/intel-pt.c:1527:17 Uninitialized value was stored to memory at #0 0x5581c005ddf8 in intel_pt_walk_insn tools/perf/util/intel-pt-decoder/intel-pt-decoder.c:1256:25 The op value is used to set branch flags for branch instructions encountered when walking the code, so fix by setting op to INTEL_PT_OP_OTHER in other cases. Fixes: 4c761d80 ("perf intel-pt: Fix intel_pt_fup_event() assumptions about setting state type") Reported-by: -
Howard Chu authored
Fix comment misspellings Signed-off-by:
Howard Chu <howardchu95@gmail.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: Kan Liang <kan.liang@linux.intel.com> 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/20240425060427.1800663-1-howardchu95@gmail.comSigned-off-by:
-
Namhyung Kim authored
dso__disassemble_filename() tries to get the filename for objdump (or capstone) using build-id. But I found sometimes it didn't disassemble some functions. It turned out that those functions belong to a DSO which has no binary type set. It seems it sets the binary type for some special files only - like kernel (kallsyms or kcore) or BPF images. And there's a logic to skip dso with DSO_BINARY_TYPE__NOT_FOUND. As it's checked the build-id cache link, it should set the binary type as DSO_BINARY_TYPE__BUILD_ID_CACHE. Fixes: 873a8373 ("perf annotate: Skip DSOs not found") Signed-off-by:
-
Namhyung Kim authored
I found some cases that capstone failed to disassemble. Probably my capstone is an old version but anyway there's a chance it can fail. And then it silently stopped in the middle. In my case, it didn't understand "RDPKRU" instruction. Let's check if the capstone disassemble reached the end of the function and fallback to objdump if not. Signed-off-by:
-
Namhyung Kim authored
As it removed the sample accounting for code when no symbol sort key is given for 'perf report' TUI, it might not have allocated the 'struct annotated_source' yet. Let's check if it's NULL first. Fixes: 6cdd977e ("perf report: Do not collect sample histogram unnecessarily") Reviewed-by:
-
Namhyung Kim authored
Add a basic 'perf annotate' test: $ ./perf test annotate -vv 76: perf annotate basic tests: --- start --- test child forked, pid 846989 fbcd0-fbd55 l noploop perf does have symbol 'noploop' Basic perf annotate test : 0 0xfbcd0 <noploop>: 0.00 : fbcd0: pushq %rbp 0.00 : fbcd1: movq %rsp, %rbp 0.00 : fbcd4: pushq %r12 0.00 : fbcd6: pushq %rbx 0.00 : fbcd7: movl $1, %ebx 0.00 : fbcdc: subq $0x10, %rsp 0.00 : fbce0: movq %fs:0x28, %rax 0.00 : fbce9: movq %rax, -0x18(%rbp) 0.00 : fbced: xorl %eax, %eax 0.00 : fbcef: testl %edi, %edi 0.00 : fbcf1: jle 0xfbd04 0.00 : fbcf3: movq (%rsi), %rdi 0.00 : fbcf6: movl $0xa, %edx 0.00 : fbcfb: xorl %esi, %esi 0.00 : fbcfd: callq 0x41920 0.00 : fbd02: movl %eax, %ebx 0.00 : fbd04: leaq -0x7b(%rip), %r12 # fbc90 <sighandler> 0.00 : fbd0b: movl $2, %edi 0.00 : fbd10: movq %r12, %rsi 0.00 : fbd13: callq 0x40a00 0.00 : fbd18: movl $0xe, %edi 0.00 : fbd1d: movq %r12, %rsi 0.00 : fbd20: callq 0x40a00 0.00 : fbd25: movl %ebx, %edi 0.00 : fbd27: callq 0x407c0 0.10 : fbd2c: movl 0x89785e(%rip), %eax # 993590 <done> 0.00 : fbd32: testl %eax, %eax 99.90 : fbd34: je 0xfbd2c 0.00 : fbd36: movq -0x18(%rbp), %rax 0.00 : fbd3a: subq %fs:0x28, %rax 0.00 : fbd43: jne 0xfbd50 0.00 : fbd45: addq $0x10, %rsp 0.00 : fbd49: xorl %eax, %eax 0.00 : fbd4b: popq %rbx 0.00 : fbd4c: popq %r12 0.00 : fbd4e: popq %rbp 0.00 : fbd4f: retq 0.00 : fbd50: callq 0x407e0 0.00 : fbcd0: pushq %rbp 0.00 : fbcd1: movq %rsp, %rbp 0.00 : fbcd4: pushq %r12 0.00 : fbcd0: push %rbp 0.00 : fbcd1: mov %rsp,%rbp 0.00 : fbcd4: push %r12 Basic annotate test [Success] ---- end(0) ---- 76: perf annotate basic tests : Ok Reviewed-by: -
Ian Rogers authored
Add comments. Pass ownership of the event name to save on a strdup. Signed-off-by:
Atish Patra <atishp@rivosinc.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Beeman Strong <beeman@rivosinc.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: James Clark <james.clark@arm.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/20240416061533.921723-17-irogers@google.comSigned-off-by:
-
Ian Rogers authored
Add comments. Ensure leader->group_name is freed before overwriting it. Signed-off-by:
-
Ian Rogers authored
Moves 352 bytes from .data to .data.rel.ro. Signed-off-by:
-
Ian Rogers authored
Use a struct/bitmap rather than a copied string from lexer. In lexer give improved error message when too many precise flags are given or repeated modifiers. Before: $ perf stat -e 'cycles:kuk' true event syntax error: 'cycles:kuk' \___ Bad modifier ... $ perf stat -e 'cycles:pppp' true event syntax error: 'cycles:pppp' \___ Bad modifier ... $ perf stat -e '{instructions:p,cycles:pp}:pp' -a true event syntax error: '..cycles:pp}:pp' \___ Bad modifier ... After: $ perf stat -e 'cycles:kuk' true event syntax error: 'cycles:kuk' \___ Duplicate modifier 'k' (kernel) ... $ perf stat -e 'cycles:pppp' true event syntax error: 'cycles:pppp' \___ Maximum precise value is 3 ... $ perf stat -e '{instructions:p,cycles:pp}:pp' true event syntax error: '..cycles:pp}:pp' \___ Maximum combined precise value is 3, adding precision to "cycles:pp" ... Signed-off-by: -
Ian Rogers authored
Inline parse_events_evlist_error that is only used in parse_events_error. Modify parse_events_error to not report a parser error unless errors haven't already been reported. Make it clearer that the latter case only happens for unrecognized input. Before: $ perf stat -e 'cycles/period=99999999999999999999/' true event syntax error: 'cycles/period=99999999999999999999/' \___ parser error event syntax error: '..les/period=99999999999999999999/' \___ Bad base 10 number "99999999999999999999" Run 'perf list' for a list of valid events Usage: perf stat [<options>] [<command>] -e, --event <event> event selector. use 'perf list' to list available events $ perf stat -e 'cycles:xyz' true event syntax error: 'cycles:xyz' \___ parser error Run 'perf list' for a list of valid events Usage: perf stat [<options>] [<command>] -e, --event <event> event selector. use 'perf list' to list available events After: $ perf stat -e 'cycles/period=99999999999999999999/xyz' true event syntax error: '..les/period=99999999999999999999/xyz' \___ Bad base 10 number "99999999999999999999" Run 'perf list' for a list of valid events Usage: perf stat [<options>] [<command>] -e, --event <event> event selector. use 'perf list' to list available events $ perf stat -e 'cycles:xyz' true event syntax error: 'cycles:xyz' \___ Unrecognized input Run 'perf list' for a list of valid events Usage: perf stat [<options>] [<command>] -e, --event <event> event selector. use 'perf list' to list available events Signed-off-by: -
Ian Rogers authored
Use the error handler from the parse_state to give a more informative error message. Before: $ perf stat -e 'cycles/period=99999999999999999999/' true event syntax error: 'cycles/period=99999999999999999999/' \___ parser error Run 'perf list' for a list of valid events Usage: perf stat [<options>] [<command>] -e, --event <event> event selector. use 'perf list' to list available events After: $ perf stat -e 'cycles/period=99999999999999999999/' true event syntax error: 'cycles/period=99999999999999999999/' \___ parser error event syntax error: '..les/period=99999999999999999999/' \___ Bad base 10 number "99999999999999999999" Run 'perf list' for a list of valid events Usage: perf stat [<options>] [<command>] -e, --event <event> event selector. use 'perf list' to list available events Signed-off-by: -
Ian Rogers authored
The helper function just wraps a splice and free. Making the free inline removes a comment, so then it just wraps a splice which we can make inline too. Signed-off-by:
-
Ian Rogers authored
It was requested that RISC-V be able to add events to the perf tool so the PMU driver didn't need to map legacy events to config encodings: https://lore.kernel.org/lkml/20240217005738.3744121-1-atishp@rivosinc.com/ This change makes the priority of events specified without a PMU the same as those specified with a PMU, namely sysfs and JSON events are checked first before using the legacy encoding. The hw_term is made more generic as a hardware_event that encodes a pair of string and int value, allowing parse_events_multi_pmu_add to fall back on a known encoding when the sysfs/JSON adding fails for core events. As this covers PE_VALUE_SYM_HW, that token is removed and related code simplified. Signed-off-by:
-
Ian Rogers authored
Allow the term list to be const so that other functions can pass const term lists. Add const as necessary to called functions. Signed-off-by:
-
Ian Rogers authored
Avoid duplicate logic for name_or_raw and PE_TERM_HW by having a rule to turn PE_TERM_HW into a name_or_raw. Signed-off-by:
-
Ian Rogers authored
Prior behavior is to not look for legacy cache names in sysfs/JSON and to create events on all core PMUs. New behavior is to look for sysfs/JSON events first on all PMUs, for core PMUs add a legacy event if the sysfs/JSON event isn't present. This is done so that there is consistency with how event names in terms are handled and their prioritization of sysfs/JSON over legacy. It may make sense to use a legacy cache event name as an event name on a non-core PMU so we should allow it. Signed-off-by:
-
Ian Rogers authored
Switch from "cache-references" to "branches" in test as Intel has a sysfs event for "cache-references" and changing the priority for sysfs over legacy causes the test to fail. Signed-off-by:
-
Ian Rogers authored
Move all implementation to pmu code. Don't allocate a fnmatch wildcard pattern, matching ignoring the suffix already handles this, and only use fnmatch if the given PMU name has a '*' in it. Signed-off-by:
-
Ian Rogers authored
In parse_events_add_pmu, delay copying the list of terms until it is known the list contains terms. Signed-off-by:
-
Ian Rogers authored
Avoid passing the name of a PMU then finding it again, just directly pass the PMU. parse_events_multi_pmu_add_or_add_pmu() is the only version that needs to find a PMU, so move the find there. Remove the error message as parse_events_multi_pmu_add_or_add_pmu will given an error at the end when a name isn't either a PMU name or event name. Without the error message being created the location in the input parameter (loc) can be removed. Signed-off-by:
-
Ian Rogers authored
Factor out the case of an event or PMU name followed by a slash based term list. This is with a view to sharing the code with new legacy hardware parsing. Use early return to reduce indentation in the code. Make parse_events_add_pmu static now it doesn't need sharing with parse-events.y. Signed-off-by:
-
Adrian Hunter authored
Add a Python script to run a perf script command multiple times in parallel, using perf script options --cpu and --time so that each job processes a different chunk of the data. Extend perf script tests to test also the new script. The script supports the use of normal 'perf script' options like --dlfilter and --script, so that the benefit of running parallel jobs naturally extends to them also. In addition, a command can be provided (refer --pipe-to option) to pipe standard output to a custom command. Refer to the script's own help text at the end of the patch for more details. The script is useful for Intel PT traces, that can be efficiently decoded by 'perf script' when split by CPU and/or time ranges. Running jobs in parallel can decrease the overall decoding time. Committer testing: Ian reported that shellcheck found some issues, I installed it as there are no warnings about it not being available, but when available it fails the build with: TEST /tmp/build/perf-tools-next/tests/shell/script.sh.shellcheck_log CC /tmp/build/perf-tools-next/util/header.o In tests/shell/script.sh line 20: rm -rf "${temp_dir}/"* ^-------------^ SC2115 (warning): Use "${var:?}" to ensure this never expands to /* . In tests/shell/script.sh line 83: output1_dir="${temp_dir}/output1" ^---------^ SC2034 (warning): output1_dir appears unused. Verify use (or export if used externally). In tests/shell/script.sh line 84: output2_dir="${temp_dir}/output2" ^---------^ SC2034 (warning): output2_dir appears unused. Verify use (or export if used externally). In tests/shell/script.sh line 86: python3 "${pp}" -o "${output_dir}" --jobs 4 --verbose -- perf script -i "${perf_data}" ^-----------^ SC2154 (warning): output_dir is referenced but not assigned (did you mean 'output1_dir'?). For more information: https://www.shellcheck.net/wiki/SC2034 -- output1_dir appears unused. Verif... https://www.shellcheck.net/wiki/SC2115 -- Use "${var:?}" to ensure this nev... https://www.shellcheck.net/wiki/SC2154 -- output_dir is referenced but not ... Did these fixes: - rm -rf "${temp_dir}/"* + rm -rf "${temp_dir:?}/"* And: @@ -83,8 +83,8 @@ test_parallel_perf() output1_dir="${temp_dir}/output1" output2_dir="${temp_dir}/output2" perf record -o "${perf_data}" --sample-cpu uname - python3 "${pp}" -o "${output_dir}" --jobs 4 --verbose -- perf script -i "${perf_data}" - python3 "${pp}" -o "${output_dir}" --jobs 4 --verbose --per-cpu -- perf script -i "${perf_data}" + python3 "${pp}" -o "${output1_dir}" --jobs 4 --verbose -- perf script -i "${perf_data}" + python3 "${pp}" -o "${output2_dir}" --jobs 4 --verbose --per-cpu -- perf script -i "${perf_data}" After that: root@number:~# perf test -vv "perf script tests" 97: perf script tests: --- start --- test child forked, pid 4084139 DB test [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.032 MB /tmp/perf-test-script.T4MJDr0L6J/perf.data (7 samples) ] <SNIP> DB test [Success] parallel-perf test Linux [ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.034 MB /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data (7 samples) ] Starting: perf script --time=,91898.301878499 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --time=91898.301878500,91898.301905999 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --time=91898.301906000,91898.301933499 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --time=91898.301933500, -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --time=91898.301878500,91898.301905999 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --time=91898.301906000,91898.301933499 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 4 jobs: 2 completed, 2 running Finished: perf script --time=,91898.301878499 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --time=91898.301933500, -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 4 jobs: 4 completed, 0 running All jobs finished successfully parallel-perf.py done Starting: perf script --cpu=0 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=1 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=2 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=3 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=0 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=1 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=2 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=3 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 4 completed, 0 running Starting: perf script --cpu=4 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=5 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=6 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=7 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=4 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=5 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=6 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=7 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 8 completed, 0 running Starting: perf script --cpu=8 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=9 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=10 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=11 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=8 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=9 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=10 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=11 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 12 completed, 0 running Starting: perf script --cpu=12 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=13 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=14 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=15 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=12 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=13 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=14 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=15 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 16 completed, 0 running Starting: perf script --cpu=16 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=17 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=18 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=19 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=16 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=17 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=18 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=19 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 20 completed, 0 running Starting: perf script --cpu=20 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=21 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=22 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=23 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=20 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=21 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=22 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=23 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 24 completed, 0 running Starting: perf script --cpu=24 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=25 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=26 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Starting: perf script --cpu=27 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=25 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=26 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data Finished: perf script --cpu=27 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 27 completed, 1 running Finished: perf script --cpu=24 -i /tmp/perf-test-script.T4MJDr0L6J/pp-perf.data There are 28 jobs: 28 completed, 0 running All jobs finished successfully parallel-perf.py done parallel-perf test [Success] --- Cleaning up --- ---- end(0) ---- 97: perf script tests : Ok root@number:~# Reviewed-by:Andi Kleen <ak@linux.intel.com> Signed-off-by:
-
