- 06 Jul, 2020 8 commits
-
-
Steven Rostedt (VMware) authored
When the offset option is set for the function plugin enabled, it will display the offset of the functions along with their names. This helps in finding exactly where a function was called by its parent. trace-cmd report -O parent -O offset [..] rcuc/163-1330 [163] 740.653251: function: _raw_spin_lock+0x0 <-- rcu_cpu_kthread+0x4d8 Link: http://lore.kernel.org/linux-trace-devel/20200702174950.123454-2-tz.stoyanov@gmail.comSigned-off-by:Steven Rostedt (VMware) <rostedt@goodmis.org> [ Ported from trace-cmd.git ] Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: linux-trace-devel@vger.kernel.org Link: http://lore.kernel.org/lkml/20200702185704.986181512@goodmis.orgSigned-off-by:
Tzvetomir Stoyanov (VMware) <tz.stoyanov@gmail.com> Signed-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Jan Kiszka authored
Exceptions require individual decoding (only feasible intercepts listed), XSETBV was missing and the AVIC brought in two new exit codes. Link: http://lkml.kernel.org/r/5741D822.3030203@web.de Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-10-tz.stoyanov@gmail.comSigned-off-by:
Jan Kiszka <jan.kiszka@siemens.com> [ Ported from trace-cmd.git ] Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: linux-trace-devel@vger.kernel.org Link: http://lore.kernel.org/lkml/20200702185704.844582602@goodmis.orgSigned-off-by:
-
Steven Rostedt (Red Hat) authored
The tlb_flush tracepoints uses enums that are not yet known by the traceevent library. Add a plugin to handle that. Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-9-tz.stoyanov@gmail.comSigned-off-by:
-
Tzvetomir Stoyanov (VMware) authored
Each time the pretty_print() function is called to print an event, the event's format string is parsed. As this format string does not change, this parsing can be done only once - when the event struct is initialized. Link: https://lore.kernel.org/linux-trace-devel/20200529134929.537110-1-tz.stoyanov@gmail.com Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-8-tz.stoyanov@gmail.comSigned-off-by:
-
Tzvetomir Stoyanov (VMware) authored
The printk format specifiers used in event's print format files extend the standard printf formats. There are a lot of new options related to printing pointers and kernel specific structures. Currently trace-cmd does not support many of them. Support for these new printk specifiers is added to the pretty_print() function: - UUID/GUID address: %pU[bBlL] - Raw buffer as a hex string: %*ph[CDN] These are improved: - MAC address: %pMF, %pM and %pmR - IPv4 adderss: %p[Ii]4[hnbl] Function pretty_print() is refactored. The logic for printing pointers %p[...] is moved to its own function. Link: https://lore.kernel.org/linux-trace-devel/20200515053754.3695335-1-tz.stoyanov@gmail.com Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-7-tz.stoyanov@gmail.com Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=207605Reported-by:
Johannes Berg <johannes@sipsolutions.net> Signed-off-by:
-
Tzvetomir Stoyanov (VMware) authored
Implement new traceevent plugin API, which can be used to add new plugins directories: enum tep_plugin_load_priority { TEP_PLUGIN_FIRST, TEP_PLUGIN_LAST, }; int tep_add_plugin_path(struct tep_handle *tep, char *path, enum tep_plugin_load_priority prio); It adds the "path" as new plugin directory, in the context of the handler "tep". The tep_load_plugins() API searches for plugins in this new location. Depending of the priority "prio", the plugins from this directory are loaded before (TEP_PLUGIN_FIRST) or after (TEP_PLUGIN_LAST) the ordinary libtraceevent plugin locations. Link: http://lore.kernel.org/linux-trace-devel/20191007114947.17104-2-tz.stoyanov@gmail.com Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-6-tz.stoyanov@gmail.comSigned-off-by: -
Tzvetomir Stoyanov (VMware) authored
Add tep_plugin_add_option() and tep_plugin_print_options() to lib traceevent library that allows plugins to have their own options. For example, the function plugin by default does not print the parent, as it uses the parent to do the indenting. The "parent" option is created by the function plugin that will print the parent of the function like it does in the trace file. The tep_plugin_print_options() will print out the list of options that a Cc: Tzvetomir Stoyanov (VMware) <tz.stoyanov@gmail.com> plugin has defined. Link: http://lore.kernel.org/linux-trace-devel/20190802110101.14759-3-tz.stoyanov@gmail.com Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-5-tz.stoyanov@gmail.comSigned-off-by:
-
Tzvetomir Stoyanov (VMware) authored
Add the API function tep_load_plugins_hook() to the traceevent API to allow tools a common method to load in the plugins that are part of the lib traceevent library. Link: http://lore.kernel.org/linux-trace-devel/20190802110101.14759-4-tz.stoyanov@gmail.com Link: http://lore.kernel.org/linux-trace-devel/20200625100516.365338-4-tz.stoyanov@gmail.comSigned-off-by:
-
- 02 Jul, 2020 2 commits
-
-
Ian Rogers authored
Rather than disable all warnings with -w, disable specific warnings. Predicate enabling the warnings on a recent version of bison. Tested with GCC 9.3.0 and clang 9.0.1. Committer testing: The full set of compilers, gcc and clang that this will be tested on will be on the signed tag when this change goes upstream. Had to add -Wno-switch-enum to build on opensuse tumbleweed: /tmp/build/perf/util/parse-events-bison.c: In function 'yydestruct': /tmp/build/perf/util/parse-events-bison.c:1200:3: error: enumeration value 'YYSYMBOL_YYEMPTY' not handled in switch [-Werror=switch-enum] 1200 | switch (yykind) | ^~~~~~ /tmp/build/perf/util/parse-events-bison.c:1200:3: error: enumeration value 'YYSYMBOL_YYEOF' not handled in switch [-Werror=switch-enum] Also replace -Wno-error=implicit-function-declaration with -Wno-implicit-function-declaration. Also needed to check just the first two levels of the bison version, as the patch was assuming that all versions were of the form x.y.z, and there are several cases where it is just x.y, breaking the build. Signed-off-by:Ian Rogers <irogers@google.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Jin Yao <yao.jin@linux.intel.com> Cc: John Garry <john.garry@huawei.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lore.kernel.org/lkml/20200619043356.90024-11-irogers@google.comSigned-off-by:
-
Ian Rogers authored
Rather than disable all warnings with -w, disable specific warnings. Predicate enabling the warnings on more recent flex versions. Tested with GCC 9.3.0 and clang 9.0.1. Committer notes: The full set of compilers, gcc and clang that this will be tested on will be on the signed tag when this change goes upstream. Added -Wno-misleading-indentation to the flex_flags to overcome this on opensuse tumbleweed when building with clang: CC /tmp/build/perf/util/parse-events-flex.o CC /tmp/build/perf/util/pmu.o /tmp/build/perf/util/parse-events-flex.c:5038:13: error: misleading indentation; statement is not part of the previous 'if' [-Werror,-Wmisleading-indentation] if ( ! yyg->yy_state_buf ) ^ /tmp/build/perf/util/parse-events-flex.c:5036:9: note: previous statement is here if ( ! yyg->yy_state_buf ) ^ And we need to use this to redirect stderr to stdin and then grep in a way that is acceptable for BusyBox shell: 2>&1 | Previously I was using: |& Which seems to be bash specific. Added -Wno-sign-compare to overcome this on systems such as centos:7: CC /tmp/build/perf/util/parse-events-flex.o CC /tmp/build/perf/util/pmu.o CC /tmp/build/perf/util/pmu-flex.o util/parse-events.l: In function 'parse_events_lex': /tmp/build/perf/util/parse-events-flex.c:193:36: error: comparison between signed and unsigned integer expressions [-Werror=sign-compare] for ( yyl = n; yyl < yyleng; ++yyl )\ ^ /tmp/build/perf/util/parse-events-flex.c:204:9: note: in expansion of macro 'YY_LESS_LINENO' Added -Wno-unused-parameter to overcome this in systems such as centos:7: CC /tmp/build/perf/util/parse-events-flex.o CC /tmp/build/perf/util/pmu.o /tmp/build/perf/util/parse-events-flex.c: In function 'yy_fatal_error': /tmp/build/perf/util/parse-events-flex.c:6265:58: error: unused parameter 'yyscanner' [-Werror=unused-parameter] static void yy_fatal_error (yyconst char* msg , yyscan_t yyscanner) ^ Added -Wno-missing-declarations to build in systems such as centos:6: /tmp/build/perf/util/parse-events-flex.c:6313: error: no previous prototype for 'parse_events_get_column' /tmp/build/perf/util/parse-events-flex.c:6389: error: no previous prototype for 'parse_events_set_column' And -Wno-missing-prototypes to cover older compilers: -Wmissing-prototypes (C only) Warn if a global function is defined without a previous prototype declaration. This warning is issued even if the definition itself provides a prototype. The aim is to detect global functions that fail to be declared in header files. -Wmissing-declarations (C only) Warn if a global function is defined without a previous declaration. Do so even if the definition itself provides a prototype. Use this option to detect global functions that are not declared in header files. Older C compilers lack -Wno-misleading-indentation, check if it is available before using it. Also needed to check just the first two levels of the flex version, as the patch was assuming that all versions were of the form x.y.z, and there are several cases where it is just x.y, breaking the build. Signed-off-by:Jiri Olsa <jolsa@redhat.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Jin Yao <yao.jin@linux.intel.com> Cc: John Garry <john.garry@huawei.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: http://lore.kernel.org/lkml/20200619043356.90024-8-irogers@google.comSigned-off-by:
-
- 23 Jun, 2020 2 commits
-
-
Ian Rogers authored
Declare bison header file output so that C files can depend upon them. As there are multiple output targets $@ is replaced by the target name. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
These will break the build as soon as we stop disabling all warnings when building flex and bison generated files, so add them before we do that to keep the tree bisectable. Noticed when building on centos:7 with NO_LIBBPF=1: util/expr.c: In function 'key_equal': util/expr.c:29:2: error: implicit declaration of function 'strcmp' [-Werror=implicit-function-declaration] return !strcmp((const char *)key1, (const char *)key2); ^ util/expr.c: In function 'expr__add_id': util/expr.c:40:3: error: implicit declaration of function 'malloc' [-Werror=implicit-function-declaration] val_ptr = malloc(sizeof(double)); ^ util/expr.c:40:13: error: incompatible implicit declaration of built-in function 'malloc' [-Werror] val_ptr = malloc(sizeof(double)); ^ util/expr.c:42:12: error: 'ENOMEM' undeclared (first use in this function) return -ENOMEM; ^ Signed-off-by:
-
- 22 Jun, 2020 28 commits
-
-
Ian Rogers authored
Declare flex header file output so that bison C files can depend upon them. As there are multiple output targets $@ is replaced by the target name. Signed-off-by:
-
Ian Rogers authored
Allow pmu parser's flex to be debugged as the parse-events and expr currently are. Enabling this requires the C code to call perf_pmu__flex_debug. Signed-off-by:
-
Ian Rogers authored
Allow pmu parser to be debugged as the parse-events and expr currently are. Enabling this requires the C code to set perf_pmu_debug. Signed-off-by:
-
Ian Rogers authored
This reduces the command line size slightly. Signed-off-by:
-
Ian Rogers authored
This reduces the command line size slightly. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
To differentiate from libperf's 'struct perf_evlist' methods. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by: -
Arnaldo Carvalho de Melo authored
To differentiate from libperf's 'struct perf_evlist' methods. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by: -
Arnaldo Carvalho de Melo authored
To differentiate from libperf's 'struct perf_evlist' methods. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by: -
Arnaldo Carvalho de Melo authored
To differentiate from libperf's 'struct perf_evlist' methods. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by: -
Arnaldo Carvalho de Melo authored
To differentiate from libperf's 'struct perf_evlist' methods. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by: -
John Garry authored
For perf list, the CPU core PMU HW event ordering is such that not all events may will be listed adjacent - consider this example: $ tools/perf/perf list List of pre-defined events (to be used in -e): duration_time [Tool event] branch-instructions OR cpu/branch-instructions/ [Kernel PMU event] branch-misses OR cpu/branch-misses/ [Kernel PMU event] bus-cycles OR cpu/bus-cycles/ [Kernel PMU event] cache-misses OR cpu/cache-misses/ [Kernel PMU event] cache-references OR cpu/cache-references/ [Kernel PMU event] cpu-cycles OR cpu/cpu-cycles/ [Kernel PMU event] cstate_core/c3-residency/ [Kernel PMU event] cstate_core/c6-residency/ [Kernel PMU event] cstate_core/c7-residency/ [Kernel PMU event] cstate_pkg/c2-residency/ [Kernel PMU event] cstate_pkg/c3-residency/ [Kernel PMU event] cstate_pkg/c6-residency/ [Kernel PMU event] cstate_pkg/c7-residency/ [Kernel PMU event] cycles-ct OR cpu/cycles-ct/ [Kernel PMU event] cycles-t OR cpu/cycles-t/ [Kernel PMU event] el-abort OR cpu/el-abort/ [Kernel PMU event] el-capacity OR cpu/el-capacity/ [Kernel PMU event] Notice in the above example how the cstate_core PMU events are mixed in the middle of the CPU core events. For my arm64 platform, all the uncore events get mixed in, making the list very disorganised: page-faults OR faults [Software event] task-clock [Software event] duration_time [Tool event] L1-dcache-load-misses [Hardware cache event] L1-dcache-loads [Hardware cache event] L1-icache-load-misses [Hardware cache event] L1-icache-loads [Hardware cache event] branch-load-misses [Hardware cache event] branch-loads [Hardware cache event] dTLB-load-misses [Hardware cache event] dTLB-loads [Hardware cache event] iTLB-load-misses [Hardware cache event] iTLB-loads [Hardware cache event] br_mis_pred OR armv8_pmuv3_0/br_mis_pred/ [Kernel PMU event] br_mis_pred_retired OR armv8_pmuv3_0/br_mis_pred_retired/ [Kernel PMU event] br_pred OR armv8_pmuv3_0/br_pred/ [Kernel PMU event] br_retired OR armv8_pmuv3_0/br_retired/ [Kernel PMU event] br_return_retired OR armv8_pmuv3_0/br_return_retired/ [Kernel PMU event] bus_access OR armv8_pmuv3_0/bus_access/ [Kernel PMU event] bus_cycles OR armv8_pmuv3_0/bus_cycles/ [Kernel PMU event] cid_write_retired OR armv8_pmuv3_0/cid_write_retired/ [Kernel PMU event] cpu_cycles OR armv8_pmuv3_0/cpu_cycles/ [Kernel PMU event] dtlb_walk OR armv8_pmuv3_0/dtlb_walk/ [Kernel PMU event] exc_return OR armv8_pmuv3_0/exc_return/ [Kernel PMU event] exc_taken OR armv8_pmuv3_0/exc_taken/ [Kernel PMU event] hisi_sccl1_ddrc0/act_cmd/ [Kernel PMU event] hisi_sccl1_ddrc0/flux_rcmd/ [Kernel PMU event] hisi_sccl1_ddrc0/flux_rd/ [Kernel PMU event] hisi_sccl1_ddrc0/flux_wcmd/ [Kernel PMU event] hisi_sccl1_ddrc0/flux_wr/ [Kernel PMU event] hisi_sccl1_ddrc0/pre_cmd/ [Kernel PMU event] hisi_sccl1_ddrc0/rnk_chg/ [Kernel PMU event] ... hisi_sccl7_l3c21/wr_hit_cpipe/ [Kernel PMU event] hisi_sccl7_l3c21/wr_hit_spipe/ [Kernel PMU event] hisi_sccl7_l3c21/wr_spipe/ [Kernel PMU event] inst_retired OR armv8_pmuv3_0/inst_retired/ [Kernel PMU event] inst_spec OR armv8_pmuv3_0/inst_spec/ [Kernel PMU event] itlb_walk OR armv8_pmuv3_0/itlb_walk/ [Kernel PMU event] l1d_cache OR armv8_pmuv3_0/l1d_cache/ [Kernel PMU event] l1d_cache_refill OR armv8_pmuv3_0/l1d_cache_refill/ [Kernel PMU event] l1d_cache_wb OR armv8_pmuv3_0/l1d_cache_wb/ [Kernel PMU event] l1d_tlb OR armv8_pmuv3_0/l1d_tlb/ [Kernel PMU event] l1d_tlb_refill OR armv8_pmuv3_0/l1d_tlb_refill/ [Kernel PMU event] So the events are list alphabetically. However, CPU core event listing is special from commit dc098b35 ("perf list: List kernel supplied event aliases"), in that the alias and full event is shown (in that order). As such, the core events may become sparse. Improve this by grouping the CPU core events and ensure that they are listed first for kernel PMU events. For the first example, above, this now looks like: duration_time [Tool event] branch-instructions OR cpu/branch-instructions/ [Kernel PMU event] branch-misses OR cpu/branch-misses/ [Kernel PMU event] bus-cycles OR cpu/bus-cycles/ [Kernel PMU event] cache-misses OR cpu/cache-misses/ [Kernel PMU event] cache-references OR cpu/cache-references/ [Kernel PMU event] cpu-cycles OR cpu/cpu-cycles/ [Kernel PMU event] cycles-ct OR cpu/cycles-ct/ [Kernel PMU event] cycles-t OR cpu/cycles-t/ [Kernel PMU event] el-abort OR cpu/el-abort/ [Kernel PMU event] el-capacity OR cpu/el-capacity/ [Kernel PMU event] el-commit OR cpu/el-commit/ [Kernel PMU event] el-conflict OR cpu/el-conflict/ [Kernel PMU event] el-start OR cpu/el-start/ [Kernel PMU event] instructions OR cpu/instructions/ [Kernel PMU event] mem-loads OR cpu/mem-loads/ [Kernel PMU event] mem-stores OR cpu/mem-stores/ [Kernel PMU event] ref-cycles OR cpu/ref-cycles/ [Kernel PMU event] topdown-fetch-bubbles OR cpu/topdown-fetch-bubbles/ [Kernel PMU event] topdown-recovery-bubbles OR cpu/topdown-recovery-bubbles/ [Kernel PMU event] topdown-slots-issued OR cpu/topdown-slots-issued/ [Kernel PMU event] topdown-slots-retired OR cpu/topdown-slots-retired/ [Kernel PMU event] topdown-total-slots OR cpu/topdown-total-slots/ [Kernel PMU event] tx-abort OR cpu/tx-abort/ [Kernel PMU event] tx-capacity OR cpu/tx-capacity/ [Kernel PMU event] tx-commit OR cpu/tx-commit/ [Kernel PMU event] tx-conflict OR cpu/tx-conflict/ [Kernel PMU event] tx-start OR cpu/tx-start/ [Kernel PMU event] cstate_core/c3-residency/ [Kernel PMU event] cstate_core/c6-residency/ [Kernel PMU event] cstate_core/c7-residency/ [Kernel PMU event] cstate_pkg/c2-residency/ [Kernel PMU event] cstate_pkg/c3-residency/ [Kernel PMU event] cstate_pkg/c6-residency/ [Kernel PMU event] cstate_pkg/c7-residency/ [Kernel PMU event] Signed-off-by:John Garry <john.garry@huawei.com> Acked-by:
Namhyung Kim <namhyung@kernel.org> Tested-by:
-
John Garry authored
In commit dc098b35 ("perf list: List kernel supplied event aliases"), the aliases for events are supplied in addition to CPU event in perf list. This relies on the name of the core PMU being "cpu", which is not the case for arm64, so arm64 has always missed this. Use generic is_pmu_core() helper which takes account of arm64 to make this feature work for arm64 (and possibly other archs). Sample, before: armv8_pmuv3_0/br_mis_pred/ [Kernel PMU event] after: br_mis_pred OR armv8_pmuv3_0/br_mis_pred/ [Kernel PMU event] Signed-off-by:
-
Mike Leach authored
Adjust the handling of the session sink selection to allow no sink to be selected on the command line. This then forwards the sink selection to the CoreSight infrastructure which will attempt to select a sink based on the default sink select priorities. Signed-off-by:
Mike Leach <mike.leach@linaro.org> Tested-by:
Leo Yan <leo.yan@linaro.org> Cc: Mathieu Poirier <mathieu.poirier@linaro.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Suzuki Poulouse <suzuki.poulose@arm.com> Cc: coresight@lists.linaro.org Cc: linux-arm-kernel@lists.infradead.org Signed-off-by:
-
Ian Rogers authored
These are broadly useful but required to handle TMA metrics. For example encoding Ports_Utilization from: https://download.01.org/perfmon/TMA_Metrics.csv requires '<'. { "BriefDescription": "This metric estimates fraction of cycles the CPU performance was potentially limited due to Core computation issues (non divider-related). Two distinct categories can be attributed into this metric: (1) heavy data-dependency among contiguous instructions would manifest in this metric - such cases are often referred to as low Instruction Level Parallelism (ILP). (2) Contention on some hardware execution unit other than Divider. For example; when there are too many multiply operations.", "MetricExpr": "( ( cpu@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ + cpu@EXE_ACTIVITY.1_PORTS_UTIL@ + ( cpu@EXE_ACTIVITY.2_PORTS_UTIL@ * ( ( ( cpu@UOPS_RETIRED.RETIRE_SLOTS@ ) / ( cpu@CPU_CLK_UNHALTED.THREAD@ ) ) / ( ( 4.000000 ) + 1.000000 ) ) ) ) / ( cpu@CPU_CLK_UNHALTED.THREAD@ ) if ( cpu@ARITH.DIVIDER_ACTIVE\\,cmask\\=1@ < cpu@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ ) else ( ( cpu@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ + cpu@EXE_ACTIVITY.1_PORTS_UTIL@ + ( cpu@EXE_ACTIVITY.2_PORTS_UTIL@ * ( ( ( cpu@UOPS_RETIRED.RETIRE_SLOTS@ ) / ( cpu@CPU_CLK_UNHALTED.THREAD@ ) ) / ( ( 4.000000 ) + 1.000000 ) ) ) ) - cpu@EXE_ACTIVITY.EXE_BOUND_0_PORTS@ ) / ( cpu@CPU_CLK_UNHALTED.THREAD@ ) )", "MetricGroup": "Topdown_Group_Ports_Utilization", "MetricName": "Topdown_Metric_Ports_Utilization" }, Signed-off-by:
-
Ian Rogers authored
d_ratio avoids division by 0 yielding infinity, such as when a counter doesn't get scheduled. An example usage is: { "BriefDescription": "DCache L1 misses", "MetricExpr": "d_ratio(MEM_LOAD_RETIRED.L1_MISS, MEM_LOAD_RETIRED.L1_HIT + MEM_LOAD_RETIRED.L1_MISS + MEM_LOAD_RETIRED.FB_HIT)", "MetricGroup": "DCache;DCache_L1", "MetricName": "DCache_L1_Miss", "ScaleUnit": "100%", } Signed-off-by: -
Arnaldo Carvalho de Melo authored
Fixups related to the introduction of libperf, where the perf_{evsel,evlist}__ prefix is reserved for functions operating on struct perf_{evsel,evlist}. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by: -
Jiri Olsa authored
Adding new metric test for frontend metric. It's stolen from x86 pmu events. Committer testing: # perf test "Parse and process metrics" 67: Parse and process metrics : Ok # perf test -v "Parse and process metrics" # 67: Parse and process metrics : --- start --- test child forked, pid 104881 metric expr inst_retired.any / cpu_clk_unhalted.thread for IPC found event inst_retired.any found event cpu_clk_unhalted.thread adding {inst_retired.any,cpu_clk_unhalted.thread}:W metric expr idq_uops_not_delivered.core / (4 * (( ( cpu_clk_unhalted.thread / 2 ) * ( 1 + cpu_clk_unhalted.one_thread_active / cpu_clk_unhalted.ref_xclk ) ))) for Frontend_Bound_SMT found event cpu_clk_unhalted.one_thread_active found event cpu_clk_unhalted.ref_xclk found event idq_uops_not_delivered.core found event cpu_clk_unhalted.thread adding {cpu_clk_unhalted.one_thread_active,cpu_clk_unhalted.ref_xclk,idq_uops_not_delivered.core,cpu_clk_unhalted.thread}:W test child finished with 0 ---- end ---- Parse and process metrics: Ok # Had to fix it to initialize that 'struct value' array sentinel with a named initializer to fix the build with some versions of clang: tests/parse-metric.c:154:7: error: missing field 'val' initializer [-Werror,-Wmissing-field-initializers] { 0 }, Signed-off-by:Jiri Olsa <jolsa@kernel.org> Acked-by:
-
Jiri Olsa authored
Adding new test that process metrics code and checks the expected results. Starting with easy ipc metric. Committer testing: # perf test "Parse and process metrics" 67: Parse and process metrics : Ok # # perf test -v "Parse and process metrics" 67: Parse and process metrics : --- start --- test child forked, pid 103402 metric expr inst_retired.any / cpu_clk_unhalted.thread for IPC found event inst_retired.any found event cpu_clk_unhalted.thread adding {inst_retired.any,cpu_clk_unhalted.thread}:W test child finished with 0 ---- end ---- Parse and process metrics: Ok # Had to fix it to initialize that 'struct value' array sentinel with a named initializer to fix the build with some versions of clang: tests/parse-metric.c:135:7: error: missing field 'val' initializer [-Werror,-Wmissing-field-initializers] { 0 }, Signed-off-by: -
Jiri Olsa authored
Adding test_generic_metric that prepares and runs given metric over the data from struct runtime_stat object. Signed-off-by:
-
Jiri Olsa authored
We don't release metric_events rblist, add the missing delete hook and call the release before leaving cmd_stat. Signed-off-by:
-
Jiri Olsa authored
Factoring out prepare_metric function so it can be used in test interface coming in following changes. Signed-off-by:
-
Jiri Olsa authored
Add the metricgroup__parse_groups_test function. It will be used as test's interface to metric parsing in following changes. Signed-off-by:
-
Jiri Olsa authored
For testing purposes we need to pass our own map of events from parse_groups() through metricgroup__add_metric. Acked-by:
-
Jiri Olsa authored
Allow to pass fake_pmu in parse_groups function so it can be used in parse_events call. It's will be passed by the upcoming metricgroup__parse_groups_test function. Committer notes: Made it a 'struct perf_pmu' pointer, in line with the changes at the start of this patchkit to avoid statics deep down in library code. Acked-by:
-
Jiri Olsa authored
Factor out the parse_groups function, it will be used for new test interface coming in following changes. Signed-off-by:
-
Jiri Olsa authored
The test goes through all metrics compiled for arch within pmu events and try to parse them. This test is different from 'test_parsing' in that we go through all the events in the current arch, not just one defined for current CPU model. Using 'fake_pmu' to parse events which do not have PMUs defined in the system. Say there's bad change in ivybridge metrics file, like: - a/tools/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json + b/tools/perf/pmu-events/arch/x86/ivybridge/ivb-metrics.json @@ -8,7 +8,7 @@ - "MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / (4 * (( + "MetricExpr": "IDQ_UOPS_NOT_DELIVERED.CORE / / (4 * the test fails with (on my kabylake laptop): $ perf test 'Parsing of PMU event table metrics with fake PMUs' -v parsing 'idq_uops_not_delivered.core / / (4 * (( ( cpu_clk_unh... syntax error, line 1 expr__parse failed test child finished with -1 ... The test also defines its own list of metrics and tries to parse them. It's handy for developing. Committer notes: Testing it: $ perf test fake 10: PMU events : 10.4: Parsing of PMU event table metrics with fake PMUs : FAILED! $ perf test -v fake |& tail parsing '(unc_p_freq_trans_cycles / unc_p_clockticks) * 100.' parsing '(unc_m_power_channel_ppd / unc_m_clockticks) * 100.' parsing '(unc_m_power_critical_throttle_cycles / unc_m_clockticks) * 100.' parsing '(unc_m_power_self_refresh / unc_m_clockticks) * 100.' parsing 'idq_uops_not_delivered.core / * (4 * cycles)' syntax error expr__parse failed test child finished with -1 ---- end ---- PMU events subtest 4: FAILED! $ And fix this error: tests/pmu-events.c:437:40: error: missing field 'idx' initializer [-Werror,-Wmissing-field-initializers] struct parse_events_error error = { 0 }; Signed-off-by: -
Arnaldo Carvalho de Melo authored
When wanting to use the support in __parse_events() for fake pmus, just pass it. Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Michael Petlan <mpetlan@redhat.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Signed-off-by: -
Arnaldo Carvalho de Melo authored
This is an alternative patch to what Jiri sent that instead of changing all callers to parse_events() for allowing to pass a fake_pmu, provide another function specifically for that. From Jiri's patch: This way it's possible to parse events from PMUs which are not present in the system. It's available only for testing purposes coming in following changes, so all the current users set fake_pmu argument as false. Based-on-a-patch-by:
-
