- 03 Aug, 2023 3 commits
-
-
Ian Rogers authored
If bison is version 3.8.2, reduce the number of bison C warnings disabled. Earlier bison versions have all C warnings disabled. Avoid implicit declarations of yylex by adding the declaration in the C file. A header can't be included as a circular dependency would occur due to the lexer using the bison defined tokens. Committer notes: Some recent versions of gcc and clang (noticed on Alpine Linux 3.17, edge, clearlinux, fedora 37, etc. Signed-off-by:
Ian Rogers <irogers@google.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andrii Nakryiko <andrii@kernel.org> Cc: Eduard Zingerman <eddyz87@gmail.com> Cc: Gaosheng Cui <cuigaosheng1@huawei.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: Nathan Chancellor <nathan@kernel.org> Cc: Nick Desaulniers <ndesaulniers@google.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rob Herring <robh@kernel.org> Cc: Tom Rix <trix@redhat.com> Cc: bpf@vger.kernel.org Cc: llvm@lists.linux.dev Link: https://lore.kernel.org/r/20230728064917.767761-6-irogers@google.comSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Ian Rogers authored
If flex is version 2.6.4, reduce the number of flex C warnings disabled. Earlier flex versions have all C warnings disabled. Committer notes: Added this to the list of ignored warnings to get it building on a Fedora 36 machine with flex 2.6.4: -Wno-misleading-indentation Noticed when building with: $ make LLVM=1 -C tools/perf NO_BPF_SKEL=1 DEBUG=1 Take two: We can't just try to canonicalize flex versions by just removing the dots, as we end up with: 2.6.4 >= 2.5.37 becoming: 264 >= 2537 Failing the build on flex 2.5.37, so instead use the back to the past added $(call version_ge3,$(FLEX_VERSION),2.6.4) variant to check for that. Making sure $(FLEX_VERSION) keeps the dots as we may want to use 'sort -V' or something nicer when available everywhere. Some other tweaks for other flex versions and combinations with gcc and clang versions were added, notes on the patch. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
The next cset needs to compare if a flex version is greater or equal/less than another, but since there is no canonical, generally available way to compare versions in the command line (sort -V, yeah, but...), just use awk to canonicalize the versions like is also done in scripts/rust_is_available.sh. There was a problem spotted in linux-next where a bashism, here documents, aka the '<<<' stdin redirector, for strings to be used as the stdin for awk. Use $(shell echo | awk ...) instead. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by:
-
- 28 Jul, 2023 26 commits
-
-
Ian Rogers authored
Commit d58ac0bf ("perf build: Add clang and llvm compile and linking support") added -Wall and -Wno-strict-aliasing for CXXFLAGS, but not -Wextra. -Wno-strict-aliasing is no longer necessary, adding -Wextra for CXXFLAGS requires adding -Wno-unused-parameter clang.cpp and clang-test.cpp for LIBCLANGLLVM=1 to build. 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: Andrii Nakryiko <andrii@kernel.org> Cc: Eduard Zingerman <eddyz87@gmail.com> Cc: Gaosheng Cui <cuigaosheng1@huawei.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: Nathan Chancellor <nathan@kernel.org> Cc: Nick Desaulniers <ndesaulniers@google.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rob Herring <robh@kernel.org> Cc: Tom Rix <trix@redhat.com> Cc: bpf@vger.kernel.org Cc: llvm@lists.linux.dev Link: https://lore.kernel.org/r/20230728064917.767761-4-irogers@google.comSigned-off-by:
-
Ian Rogers authored
Commit 6a40cd90 ("perf tools: Add libunwind dependency for DWARF CFI unwinding") added libunwind support but also -funwind-tables and -ggdb3 to the standard build. These build flags aren't necessary so remove, set -g when DEBUG is enabled for the build. Signed-off-by:
-
Ian Rogers authored
Added during the progress to libbpf 1.0 the deprecated functions are no longer used and so the pragma can be removed. Signed-off-by:
-
Jing Zhang authored
Alibaba's T-Head ali_drw PMU supports DDR bandwidth metrics. Update its usage in the documentation. Reviewed-by:
John Garry <john.g.garry@oracle.com> Signed-off-by:
Jing Zhang <renyu.zj@linux.alibaba.com> Acked-by:
-
Jing Zhang authored
Add JSON metrics for T-HEAD Yitian 710 SoC DDR. Reviewed-by:
-
Jing Zhang authored
Add alias support for T-HEAD Yitian 710 SoC DDR PMU events. Reviewed-by:
Shuai Xue <xueshuai@linux.alibaba.com> Signed-off-by:
-
James Clark authored
Defconfig doesn't give full coverage for a perf test run, so these can be merged with defconfig to do so. It's not complete yet, but is a starting point as a place to add to when a specific test needs something extra to run. Signed-off-by:
-
Anup Sharma authored
This will enable the execution of gecko.py script using record and report commands in 'perf script'. And this will be also reflected at "perf script -l" command. For Example: perf script record gecko perf script report gecko Committer notes: As discussed on the perf tools office hours, I made -F 99 the default for the record script and removed the double -- on the report script so that the existing 'perf script' protocol for the combined operation: # perf script gecko Works, i.e. the record script pipes its stdout into the stdin of the report script, basically: /bin/sh /usr/libexec/perf-core/scripts/python/bin/gecko-record -F 99 -g -a -q -o - | \ /bin/sh /usr/libexec/perf-core/scripts/python/bin/gecko-report -i - Testing it: The resulting JSON file needs to be uploaded to https://profiler.firefox.com, Anup already has code to start a local http server on the trace_begin handler of the gecko python script, start firefox and feed it the JSON. The example below only collects sample for the specified workload, so that we don't produce thousands of lines, to collect system wide samples, use instead: # perf script gecko -a sleep 0.5 # nohup perf script gecko sleep 0.5 { "meta": { "interval": 1, "processType": 0, "product": "x86_64 GNU/Linux", "stackwalk": 1, "debug": 0, "gcpoison": 0, "asyncstack": 1, "startTime": 274601692.636, "shutdownTime": null, "version": 24, "presymbolicated": true, "categories": [ { "name": "User", "color": "yellow", "subcategories": [ "Other" ] }, { "name": "Kernel", "color": "orange", "subcategories": [ "Other" ] } ], "markerSchema": [] }, "libs": [], "threads": [ { "tid": 3344498, "pid": 3344498, "name": "sleep", "markers": { "schema": { "name": 0, "startTime": 1, "endTime": 2, "phase": 3, "category": 4, "data": 5 }, "data": [] }, "samples": { "schema": { "stack": 0, "time": 1, "responsiveness": 2 }, "data": [ [ 21, 274601692.636, 0 ], [ 23, 274601692.641, 0 ], [ 29, 274601692.643, 0 ], [ 42, 274601692.648, 0 ] ] }, "frameTable": { "schema": { "location": 0, "relevantForJS": 1, "innerWindowID": 2, "implementation": 3, "optimizations": 4, "line": 5, "column": 6, "category": 7, "subcategory": 8 }, "data": [ [ 0, false, 0, null, null, null, null, 1, null ], [ 1, false, 0, null, null, null, null, 1, null ], [ 2, false, 0, null, null, null, null, 1, null ], [ 3, false, 0, null, null, null, null, 1, null ], [ 4, false, 0, null, null, null, null, 1, null ], [ 5, false, 0, null, null, null, null, 1, null ], [ 6, false, 0, null, null, null, null, 1, null ], [ 7, false, 0, null, null, null, null, 1, null ], [ 8, false, 0, null, null, null, null, 1, null ], [ 9, false, 0, null, null, null, null, 1, null ], [ 10, false, 0, null, null, null, null, 1, null ], [ 11, false, 0, null, null, null, null, 1, null ], [ 12, false, 0, null, null, null, null, 1, null ], [ 13, false, 0, null, null, null, null, 1, null ], [ 14, false, 0, null, null, null, null, 1, null ], [ 15, false, 0, null, null, null, null, 1, null ], [ 16, false, 0, null, null, null, null, 1, null ], [ 17, false, 0, null, null, null, null, 1, null ], [ 18, false, 0, null, null, null, null, 1, null ], [ 19, false, 0, null, null, null, null, 1, null ], [ 20, false, 0, null, null, null, null, 1, null ], [ 21, false, 0, null, null, null, null, 1, null ], [ 22, false, 0, null, null, null, null, 1, null ], [ 23, false, 0, null, null, null, null, 1, null ], [ 24, false, 0, null, null, null, null, 1, null ], [ 25, false, 0, null, null, null, null, 1, null ], [ 26, false, 0, null, null, null, null, 1, null ], [ 27, false, 0, null, null, null, null, 1, null ], [ 28, false, 0, null, null, null, null, 1, null ], [ 29, false, 0, null, null, null, null, 1, null ], [ 30, false, 0, null, null, null, null, 1, null ], [ 31, false, 0, null, null, null, null, 1, null ], [ 32, false, 0, null, null, null, null, 1, null ], [ 33, false, 0, null, null, null, null, 1, null ], [ 34, false, 0, null, null, null, null, 1, null ], [ 35, false, 0, null, null, null, null, 1, null ], [ 36, false, 0, null, null, null, null, 1, null ], [ 37, false, 0, null, null, null, null, 1, null ], [ 38, false, 0, null, null, null, null, 1, null ] ] }, "stackTable": { "schema": { "prefix": 0, "frame": 1 }, "data": [ [ null, 0 ], [ 0, 1 ], [ 1, 2 ], [ 2, 3 ], [ 3, 4 ], [ 4, 5 ], [ 5, 6 ], [ 6, 7 ], [ 7, 8 ], [ 8, 9 ], [ 9, 10 ], [ 10, 11 ], [ 11, 12 ], [ 12, 13 ], [ 13, 14 ], [ 14, 15 ], [ 15, 16 ], [ 16, 17 ], [ 17, 18 ], [ 18, 19 ], [ 19, 20 ], [ 20, 21 ], [ 20, 22 ], [ 22, 23 ], [ 11, 24 ], [ 24, 25 ], [ 25, 26 ], [ 26, 27 ], [ 27, 28 ], [ 28, 29 ], [ 9, 11 ], [ 30, 24 ], [ 31, 25 ], [ 32, 30 ], [ 33, 31 ], [ 34, 32 ], [ 35, 29 ], [ 36, 33 ], [ 37, 34 ], [ 38, 35 ], [ 39, 36 ], [ 40, 37 ], [ 41, 38 ] ] }, "stringTable": [ "__func__.0 (in [kernel.kallsyms].rodata)", "perf_trace_ext4_fc_track_inode (in [kernel.kallsyms])", "perf_trace_ext4_es_insert_delayed_block (in [kernel.kallsyms])", "ext4_es_show_pblock (in [kernel.kallsyms])", "perf_trace_ext4_ext_rm_leaf (in [kernel.kallsyms])", "devcgroup_access_write (in [kernel.kallsyms])", "devcgroup_update_access (in [kernel.kallsyms])", "propagate_exception (in [kernel.kallsyms])", "revalidate_active_exceptions (in [kernel.kallsyms])", "perf_trace_ext4_fc_commit_stop (in [kernel.kallsyms])", "perf_fetch_caller_regs (in [kernel.kallsyms])", "khugepaged (in [kernel.kallsyms])", "khugepaged_wait_work (in [kernel.kallsyms])", "freezable_schedule_timeout (in [kernel.kallsyms])", "freezer_count (in [kernel.kallsyms])", "try_to_freeze (in [kernel.kallsyms])", "try_to_freeze_unsafe (in [kernel.kallsyms])", "split_huge_pages_write (in [kernel.kallsyms])", "migrate_pages (in [kernel.kallsyms])", "unmap_and_move (in [kernel.kallsyms])", "__unmap_and_move (in [kernel.kallsyms])", "collect_events (in [kernel.kallsyms])", "uncore_down_prepare (in [kernel.kallsyms])", "perf_iommu_read (in [kernel.kallsyms])", "khugepaged_do_scan (in [kernel.kallsyms])", "khugepaged_scan_mm_slot (in [kernel.kallsyms])", "khugepaged_scan_file (in [kernel.kallsyms])", "need_resched (in [kernel.kallsyms])", "get_current (in [kernel.kallsyms])", "move_to_new_page (in [kernel.kallsyms])", "khugepaged_scan_pmd (in [kernel.kallsyms])", "trace_mm_khugepaged_scan_pmd (in [kernel.kallsyms])", "migrate_huge_page_move_mapping (in [kernel.kallsyms])", "do_huge_pmd_numa_page (in [kernel.kallsyms])", "pmd_pfn (in [kernel.kallsyms])", "protnone_mask (in [kernel.kallsyms])", "__pte_needs_invert (in [kernel.kallsyms])", "reclaim_high (in [kernel.kallsyms])", "memcg_memory_event (in [kernel.kallsyms])" ], "registerTime": 0, "unregisterTime": null, "processType": "default" } ], "processes": [], "pausedRanges": [] } # Signed-off-by:Anup Sharma <anupnewsmail@gmail.com> Tested-by:
-
Anup Sharma authored
The stack has been created for storing func and dso from the callchain. The sample has been added to a specific thread. It first checks if the thread exists in the Thread class. Then it call _add_sample function which is responsible for appending a new entry to the samples list. Also callchain parsing and storing part is implemented. Moreover removed the comment from thread. Signed-off-by:
-
Anup Sharma authored
The intern_stack function is responsible for retrieving or creating a stack_id based on the provided frame_id and prefix_id. It first generates a key using the frame_id and prefix_id values. If the stack corresponding to the key is found in the stackMap, it is returned. Otherwise, a new stack is created by appending the prefix_id and frame_id to the stackTable. The key and the index of the newly created stack are added to the stackMap for future reference. The _intern_frame function is responsible for retrieving or creating a frame_id based on the provided frame string. If the frame_id corresponding to the frameString is found in the frameMap, it is returned. Otherwise, a new frame is created by appending relevant information to the frameTable and adding the frameString to the string_id through _intern_string. The _intern_string function will gets a matching string, or saves the new string and returns a String ID. Signed-off-by:
-
Anup Sharma authored
The final output will now be presented in JSON format following the Gecko profile structure. Additionally, the inclusion of PRODUCT allows easy retrieval of header information for UI. Furthermore, CATEGORIES have been introduced to enable customization of kernel and user colors using input arguments. To facilitate this functionality, an argparse-based parser has been implemented. Note: The implementation of threads will be addressed in subsequent commits for now I have commented it out. Signed-off-by:
-
Anup Sharma authored
This commit introduces new classes and conversion functions to facilitate the representation of Gecko profile information. The new classes Frame, Stack, Sample, and Thread are added to handle specific components of the profile data, also link to the origin docs has been commented out. Additionally, Inside the Thread class _to_json_dict() method has been created that converts the current thread data into the corresponding format expected by the GeckoThread JSON schema, as per the Gecko profile format specification. Signed-off-by:
-
Anup Sharma authored
The script takes in a sample event dictionary(param_dict) and retrieves relevant data such as time stamp, PID, TID, and comm for each event. Also start time is defined as a global variable as it need to be passed to trace_end for gecko meta information field creation. Signed-off-by:
-
Anup Sharma authored
Added necessary modules, including the Perf-Trace-Util library, and defines the required functions and variables for using perf script python. The perf_trace_context and Core modules for tracing and processing events has been also imported. Added usage information. Signed-off-by:
-
Xiu Jianfeng authored
The 'it' should be 'is' here, fix it. Signed-off-by:
Xiu Jianfeng <xiujianfeng@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/20230727105001.261420-1-xiujianfeng@huawei.comSigned-off-by:
-
Namhyung Kim authored
When 'perf record -d' is used, it needs data mmaps to symbolize global data. But it missed to collect kernel data maps so it cannot symbolize them. Instead of having a separate map, just increase the kernel map size to include the data section. Probably we can have a separate kernel map for data, but the current code assumes a single kernel map. So it'd require more changes in other places and looks error-prone. I decided not to go that way for now. Also it seems the kernel module size already includes the data section. For example, my system has the following. $ grep -e _stext -e _etext -e _edata /proc/kallsyms ffffffff99800000 T _stext ffffffff9a601ac8 T _etext ffffffff9b446a00 D _edata Size of the text section is (0x9a601ac8 - 0x99800000 = 0xe01ac8) and size including data section is (0x9b446a00 - 0x99800000 = 0x1c46a00). Before: $ perf record -d true $ perf report -D | grep MMAP | head -1 0 0 0x460 [0x60]: PERF_RECORD_MMAP -1/0: [0xffffffff99800000(0xe01ac8) @ 0xffffffff99800000]: x [kernel.kallsyms]_text ^^^^^^^^ here After: $ perf report -D | grep MMAP | head -1 0 0 0x460 [0x60]: PERF_RECORD_MMAP -1/0: [0xffffffff99800000(0x1c46a00) @ 0xffffffff99800000]: x [kernel.kallsyms]_text ^^^^^^^^^ Instead of just replacing it to _edata, try _edata first and then fall back to _etext just in case. Signed-off-by:Namhyung Kim <namhyung@kernel.org> Acked-by:
Adrian Hunter <adrian.hunter@intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@kernel.org> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Link: https://lore.kernel.org/r/20230725001929.368041-2-namhyung@kernel.orgSigned-off-by:
-
Namhyung Kim authored
The kallsyms__get_symbol_start() to get any symbol address from kallsyms. The existing kallsyms__get_function_start() only allows text symbols so create this to allow data symbols too. Signed-off-by:
-
Ian Rogers authored
Prefer informative messages rather than none with ABORT_ON. Document one failure mode and add an error message for another. Signed-off-by:
-
Ian Rogers authored
Improve the location for add PMU for cases when PMUs aren't found. Signed-off-by:
-
Ian Rogers authored
Follow convention from parse_events_terms__num/str and pass the YYLTYPE for the location. Signed-off-by:
-
Ian Rogers authored
When no events or PMUs match report an error for event_pmu: Before: ``` $ perf stat -e 'asdfasdf' -a sleep 1 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 'asdfasdf' -a sleep 1 event syntax error: 'asdfasdf' \___ Bad event name Unabled to find PMU or event on a PMU of 'asdfasdf' 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 ``` Fixes the inadvertent removal when hybrid parsing was modified. Fixes: 70c90e4a ("perf parse-events: Avoid scanning PMUs before parsing") Signed-off-by: -
Ian Rogers authored
Add PE_ABORT that will YYNOMEM or YYABORT accordingly. Signed-off-by:
-
Ian Rogers authored
Migration to improve error reporting as YYABORT cases should carry event parsing errors. Signed-off-by:
-
Ian Rogers authored
Split cases in event_pmu for greater accuracy. Signed-off-by:
-
Ian Rogers authored
If the name memory allocation fails then propagate to the parser. Committer notes: Use $(BISON_FALLBACK_FLAGS) on the bison call so that we continue building with older bison versions, before 3.81, where YYNOMEM isn't present. Signed-off-by:
-
Arnaldo Carvalho de Melo authored
YYNOMEM was introduced in bison 3.81, so define it as YYABORT for older versions, which should provide the previous perf behaviour. Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Signed-off-by:
-
- 26 Jul, 2023 5 commits
-
-
Ian Rogers authored
There is logic to avoid printing the regrouping warning for wild card PMUs, this logic also needs to apply for wild card events. Before: ``` $ perf stat -e '{data_read,data_write}' -a sleep 1 WARNING: events were regrouped to match PMUs Performance counter stats for 'system wide': 2,979.16 MiB data_read 410.26 MiB data_write 1.001541923 seconds time elapsed ``` After: ``` $ perf stat -e '{data_read,data_write}' -a sleep 1 Performance counter stats for 'system wide': 2,975.94 MiB data_read 432.05 MiB data_write 1.001119499 seconds time elapsed ``` Signed-off-by: -
Ian Rogers authored
Improve documentation of 'struct parse_events_state'. Signed-off-by:
-
Ian Rogers authored
The tokens PE_PREFIX_RAW and PE_PREFIX_GROUP are unused so remove them. Signed-off-by:
-
Ian Rogers authored
Removed by commit 70c90e4a ("perf parse-events: Avoid scanning PMUs before parsing"). Signed-off-by:
-
Ian Rogers authored
Removed by commit 70c90e4a ("perf parse-events: Avoid scanning PMUs before parsing"). Signed-off-by:
-
- 24 Jul, 2023 4 commits
-
-
Ian Rogers authored
Add an LTO build option, that sets the appropriate CFLAGS and CXXFLAGS values. Signed-off-by:
-
Ian Rogers authored
GCC LTO will complain that the array length varies for the arch_tests weak symbol. Use extern/static and architecture determining #if to workaround this problem. Signed-off-by:
-
Ian Rogers authored
With GCC LTO a potential use uninitialized is spotted: ``` In function ‘parse_events_config_bpf’, inlined from ‘parse_events_load_bpf’ at util/parse-events.c:874:8: util/parse-events.c:792:37: error: ‘error_pos’ may be used uninitialized [-Werror=maybe-uninitialized] 792 | idx = term->err_term + error_pos; | ^ util/parse-events.c: In function ‘parse_events_load_bpf’: util/parse-events.c:765:13: note: ‘error_pos’ was declared here 765 | int error_pos; | ^ ``` So initialize at declaration. Reviewed-by:Nick Desaulniers <ndesaulniers@google.com> Signed-off-by:
-
Ian Rogers authored
perf_event__read_stat_config will assign values based on number of tags and tag values. Initialize the structs to zero before they are assigned so that no uninitialized values can be seen. This potential error was reported by GCC with LTO enabled. Reviewed-by:
-
- 20 Jul, 2023 2 commits
-
-
Lu Hongfei authored
When wrapping code, use ';' better than using ',' which is more in line with the coding habits of most engineers. Signed-off-by:
Lu Hongfei <luhongfei@vivo.com> Acked-by:
-
Arnaldo Carvalho de Melo authored
[root@five ~]# perf bench uprobe all # Running uprobe/baseline benchmark... # Executed 1,000 usleep(1000) calls Total time: 1,053,963 usecs 1,053.963 usecs/op # Running uprobe/empty benchmark... # Executed 1,000 usleep(1000) calls Total time: 1,056,293 usecs +2,330 to baseline 1,056.293 usecs/op 2.330 usecs/op to baseline # Running uprobe/trace_printk benchmark... # Executed 1,000 usleep(1000) calls Total time: 1,056,977 usecs +3,014 to baseline +684 to previous 1,056.977 usecs/op 3.014 usecs/op to baseline 0.684 usecs/op to previous [root@five ~]# Acked-by:
-
