- 16 Nov, 2017 31 commits
-
-
Andi Kleen authored
When using leader sampling the values of the not sampled but counted events are shown by perf script in "period". Currently printing period is only allowed when the main event has a period, that is it is in frequency mode. This implies that we cannot dump the values of counted events when the leader event is not in frequency mode. Just remove the check that the period must be set on all events. It will just be printed as 0 instead if it's not available. This fixes the following: $ perf record -c 100000 -e '{cycles,branches}:S' $ perf script -F event,period Further commentary by Jiri Olsa: The period will be the value of configured period, not 0: int perf_evsel__parse_sample(struct ... ... data->period = evsel->attr.sample_period; $ perf record -c 100000 $ perf script -F event,period | head -3 Failed to open /tmp/perf-2048.map, continuing without symbols 100000 cycles:ppp: 100000 cycles:ppp: other than that I think we can remove that check, because we will have always sane number in period Signed-off-by:Andi Kleen <ak@linux.intel.com> Acked-by:
Jiri Olsa <jolsa@kernel.org> Link: http://lkml.kernel.org/r/20171109145528.23371-4-andi@firstfloor.orgSigned-off-by:
Arnaldo Carvalho de Melo <acme@redhat.com>
-
Andi Kleen authored
Document STAT and CACHE header entries. Signed-off-by:
-
Thomas-Mich Richter authored
Clarify the perf buildid-cache help text for the purge operation. The purge subcommand takes a list of files (binaries) as option parameter. Make the wording the same as for the add and remove operation. Signed-off-by:
Thomas-Mich Richter <tmricht@linux.vnet.ibm.com> Reviewed-by:
Hendrik Brueckner <brueckner@linux.vnet.ibm.com> Acked-by:
Masami Hiramatsu <mhiramat@kernel.org> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com> LPU-Reference: 20171107144853.12925-1-tmricht@linux.vnet.ibm.com Signed-off-by:
-
Sukadev Bhattiprolu authored
The POWER9 hardware has dropped support for several events, added a few new events and changed the category for a couple of events. Update the POWER9 events in Linux to reflect these changes. Signed-off-by:
Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com> Cc: Jiri Olsa <jolsa@redhat.com> Cc: Michael Ellerman <mpe@ellerman.id.au> Cc: Madhavan Srinivasan <maddy@linux.vnet.ibm.com> Link: http://lkml.kernel.org/r/20171108201938.GA10985@us.ibm.comSigned-off-by:
-
Arnaldo Carvalho de Melo authored
When processing PERF_RECORD_AUXTRACE_INFO several perf_evsel entries will be synthesized and inserted into session->evlist, eventually ending in perf_script.tool.sample(), which ends up calling builtin-script.c's process_event(), that expects evsel->priv to be a perf_evsel_script object with a valid FILE pointer in fp. So we need to intercept the processing of PERF_RECORD_AUXTRACE_INFO and then setup evsel->priv for these newly created perf_evsel instances, do it to fix the segfault in process_event() trying to use a NULL for that FILE pointer. Reported-by:
Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ravi Bangoria <ravi.bangoria@linux.vnet.ibm.com> Cc: Wang Nan <wangnan0@huawei.com> Cc: yuzhoujian <yuzhoujian@didichuxing.com> Fixes: a14390fd ("perf script: Allow creating per-event dump files") Link: http://lkml.kernel.org/n/tip-bthnur8r8de01gxvn2qayx6e@git.kernel.org [ Merge fix by Ravi Bangoria before pushing upstream to preserv bisectability ] Signed-off-by:
-
Seonghyun Park authored
Include newly added fields 'mmap2', 'comm_exec', 'use_clockid', 'namespaces', 'write_backward' and 'context_switch' from perf_event_attr to store_event(). Signed-off-by:
Seonghyun Park <seonghyun0p@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Seonghyun Park <seonghyun0p@gmail.com> Link: http://lkml.kernel.org/n/tip-vltn7pqhcv8h5fmo9cthk87q@git.kernel.org [ Fix log message to add 'write_backward', fix the patch to add 'use_clock_id' ] Signed-off-by:
-
Arnaldo Carvalho de Melo authored
I forgot one conversion, which got noticed by Thomas when running: $ perf stat -e '{cpu-clock,instructions}' kill kill: not enough arguments Segmentation fault (core dumped) $ Fix it, those stats are in evsel->stats, not anymore in evsel->priv. Reported-by: -
Jiri Olsa authored
Using the shell function for perl CFLAGS retrieval instead of back quotes (``). Both execute shell with the command, but the latter is more explicit and seems to be the preferred way. Also we don't have any other use of the back quotes in perf Makefiles. Signed-off-by:
-
Andrei Vagin authored
Currently if trace_event__register_resolver() fails, we return -errno, but we can't be sure that errno isn't zero in this case. Signed-off-by:
Andrei Vagin <avagin@openvz.org> Reviewed-by:
Jiri Olsa <jolsa@redhat.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Vasily Averin <vvs@virtuozzo.com> Link: http://lkml.kernel.org/r/20171108002246.8924-2-avagin@openvz.orgSigned-off-by:
-
Andi Kleen authored
Use a typed enum for the perf_evsel_config_term type enum. This allows gcc to do much stronger type checks, and also check for missing case statements. I removed the unused _MAX member from the number. It found one missing case. I'm not sure it's a real problem, so I just turned it into a BUG_ON for now. Signed-off-by:
-
Andi Kleen authored
The Intel PMU event aliases have a implicit period= specifier to set the default period. Unfortunately this breaks overriding these periods with -c or -F, because the alias terms look like they are user specified to the internal parser, and user specified event qualifiers override the command line options. Track that they are coming from aliases by adding a "weak" state to the term. Any weak terms don't override command line options. I only did it for -c/-F for now, I think that's the only case that's broken currently. Before: $ perf record -c 1000 -vv -e uops_issued.any ... { sample_period, sample_freq } 2000003 After: $ perf record -c 1000 -vv -e uops_issued.any ... { sample_period, sample_freq } 1000 Signed-off-by: -
Arnaldo Carvalho de Melo authored
Two more, that were just in perf/core and thus weren't covered by Ingo's latest headers synch, kcmp.h and prctl.h, silencing this: Warning: Kernel ABI header at 'tools/include/uapi/linux/kcmp.h' differs from latest version at 'include/uapi/linux/kcmp.h' Warning: Kernel ABI header at 'tools/include/uapi/linux/prctl.h' differs from latest version at 'include/uapi/linux/prctl.h' Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: David Ahern <dsahern@gmail.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Wang Nan <wangnan0@huawei.com> Link: http://lkml.kernel.org/n/tip-2a0r7iybyqpkftllyy5t9hfk@git.kernel.orgSigned-off-by:
-
Jiri Olsa authored
Align source with offset lines, which are more advanced, because of the address column. Before: : static void *worker_thread(void *__tdata) : { 0.00 : 48a971: push %rbp 0.00 : 48a972: mov %rsp,%rbp 0.00 : 48a975: sub $0x30,%rsp 0.00 : 48a979: mov %rdi,-0x28(%rbp) 0.00 : 48a97d: mov %fs:0x28,%rax 0.00 : 48a986: mov %rax,-0x8(%rbp) 0.00 : 48a98a: xor %eax,%eax : struct thread_data *td = __tdata; 0.00 : 48a98c: mov -0x28(%rbp),%rax 0.00 : 48a990: mov %rax,-0x10(%rbp) : int m = 0, i; 0.00 : 48a994: movl $0x0,-0x1c(%rbp) : int ret; : : for (i = 0; i < loops; i++) { 0.00 : 48a99b: movl $0x0,-0x18(%rbp) After: : static void *worker_thread(void *__tdata) : { 0.00 : 48a971: push %rbp 0.00 : 48a972: mov %rsp,%rbp 0.00 : 48a975: sub $0x30,%rsp 0.00 : 48a979: mov %rdi,-0x28(%rbp) 0.00 : 48a97d: mov %fs:0x28,%rax 0.00 : 48a986: mov %rax,-0x8(%rbp) 0.00 : 48a98a: xor %eax,%eax : struct thread_data *td = __tdata; 0.00 : 48a98c: mov -0x28(%rbp),%rax 0.00 : 48a990: mov %rax,-0x10(%rbp) : int m = 0, i; 0.00 : 48a994: movl $0x0,-0x1c(%rbp) : int ret; : : for (i = 0; i < loops; i++) { 0.00 : 48a99b: movl $0x0,-0x18(%rbp) It makes bigger different when displaying script sources, where the comment lines looks oddly shifted from the lines which actually hold code. I'll send script support separately. Committer note: Do not use a fixed column width for the addresses, as kernel ones se more than 10 columns, look at the last offset and get the right width. Signed-off-by: -
Jiri Olsa authored
Factor disasm_line__write function from annotate_browser__write, which now keeps only generic display code. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line in browser::b::top. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line in find functions: annotate_browser__find_string annotate_browser__find_string_reverse Signed-off-by:
-
Jiri Olsa authored
Using struct annotation_line arg in browser_line function to make it generic. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line as a browser::offsets array entry. Signed-off-by:
-
Jiri Olsa authored
Use struct annotation_line as a browser::selection. We want to be able to use the annotate_browser for all sorts of source data, so it needs to be able to work over the generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Rename disasm_line__browser function to browser_line, because the browser got generic and is no longer disasm specific. Signed-off-by:
-
Jiri Olsa authored
Rename struct browser_disasm_line to browser_line, because the browser operates now on generic lines and no longer on disasm lines. Signed-off-by:
-
Jiri Olsa authored
We now keep samples_nr in struct annotation_line, so there's no need to pass nr_events to disasm_rb_tree__insert function. Signed-off-by:
-
Jiri Olsa authored
We now carry the data in 'struct annotation_line', so using it instead of samples from 'struct browser_disasm_line' and removing it and its setup. Signed-off-by:
-
Jiri Olsa authored
Move generic annotation line display code into annotation_line__print function. Signed-off-by:
-
Jiri Olsa authored
Separating struct annotation_line display function, it will hold the generic line display code. Signed-off-by:
-
Jiri Olsa authored
Remove struct source_line*, no longer needed. Signed-off-by:
-
Jiri Olsa authored
Remove disasm__calc_percent() function, because it's no longer needed. Signed-off-by:
-
Jiri Olsa authored
Remove disasm__calc_percent() from annotate_browser__calc_percent(), because we already have the data calculated in struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Remove disasm__calc_percent() from disasm_line__print(), because we already have the data calculated in struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Replace symbol__get_source_line() with symbol__calc_lines(), which calculates the source line tree over the struct annotation_line. This will allow us to remove redundant struct source_line in following patches. Signed-off-by:
-
Jiri Olsa authored
Add symbol__calc_percent function, that calculates annotation data for symbol and put the data in the struct annotation_line::samples array. Committer notes: Made symbol__calc_percent non static to be used in the next two patches, which will get some fixups from jolsa, doing it this way to keep this bisectable. Signed-off-by:
-
- 13 Nov, 2017 9 commits
-
-
Jiri Olsa authored
Add samples array into struct annotation_line to hold the annotation data. The data is populated in the following patches. Signed-off-by:
-
Jiri Olsa authored
Mov disasm__purge() to annotated_source__purge() to make it work over a generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Changing the way the annotation lines are allocated and adding annotation_line__(new|delete) functions to deal with this. Before the allocation schema was as follows: ----------------------------------------------------------- struct disasm_line | struct annotation_line | private space ----------------------------------------------------------- Where the private space is used in TUI code to store computed annotation data for events. The stdio code computes the data on the fly. The goal is to compute and store annotation line's data directly in the struct annotation_line itself, so this patch changes the line allocation schema as follows: ------------------------------------------------------------ privsize space | struct disasm_line | struct annotation_line ------------------------------------------------------------ Moving struct annotation_line to the end, because in following changes we will move here the non-fixed length event's data. Signed-off-by:
-
Jiri Olsa authored
Move rb_node to struct annotation_line to make struct annotation_line the rb tree node for sorted lines used in both stdio and TUI code. This way we can unite the sorted lines lines codes for both TUI and stdio in the following patches. Signed-off-by:
-
Jiri Olsa authored
Rename disasm__add() into annotation_line__add() to make it work over a generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Rename disasm__get_next_ip_line() to annotation_line__next() to make it work over a generic struct annotation_line. Signed-off-by:
-
Jiri Olsa authored
Add evsel into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. This change also allow us to move the arch name initialization under symbol__annotate function. Link: http://lkml.kernel.org/n/tip-a9ok53rrgt1s5e8uglyvy6qt@git.kernel.orgSigned-off-by:
-
Jiri Olsa authored
Add offset/line/line_nr into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. Signed-off-by:
-
Jiri Olsa authored
Add map into struct annotate_args to reduce the number of arguments that need to travel all the way to line allocation. Signed-off-by:
-
