kernel/tools/perf/Documentation/perf-config.txt, branch linux-6.2.y

perf auxtrace: Add itrace option flag d+e to log on error

2022-10-04T11:55:21Z

Add flag +e to the itrace d (decoder debug log) option to get output only on decoding errors. The log can be very big so reducing the output to where there are decoding errors can be useful for analyzing errors. By default, the log size in that case is 16384 bytes, but can be altered by perf config e.g. perf config itrace.debug-log-buffer-size=30000 Reviewed-by: Andi Kleen Reviewed-by: Namhyung Kim Signed-off-by: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Link: https://lore.kernel.org/r/20220905073424.3971-3-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Disable debuginfod by default

2022-01-15T20:41:25Z

Fedora 35 sets DEBUGINFOD_URLS by default, which might lead to unexpected stalls in perf record exit path, when we try to cache profiled binaries. # DEBUGINFOD_PROGRESS=1 ./perf record -a ^C[ perf record: Woken up 1 times to write data ] Downloading from https://debuginfod.fedoraproject.org/ 447069 Downloading from https://debuginfod.fedoraproject.org/ 1502175 Downloading \^Z Disabling DEBUGINFOD_URLS by default in perf record and adding debuginfod option and .perfconfig variable support to enable id. Default without debuginfo processing: # perf record -a Using system debuginfod setup: # perf record -a --debuginfod Using custom debuginfd url: # perf record -a --debuginfod='https://evenbetterdebuginfodserver.krava' Adding single perf_debuginfod_setup function and using it also in perf buildid-cache command. Signed-off-by: Jiri Olsa Cc: Alexander Shishkin Cc: Frank Ch. Eigler Cc: Ian Rogers Cc: Mark Rutland Cc: Michael Petlan Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lore.kernel.org/lkml/20211209200425.303561-1-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf intel-pt: Add a config for max loops without consuming a packet

2021-07-07T14:40:56Z

The Intel PT decoder limits the number of unconditional branches (e.g. jmps) decoded without consuming any trace packets. Generally, a loop needs a conditional branch which generates a TNT packet, whereas a "ret" instruction will generate a TIP or TNT packet. So exceeding the limit is assumed to be a never-ending loop, which can happen if there has been a decoding error putting the decoder at the wrong place in the code. Up until now, the limit of 10000 has been enough but some analytic purposes have been reported to exceed that. Increase the limit to 100000, and make it configurable via perf config intel-pt.max-loops. Also amend the "Never-ending loop" message to mention the configuration entry. Signed-off-by: Adrian Hunter Reviewed-by: Andi Kleen Cc: Jiri Olsa Link: http://lore.kernel.org/lkml/20210701175132.3977-1-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf report: Make --skip-empty as default

2021-04-29T13:30:59Z

so that the compact output is shown by default. Also add 'report.skip-empty' config option to override the default. Users can also use --no-skip-empty command line option to change the behavior anytime. Committer testing: $ perf report --stat Aggregated stats: TOTAL events: 19 COMM events: 2 EXIT events: 1 SAMPLE events: 8 MMAP2 events: 4 FINISHED_ROUND events: 1 THREAD_MAP events: 1 CPU_MAP events: 1 TIME_CONV events: 1 cycles:u stats: SAMPLE events: 8 $ perf config report.skip-empty=false $ perf report --stat Aggregated stats: TOTAL events: 19 MMAP events: 0 LOST events: 0 COMM events: 2 EXIT events: 1 THROTTLE events: 0 UNTHROTTLE events: 0 FORK events: 0 READ events: 0 SAMPLE events: 8 MMAP2 events: 4 AUX events: 0 ITRACE_START events: 0 LOST_SAMPLES events: 0 SWITCH events: 0 SWITCH_CPU_WIDE events: 0 NAMESPACES events: 0 KSYMBOL events: 0 BPF_EVENT events: 0 CGROUP events: 0 TEXT_POKE events: 0 ATTR events: 0 EVENT_TYPE events: 0 TRACING_DATA events: 0 BUILD_ID events: 0 FINISHED_ROUND events: 1 ID_INDEX events: 0 AUXTRACE_INFO events: 0 AUXTRACE events: 0 AUXTRACE_ERROR events: 0 THREAD_MAP events: 1 CPU_MAP events: 1 STAT_CONFIG events: 0 STAT events: 0 STAT_ROUND events: 0 EVENT_UPDATE events: 0 TIME_CONV events: 1 FEATURE events: 0 COMPRESSED events: 0 cycles:u stats: SAMPLE events: 8 $ perf config report.skip-empty report.skip-empty=false $ Reviewed-by: Andi Kleen Signed-off-by: Namhyung Kim Tested-by: Arnaldo Carvalho de Melo Acked-by: Jiri Olsa Cc: Alexander Shishkin Cc: Ian Rogers Cc: Ingo Molnar Cc: Mark Rutland Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20210427013717.1651674-6-namhyung@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf config: Add annotate.demangle{,_kernel}

2021-03-06T19:42:31Z

Committer notes: This allows setting this in from the command line: $ perf config annotate.demangle $ perf config annotate.demangle=yes $ perf config annotate.demangle annotate.demangle=yes $ cat ~/.perfconfig # this file is auto-generated. [report] sort-order = srcline [annotate] demangle = yes $ $ $ perf config annotate.demangle_kernel $ perf config annotate.demangle_kernel=yes $ perf config annotate.demangle_kernel annotate.demangle_kernel=yes $ cat ~/.perfconfig # this file is auto-generated. [report] sort-order = srcline [annotate] demangle = yes demangle_kernel = yes $ Signed-off-by: Martin Liška Tested-by: Arnaldo Carvalho de Melo Link: https://lore.kernel.org/r/c96aabe7-791f-9503-295f-3147a9d19b60@suse.cz Signed-off-by: Arnaldo Carvalho de Melo

perf daemon: Add config file support

2021-02-11T13:02:54Z

Adding support to configure daemon with config file. Each client or server invocation of perf daemon needs to know the base directory, where all sessions data is stored. The base is defined with: daemon.base Base path for daemon data. All sessions data are stored under this path. The daemon allows to create record sessions. Each session is a record command spawned and monitored by perf daemon. The session is defined with: session-.run Defines new record session for daemon. The value is record's command line without the 'record' keyword. Example: # cat ~/.perfconfig [daemon] base=/opt/perfdata [session-cycles] run = -m 10M -e cycles --overwrite --switch-output -a [session-sched] run = -m 20M -e sched:* --overwrite --switch-output -a The example above defines '/opt/perfdata' as the base directory and 2 record sessions. # perf daemon start [2021-01-28 19:47:33.454413] daemon started (pid 16015) [2021-01-28 19:47:33.455910] reconfig: ruining session [cycles:16016]: -m 10M -e cycles --overwrite --switch-output -a [2021-01-28 19:47:33.456599] reconfig: ruining session [sched:16017]: -m 20M -e sched:* --overwrite --switch-output -a # ps -ef | grep perf ... perf daemon start ... /home/jolsa/.../perf record -m 20M -e cycles --overwrite --switch-output -a ... /home/jolsa/.../perf record -m 20M -e sched:* --overwrite --switch-output -a The base directory is populated with: # find /opt/perfdata/ /opt/perfdata/ /opt/perfdata/control <- control socket /opt/perfdata/session-cycles <- data for session 'cycles': /opt/perfdata/session-cycles/output <- perf record output /opt/perfdata/session-cycles/perf.data <- perf data /opt/perfdata/session-sched <- ditto for session 'sched' /opt/perfdata/session-sched/output /opt/perfdata/session-sched/perf.data Signed-off-by: Jiri Olsa Cc: Alexander Shishkin Cc: Alexei Budankov Cc: Ian Rogers Cc: Ingo Molnar Cc: Mark Rutland Cc: Michael Petlan Cc: Namhyung Kim Cc: Peter Zijlstra Link: https://lore.kernel.org/r/20210208200908.1019149-7-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf buildid-cache: Add --debuginfod option to specify a server to fetch debug files

2020-12-28T15:20:39Z

Add the --debuginfod option to specify debuginfod URL and support to do that through config file as well. Use the following in ~/.perfconfig file: [buildid-cache] debuginfod=http://192.168.122.174:8002 Signed-off-by: Jiri Olsa Acked-by: Ian Rogers Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexei Budankov Cc: Andi Kleen Cc: Mark Rutland Cc: Michael Petlan Cc: Peter Zijlstra Cc: Song Liu Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20201214105457.543111-14-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf record: Add --buildid-mmap option to enable PERF_RECORD_MMAP2's build id

2020-12-28T14:35:57Z

Add --buildid-mmap option to enable build id in PERF_RECORD_MMAP2 events. It will only work if there's kernel support for that and it disables build id cache (implies --no-buildid). It's also possible to enable it permanently via config option in ~/.perfconfig file: [record] build-id=mmap Also added build_id bit in the verbose output for perf_event_attr: # perf record --buildid-mmap -vv ... perf_event_attr: type 1 size 120 ... build_id 1 Adding also missing text_poke bit. Committer testing: $ perf record -h build Usage: perf record [] [] or: perf record [] -- [] -B, --no-buildid do not collect buildids in perf.data -N, --no-buildid-cache do not update the buildid cache --buildid-all Record build-id of all DSOs regardless of hits --buildid-mmap Record build-id in map events $ $ perf record --buildid-mmap sleep 1 Failed: no support to record build id in mmap events, update your kernel. $ After adding the needed kernel bits in a test kernel: $ perf record -vv --buildid-mmap sleep 1 |& grep -m1 build Enabling build id in mmap2 events. $ perf evlist -v cycles:u: size: 120, { sample_period, sample_freq }: 4000, sample_type: IP|TID|TIME|PERIOD, read_format: ID, disabled: 1, inherit: 1, exclude_kernel: 1, mmap: 1, comm: 1, freq: 1, enable_on_exec: 1, task: 1, sample_id_all: 1, exclude_guest: 1, mmap2: 1, comm_exec: 1, ksymbol: 1, bpf_event: 1, build_id: 1 $ Signed-off-by: Jiri Olsa Tested-by: Arnaldo Carvalho de Melo Acked-by: Ian Rogers Acked-by: Namhyung Kim Cc: Adrian Hunter Cc: Alexander Shishkin Cc: Alexei Budankov Cc: Andi Kleen Cc: Mark Rutland Cc: Michael Petlan Cc: Peter Zijlstra Cc: Song Liu Cc: Stephane Eranian Link: http://lore.kernel.org/lkml/20201214105457.543111-16-jolsa@kernel.org Signed-off-by: Arnaldo Carvalho de Melo

perf config: Fix example command in manpage to conform to syntax specified in the SYNOPSIS section.

2020-12-17T17:36:17Z

Committer testing: With the previously documented example: $ perf config --user report sort-order=srcline The config variable does not contain a section name: report $ With the fixed example line: $ perf config --user report.sort-order=srcline $ perf config --user report.sort-order report.sort-order=srcline $ Signed-off-by: Nick Thompson Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Link: https://lore.kernel.org/linux-perf-users/20201217142619.GA14524@redhat.com/ Signed-off-by: Arnaldo Carvalho de Melo

perf annotate: Allow configuring the 'disassembler_style' knob via 'perf config'

2020-09-04T19:07:23Z

# perf annotate --stdio2 acpi_processor_ffh_cstate_enter > default # perf config annotate.disassembler_style=intel # perf config annotate.disassembler_style annotate.disassembler_style=intel # perf annotate --stdio2 acpi_processor_ffh_cstate_enter > intel # diff -u default intel --- default 2020-09-04 13:09:26.019205732 -0300 +++ intel 2020-09-04 13:09:52.823795081 -0300 @@ -1,42 +1,42 @@ Samples: 1K of event 'cycles', 4000 Hz, Event count (approx.): 990065316, [percent: local period] acpi_processor_ffh_cstate_enter() /lib/modules/5.9.0-rc3/build/vmlinux -Percent → callq __fentry__ - mov cpu_number,%edx - mov %edx,%edx - mov cpu_cstate_entry,%rax - add -0x7dbe9700(,%rdx,8),%rax - movzbl 0x9(%rdi),%edx - mov 0x4(%rax,%rdx,8),%edi - mov (%rax,%rdx,8),%esi - → jmpq 137ccc6 - 2d: → jmpq 137ccd8 +Percent → call __fentry__ + mov edx,DWORD PTR gs:[rip+0x7e541d74] + mov edx,edx + mov rax,QWORD PTR [rip+0x152b8fb] + add rax,QWORD PTR [rdx*8-0x7dbe9700] + movzx edx,BYTE PTR [rdi+0x9] + mov edi,DWORD PTR [rax+rdx*8+0x4] + mov esi,DWORD PTR [rax+rdx*8] + → jmp 137ccc6 + 2d: → jmp 137ccd8 mfence - mov %gs:0x17bc0,%rax - clflush (%rax) + mov rax,QWORD PTR gs:0x17bc0 + clflush BYTE PTR [rax] mfence - xor %edx,%edx - mov %rdx,%rcx - mov %gs:0x17bc0,%rax - 0.00 monitor %rax,%ecx,%edx - mov (%rax),%rax - test $0x8,%al + xor edx,edx + mov rcx,rdx + mov rax,QWORD PTR gs:0x17bc0 + 0.00 monitor + mov rax,QWORD PTR [rax] + test al,0x8 ↓ jne 71 - ↓ jmpq 68 - verw 0x538b08(%rip) # ffffffff82008150 - 68: mov %rsi,%rax - mov %rdi,%rcx -100.00 mwait %eax,%ecx - 71: mov %gs:0x17bc0,%rax - lock andb $0xdf,0x2(%rax) - lock addl $0x0,-0x4(%rsp) - mov (%rax),%rax - test $0x8,%al + ↓ jmp 68 + verw WORD PTR [rip+0x538b08] # ffffffff82008150 + 68: mov rax,rsi + mov rcx,rdi +100.00 mwait + 71: mov rax,QWORD PTR gs:0x17bc0 + lock and BYTE PTR [rax+0x2],0xdf + lock add DWORD PTR [rsp-0x4],0x0 + mov rax,QWORD PTR [rax] + test al,0x8 ↓ je 97 - andl $0x7fffffff,__preempt_count - 97: ← retq - mov %gs:0x17bc0,%rax - lock orb $0x20,0x2(%rax) - mov (%rax),%rax - test $0x8,%al + and DWORD PTR gs:[rip+0x7e548509],0x7fffffff + 97: ret + mov rax,QWORD PTR gs:0x17bc0 + lock or BYTE PTR [rax+0x2],0x20 + mov rax,QWORD PTR [rax] + test al,0x8 ↑ jne 71 - ↑ jmpq 2d + ↑ jmp 2d # Requested-by: Matt P. Dziubinski Cc: Adrian Hunter Cc: Ian Rogers Cc: Jiri Olsa Cc: Namhyung Kim Signed-off-by: Arnaldo Carvalho de Melo