kernel/git/stable/linux.git/tools/perf/perf.h, branch linux-5.3.y

perf tools: Increase MAX_NR_CPUS and MAX_CACHES

2019-06-25T11:47:10+00:00

Attempting to profile 1024 or more CPUs with perf causes two errors: perf record -a [ perf record: Woken up X times to write data ] way too many cpu caches.. [ perf record: Captured and wrote X MB perf.data (X samples) ] perf report -C 1024 Error: failed to set cpu bitmap Requested CPU 1024 too large. Consider raising MAX_NR_CPUS Increasing MAX_NR_CPUS from 1024 to 2048 and redefining MAX_CACHES as MAX_NR_CPUS * 4 returns normal functionality to perf: perf record -a [ perf record: Woken up X times to write data ] [ perf record: Captured and wrote X MB perf.data (X samples) ] perf report -C 1024 ... Signed-off-by: Kyle Meyer Cc: Alexander Shishkin Cc: Jiri Olsa Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/20190620193630.154025-1-meyerk@stormcage.eag.rdlabs.hpecorp.net Signed-off-by: Arnaldo Carvalho de Melo

perf record: Add support to collect callchains from kernel or user space only

2019-06-10T18:50:01+00:00

One can just record callchains in the kernel or user space with this new options. We can use it together with "--all-kernel" options. This two options is used just like print_stack(sys) or print_ustack(usr) for systemtap. Shown below is the usage of this new option combined with "--all-kernel" options: 1. Configure all used events to run in kernel space and just collect kernel callchains. $ perf record -a -g --all-kernel --kernel-callchains 2. Configure all used events to run in kernel space and just collect user callchains. $ perf record -a -g --all-kernel --user-callchains Committer notes: Improved documentation to state that asking for kernel callchains really is asking for excluding user callchains, and vice versa. Further mentioned that using both won't get both, but nothing, as both will be excluded. Signed-off-by: yuzhoujian Acked-by: Jiri Olsa Cc: Adrian Hunter Cc: Alexander Shishkin Cc: David Ahern Cc: Milian Wolff Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Wang Nan Link: http://lkml.kernel.org/r/1559222962-22891-1-git-send-email-ufo19890607@gmail.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Implement COMPRESSED event record and its attributes

2019-05-15T19:36:49+00:00

Implemented PERF_RECORD_COMPRESSED event, related data types, header feature and functions to write, read and print feature attributes from the trace header section. comp_mmap_len preserves the size of mmaped kernel buffer that was used during collection. comp_mmap_len size is used on loading stage as the size of decomp buffer for decompression of COMPRESSED events content. Committer notes: Fixed up conflict with BPF_PROG_INFO and BTF_BTF header features. Signed-off-by: Alexey Budankov Reviewed-by: Jiri Olsa Cc: Alexander Shishkin Cc: Andi Kleen Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/ebbaf031-8dda-3864-ebc6-7922d43ee515@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Implement --mmap-flush= option

2019-04-01T18:18:10+00:00

Implement a --mmap-flush option that specifies minimal number of bytes that is extracted from mmaped kernel buffer to store into a trace. The default option value is 1 byte what means every time trace writing thread finds some new data in the mmaped buffer the data is extracted, possibly compressed and written to a trace. $ tools/perf/perf record --mmap-flush 1024 -e cycles -- matrix.gcc $ tools/perf/perf record --aio --mmap-flush 1K -e cycles -- matrix.gcc The option is independent from -z setting, doesn't vary with compression level and can serve two purposes. The first purpose is to increase the compression ratio of a trace data. Larger data chunks are compressed more effectively so the implemented option allows specifying data chunk size to compress. Also at some cases executing more write syscalls with smaller data size can take longer than executing less write syscalls with bigger data size due to syscall overhead so extracting bigger data chunks specified by the option value could additionally decrease runtime overhead. The second purpose is to avoid self monitoring live-lock issue in system wide (-a) profiling mode. Profiling in system wide mode with compression (-a -z) can additionally induce data into the kernel buffers along with the data from monitored processes. If performance data rate and volume from the monitored processes is high then trace streaming and compression activity in the tool is also high. High tool process activity can lead to subtle live-lock effect when compression of single new byte from some of mmaped kernel buffer leads to generation of the next single byte at some mmaped buffer. So perf tool process ends up in endless self monitoring. Implemented synch parameter is the mean to force data move independently from the specified flush threshold value. Despite the provided flush value the tool needs capability to unconditionally drain memory buffers, at least in the end of the collection. Committer testing: Running with the default value, i.e. as soon as there is something to read go on consuming, we first write the synthesized events, small chunks of about 128 bytes: # perf trace -m 2048 --call-graph dwarf -e write -- perf record 101.142 ( 0.004 ms): perf/25821 write(fd: 3, buf: 0x210db60, count: 120) = 120 __libc_write (/usr/lib64/libpthread-2.28.so) ion (/home/acme/bin/perf) record__write (inlined) process_synthesized_event (/home/acme/bin/perf) perf_tool__process_synth_event (inlined) perf_event__synthesize_mmap_events (/home/acme/bin/perf) Then we move to reading the mmap buffers consuming the events put there by the kernel perf infrastructure: 107.561 ( 0.005 ms): perf/25821 write(fd: 3, buf: 0x7f1befc02000, count: 336) = 336 __libc_write (/usr/lib64/libpthread-2.28.so) ion (/home/acme/bin/perf) record__write (inlined) record__pushfn (/home/acme/bin/perf) perf_mmap__push (/home/acme/bin/perf) record__mmap_read_evlist (inlined) record__mmap_read_all (inlined) __cmd_record (inlined) cmd_record (/home/acme/bin/perf) 12919.953 ( 0.136 ms): perf/25821 write(fd: 3, buf: 0x7f1befc83150, count: 184984) = 184984 12920.094 ( 0.155 ms): perf/25821 write(fd: 3, buf: 0x7f1befc02150, count: 261816) = 261816 12920.253 ( 0.093 ms): perf/25821 write(fd: 3, buf: 0x7f1befb81120, count: 170832) = 170832 If we limit it to write only when more than 16MB are available for reading, it throttles that to a quarter of the --mmap-pages set for 'perf record', which by default get to 528384 bytes, found out using 'record -v': mmap flush: 132096 mmap size 528384B With that in place all the writes coming from record__mmap_read_evlist(), i.e. from the mmap buffers setup by the kernel perf infrastructure were at least 132096 bytes long. Trying with a bigger mmap size: perf trace -e write perf record -v -m 2048 --mmap-flush 16M 74982.928 ( 2.471 ms): perf/26500 write(fd: 3, buf: 0x7ff94a6cc000, count: 3580888) = 3580888 74985.406 ( 2.353 ms): perf/26500 write(fd: 3, buf: 0x7ff949ecb000, count: 3453256) = 3453256 74987.764 ( 2.629 ms): perf/26500 write(fd: 3, buf: 0x7ff9496ca000, count: 3859232) = 3859232 74990.399 ( 2.341 ms): perf/26500 write(fd: 3, buf: 0x7ff948ec9000, count: 3769032) = 3769032 74992.744 ( 2.064 ms): perf/26500 write(fd: 3, buf: 0x7ff9486c8000, count: 3310520) = 3310520 74994.814 ( 2.619 ms): perf/26500 write(fd: 3, buf: 0x7ff947ec7000, count: 4194688) = 4194688 74997.439 ( 2.787 ms): perf/26500 write(fd: 3, buf: 0x7ff9476c6000, count: 4029760) = 4029760 Was again limited to a quarter of the mmap size: mmap flush: 2098176 mmap size 8392704B A warning about that would be good to have but can be added later, something like: "max flush is a quarter of the mmap size, if wanting to bump the mmap flush further, bump the mmap size as well using -m/--mmap-pages" Also rename the 'sync' parameters to 'synch' to keep tools/perf building with older glibcs: cc1: warnings being treated as errors builtin-record.c: In function 'record__mmap_read_evlist': builtin-record.c:775: warning: declaration of 'sync' shadows a global declaration /usr/include/unistd.h:933: warning: shadowed declaration is here builtin-record.c: In function 'record__mmap_read_all': builtin-record.c:856: warning: declaration of 'sync' shadows a global declaration /usr/include/unistd.h:933: warning: shadowed declaration is here Signed-off-by: Alexey Budankov Reviewed-by: Jiri Olsa Tested-by: Arnaldo Carvalho de Melo Cc: Alexander Shishkin Cc: Andi Kleen Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/f6600d72-ecfa-2eb7-7e51-f6954547d500@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Replace option --bpf-event with --no-bpf-event

2019-03-19T19:52:06+00:00

Currently, monitoring of BPF programs through bpf_event is off by default for 'perf record'. To turn it on, the user need to use option "--bpf-event". As BPF gets wider adoption in different subsystems, this option becomes inconvenient. This patch makes bpf_event on by default, and adds option "--no-bpf-event" to turn it off. Since option --bpf-event is not released yet, it is safe to remove it. Signed-off-by: Song Liu Reviewed-by: Jiri Olsa Cc: Alexei Starovoitov Cc: Daniel Borkmann Cc: kernel-team@fb.com Cc: Namhyung Kim Cc: Peter Zijlstra Cc: Stanislav Fomichev Link: http://lkml.kernel.org/r/20190312053051.2690567-2-songliubraving@fb.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Allocate affinity masks

2019-02-06T13:00:39+00:00

Allocate affinity option and masks for mmap data buffers and record thread as well as initialize allocated objects. Signed-off-by: Alexey Budankov Reviewed-by: Jiri Olsa Cc: Alexander Shishkin Cc: Andi Kleen Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/526fa2b0-07de-6dbd-a7e9-26ba875593c9@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Handle PERF_RECORD_BPF_EVENT

2019-01-21T20:00:57+00:00

This patch adds basic handling of PERF_RECORD_BPF_EVENT. Tracking of PERF_RECORD_BPF_EVENT is OFF by default. Option --bpf-event is added to turn it on. Committer notes: Add dummy machine__process_bpf_event() variant that returns zero for systems without HAVE_LIBBPF_SUPPORT, such as Alpine Linux, unbreaking the build in such systems. Remove the needless include from bpf->event.h, provide just forward declarations for the structs and unions in the parameters, to reduce compilation time and needless rebuilds when machine.h gets changed. Committer testing: When running with: # perf record --bpf-event On an older kernel where PERF_RECORD_BPF_EVENT and PERF_RECORD_KSYMBOL is not present, we fallback to removing those two bits from perf_event_attr, making the tool to continue to work on older kernels: perf_event_attr: size 112 { sample_period, sample_freq } 4000 sample_type IP|TID|TIME|PERIOD read_format ID disabled 1 inherit 1 mmap 1 comm 1 freq 1 enable_on_exec 1 task 1 precise_ip 3 sample_id_all 1 exclude_guest 1 mmap2 1 comm_exec 1 ksymbol 1 bpf_event 1 ------------------------------------------------------------ sys_perf_event_open: pid 5779 cpu 0 group_fd -1 flags 0x8 sys_perf_event_open failed, error -22 switching off bpf_event ------------------------------------------------------------ perf_event_attr: size 112 { sample_period, sample_freq } 4000 sample_type IP|TID|TIME|PERIOD read_format ID disabled 1 inherit 1 mmap 1 comm 1 freq 1 enable_on_exec 1 task 1 precise_ip 3 sample_id_all 1 exclude_guest 1 mmap2 1 comm_exec 1 ksymbol 1 ------------------------------------------------------------ sys_perf_event_open: pid 5779 cpu 0 group_fd -1 flags 0x8 sys_perf_event_open failed, error -22 switching off ksymbol ------------------------------------------------------------ perf_event_attr: size 112 { sample_period, sample_freq } 4000 sample_type IP|TID|TIME|PERIOD read_format ID disabled 1 inherit 1 mmap 1 comm 1 freq 1 enable_on_exec 1 task 1 precise_ip 3 sample_id_all 1 exclude_guest 1 mmap2 1 comm_exec 1 ------------------------------------------------------------ And then proceeds to work without those two features. As passing --bpf-event is an explicit action performed by the user, perhaps we should emit a warning telling that the kernel has no such feature, but this can be done on top of this patch. Now with a kernel that supports these events, start the 'record --bpf-event -a' and then run 'perf trace sleep 10000' that will use the BPF augmented_raw_syscalls.o prebuilt (for another kernel version even) and thus should generate PERF_RECORD_BPF_EVENT events: [root@quaco ~]# perf record -e dummy -a --bpf-event ^C[ perf record: Woken up 1 times to write data ] [ perf record: Captured and wrote 0.713 MB perf.data ] [root@quaco ~]# bpftool prog 13: cgroup_skb tag 7be49e3934a125ba gpl loaded_at 2019-01-19T09:09:43-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 13,14 14: cgroup_skb tag 2a142ef67aaad174 gpl loaded_at 2019-01-19T09:09:43-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 13,14 15: cgroup_skb tag 7be49e3934a125ba gpl loaded_at 2019-01-19T09:09:43-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 15,16 16: cgroup_skb tag 2a142ef67aaad174 gpl loaded_at 2019-01-19T09:09:43-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 15,16 17: cgroup_skb tag 7be49e3934a125ba gpl loaded_at 2019-01-19T09:09:44-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 17,18 18: cgroup_skb tag 2a142ef67aaad174 gpl loaded_at 2019-01-19T09:09:44-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 17,18 21: cgroup_skb tag 7be49e3934a125ba gpl loaded_at 2019-01-19T09:09:45-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 21,22 22: cgroup_skb tag 2a142ef67aaad174 gpl loaded_at 2019-01-19T09:09:45-0300 uid 0 xlated 296B jited 229B memlock 4096B map_ids 21,22 31: tracepoint name sys_enter tag 12504ba9402f952f gpl loaded_at 2019-01-19T09:19:56-0300 uid 0 xlated 512B jited 374B memlock 4096B map_ids 30,29,28 32: tracepoint name sys_exit tag c1bd85c092d6e4aa gpl loaded_at 2019-01-19T09:19:56-0300 uid 0 xlated 256B jited 191B memlock 4096B map_ids 30,29 # perf report -D | grep PERF_RECORD_BPF_EVENT | nl 1 0 55834574849 0x4fc8 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 13 2 0 60129542145 0x5118 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 14 3 0 64424509441 0x5268 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 15 4 0 68719476737 0x53b8 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 16 5 0 73014444033 0x5508 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 17 6 0 77309411329 0x5658 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 18 7 0 90194313217 0x57a8 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 21 8 0 94489280513 0x58f8 [0x18]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 22 9 7 620922484360 0xb6390 [0x30]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 29 10 7 620922486018 0xb6410 [0x30]: PERF_RECORD_BPF_EVENT bpf event with type 2, flags 0, id 29 11 7 620922579199 0xb6490 [0x30]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 30 12 7 620922580240 0xb6510 [0x30]: PERF_RECORD_BPF_EVENT bpf event with type 2, flags 0, id 30 13 7 620922765207 0xb6598 [0x30]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 31 14 7 620922874543 0xb6620 [0x30]: PERF_RECORD_BPF_EVENT bpf event with type 1, flags 0, id 32 # There, the 31 and 32 tracepoint BPF programs put in place by 'perf trace'. Signed-off-by: Song Liu Reviewed-by: Arnaldo Carvalho de Melo Tested-by: Arnaldo Carvalho de Melo Cc: Alexei Starovoitov Cc: Daniel Borkmann Cc: Peter Zijlstra Cc: kernel-team@fb.com Cc: netdev@vger.kernel.org Link: http://lkml.kernel.org/r/20190117161521.1341602-7-songliubraving@fb.com Signed-off-by: Arnaldo Carvalho de Melo

perf tools: Allow specifying proc-map-timeout in config file

2018-12-17T17:56:57+00:00

The default timeout of 500ms for parsing /proc//maps files is too short for profiling many of our services. This can be overridden by passing --proc-map-timeout to the relevant command but it'd be nice to globally increase our default value. This patch permits setting a different default with the core.proc-map-timeout config file parameter. Signed-off-by: Mark Drayton Acked-by: Song Liu Acked-by: Namhyung Kim Cc: Alexander Shishkin Cc: Jiri Olsa Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/20181204203420.1683114-1-mbd@fb.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Enable asynchronous trace writing

2018-12-17T17:55:08+00:00

The trace file offset is read once before mmaps iterating loop and written back after all performance data is enqueued for aio writing. The trace file offset is incremented linearly after every successful aio write operation. record__aio_sync() blocks till completion of the started AIO operation and then proceeds. record__aio_mmap_read_sync() implements a barrier for all incomplete aio write requests. Signed-off-by: Alexey Budankov Reviewed-by: Jiri Olsa Acked-by: Namhyung Kim Cc: Alexander Shishkin Cc: Andi Kleen Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/ce2d45e9-d236-871c-7c8f-1bed2d37e8ac@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo

perf record: Encode -k clockid frequency into Perf trace

2018-10-18T14:16:38+00:00

Store -k clockid frequency into Perf trace to enable timestamps derived metrics conversion into wall clock time on reporting stage. Below is the example of perf report output: tools/perf/perf record -k raw -- ../../matrix/linux/matrix.gcc ... [ perf record: Captured and wrote 31.222 MB perf.data (818054 samples) ] tools/perf/perf report --header # ======== ... # event : name = cycles:ppp, , size = 112, { sample_period, sample_freq } = 4000, sample_type = IP|TID|TIME|PERIOD, disabled = 1, inherit = 1, mmap = 1, comm = 1, freq = 1, enable_on_exec = 1, task = 1, precise_ip = 3, sample_id_all = 1, exclude_guest = 1, mmap2 = 1, comm_exec = 1, use_clockid = 1, clockid = 4 ... # clockid frequency: 1000 MHz ... # ======== Signed-off-by: Alexey Budankov Cc: Alexander Shishkin Cc: Andi Kleen Cc: Jiri Olsa Cc: Namhyung Kim Cc: Peter Zijlstra Link: http://lkml.kernel.org/r/23a4a1dc-b160-85a0-347d-40a2ed6d007b@linux.intel.com Signed-off-by: Arnaldo Carvalho de Melo