Previously, this command returns no help message on aarch64:
-> ./perf record --user-regs=?
available registers:
Usage: perf record [<options>] [<command>]
or: perf record [<options>] -- <command> [<options>]
With this change, the registers are listed.
-> ./perf record --user-regs=?
available registers: x0 x1 x2 x3 x4 x5 x6 x7 x8 x9 x10 x11 x12 x13 x14 x15 x16 x17 x18 x19 x20 x21 x22 x23 x24 x25 x26 x27 x28 x29 lr sp pc
It's also now possible to record subsets of registers on aarch64:
-> ./perf record --user-regs=x4,x5 ls
-> ./perf report --dump-raw-trace
12801163749305260 0xc70 [0x40]: PERF_RECORD_SAMPLE(IP, 0x2): 51956/51956: 0xffffaa6571f0 period: 145785 addr: 0
... user regs: mask 0x30 ABI 64-bit
.... x4 0x000000000000006c
.... x5 0x0000001001000001
... thread: ls:51956
...... dso: /usr/lib64/ld-2.17.so
Signed-off-by: Alexandre Truong <alexandre.truong@arm.com>
Tested-by: James Clark <james.clark@arm.com>
Acked-by: John Garry <john.garry@huawei.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Link: http://lore.kernel.org/lkml/20201127153923.26717-1-alexandre.truong@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
It calculates IPC from the cycles and instruction counts and compares it
with the shadow stat for both global aggregation (default) and no
aggregation mode.
$ perf stat -a -A -e cycles,instructions sleep 1
Performance counter stats for 'system wide':
CPU0 39,580,880 cycles
CPU1 45,426,945 cycles
CPU2 31,151,685 cycles
CPU3 55,167,421 cycles
CPU0 17,073,564 instructions # 0.43 insn per cycle
CPU1 34,955,764 instructions # 0.77 insn per cycle
CPU2 15,688,459 instructions # 0.50 insn per cycle
CPU3 34,699,217 instructions # 0.63 insn per cycle
1.003275495 seconds time elapsed
In this example, the 'insn per cycle' should be matched to the number
for each cpu. For CPU2, 0.50 = 15,688,459 / 31,151,685 .
Committer testing:
# perf test shadow
78: perf stat metrics (shadow stat) test : Ok
#
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20201127041404.390276-2-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Since some gcc generates a broken DWARF which lacks DW_AT_declaration
attribute from the subprogram DIE of function prototype.
(https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97060)
So, in addition to the DW_AT_declaration check, we also check the
subprogram DIE has DW_AT_inline or actual entry pc.
Committer testing:
# cat /etc/fedora-release
Fedora release 33 (Thirty Three)
#
Before:
# perf test vfs_getname
78: Use vfs_getname probe to get syscall args filenames : FAILED!
79: Check open filename arg using perf trace + vfs_getname : FAILED!
81: Add vfs_getname probe to get syscall args filenames : FAILED!
#
After:
# perf test vfs_getname
78: Use vfs_getname probe to get syscall args filenames : Ok
79: Check open filename arg using perf trace + vfs_getname : Ok
81: Add vfs_getname probe to get syscall args filenames : Ok
#
Reported-by: Thomas Richter <tmricht@linux.ibm.com>
Signed-off-by: Masami Hiramatsu <mhiramat@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Sumanth Korikkar <sumanthk@linux.ibm.com>
Link: http://lore.kernel.org/lkml/160645613571.2824037.7441351537890235895.stgit@devnote2
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Currently perf stat shows some metrics (like IPC) for defined events.
But when no aggregation mode is used (-A option), it shows incorrect
values since it used a value from a different cpu.
Before:
$ perf stat -aA -e cycles,instructions sleep 1
Performance counter stats for 'system wide':
CPU0 116,057,380 cycles
CPU1 86,084,722 cycles
CPU2 99,423,125 cycles
CPU3 98,272,994 cycles
CPU0 53,369,217 instructions # 0.46 insn per cycle
CPU1 33,378,058 instructions # 0.29 insn per cycle
CPU2 58,150,086 instructions # 0.50 insn per cycle
CPU3 40,029,703 instructions # 0.34 insn per cycle
1.001816971 seconds time elapsed
So the IPC for CPU1 should be 0.38 (= 33,378,058 / 86,084,722)
but it was 0.29 (= 33,378,058 / 116,057,380) and so on.
After:
$ perf stat -aA -e cycles,instructions sleep 1
Performance counter stats for 'system wide':
CPU0 109,621,384 cycles
CPU1 159,026,454 cycles
CPU2 99,460,366 cycles
CPU3 124,144,142 cycles
CPU0 44,396,706 instructions # 0.41 insn per cycle
CPU1 120,195,425 instructions # 0.76 insn per cycle
CPU2 44,763,978 instructions # 0.45 insn per cycle
CPU3 69,049,079 instructions # 0.56 insn per cycle
1.001910444 seconds time elapsed
Fixes: 44d49a6002 ("perf stat: Support metrics in --per-core/socket mode")
Reported-by: Sam Xi <xyzsam@google.com>
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Reviewed-by: Andi Kleen <ak@linux.intel.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20201127041404.390276-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
It didn't check the tool->cgroup_events bit which is set when the
--all-cgroups option is given. Without it, samples will not have cgroup
info so no reason to synthesize.
We can check the PERF_RECORD_CGROUP records after running perf record
*WITHOUT* the --all-cgroups option:
Before:
$ perf report -D | grep CGROUP
0 0 0x8430 [0x38]: PERF_RECORD_CGROUP cgroup: 1 /
CGROUP events: 1
CGROUP events: 0
CGROUP events: 0
After:
$ perf report -D | grep CGROUP
CGROUP events: 0
CGROUP events: 0
CGROUP events: 0
Committer testing:
Before:
# perf record -a sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 2.208 MB perf.data (10003 samples) ]
# perf report -D | grep "CGROUP events"
CGROUP events: 146
CGROUP events: 0
CGROUP events: 0
#
After:
# perf record -a sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 2.208 MB perf.data (10448 samples) ]
# perf report -D | grep "CGROUP events"
CGROUP events: 0
CGROUP events: 0
CGROUP events: 0
#
With all-cgroups:
# perf record --all-cgroups -a sleep 1
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 2.374 MB perf.data (11526 samples) ]
# perf report -D | grep "CGROUP events"
CGROUP events: 146
CGROUP events: 0
CGROUP events: 0
#
Fixes: 8fb4b67939 ("perf record: Add --all-cgroups option")
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20201127054356.405481-1-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
To pick the changes in:
7a078d2d18 ("libbpf, hashmap: Fix undefined behavior in hash_bits")
That don't entail any changes in tools/perf.
This addresses this perf build warning:
Warning: Kernel ABI header at 'tools/perf/util/hashmap.h' differs from latest version at 'tools/lib/bpf/hashmap.h'
diff -u tools/perf/util/hashmap.h tools/lib/bpf/hashmap.h
Not a kernel ABI, its just that this uses the mechanism in place for
checking kernel ABI files drift.
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
In the Armv8 ARM (ARM DDI 0487F.c), chapter "D10.2.6 Events packet", it
describes the event bit is valid with specific payload requirement. For
example, the Last Level cache access event, the bit is defined as:
E[8], byte 1 bit [0], when SZ == 0b01 , when SZ == 0b10 ,
or when SZ == 0b11
It requires the payload size is at least 2 bytes, when byte 1 (start
counting from 0) is valid, E[8] (bit 0 in byte 1) can be used for LLC
access event type. For safety, the code checks the condition for
payload size firstly, if meet the requirement for payload size, then
continue to parse event type.
If review function arm_spe_get_payload(), it has used cast, so any bytes
beyond the valid size have been set to zeros.
For this reason, we don't need to check payload size anymore afterwards
when parse events, thus this patch removes payload size conditions.
Suggested-by: Andre Przywara <andre.przywara@arm.com>
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Reviewed-by: Andre Przywara <andre.przywara@arm.com>
Acked-by: Will Deacon <will@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Al Grant <Al.Grant@arm.com>
Cc: Arnaldo Carvalho de Melo <acme@kernel.org>
Cc: Dave Martin <Dave.Martin@arm.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@arm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: John Garry <john.garry@huawei.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Mathieu Poirier <mathieu.poirier@linaro.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Wei Li <liwei391@huawei.com>
Link: https://lore.kernel.org/r/20201119152441.6972-12-leo.yan@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
