The recent kernel added lost count can be read from either read(2) or
ring buffer data with PERF_SAMPLE_READ. As it's a variable length data
we need to access it according to the format info.
But for perf tools use cases, PERF_FORMAT_ID is always set. So we can
only check PERF_FORMAT_LOST bit to determine the data format.
Add sample_read_value_size() and next_sample_read_value() helpers to
make it a bit easier to access. Use them in all places where it reads
the struct sample_read_value.
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220819003644.508916-5-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
A mask encoding of a cpu map is laid out as:
u16 nr
u16 long_size
unsigned long mask[];
However, the mask may be 8-byte aligned meaning there is a 4-byte pad
after long_size. This means 32-bit and 64-bit builds see the mask as
being at different offsets. On top of this the structure is in the byte
data[] encoded as:
u16 type
char data[]
This means the mask's struct isn't the required 4 or 8 byte aligned, but
is offset by 2. Consequently the long reads and writes are causing
undefined behavior as the alignment is broken.
Fix the mask struct by creating explicit 32 and 64-bit variants, use a
union to avoid data[] and casts; the struct must be packed so the
layout matches the existing perf.data layout. Taking an address of a
member of a packed struct breaks alignment so pass the packed
perf_record_cpu_map_data to functions, so they can access variables with
the right alignment.
As the 64-bit version has 4 bytes of padding, optimizing writing to only
write the 32-bit version.
Committer notes:
Disable warnings about 'packed' that break the build in some arches like
riscv64, but just around that specific struct.
Signed-off-by: Ian Rogers <irogers@google.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Bayduraev <alexey.v.bayduraev@linux.intel.com>
Cc: Athira Jajeev <atrajeev@linux.vnet.ibm.com>
Cc: Colin Ian King <colin.king@intel.com>
Cc: Dave Marchevsky <davemarchevsky@fb.com>
Cc: German Gomez <german.gomez@arm.com>
Cc: Gustavo A. R. Silva <gustavoars@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: James Clark <james.clark@arm.com>
Cc: Kees Kook <keescook@chromium.org>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Riccardo Mancini <rickyman7@gmail.com>
Cc: Song Liu <songliubraving@fb.com>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20220614143353.1559597-5-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Some processes store jitted code in memfd mappings to avoid having rwx
mappings. These processes map the code with a writeable mapping and a
read-execute mapping. They write the code using the writeable mapping
and then unmap the writeable mapping. All subsequent execution is
through the read-execute mapping.
perf inject --jit ignores //anon* mappings for each process where a
jitdump is present because it expects to inject mmap events for each
jitted code range, and said jitted code ranges will overlap with the
//anon* mappings.
Ignore /memfd: and [anon:* mappings so that jitted code contained in
/memfd: and [anon:* mappings is treated the same way as jitted code
contained in //anon* mappings.
Signed-off-by: Brian Robbins <brianrob@linux.microsoft.com>
Acked-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.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/20220805220645.95855-1-brianrob@linux.microsoft.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
First noticed with fedora:rawhide:
48 11.10 fedora:rawhide : FAIL gcc version 12.1.1 20220628 (Red Hat 12.1.1-3) (GCC)
util/scripting-engines/trace-event-python.c: In function 'python_start_script':
util/scripting-engines/trace-event-python.c:1899:9: error: 'PySys_SetArgv' is deprecated [-Werror=deprecated-declarations]
1899 | PySys_SetArgv(argc + 1, command_line);
No time now to address this warning, so don't make it an error, in time
we should either add yet more ifdefs to continue supporting older
systems or just convert to whatever new infra python put in place for
argv processing, sigh.
Acked-by: Ian Rogers <irogers@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Move print_*_events functions out of parse-events.c into a new
print-events.c. Move tracepoint code into tracepoint.c or
trace-event-info.c (sole user). This reduces the dependencies of
parse-events.c and makes it more amenable to being a library in the
future.
Remove some unnecessary definitions from parse-events.h. Fix a
checkpatch.pl warning on using unsigned rather than unsigned int. Fix
some line length warnings too.
Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.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>
Cc: Stephane Eranian <eranian@google.com>
Link: https://lore.kernel.org/r/20220729204217.250166-3-irogers@google.com
[ Add include linux/stddef.h before perf_events.h for systems where __always_inline isn't pulled in before used, such as older Alpine Linux ]
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
binutils changed the signature of init_disassemble_info(), which now causes
compilation failures for tools/perf/util/annotate.c, e.g. on debian
unstable.
Relevant binutils commit:
https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=60a3da00bd5407f07
Wire up the feature test and switch to init_disassemble_info_compat(),
which were introduced in prior commits, fixing the compilation failure.
I verified that perf can still disassemble bpf programs by using bpftrace
under load, recording a perf trace, and then annotating the bpf "function"
with and without the changes. With old binutils there's no change in output
before/after this patch. When comparing the output from old binutils (2.35)
to new bintuils with the patch (upstream snapshot) there are a few output
differences, but they are unrelated to this patch. An example hunk is:
1.15 : 55:mov %rbp,%rdx
0.00 : 58:add $0xfffffffffffffff8,%rdx
0.00 : 5c:xor %ecx,%ecx
- 1.03 : 5e:callq 0xffffffffe12aca3c
+ 1.03 : 5e:call 0xffffffffe12aca3c
0.00 : 63:xor %eax,%eax
- 2.18 : 65:leaveq
- 2.82 : 66:retq
+ 2.18 : 65:leave
+ 2.82 : 66:ret
Signed-off-by: Andres Freund <andres@anarazel.de>
Acked-by: Quentin Monnet <quentin@isovalent.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Ben Hutchings <benh@debian.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Sedat Dilek <sedat.dilek@gmail.com>
Cc: bpf@vger.kernel.org
Link: http://lore.kernel.org/lkml/20220622181918.ykrs5rsnmx3og4sv@alap3.anarazel.de
Link: https://lore.kernel.org/r/20220801013834.156015-5-andres@anarazel.de
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Some functions we use for bpf prologue generation are going to be
deprecated. This change reworks current code not to use them.
We need to replace following functions/struct:
bpf_program__set_prep
bpf_program__nth_fd
struct bpf_prog_prep_result
Currently we use bpf_program__set_prep to hook perf callback before
program is loaded and provide new instructions with the prologue.
We replace this function/ality by taking instructions for specific
program, attaching prologue to them and load such new ebpf programs
with prologue using separate bpf_prog_load calls (outside libbpf
load machinery).
Before we can take and use program instructions, we need libbpf to
actually load it. This way we get the final shape of its instructions
with all relocations and verifier adjustments).
There's one glitch though.. perf kprobe program already assumes
generated prologue code with proper values in argument registers,
so loading such program directly will fail in the verifier.
That's where the fallback pre-load handler fits in and prepends
the initialization code to the program. Once such program is loaded
we take its instructions, cut off the initialization code and prepend
the prologue.
I know.. sorry ;-)
To have access to the program when loading this patch adds support to
register 'fallback' section handler to take care of perf kprobe programs.
The fallback means that it handles any section definition besides the
ones that libbpf handles.
The handler serves two purposes:
- allows perf programs to have special arguments in section name
- allows perf to use pre-load callback where we can attach init
code (zeroing all argument registers) to each perf program
Suggested-by: Andrii Nakryiko <andrii@kernel.org>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <songliubraving@fb.com>
Cc: Yonghong Song <yhs@fb.com>
Cc: bpf@vger.kernel.org
Cc: netdev@vger.kernel.org
Link: https://lore.kernel.org/r/20220616202214.70359-2-jolsa@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
The perf jvmti agent doesn't create program headers, in this case
fallback on section headers as happened previously.
Committer notes:
To test this, from a public post by Ian:
1) download a Java workload dacapo-9.12-MR1-bach.jar from
https://sourceforge.net/projects/dacapobench/
2) build perf such as "make -C tools/perf O=/tmp/perf NO_LIBBFD=1" it
should detect Java and create /tmp/perf/libperf-jvmti.so
3) run perf with the jvmti agent:
perf record -k 1 java -agentpath:/tmp/perf/libperf-jvmti.so -jar dacapo-9.12-MR1-bach.jar -n 10 fop
4) run perf inject:
perf inject -i perf.data -o perf-injected.data -j
5) run perf report
perf report -i perf-injected.data | grep org.apache.fop
With this patch reverted I see lots of symbols like:
0.00% java jitted-388040-4656.so [.] org.apache.fop.fo.FObj.bind(org.apache.fop.fo.PropertyList)
With the patch (2d86612aac ("perf symbol: Correct address for bss
symbols")) I see lots of:
dso__load_sym_internal: failed to find program header for symbol:
Lorg/apache/fop/fo/FObj;bind(Lorg/apache/fop/fo/PropertyList;)V
st_value: 0x40
Fixes: 2d86612aac ("perf symbol: Correct address for bss symbols")
Reviewed-by: Leo Yan <leo.yan@linaro.org>
Signed-off-by: Ian Rogers <irogers@google.com>
Tested-by: Leo Yan <leo.yan@linaro.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Leo Yan <leo.yan@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lore.kernel.org/lkml/20220731164923.691193-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Add -b/--use-bpf option to use BPF to collect lock contention stats.
For simplicity it now runs system-wide and requires C-c to stop.
Upcoming changes will add the usual filtering.
$ sudo perf lock con -b
^C
contended total wait max wait avg wait type caller
42 192.67 us 13.64 us 4.59 us spinlock queue_work_on+0x20
23 85.54 us 10.28 us 3.72 us spinlock worker_thread+0x14a
6 13.92 us 6.51 us 2.32 us mutex kernfs_iop_permission+0x30
3 11.59 us 10.04 us 3.86 us mutex kernfs_dop_revalidate+0x3c
1 7.52 us 7.52 us 7.52 us spinlock kthread+0x115
1 7.24 us 7.24 us 7.24 us rwlock:W sys_epoll_wait+0x148
2 7.08 us 3.99 us 3.54 us spinlock delayed_work_timer_fn+0x1b
1 6.41 us 6.41 us 6.41 us spinlock idle_balance+0xa06
2 2.50 us 1.83 us 1.25 us mutex kernfs_iop_lookup+0x2f
1 1.71 us 1.71 us 1.71 us mutex kernfs_iop_getattr+0x2c
Signed-off-by: Namhyung Kim <namhyung@kernel.org>
Cc: Blake Jones <blakejones@google.com>
Cc: Boqun Feng <boqun.feng@gmail.com>
Cc: Davidlohr Bueso <dave@stgolabs.net>
Cc: Ian Rogers <irogers@google.com>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Waiman Long <longman@redhat.com>
Cc: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20220729200756.666106-3-namhyung@kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Current perf stat uses the evlist__add_default_attrs() to add the
generic default attrs, and uses arch_evlist__add_default_attrs() to add
the Arch specific default attrs, e.g., Topdown for x86.
It works well for the non-hybrid platforms. However, for a hybrid
platform, the hard code generic default attrs don't work.
Uses arch_evlist__add_default_attrs() to replace the
evlist__add_default_attrs(). The arch_evlist__add_default_attrs() is
modified to invoke the same __evlist__add_default_attrs() for the
generic default attrs. No functional change.
Add default_null_attrs[] to indicate the arch specific attrs.
No functional change for the arch specific default attrs either.
Signed-off-by: Kan Liang <kan.liang@linux.intel.com>
Acked-by: Ian Rogers <irogers@google.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220721065706.2886112-4-zhengjun.xing@linux.intel.com
Signed-off-by: Xing Zhengjun <zhengjun.xing@linux.intel.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
When using 'perf mem' and 'perf c2c', an issue is observed that tool
reports the wrong offset for global data symbols. This is a common
issue on both x86 and Arm64 platforms.
Let's see an example, for a test program, below is the disassembly for
its .bss section which is dumped with objdump:
...
Disassembly of section .bss:
0000000000004040 <completed.0>:
...
0000000000004080 <buf1>:
...
00000000000040c0 <buf2>:
...
0000000000004100 <thread>:
...
First we used 'perf mem record' to run the test program and then used
'perf --debug verbose=4 mem report' to observe what's the symbol info
for 'buf1' and 'buf2' structures.
# ./perf mem record -e ldlat-loads,ldlat-stores -- false_sharing.exe 8
# ./perf --debug verbose=4 mem report
...
dso__load_sym_internal: adjusting symbol: st_value: 0x40c0 sh_addr: 0x4040 sh_offset: 0x3028
symbol__new: buf2 0x30a8-0x30e8
...
dso__load_sym_internal: adjusting symbol: st_value: 0x4080 sh_addr: 0x4040 sh_offset: 0x3028
symbol__new: buf1 0x3068-0x30a8
...
The perf tool relies on libelf to parse symbols, in executable and
shared object files, 'st_value' holds a virtual address; 'sh_addr' is
the address at which section's first byte should reside in memory, and
'sh_offset' is the byte offset from the beginning of the file to the
first byte in the section. The perf tool uses below formula to convert
a symbol's memory address to a file address:
file_address = st_value - sh_addr + sh_offset
^
` Memory address
We can see the final adjusted address ranges for buf1 and buf2 are
[0x30a8-0x30e8) and [0x3068-0x30a8) respectively, apparently this is
incorrect, in the code, the structure for 'buf1' and 'buf2' specifies
compiler attribute with 64-byte alignment.
The problem happens for 'sh_offset', libelf returns it as 0x3028 which
is not 64-byte aligned, combining with disassembly, it's likely libelf
doesn't respect the alignment for .bss section, therefore, it doesn't
return the aligned value for 'sh_offset'.
Suggested by Fangrui Song, ELF file contains program header which
contains PT_LOAD segments, the fields p_vaddr and p_offset in PT_LOAD
segments contain the execution info. A better choice for converting
memory address to file address is using the formula:
file_address = st_value - p_vaddr + p_offset
This patch introduces elf_read_program_header() which returns the
program header based on the passed 'st_value', then it uses the formula
above to calculate the symbol file address; and the debugging log is
updated respectively.
After applying the change:
# ./perf --debug verbose=4 mem report
...
dso__load_sym_internal: adjusting symbol: st_value: 0x40c0 p_vaddr: 0x3d28 p_offset: 0x2d28
symbol__new: buf2 0x30c0-0x3100
...
dso__load_sym_internal: adjusting symbol: st_value: 0x4080 p_vaddr: 0x3d28 p_offset: 0x2d28
symbol__new: buf1 0x3080-0x30c0
...
Fixes: f17e04afaf ("perf report: Fix ELF symbol parsing")
Reported-by: Chang Rui <changruinj@gmail.com>
Suggested-by: Fangrui Song <maskray@google.com>
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Acked-by: Namhyung Kim <namhyung@kernel.org>
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: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220724060013.171050-2-leo.yan@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Implements workqueue trace bpf function.
Test cases:
# perf kwork -k workqueue lat -b
Starting trace, Hit <Ctrl+C> to stop and report
^C
Kwork Name | Cpu | Avg delay | Count | Max delay | Max delay start | Max delay end |
--------------------------------------------------------------------------------------------------------------------------------
(w)addrconf_verify_work | 0002 | 5.856 ms | 1 | 5.856 ms | 111994.634313 s | 111994.640169 s |
(w)vmstat_update | 0001 | 1.247 ms | 1 | 1.247 ms | 111996.462651 s | 111996.463899 s |
(w)neigh_periodic_work | 0001 | 1.183 ms | 1 | 1.183 ms | 111996.462789 s | 111996.463973 s |
(w)neigh_managed_work | 0001 | 0.989 ms | 2 | 1.635 ms | 111996.462820 s | 111996.464455 s |
(w)wb_workfn | 0000 | 0.667 ms | 1 | 0.667 ms | 111996.384273 s | 111996.384940 s |
(w)bpf_prog_free_deferred | 0001 | 0.495 ms | 1 | 0.495 ms | 111986.314201 s | 111986.314696 s |
(w)mix_interrupt_randomness | 0002 | 0.421 ms | 6 | 0.749 ms | 111995.927750 s | 111995.928499 s |
(w)vmstat_shepherd | 0000 | 0.374 ms | 2 | 0.385 ms | 111991.265242 s | 111991.265627 s |
(w)e1000_watchdog | 0002 | 0.356 ms | 5 | 0.390 ms | 111994.528380 s | 111994.528770 s |
(w)vmstat_update | 0000 | 0.231 ms | 2 | 0.365 ms | 111996.384407 s | 111996.384772 s |
(w)flush_to_ldisc | 0006 | 0.165 ms | 1 | 0.165 ms | 111995.930606 s | 111995.930771 s |
(w)flush_to_ldisc | 0000 | 0.094 ms | 2 | 0.095 ms | 111996.460453 s | 111996.460548 s |
--------------------------------------------------------------------------------------------------------------------------------
# perf kwork -k workqueue rep -b
Starting trace, Hit <Ctrl+C> to stop and report
^C
Kwork Name | Cpu | Total Runtime | Count | Max runtime | Max runtime start | Max runtime end |
--------------------------------------------------------------------------------------------------------------------------------
(w)e1000_watchdog | 0002 | 0.627 ms | 2 | 0.324 ms | 112002.720665 s | 112002.720989 s |
(w)flush_to_ldisc | 0007 | 0.598 ms | 2 | 0.534 ms | 112000.875226 s | 112000.875761 s |
(w)wq_barrier_func | 0007 | 0.492 ms | 1 | 0.492 ms | 112000.876981 s | 112000.877473 s |
(w)flush_to_ldisc | 0007 | 0.281 ms | 1 | 0.281 ms | 112005.826882 s | 112005.827163 s |
(w)mix_interrupt_randomness | 0002 | 0.229 ms | 3 | 0.102 ms | 112005.825671 s | 112005.825774 s |
(w)vmstat_shepherd | 0000 | 0.202 ms | 1 | 0.202 ms | 112001.504511 s | 112001.504713 s |
(w)bpf_prog_free_deferred | 0001 | 0.181 ms | 1 | 0.181 ms | 112000.883251 s | 112000.883432 s |
(w)wb_workfn | 0007 | 0.130 ms | 1 | 0.130 ms | 112001.505195 s | 112001.505325 s |
(w)vmstat_update | 0000 | 0.053 ms | 1 | 0.053 ms | 112001.504763 s | 112001.504815 s |
--------------------------------------------------------------------------------------------------------------------------------
Signed-off-by: Yang Jihong <yangjihong1@huawei.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.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: Paul Clarke <pc@us.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: https://lore.kernel.org/r/20220709015033.38326-18-yangjihong1@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
