Skip to content

K
Kernel

openeuler / Kernel
接近 2 年 前同步成功

通知 8
Star 0
Fork 0
K
Kernel
切换分支/标签
  1. 08 9月, 2016 2 次提交
  3. 05 9月, 2016 3 次提交
  5. 01 9月, 2016 10 次提交
    • R
      perf probe: Move dwarf specific functions to dwarf-aux.c · 6243b9dc
      由 Ravi Bangoria 提交于 
      Move generic dwarf related functions from util/probe-finder.c to
util/dwarf-aux.c. Functions name and their prototype are also changed
accordingly. No functionality changes.
Suggested-and-Acked-by: NMasami Hiramatsu <mhiramat@kernel.org>
Signed-off-by: NRavi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Hemant Kumar <hemant@linux.vnet.ibm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1472546377-25612-1-git-send-email-ravi.bangoria@linux.vnet.ibm.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      6243b9dc
    • R
      perf uprobe: Skip prologue if program compiled without optimization · e47392bf
      由 Ravi Bangoria 提交于 
      The function prologue prepares stack and registers before executing
function logic.

When target program is compiled without optimization, function parameter
information is only valid after the prologue.

When we probe entrypc of the function, and try to record a function
parameter, it contains a garbage value.

For example:

  $ vim test.c
    #include <stdio.h>

    void foo(int i)
    {
       printf("i: %d\n", i);
    }

    int main()
    {
      foo(42);
      return 0;
    }

  $ gcc -g test.c -o test
  $ objdump -dl test | less
    foo():
    /home/ravi/test.c:4
      400536:       55                      push   %rbp
      400537:       48 89 e5                mov    %rsp,%rbp
      40053a:       48 83 ec 10             sub    -bashx10,%rsp
      40053e:       89 7d fc                mov    %edi,-0x4(%rbp)
    /home/ravi/test.c:5
      400541:       8b 45 fc                mov    -0x4(%rbp),%eax
    ...
    ...
    main():
    /home/ravi/test.c:9
      400558:       55                      push   %rbp
      400559:       48 89 e5                mov    %rsp,%rbp
    /home/ravi/test.c:10
      40055c:       bf 2a 00 00 00          mov    -bashx2a,%edi
      400561:       e8 d0 ff ff ff          callq  400536 <foo>

  $ perf probe -x ./test 'foo i'
  $ cat /sys/kernel/debug/tracing/uprobe_events
     p:probe_test/foo /home/ravi/test:0x0000000000000536 i=-12(%sp):s32

  $ perf record -e probe_test:foo ./test
  $ perf script
     test  5778 [001]  4918.562027: probe_test:foo: (400536) i=0

Here variable 'i' is passed via stack which is pushed on stack at
0x40053e. But we are probing at 0x400536.

To resolve this issues, we need to probe on next instruction after
prologue.  gdb and systemtap also does same thing. I've implemented this
patch based on approach systemtap has used.

After applying patch:

  $ perf probe -x ./test 'foo i'
  $ cat /sys/kernel/debug/tracing/uprobe_events
    p:probe_test/foo /home/ravi/test:0x0000000000000541 i=-4(%bp):s32

  $ perf record -e probe_test:foo ./test
  $ perf script
    test  6300 [001]  5877.879327: probe_test:foo: (400541) i=42

No need to skip prologue for optimized case since debug info is correct
for each instructions for -O2 -g. For more details please visit:

        https://bugzilla.redhat.com/show_bug.cgi?id=612253#c6

Changes in v2:

- Skipping prologue only when any ARG is either C variable, $params or
  $vars.

- Probe on line(:1) may not be always possible. Recommend only address
  to force probe on function entry.

Committer notes:

Testing it with 'perf trace':

  # perf probe -x ./test foo i
  Added new event:
    probe_test:foo       (on foo in /home/acme/c/test with i)

  You can now use it in all perf tools, such as:

	  perf record -e probe_test:foo -aR sleep 1

  # cat /sys/kernel/debug/tracing/uprobe_events
  p:probe_test/foo /home/acme/c/test:0x0000000000000526 i=-12(%sp):s32
  # trace --no-sys --event probe_*:* ./test
  i: 42
     0.000 probe_test:foo:(400526) i=0)
  #

After the patch:

  # perf probe -d *:*
  Removed event: probe_test:foo
  # perf probe -x ./test foo i
  Target program is compiled without optimization. Skipping prologue.
  Probe on address 0x400526 to force probing at the function entry.

  Added new event:
    probe_test:foo       (on foo in /home/acme/c/test with i)

  You can now use it in all perf tools, such as:

	perf record -e probe_test:foo -aR sleep 1

  # cat /sys/kernel/debug/tracing/uprobe_events
  p:probe_test/foo /home/acme/c/test:0x0000000000000531 i=-4(%bp):s32
  # trace --no-sys --event probe_*:* ./test
  i: 42
     0.000 probe_test:foo:(400531) i=42)
  #
Reported-by: NMichael Petlan <mpetlan@redhat.com>
Report-Link: https://www.mail-archive.com/linux-perf-users@vger.kernel.org/msg02348.htmlSigned-off-by: NRavi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: NJiri Olsa <jolsa@kernel.org>
Acked-by: NMasami Hiramatsu <mhiramat@kernel.org>
Acked-by: NNaveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Hemant Kumar <hemant@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Wang Nan <wangnan0@huawei.com>
Cc: Yauheni Kaliuta <yauheni.kaliuta@redhat.com>
Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1299021
Link: http://lkml.kernel.org/r/1470214725-5023-2-git-send-email-ravi.bangoria@linux.vnet.ibm.com
[ Rename 'die' to 'cu_die' to avoid shadowing a die() definition on at least centos 5, Debian 7 and ubuntu:12.04.5]
[ Use PRIx64 instead of lx to format a Dwarf_Addr, aka long long unsigned int, fixing the build on 32-bit systems ]
[ dwarf_getsrclines() expects a size_t * argument ]
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      e47392bf
    • R
      perf probe: Add helper function to check if probe with variable · b3f33f93
      由 Ravi Bangoria 提交于 
      Introduce helper function instead of inline code and replace hardcoded
strings "$vars" and "$params" with their corresponding macros.

perf_probe_with_var() is not declared as static since it will be called
from different file in subsequent patch.
Signed-off-by: NRavi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Acked-by: NMasami Hiramatsu <mhiramat@kernel.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Hemant Kumar <hemant@linux.vnet.ibm.com>
Cc: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Wang Nan <wangnan0@huawei.com>
Link: http://lkml.kernel.org/r/1470214725-5023-1-git-send-email-ravi.bangoria@linux.vnet.ibm.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      b3f33f93
    • A
      perf symbols: Fixup symbol sizes before picking best ones · 432746f8
      由 Arnaldo Carvalho de Melo 提交于 
      When we call symbol__fixup_duplicate() we use algorithms to pick the
"best" symbols for cases where there are various functions/aliases to an
address, and those check zero size symbols, which, before calling
symbol__fixup_end() are _all_ symbols in a just parsed kallsyms file.

So first fixup the end, then fixup the duplicates.

Found while trying to figure out why 'perf test vmlinux' failed, see the
output of 'perf test -v vmlinux' to see cases where the symbols picked
as best for vmlinux don't match the ones picked for kallsyms.

Cc: Anton Blanchard <anton@samba.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Fixes: 694bf407 ("perf symbols: Add some heuristics for choosing the best duplicate symbol")
Link: http://lkml.kernel.org/n/tip-rxqvdgr0mqjdxee0kf8i2ufn@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      432746f8
    • A
      perf symbols: Check symbol_conf.allow_aliases for kallsyms loading too · c97b40e4
      由 Arnaldo Carvalho de Melo 提交于 
      We can allow aliases to be kept, but we were checking this just when
loading vmlinux files, be consistent, do it for any symbol table loading
code that calls symbol__fixup_duplicate() by making this function check
.allow_aliases instead.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: David Ahern <dsahern@gmail.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Wang Nan <wangnan0@huawei.com>
Fixes: 680d926a ("perf symbols: Allow symbol alias when loading map for symbol name")
Link: http://lkml.kernel.org/n/tip-z0avp0s6cfjckc4xj3pdfjdz@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      c97b40e4
    • M
      perf probe: Support probing on offline cross-arch binary · 293d5b43
      由 Masami Hiramatsu 提交于 
      Support probing on offline cross-architecture binary by adding getting
the target machine arch from ELF and choose correct register string for
the machine.

Here is an example:
  -----
  $ perf probe --vmlinux=./vmlinux-arm --definition 'do_sys_open $params'
  p:probe/do_sys_open do_sys_open+0 dfd=%r5:s32 filename=%r1:u32 flags=%r6:s32 mode=%r3:u16
  -----

Here, we can get probe/do_sys_open from above and append it to to the target
machine's tracing/kprobe_events file in the tracefs mountput, usually
/sys/kernel/debug/tracing/kprobe_events (or /sys/kernel/tracing/kprobe_events).
Signed-off-by: NMasami Hiramatsu <mhiramat@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/147214229717.23638.6440579792548044658.stgit@devbox
[ Add definition for EM_AARCH64 to fix the build on at least centos 6, debian 7 & ubuntu 12.04.5 ]
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      293d5b43
    • M
      perf probe: Ignore vmlinux buildid if offline kernel is given · 428aff82
      由 Masami Hiramatsu 提交于 
      Ignore the buildid of running kernel when both of --definition and
--vmlinux is given because that kernel should be off-line.

This also skips post-processing of kprobe event for relocating symbol
and checking blacklist, because it can not be done on off-line kernel.

E.g. without this fix perf shows an error as below
  ----
  $ perf probe --vmlinux=./vmlinux-arm --definition do_sys_open
  ./vmlinux-arm with build id 7a1f76dd56e9c4da707cd3d6333f50748141434b not found, continuing without symbols
  Failed to find symbol do_sys_open in kernel
    Error: Failed to add events.
  ----
with this fix, we can get the definition
  ----
  $ perf probe --vmlinux=./vmlinux-arm --definition do_sys_open
  p:probe/do_sys_open do_sys_open+0
  ----
Signed-off-by: NMasami Hiramatsu <mhiramat@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/147214228193.23638.12581984840822162131.stgit@devboxSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      428aff82
    • M
      perf probe: Show trace event definition · 1c20b1d1
      由 Masami Hiramatsu 提交于 
      Add --definition/-D option for showing the trace-event definition in
stdout. This can be useful in debugging or combined with a shell script.

e.g.
  ----
  # perf probe --definition 'do_sys_open $params'
  p:probe/do_sys_open _text+2261728 dfd=%di:s32 filename=%si:u64 flags=%dx:s32 mode=%cx:u16
  ----
Suggested-and-Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NMasami Hiramatsu <mhiramat@kernel.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/147214226712.23638.2240534040014013658.stgit@devboxSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      1c20b1d1
    • M
      perf symbols: Demangle symbols for synthesized @plt entries. · 2a8d41b4
      由 Milian Wolff 提交于 
      The symbols in the synthesized @plt entries where not demangled before,
i.e. we could end up with entries such as:

    $ perf report
    Samples: 7K of event 'cycles:ppp', Event count (approx.): 6223833141
    Children      Self  Command          Shared Object           Symbol
    -   93.63%    28.89%  lab_mandelbrot   lab_mandelbrot        [.] main
        - 73.81% main
            - 33.57% hypot
              27.76% __hypot_finite
              15.97% __muldc3
               2.90% __muldc3@plt
               2.40% _ZNK6QImage6heightEv@plt
             + 2.14% QColor::rgb
               1.94% _ZNK6QImage5widthEv@plt
               1.92% cabs@plt

This patch remedies this issue by also applying demangling to the
synthesized symbols. The output for the above is now:

    $ perf report
    Samples: 7K of event 'cycles:ppp', Event count (approx.): 6223833141
    Children      Self  Command          Shared Object           Symbol
    -   93.63%    28.89%  lab_mandelbrot   lab_mandelbrot        [.] main
        - 73.81% main
            - 33.57% hypot
              27.76% __hypot_finite
              15.97% __muldc3
               2.90% __muldc3@plt
               2.40% QImage::height() const@plt
             + 2.14% QColor::rgb
               1.94% QImage::width() const@plt
               1.92% cabs@plt
Signed-off-by: NMilian Wolff <milian.wolff@kdab.com>
LPU-Reference: 20160830114102.30863-1-milian.wolff@kdab.com
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      2a8d41b4
    • A
      perf probe: Do not use map_load filters for function · fd227598
      由 Arnaldo Carvalho de Melo 提交于 
      It is simpler to just do the loop, no need for globals and the last user
of such facility disappears.

Testing:

  # perf probe -F [a-z]*recvmsg
  aead_recvmsg
  compat_SyS_recvmsg
  compat_sys_recvmsg
  hash_recvmsg
  inet_recvmsg
  kernel_recvmsg
  netlink_recvmsg
  packet_recvmsg
  ping_recvmsg
  raw_recvmsg
  rawv6_recvmsg
  rng_recvmsg
  security_socket_recvmsg
  selinux_socket_recvmsg
  skcipher_recvmsg
  sock_common_recvmsg
  sock_no_recvmsg
  sock_recvmsg
  sys_recvmsg
  tcp_recvmsg
  udp_recvmsg
  udpv6_recvmsg
  unix_dgram_recvmsg
  unix_seqpacket_recvmsg
  unix_stream_recvmsg
  #

Without filters:

  # perf probe -F | tail -5
  zswap_pool_create
  zswap_pool_current
  zswap_update_total_size
  zswap_writeback_entry
  zswap_zpool_param_set
  #
  # perf probe -F | wc -l
  33311
  #
Acked-by: NMasami Hiramatsu <mhiramat@kernel.org>
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/r/20160831130427.GA13095@kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      fd227598
  7. 30 8月, 2016 2 次提交
  9. 24 8月, 2016 16 次提交
  11. 19 8月, 2016 1 次提交
    • A
      perf evsel: Do not access outside hw cache name arrays · c53412ee
      由 Arnaldo Carvalho de Melo 提交于 
      We have to check if the values are >= *_MAX, not just >, fix it.

From the bugzilla report:

''In file /tools/perf/util/evsel.c  function __perf_evsel__hw_cache_name
it appears that there is a bug that reads beyond the end of the buffer.
The statement "if (type > PERF_COUNT_HW_CACHE_MAX)" allows type to be
equal to the maximum value. Later, when statement "if
(!perf_evsel__is_cache_op_valid(type, op))" is executed, the function
can access array perf_evsel__hw_cache_stat[type] beyond the end of the
buffer.

It appears to me that the statement "if (type > PERF_COUNT_HW_CACHE_MAX)"
should be "if (type >= PERF_COUNT_HW_CACHE_MAX)"

Bug found with Coverity and manual code review. No attempts were made to
execute the code with a maximum type value.''

Committer note:

Testing it:

  $ perf record -e $(echo $(perf list cache | cut -d \[ -f1) | sed 's/ /,/g') usleep 1
  [ perf record: Woken up 16 times to write data ]
  [ perf record: Captured and wrote 0.023 MB perf.data (34 samples) ]
  $ perf evlist
  L1-dcache-load-misses
  L1-dcache-loads
  L1-dcache-stores
  L1-icache-load-misses
  LLC-load-misses
  LLC-loads
  LLC-store-misses
  LLC-stores
  branch-load-misses
  branch-loads
  dTLB-load-misses
  dTLB-loads
  dTLB-store-misses
  dTLB-stores
  iTLB-load-misses
  iTLB-loads
  node-load-misses
  node-loads
  node-store-misses
  node-stores
  $ perf list cache

  List of pre-defined events (to be used in -e):

    L1-dcache-load-misses        [Hardware cache event]
    L1-dcache-loads              [Hardware cache event]
    L1-dcache-stores             [Hardware cache event]
    L1-icache-load-misses        [Hardware cache event]
    LLC-load-misses              [Hardware cache event]
    LLC-loads                    [Hardware cache event]
    LLC-store-misses             [Hardware cache event]
    LLC-stores                   [Hardware cache event]
    branch-load-misses           [Hardware cache event]
    branch-loads                 [Hardware cache event]
    dTLB-load-misses             [Hardware cache event]
    dTLB-loads                   [Hardware cache event]
    dTLB-store-misses            [Hardware cache event]
    dTLB-stores                  [Hardware cache event]
    iTLB-load-misses             [Hardware cache event]
    iTLB-loads                   [Hardware cache event]
    node-load-misses             [Hardware cache event]
    node-loads                   [Hardware cache event]
    node-store-misses            [Hardware cache event]
    node-stores                  [Hardware cache event]
  $
Reported-by: NBrian Sweeney <bsweeney@lgsinnovations.com>
Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=153351Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      c53412ee
  13. 17 8月, 2016 1 次提交
    • M
      perf unwind: Use addr_location::addr instead of ip for entries · 67540759
      由 Milian Wolff 提交于 
      This fixes the srcline translation for call chains of user space
applications.

Before we got:

    perf report --stdio --no-children -s sym,srcline -g address
     8.92%  [.] main                                 mandelbrot.h:41
            |
            |--3.70%--main +8390240
            |          __libc_start_main +139950056726769
            |          _start +8388650
            |
            |--2.74%--main +8390189
            |
             --2.08%--main +8390296
                       __libc_start_main +139950056726769
                       _start +8388650

     7.59%  [.] main                                 complex:1326
            |
            |--4.79%--main +8390203
            |          __libc_start_main +139950056726769
            |          _start +8388650
            |
             --2.80%--main +8390219

     7.12%  [.] __muldc3                             libgcc2.c:1945
            |
            |--3.76%--__muldc3 +139950060519490
            |          main +8390224
            |          __libc_start_main +139950056726769
            |          _start +8388650
            |
             --3.32%--__muldc3 +139950060519512
                       main +8390224

With this patch applied, we instead get:

    perf report --stdio --no-children -s sym,srcline -g address
     8.92%  [.] main                                 mandelbrot.h:41
            |
            |--3.70%--main mandelbrot.h:41
            |          __libc_start_main +241
            |          _start +4194346
            |
            |--2.74%--main mandelbrot.h:41
            |
             --2.08%--main mandelbrot.h:41
                       __libc_start_main +241
                       _start +4194346

     7.59%  [.] main                                 complex:1326
            |
            |--4.79%--main complex:1326
            |          __libc_start_main +241
            |          _start +4194346
            |
             --2.80%--main complex:1326

     7.12%  [.] __muldc3                             libgcc2.c:1945
            |
            |--3.76%--__muldc3 libgcc2.c:1945
            |          main mandelbrot.h:39
            |          __libc_start_main +241
            |          _start +4194346
            |
             --3.32%--__muldc3 libgcc2.c:1945
                       main mandelbrot.h:39
Suggested-and-Acked-by: NNamhyung Kim <namhyung@kernel.org>
Signed-off-by: NMilian Wolff <milian.wolff@kdab.com>
Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
LPU-Reference: 20160816153926.11288-1-milian.wolff@kdab.com
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      67540759
  15. 16 8月, 2016 4 次提交
  17. 13 8月, 2016 1 次提交