Skip to content

R
raspberrypi-kernel

openeuler / raspberrypi-kernel

通知 14
Star 1
Fork 0
R
raspberrypi-kernel
切换分支/标签
  1. 27 12月, 2019 2 次提交
    • J
      perf annotate: Remove hist__account_cycles() from callback · f83e4568
      由 Jin Yao 提交于 
      mainline inclusion
from mainline-5.2-rc1
commit bdd1666b3d03
category: bugfix
bugzilla: 16426
CVE: NA

-------------------------------------------------

The hist__account_cycles() function is executed when the
hist_iter__branch_callback() is called.

But it looks it's not necessary.  In hist__account_cycles, it already
walks on all branch entries.

This patch moves the hist__account_cycles out of callback, now the data
processing is much faster than before.

Previous code has an issue that the ch[offset].num++ (in
__symbol__account_cycles) is executed repeatedly since
hist__account_cycles is called in each hist_iter__branch_callback, so
the counting of ch[offset].num is not correct (too big).

With this patch, the issue is fixed. And we don't need the code of
"ch->reset >= ch->num / 2" to check if there are too many overlaps (in
annotation__count_and_fill), otherwise some data would be hidden.

Now, we can try, for example:

  perf record -b ...
  perf annotate or perf report -s symbol

The before/after output should be no change.

 v3:
 ---
 Fix the crash in stdio mode.
 Like previous code, it needs the checking of ui__has_annotation()
 before hist__account_cycles()

 v2:
 ---
 1. Cover the similar perf report
 2. Remove the checking code "ch->reset >= ch->num / 2"
Signed-off-by: NJin Yao <yao.jin@linux.intel.com>
Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1552684577-29041-1-git-send-email-yao.jin@linux.intel.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: NWei Li <liwei391@huawei.com>
Reviewed-by: NJian Cheng <cj.chengjian@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
      f83e4568
    • W
      perf: fix incorrect change in annotate result · 7e5c7786
      由 Wei Li 提交于 
      hulk inclusion
category: bugfix
bugzilla: 9583
CVE: NA

-------------------------------------------------

The output of "perf annotate -l --stdio xxx" changed since commit 425859ff ("perf
annotate: No need to calculate notes->start twice") removed notes->start assignment
in symbol__calc_lines(). It will get failed in find_address_in_section() from
symbol__tty_annotate() subroutine as the a2l->addr is wrong. So the annotate summary
doesn't report the line number of source code correctly.

Before fix:

[root@localhost tools]# ./perf/perf record -o perf.data /root/common_while_1
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.261 MB perf.data (6538 samples) ]

[root@localhost tools]# ./perf/perf annotate -i perf.data -l -s hotspot_1 --stdio | less

Sorted summary for file /root/common_while_1
----------------------------------------------

   32.53 common_while_1[3c]
   32.26 common_while_1[14]
   15.59 common_while_1[38]
   14.89 common_while_1[2c]
    2.48 common_while_1[48]
    1.10 common_while_1[c]
    0.58 common_while_1[8]
 Percent |      Source code & Disassembly of common_while_1 for cycles:ppp (2457 samples, percent: local period)
----------------------------------------------------------------------------------------------------------------
         :
         :
         :
         :            Disassembly of section .text:
         :
         :            00000000004008b0 <hotspot_1>:
         :            hotspot_1():
         :                for(i = 0; i<=100; i++) {
         :                    s_cnt2++;
         :                }
         :            }
         :
         :            void hotspot_1() {
    0.00 :   4008b0:       stp     x29, x30, [sp,#-32]!
    0.00 :   4008b4:       mov     x29, sp
         :                int i;
         :                while(1) {
         :                    hotspot_2();
 common_while_1[8]    0.58 :   4008b8:       bl      400864 <hotspot_2>
         :                    for(i = 0; i<=100; i++) {
 common_while_1[c]    1.10 :   4008bc:       str     wzr, [x29,#28]
    0.00 :   4008c0:       b       4008ec <hotspot_1+0x3c>
         :                        s_cnt1++;
 common_while_1[14]   32.26 :   4008c4:       adrp    x0, 420000 <fclose@GLIBC_2.17>
    0.00 :   4008c8:       add     x0, x0, #0x7c
    0.16 :   4008cc:       ldr     w0, [x0]
    0.00 :   4008d0:       add     w1, w0, #0x1
...

After fix:

[root@localhost tools]# ./perf/perf record -o perf.data /root/common_while_1
^C[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.261 MB perf.data (6538 samples) ]

[root@localhost tools]# ./perf/perf annotate -i perf.data -l -s hotspot_1 --stdio | less

Sorted summary for file /root/common_while_1
----------------------------------------------

   50.27 common_while_1.c:37
   46.46 common_while_1.c:38
    2.79 common_while_1.c:36
 Percent |      Source code & Disassembly of common_while_1 for cycles:ppp (3149 samples, percent: local period)
----------------------------------------------------------------------------------------------------------------
         :
         :
         :
         :            Disassembly of section .text:
         :
         :            00000000004008b0 <hotspot_1>:
         :            hotspot_1():
         :                for(i = 0; i<=100; i++) {
         :                    s_cnt2++;
         :                }
         :            }
         :
         :            void hotspot_1() {
    0.00 :   4008b0:       stp     x29, x30, [sp,#-32]!
    0.00 :   4008b4:       mov     x29, sp
         :                int i;
         :                while(1) {
         :                    hotspot_2();
 common_while_1.c:36    0.82 :   4008b8:       bl      400864 <hotspot_2>
         :                    for(i = 0; i<=100; i++) {
 common_while_1.c:37    1.17 :   4008bc:       str     wzr, [x29,#28]
    0.00 :   4008c0:       b       4008ec <hotspot_1+0x3c>
         :                        s_cnt1++;
 common_while_1.c:38   31.33 :   4008c4:       adrp    x0, 420000 <fclose@GLIBC_2.17>
    0.00 :   4008c8:       add     x0, x0, #0x7c
    0.19 :   4008cc:       ldr     w0, [x0]
    0.00 :   4008d0:       add     w1, w0, #0x1
...

Fixes: 425859ff ("perf annotate: No need to calculate notes->start twice")
Signed-off-by: NWei Li <liwei391@huawei.com>
Signed-off-by: NWei Li <liwei391@huawei.com>
Reviewed-by: NLi Bin <huawei.libin@huawei.com>
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
      7e5c7786
  3. 31 8月, 2018 2 次提交
    • K
      perf annotate: Fix parsing aarch64 branch instructions after objdump update · 4e67b2a5
      由 Kim Phillips 提交于 
      Starting with binutils 2.28, aarch64 objdump adds comments to the
disassembly output to show the alternative names of a condition code
[1].

It is assumed that commas in objdump comments could occur in other
arches now or in the future, so this fix is arch-independent.

The fix could have been done with arm64 specific jump__parse and
jump__scnprintf functions, but the jump__scnprintf instruction would
have to have its comment character be a literal, since the scnprintf
functions cannot receive a struct arch easily.

This inconvenience also applies to the generic jump__scnprintf, which is
why we add a raw_comment pointer to struct ins_operands, so the __parse
function assigns it to be re-used by its corresponding __scnprintf
function.

Example differences in 'perf annotate --stdio2' output on an aarch64
perf.data file:

BEFORE: → b.cs   ffff200008133d1c <unwind_frame+0x18c>  // b.hs, dffff7ecc47b
AFTER : ↓ b.cs   18c

BEFORE: → b.cc   ffff200008d8d9cc <get_alloc_profile+0x31c>  // b.lo, b.ul, dffff727295b
AFTER : ↓ b.cc   31c

The branch target labels 18c and 31c also now appear in the output:

BEFORE:        add    x26, x29, #0x80
AFTER : 18c:   add    x26, x29, #0x80

BEFORE:        add    x21, x21, #0x8
AFTER : 31c:   add    x21, x21, #0x8

The Fixes: tag below is added so stable branches will get the update; it
doesn't necessarily mean that commit was broken at the time, rather it
didn't withstand the aarch64 objdump update.

Tested no difference in output for sample x86_64, power arch perf.data files.

[1] https://sourceware.org/git/gitweb.cgi?p=binutils-gdb.git;a=commit;h=bb7eff5206e4795ac79c177a80fe9f4630aaf730Signed-off-by: NKim Phillips <kim.phillips@arm.com>
Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Anton Blanchard <anton@samba.org>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ravi Bangoria <ravi.bangoria@linux.vnet.ibm.com>
Cc: Robin Murphy <robin.murphy@arm.com>
Cc: Taeung Song <treeze.taeung@gmail.com>
Cc: linux-arm-kernel@lists.infradead.org
Fixes: b13bbeee ("perf annotate: Fix branch instruction with multiple operands")
Link: http://lkml.kernel.org/r/20180827125340.a2f7e291901d17cea05daba4@arm.comSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      4e67b2a5
    • M
      perf annotate: Properly interpret indirect call · 1dc27f63
      由 Martin Liška 提交于 
      The patch changes the parsing of:

	callq  *0x8(%rbx)

from:

  0.26 │     → callq  *8

to:

  0.26 │     → callq  *0x8(%rbx)

in this case an address is followed by a register, thus one can't parse
only the address.

Committer testing:

1) run 'perf record sleep 10'
2) before applying the patch, run:

     perf annotate --stdio2 > /tmp/before

3) after applying the patch, run:

     perf annotate --stdio2 > /tmp/after

4) diff /tmp/before /tmp/after:
  --- /tmp/before 2018-08-28 11:16:03.238384143 -0300
  +++ /tmp/after  2018-08-28 11:15:39.335341042 -0300
  @@ -13274,7 +13274,7 @@
                ↓ jle    128
                  hash_value = hash_table->hash_func (key);
                  mov    0x8(%rsp),%rdi
  -  0.91       → callq  *30
  +  0.91       → callq  *0x30(%r12)
                  mov    $0x2,%r8d
                  cmp    $0x2,%eax
                  node_hash = hash_table->hashes[node_index];
  @@ -13848,7 +13848,7 @@
                   mov    %r14,%rdi
                   sub    %rbx,%r13
                   mov    %r13,%rdx
  -              → callq  *38
  +              → callq  *0x38(%r15)
                   cmp    %rax,%r13
     1.91        ↓ je     240
            1b4:   mov    $0xffffffff,%r13d
  @@ -14026,7 +14026,7 @@
                   mov    %rcx,-0x500(%rbp)
                   mov    %r15,%rsi
                   mov    %r14,%rdi
  -              → callq  *38
  +              → callq  *0x38(%rax)
                   mov    -0x500(%rbp),%rcx
                   cmp    %rax,%rcx
                 ↓ jne    9b0
<SNIP tons of other such cases>
Signed-off-by: NMartin Liška <mliska@suse.cz>
Tested-by: NArnaldo Carvalho de Melo <acme@redhat.com>
Tested-by: NKim Phillips <kim.phillips@arm.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Link: http://lkml.kernel.org/r/bd1f3932-be2b-85f9-7582-111ee0a43b07@suse.czSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      1dc27f63
  5. 20 8月, 2018 1 次提交
  7. 09 8月, 2018 17 次提交
  9. 06 6月, 2018 1 次提交
    • A
      perf annnotate: Make __symbol__inc_addr_samples handle src->histograms == NULL · 8d628d26
      由 Arnaldo Carvalho de Melo 提交于 
      Making it a bit more robust, this took place here when a sample appeared
right after:

  ffffffff8a925000 D __nosave_end

And before the next considered symbol, which, using kallsyms make us
over guess the size of __nosave_end, and then the sequence:

  hist_entry__inc_addr_samples ->
    symbol__inc_addr_samples ->
      symbol__hists ->
        annotated_source__alloc_histograms

Ends up not liking to allocate gigabytes of ram for annotation...

This will be alleviated by considering BSS symbols, which we should but
don't so far, and then we should investigate those samples further.

The testcase was to have:

   perf top -e cycles/call-graph=fp/,cache-misses/call-graph=dwarf/,instructions

Running for a while till it segfaulted trying to access NULL notes->src->histograms.

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: https://lkml.kernel.org/n/tip-ndfjtpiop3tdcnyjgp320ra8@git.kernel.orgSigned-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      8d628d26
  11. 04 6月, 2018 15 次提交
  13. 23 5月, 2018 1 次提交
  15. 19 5月, 2018 1 次提交
    • J
      perf annotate: Create hotkey 'c' to show min/max cycles · 3e71fc03
      由 Jin Yao 提交于 
      In the 'perf annotate' view, a new hotkey 'c' is created for showing the
min/max cycles.

For example, when press 'c', the annotate view is:

  Percent│ IPC     Cycle(min/max)
         │
         │
         │                             Disassembly of section .text:
         │
         │                             000000000003aab0 <random@@GLIBC_2.2.5>:
    8.22 │3.92                           sub    $0x18,%rsp
         │3.92                           mov    $0x1,%esi
         │3.92                           xor    %eax,%eax
         │3.92                           cmpl   $0x0,argp_program_version_hook@@G
         │3.92             1(2/1)      ↓ je     20
         │                               lock   cmpxchg %esi,__abort_msg@@GLIBC_P
         │                             ↓ jne    29
         │                             ↓ jmp    43
         │1.10                     20:   cmpxchg %esi,__abort_msg@@GLIBC_PRIVATE+
    8.93 │1.10             1(5/1)      ↓ je     43

When press 'c' again, the annotate view is switched back:

  Percent│ IPC Cycle
         │
         │
         │                Disassembly of section .text:
         │
         │                000000000003aab0 <random@@GLIBC_2.2.5>:
    8.22 │3.92              sub    $0x18,%rsp
         │3.92              mov    $0x1,%esi
         │3.92              xor    %eax,%eax
         │3.92              cmpl   $0x0,argp_program_version_hook@@GLIBC_2.2.5+0x
         │3.92     1      ↓ je     20
         │                  lock   cmpxchg %esi,__abort_msg@@GLIBC_PRIVATE+0x8a0
         │                ↓ jne    29
         │                ↓ jmp    43
         │1.10        20:   cmpxchg %esi,__abort_msg@@GLIBC_PRIVATE+0x8a0
    8.93 │1.10     1      ↓ je     43
Signed-off-by: NJin Yao <yao.jin@linux.intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/1526569118-14217-3-git-send-email-yao.jin@linux.intel.com
[ Rename all maxmin to minmax ]
Signed-off-by: NArnaldo Carvalho de Melo <acme@redhat.com>
      3e71fc03