linux/tools/perf
Björn Töpel 7598f8bc13 perf probe: Fix probe definition for inlined functions
In commit 613f050d68 ("perf probe: Fix to probe on gcc generated
functions in modules"), the offset from symbol is, incorrectly, added
to the trace point address. This leads to incorrect probe trace points
for inlined functions and when using relative line number on symbols.

Prior this patch:
  $ perf probe -m nf_nat -D in_range
  p:probe/in_range nf_nat:in_range.isra.9+0
  $ perf probe -m i40e -D i40e_clean_rx_irq
  p:probe/i40e_clean_rx_irq i40e:i40e_napi_poll+2212
  $ perf probe -m i40e -D i40e_clean_rx_irq:16
  p:probe/i40e_clean_rx_irq i40e:i40e_lan_xmit_frame+626

After:
  $ perf probe -m nf_nat -D in_range
  p:probe/in_range nf_nat:in_range.isra.9+0
  $ perf probe -m i40e -D i40e_clean_rx_irq
  p:probe/i40e_clean_rx_irq i40e:i40e_napi_poll+1106
  $ perf probe -m i40e -D i40e_clean_rx_irq:16
  p:probe/i40e_clean_rx_irq i40e:i40e_napi_poll+2665

Committer testing:

Using 'pfunct', a tool found in the 'dwarves' package [1], one can ask what are
the functions that while not being explicitely marked as inline, were inlined
by the compiler:

  # pfunct --cc_inlined /lib/modules/4.12.0-rc4+/kernel/drivers/net/ethernet/intel/e1000e/e1000e.ko | head
  __ew32
  e1000_regdump
  e1000e_dump_ps_pages
  e1000_desc_unused
  e1000e_systim_to_hwtstamp
  e1000e_rx_hwtstamp
  e1000e_update_rdt_wa
  e1000e_update_tdt_wa
  e1000_put_txbuf
  e1000_consume_page

Then ask 'perf probe' to produce the kprobe_tracer probe definitions for two of
them:

  # perf probe -m e1000e -D e1000e_rx_hwtstamp
  p:probe/e1000e_rx_hwtstamp e1000e:e1000_receive_skb+74

  # perf probe -m e1000e -D e1000_consume_page
  p:probe/e1000_consume_page e1000e:e1000_clean_jumbo_rx_irq+876
  p:probe/e1000_consume_page_1 e1000e:e1000_clean_jumbo_rx_irq+1506
  p:probe/e1000_consume_page_2 e1000e:e1000_clean_rx_irq_ps+1074

Now lets concentrate on the 'e1000_consume_page' one, that was inlined twice in
e1000_clean_jumbo_rx_irq(), lets see what readelf says about the DWARF tags for
that function:

  $ readelf -wi /lib/modules/4.12.0-rc4+/kernel/drivers/net/ethernet/intel/e1000e/e1000e.ko
  <SNIP>
  <1><13e27b>: Abbrev Number: 121 (DW_TAG_subprogram)
    <13e27c>   DW_AT_name        : (indirect string, offset: 0xa8945): e1000_clean_jumbo_rx_irq
    <13e287>   DW_AT_low_pc      : 0x17a30
  <3><13e6ef>: Abbrev Number: 119 (DW_TAG_inlined_subroutine)
    <13e6f0>   DW_AT_abstract_origin: <0x13ed2c>
    <13e6f4>   DW_AT_low_pc      : 0x17be6
  <SNIP>
  <1><13ed2c>: Abbrev Number: 142 (DW_TAG_subprogram)
     <13ed2e>   DW_AT_name        : (indirect string, offset: 0xa54c3): e1000_consume_page

So, the first time in e1000_clean_jumbo_rx_irq() where e1000_consume_page() is
inlined is at PC 0x17be6, which subtracted from e1000_clean_jumbo_rx_irq()'s
address, gives us the offset we should use in the probe definition:

  0x17be6 - 0x17a30 = 438

but above we have 876, which is twice as much.

Lets see the second inline expansion of e1000_consume_page() in
e1000_clean_jumbo_rx_irq():

  <3><13e86e>: Abbrev Number: 119 (DW_TAG_inlined_subroutine)
    <13e86f>   DW_AT_abstract_origin: <0x13ed2c>
    <13e873>   DW_AT_low_pc      : 0x17d21

  0x17d21 - 0x17a30 = 753

So we where adding it at twice the offset from the containing function as we
should.

And then after this patch:

  # perf probe -m e1000e -D e1000e_rx_hwtstamp
  p:probe/e1000e_rx_hwtstamp e1000e:e1000_receive_skb+37

  # perf probe -m e1000e -D e1000_consume_page
  p:probe/e1000_consume_page e1000e:e1000_clean_jumbo_rx_irq+438
  p:probe/e1000_consume_page_1 e1000e:e1000_clean_jumbo_rx_irq+753
  p:probe/e1000_consume_page_2 e1000e:e1000_clean_jumbo_rx_irq+1353
  #

Which matches the two first expansions and shows that because we were
doubling the offset it would spill over the next function:

  readelf -sw /lib/modules/4.12.0-rc4+/kernel/drivers/net/ethernet/intel/e1000e/e1000e.ko
   673: 0000000000017a30  1626 FUNC    LOCAL  DEFAULT    2 e1000_clean_jumbo_rx_irq
   674: 0000000000018090  2013 FUNC    LOCAL  DEFAULT    2 e1000_clean_rx_irq_ps

This is the 3rd inline expansion of e1000_consume_page() in
e1000_clean_jumbo_rx_irq():

   <3><13ec77>: Abbrev Number: 119 (DW_TAG_inlined_subroutine)
    <13ec78>   DW_AT_abstract_origin: <0x13ed2c>
    <13ec7c>   DW_AT_low_pc      : 0x17f79

  0x17f79 - 0x17a30 = 1353

 So:

   0x17a30 + 2 * 1353 = 0x184c2

  And:

   0x184c2 - 0x18090 = 1074

Which explains the bogus third expansion for e1000_consume_page() to end up at:

   p:probe/e1000_consume_page_2 e1000e:e1000_clean_rx_irq_ps+1074

All fixed now :-)

[1] https://git.kernel.org/pub/scm/devel/pahole/pahole.git/

Signed-off-by: Björn Töpel <bjorn.topel@intel.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Acked-by: Masami Hiramatsu <mhiramat@kernel.org>
Cc: stable@vger.kernel.org
Fixes: 613f050d68 ("perf probe: Fix to probe on gcc generated functions in modules")
Link: http://lkml.kernel.org/r/20170621164134.5701-1-bjorn.topel@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2017-06-22 16:08:09 -03:00
..
arch perf tools: Fix build with ARCH=x86_64 2017-06-14 15:44:29 -03:00
bench perf tools: Move extra string util functions to util/string2.h 2017-04-19 13:01:51 -03:00
Documentation perf script python: Remove dups in documentation examples 2017-06-07 20:36:12 -03:00
jvmti perf kvmti: Remove unused Makefile file 2016-11-14 12:42:56 -03:00
pmu-events perf tools: Fix build with ARCH=x86_64 2017-06-14 15:44:29 -03:00
python perf python: Add tracepoint example 2016-07-12 16:23:35 -03:00
scripts perf/core improvements and fixes: 2016-08-04 11:02:38 +02:00
tests perf tools: Fix build with ARCH=x86_64 2017-06-14 15:44:29 -03:00
trace perf tools: Add signal.h to places using its definitions 2017-04-20 13:22:43 -03:00
ui perf tools: Put caller above callee in --children mode 2017-05-24 08:41:49 +02:00
util perf probe: Fix probe definition for inlined functions 2017-06-22 16:08:09 -03:00
.gitignore perf tools: Ignore generated files pmu-events/{jevents,pmu-events.c} for git 2017-03-13 10:59:36 -03:00
Build perf trace: Beautify statx syscall 'flag' and 'mask' arguments 2017-03-31 14:42:31 -03:00
builtin-annotate.c perf tools: Include errno.h where needed 2017-04-19 13:01:51 -03:00
builtin-bench.c perf tools: Remove unused 'prefix' from builtin functions 2017-03-27 11:58:09 -03:00
builtin-buildid-cache.c perf symbols: Accept symbols starting at address 0 2017-05-02 18:23:04 -03:00
builtin-buildid-list.c perf tools: Include errno.h where needed 2017-04-19 13:01:51 -03:00
builtin-c2c.c perf tools: Move event prototypes from util.h to event.h 2017-04-25 15:30:47 -03:00
builtin-config.c perf config: Refactor a duplicated code for obtaining config file name 2017-05-02 18:23:12 -03:00
builtin-data.c perf tools: Remove unused 'prefix' from builtin functions 2017-03-27 11:58:09 -03:00
builtin-diff.c perf tools: Include errno.h where needed 2017-04-19 13:01:51 -03:00
builtin-evlist.c perf tools: Remove unused 'prefix' from builtin functions 2017-03-27 11:58:09 -03:00
builtin-ftrace.c perf tools: Remove poll.h and wait.h from util.h 2017-04-24 13:43:34 -03:00
builtin-help.c perf tools: Remove string.h, unistd.h and sys/stat.h from util.h 2017-04-24 13:43:33 -03:00
builtin-inject.c perf tools: Use just forward declarations for struct thread where possible 2017-04-24 13:43:35 -03:00
builtin-kallsyms.c perf tools: Including missing inttypes.h header 2017-04-19 13:01:46 -03:00
builtin-kmem.c perf tools: Include errno.h where needed 2017-04-19 13:01:51 -03:00
builtin-kvm.c perf tools: Remove poll.h and wait.h from util.h 2017-04-24 13:43:34 -03:00
builtin-list.c perf tools: Remove unused 'prefix' from builtin functions 2017-03-27 11:58:09 -03:00
builtin-lock.c perf tools: Include errno.h where needed 2017-04-19 13:01:51 -03:00
builtin-mem.c perf tools: Use just forward declarations for struct thread where possible 2017-04-24 13:43:35 -03:00
builtin-probe.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jikos/trivial 2017-05-02 19:09:35 -07:00
builtin-record.c perf tools: Remove poll.h and wait.h from util.h 2017-04-24 13:43:34 -03:00
builtin-report.c perf tools: Remove string.h, unistd.h and sys/stat.h from util.h 2017-04-24 13:43:33 -03:00
builtin-sched.c perf tools: Include errno.h where needed 2017-04-19 13:01:51 -03:00
builtin-script.c perf script: Add --inline option for debugging 2017-05-24 08:41:48 +02:00
builtin-stat.c perf stat: Only print NMI watchdog hint when enabled 2017-06-02 11:15:34 -03:00
builtin-timechart.c perf tools: Use just forward declarations for struct thread where possible 2017-04-24 13:43:35 -03:00
builtin-top.c perf tools: Move event prototypes from util.h to event.h 2017-04-25 15:30:47 -03:00
builtin-trace.c perf trace: Add mmap alias for s390 2017-06-01 10:13:21 -03:00
builtin-version.c perf tools: Remove string.h, unistd.h and sys/stat.h from util.h 2017-04-24 13:43:33 -03:00
builtin.h perf tools: Remove stale prototypes from builtin.h 2017-04-24 13:43:33 -03:00
check-headers.sh tools include uapi: Grab copies of stat.h and fcntl.h 2017-03-31 11:26:03 -03:00
command-list.txt perf tools: Missing c2c command in command-list 2017-03-13 10:59:31 -03:00
CREDITS
design.txt
Makefile
Makefile.config perf tools: Fix build with ARCH=x86_64 2017-06-14 15:44:29 -03:00
Makefile.perf perf tools: Fix build with ARCH=x86_64 2017-06-14 15:44:29 -03:00
MANIFEST tools include: Introduce linux/bug.h, from the kernel sources 2017-04-19 13:01:42 -03:00
perf-archive.sh
perf-completion.sh
perf-read-vdso.c
perf-sys.h perf powerpc: Fix build-test failure 2016-09-08 13:44:07 -03:00
perf-with-kcore.sh
perf.c perf tools: Move event prototypes from util.h to event.h 2017-04-25 15:30:47 -03:00
perf.h perf tools: Add PERF_RECORD_NAMESPACES to include namespaces related info 2017-03-14 11:38:23 -03:00