在 2023/8/17 4:20, Steven Rostedt 写道:

On Mon, 14 Aug 2023 10:59:18 +0800
Yipeng Zou [off-list ref] wrote:

quoted

So I directly dump the trace file before md5sum, the diff shows that:

[root@localhost]# diff trace_1.log trace_2.log -y --suppress-common-lines
dockerd-12285   [036] d.... 18385.510290: sched_stat | <...>-12285   [036] d.... 18385.510290: sched_stat
dockerd-12285   [036] d.... 18385.510291: sched_swit | <...>-12285   [036] d.... 18385.510291: sched_swit
<...>-740       [044] d.... 18385.602859: sched_stat | kworker/44:1-740 [044] d.... 18385.602859: sched_stat
<...>-740       [044] d.... 18385.602860: sched_swit | kworker/44:1-740 [044] d.... 18385.602860: sched_swit

So the problem is the <..> being added?

quoted

And we can see that the saved_cmdlines in kernel was changed.

So Maybe there is no safe to use md5sum here to verify whether
the file whether it has changed.

quoted

From my point of view, we only need to verify traceoff event  

is valid there.

Maybe it is enough to check first that the tracing_on equals 0,
and then check that the total number of rows of trace does not
increase?

Fixes: d87b29179aa0 ("selftests: ftrace: Use md5sum to take less time of checking logs")
Signed-off-by: Yipeng Zou <redacted>
---
  .../test.d/ftrace/func_traceonoff_triggers.tc    | 16 ++++------------
  1 file changed, 4 insertions(+), 12 deletions(-)

diff --git a/tools/testing/selftests/ftrace/test.d/ftrace/func_traceonoff_triggers.tc b/tools/testing/selftests/ftrace/test.d/ftrace/func_traceonoff_triggers.tc
index aee22289536b..180c60605d7b 100644
--- a/tools/testing/selftests/ftrace/test.d/ftrace/func_traceonoff_triggers.tc
+++ b/tools/testing/selftests/ftrace/test.d/ftrace/func_traceonoff_triggers.tc

@@ -77,25 +77,17 @@ if [ $cnt -ne 1 ]; then
     fail "Did not find traceoff trigger"
  fi
  
-cnt=`cnt_trace`
-sleep $SLEEP_TIME
-cnt2=`cnt_trace`
-
-if [ $cnt -ne $cnt2 ]; then
-   fail "Tracing is not stopped"
-fi
-
  on=`cat tracing_on`
  if [ $on != "0" ]; then
      fail "Tracing is not off"
  fi
  
-csum1=`md5sum trace`

What about just changing the above to:

  # Can not rely on names being around as they are only cached
  csum1=`cat trace | sed -e 's/^ *[^ ]*\(-[0-9][0-9]*\)/\1/' | md5sum`

quoted

+cnt=`cnt_trace`
  sleep $SLEEP_TIME
-csum2=`md5sum trace`

And this:

  csum2=`cat trace | sed -e 's/^ *[^ ]*\(-[0-9][0-9]*\)/\1/' | md5sum`

That will strip off the names an leave just the -[PID]... part of the line?

-- Steve

Hi Steve:

     Thanks for your suggests, this is indeed more effective.

     The problem was that <...> filed be filled with names, and it 
causes the md5 value changed.

     I will send v2 in that ways, thanks !

-- Yipeng

quoted

+cnt2=`cnt_trace`
  
-if [ "$csum1" != "$csum2" ]; then
-    fail "Tracing file is still changing"
+if [ $cnt -ne $cnt2 ]; then
+   fail "Tracing is not stopped"
  fi
  
  clear_trace

-- 
Regards,
Yipeng Zou