lkml.org 
[lkml]   [2013]   [Feb]   [4]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 6/6] uprobes/perf: Always increment trace_uprobe->nhit
On 02/04, Srikar Dronamraju wrote:
>
> * Oleg Nesterov <oleg@redhat.com> [2013-01-31 20:18:32]:
>
> > Move tu->nhit++ from uprobe_trace_func() to uprobe_dispatcher().
> >
> > ->nhit counts how many time we hit the breakpoint inserted by this
> > uprobe, we do not want to loose this info if uprobe was enabled by
> > sys_perf_event_open().
> >
>
> Though I dont see a problem with this change, It seems unnecessary for
> me.
>
> Info from nhits is mostly for /sys/kernel/debug/tracing/uprobe_profile

It is only for uprobe_profile, yes, and it is useful. Why should we hide
this info if this uprobe is used by perf?

> I am not sure how sys_perf_event_open() is making use of this?

I hope I'll send the final series today. From the changelog of the patch
which actually turns the filtering on:

Testing:

# perf probe -x /lib/libc.so.6 syscall

# perl -e 'syscall -1 while 1' &
[1] 530

# perf record -e probe_libc:syscall perl -e 'syscall -1 for 1..10; sleep 1'

# perf report --show-total-period
100.00% 10 perl libc-2.8.so [.] syscall

Before this patch:

# cat /sys/kernel/debug/tracing/uprobe_profile
/lib/libc.so.6 syscall 79291

A huge ->nrhit == 79291 reflects the fact that the background process
530 constantly hits this breakpoint too, even if doesn't contribute to
the output.

After the patch:

# cat /sys/kernel/debug/tracing/uprobe_profile
/lib/libc.so.6 syscall 10

This shows that only the target process was punished by int3.

Oleg.



\
 
 \ /
  Last update: 2013-02-04 17:03    [W:0.071 / U:0.780 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site