Re: perf/ftrace: fuzzer triggers another warning with PERF_EVENT_IOC_SET_FILTER

[Vince Weaver]

On Thu, 18 Jun 2015, Steven Rostedt wrote:

> On Wed, 17 Jun 2015 19:32:26 -0400 (EDT)
> Vince Weaver <vincent.weaver@xxxxxxxxx> wrote:
> 
> > 
> > The fuzzer triggered another warning in the PERF_EVENT_IOC_SET_FILTER 
> > code.  This is with the previous filter patch applied.
> > 
> > >From what I can tell the filter was:
> > nfs4:nfs4_fsinfo 
> > (((fhandle~601)&&valid&809)&&common_type>=860) 
> 
> Hmm, I can't trigger this with the filter file, can you?
> 
>  echo '(((fhandle~601)&&valid&809)&&common_type>=860)' > \
>   /sys/kernel/debug/tracing/events/nfs4/nfs4_fsinfo/filter
> 
> I'll write a user app to use the perf syscall and see if that is
> different, but in the mean time, can you try the above.

I can't seem to replicate this with perf or echo.

Unfortunately I made many changes to the fuzzer since hitting the bug so 
it's a bit hard to replicate with the fuzzer.  I'll keep fuzzing and see 
if I can hit it again.

As an aside, I can't find any documentation of how the "~" glob operator 
works in ftrace filters.  Can you give a quick example of a valid filter 
that uses it?

Thanks,

Vince
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/