Re: [PATCH 1/2] tracing: Check total refcount before releasingbufs in profile_enable failure

[Frederic Weisbecker]

On Wed, Oct 07, 2009 at 02:08:27PM +1100, Paul Mackerras wrote:
> Frederic Weisbecker writes:
> 
> > When we call the profile_enable() callback of an event, we release the
> > shared perf event tracing buffers unconditionnaly in the failure path.
> > This is wrong because there may be other users of these. Then check the
> > total refcount before doing this.
> 
> [snip]
> 
> > -	kfree(trace_profile_buf_nmi);
> >  fail_buf_nmi:
> > -	kfree(trace_profile_buf);
> > +	if (!total_profile_count) {
> 
> A small problem here: total_profile_count will be 1, not 0, in the
> case where we need to free...
> 
> > +		kfree(trace_profile_buf_nmi);
> > +		kfree(trace_profile_buf);
> > +		trace_profile_buf_nmi = NULL;
> > +		trace_profile_buf = NULL;
> > +	}
> >  fail_buf:
> > -	total_profile_count--;
> 
> since we don't decrement total_profile_count until here.
> 
> Paul.


No, because now we only increment total_profile_count if
nothing has failed. This is even the last thing done
in the succeeded path. So if we fail and need to free, it
means its values is still 0.

--
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/