Re: blktrace segfault.

[Jens Axboe <jens.axboe@oracle.com>]

On Wed, Nov 14 2007, Jan Blunck wrote:
> On Sun, Oct 28, Aneesh Kumar K.V wrote:
> 
> >
> > the below diff fix the same.
> >
> > diff --git a/blktrace.c b/blktrace.c
> > index 42613a7..4f4fcde 100644
> > --- a/blktrace.c
> > +++ b/blktrace.c
> > @@ -1220,6 +1220,7 @@ static int start_devices(void)
> > 	size = ncpus * sizeof(struct thread_information);
> > 	thread_information = malloc(size * ndevs);
> > +	memset(thread_information, 0, size*ndevs);
> > 	if (!thread_information) {
> > 		fprintf(stderr, "Out of memory, threads (%d)\n", size * ndevs);
> > 		return 1;
> 
> I'm pretty sure you want to make the memset after the NULL-pointer check. If
> malloc fails this is segfaulting too.

Ehm that's indeed a good point, I'll check in a fix for that.

-- 
Jens Axboe