On Wed, 2010-10-27 at 13:31 +0200, Karel Zak wrote:
> On Wed, Oct 27, 2010 at 06:46:42AM -0300, Davidlohr Bueso wrote:
> > From: Davidlohr Bueso <dave@xxxxxxx>
> > Date: Wed, 27 Oct 2010 06:44:17 -0300
> > Subject: [PATCH] taskset: fix memory leaks when exiting on error
> >
> > Signed-off-by: Davidlohr Bueso <dave@xxxxxxx>
> > ---
> > schedutils/taskset.c | 48 +++++++++++++++++++++++++++++++++++-------------
> > 1 files changed, 35 insertions(+), 13 deletions(-)
> >
> > diff --git a/schedutils/taskset.c b/schedutils/taskset.c
> > index 201fc15..8094c5d 100644
> > --- a/schedutils/taskset.c
> > +++ b/schedutils/taskset.c
> > @@ -125,8 +125,10 @@ int main(int argc, char *argv[])
> >
> > buflen = 7 * cur_nbits;
> > buf = malloc(buflen);
> > - if (!buf)
> > + if (!buf) {
> > + cpuset_free(cur_set);
> > err(EXIT_FAILURE, _("malloc failed"));
> > + }
> >
> > /*
> > * new_set is always used for the sched_setaffinity call
> > @@ -134,12 +136,17 @@ int main(int argc, char *argv[])
> > * cpumask_t if the user's mask is shorter.
> > */
> > new_set = cpuset_alloc(ncpus, &new_setsize, NULL);
> > - if (!new_set)
> > + if (!new_set) {
> > + free(buf);
> > + cpuset_free(cur_set);
> > err(EXIT_FAILURE, _("cpuset_alloc failed"));
> > + }
> >
> > if (pid) {
> > - if (sched_getaffinity(pid, cur_setsize, cur_set) < 0)
> > - err(EXIT_FAILURE, _("failed to get pid %d's affinity"), pid);
> > + if (sched_getaffinity(pid, cur_setsize, cur_set) < 0) {
> > + fprintf(stderr, _("failed to get pid %d's affinity"), pid);
> > + goto err_free;
>
> I know about this problem, but I have doubts that this pedantic point
> of view and the extra code makes any real effect. After exit() all
> memory will be deallocated... I think it's better in this example to
> call err() and keep the code simple and readable.
>
Fair enough. I ran into this when looking to use the xalloc stuff, so we
just might want to change cpuset_alloc() to use xmalloc instead of
malloc and be done with it.
Davidlohr
--
To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
