Re: Re: alsa-lib patches

[James Courtier-Dutton <James@superbug.co.uk>]

Sergey Vlasov wrote:
> 
> mixer->pelems is an array of pointers to snd_mixer_elem_t, and qsort()
> passes pointers to array elements as arguments to the compare
> function - which becomes (snd_mixer_elem_t **).  But mixer->compare
> expects to get (snd_mixer_elem_t *) as arguments.
> 
Ok, so not a security issue, just a messed up pointer issue.


> However, seems that mixer elements have an indirect pointer to the
> mixer they belong to - elem->class->mixer.  So, how about this:
> 
> static int snd_mixer_qsort_compare(const void *a, const void *b)
> {
> 	snd_mixer_elem_t *elem_a = *(snd_mixer_elem_t **)a;
> 	snd_mixer_elem_t *elem_b = *(snd_mixer_elem_t **)b;
> 
> 	return elem_a->class->mixer->compare(elem_a, elem_b);
> }
> 
> static int snd_mixer_sort(snd_mixer_t *mixer)
> {
> 	unsigned int k;
> 	assert(mixer);
> 	assert(mixer->compare);
> 	INIT_LIST_HEAD(&mixer->elems);
> 	qsort(mixer->pelems, mixer->count, sizeof(snd_mixer_elem_t*), snd_mixer_qsort_compare);
> 	for (k = 0; k < mixer->count; k++)
> 		list_add_tail(&mixer->pelems[k]->list, &mixer->elems);
> 	return 0;
> }

That seems like a better solution. This sort routine does not have to be
high performance as it is not used often.

I suppose this begs the question, why has it work ok up until now?

James




-------------------------------------------------------
This SF.Net email is sponsored by Oracle Space Sweepstakes
Want to be the first software developer in space?
Enter now for the Oracle Space Sweepstakes!
http://ads.osdn.com/?ad_id=7393&alloc_id=16281&op=click