Am Mittwoch, 4. Juli 2007 schrieb Rusty Russell: > + vbr = mempool_alloc(vblk->pool, GFP_ATOMIC); > + if (!vbr) > + goto stop; [...] > + BUG_ON(req->nr_phys_segments > ARRAY_SIZE(vblk->sg)); > + vbr->req = req; > + if (!do_req(q, vblk, vbr)) > + goto stop; [...] > +stop: > + /* Queue full? Wait. */ > + blk_stop_queue(q); > + mempool_free(vbr, vblk->pool);
Hmm, can mempool_free really handle NULL as its first argument? (first goto). As far as I can see we have a problem here. Our pool has preallocated 1 element. So what about: - first alloc fails in mempool_alloc fails -> we get the pre-allocated element, no more elements left - second alloc fails in mempool_alloc fails -> we get NULL -> we return NULL to mempool_free which adds NULL back to the pool as the pool is empty. Now the pool has an NULL entry, no? Christian _______________________________________________ Virtualization mailing list [email protected] https://lists.linux-foundation.org/mailman/listinfo/virtualization
