On Sep 5, 2014 3:55 AM, "Paolo Bonzini" [off-list ref] wrote:
Il 03/09/2014 06:29, Rusty Russell ha scritto:
quoted
+ desc = kmalloc(total_sg * sizeof(struct vring_desc), gfp);
+ if (!desc)
+ return NULL;
- return head;
+ for (i = 0; i < total_sg; i++)
+ desc[i].next = i+1;
+ return desc;
}
Would it make sense to keep a cache of a few (say) 8- or 16-element
indirect descriptors? You'd only have to do this ugly (and slowish) for
loop on the first allocation.
Also, since this is mostly an aesthetic patch,
quoted
+ if (indirect)
+ vq->free_head = vq->vring.desc[head].next;
+ else
+ vq->free_head = i;
I'd move the indirect case above, where the vring.desc[head] is actually
allocated.
Please don't: I'll just have to undo it for the DMA API stuff, since
descriptor setup will be able to fail.
--Andy