Messages in this thread | | | Subject | Re: [PATCH v2 6/8] blk-mq: Rework blk_mq_realloc_hw_ctxs() | From | Sagi Grimberg <> | Date | Thu, 6 Oct 2016 13:11:32 +0300 |
| |
On 06/10/16 11:25, Alexander Gordeev wrote: > On Thu, Oct 06, 2016 at 12:47:26AM +0300, Sagi Grimberg wrote: >> >>> @@ -1908,33 +1909,36 @@ static void blk_mq_realloc_hw_ctxs(struct blk_mq_tag_set *set, >>> if (node == NUMA_NO_NODE) >>> node = set->numa_node; >>> >>> - hctxs[i] = kzalloc_node(sizeof(struct blk_mq_hw_ctx), >>> - GFP_KERNEL, node); >>> - if (!hctxs[i]) >>> + hctx = kzalloc_node(sizeof(*hctx), GFP_KERNEL, node); >>> + if (!hctx) >>> break; >>> >>> - if (blk_mq_init_hctx(q, set, hctxs[i], i, node)) { >>> - kfree(hctxs[i]); >>> - hctxs[i] = NULL; >>> + if (blk_mq_init_hctx(q, set, hctx, i, node)) { >>> + kfree(hctx); >>> break; >>> } >>> - blk_mq_hctx_kobj_init(hctxs[i]); >>> + >>> + blk_mq_hctx_kobj_init(hctx); >>> + hctxs[i] = hctx; >>> } >>> for (j = i; j < q->nr_hw_queues; j++) { >>> - struct blk_mq_hw_ctx *hctx = hctxs[j]; >>> + hctx = hctxs[i]; >> >> Didn't you mean hctx[j]? > > Surely, I did. > Thanks!
Maybe it would be cleaner to do:
while (--i >= 0) { hctx = hctxs[i]; ... }
| |