[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CACVXFVP+9w-aAn+ZprrpEy0XKqgiY5i+eQWWxDKiMKBPJGRKEA@mail.gmail.com>
Date: Thu, 4 Jun 2015 21:19:29 +0800
From: Ming Lei <tom.leiming@...il.com>
To: Jens Axboe <axboe@...com>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Cc: Stefan Seyfried <stefan.seyfried@...glemail.com>,
Ming Lei <tom.leiming@...il.com>, NeilBrown <neilb@...e.de>,
Christoph Hellwig <hch@....de>, "v4.0" <stable@...r.kernel.org>
Subject: Re: [PATCH] blk-mq: free hctx->ctxs in queue's release handler
On Thu, Jun 4, 2015 at 9:14 PM, Ming Lei <tom.leiming@...il.com> wrote:
> Now blk_cleanup_queue() can be called before calling
> del_gendisk()[1], inside which hctx->ctxs is touched
> from blk_mq_unregister_hctx(), but the variable has
> been freed by blk_cleanup_queue() at that time.
>
> So this patch moves freeing of hctx->ctxs into queue's
> release handler for fixing the oops reported by Stefan.
>
> Also ctx's kobject is embedded into each ctx pointed by
> hctx->ctxs[], which shouldn't have been released so early.
Hammm, the above line is wrong and should be removed, and
hctx->ctxs[] just stores the pointer of ctx.
Jens, if you need v1, please let me know.
>
> [1], 6cd18e711dd8075 (block: destroy bdi before blockdev is
> unregistered)
>
> Reported-by: Stefan Seyfried <stefan.seyfried@...glemail.com>
> Cc: NeilBrown <neilb@...e.de>
> Cc: Christoph Hellwig <hch@....de>
> Cc: stable@...r.kernel.org (v4.0)
> Signed-off-by: Ming Lei <tom.leiming@...il.com>
> ---
> block/blk-mq.c | 8 ++++++--
> 1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/block/blk-mq.c b/block/blk-mq.c
> index e68b71b..594eea0 100644
> --- a/block/blk-mq.c
> +++ b/block/blk-mq.c
> @@ -1600,6 +1600,7 @@ static int blk_mq_hctx_notify(void *data, unsigned long action,
> return NOTIFY_OK;
> }
>
> +/* hctx->ctxs will be freed in queue's release handler */
> static void blk_mq_exit_hctx(struct request_queue *q,
> struct blk_mq_tag_set *set,
> struct blk_mq_hw_ctx *hctx, unsigned int hctx_idx)
> @@ -1618,7 +1619,6 @@ static void blk_mq_exit_hctx(struct request_queue *q,
>
> blk_mq_unregister_cpu_notifier(&hctx->cpu_notifier);
> blk_free_flush_queue(hctx->fq);
> - kfree(hctx->ctxs);
> blk_mq_free_bitmap(&hctx->ctx_map);
> }
>
> @@ -1891,8 +1891,12 @@ void blk_mq_release(struct request_queue *q)
> unsigned int i;
>
> /* hctx kobj stays in hctx */
> - queue_for_each_hw_ctx(q, hctx, i)
> + queue_for_each_hw_ctx(q, hctx, i) {
> + if (!hctx)
> + continue;
> + kfree(hctx->ctxs);
> kfree(hctx);
> + }
>
> kfree(q->queue_hw_ctx);
>
> --
> 1.9.1
>
--
Ming Lei
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists