From: hch@infradead.org (Christoph Hellwig)
Subject: [PATCHv2-4.5 08/10] NVMe: Move error handling to failed reset handler
Date: Sat, 13 Feb 2016 01:46:34 -0800 [thread overview]
Message-ID: <20160213094634.GB15318@infradead.org> (raw)
In-Reply-To: <1455221147-24228-9-git-send-email-keith.busch@intel.com>
On Thu, Feb 11, 2016@01:05:45PM -0700, Keith Busch wrote:
> This moves failed queue handling out of the namespace removal path and into
> the reset failure path, fixing a deadlock condition if the controller
> fails or link down during del_gendisk. Previously the driver had to see
> the controller as degraded prior to calling del_gendisk to setup the
> queues to fail. If the controller happened to fail after this though,
> there was no task to end the request_queue.
>
> On failure, all namespace states are set to 'dead'. This has capacity
> revalidate to 0, and ends all new requests with error status.
I like this a lot, but a few comments below:
> +/**
> + * nvme_kill_ns_queues(): Ends all namespace queues
> + * @ctrl: the dead controller that needs to end
> + *
> + * Call this function when the driver determines it is unable to get the
> + * controller in a state capable of servicing IO.
> + */
> +void nvme_kill_ns_queues(struct nvme_ctrl *ctrl)
Why do we have a separate function for this instead of driving this
from nvme_remove_namespaces?
> @@ -678,7 +678,9 @@ static int nvme_queue_rq(struct blk_mq_hw_ctx *hctx,
>
> spin_lock_irq(&nvmeq->q_lock);
> if (unlikely(nvmeq->cq_vector < 0)) {
> - ret = BLK_MQ_RQ_QUEUE_BUSY;
> + ret = test_bit(NVME_NS_DEAD, &ns->flags) ?
> + BLK_MQ_RQ_QUEUE_ERROR :
> + BLK_MQ_RQ_QUEUE_BUSY;
This would be easier to read with a good old 'if'.
next prev parent reply other threads:[~2016-02-13 9:46 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-02-11 20:05 [PATCHv2-4.5 00/10] NVMe fixes Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 01/10] blk-mq: End unstarted requests on dying queue Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 02/10] NVMe: Fix io incapable return values Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 03/10] NVMe: Allow request merges Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 04/10] NVMe: Fix namespace removal deadlock Keith Busch
2016-02-13 9:44 ` Christoph Hellwig
2016-02-11 20:05 ` [PATCHv2-4.5 05/10] NVMe: Requeue requests on suspended queues Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 06/10] NVMe: Poll device while still active during remove Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 07/10] NVMe: Simplify device reset failure Keith Busch
2016-02-13 9:21 ` Christoph Hellwig
2016-02-11 20:05 ` [PATCHv2-4.5 08/10] NVMe: Move error handling to failed reset handler Keith Busch
2016-02-13 9:46 ` Christoph Hellwig [this message]
2016-02-16 21:57 ` Keith Busch
2016-02-17 8:06 ` Christoph Hellwig
2016-02-11 20:05 ` [PATCHv2-4.5 09/10] NVMe: Mark queues as dead on degraded controller Keith Busch
2016-02-11 20:05 ` [PATCHv2-4.5 10/10] NVMe: Rate limit nvme IO warnings Keith Busch
2016-02-12 8:16 ` Johannes Thumshirn
2016-02-11 22:28 ` [PATCHv2-4.5 00/10] NVMe fixes Keith Busch
2016-02-11 22:38 ` Jens Axboe
2016-02-12 8:35 ` Christoph Hellwig
2016-02-12 15:09 ` Jens Axboe
2016-02-12 15:24 ` Keith Busch
2016-02-13 9:49 ` Christoph Hellwig
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20160213094634.GB15318@infradead.org \
--to=hch@infradead.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).