linux-nvme.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
From: hch@lst.de (Christoph Hellwig)
Subject: [PATCH 6/6] nvme: ignore retries for multipath devices
Date: Tue, 3 Oct 2017 13:53:55 +0200	[thread overview]
Message-ID: <20171003115355.GB24650@lst.de> (raw)
In-Reply-To: <9d603720-2f86-ab3c-6ba3-9d57afb3568e@suse.de>

On Tue, Oct 03, 2017@12:02:38PM +0200, Hannes Reinecke wrote:
> >>  	if (nvme_req(req)->status & NVME_SC_DNR)
> >>  		return false;
> >> -	if (nvme_req(req)->retries >= nvme_max_retries)
> >> +	if (nvme_req(req)->retries >= nvme_max_retries &&
> >> +	    !(req->cmd_flags & REQ_NVME_MPATH))
> >>  		return false;
> >>  	return true;
> > 
> > All failover logic is inside a nvme_req_needs_retry() conditional,
> > so this change looks completely broken - it basically disables
> > failover.
> > 
> Not in our tests.
> Without this patch we'd been seeing I/O errors during failover; with
> this patch I/O continues on the failover path.

http://git.infradead.org/users/hch/block.git/blob/refs/heads/nvme-mpath:/drivers/nvme/host/core.c#l208

210	if (unlikely(nvme_req(req)->status && nvme_req_needs_retry(req))) {
211		if (nvme_req_needs_failover(req)) {
212			nvme_failover_req(req);
213			return;
214		 }

The only call to nvme_failover_req is guarded by nvme_req_needs_retry,
and you change needs_retry to return true for MPATH requests that
exceed the number of retries.  I just don't see how we'd hit the
max_retries count, as each retry before should have already taken
nvme_req_needs_failover before.  What error code do you see this
with?  What kinds of device/setup?

  reply	other threads:[~2017-10-03 11:53 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-02 13:55 [PATCH 0/6] nvme: sanitize multipathing Hannes Reinecke
2017-10-02 13:55 ` [PATCH 1/6] nvme: display 'CMIC' controller attribute Hannes Reinecke
2017-10-02 16:15   ` Christoph Hellwig
2017-10-02 13:55 ` [PATCH 2/6] nvme: use 'nvmsXnZ' instead of 'nvm-subXnZ' Hannes Reinecke
2017-10-02 16:16   ` Christoph Hellwig
2017-10-02 16:20     ` Keith Busch
2017-10-11 14:32       ` Guan Junxiong
2017-10-02 13:55 ` [PATCH 3/6] nvme: claim block devices Hannes Reinecke
2017-10-02 16:42   ` Christoph Hellwig
2017-10-03 10:08     ` Hannes Reinecke
2017-10-03 11:55       ` Christoph Hellwig
2017-10-04  5:42         ` Hannes Reinecke
2017-10-04  6:15           ` Christoph Hellwig
2017-10-04  6:33             ` Hannes Reinecke
2017-10-04  7:13               ` Christoph Hellwig
2017-10-04  7:26                 ` Hannes Reinecke
2017-10-05  1:08                 ` Martin K. Petersen
2017-10-05  6:51                   ` Johannes Thumshirn
2017-10-05 14:05                     ` Keith Busch
2017-10-05 14:39                       ` Hannes Reinecke
2017-10-05 14:47                         ` Christoph Hellwig
2017-10-02 13:55 ` [PATCH 4/6] nvme: display 'NMIC' namespace attribute Hannes Reinecke
2017-10-02 16:16   ` Christoph Hellwig
2017-10-03 10:00     ` Hannes Reinecke
2017-10-03 11:49       ` Christoph Hellwig
2017-10-03 16:01         ` Hannes Reinecke
2017-10-02 13:55 ` [PATCH 5/6] nvme: Export subsystems to /sys/class/nvme-subsys Hannes Reinecke
2017-10-02 16:18   ` Christoph Hellwig
2017-10-02 16:53     ` Keith Busch
2017-10-02 16:56       ` Christoph Hellwig
2017-10-02 17:15         ` Keith Busch
2017-10-02 13:55 ` [PATCH 6/6] nvme: ignore retries for multipath devices Hannes Reinecke
2017-10-02 16:22   ` Christoph Hellwig
2017-10-03 10:02     ` Hannes Reinecke
2017-10-03 11:53       ` Christoph Hellwig [this message]
2017-10-02 14:35 ` [PATCH] nvme: reset retires after path failover Johannes Thumshirn
2017-10-02 14:43   ` Keith Busch
2017-10-02 15:08     ` Hannes Reinecke
2017-10-02 14:45   ` Johannes Thumshirn
2017-10-02 16:25     ` Christoph Hellwig
2017-10-02 14:46   ` Johannes Thumshirn
2017-10-02 16:19   ` 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=20171003115355.GB24650@lst.de \
    --to=hch@lst.de \
    /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).