Linux block layer
 help / color / mirror / Atom feed
From: Jens Axboe <axboe@kernel.dk>
To: "linux-block@vger.kernel.org" <linux-block@vger.kernel.org>
Subject: Re: [PATCH] block: retry blk_rq_map_user_iov() on failure
Date: Thu, 31 Oct 2019 13:30:17 -0600	[thread overview]
Message-ID: <fec5226c-d13b-a7ef-cb3a-217b26a1aa22@kernel.dk> (raw)
In-Reply-To: <c8718f10-0442-de4f-edfc-ecc6fe28b768@kernel.dk>

On 10/31/19 1:26 PM, Jens Axboe wrote:
> We ran into an issue in production where reading an NVMe drive log
> randomly fails. The request is a big one, 1056K, and the application
> (nvme-cli) nicely aligns the buffer. This means the kernel will attempt
> to map the pages in for zero-copy DMA, but we ultimately fail adding
> enough pages to the request to satisfy the size requirement as we ran
> out of segments supported by the hardware.
> 
> If we fail a user map that attempts to map pages in directly, retry with
> copy == true set.
> 
> Signed-off-by: Jens Axboe <axboe@kernel.dk>

I did actually add comments to this, but apparently sent out the version
without comments... Current one below, only difference is the addition
of the comment explaining why we retry.


diff --git a/block/blk-map.c b/block/blk-map.c
index 3a62e471d81b..4ca7c5db7d78 100644
--- a/block/blk-map.c
+++ b/block/blk-map.c
@@ -137,6 +137,7 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 	else if (queue_virt_boundary(q))
 		copy = queue_virt_boundary(q) & iov_iter_gap_alignment(iter);
 
+retry:
 	i = *iter;
 	do {
 		ret =__blk_rq_map_user_iov(rq, map_data, &i, gfp_mask, copy);
@@ -154,6 +155,18 @@ int blk_rq_map_user_iov(struct request_queue *q, struct request *rq,
 	__blk_rq_unmap_user(bio);
 fail:
 	rq->bio = NULL;
+	/*
+	 * If we fail and we're in direct map mode, retry once with copying
+	 * enabled instead. This can fix the case where userspace wants to
+	 * do a large IO, but fails because the application memory is
+	 * fragmented and we end up exceeding the number of segments for the
+	 * request. With a copy == true retry, we aren't at the mercy of
+	 * application memory layout.
+	 */
+	if (ret && !copy) {
+		copy = true;
+		goto retry;
+	}
 	return ret;
 }
 EXPORT_SYMBOL(blk_rq_map_user_iov);

-- 
Jens Axboe


  reply	other threads:[~2019-10-31 19:30 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-10-31 19:26 [PATCH] block: retry blk_rq_map_user_iov() on failure Jens Axboe
2019-10-31 19:30 ` Jens Axboe [this message]
2019-11-01  1:35   ` Jens Axboe

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=fec5226c-d13b-a7ef-cb3a-217b26a1aa22@kernel.dk \
    --to=axboe@kernel.dk \
    --cc=linux-block@vger.kernel.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