Linux RAID subsystem development
 help / color / mirror / Atom feed
From: NeilBrown <neilb@suse.com>
To: Shaohua Li <shli@kernel.org>
Cc: linux-raid@vger.kernel.org
Subject: [md PATCH 2/5] md: don't call bitmap_create() while array is quiesced.
Date: Tue, 17 Oct 2017 13:46:43 +1100	[thread overview]
Message-ID: <150820840345.1646.5139051529060216131.stgit@noble> (raw)
In-Reply-To: <150820826980.1646.6380214598725492144.stgit@noble>

bitmap_create() allocates memory with GFP_KERNEL and
so can wait for IO.
If called while the array is quiesced, it could wait indefinitely
for write out to the array - deadlock.
So call bitmap_create() before quiescing the array.

Signed-off-by: NeilBrown <neilb@suse.com>
---
 drivers/md/md.c |   16 ++++++++++------
 1 file changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/md/md.c b/drivers/md/md.c
index 04538b60f8f3..308456842d3e 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -6619,22 +6619,26 @@ static int set_bitmap_file(struct mddev *mddev, int fd)
 		return -ENOENT; /* cannot remove what isn't there */
 	err = 0;
 	if (mddev->pers) {
-		mddev->pers->quiesce(mddev, 1);
 		if (fd >= 0) {
 			struct bitmap *bitmap;
 
 			bitmap = bitmap_create(mddev, -1);
+			mddev->pers->quiesce(mddev, 1);
 			if (!IS_ERR(bitmap)) {
 				mddev->bitmap = bitmap;
 				err = bitmap_load(mddev);
 			} else
 				err = PTR_ERR(bitmap);
-		}
-		if (fd < 0 || err) {
+			if (err) {
+				bitmap_destroy(mddev);
+				fd = -1;
+			}
+			mddev->pers->quiesce(mddev, 0);
+		} else if (fd < 0) {
+			mddev->pers->quiesce(mddev, 1);
 			bitmap_destroy(mddev);
-			fd = -1; /* make sure to put the file */
+			mddev->pers->quiesce(mddev, 0);
 		}
-		mddev->pers->quiesce(mddev, 0);
 	}
 	if (fd < 0) {
 		struct file *f = mddev->bitmap_info.file;
@@ -6918,8 +6922,8 @@ static int update_array_info(struct mddev *mddev, mdu_array_info_t *info)
 				mddev->bitmap_info.default_offset;
 			mddev->bitmap_info.space =
 				mddev->bitmap_info.default_space;
-			mddev->pers->quiesce(mddev, 1);
 			bitmap = bitmap_create(mddev, -1);
+			mddev->pers->quiesce(mddev, 1);
 			if (!IS_ERR(bitmap)) {
 				mddev->bitmap = bitmap;
 				rv = bitmap_load(mddev);



      parent reply	other threads:[~2017-10-17  2:46 UTC|newest]

Thread overview: 17+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-17  2:46 [md PATCH 0/5] Address deadlock associated with setting suspend_lo NeilBrown
2017-10-17  2:46 ` [md PATCH 1/5] md: always hold reconfig_mutex when calling mddev_suspend() NeilBrown
2017-10-18  6:11   ` Shaohua Li
2017-10-18  7:35     ` NeilBrown
2017-10-19  1:17       ` [md PATCH 1/5 v2] " NeilBrown
2017-10-19  3:45         ` Shaohua Li
2017-10-19  6:29           ` NeilBrown
2017-10-20  4:37             ` Shaohua Li
2017-10-23  0:02               ` NeilBrown
2017-10-23  1:48                 ` Shaohua Li
2017-10-17  2:46 ` [md PATCH 4/5] md: use mddev_suspend/resume instead of ->quiesce() NeilBrown
2017-10-17  2:46 ` [md PATCH 5/5] md: allow metadata update while suspending NeilBrown
2017-10-17  2:46 ` [md PATCH 3/5] md: move suspend_hi/lo handling into core md code NeilBrown
2017-10-18  6:16   ` Shaohua Li
2017-10-18  7:40     ` NeilBrown
2017-10-19  1:49       ` [md PATCH 6/5] md: remove special meaning of ->quiesce(.., 2) NeilBrown
2017-10-17  2:46 ` NeilBrown [this message]

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=150820840345.1646.5139051529060216131.stgit@noble \
    --to=neilb@suse.com \
    --cc=linux-raid@vger.kernel.org \
    --cc=shli@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