linux-block.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Christoph Hellwig <hch@lst.de>
To: Jens Axboe <axboe@kernel.dk>
Cc: Hillf Danton <hdanton@sina.com>,
	Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>,
	Pavel Tatashin <pasha.tatashin@soleen.com>,
	"Reviewed-by : Tyler Hicks" <tyhicks@linux.microsoft.com>,
	linux-block@vger.kernel.org
Subject: [PATCH 8/8] loop: avoid holding loop_ctl_mutex over add_disk
Date: Thu, 26 Aug 2021 15:38:10 +0200	[thread overview]
Message-ID: <20210826133810.3700-9-hch@lst.de> (raw)
In-Reply-To: <20210826133810.3700-1-hch@lst.de>

To avoid complex lock ordering issues loop_ctl_mutex should not
be held over add_disk.  Add a new Lo_new state for a loop device
that has just been created but which is not live yet.

Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
 drivers/block/loop.c | 25 +++++++++++++++----------
 drivers/block/loop.h |  1 +
 2 files changed, 16 insertions(+), 10 deletions(-)

diff --git a/drivers/block/loop.c b/drivers/block/loop.c
index cb857d5e8313..c8a24e06c259 100644
--- a/drivers/block/loop.c
+++ b/drivers/block/loop.c
@@ -2327,7 +2327,11 @@ static int loop_add(int i)
 	lo = kzalloc(sizeof(*lo), GFP_KERNEL);
 	if (!lo)
 		goto out;
-	lo->lo_state = Lo_unbound;
+	lo->lo_state = Lo_new;
+	atomic_set(&lo->lo_refcnt, 0);
+	mutex_init(&lo->lo_mutex);
+	spin_lock_init(&lo->lo_lock);
+	spin_lock_init(&lo->lo_work_lock);
 
 	err = mutex_lock_killable(&loop_ctl_mutex);
 	if (err)
@@ -2341,9 +2345,11 @@ static int loop_add(int i)
 	} else {
 		err = idr_alloc(&loop_index_idr, lo, 0, 0, GFP_KERNEL);
 	}
-	if (err < 0)
-		goto out_unlock;
 	i = err;
+	lo->lo_number = i;
+	mutex_unlock(&loop_ctl_mutex);
+	if (err < 0)
+		goto out_free_dev;
 
 	err = -ENOMEM;
 	lo->tag_set.ops = &loop_mq_ops;
@@ -2397,11 +2403,6 @@ static int loop_add(int i)
 	if (!part_shift)
 		disk->flags |= GENHD_FL_NO_PART_SCAN;
 	disk->flags |= GENHD_FL_EXT_DEVT;
-	atomic_set(&lo->lo_refcnt, 0);
-	mutex_init(&lo->lo_mutex);
-	lo->lo_number		= i;
-	spin_lock_init(&lo->lo_lock);
-	spin_lock_init(&lo->lo_work_lock);
 	disk->major		= LOOP_MAJOR;
 	disk->first_minor	= i << part_shift;
 	disk->minors		= 1 << part_shift;
@@ -2412,14 +2413,18 @@ static int loop_add(int i)
 	disk->event_flags	= DISK_EVENT_FLAG_UEVENT;
 	sprintf(disk->disk_name, "loop%d", i);
 	add_disk(disk);
-	mutex_unlock(&loop_ctl_mutex);
+
+	mutex_lock(&lo->lo_mutex);
+	lo->lo_state = Lo_unbound;
+	mutex_unlock(&lo->lo_mutex);
+
 	return i;
 
 out_cleanup_tags:
 	blk_mq_free_tag_set(&lo->tag_set);
 out_free_idr:
+	mutex_lock(&loop_ctl_mutex);
 	idr_remove(&loop_index_idr, i);
-out_unlock:
 	mutex_unlock(&loop_ctl_mutex);
 out_free_dev:
 	kfree(lo);
diff --git a/drivers/block/loop.h b/drivers/block/loop.h
index d14ce6bdc014..608a20c23c64 100644
--- a/drivers/block/loop.h
+++ b/drivers/block/loop.h
@@ -24,6 +24,7 @@ enum {
 	Lo_bound,
 	Lo_rundown,
 	Lo_deleting,
+	Lo_new,
 };
 
 struct loop_func_table;
-- 
2.30.2


  parent reply	other threads:[~2021-08-26 13:49 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-26 13:38 sort out the lock order in the loop driver v2 Christoph Hellwig
2021-08-26 13:38 ` [PATCH 1/8] cryptoloop: fix a sparse annotation Christoph Hellwig
2021-08-26 13:38 ` [PATCH 2/8] loop: remove the unused idx argument to loop_control_get_free Christoph Hellwig
2021-08-26 13:38 ` [PATCH 3/8] loop: remove the ->ioctl method in loop_func_table Christoph Hellwig
2021-08-26 13:38 ` [PATCH 4/8] loop: return void from the ->release " Christoph Hellwig
2021-08-26 13:38 ` [PATCH 5/8] loop: merge the cryptoloop module into the main loop module Christoph Hellwig
2021-08-26 16:31   ` Milan Broz
2021-08-26 16:34     ` Christoph Hellwig
2021-08-26 16:44       ` Milan Broz
2021-08-27  6:45         ` Christoph Hellwig
2021-08-27 11:33           ` Tetsuo Handa
2021-08-27 15:38             ` Christoph Hellwig
2021-08-26 13:38 ` [PATCH 6/8] loop: devirtualize transfer transformations Christoph Hellwig
2021-08-26 13:38 ` [PATCH 7/8] loop: move loop device deletion out of loop_ctl_mutex Christoph Hellwig
2021-08-26 13:38 ` Christoph Hellwig [this message]
2021-08-27  0:30 ` sort out the lock order in the loop driver v2 Tetsuo Handa
2021-08-27  6:40   ` Christoph Hellwig
2021-08-27  7:46     ` Tetsuo Handa
2021-08-27 15:34       ` Christoph Hellwig
     [not found] ` <20210827130259.2622-1-hdanton@sina.com>
2021-08-27 14:10   ` Tetsuo Handa
     [not found]   ` <20210828035114.2762-1-hdanton@sina.com>
2021-08-28  5:17     ` Tetsuo Handa

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=20210826133810.3700-9-hch@lst.de \
    --to=hch@lst.de \
    --cc=axboe@kernel.dk \
    --cc=hdanton@sina.com \
    --cc=linux-block@vger.kernel.org \
    --cc=pasha.tatashin@soleen.com \
    --cc=penguin-kernel@i-love.sakura.ne.jp \
    --cc=tyhicks@linux.microsoft.com \
    /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).