mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
To: linux-kernel@vger.kernel.org
Subject: [PATCH] Fix a race condition between ->i_mapping and iput()
Date: Fri, 10 Mar 2006 01:23:25 +0900	[thread overview]
Message-ID: <877j73ziwy.fsf@duaron.myhome.or.jp> (raw)

Hi,

This race became a cause of oops, and can reproduce by the following.

    while true; do
	dd if=/dev/zero of=/dev/.static/dev/hdg1 bs=512 count=1000 & sync
    done


This race condition was between __sync_single_inode() and iput().

          cpu0 (fs's inode)                 cpu1 (bdev's inode)
------------------------------------------------------------------------
                                       close("/dev/hda2")
                                       [...]
__sync_single_inode()
   /* copy the bdev's ->i_mapping */
   mapping = inode->i_mapping;

                                       generic_forget_inode()
                                          bdev_clear_inode()
					     /* restre the fs's ->i_mapping */
				             inode->i_mapping = &inode->i_data;
				          /* bdev's inode was freed */
                                          destroy_inode(inode);

   if (wait) {
      /* dereference a freed bdev's mapping->host */
      filemap_fdatawait(mapping);  /* Oops */

Since __sync_signle_inode() is only taking a ref-count of fs's inode,
the another process can be close() and freeing the bdev's inode while
writing fs's inode.  So, __sync_signle_inode() accesses the freed
->i_mapping, oops.

This patch takes ref-count of bdev's inode for fs's inode before
setting a ->i_mapping, and the clear_inode() of fs's inode does iput().
So, if fs's inode is still living, bdev's inode shouldn't be freed.

This lifetime rule may be a poor, but very simple.

Umm... should we use an another rule to free it more early?
(e.g. if bdev's inode become I_FREEING, it should call bd_forget()
before releasing the inode_lock. And some place should call
igrab(->i_mapping->host->i_count) and iput())


What do you think, comment?

Signed-off-by: OGAWA Hirofumi <hirofumi@mail.parknet.co.jp>
---

 fs/block_dev.c |   21 +++++++++++++++++++--
 1 file changed, 19 insertions(+), 2 deletions(-)

diff -puN fs/block_dev.c~i_mapping-race-fix-2 fs/block_dev.c
--- linux-2.6/fs/block_dev.c~i_mapping-race-fix-2	2006-03-09 03:08:54.000000000 +0900
+++ linux-2.6-hirofumi/fs/block_dev.c	2006-03-09 03:08:54.000000000 +0900
@@ -441,13 +441,22 @@ static struct block_device *bd_acquire(s
 	spin_unlock(&bdev_lock);
 	bdev = bdget(inode->i_rdev);
 	if (bdev) {
+		struct block_device *old = NULL;
+
+		BUG_ON(inode->i_sb == blockdev_superblock);
 		spin_lock(&bdev_lock);
-		if (inode->i_bdev)
+		if (inode->i_bdev) {
+			old = inode->i_bdev;
 			__bd_forget(inode);
+		}
+		atomic_inc(&bdev->bd_inode->i_count);
 		inode->i_bdev = bdev;
 		inode->i_mapping = bdev->bd_inode->i_mapping;
 		list_add(&inode->i_devices, &bdev->bd_inodes);
 		spin_unlock(&bdev_lock);
+
+		if (old)
+			iput(old->bd_inode);
 	}
 	return bdev;
 }
@@ -456,10 +465,18 @@ static struct block_device *bd_acquire(s
 
 void bd_forget(struct inode *inode)
 {
+	struct block_device *old = NULL;
+
 	spin_lock(&bdev_lock);
-	if (inode->i_bdev)
+	if (inode->i_bdev) {
+		if (inode->i_sb != blockdev_superblock)
+			old = inode->i_bdev;
 		__bd_forget(inode);
+	}
 	spin_unlock(&bdev_lock);
+
+	if (old)
+		iput(old->bd_inode);
 }
 
 int bd_claim(struct block_device *bdev, void *holder)
_

             reply	other threads:[~2006-03-09 16:23 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-03-09 16:23 OGAWA Hirofumi [this message]
2006-03-10  4:27 ` Andrew Morton
2006-03-10 15:30   ` OGAWA Hirofumi
2006-03-11 16:00 OGAWA Hirofumi
2006-03-16 22:49 ` Andrew Morton
2006-03-17 14:13   ` OGAWA Hirofumi

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=877j73ziwy.fsf@duaron.myhome.or.jp \
    --to=hirofumi@mail.parknet.co.jp \
    --cc=linux-kernel@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

all inboxes | Powered by JetHome®