All of lore.kernel.org
 help / color / mirror / Atom feed
From: Konstantin Khlebnikov <khlebnikov@openvz.org>
To: Hugh Dickins <hughd@google.com>
Cc: <linux-mm@kvack.org>, <linux-kernel@vger.kernel.org>
Subject: [PATCH] tmpfs: fix race between umount and writepage
Date: Tue, 5 Apr 2011 14:34:52 +0400	[thread overview]
Message-ID: <20110405103452.18737.28363.stgit@localhost6> (raw)

shmem_writepage() call igrab() on the inode for the page which is came from
reclaimer to add it later into shmem_swaplist for swap-unuse operation.

This igrab() can race with super-block deactivating process:

shrink_inactive_list()		deactivate_super()
pageout()			tmpfs_fs_type->kill_sb()
shmem_writepage()		kill_litter_super()
				generic_shutdown_super()
				 evict_inodes()
 igrab()
				  atomic_read(&inode->i_count)
				   skip-inode
 iput()
				 if (!list_empty(&sb->s_inodes))
					printk("VFS: Busy inodes after...

To avoid this race after this patch shmem_writepage() also try grab sb->s_active.

If sb->s_active == 0 adding to the shmem_swaplist not required, because
super-block deactivation in progress and swap-entries will be released soon.

Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
---
 mm/shmem.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/mm/shmem.c b/mm/shmem.c
index 58da7c1..1f49c03 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1038,11 +1038,13 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc)
 	struct address_space *mapping;
 	unsigned long index;
 	struct inode *inode;
+	struct super_block *sb;
 
 	BUG_ON(!PageLocked(page));
 	mapping = page->mapping;
 	index = page->index;
 	inode = mapping->host;
+	sb = inode->i_sb;
 	info = SHMEM_I(inode);
 	if (info->flags & VM_LOCKED)
 		goto redirty;
@@ -1083,7 +1085,10 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc)
 		delete_from_page_cache(page);
 		shmem_swp_set(info, entry, swap.val);
 		shmem_swp_unmap(entry);
-		if (list_empty(&info->swaplist))
+		if (!list_empty(&info->swaplist) ||
+				!atomic_inc_not_zero(&sb->s_active))
+			sb = NULL;
+		if (sb)
 			inode = igrab(inode);
 		else
 			inode = NULL;
@@ -1098,6 +1103,8 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc)
 			mutex_unlock(&shmem_swaplist_mutex);
 			iput(inode);
 		}
+		if (sb)
+			deactivate_super(sb);
 		return 0;
 	}
 


WARNING: multiple messages have this Message-ID (diff)
From: Konstantin Khlebnikov <khlebnikov@openvz.org>
To: Hugh Dickins <hughd@google.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org
Subject: [PATCH] tmpfs: fix race between umount and writepage
Date: Tue, 5 Apr 2011 14:34:52 +0400	[thread overview]
Message-ID: <20110405103452.18737.28363.stgit@localhost6> (raw)

shmem_writepage() call igrab() on the inode for the page which is came from
reclaimer to add it later into shmem_swaplist for swap-unuse operation.

This igrab() can race with super-block deactivating process:

shrink_inactive_list()		deactivate_super()
pageout()			tmpfs_fs_type->kill_sb()
shmem_writepage()		kill_litter_super()
				generic_shutdown_super()
				 evict_inodes()
 igrab()
				  atomic_read(&inode->i_count)
				   skip-inode
 iput()
				 if (!list_empty(&sb->s_inodes))
					printk("VFS: Busy inodes after...

To avoid this race after this patch shmem_writepage() also try grab sb->s_active.

If sb->s_active == 0 adding to the shmem_swaplist not required, because
super-block deactivation in progress and swap-entries will be released soon.

Signed-off-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
---
 mm/shmem.c |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

diff --git a/mm/shmem.c b/mm/shmem.c
index 58da7c1..1f49c03 100644
--- a/mm/shmem.c
+++ b/mm/shmem.c
@@ -1038,11 +1038,13 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc)
 	struct address_space *mapping;
 	unsigned long index;
 	struct inode *inode;
+	struct super_block *sb;
 
 	BUG_ON(!PageLocked(page));
 	mapping = page->mapping;
 	index = page->index;
 	inode = mapping->host;
+	sb = inode->i_sb;
 	info = SHMEM_I(inode);
 	if (info->flags & VM_LOCKED)
 		goto redirty;
@@ -1083,7 +1085,10 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc)
 		delete_from_page_cache(page);
 		shmem_swp_set(info, entry, swap.val);
 		shmem_swp_unmap(entry);
-		if (list_empty(&info->swaplist))
+		if (!list_empty(&info->swaplist) ||
+				!atomic_inc_not_zero(&sb->s_active))
+			sb = NULL;
+		if (sb)
 			inode = igrab(inode);
 		else
 			inode = NULL;
@@ -1098,6 +1103,8 @@ static int shmem_writepage(struct page *page, struct writeback_control *wbc)
 			mutex_unlock(&shmem_swaplist_mutex);
 			iput(inode);
 		}
+		if (sb)
+			deactivate_super(sb);
 		return 0;
 	}
 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

             reply	other threads:[~2011-04-05 10:34 UTC|newest]

Thread overview: 34+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-04-05 10:34 Konstantin Khlebnikov [this message]
2011-04-05 10:34 ` [PATCH] tmpfs: fix race between umount and writepage Konstantin Khlebnikov
2011-04-08 12:27 ` Konstantin Khlebnikov
2011-04-08 12:27   ` Konstantin Khlebnikov
2011-04-20 20:04 ` Andrew Morton
2011-04-20 20:04   ` Andrew Morton
2011-04-21  6:37   ` Konstantin Khlebnikov
2011-04-21  6:37     ` Konstantin Khlebnikov
2011-04-21  6:41     ` [PATCH v2] " Konstantin Khlebnikov
2011-04-21  6:41       ` Konstantin Khlebnikov
2011-04-21 19:44       ` Andrew Morton
2011-04-21 19:44         ` Andrew Morton
2011-04-22  4:05         ` Konstantin Khlebnikov
2011-04-22  4:05           ` Konstantin Khlebnikov
2011-05-03 20:06           ` Hugh Dickins
2011-05-03 20:06             ` Hugh Dickins
2011-05-07  5:33             ` Konstantin Khlebnikov
2011-05-07  5:33               ` Konstantin Khlebnikov
2011-05-07 23:56               ` Hugh Dickins
2011-05-07 23:56                 ` Hugh Dickins
2011-05-08 12:51                 ` Konstantin Khlebnikov
2011-05-08 12:51                   ` Konstantin Khlebnikov
2011-05-08 19:36                   ` Hugh Dickins
2011-05-08 19:36                     ` Hugh Dickins
2011-05-10  9:52                     ` Konstantin Khlebnikov
2011-05-10  9:52                       ` Konstantin Khlebnikov
2011-05-10 18:55                       ` Hugh Dickins
2011-05-10 18:55                         ` Hugh Dickins
2011-05-08 19:41                   ` [PATCH 1/3] " Hugh Dickins
2011-05-08 19:41                     ` Hugh Dickins
2011-05-08 19:43                     ` [PATCH 2/3] tmpfs: fix race between umount and swapoff Hugh Dickins
2011-05-08 19:43                       ` Hugh Dickins
2011-05-08 19:45                     ` [PATCH 3/3] tmpfs: fix spurious ENOSPC when racing with unswap Hugh Dickins
2011-05-08 19:45                       ` Hugh Dickins

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=20110405103452.18737.28363.stgit@localhost6 \
    --to=khlebnikov@openvz.org \
    --cc=hughd@google.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.