diff mbox

[3.8.y.z,extended,stable] Patch "Btrfs: fix unlock after free on rewinded tree blocks" has been added to staging queue

Message ID 1374015238-27459-1-git-send-email-kamal@canonical.com
State New
Headers show

Commit Message

Kamal Mostafa July 16, 2013, 10:53 p.m. UTC
This is a note to let you know that I have just added a patch titled

    Btrfs: fix unlock after free on rewinded tree blocks

to the linux-3.8.y-queue branch of the 3.8.y.z extended stable tree 
which can be found at:

 http://kernel.ubuntu.com/git?p=ubuntu/linux.git;a=shortlog;h=refs/heads/linux-3.8.y-queue

This patch is scheduled to be released in version 3.8.13.5.

If you, or anyone else, feels it should not be added to this tree, please 
reply to this email.

For more information about the 3.8.y.z tree, see
https://wiki.ubuntu.com/Kernel/Dev/ExtendedStable

Thanks.
-Kamal

------

From 5c1832cf6669db6f5185027e7fb4464526d8b26e Mon Sep 17 00:00:00 2001
From: Jan Schmidt <list.btrfs@jan-o-sch.net>
Date: Sat, 13 Apr 2013 13:19:55 +0000
Subject: Btrfs: fix unlock after free on rewinded tree blocks

commit 47fb091fb787420cd195e66f162737401cce023f upstream.

When tree_mod_log_rewind decides to make a copy of the current tree buffer
for its modifications, it subsequently freed the buffer before unlocking it.
Obviously, those operations are required in reverse order.

Signed-off-by: Jan Schmidt <list.btrfs@jan-o-sch.net>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
Signed-off-by: Kamal Mostafa <kamal@canonical.com>
---
 fs/btrfs/ctree.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

--
1.8.1.2
diff mbox

Patch

diff --git a/fs/btrfs/ctree.c b/fs/btrfs/ctree.c
index ce1c169..f28b330 100644
--- a/fs/btrfs/ctree.c
+++ b/fs/btrfs/ctree.c
@@ -1189,6 +1189,13 @@  __tree_mod_log_rewind(struct extent_buffer *eb, u64 time_seq,
 	btrfs_set_header_nritems(eb, n);
 }

+/*
+ * Called with eb read locked. If the buffer cannot be rewinded, the same buffer
+ * is returned. If rewind operations happen, a fresh buffer is returned. The
+ * returned buffer is always read-locked. If the returned buffer is not the
+ * input buffer, the lock on the input buffer is released and the input buffer
+ * is freed (its refcount is decremented).
+ */
 static struct extent_buffer *
 tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 		    u64 time_seq)
@@ -1222,8 +1229,11 @@  tree_mod_log_rewind(struct btrfs_fs_info *fs_info, struct extent_buffer *eb,
 	}

 	extent_buffer_get(eb_rewin);
+	btrfs_tree_read_unlock(eb);
 	free_extent_buffer(eb);

+	extent_buffer_get(eb_rewin);
+	btrfs_tree_read_lock(eb_rewin);
 	__tree_mod_log_rewind(eb_rewin, time_seq, tm);
 	WARN_ON(btrfs_header_nritems(eb_rewin) >
 		BTRFS_NODEPTRS_PER_BLOCK(fs_info->fs_root));
@@ -2800,15 +2810,9 @@  again:
 				btrfs_clear_path_blocking(p, b,
 							  BTRFS_READ_LOCK);
 			}
+			b = tree_mod_log_rewind(root->fs_info, b, time_seq);
 			p->locks[level] = BTRFS_READ_LOCK;
 			p->nodes[level] = b;
-			b = tree_mod_log_rewind(root->fs_info, b, time_seq);
-			if (b != p->nodes[level]) {
-				btrfs_tree_unlock_rw(p->nodes[level],
-						     p->locks[level]);
-				p->locks[level] = 0;
-				p->nodes[level] = b;
-			}
 		} else {
 			p->slots[level] = slot;
 			unlock_up(p, level, lowest_unlock, 0, NULL);