Btrfs: send, fix corrupted path strings for long paths
authorFilipe Manana <fdmanana@gmail.com>
Wed, 21 May 2014 16:38:13 +0000 (17:38 +0100)
committerChris Mason <clm@fb.com>
Fri, 6 Jun 2014 19:00:46 +0000 (12:00 -0700)
If a path has more than 230 characters, we allocate a new buffer to
use for the path, but we were forgotting to copy the contents of the
previous buffer into the new one, which has random content from the
kmalloc call.

Test:

    mkfs.btrfs -f /dev/sdd
    mount /dev/sdd /mnt

    TEST_PATH="/mnt/fdmanana/.config/google-chrome-mysetup/Default/Pepper_Data/Shockwave_Flash/WritableRoot/#SharedObjects/JSHJ4ZKN/s.wsj.net/[[IMPORT]]/players.edgesuite.net/flash/plugins/osmf/advanced-streaming-plugin/v2.7/osmf1.6/Ak#"
    mkdir -p $TEST_PATH
    echo "hello world" > $TEST_PATH/amaiAdvancedStreamingPlugin.txt

    btrfs subvolume snapshot -r /mnt /mnt/mysnap1
    btrfs send /mnt/mysnap1 -f /tmp/1.snap

A test for xfstests follows.

Signed-off-by: Filipe David Borba Manana <fdmanana@gmail.com>
Cc: Marc Merlin <marc@merlins.org>
Tested-by: Marc MERLIN <marc@merlins.org>
Signed-off-by: Chris Mason <clm@fb.com>
fs/btrfs/send.c

index fd38b50..484aaca 100644 (file)
@@ -360,10 +360,13 @@ static int fs_path_ensure_buf(struct fs_path *p, int len)
        /*
         * First time the inline_buf does not suffice
         */
-       if (p->buf == p->inline_buf)
+       if (p->buf == p->inline_buf) {
                tmp_buf = kmalloc(len, GFP_NOFS);
-       else
+               if (tmp_buf)
+                       memcpy(tmp_buf, p->buf, old_buf_len);
+       } else {
                tmp_buf = krealloc(p->buf, len, GFP_NOFS);
+       }
        if (!tmp_buf)
                return -ENOMEM;
        p->buf = tmp_buf;