[Top] [All Lists]

Re: [PATCH] xfstests: add test for btrfs data corruption when using comp

To: Filipe David Borba Manana <fdmanana@xxxxxxxxx>, <xfs@xxxxxxxxxxx>
Subject: Re: [PATCH] xfstests: add test for btrfs data corruption when using compression
From: Josef Bacik <jbacik@xxxxxx>
Date: Mon, 10 Feb 2014 16:32:50 -0500
Cc: <linux-btrfs@xxxxxxxxxxxxxxx>, <clm@xxxxxx>
Delivered-to: xfs@xxxxxxxxxxx
Dkim-signature: v=1; a=rsa-sha256; c=relaxed/simple; d=fb.com; h=message-id : date : from : mime-version : to : cc : subject : references : in-reply-to : content-type : content-transfer-encoding; s=facebook; bh=5BBQDiDsiP+XX0uzNTLXsMpIeVP7+9Sx7WYGhe7BauI=; b=KUpprHAgZCKr+a6xwBKYKzPQ+bt2cg05TcyweMeX10nD0gpeJ1tIJfUKu+w0V/cKd5nt 3687D8PIuNkUCGrEaDWRqkx0qaUbvXUeiDvhUCjDCFSdJXNRNLxijyEyNGzXLi5fsgr6 3N/1tqx5cMD9p2u5xj9WXhP8GOLibGbCLdw=
In-reply-to: <1391874609-19721-1-git-send-email-fdmanana@xxxxxxxxx>
References: <1391874609-19721-1-git-send-email-fdmanana@xxxxxxxxx>
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:24.0) Gecko/20100101 Thunderbird/24.2.0

On 02/08/2014 10:50 AM, Filipe David Borba Manana wrote:
Test for a btrfs data corruption when using compressed files/extents.
Under certain cases, it was possible for reads to return random data
(content from a previously used page) instead of zeroes. This also
caused partial updates to those regions that were supposed to be filled
with zeroes to save random (and invalid) data into the file extents.

This is fixed by the commit for the linux kernel titled:

    Btrfs: fix data corruption when reading/updating compressed extents

Ran with and without the corresponding fix and all worked as expected. You can add

Reviewed-by: Josef Bacik <jbacik@xxxxxx>



<Prev in Thread] Current Thread [Next in Thread>