Re: [PATCH 13/23] ext4: Add new feature to make block group checksums use metadata_csum algorithm
From: Ted Ts'o <tytso@mit.edu>
Date: 2012-02-13 22:28:28
Also in:
linux-fsdevel, lkml
Possibly related (same subject, not in this thread)
- 2012-01-07 · [PATCH 13/23] ext4: Add new feature to make block group checksums use metadata_csum algorithm · Darrick J. Wong <hidden>
- 2011-12-14 · [PATCH 13/23] ext4: Add new feature to make block group checksums use metadata_csum algorithm · Darrick J. Wong <hidden>
On Mon, Feb 13, 2012 at 02:40:26PM -0700, Andreas Dilger wrote:
If a kernel understands METADATA_CSUM, it will check this first and ignore whether GDT_CSUM is set or not (though it shouldn't ever be set at the same time). Either of these features will cause an older kernel to mount the filesystem read-only, which is all that is needed.
... note that when we check for the uninit bits, this will have to be done if GDT_CSUM || METADATA_CSUM are set. - Ted