On Fri, 2011-03-04 at 16:55 -0800, Greg KH wrote:
> 2.6.37-stable review patch.  If anyone has any objections, please let us know.
> 
> ------------------
> 
> From: NeilBrown <[email protected]>
> 
> commit 93b270f76e7ef3b81001576860c2701931cdc78b upstream.
> 
> There are two cases when we call flush_disk.
> In one, the device has disappeared (check_disk_change) so any
> data will hold becomes irrelevant.
> In the oter, the device has changed size (check_disk_size_change)
> so data we hold may be irrelevant.
[...]
> This regression was introduced by commit 608aeef17a which causes
> check_disk_size_change to call flush_disk, so it is suitable for any
> kernel since 2.6.27.
[...]

Unfortunately this fix appears to depend on at least these changes in
2.6.37, and so is not immediately applicable to earlier versions:

commit 99a38919241fd051b8d93b2e4d0c05ef0556d795
Author: Christoph Hellwig <[email protected]>
Date:   Sat Oct 23 19:07:20 2010 +0200

    fs: fix buffer invalidation in invalidate_list

commit d895a1c96af8c2a0f6a5e0119695a7c6b92df8db
Author: Christoph Hellwig <[email protected]>
Date:   Sun Oct 24 19:40:24 2010 +0200

    fs: do not drop inode_lock in dispose_list
    
commit a031878670ac8fe466859d4c1506bd91ae48678c
Author: Christoph Hellwig <[email protected]>
Date:   Sun Oct 24 19:40:33 2010 +0200

    fs: fold invalidate_list into invalidate_inodes

commit 63997e98a3be68d7cec806d22bf9b02b2e1daabb
Author: Al Viro <[email protected]>
Date:   Mon Oct 25 20:49:35 2010 -0400

    split invalidate_inodes()

Are you working on or have you submitted a backport yet?

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

Attachment: signature.asc
Description: This is a digitally signed message part

_______________________________________________
stable mailing list
[email protected]
http://linux.kernel.org/mailman/listinfo/stable

Reply via email to