xfs
[Top] [All Lists]

Re: [PATCH] xfs: i_state of inode is changed after the inode is freed

To: David Chinner <dgc@xxxxxxx>
Subject: Re: [PATCH] xfs: i_state of inode is changed after the inode is freed
From: Masayuki Saito <m-saito@xxxxxxxxxxxxxx>
Date: Fri, 7 Jul 2006 21:41:31 +0900
Cc: xfs@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx
In-reply-to: <20060704204145.GU15160733@xxxxxxxxxxxxxxxxx>
References: <20060704204145.GU15160733@xxxxxxxxxxxxxxxxx>
Sender: xfs-bounce@xxxxxxxxxxx
Thank you for comments.

>You'd be talking about xfs_iunpin(), wouldn't you ;)
Yes, of course.

>http://kernel.org/git/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commit;h=714250879ea61cdb1a39bb96fe9d934ee0c669a2
>
>This fixed the reproducable test case I had for the problem.
>Can you see if it fixes your problem as well?
We applied the above TAKE to linux-2.6.17.1 and tested it.
However, we confirm the case that i_state of the inode was changed
when the inode was freed in xfs filesystem.

We think that the TAKE reduces the occurrence only.
And we think that our patch fixes the problem.

Could you review our patch again?


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