From: Jan Kara <jack@suse.cz>
Subject: Re: Lazytime feature bugs
Date: Tue, 24 Feb 2015 19:31:03 +0100
Message-ID: <20150224183103.GA13024@quack.suse.cz>
References: <20150218131931.GA10424@quack.suse.cz>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Cc: linux-fsdevel@vger.kernel.org,
	Alexander Viro <viro@zeniv.linux.org.uk>,
	Linus Torvalds <torvalds@linux-foundation.org>
To: tytso@mit.edu
Content-Disposition: inline
In-Reply-To: <20150218131931.GA10424@quack.suse.cz>
Sender: linux-fsdevel-owner@vger.kernel.org

  Hello,

  any reaction here?

								Honza

On Wed 18-02-15 14:19:31, Jan Kara wrote:
>   I had a look at what got merged as a lazytime series from Ted and I have
> found a couple of issues (I think I've pointed them out in my emails back
> in November but admittedly things were somewhat confused back then since
> Ted was submitting new versions pretty fast). Anyway here are the issues
> I've found in the merged code:
> 
> 1) Inode that gets periodically dirtied with I_DIRTY_PAGES, cleaned and
>    dirtied again will have inode with updated timestamps never written due
>    to age since inode->dirtied_when gets reset on each redirtying with
>    I_DIRTY_PAGES.
> 2) The code won't maintain time ordering of b_dirty_time list by
>    inode->dirtied_when - this happens because requeue_inode() moves inode
>    at the head of the b_dirty_time list but inodes in b_io list from which
>    we move are no longer ordered by dirtied_when (due to that list being
>    combined from several lists and also because we sort the list by
>    superblock). As a result terminating logic in move_expired_inodes() may
>    terminate the scan too early for b_dirty_time list.
> 3) This is mostly cosmetic currently but is a potential landmine for
>    future: If you dirty with I_DIRTY_PAGES | I_DIRTY_TIME inode gets filed
>    to b_dirty_time list instead of b_dirty list.
> 4) Another mostly cosmetic issue: move_expired_inodes() should use
>    work->for_sync instead of work->reason == WB_REASON_SYNC (work->reason is
>    there only for tracing purposes).
-- 
Jan Kara <jack@suse.cz>
SUSE Labs, CR