xfs
[Top] [All Lists]

Re: [PATCH v2] xfs: re-enable xfsaild idle mode and fix associated races

To: Dave Chinner <david@xxxxxxxxxxxxx>
Subject: Re: [PATCH v2] xfs: re-enable xfsaild idle mode and fix associated races
From: Christoph Hellwig <hch@xxxxxxxxxxxxx>
Date: Thu, 21 Jun 2012 03:11:36 -0400
Cc: Brian Foster <bfoster@xxxxxxxxxx>, Christoph Hellwig <hch@xxxxxxxxxxxxx>, xfs@xxxxxxxxxxx
In-reply-to: <20120620223548.GK30705@dastard>
References: <1339087793-45731-1-git-send-email-bfoster@xxxxxxxxxx> <20120620080523.GA26167@xxxxxxxxxxxxx> <4FE1F37E.6090706@xxxxxxxxxx> <20120620223548.GK30705@dastard>
User-agent: Mutt/1.5.21 (2010-09-15)
On Thu, Jun 21, 2012 at 08:35:48AM +1000, Dave Chinner wrote:
> writes are single instructions on 64 bit systems. The lock is
> required for 32 bit systems because the write requires separate 32
> bit writes to the LSN which can result in unlocked accesses seeing
> partially updated (and hence incorrect) LSN values.
> 
> So the memory barriers are definitely needed for 64 bit machines
> because there is no locking on the update and spinlocks only provide
> memory barriers via unlock->lock transitions, not via a single
> spin_lock() call.

Indeed.  So we'll either need the barriers, or just always take xa_lock
in xfs_ail_push.  Given that xa_lock and xa_target appear in the same
cache line it probably wouldn't even make much of a difference.

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