[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <0C7297FA1D2D244A9C7F6959C0BF1E52027D909B@azsmsx413.amr.corp.intel.com>
Date: Fri, 21 Sep 2007 13:52:24 -0700
From: "Williams, Dan J" <dan.j.williams@...el.com>
To: "Andrew Morton" <akpm@...ux-foundation.org>
Cc: <linux-kernel@...r.kernel.org>, <linux-raid@...r.kernel.org>,
<neilb@...e.de>, <yur@...raft.com>, <saeed.bishara@...il.com>,
"Nelson, Shannon" <shannon.nelson@...el.com>
Subject: RE: [PATCH 2.6.23-rc7 0/3] async_tx and md-accel fixes for 2.6.23
> From: Andrew Morton [mailto:akpm@...ux-foundation.org]
> > Fix a couple bugs and provide documentation for the async_tx api.
> >
> > Neil, please 'ack' patch #3.
> >
> > git://lost.foo-projects.org/~dwillia2/git/iop
async-tx-fixes-for-linus
>
> Well it looks like Neil is on vacation or is hiding from us or
something.
>
> In which case our options would be to go ahead and merge your #3 with
> our fingers crossed, or wait and do it in 2.6.23.1 or .2.
>
> How serious is this bug which you're fixing? Will it affect users
> of "legacy" MD setups, or only those who have enabled fancy dma-engine
> features?
>
The symptom is raid array that fails to complete a read request.
It has only been triggered by Yuri when he tried a dma-engine
configuration on his PPC platform with PAGE_SIZE=64K, this opens a large
window for the bug to hit. The window is much smaller but not 100%
closed in the !dma-engine case, although no one else has hit it.
2.6.22 returned reads from handle_stripe under the stripe lock so this
patch can be seen as 'reverting a botched micro-optimization' versus
'introducing a new way to handle reads'.
> Either way, we need to get #2 (at least) in to Linus.
#2 is less critical and cannot hit with any of the current drivers in
the kernel, so it is ok if this one misses 2.6.23.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists