linux-kernel - Re: [PATCH] aoe: end barrier bios with EOPNOTSUPP

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20090910074820.GS18599@kernel.dk>
Date:	Thu, 10 Sep 2009 09:48:20 +0200
From:	Jens Axboe <jens.axboe@...cle.com>
To:	Ed Cashin <ecashin@...aid.com>
Cc:	alan@...rguk.ukuu.org.uk, akpm@...ux-foundation.org,
	apw@...onical.com, bonbons@...ux-vserver.org,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH] aoe: end barrier bios with EOPNOTSUPP

On Wed, Sep 09 2009, Ed Cashin wrote:
> On Wed Sep  9 16:59:30 EDT 2009, jens.axboe@...cle.com wrote:
> > On Wed, Sep 09 2009, Ed Cashin wrote:
> ...
> > > To implement the barrier, [aoe] would stop generating new AoE write
> > > commands, wait for AoE write responses for all the outstanding write
> > > commands, issue the ATA cache flush command, wait for the response to
> > > the flush, and then resume normal activity.
> > 
> > That's how barriers work with sata to begin with, so I'd very much
> > recommend that you do the same in aoe instead of just not supporting it.
> 
> This patch with EOPNOTSUPP just fixes the regression in 2.6.31-rc9.
> 
> A patch adding support for barriers would be nice.  Since I have
> limited time I tried to motivate that work by using the torture tests
> that Chris Mason (I believe it was) posted a while ago, but I could
> not get any problems to manifest using that torture test.
> 
> If I had a reproducable problem case that the barrier implementation
> could fix, it would be very helpful.

Depending on timing and the other end, it may not be easy to reproduce.
But the problem is indeed real, so I think you should just add the
proper barrier implementation instead of spending too much time trying
to create a reproducable problem. By the time you get there, you could
have fixed it many times over :-)

-- 
Jens Axboe

--
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/