From: James Bottomley Subject: Re: [PATCH v5] SCSI host lock push-down Date: Tue, 16 Nov 2010 17:36:52 +0000 Message-ID: <1289929012.11405.12.camel@mulgrave.site> References: <1289606118.3015.539.camel@mulgrave.site> <20101113051635.GA11613@havoc.gtf.org> <20101116071029.GA13540@havoc.gtf.org> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-kernel-owner@vger.kernel.org To: Linus Torvalds Cc: Jeff Garzik , "Nicholas A. Bellinger" , Andrew Morton , linux-scsi , linux-kernel List-Id: linux-scsi@vger.kernel.org On Tue, 2010-11-16 at 09:25 -0800, Linus Torvalds wrote: > On Mon, Nov 15, 2010 at 11:10 PM, Jeff Garzik wrote: > > > > Move the mid-layer's ->queuecommand() invocation from being locked > > with the host lock to being unlocked to facilitate speeding up the > > critical path for drivers who don't need this lock taken anyway. > > Looks ok to me. How should I take this? Just as a patch? Or should it > go through the scsi tree? I was thinking through next for a few days to see if anything breaks in the architectures with scsi drivers we can't compile on x86 ... but realistically that will happen in -rcX as well, so I'm happy either way. James