From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Moyer Subject: Re: Updated direct IO optimization patchkit v2 Date: Thu, 18 Aug 2011 13:53:16 -0400 Message-ID: References: <1312259893-4548-1-git-send-email-andi@firstfloor.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org, hch@infradead.org To: Andi Kleen Return-path: In-Reply-To: <1312259893-4548-1-git-send-email-andi@firstfloor.org> (Andi Kleen's message of "Mon, 1 Aug 2011 21:38:02 -0700") Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org Andi Kleen writes: > Various micro optimizations of the block direct IO path. > > This addresses all review comments and adds an additional prefetch > optimization that gives another about 0.3% in a large macro benchmark. > Overall the patchkit is worht about 2.2% in the same macro benchmark. > > I also ported it to current mainline, resolving some conflicts with > hch's recent changes to direct-io. Hi, Andi, If you get rid of the unlikely I pointed out in patch 11 and address my question on patch 6 (about always_inline) and repost, I'll ack the series. I was hoping hch or Jens would chime in on patch 10, but it seems they're busy. Cheers, Jeff