From: David Brownell Subject: Re: [patch 2.6.30-rc2 2/2] palm_bk3710: UDMA performance fix Date: Wed, 22 Apr 2009 13:32:24 -0700 Message-ID: <200904221332.25200.david-b@pacbell.net> References: <200904201841.10989.david-b@pacbell.net> <49ED99AE.3020100@ru.mvista.com> <200904222026.11495.bzolnier@gmail.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: Received: from smtp128.sbc.mail.sp1.yahoo.com ([69.147.65.187]:32767 "HELO smtp128.sbc.mail.sp1.yahoo.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S1754977AbZDVUc0 (ORCPT ); Wed, 22 Apr 2009 16:32:26 -0400 In-Reply-To: <200904222026.11495.bzolnier@gmail.com> Content-Disposition: inline Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Bartlomiej Zolnierkiewicz Cc: Sergei Shtylyov , linux-ide@vger.kernel.org, DaVinci On Wednesday 22 April 2009, Bartlomiej Zolnierkiewicz wrote: > On Tuesday 21 April 2009 12:02:22 Sergei Shtylyov wrote: > > David Brownell wrote: > > > > > Fix UDMA throughput bug: tCYC averages t2CYCTYP/2, but the code > > > previously assumed it was the same as t2CYCTYP. > > > > Wow, thanks for finding it! > > .... > > > > > Signed-off-by: David Brownell > > > > The patch itself is: > > > > Acked-by: Sergei Shtylyov > > applied Thanks. Further experimentation suggests (a) doing timing computations in picoseconds, rather than nanoseconds, gives better answers and makes that "slow" system faster by 25% or more; and (b) increasing all the UDMA timings by one clock cycle gives another few percent. Now (a) makes obvious sense; but (b) seems odd ... - Dave