From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752868AbZHXQwv (ORCPT ); Mon, 24 Aug 2009 12:52:51 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1752510AbZHXQwt (ORCPT ); Mon, 24 Aug 2009 12:52:49 -0400 Received: from srv5.dvmed.net ([207.36.208.214]:58439 "EHLO mail.dvmed.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752482AbZHXQwt (ORCPT ); Mon, 24 Aug 2009 12:52:49 -0400 Message-ID: <4A92C55F.6080506@garzik.org> Date: Mon, 24 Aug 2009 12:52:47 -0400 From: Jeff Garzik User-Agent: Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.9.1.1) Gecko/20090814 Fedora/3.0-2.6.b3.fc11 Thunderbird/3.0b3 MIME-Version: 1.0 To: John Stoffel CC: Jens Axboe , linux-kernel@vger.kernel.org, benh@kernel.crashing.org, htejun@gmail.com, bzolnier@gmail.com, alan@lxorguk.ukuu.org.uk, akpm@linux-foundation.org Subject: Re: [PATCH 4/7] libata: use lazy workqueues for the pio task References: <1251100616-28604-1-git-send-email-jens.axboe@oracle.com> <1251100616-28604-5-git-send-email-jens.axboe@oracle.com> <4A92C081.50207@garzik.org> <19090.50074.62901.633104@stoffel.org> In-Reply-To: <19090.50074.62901.633104@stoffel.org> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit X-Spam-Score: -4.4 (----) X-Spam-Report: SpamAssassin version 3.2.5 on srv5.dvmed.net summary: Content analysis details: (-4.4 points, 5.0 required) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On 08/24/2009 12:45 PM, John Stoffel wrote: >>>>>> "Jeff" == Jeff Garzik writes: > Jeff> No objections to the code, operationally... > > Jeff> But it is disappointing that the "1 thread on UP" problem is not > Jeff> solved while changing this libata area. Is there no way to > Jeff> specify a minimum lazy-thread count? > > Jeff> A key problem continues to be tying to the number of CPUs, which > Jeff> is quite inappropriate for libata. > > So should the minimum number be the NumATADisks on the system? Actual > or potential? I've got a system with dual CPUs and two IDE disk, two > SATA disks and two SCSI disks, plus two SCSI Tape drives. All on > seperate controllers... how would that work? Technically speaking, the maximum is the number of PIO-polling devices. Theoretically this can change with hotplugging, but that is _very_ rare -- mainly PATA+media bay situations, or bridged SATA with an ancient PATA device. Jeff