From mboxrd@z Thu Jan 1 00:00:00 1970 From: Jeff Garzik Subject: Re: [PATCH] sata_mv: new exception handling (hotplug, NCQ framework) Date: Fri, 01 Jun 2007 16:17:51 -0400 Message-ID: <46607EEF.8080602@garzik.org> References: <20070519054201.GA32686@havoc.gtf.org> <464EE3E9.80708@wpkg.org> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Return-path: Received: from srv5.dvmed.net ([207.36.208.214]:55425 "EHLO mail.dvmed.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1764166AbXFAUR4 (ORCPT ); Fri, 1 Jun 2007 16:17:56 -0400 In-Reply-To: <464EE3E9.80708@wpkg.org> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Tomasz Chmielewski Cc: linux-ide@vger.kernel.org, LKML , Dave Dillow , dean gaudet Tomasz Chmielewski wrote: > Jeff Garzik schrieb: >> Below is a refresh of my on-going effort to convert sata_mv to the new >> exception handling framework. sata_mv is one of the last hold-outs, and >> its old-EH implementation blocks new features like hotplug and NCQ. >> >> It works for me on the one 50xx and one 60xx card I tested it on, but >> other testers reported regressions, which is why it is not yet upstream. > > Hi, > > If I'm correct, this patch won't make to 2.6.22, and the first possible > inclusion would be 2.6.23? Correct... if that. No guarantee it will make 2.6.23 either, since it is low priority :( > Could you summarize what other regressions were reported? I can't find > much information about sata_mv regressiobs on linux-ide list (at least > when looking at the subjects: lots of patches from you, and two reports > from me). Dave Dillow and dean gaudet responded with reports that the BUG_ON/WARN_ON traps in would trigger: WARNING: at drivers/ata/sata_mv.c:1287 mv_qc_issue() and WARNING: at drivers/ata/sata_mv.c:1333 mv_get_crpb_status() which leads me to think that my sata_mv new-EH patches are accidentally turning on NCQ mode. I have a feeling that I will be able to reproduce this once I attach an NCQ-capable drive and re-test on my own systems, but again, haven't found the time :/ Jeff