From mboxrd@z Thu Jan 1 00:00:00 1970 From: Tejun Heo Subject: Re: sata_mv port lockup on hotplug (kernel 2.6.38.2) Date: Tue, 26 Apr 2011 15:50:27 +0200 Message-ID: <20110426135027.GI878@htj.dyndns.org> References: <4D9CD275.9000002@teksavvy.com> <4D9FACC9.7020200@teksavvy.com> <4DA45CA7.9040102@teksavvy.com> <4DA467FB.6020905@teksavvy.com> <20110423005610.GC1576@mtj.dyndns.org> <20110425162242.GB30828@mtj.dyndns.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from mail-fx0-f46.google.com ([209.85.161.46]:43980 "EHLO mail-fx0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1755515Ab1DZNuc (ORCPT ); Tue, 26 Apr 2011 09:50:32 -0400 Content-Disposition: inline In-Reply-To: Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Bruce Stenning Cc: Mark Lord , "linux-kernel@vger.kernel.org" , "linux-ide@vger.kernel.org" On Tue, Apr 26, 2011 at 02:13:22PM +0100, Bruce Stenning wrote: > Tejun wrote: > > ata4 is reporting link down and the device is detached and port > > disabled afterwards. Doesn't seem to have much to do with > > freezing/thawing? > > After all processing of a normal unplug event has completed, the port is left > thawed (and the irq enabled) to allow future hotplug events on this port to be > actioned. In the case of the port lockups that I have seen the port appears > to be left frozen, with the irq disabled, and I am hypothesising that this is > due to a race between the freeze and the thaw. > > Below is some additional tracing leading up to a port lockup, with some more > DPRINTK statements added. The port freeze appears to be happening after the > port thaw. The irq mask writes that are output by mv_set_main_irq_mask seem > to show the irqs left masked off. Does the following patch resolve the problem? Thanks. diff --git a/drivers/ata/libata-eh.c b/drivers/ata/libata-eh.c index 88cd22f..73f0589 100644 --- a/drivers/ata/libata-eh.c +++ b/drivers/ata/libata-eh.c @@ -2802,19 +2802,6 @@ int ata_eh_reset(struct ata_link *link, int classify, } /* - * Some controllers can't be frozen very well and may set - * spuruious error conditions during reset. Clear accumulated - * error information. As reset is the final recovery action, - * nothing is lost by doing this. - */ - spin_lock_irqsave(link->ap->lock, flags); - memset(&link->eh_info, 0, sizeof(link->eh_info)); - if (slave) - memset(&slave->eh_info, 0, sizeof(link->eh_info)); - ap->pflags &= ~ATA_PFLAG_EH_PENDING; - spin_unlock_irqrestore(link->ap->lock, flags); - - /* * Make sure onlineness and classification result correspond. * Hotplug could have happened during reset and some * controllers fail to wait while a drive is spinning up after