From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dan Aloni Subject: Re: sata_mv: module reloading doesn't work Date: Tue, 4 Apr 2006 09:08:28 +0300 Message-ID: <20060404060828.GA11996@localdomain> References: <20060402155647.GB20270@localdomain> <311601c90604021059jcdf56e4ja35e3507ab291179@mail.gmail.com> <20060403215729.GA17731@localdomain> <4431DE2B.1020306@rtr.ca> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from noname.neutralserver.com ([70.84.186.210]:32360 "EHLO noname.neutralserver.com") by vger.kernel.org with ESMTP id S1751290AbWDDGHC (ORCPT ); Tue, 4 Apr 2006 02:07:02 -0400 Content-Disposition: inline In-Reply-To: <4431DE2B.1020306@rtr.ca> Sender: linux-ide-owner@vger.kernel.org List-Id: linux-ide@vger.kernel.org To: Mark Lord Cc: "Eric D. Mudama" , Linux Kernel List , Jeff Garzik , Mark Lord , IDE/ATA development list On Mon, Apr 03, 2006 at 10:47:07PM -0400, Mark Lord wrote: > Dan Aloni wrote: > > > > * Normal boot > > * insmod sata_mv > > * all is okay, as expected > > * rmmod sata_mv > > * insmod sata_mv > > * all is bad, as expected > > * kexec > > * insmod sata_mv > > * all is bad! > > > >Conclusion: sata_mv's shutdown does something bad. > > sata_mv seems to just use the default libata shutdown sequence, > so perhaps it's leaving the device in EDMA mode with interrupt > coalescing still on (from the BIOS), and interrupts are still > coming in or something.. > > I suppose it really ought to shut down the device before exiting, > and maybe the default of pci_disable_device() is not enough.. ? BTW, can you tell what firmware version are you using? I'm using rev 9, perhaps it also plays some role in this problem. -- Dan Aloni, Linux specialist XIV LTD, http://www.xivstorage.com da-x@monatomic.org, da-x@colinux.org, da-x@gmx.net, dan@xiv.co.il