From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Sun, 3 Mar 2002 16:16:44 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Sun, 3 Mar 2002 16:16:35 -0500 Received: from smtp02.web.de ([217.72.192.151]:2332 "EHLO smtp.web.de") by vger.kernel.org with ESMTP id ; Sun, 3 Mar 2002 16:16:23 -0500 Message-ID: <3C82929B.48E5E77A@web.de> Date: Sun, 03 Mar 2002 22:16:11 +0100 From: Ulrich Hahn X-Mailer: Mozilla 4.76 (Macintosh; U; PPC) X-Accept-Language: en-US,de MIME-Version: 1.0 To: Charles Briscoe-Smith CC: linux-kernel@vger.kernel.org Subject: Re: PCMCIA-related IDE problems - "hda: lost interrupt" In-Reply-To: <20020302122339.A31436@merry.bs.lan> Content-Type: text/plain; charset=us-ascii; x-mac-type="54455854"; x-mac-creator="4D4F5353" Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Hi there - I was lucky today finding another solution that helped me out: It seems the 2.4 kernel series produce the problem of messing with PCI interrupts. I had no problem with one and the same pcmcia_cs package with a 2.2.9 kernel which locked up ALL of the 2.4 kernels I tried meanwile (I am on 2.4.18 now using the kernel-owned pcmcia modules, yenta_socket, which does not seem to know any parameters at all). Today I found a hint on giving the kernel an irqmask on bootup in the lilo.conf file: append="pci=irqmask=0xafff" This prevents the vital IRQs of ide or mouse from being taken by the PCI bridge when the yenta_socket or i82365 module is loaded. (Unloading the module again would not give back interrupt control - so a reboot was the final step) Charles Briscoe-Smith wrote: > > Personal question: did you find a solution? > > Yes, I did, and it has since been documented in the PCMCIA HOWTO, > section 2.3, subsection "Card readers for desktop systems": > > For Chase CardPORT and Altec ISA card readers using the Cirrus PD6722 > ISA-to-PCMCIA bridge, the i82365 driver should be loaded with a > ``has_ring=0'' parameter to prevent irq 15 conflicts. > > I had been trying the option "has_ring=1", which I didn't know was > the default. > > [ CC'ed linux-kernel so that this gets into its archives. I am not on > linux-kernel so, if replying, please CC me (and, I presume, Ulrich). ] Thanks for your hint!