From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261180AbULJLde (ORCPT ); Fri, 10 Dec 2004 06:33:34 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261183AbULJLdd (ORCPT ); Fri, 10 Dec 2004 06:33:33 -0500 Received: from i-194-106-33-239.freedom2surf.net ([194.106.33.239]:15294 "EHLO webmail.freedom2surf.net") by vger.kernel.org with ESMTP id S261180AbULJLdb (ORCPT ); Fri, 10 Dec 2004 06:33:31 -0500 Message-ID: <1102678410.41b9898a485ad@webmail.freedom2surf.net> Date: Fri, 10 Dec 2004 11:33:30 +0000 From: bugzilla@emiller.f2s.com To: linux-kernel@vger.kernel.org Subject: ACPI and APIC on nforce2 with 2.4 kernel - is there hope? MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT User-Agent: Internet Messaging Program (IMP) 3.2.3 X-Originating-IP: 81.171.131.10 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Is there any hope of seeing the acpi_skip_timer_override patch in the 2.4 branch? It has been in the 2.6 branch since around 2.6.5-rcfoo. This patch addresses a problem specific to nforce2 motherboards that results in an XT-PIC timer on ACPI and APIC enabled 2.4.x kernels. I really don't know how much of a Bad Thing (TM) this is but my anxieties are due to: 1) The fact that I see clock gain of about one minute per week; 2) The fact that nforce2 has suffered from notorious instability and I just want to feel comfort that all known problems are going to get ironed out. I acknowledge that this problem is: 1) a hardware problem; 2) fixed in the latest stable branch (as of the 2.6.6 release). If the answer to my question is no, I would appreciate guidance on: 1) whether this issue could cause me problems (is it likely to result in clock gain; could it cause general instability or other problems); 2) what people recommend as a 2.4.x kernel configuration for nforce2 motherboards (mine is Abit NF7-S Rev2) in terms of ACPI, IOAPIC and LAPIC parameters. To demonstrate, with an ACPI and APIC enabled 2.4.26 Debianised kernel, I see: cat /proc/interrupts CPU0 0: 1180749 XT-PIC timer 1: 1834 IO-APIC-edge keyboard 2: 0 XT-PIC cascade 8: 4 IO-APIC-edge rtc 9: 0 IO-APIC-level acpi 12: 234451 IO-APIC-edge PS/2 Mouse 14: 20449 IO-APIC-edge ide0 15: 23 IO-APIC-edge ide1 19: 0 IO-APIC-level mgacore 20: 1192269 IO-APIC-level usb-ohci, eth0 21: 443 IO-APIC-level ehci_hcd, NVidia nForce2 22: 3 IO-APIC-level ohci1394, usb-ohci NMI: 0 LOC: 1180732 ERR: 0 MIS: 0 N.B. I understand that the separate issue of the non-availability of IRQ2 because of the spurious reservation by cascade has now been patched, as of 2.4.27-pre2 (and 2.6.6). See also http://bugzilla.kernel.org/show_bug.cgi?id=1203 Please cc me as I am not subscribed to this list.