From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262469AbTIELgb (ORCPT ); Fri, 5 Sep 2003 07:36:31 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262472AbTIELgb (ORCPT ); Fri, 5 Sep 2003 07:36:31 -0400 Received: from mta05-svc.ntlworld.com ([62.253.162.45]:43224 "EHLO mta05-svc.ntlworld.com") by vger.kernel.org with ESMTP id S262469AbTIELgT (ORCPT ); Fri, 5 Sep 2003 07:36:19 -0400 From: Robert Davies To: linux-kernel@vger.kernel.org Subject: Regression? 2.6.0-test4{,-mm5} SMP interrupt distribution according to /proc/interrupts Date: Fri, 5 Sep 2003 12:36:16 +0100 User-Agent: KMail/1.5.2 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200309051236.17259.rob_davies@ntlworld.com> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Please personally Cc: to me, for any additional information or responses to this post, I am not subscribed to kernel list. Hardware is dual Athlon-MP, AMD762 chipset for most part 2.6.0-test has performed well for me. I've noticed however according to /proc/interrupts, CPU0 is getting all the interrupt service action (like when noapic is set), is this expected? 516rob@ash$ cat proc_interrupts-2.6.0-test4-mm5 + uname -a Linux oak 2.6.0-test4-mm5 #2 SMP Thu Sep 4 18:13:52 BST 2003 i686 AMD Athlon(tm) MP 2000+ AuthenticAMD GNU/Linux + cat /proc/interrupts CPU0 CPU1 0: 354719 93 IO-APIC-edge timer 1: 457 1 IO-APIC-edge i8042 2: 0 0 XT-PIC cascade 8: 52 1 IO-APIC-edge rtc 9: 0 0 IO-APIC-level acpi 12: 20562 1 IO-APIC-edge i8042 14: 6053 1 IO-APIC-edge ide0 16: 7689 1 IO-APIC-level ide2 17: 203 1 IO-APIC-level eth0, mga@PCI:1:5:0 18: 2270 0 IO-APIC-level EMU10K1 19: 0 0 IO-APIC-level ohci-hcd NMI: 0 0 LOC: 354664 354663 ERR: 0 MIS: 0 The CPU1 numbers don't seem to change, except 0: and LOC:. I have ACPI active in both Gentoo with vanilla 2.6.0-test, 2.6.0-test-mm5 and 2.4.22, and also under SuSE with their 2.4.20 kernel, 2.4 and 2.6 are consistent. /proc/cmdline shows root=/dev/hde5 reboot=warm, using acpi=off option makes no change in 2.6. Under 2.4 kernels the interrupts are distributed : 515rob@ash$ cat proc_interrupts-2.4* CPU0 CPU1 0: 25148 137206 IO-APIC-edge timer 1: 6 1076 IO-APIC-edge keyboard 2: 0 0 XT-PIC cascade 8: 65 153 IO-APIC-edge rtc 9: 0 0 IO-APIC-level acpi 14: 6619 22404 IO-APIC-edge ide0 16: 12 373 IO-APIC-level ide2 17: 192602 49262 IO-APIC-level eth0, eth1 18: 0 0 IO-APIC-level EMU10K1 19: 0 0 IO-APIC-level usb-ohci NMI: 0 0 LOC: 162293 162274 ERR: 0 MIS: 0 CPU0 CPU1 0: 70700 62464 IO-APIC-edge timer 1: 2210 2328 IO-APIC-edge keyboard 2: 0 0 XT-PIC cascade 8: 92 91 IO-APIC-edge rtc 9: 0 0 IO-APIC-level acpi 14: 2652 934 IO-APIC-edge ide0 16: 8964 8723 IO-APIC-level ide2 17: 1057 1048 IO-APIC-level eth0 NMI: 0 0 LOC: 133097 133096 ERR: 0 MIS: 0 Regards Rob