From nobody@FreeBSD.org Thu Aug 21 09:10:16 2008 Return-Path: Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id DB74F1065678 for ; Thu, 21 Aug 2008 09:10:16 +0000 (UTC) (envelope-from nobody@FreeBSD.org) Received: from www.freebsd.org (www.freebsd.org [IPv6:2001:4f8:fff6::21]) by mx1.freebsd.org (Postfix) with ESMTP id CB9278FC22 for ; Thu, 21 Aug 2008 09:10:16 +0000 (UTC) (envelope-from nobody@FreeBSD.org) Received: from www.freebsd.org (localhost [127.0.0.1]) by www.freebsd.org (8.14.2/8.14.2) with ESMTP id m7L9AGM1070858 for ; Thu, 21 Aug 2008 09:10:16 GMT (envelope-from nobody@www.freebsd.org) Received: (from nobody@localhost) by www.freebsd.org (8.14.2/8.14.1/Submit) id m7L9AG1K070857; Thu, 21 Aug 2008 09:10:16 GMT (envelope-from nobody) Message-Id: <200808210910.m7L9AG1K070857@www.freebsd.org> Date: Thu, 21 Aug 2008 09:10:16 GMT From: Jason Chambers To: freebsd-gnats-submit@FreeBSD.org Subject: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset X-Send-Pr-Version: www-3.1 X-GNATS-Notify: >Number: 126709 >Category: kern >Synopsis: [ata] Regression of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset [regression] >Confidential: no >Severity: serious >Priority: low >Responsible: freebsd-bugs >State: closed >Quarter: >Keywords: >Date-Required: >Class: sw-bug >Submitter-Id: current-users >Arrival-Date: Thu Aug 21 09:20:02 UTC 2008 >Closed-Date: Wed Jan 28 01:15:06 UTC 2009 >Last-Modified: Wed Jan 28 01:15:06 UTC 2009 >Originator: Jason Chambers >Release: 6.2 >Organization: UCLA >Environment: FreeBSD x.y.z 6.2-RELEASE-p12 FreeBSD 6.2-RELEASE-p12 #0: Thu Jul 17 17:32:02 PDT 2008 asdfs@x.y.z:/usr/obj/usr/src/sys/MYKERN i386 >Description: The previous PR for this issue seems to be abandoned. See http://www.freebsd.org/cgi/query-pr.cgi?pr=kern/122368 for background information. I've also noticed the issue with the amd64 port as well. I was not able to get the 6.2 release of amd64 to install for what I suspect is the same issue. I'm willing to help as much as possible and have a spare drive committed for testing. Thanks. --Jason >How-To-Repeat: Upgrade from 6.2 to >= 6.3 or try to install >= 6.3. >Fix: >Release-Note: >Audit-Trail: From: "Andrey V. Elsukov" To: Jason Chambers Cc: freebsd-gnats-submit@FreeBSD.org Subject: Re: kern/126709: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset Date: Thu, 21 Aug 2008 13:44:41 +0400 Jason Chambers wrote: > I'm willing to help as much as possible and have a spare drive > committed for testing. Hi, Jason. Can you try last CURRENT snapshot, probably this commit fixes your problem: http://lists.freebsd.org/pipermail/cvs-src/2008-July/093202.html You can download one here: ftp://ftp.freebsd.org/pub/FreeBSD/snapshots/200808/ -- WBR, Andrey V. Elsukov From: Jason Chambers To: "Andrey V. Elsukov" Cc: freebsd-gnats-submit@FreeBSD.org Subject: Re: kern/126709: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset Date: Thu, 21 Aug 2008 03:16:57 -0700 Jason Chambers wrote: > I'm only seeing ia64 and powerpc snapshots for CURRENT. > I'll keep an eye on this until one is available. Thanks for the reply. --Jason From: Jason Chambers To: "Andrey V. Elsukov" Cc: freebsd-gnats-submit@FreeBSD.org Subject: Re: kern/126709: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset Date: Thu, 21 Aug 2008 03:14:47 -0700 Andrey V. Elsukov wrote: > You can download one here: > ftp://ftp.freebsd.org/pub/FreeBSD/snapshots/200808/ I'm only seeing ia64 and powerpc snapshots for CURRENT. Regards, --Jason From: linimon@lonesome.com (Mark Linimon) To: bug-followup@FreeBSD.org Cc: Subject: Re: kern/126709: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset Date: Thu, 2 Oct 2008 13:13:54 -0500 ----- Forwarded message from Jason Chambers ----- From: Jason Chambers To: Garrett Cooper Cc: freebsd-bugs@freebsd.org Subject: Re: kern/126709: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset Jason Chambers wrote: > Once the 200808 snapshots are generated I'll try both i386 and amd64 and > report the results. > 7.1-BETA i386 boots fine on this system. Thanks to all involved. Regards, --Jason ----- End forwarded message ----- State-Changed-From-To: open->patched State-Changed-By: linimon State-Changed-When: Thu Oct 2 18:45:12 UTC 2008 State-Changed-Why: Submitter notes the latest 7.x seems to have fixed the problem. http://www.freebsd.org/cgi/query-pr.cgi?pr=126709 From: Jason Chambers To: Garrett Cooper Cc: bug-followup@FreeBSD.org Subject: Re: kern/126709: Regession of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset Date: Thu, 02 Oct 2008 21:47:17 -0700 Looks like I spoke too soon. Intense disk usage seems to slow everything down. On the extreme end it seems I can lock up the system by compiling a kernel while doing a sftp transfer of 20MB/sec. It will still respond to pings but no keyboard or ssh access. --Jason From: Jason Chambers To: bug-followup@FreeBSD.org, jchambers@ucla.edu Cc: Subject: Re: kern/126709: [ata] Regression of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset [regression] Date: Mon, 06 Oct 2008 11:28:01 -0700 Hi all. I believe there is a bug in the fix to resolve the regression. Can someone change the status on this from patched to open ? Heavy disk IO slows everything down and on the extreme end it locks up the system. Could this be due to an interrupt storm ? How would I help troubleshoot this ? Thanks. State-Changed-From-To: patched->analyzed State-Changed-By: linimon State-Changed-When: Mon Oct 6 22:36:03 UTC 2008 State-Changed-Why: Submitter notes that the patch does not fix the problem. http://www.freebsd.org/cgi/query-pr.cgi?pr=126709 From: Jason Chambers To: linimon@FreeBSD.org Cc: bug-followup@FreeBSD.org Subject: Re: kern/126709: [ata] Regression of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset [regression] Date: Mon, 24 Nov 2008 18:51:33 +0000 Hello all, Were any changes made between 7.1-BETA-1 and 7.2-BETA-2 that might have resolved this problem ? To recap, 7.1-BETA1 seemed to fix the recognition problem with the controller but heavy disk IO made the system unresponsive. Thanks again, --Jason From: Jason Chambers To: bug-followup@FreeBSD.org Cc: Subject: Re: kern/126709: [ata] Regression of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset [regression] Date: Mon, 12 Jan 2009 18:53:00 -0800 Whatever changes made it into 7.1-RELEASE appear to have mostly fixed the problem. With 7.1 the system runs fine except for the problem described below. After about 7 hours of uptime the system fails due to an interrupt storm between what looks like the nVidia controller, network interface, and USB controller. Changing the em0 interface status to down stops the interrupt storm however it returns as soon as the interface is enabled again. Let me know if you need more data. Thanks, --Jason dmesg log: ============ 16:02:44 -- Starts generating interrupt errors: "interrupt storm detected on "irq21:"; throttling interrupt source" 16:02:45 -- em0: watchdog timeout -- resetting 16:02:45 -- interrupt storm detected.... 16:02:45 -- em0: link state changed to DOWN (...) 16:02:49 -- em0: link state changed to UP The above repeats regularly but is partially squelched by "last message repeated x times" messages. vmstat -i shows: ====================== interrupt total rate irq1: atkbd0 138 0 irq14: ata0 91 0 irq18: fwohci0 3 0 irq19: em0 150708 5 irq21: ohci0+ 853571283 31349 irq22: ehci0+ 267 0 cpu0: timer 54455483 1999 cpu0: timer 54446087 1999 Total 962624060 35354 Checking the dmesg boot log - bzcat /var/log/messages.0.bz2 | grep "irq 21" ================================================ ohci0: mem 0xdfffb000-0xdfffbfff irq 21 at device 11.0 on pci0 atapci1: port 0xfe00-0xfe07,0xfe10-0xfe13,0xfe20-0xfe27,0xfe30-0xfe33,0xfea0-0xfeaf mem 0xdfffd000-0xdfffdfff irq 21 at device 16.0 on pci0 Network card boot message: ============================ em0: port 0xdcc0-0xdcff mem 0xdcec0000-0xdcedffff,0xdcee0000-0xdcefffff irq 19 at device 3.0 on pci3 From: Jason Chambers To: bug-followup@FreeBSD.org, jchambers@ucla.edu Cc: Subject: Re: kern/126709: [ata] Regression of SATA/PATA support with 6.3 and 7.0 (x86/amd64) -- Nvidia nForce4 chipset [regression] Date: Tue, 27 Jan 2009 15:44:49 -0800 This appears to be a resolved bug. I've successfully used the system for a while now without incident. The previous followup I posted must have been due to another issue altogether and has not repeated since that time. Thanks to all involved. --Jason State-Changed-From-To: analyzed->closed State-Changed-By: linimon State-Changed-When: Wed Jan 28 01:14:35 UTC 2009 State-Changed-Why: Submitter notes the problem is resolved with the latest release. http://www.freebsd.org/cgi/query-pr.cgi?pr=126709 >Unformatted: