From nobody@FreeBSD.org  Thu Dec 23 18:07:13 2004
Return-Path: <nobody@FreeBSD.org>
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 157ED16A4E0
	for <freebsd-gnats-submit@FreeBSD.org>; Thu, 23 Dec 2004 18:07:13 +0000 (GMT)
Received: from www.freebsd.org (www.freebsd.org [216.136.204.117])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 0054843D3F
	for <freebsd-gnats-submit@FreeBSD.org>; Thu, 23 Dec 2004 18:07:13 +0000 (GMT)
	(envelope-from nobody@FreeBSD.org)
Received: from www.freebsd.org (localhost [127.0.0.1])
	by www.freebsd.org (8.13.1/8.13.1) with ESMTP id iBNI7CBO060869
	for <freebsd-gnats-submit@FreeBSD.org>; Thu, 23 Dec 2004 18:07:12 GMT
	(envelope-from nobody@www.freebsd.org)
Received: (from nobody@localhost)
	by www.freebsd.org (8.13.1/8.13.1/Submit) id iBNI7CJp060856;
	Thu, 23 Dec 2004 18:07:12 GMT
	(envelope-from nobody)
Message-Id: <200412231807.iBNI7CJp060856@www.freebsd.org>
Date: Thu, 23 Dec 2004 18:07:12 GMT
From: Peter Trifonov <petert@dcn.infos.ru>
To: freebsd-gnats-submit@FreeBSD.org
Subject: fxp device timeout
X-Send-Pr-Version: www-2.3

>Number:         75441
>Category:       i386
>Synopsis:       [if_fxp]fxp device timeout
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    freebsd-i386
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Thu Dec 23 18:10:31 GMT 2004
>Closed-Date:    Fri Sep 16 07:59:32 GMT 2005
>Last-Modified:  Fri Sep 16 07:59:32 GMT 2005
>Originator:     Peter Trifonov
>Release:        5.3
>Organization:
St.Petersburg State Polytechnic University
>Environment:
FreeBSD firewall.xj.dcn 5.3-RELEASE FreeBSD 5.3-RELEASE #0: Sat Dec 18 11:45:40 MSK 2004     bugman@firewall.xj.dcn:/usr/src/sys/i386/compile/FIREWALL  i386
>Description:
      The system is dual processor PentiumPro with both IDE& SCSI hard drives and 3 NICs xl0,fxp0,fxp1
xl0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        options=9<RXCSUM,VLAN_MTU>
        inet 10.0.103.3 netmask 0xffffff00 broadcast 10.0.103.255
        inet  xxxxxxxxxx netmask 0xfffffff0 broadcast xxxxxxxxxxx
        ether 00:50:da:39:91:99
        media: Ethernet autoselect (100baseTX <full-duplex>)
        status: active
fxp0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        options=8<VLAN_MTU>
        inet xxxxxxxxx netmask 0xffffffc0 broadcast xxxxxxxxxxxxxx
        ether 00:60:b0:67:21:6f
        media: Ethernet autoselect (100baseTX <full-duplex>)
        status: active
fxp1: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
        options=8<VLAN_MTU>
        inet xxxxxxxxxxxxxxxx netmask 0xfffffffc broadcast xxxxxxxxxxxxxxx
        ether 00:60:b0:67:31:ec
        media: Ethernet autoselect (10baseT/UTP)
        status: active

 From time to time both fxp's SIMULTANEOUSLY stop working and produce messages fxp*: device timeout. 
        After doing ifconfig fxp*  down& up the system said 
        fxp1: SCB timeout: 0x70 0x0 0x50 0x0 
        fxp1: SCB timeout: 0x20 0x0 0x50 0x0  
        fxp0: SCB timeout: 0x70 0x0 0x50 0x0
        fxp1: SCB timeout: 0x20 0x0 0x50 0x0

The system resumes normal operation only after reboot.
xl0 does not experience any problems.



>How-To-Repeat:
      The problem occurs sporadically approximately two times per day. 
>Fix:
      
>Release-Note:
>Audit-Trail:

From: Peter Trifonov <petert@dcn.infos.ru>
To: freebsd-gnats-submit@FreeBSD.org, petert@dcn.infos.ru
Cc:  
Subject: Re: i386/75441: fxp device timeout
Date: Sat, 25 Dec 2004 13:43:04 +0300

 After further testing I have found a deterministic way to reproduce the 
 problem. Doing flood ping over both fxp interfaces always causes the system 
 to say "device timeout" on both of them. Moreover, bringing both fxp's down, 
 waiting for a few seconds and bringing them up in most cases fixes the 
 problem.
 
 
 

From: Peter Trifonov <petert@dcn.infos.ru>
To: freebsd-gnats-submit@FreeBSD.org, petert@dcn.infos.ru
Cc:  
Subject: Re: i386/75441: fxp device timeout
Date: Thu, 30 Dec 2004 10:44:25 +0300

 I have replaced fxp0 and fxp1 NICs with 3Com network cards=20
 Now I have xl0 utilizing IRQ 10 and xl1 and xl2 sharing IRQ 11
 After this "fxp{0,1} device timeout" messages changed to "xl{1,2} watchdog=
 =20
 timeout". xl0 still works fine.  Again, xl1 and xl2 can be brought back to=
 =20
 life by doing ifconfig xl{1,2} down and ifconfig xl{1,2} up.
 =46rom this I guess that the problem is not related to fxp or xl drivers, b=
 ut to=20
 the kernel interrupt management system. It seems that on SMP systems IRQ=20
 sharing may cause some of the interrupts to be lost.
 

From: "Peter Trifonov" <pvtrifonov@mail.ru>
To: <freebsd-gnats-submit@FreeBSD.org>
Cc:  
Subject: Re: i386/75441: fxp device timeout
Date: Mon, 31 Jan 2005 15:13:55 +0300

 The problem was fixed by a patch by John Baldwin
 
 See http://lists.freebsd.org/pipermail/freebsd-smp/2005-January/000751.html
 and the related thread.
 
 
 
 With best regards,
 P. Trifonov 
 
State-Changed-From-To: open->closed 
State-Changed-By: linimon 
State-Changed-When: Fri Sep 16 07:59:18 GMT 2005 
State-Changed-Why:  
Submitter notes problem has been fixed. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=75441 
>Unformatted:
