From clint@0lsen.net  Sun Aug  1 17:31:48 2004
Return-Path: <clint@0lsen.net>
Received: from mx1.FreeBSD.org (mx1.freebsd.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP id 83ABD16A4CE
	for <FreeBSD-gnats-submit@freebsd.org>; Sun,  1 Aug 2004 17:31:48 +0000 (GMT)
Received: from rwcrmhc11.comcast.net (rwcrmhc11.comcast.net [204.127.198.35])
	by mx1.FreeBSD.org (Postfix) with ESMTP id 5D6AF43D5A
	for <FreeBSD-gnats-submit@freebsd.org>; Sun,  1 Aug 2004 17:31:48 +0000 (GMT)
	(envelope-from clint@0lsen.net)
Received: from 0lsen.net ([24.20.127.157])
          by comcast.net (rwcrmhc11) with ESMTP
          id <2004080117314701300bfb1ve>; Sun, 1 Aug 2004 17:31:47 +0000
Received: by 0lsen.net (Postfix, from userid 1001)
	id 77964176D8; Sun,  1 Aug 2004 10:31:47 -0700 (PDT)
Message-Id: <20040801173147.77964176D8@0lsen.net>
Date: Sun,  1 Aug 2004 10:31:47 -0700 (PDT)
From: Clint Olsen <clint@0lsen.net>
Reply-To: Clint Olsen <clint@0lsen.net>
To: FreeBSD-gnats-submit@freebsd.org
Cc:
Subject: sk0 hangs moving lots of data
X-Send-Pr-Version: 3.113
X-GNATS-Notify:

>Number:         69879
>Category:       kern
>Synopsis:       sk0 hangs moving lots of data
>Confidential:   no
>Severity:       non-critical
>Priority:       medium
>Responsible:    freebsd-bugs
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Sun Aug 01 17:40:23 GMT 2004
>Closed-Date:    Sat Aug 21 00:02:10 GMT 2004
>Last-Modified:  Tue Oct 12 16:20:22 GMT 2004
>Originator:     Clint Olsen
>Release:        FreeBSD 5.2.1-RELEASE i386
>Organization:
>Environment:
System: FreeBSD belle.0lsen.net 5.2.1-RELEASE FreeBSD 5.2.1-RELEASE #0: Mon Feb 23 20:45:55 GMT 2004 root@wv1u.btc.adaptec.com:/usr/obj/usr/src/sys/GENERIC i386
    skc0: Yukon Gigabit Ethernet 10/100/1000Base-T Adapter
    sk0: <Marvell Semiconductor, Inc. Yukon> on skc0
    sk0: Ethernet address: 00:0e:a6:96:8f:72
    miibus0: <MII bus> on sk0
    e1000phy0: <Marvell 88E1000 Gigabit PHY> on miibus0
    e1000phy0:  10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, 1000baseTX-FDX, auto
>Description:
    Network (sk0) hangs when moving a large amount of data across the
    interface.  What I'm looking for is help on how to track down whether
    this is a hardware or a software problem.
>How-To-Repeat:
    Using ssh and tar, I copied over about 2G of data from one FreeBSD box
    to another via:

    source% tar cf - . | ssh target "tar xvf - -p --atime-preserve"

    Eventually the network hangs.
>Fix:
    Currently when the network hangs, I am able to take the network down
    and immediately bring it back up using 'ifconfig sk0 down; ifconfig sk0
    up'
>Release-Note:
>Audit-Trail:

From: Clint Olsen <clint@0lsen.net>
To: FreeBSD-gnats-submit@FreeBSD.org
Cc:  
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Sun, 1 Aug 2004 13:33:11 -0700

 I think I may have some more clues as to what's happening.  In the case
 when I was moving files, I was actually logged into the target machine, and
 I used ssh to get back to the source machine before running the command.
 Since I was copying all the files out of my home directory, I also would
 have clobbered my active .ssh directory and all the files in it.  This may
 have been what led up to the hang.  Nonetheless, I don't think it should
 put the network interface into a tizzy, though...
 
 Thanks,
 
 -Clint

From: Petter =?iso-8859-1?q?Holmstr=F6m?= <peholmst@abo.fi>
To: freebsd-gnats-submit@FreeBSD.org, clint@0lsen.net
Cc:  
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Mon, 16 Aug 2004 15:00:21 +0300

 =2D----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA1
 
 Hello,
 
 We have a similar problem with out onboard NIC (Asus A7V880, Marvell Gigabi=
 t LAN controller in 100 mbps network).
 It hangs every now and then, mostly while transmitting lots of text (logfil=
 es, etc) through SSH.=20
 It has also hung while copying many files over SAMBA, even though this is n=
 ot very common. I've noticed that the best=20
 way to get the network to hung is to SSH into the server and then issue 'yp=
 cat passwd' (will return a few thousand lines).
 I can get the network back up by issuing several 'ifconfig sk0 down' and 'i=
 fconfig sk0 up'
 commands, but I usually have to type these commands at least 10 times.
 
 Our Environment:
 
 =46reeBSD spider3.abo.fi 5.2.1-RELEASE-p9 FreeBSD 5.2.1-RELEASE-p9 #1: Sat =
 Aug 14 05:29:38 EEST 2004     root@spider3.abo.fi:/usr/obj/usr/src/sys/SPID=
 ER  i386
 skc0: <Marvell Gigabit Ethernet> port 0xed00-0xedff mem 0xfe700000-0xfe703f=
 ff irq 18 at device 9.0 on pci0
 skc0: Yukon Gigabit Ethernet 10/100/1000Base-T Adapter
 sk0: <Marvell Semiconductor, Inc. Yukon> on skc0
 sk0: flags=3D8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
         inet 130.232.82.226 netmask 0xfffffc00 broadcast 130.232.83.255
         inet6 fe80::211:2fff:fe11:f139%sk0 prefixlen 64 scopeid 0x1
         ether 00:11:2f:11:f1:39
         media: Ethernet 100baseTX (100baseTX <full-duplex>)
         status: active
 
 =2D -Petter-
 
 =2D --=20
 Petter Holmstr=F6m  <peholmst@abo.fi>
 =C5bo Akademi University
 Axelia, Biskopsgatan 8, 20500 =C5bo, Finland
 Tel +358 2 215 3508,  GSM +358 40 8395765
 Get my GnuPG public key from http://www.abo.fi/~peholmst/
 =2D----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.2.4 (GNU/Linux)
 
 iD8DBQFBIKHV+hJtFPUEs/oRAq4SAJ4hRB8MN/ywz6YyGrm8Qg3NN9fiOACghPmt
 RJ0V1xgshJcZnMYPWMQQuTQ=3D
 =3DoUjV
 =2D----END PGP SIGNATURE-----

From: Clint Olsen <clint@0lsen.net>
To: Petter Holmstr?m <peholmst@abo.fi>
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Mon, 16 Aug 2004 10:13:25 -0700

 Hi:
 
 Thanks for at least confirming the problem :)  I just had an incident last
 night with this, and it does appear that sk0 does come back by itself after
 a couple minutes of hanging (no intervention on my part).  This time it had
 nothing to do with SSH.  I was running some simultaneous downloads off the
 WAN which finally got the network load up high enough that it tripped this
 'bug'.  So, it's probably a low-level networking problem rather than being
 protocol specific.
 
 -Clint

From: "Bjoern A. Zeeb" <bzeeb-lists@lists.zabbadoz.net>
To: freebsd-gnats-submit@FreeBSD.org
Cc:  
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Mon, 16 Aug 2004 17:53:10 +0000 (UTC)

 I also have some problems with if_sk on Asus K8V SE Deluxe, Marvel.
 I only get it "hanging" with outgoing traffic:
 - cvs co/up from another host with server on the host with the sk
 - cvsup, scpt,sftp, ftp, ... --""--
 
 It seems not to be protocol specific.
 On some longer hangs ifconfig sk0 down/up didn't help either
 but clearing arp caches helped (or sk just recovered that moments).

From: Rainer Goellner <rainer@jabbe.de>
To: freebsd-gnats-submit@FreeBSD.org, clint@0lsen.net
Cc:  
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Tue, 17 Aug 2004 22:37:26 +0200

 I've got the same board as Petter and can confirm the problem,
 though with a very low load (just three computers in a home
 network).
 
 As Bjoern, I do have the impression that hangs occur while
 traffic is going out, i.e. getting something per ftp from
 that machine.
 
 I'm not shure about that, but it seems to me making some
 noise the other way round helps, even as little as just
 getting the replies from pings to the router.
 
 BTW, it's a 5.2.1-RELEASE-p9 system.
 
   Rainer

From: Clint Olsen <clint@0lsen.net>
To: Rainer Goellner <rainer@jabbe.de>
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Tue, 17 Aug 2004 16:44:22 -0700

 Well, I can say with certainty that downloading is also a problem running
 cvsup using the standard-supfile got me a load of hanging:
 
 Aug 16 02:52:45 belle kernel: sk0: watchdog timeout
 Aug 16 03:01:11 belle kernel: sk0: watchdog timeout
 Aug 16 03:12:17 belle kernel: sk0: watchdog timeout
 Aug 16 03:21:46 belle kernel: sk0: watchdog timeout
 Aug 17 16:38:46 belle kernel: sk0: watchdog timeout
 
 I just now got back, but cvsup isn't finished so I'm in for more fun an
 excitement.
 
 I think it's time to ditch the onboard ethernet in favor of something
 that's better supported...
 
 -Clint
State-Changed-From-To: open->feedback 
State-Changed-By: jmg 
State-Changed-When: Fri Aug 20 06:23:57 GMT 2004 
State-Changed-Why:  
could you test that it continues to hang?  I transfered over 31gigs of 
data via ssh, and I didn't have any problems...  I also fixed the LOR's.. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=69879 
State-Changed-From-To: feedback->closed 
State-Changed-By: jmg 
State-Changed-When: Sat Aug 21 00:01:10 GMT 2004 
State-Changed-Why:  
reporter confirms that there are no hangs on 6.0-CURRENT (aka HEAD), I plan 
to back port the LOR patches to RELENG_5_3 (w/ re@'s blessings of course).. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=69879 

From: Unreal HSHH <hunreal@gmail.com>
To: freebsd-gnats-submit@FreeBSD.org, clint@0lsen.net
Cc:  
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Tue, 31 Aug 2004 14:39:25 +0800

 I cvsup the src to the lastest.
 But still got sk0 error,
 
 Aug 30 20:32:01 tv kernel: sk0: discard frame w/o leading ethernet
 header (len 0 pkt len 0)
 Aug 30 20:32:01 tv kernel: sk0: discard oversize frame (ether type
 7201 flags 3 len 1738 > max 1514)
 
 My box is,
 FreeBSD my.net 6.0-CURRENT FreeBSD 6.0-CURRENT #0: Mon Aug 30 19:21:19
 CST 2004     root@tv:/usr/obj/usr/src/sys/my  i386

From: Sean Farley <sean-freebsd@farley.org>
To: freebsd-gnats-submit@FreeBSD.org, clint@0lsen.net
Cc: peholmst@abo.fi, rainer@jabbe.de, bzeeb-lists@lists.zabbadoz.net,
	hunreal@gmail.com
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Wed, 1 Sep 2004 16:02:41 -0500 (CDT)

 I was searching for opinions on this motherboard with FreeBSD when I
 came across the bug.  E-mail me separately with words of love or hate
 for this board.  My board (SolTek 75DRV5) is dying from bad caps, and I
 need a replacement.
 
 Back to the bug.  According to this, there was a network hanging bug in
 earlier BIOS versions (prior to 1004.001 beta) for this board:
 http://forums.pcper.com/showthread.php?p=2921031#post2921031
 http://forums.pcper.com/showthread.php?p=2921073#post2921073
 
 Maybe a BIOS upgrade will help, even for other ASUS boards (K8V).
 
 Sean
 -----------------------
 sean-freebsd@farley.org

From: Clint Olsen <clint@0lsen.net>
To: Sean Farley <sean-freebsd@farley.org>
Cc: freebsd-gnats-submit@FreeBSD.org, peholmst@abo.fi,
	rainer@jabbe.de, bzeeb-lists@lists.zabbadoz.net, hunreal@gmail.com
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Wed, 1 Sep 2004 14:09:43 -0700

 IMO of course:
 
 If it were truly BIOS specific, then the 5.3-beta and 6.0-current would not
 have been able to fix this problem.
 
 -Clint

From: Sean Farley <sean-freebsd@farley.org>
To: Clint Olsen <clint@0lsen.net>
Cc: freebsd-gnats-submit@FreeBSD.org, peholmst@abo.fi,
	rainer@jabbe.de, bzeeb-lists@lists.zabbadoz.net, hunreal@gmail.com
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Wed, 1 Sep 2004 16:17:31 -0500 (CDT)

 On Wed, 1 Sep 2004, Clint Olsen wrote:
 
 > IMO of course:
 >
 > If it were truly BIOS specific, then the 5.3-beta and 6.0-current
 > would not have been able to fix this problem.
 
 I missed the "reporter confirms that there are no hangs ..." part.
 Sorry.
 
 Sean
 -----------------------
 sean-freebsd@farley.org

From: Ceri Davies <ceri@submonkey.net>
To: Clint Olsen <clint@0lsen.net>
Cc: FreeBSD Gnats Submit <freebsd-gnats-submit@FreeBSD.org>
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Wed, 1 Sep 2004 22:18:26 +0100

 On Wed, Sep 01, 2004 at 09:10:27PM +0000, Clint Olsen wrote:
 
 >  IMO of course:
 >  
 >  If it were truly BIOS specific, then the 5.3-beta and 6.0-current would not
 >  have been able to fix this problem.
 
 Apparently they haven't.
 
 Ceri

From: Rong-En Fan <rafan@infor.org>
To: freebsd-gnats-submit@FreeBSD.org, clint@0lsen.net
Cc:  
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Tue, 12 Oct 2004 23:45:19 +0800 (CST)

 I have a sk(4), 3C940 and got a hang on 4.10-RELEASE
 while lots of NFS traffic. I have sent a mail to
 -stable, please see the following link.
 
 http://lists.freebsd.org/pipermail/freebsd-stable/2004-October/008885.html
 
 The solution is the same as this one,
 just `ifconfig sk0 down up`
 

From: Clint Olsen <clint@0lsen.net>
To: Rong-En Fan <rafan@infor.org>
Cc: freebsd-gnats-submit@FreeBSD.org
Subject: Re: kern/69879: sk0 hangs moving lots of data
Date: Tue, 12 Oct 2004 09:12:58 -0700

 On Oct 12, Rong-En Fan wrote:
 > 
 > http://lists.freebsd.org/pipermail/freebsd-stable/2004-October/008885.html
 > 
 > The solution is the same as this one,
 > just `ifconfig sk0 down up`
 
 This did not work for me this time...
 
 -Clint
 
>Unformatted:
