From nobody@FreeBSD.org  Wed Dec 28 06:58:44 2005
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 35B5A16A420
	for <freebsd-gnats-submit@FreeBSD.org>; Wed, 28 Dec 2005 06:58:44 +0000 (GMT)
	(envelope-from nobody@FreeBSD.org)
Received: from www.freebsd.org (www.freebsd.org [216.136.204.117])
	by mx1.FreeBSD.org (Postfix) with ESMTP id C000443D53
	for <freebsd-gnats-submit@FreeBSD.org>; Wed, 28 Dec 2005 06:58:43 +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 jBS6whuE041900
	for <freebsd-gnats-submit@FreeBSD.org>; Wed, 28 Dec 2005 06:58:43 GMT
	(envelope-from nobody@www.freebsd.org)
Received: (from nobody@localhost)
	by www.freebsd.org (8.13.1/8.13.1/Submit) id jBS6whaG041899;
	Wed, 28 Dec 2005 06:58:43 GMT
	(envelope-from nobody)
Message-Id: <200512280658.jBS6whaG041899@www.freebsd.org>
Date: Wed, 28 Dec 2005 06:58:43 GMT
From: Joo <joao@matik.com.br>
To: freebsd-gnats-submit@FreeBSD.org
Subject: sk connection shut down after some time of inactivity
X-Send-Pr-Version: www-2.3

>Number:         91000
>Category:       kern
>Synopsis:       [sk] sk connection shut down after some time of inactivity
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    bz
>State:          closed
>Quarter:        
>Keywords:       
>Date-Required:  
>Class:          sw-bug
>Submitter-Id:   current-users
>Arrival-Date:   Wed Dec 28 07:00:26 GMT 2005
>Closed-Date:    Sun Jun 18 00:50:35 GMT 2006
>Last-Modified:  Sun Jun 18 00:50:35 GMT 2006
>Originator:     Joo
>Release:        releng_6
>Organization:
Infomatik
>Environment:
FreeBSD gw.lucenet.com.br 6.0-STABLE FreeBSD 6.0-STABLE #0: Thu Dec 22 17:14:58 BRST 2005     hmm@gw.matik.com.br:/usr/obj/mnt/src/sys/wipgw-64  amd64
>Description:
              after some time of network inactivity the sk interface stops comunicating showing:

        media: Ethernet 100baseTX (none)
        
and in messages:

Dec 26 03:37:44 gw kernel: sk0: watchdog timeout
Dec 26 03:37:44 gw kernel: EXT: link state changed to DOWN

**** on 6.0-R I got also watchdog timeout and link down but not media-none and the board still did rx/tx and after some time i got in messages: link state changed to UP


I tried dev.skc.0.int_mod with other higher values what do not make any difference


the interface is a onboard nic on Asus A8V

skc0: <Marvell Gigabit Ethernet> port 0xb400-0xb4ff mem 0xfa500000-0xfa503fff irq 17 at device 10.0 on pci0
skc0: Marvell Yukon Lite Gigabit Ethernet rev. (0x9)
sk0: <Marvell Semiconductor, Inc. Yukon> on skc0


skc0@pci0:10:0: class=0x020000 card=0x811a1043 chip=0x432011ab rev=0x13 hdr=0x00
    vendor   = 'Marvell Semiconductor (Was: Galileo Technology Ltd)'
    device   = '88E8001/8003/8010 Gigabit Ethernet Controller with Integrated PHY (copper)'
    class    = network
    subclass = ethernet

>How-To-Repeat:
              
>Fix:
              
>Release-Note:
>Audit-Trail:
Responsible-Changed-From-To: freebsd-bugs->bz 
Responsible-Changed-By: bz 
Responsible-Changed-When: Wed Dec 28 09:36:18 UTC 2005 
Responsible-Changed-Why:  
I'll take it. More information and more up to date then in i386/67818. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=91000 

From: JoaoBR <joao@matik.com.br>
To: bug-followup@freebsd.org, joao@matik.com.br
Cc:  
Subject: Re: kern/91000: [sk] sk connection shut down after some time of inactivity
Date: Fri, 30 Dec 2005 06:57:15 -0200

 Hi, don't know if this helps but with dev.skc.0.int_mod=3D2000  the interfa=
 ce do=20
 not stop anymore but still get this in strange intervals
 
 Dec 28 04:59:32 gw kernel: sk0: watchdog timeout
 Dec 28 04:59:32 gw kernel: EXT: link state changed to DOWN
 Dec 28 20:00:24 gw kernel: EXT: link state changed to UP
 Dec 29 02:44:31 gw kernel: sk0: watchdog timeout
 Dec 29 02:44:31 gw kernel: EXT: link state changed to DOWN
 Dec 30 00:45:50 gw kernel: EXT: link state changed to UP
 
 Perhaps a short time this two days to say this is a valid workaround
 
 Jo=E3o
 
 
 
 
 
 
 
 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
 Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br

From: JoaoBR <joao@matik.com.br>
To: bug-followup@freebsd.org, joao@matik.com.br
Cc:  
Subject: Re: kern/91000: [sk] sk connection shut down after some time of inactivity
Date: Sat, 7 Jan 2006 10:44:25 -0200

 Hi
 
 as I guessed in my former messages it was a short time
 in fact setting dev.skc.0.int_mod to any higher value does cause less=20
 interface failures but they still ocurre in more random manner.
 this happens in idle times, sometimes imediatly and sometimes after an hour=
  or=20
 two, ifconfig shows no carrier then
 
 Jo=E3o
 
 
 
 
 
 
 
 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
 Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br

From: JoaoBR <joao@matik.com.br>
To: bug-followup@freebsd.org, joao@matik.com.br
Cc:  
Subject: Re: kern/91000: [sk] sk connection shut down after some time of inactivity
Date: Sat, 7 Jan 2006 10:51:51 -0200

 I was not remembering and just found some servers with the same hardware an=
 d=20
 5.4-R running as i386 and amd64
 
 they do not have this problem, no problem at all with the sk driver
 
 
 Jo=E3o
 
 
 
 
 
 
 
 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
 Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br

From: Pyun YongHyeon <pyunyh@gmail.com>
To: bug-followup@FreeBSD.org,  joao@matik.com.br
Cc:  
Subject: Re: kern/91000: [sk] sk connection shut down after some time of inactivity
Date: Wed, 18 Jan 2006 11:12:23 +0900

 There is a modified sk(4) driver available at
 http://people.freebsd.org/~yongari/sk/if_sk.c
 http://people.freebsd.org/~yongari/sk/if_skreg.h
 
 Would you please give it try?
 
 -- 
 Regards,
 Pyun YongHyeon

From: JoaoBR <joao@matik.com.br>
To: pyunyh@gmail.com
Cc: bug-followup@freebsd.org
Subject: Re: kern/91000: [sk] sk connection shut down after some time of inactivity
Date: Wed, 18 Jan 2006 20:36:52 -0200

 On Wednesday 18 January 2006 00:12, Pyun YongHyeon wrote:
 > There is a modified sk(4) driver available at
 > http://people.freebsd.org/~yongari/sk/if_sk.c
 > http://people.freebsd.org/~yongari/sk/if_skreg.h
 >
 > Would you please give it try?
 
 
 Hi, compiled fine and its up now for 12 hours without any event
 I keep watching and let you know
 
 what you say getting watchdog timeout on fxp I have the same problem with=20
 other cards I tried, xl and rl either w/wo polling on, these are following=
 =20
 imediatly when sk starts - however - this is happening for me only on SMP=20
 kernel (amd64) even causing sudden panic crashs - on UP kernel the sk does =
 it=20
 without affecting any other card on the system
 
 thanks
 Jo=E3o
 
 
 
 
 
 
 
 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
 Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br

From: JoaoBR <joao@matik.com.br>
To: pyunyh@gmail.com
Cc: bug-followup@freebsd.org
Subject: Re: kern/91000: [sk] sk connection shut down after some time of inactivity
Date: Thu, 19 Jan 2006 08:19:52 -0200

 On Wednesday 18 January 2006 23:03, Pyun YongHyeon wrote:
 >  > Hi, compiled fine and its up now for 12 hours without any event
 >  > I keep watching and let you know
 >
 
 24h without any event, seems I get sleep this weekend thank's to you!
 
 >
 > It seems that current sk(4) including mine has two seperate issues.
 > 1. Occasional watchdog timeout.
 >  This is a Tx issue and I hadn't encounterd it. Since many users reports
 >  this kind of error I guess it still happens in certain environments.
 >  I have an idea to fix this one. Would be included in next version of
 >  my patch.
 >
 
 I do not know if it helps, we analised here better and we discovered that w=
 e=20
 got much more watchdogs on cards connected to older /10 hubs and with simpl=
 e=20
 straigh through pinned TP cables. We changed to 10/100 HUBs and corrected t=
 he=20
 cables and most watchdogs went away, we only got that ones which are follow=
 ed=20
 by sk0 down. With GB port switches I never got adtional watchdog timeouts.=
 =20
 
 
 > 2. Your report for the issue on SMP indicates possible (not known) locking
 >    bug in my patch. If you can get crash dump it would be really great
 >    to narrow down the cause of the issue. Of course, you should compile
 >    your kernel with debugging symbol.
 
 
 hum, I know but indeed it happens with all former drivers up to the last=20
 dez/05 update in releng_6 with SMP. It seems that your driver update resolv=
 ed=20
 this issue for me. But let's not sing too loud, I am only 24h without=20
 problems now. Soon I get some I let you know - but hope not ;)
 
 thank's, Jo=E3o
 
 
 
 
 
 
 
 
 A mensagem foi scaneada pelo sistema de e-mail e pode ser considerada segura.
 Service fornecido pelo Datacenter Matik  https://datacenter.matik.com.br
State-Changed-From-To: open->patched 
State-Changed-By: bz 
State-Changed-When: Fri Apr 28 08:33:40 UTC 2006 
State-Changed-Why:  
Pyun has committed his work to HEAD. sk now lives in sys/dev/sk 
and the watchdog timeouts should be gone. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=91000 
State-Changed-From-To: patched->closed 
State-Changed-By: bz 
State-Changed-When: Sun Jun 18 00:49:25 UTC 2006 
State-Changed-Why:  
The patches were MFCed to RELENG_6. Thanks to all reporting 
and testing Pyun's update. 

http://www.freebsd.org/cgi/query-pr.cgi?pr=91000 
>Unformatted:
