NetBSD Problem Report #20467

Received: (qmail 8844 invoked by uid 605); 22 Feb 2003 11:50:15 -0000
Message-Id: <200302221150.h1MBoAd01404@bundy.zhadum.de>
Date: Sat, 22 Feb 2003 12:50:10 +0100 (CET)
From: Mario Kemper <magick@bundy.zhadum.de>
Sender: gnats-bugs-owner@netbsd.org
Reply-To: magick@bundy.zhadum.de
To: gnats-bugs@gnats.netbsd.org
Subject: timeout delayed on tl(4) and system slowdown
X-Send-Pr-Version: 3.95

>Number:         20467
>Category:       kern
>Synopsis:       the tl driver slows down the system
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    kern-bug-people
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sat Feb 22 11:51:00 +0000 2003
>Closed-Date:    
>Last-Modified:  Fri Feb 28 20:08:00 +0000 2003
>Originator:     Mario Kemper
>Release:        NetBSD 1.6O  Sources of feb 21 2003
>Organization:
Mario Kemper
magick@zhadum.de
>Environment:


System: NetBSD bundy 1.6O NetBSD 1.6O (BUNDY) #3: Sat Feb 22 10:25:27 CET 2003 root@bundy:/usr/netbsd-src/src-obj/sys/arch/i386/compile/BUNDY i386
Architecture: i386
Machine: i386
>Description:

i installed a Compaq Dual Port Netelligent today. The first thing
i noticed are hundreds of "timeout delayed -x" Messages. And thats a few
hundreds a minute continously.
The second thing is, that the mouse hangs when there's traffic on the line.
I use one port of the card for DSL via pppoe. The second port ist connected
to a switch.

This is what dmesg says:

tl0 at pci3 dev 0 function 0
tl0: Compaq Dual Port Netelligent 10/100 TX
tl0: Ethernet address 00:08:c7:5d:66:80
tl0: interrupting at apic 2 int 2 (irq 5)
nsphy0 at tl0 phy 1: DP83840 10/100 media interface, rev. 1
nsphy0: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
tlphy0 at tl0 phy 31: ThunderLAN 10BASE-T media interface, rev. 5
tlphy0: 10base5
tl1 at pci3 dev 1 function 0
tl1: Compaq Dual Port Netelligent 10/100 TX
tl1: Ethernet address 00:08:c7:5d:66:00
tl1: interrupting at apic 2 int 4 (irq 9)
nsphy1 at tl1 phy 1: DP83840 10/100 media interface, rev. 1
nsphy1: 10baseT, 10baseT-FDX, 100baseTX, 100baseTX-FDX, auto
tlphy1 at tl1 phy 31: ThunderLAN 10BASE-T media interface, rev. 5
tlphy1: 10base5

irq 9 and 5 are shared with the USB controllers which could explain the
mouse hangs:
tl0: interrupting at apic 2 int 2 (irq 5)
tl1: interrupting at apic 2 int 4 (irq 9)
siop0: interrupting at apic 2 int 22 (irq 9)
pciide0: using apic 2 int 23 (irq 9) for native-PCI interrupt
bktr0: interrupting at apic 2 int 21 (irq 5)
uhci1: interrupting at apic 2 int 23 (irq 9)
ohci0: interrupting at apic 2 int 21 (irq 5)
ohci1: interrupting at apic 2 int 22 (irq 9)
ehci0: interrupting at apic 2 int 23 (irq 9)
tl0: interrupting at apic 2 int 2 (irq 5)
tl1: interrupting at apic 2 int 4 (irq 9)
siop0: interrupting at apic 2 int 22 (irq 9)
pciide0: using apic 2 int 23 (irq 9) for native-PCI interrupt
bktr0: interrupting at apic 2 int 21 (irq 5)
uhci1: interrupting at apic 2 int 23 (irq 9)

I also put the card into another slot. The jumpy mouse disappeared.
I did a kernel compile and whenever there was traffic on the line
the compile job almost stopped.

The whole system feels slow when the card is active. I suppose that there
are too many interrupts and that the "timeout delayed" messages are only
a symptom. I recompiled a kernel without DEBUG. The messages disappeared but
the system behaviour remained. When i do "ifconfig tl0 down" the
messages stop.

>How-To-Repeat:
	Insert a Compaq Dual Port Netelligent and use it.
>Fix:
	Not known.
>Release-Note:
>Audit-Trail:

From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: Mario Kemper <magick@bundy.zhadum.de>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Sun, 23 Feb 2003 20:09:03 +0100

 On Sat, Feb 22, 2003 at 12:50:10PM +0100, Mario Kemper wrote:
 > 
 > >Number:         20467
 > >Category:       kern
 > >Synopsis:       the tl driver slows down the system
 > >Confidential:   no
 > >Severity:       critical
 > >Priority:       medium
 > >Responsible:    kern-bug-people
 > >State:          open
 > >Class:          sw-bug
 > >Submitter-Id:   net
 > >Arrival-Date:   Sat Feb 22 03:51:00 PST 2003
 > >Closed-Date:
 > >Last-Modified:
 > >Originator:     Mario Kemper
 > >Release:        NetBSD 1.6O  Sources of feb 21 2003
 > >Organization:
 > Mario Kemper
 > magick@zhadum.de
 > >Environment:
 > 	
 > 	
 > System: NetBSD bundy 1.6O NetBSD 1.6O (BUNDY) #3: Sat Feb 22 10:25:27 CET 2003 root@bundy:/usr/netbsd-src/src-obj/sys/arch/i386/compile/BUNDY i386
 > Architecture: i386
 > Machine: i386
 > >Description:
 > 
 > i installed a Compaq Dual Port Netelligent today. The first thing
 > i noticed are hundreds of "timeout delayed -x" Messages. And thats a few
 > hundreds a minute continously.
 > The second thing is, that the mouse hangs when there's traffic on the line.
 > I use one port of the card for DSL via pppoe. The second port ist connected
 > to a switch.

 I have such cards in a router here (3 of them) and I didn't notice such
 problems.
 What does 'systat vm' report (number of interrupts and CPU usage) when the
 card is active ?

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 23 ans d'experience feront toujours la difference
 --

From: Mario Kemper <magick@netbsd.org>
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Sun, 23 Feb 2003 22:26:22 +0100

 On Sun, Feb 23, 2003 at 08:09:03PM +0100, Manuel Bouyer wrote:
 > On Sat, Feb 22, 2003 at 12:50:10PM +0100, Mario Kemper wrote:
 > > 
 > > >Number:         20467
 > > >Category:       kern
 > > >Synopsis:       the tl driver slows down the system
 > > >Confidential:   no
 > > >Severity:       critical
 > > >Priority:       medium
 > > >Responsible:    kern-bug-people
 > > >State:          open
 > > >Class:          sw-bug
 > > >Submitter-Id:   net
 > > >Arrival-Date:   Sat Feb 22 03:51:00 PST 2003
 > > >Closed-Date:
 > > >Last-Modified:
 > > >Originator:     Mario Kemper
 > > >Release:        NetBSD 1.6O  Sources of feb 21 2003
 > > >Organization:
 > > Mario Kemper
 > > magick@zhadum.de
 > > >Environment:
 > > 	
 > > 	
 > > System: NetBSD bundy 1.6O NetBSD 1.6O (BUNDY) #3: Sat Feb 22 10:25:27 CET 2003 root@bundy:/usr/netbsd-src/src-obj/sys/arch/i386/compile/BUNDY i386
 > > Architecture: i386
 > > Machine: i386
 > > >Description:
 > > 
 > > i installed a Compaq Dual Port Netelligent today. The first thing
 > > i noticed are hundreds of "timeout delayed -x" Messages. And thats a few
 > > hundreds a minute continously.
 > > The second thing is, that the mouse hangs when there's traffic on the line.
 > > I use one port of the card for DSL via pppoe. The second port ist connected
 > > to a switch.
 > 
 > I have such cards in a router here (3 of them) and I didn't notice such
 > problems.
 I suppose you don't have -current on the routers? The card made no
 problems under 1.6beta in a friends system. 
 > What does 'systat vm' report (number of interrupts and CPU usage) when the
 > card is active ?
 > 
 I did a ping flood to my AccessPoint (the only thing currently
 connected) which led to about 1500-2000 Interrupts. 
 CPU load is not rising significantly while i do this.

 -- 
 Mario Kemper
 magick@zhadum.de


From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: Mario Kemper <magick@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Mon, 24 Feb 2003 21:23:54 +0100

 On Sun, Feb 23, 2003 at 10:26:22PM +0100, Mario Kemper wrote:
 > I did a ping flood to my AccessPoint (the only thing currently
 > connected) which led to about 1500-2000 Interrupts. 

 Looks good

 > CPU load is not rising significantly while i do this.

 Well, I don't understant what is causing the problems then ...
 Maybe it's a compatibility problem between your motherboard and the
 tl chip ?

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 23 ans d'experience feront toujours la difference
 --

From: Mario Kemper <magick@netbsd.org>
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
Cc: Mario Kemper <magick@netbsd.org>, gnats-bugs@gnats.netbsd.org,
  netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Mon, 24 Feb 2003 22:16:30 +0100

 On Mon, Feb 24, 2003 at 09:23:54PM +0100, Manuel Bouyer wrote:
 > On Sun, Feb 23, 2003 at 10:26:22PM +0100, Mario Kemper wrote:
 > > I did a ping flood to my AccessPoint (the only thing currently
 > > connected) which led to about 1500-2000 Interrupts. 
 > 
 > Looks good
 > 
 > > CPU load is not rising significantly while i do this.
 > 
 > Well, I don't understant what is causing the problems then ...
 > Maybe it's a compatibility problem between your motherboard and the
 > tl chip ?
 > 
 Its a P4B266 with 845D-Chipset. Nothing real special.
 -- 
 Mario Kemper
 magick@zhadum.de


From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: Mario Kemper <magick@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Wed, 26 Feb 2003 19:43:23 +0100

 On Mon, Feb 24, 2003 at 10:16:30PM +0100, Mario Kemper wrote:
 > On Mon, Feb 24, 2003 at 09:23:54PM +0100, Manuel Bouyer wrote:
 > > On Sun, Feb 23, 2003 at 10:26:22PM +0100, Mario Kemper wrote:
 > > > I did a ping flood to my AccessPoint (the only thing currently
 > > > connected) which led to about 1500-2000 Interrupts. 
 > > 
 > > Looks good
 > > 
 > > > CPU load is not rising significantly while i do this.
 > > 
 > > Well, I don't understant what is causing the problems then ...
 > > Maybe it's a compatibility problem between your motherboard and the
 > > tl chip ?
 > > 
 > Its a P4B266 with 845D-Chipset. Nothing real special.

 I think this is what I have too. I'm not sure about the exact chipset model,
 but it's a an asus with intel P4 (with rambus memory).

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 23 ans d'experience feront toujours la difference
 --

From: Mario Kemper <magick@netbsd.org>
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
Cc: Mario Kemper <magick@netbsd.org>, gnats-bugs@gnats.netbsd.org,
  netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Wed, 26 Feb 2003 22:51:29 +0100

 On Wed, Feb 26, 2003 at 07:43:23PM +0100, Manuel Bouyer wrote:
 > > > 
 > > Its a P4B266 with 845D-Chipset. Nothing real special.
 > 
 > I think this is what I have too. I'm not sure about the exact chipset model,
 > but it's a an asus with intel P4 (with rambus memory).
 > 
 Mine is the SDRAM-Version but i don't think tha really matters. It
 could really be a -current problem.
 -- 
 Mario Kemper
 magick@zhadum.de


From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: Mario Kemper <magick@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Thu, 27 Feb 2003 19:53:23 +0100

 On Wed, Feb 26, 2003 at 10:51:29PM +0100, Mario Kemper wrote:
 > On Wed, Feb 26, 2003 at 07:43:23PM +0100, Manuel Bouyer wrote:
 > > > > 
 > > > Its a P4B266 with 845D-Chipset. Nothing real special.
 > > 
 > > I think this is what I have too. I'm not sure about the exact chipset model,
 > > but it's a an asus with intel P4 (with rambus memory).
 > > 
 > Mine is the SDRAM-Version but i don't think tha really matters. It

 It may, as the chipset is different.

 > could really be a -current problem.

 Can you try a 1.6 kernel ?

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 23 ans d'experience feront toujours la difference
 --

From: Mario Kemper <magick@netbsd.org>
To: Manuel Bouyer <bouyer@antioche.lip6.fr>
Cc: Mario Kemper <magick@netbsd.org>, gnats-bugs@gnats.netbsd.org,
  netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Fri, 28 Feb 2003 01:09:55 +0100

 On Thu, Feb 27, 2003 at 07:53:23PM +0100, Manuel Bouyer wrote:
 > 
 > Can you try a 1.6 kernel ?
 > 
 This is the next thing i want to try. Unfortunately i don't have a
 machine with 1.6. I have to ask a friend to test ist which will take
 some time.
 -- 
 Mario Kemper
 magick@zhadum.de


From: Manuel Bouyer <bouyer@antioche.lip6.fr>
To: Mario Kemper <magick@netbsd.org>
Cc: gnats-bugs@gnats.netbsd.org, netbsd-bugs@netbsd.org
Subject: Re: kern/20467: timeout delayed on tl(4) and system slowdown
Date: Fri, 28 Feb 2003 19:42:31 +0100

 On Fri, Feb 28, 2003 at 01:09:55AM +0100, Mario Kemper wrote:
 > On Thu, Feb 27, 2003 at 07:53:23PM +0100, Manuel Bouyer wrote:
 > > 
 > > Can you try a 1.6 kernel ?
 > > 
 > This is the next thing i want to try. Unfortunately i don't have a
 > machine with 1.6. I have to ask a friend to test ist which will take
 > some time.

 It would be good if you could try it on your P4 too, to rule out hardware
 interraction.

 -- 
 Manuel Bouyer <bouyer@antioche.eu.org>
      NetBSD: 23 ans d'experience feront toujours la difference
 --
>Unformatted:

Home	PR Database Search