Bug 13757

Summary: tl0: adapter check: 180005 mesages keep coming up from kernel
Product: Base System Reporter: abur <abur>
Component: kernAssignee: Bill Paul <wpaul>
Status: Closed FIXED    
Severity: Affects Only Me    
Priority: Normal    
Version: 3.2-RELEASE   
Hardware: Any   
OS: Any   

Description abur 1999-09-15 14:10:01 UTC
	Here is a piece of /var/log/messages:

Sep 13 14:03:26 fw /kernel: tl0: adapter check: 180005
Sep 13 14:06:42 fw /kernel: tl0: adapter check: 180005
Sep 13 14:07:37 fw /kernel: tl0: adapter check: 180005
Sep 13 14:08:19 fw /kernel: xl1: transmission error: 90
Sep 13 14:08:19 fw /kernel: xl1: tx underrun, increasing tx start threshold to 1 20 bytes
Sep 13 14:37:49 fw /kernel: tl0: adapter check: 180005
Sep 13 14:39:20 fw /kernel: tl0: adapter check: 180005
Sep 13 14:40:22 fw /kernel: tl0: adapter check: 180005

	I could not find any reference (web/news-groups) to
	"adapter check: " message. The server rebooted itself a few times.
	I seem to be able to get rid of underruns by increasing
	XL_MIN_FRAMELEN from 60 to 300 in /usr/src/sys/pci/if_xlreg.h
	Reboots did not happen since than, but "adapter check: " messages
	persist.

	Here are versions of driver source files:
# grep Id if_[tx]l*
if_tl.c: *      $Id: if_tl.c,v 1.24.2.5 1999/05/06 15:39:38 wpaul Exp $
if_tlreg.h: *   $Id: if_tlreg.h,v 1.7.2.2 1999/03/31 04:12:29 wpaul Exp $
if_xl.c: *      $Id: if_xl.c,v 1.22.2.12 1999/05/05 17:10:48 wpaul Exp $
if_xlreg.h: *   $Id: if_xlreg.h,v 1.9.2.6 1999/05/05 17:10:49 wpaul Exp $

How-To-Repeat: 
	My guess is that peaks of network traffic cause the problem to show,
	but I do not really know.
Comment 1 ru freebsd_committer freebsd_triage 1999-09-15 14:38:21 UTC
Responsible Changed
From-To: freebsd-bugs->wpaul

Over to the author. 

Comment 2 Bill Paul 1999-09-17 17:56:49 UTC
Of all the gin joints in all the towns in all the world, 
abur@softchoice.com had to walk into mine and say: > 
> 
> 	FreeBSD 3.2-RELEASE on Compaq 850R
> 	NIC Card tl0: Compaq NetFlex-3/P Integrated rev 0x10 int a irq 7
> 		on pci0.16.0
> 	NIC Card xl0: 3Com 3c905-TX Fast Etherlink XL rev 0x00 int a irq 3
> 		on pci0.11.0
> 	NIC Card xl1: 3Com 3c905-TX Fast Etherlink XL rev 0x00 int a irq 11
> 		on pci0.12.0
> 	The server works as IP router/packet filter.

At what speed are all the interfaces running? Are they all at 100Mbps?
Are some 10 and others 100? Which ones?
 
> >Description:
> 
> 	Here is a piece of /var/log/messages:
> 
> Sep 13 14:03:26 fw /kernel: tl0: adapter check: 180005
> Sep 13 14:06:42 fw /kernel: tl0: adapter check: 180005
> Sep 13 14:07:37 fw /kernel: tl0: adapter check: 180005
> Sep 13 14:08:19 fw /kernel: xl1: transmission error: 90
> Sep 13 14:08:19 fw /kernel: xl1: tx underrun, increasing tx start threshold to 1 20 bytes
> Sep 13 14:37:49 fw /kernel: tl0: adapter check: 180005
> Sep 13 14:39:20 fw /kernel: tl0: adapter check: 180005
> Sep 13 14:40:22 fw /kernel: tl0: adapter check: 180005
> 
> 	I could not find any reference (web/news-groups) to
> 	"adapter check: " message.

Well you could read the ThunderLAN Programmer's Guide, but I don't expect
many people to have it handy.

> The server rebooted itself a few times.

No, the server panicked a few times and printed several informational
messages on the console, which you either ignored or decided not to
share with us. Don't do that. When the system panics, it tells you
important things: don't ignore them, and don't keep them to yourself.

> 	I seem to be able to get rid of underruns by increasing
> 	XL_MIN_FRAMELEN from 60 to 300 in /usr/src/sys/pci/if_xlreg.h

Grrr. *Don't* do that. The driver is telling you that it's detected
the transmit underrun condition and is *automatically* adjusting the
TX FIFO threshold for you. If you leave it alone, it will eventually
adjust the threshold high enough to make the problem go away and the
messages will stop.

> 	Reboots did not happen since than, but "adapter check: " messages
> 	persist.

It would be nice if you could also tell us what, if anything, you notice
happening to the system when you see the adapter check messages. Telling
us that you see a message leaves a lot of the imagination: does the 
interface get stuck? Does it stop for a moment and then keep going?

> 	Here are versions of driver source files:
> # grep Id if_[tx]l*
> if_tl.c: *      $Id: if_tl.c,v 1.24.2.5 1999/05/06 15:39:38 wpaul Exp $
> if_tlreg.h: *   $Id: if_tlreg.h,v 1.7.2.2 1999/03/31 04:12:29 wpaul Exp $
> if_xl.c: *      $Id: if_xl.c,v 1.22.2.12 1999/05/05 17:10:48 wpaul Exp $
> if_xlreg.h: *   $Id: if_xlreg.h,v 1.9.2.6 1999/05/05 17:10:49 wpaul Exp $

Well, these are all a little old. What I want you to do is
the following:

- Download if_xl.c and if_xlreg.h from http://www.freebsd.org/~wpaul/3Com/3.0
  and put them in /sys/pci. This is the driver code from 3.3-RELEASE.
- Download if_tl.c and if_tlreg.h from
  http://www.freebsd.org/~wpaul/ThunderLAN/test. This is a special version
  of the ThunderLAN driver with some tweaks which I hope will address your
  problem. Copy these to /sys/pci as well.
- Compile a new kernel with these driver sources and boot it.

Test the system under load and report what happens. Be *detailed*.
Details are good. Some day, they could save your life.

-Bill

-- 
=============================================================================
-Bill Paul            (212) 854-6020 | System Manager, Master of Unix-Fu
Work:         wpaul@ctr.columbia.edu | Center for Telecommunications Research
Home:  wpaul@skynet.ctr.columbia.edu | Columbia University, New York City
=============================================================================
 "It is not I who am crazy; it is I who am mad!" - Ren Hoek, "Space Madness"
=============================================================================
Comment 3 wilko freebsd_committer freebsd_triage 2001-11-24 11:52:21 UTC
State Changed
From-To: open->closed

Given that 
- there was never a followup on wpaul's request for more info 
- the PR dates from 1999 & FreeBSD 3.2R 
- I myself have multiple tl cards running without problems at all. 

this PR is considered dealt with