Bug 182541

Summary: sky2 failure on heavy load
Product: [openSUSE] SUSE Linux 10.1 Reporter: Michael Colligan <mcolligan>
Component: KernelAssignee: Karsten Keil <karsten.keil>
Status: RESOLVED FIXED QA Contact: E-mail List <qa-bugs>
Severity: Normal    
Priority: P1 - Urgent CC: pizza, R.Eggermont, renato_yamane, suse-beta
Version: Final   
Target Milestone: ---   
Hardware: i686   
OS: SuSE Linux 10.1   
Whiteboard:
Found By: Other Services Priority:
Business Priority: Blocker: ---
Marketing QA Status: --- IT Deployment: ---
Bug Depends on: 182512    
Bug Blocks:    
Attachments: Dmesg output
hwinfo --netcard

Description Michael Colligan 2006-06-07 18:01:36 UTC
+++ This bug was initially created as a copy of Bug #182512 +++

The sky2 module fails (hangs) upon heavy network traffic.

It appears this bug can be reolved by using v1.4 (latest in 2.6 git) for this module.  This should also solve Bug#182512 I submitted.
Comment 1 Karsten Keil 2006-06-09 17:13:51 UTC
I build a test kernel, can you please test this from
ftp://ftp.suse.com/pub/people/kkeil/testing/10.1/<arch>/kernel-<flavor>-2.6.16.20-1.8.<arch>.rpm
Comment 2 Michael Colligan 2006-06-09 22:52:37 UTC
(In reply to comment #1)

Upon initial testing the test kernel does indeed appear to solve the heavy load hang problem for the sky2 module.  I have copied several GB's across my LAN a couple of different times without problem.

However, it introduced a far worse bug that relates to  Bug#182512.  I get a kernel panic when I attempt to hibernate the laptop with the test kernel.  Please let me now what other information I can provide but I'd rather not try it again for fear of losing data!
Comment 3 Karsten Keil 2006-06-10 00:04:38 UTC
Hmm, the test kernel has sky2 v1.4, I assume that this problem did not occur
with you modified 10.1 kernel. So it seems the suspend issue is outside the driver itself.
Do you have a log from the panic ?
Comment 4 Michael Colligan 2006-06-10 01:53:13 UTC
(In reply to comment #3)

I don't have a log from the panic.  If you tell me the best way to capture that information I'm willing to try and duplicate the panic.  I was in a hurry to get somewhere when it happened.  Thats why I was hibernating the laptop.  The screen dump of the problem listed the sky2 module as the cause.  This was with the test kernel you provided.  This panic does not occur with the stock ( 2.6.16.13-4-default) 10.1 kernel. But kernel 2.6.16.13-4-default does not resume after hibernate as documented in Bug#182512

Please tell me how I can assist further.
Comment 5 Karsten Keil 2006-06-10 10:51:46 UTC
Easiest way is to use a serial console (if you laptop has a serial port)
console=ttyS0,115200n8
on the bootprompt is enough, you need a NULL modem cable to connect it to a other PC and use some terminal program (e.g. minicom) to capture the output.

If not maybe a digital camera can do the job for a screenshoot.
And again the question:
Did you try to hibernate with the version from your bug description (sky2 mainline 1.4) ? Which exact kernel did you try with sky2 1.4, mainline 2.6.17-rcX-gitY or a patched 10.1 source where you only updated sky2 ?
Comment 6 Michael Colligan 2006-06-10 11:11:01 UTC
(In reply to comment #5)

I used the kernel you provided in comment #1.
Comment 7 Karsten Keil 2006-06-11 13:17:53 UTC
Some misunderstanding, I refer to your initial description:
>It appears this bug can be reolved by using v1.4 (latest in 2.6 git) for this
>module.  This should also solve  Bug#182512 I submitted.

I assume that you did try this version somehow, or did you only found some comments about this e.g. in mailinglists ? I only want to know, why you assume, that 1.4 will fix your problems.

If you try this version (before my test kernel) I want some details what exactly you try.
Comment 8 Renato Yamane 2006-06-12 18:20:35 UTC
I have problem with this module (sky2) when I transfer files by ethernet (samba).
When I transfer files, the network broken and I need restart.
I attachment dmesg output and hwinfo.
Thanks,
Renato
BRAZIL
Comment 9 Renato Yamane 2006-06-12 18:22:09 UTC
Created attachment 88761 [details]
Dmesg output
Comment 10 Renato Yamane 2006-06-12 18:23:37 UTC
Created attachment 88762 [details]
hwinfo --netcard
Comment 11 Michael Colligan 2006-06-13 12:26:14 UTC
(In reply to comment #7)

I read from mailing list and thought this was the version used in Ubuntu 6.06 which works well.  I have not had time to verify the exact version of the module in use in the Ubuntu kernel.  I have not personally recompiled any kernel and attempted the v1.4 driver myself.
Comment 12 Renato Yamane 2006-07-07 20:31:17 UTC
Hi,
I tested Kernel 2.6.16.20-1.8 mentioned in comment #1 by Karsten Keil (from ftp://ftp.suse.com/pub/people/kkeil/testing/10.1/i386/kernel-default-2.6.16.20-1.8.i586.rpm) but the SAME problem still exists.
When I try transfer files with more than 3Mb (e.g. MP3) the ethernet failure.
Best Regards,
Renato Yamane
BRAZIL (out of World Cups ehehehe)
Comment 13 Jose Pissin 2006-07-26 14:08:33 UTC
Ethernet Interface
07:00.0 Ethernet controller: Marvell Technology Group Ltd. 88E8036 PCI-E Fast Ethernet Controller (rev 15)
        Subsystem: Sony Corporation Unknown device 81e6
        Flags: bus master, fast devsel, latency 0, IRQ 185
        Memory at d8000000 (64-bit, non-prefetchable) [size=16K]
        I/O ports at 3000 [size=256]
        Capabilities: [48] Power Management version 2
        Capabilities: [50] Vital Product Data
        Capabilities: [5c] Message Signalled Interrupts: 64bit+ Queue=0/1 Enable-
        Capabilities: [e0] Express Legacy Endpoint IRQ 0
        Capabilities: [100] Advanced Error Reporting

Kernel Instaled
cat /proc/version
Linux version 2.6.16.21-0.13-smp (geeko@buildhost) (gcc version 4.1.0 (SUSE Linux)) #1 SMP Mon Jul 17 17:22:44 UTC 2006

Module Version
ethtool -i eth0
driver: sky2
version: 0.15
firmware-version: N/A
bus-info: 0000:07:00.0


The adapter works, start using networkmanager, but if start the transfer files to another machine, or send printing, or send email to attach file (TX mode) the  adapter died.
I need to down an up interface.

This problem is present in ANY times to transfer files.
Comment 14 Renato Yamane 2006-08-25 12:35:34 UTC
I change priority to P1 because is impossible use my laptop to transfer files to another machine in network!
I can receive files but I can't send files (sky2 crash when I send files).
Best regards,
Renato
Comment 15 Karsten Keil 2006-08-25 13:49:15 UTC
Since we do not have any info about these chips we cannot do so much.
But fortunately a other open source driver from the manufacturer exist (modified sk98lin driver), you can try it from:
ftp://ftp.suse.com/pub/people/kkeil/testing/10.1/
Comment 16 Renato Yamane 2006-08-25 16:50:17 UTC
Keil,
one user fix this problem when turn-off "autonegotiation".
I test this work-arround and realy fix the problem:

# ethtool -i eth1
driver: sky2
version: 1.6
firmware-version: N/A
bus-info: 0000:01:00.0

# ethtool -A eth1 autoneg off rx on tx on

Check if autonegotiate is off:
# ethtool -a eth1
Pause parameters for eth1:
Autonegotiate:  off
RX:             on
TX:             on

Best regards,
Renato S. Yamane
Comment 17 Renato Yamane 2006-10-21 11:19:23 UTC
Patch available in:
<http://bugzilla.kernel.org/show_bug.cgi?id=6839>
--------> "Yukon FE ram buffer initialization" <---------

Tested on Kernel 2.6.18.1 and all work perfectly!

Thanks Stephen Hemminger!

Best regards,
Renato S. Yamane
Comment 18 Robbert Eggermont 2006-12-04 08:36:38 UTC
I got bitten by this bug just last week. Will this patch be applied to 10.1 or do I need to move to 10.2?
Comment 19 Renato Yamane 2006-12-04 09:49:28 UTC
This patch still don't be applied in OpenSuSE 10.1
I think that is necessary wait OpenSuSE 10.2 or compile yourself a new Kernel (>=2.6.18.2 don't necessary use patch, because sky2 is perfect!)
Comment 20 Karsten Keil 2006-12-08 15:05:39 UTC
It's fixed in 10.2 
Comment 21 Renato Yamane 2006-12-13 12:50:54 UTC
I change this bug as FIXED because OpenSuSE 10.2 don't have this problem, BUT OpenSuSE 10.1 still have problem!

For OpenSuSE 10.1 is necessary apply patch available in:
http://bugzilla.kernel.org/show_bug.cgi?id=6839 (Yukon FE ram buffer initialization)
Comment 22 Robbert Eggermont 2006-12-13 14:23:58 UTC
Will this patch be included in the next kernel update for 10.1?
I'm not looking forward to having to reapply this patch after every kernel update (the driver is already broken so introducing this patch can't break it again).