Bug 5451

Summary: 2.6.13.x ''locks'' network card.
Product: Drivers Reporter: Arno Wagner (wagner)
Component: NetworkAssignee: Stephen Hemminger (stephen)
Status: REJECTED DUPLICATE    
Severity: normal CC: bunk, romieu, stephen
Priority: P2    
Hardware: i386   
OS: Linux   
Kernel Version: 2.6.13.3 Subsystem:
Regression: --- Bisected commit-id:

Description Arno Wagner 2005-10-15 23:01:27 UTC
Most recent kernel where this bug did not occur: 
   none that I know of

Distribution: 
  Debian etch

Hardware Environment: 
  VIA KT880 (Asus A7V880), Athlon XP 3000+, 
  affected component: 
  On-board  Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller (rev 13)

Problem Description: 
  The Card works fine under Linux. However when I (warm) reboot to Windows XP, it does 
  not detect that a link is present. It is neccessary to turn off the computer, and, I think
  disconnect the PSU from power to restore functionality for Windows. Now, I realise
  that this is probably the MS driver not doing proper initialisation, but the problem first turned 
  up with 2.6.13.1 for me and 2.6.12.2 (last one I used before) works fine, so there was 
  some driver change that causes this new behaviour.

Steps to reproduce:
  Boot to 2.6.13.3, bring up NIC, reboot to Windows, see NIC fail.

Card description from 'lspci':

0000:00:09.0 Ethernet controller: Marvell Technology Group Ltd. 88E8001 Gigabit Ethernet Controller 
(rev 13)
        Subsystem: ASUSTeK Computer Inc. Marvell 88E8001 Gigabit Ethernet Controller (Asus)
        Flags: bus master, 66MHz, medium devsel, latency 64, IRQ 18
        Memory at fe200000 (32-bit, non-prefetchable) [size=16K]
        I/O ports at ed00 [size=256]
        Expansion ROM at fe100000 [disabled] [size=128K]
        Capabilities: [48] Power Management version 2
        Capabilities: [50] Vital Product Data
Comment 1 Arno Wagner 2005-11-02 18:30:09 UTC
I tested this again with 2.6.14. The problem is still present.

I have confirmed that turning of standby power (+5V_SB) is needed to reset this
behaviour, i.e. pulling the plug or switching the PSU off with the switch on 
that back of the PSU. Alternatively booting 2.6.12.2 also resets this behaviour.

Comment 2 Arno Wagner 2005-11-05 18:38:53 UTC
Here is a work-around for 2.6.14: Replace drivers/net/sk98lin from 2.6.14 with the same 
directory from 2.6.12.2. Seems to work without problem. 
Comment 3 Francois Romieu 2005-11-21 15:04:03 UTC
Can you give the skge module a try and report if it exhibits the same behavior ?

If so it will possible to add its maintainer to the Cc: list :o)

--
Ueimor
Comment 4 Arno Wagner 2005-11-21 18:43:15 UTC
skge behaves similarly. In fact skge is worse: When booting with skge from 2.6.14, neither 
Windows XP nor sk98lin from 2.6.12.2 are able to get the card up and the LEDs remain off, 
i.e. the physical link stays down. At least with sk98lin the version from 2.6.12.2 can revitalize 
the card after it has been booted with sk98lin from 2.6.14.

As before only completely turning power off helps.

BTW, there is a bug in the config infor for skge in 2.6.14: The "help" page does not give the module
name.
Comment 5 Arno Wagner 2005-12-05 17:08:10 UTC
Small update: Problem still present in 2.6.15-rc5. Just to document it.
Comment 6 Adrian Bunk 2005-12-11 12:54:54 UTC
Stephen, can you look at the skge part of this bug?
Comment 7 Stephen Hemminger 2005-12-12 09:12:26 UTC
I need to know the chip version and revision, from the kernel console output
(dmesg).

Something like:
skge 1.2 addr 0xf2000000 irq 17 chip Yukon rev 1
skge eth0: addr 00:0c:41:1a:a5:1f
Comment 8 Arno Wagner 2005-12-12 16:58:57 UTC
skge 1.2 addr 0xfe200000 irq 16 chip Yukon-Lite rev 9
skge eth0: addr 00:11:d8:54:6d:60

Is this sufficient?
Comment 9 Adrian Bunk 2006-03-08 13:37:41 UTC
Stephen, what is the status of this issue?
Comment 10 Stephen Hemminger 2006-03-08 14:51:53 UTC
No new progress, I couldn't reproduce it on the one dual boot system I tried.
It may depend on hardware and windows driver version.
Comment 11 Arno Wagner 2006-03-08 16:29:46 UTC
Well, I cannot help you, since I don't have the mainboard anymore. Sorry. 

Still, It is not a windows driver issue. Don't forget that 2.6.12.2 was not able to bring the card up as 
well. So unless there are Windows drivers in 2.6.12.2.....
Comment 12 Stephen Hemminger 2006-04-03 15:01:58 UTC
This matches symptoms of other bug. So mark as duplicate

*** This bug has been marked as a duplicate of 5204 ***