Re: D945GCLF intel atom board & r8169 ethernet broken in linux-2.6.27-rc5??

Previous thread: dmaengine: DMA_CTRL_ACK flag signification by Nicolas Ferre on Wednesday, September 24, 2008 - 4:59 am. (3 messages)

Next thread: Re: parport_pc disables parallel port on unload. by Bodo Eggert on Wednesday, September 24, 2008 - 5:47 am. (2 messages)
From: John Gumb
Date: Wednesday, September 24, 2008 - 4:26 am

Hi

I have an Intel Atom D945GCLF based system. It has a realtek RTL8101E
PCI network interface.

With linux-2.6.27-rc4 - everything seems fine. I think we've been
working OK since 2.6.26ish.

With linux-2.6.27-rc5, "ifconfig eth0" just hangs. lspci -vvx reports,
00: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

10: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

20: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff

30: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff =20

Before I load the driver lspci -vvx looks OK.

I do not have (much) time to look into this any further but hopefully
this might help someone.

I know there are problems with this device on the atom board
particularly around warm start but thus far I seem to have avoided this
issue.

A cold start in this case does not help.

Lspci.bad comes from -rc5 with the r8169 driver loaded.
Lspci.good comes from -rc4 with the r8169 driver loaded

Cheers

John
From: Kok, Auke
Date: Wednesday, September 24, 2008 - 11:37 am

did you check if falling back to the older kernel makes the device work again?
--

From: Francois Romieu
Date: Wednesday, September 24, 2008 - 1:19 pm

John Gumb <john.gumb@tandberg.com> :


Can you try patches #0001 .. #0008 at
http://userweb.kernel.org/~romieu/r8169/2.6.27-rc6/20080913/ ?

-- 
Ueimor
--

From: Rafael J. Wysocki
Date: Friday, September 26, 2008 - 2:45 pm

FWIW, these patches make the Ethernet card in the MSI Wind I've been recently
testing work much much more reliably. :-)

Thanks,
Rafael
--

From: Francois Romieu
Date: Friday, September 26, 2008 - 2:52 pm

The whole series or a part of it ?

-- 
Ueimor
--

From: Rafael J. Wysocki
Date: Friday, September 26, 2008 - 3:27 pm

I've only tried all of them at once.
--

From: Jiri Kosina
Date: Wednesday, September 24, 2008 - 3:26 pm

Hmm, probably a lot of 0xff stored in NVRAM of the card? There seems to be 
some similarity to the problem that seemed to be e1000e-specific so far.

        http://lkml.org/lkml/2008/8/8/123
        http://lkml.org/lkml/2008/9/22/23
        http://bugzilla.kernel.org/show_bug.cgi?id=11382
	https://bugzilla.novell.com/show_bug.cgi?id=425480

-- 
Jiri Kosina
--

From: Jiri Kosina
Date: Wednesday, September 24, 2008 - 3:27 pm

BTW does by any chance booting with 'nopat' kernel commandline parameter 
fix the problem?

-- 
Jiri Kosina
SUSE Labs

--

Previous thread: dmaengine: DMA_CTRL_ACK flag signification by Nicolas Ferre on Wednesday, September 24, 2008 - 4:59 am. (3 messages)

Next thread: Re: parport_pc disables parallel port on unload. by Bodo Eggert on Wednesday, September 24, 2008 - 5:47 am. (2 messages)