Skip to contentSkip to conversation

2.6 vs 2.4: pxe booting system won't restart

DISCUSSION Brian Jackson 3 participants 5 messages Nov 10, 2004 – Nov 11, 2004

lore
Share & tools
lore
#0 lore raw

I'm having a problem with 2.6 (many versions), and my Via Epia M10000 not rebooting correctly. 2.4 works fine. The problem is after the computer restarts, and the pxe stuff from the bios tries to do it's thing, it fails. I get the following error:
PXE_M0F: Exiting Intel PXE ROM.

Then the bios tries to fallback to other means of booting, and there are none. Anybody got any clues where to start looking for fixes?

#1 lore raw
Quoted 8 lines · Brian Jackson

I'm having a problem with 2.6 (many versions), and my Via Epia M10000 not rebooting correctly. 2.4 works fine. The problem is after the computer restarts, and the pxe stuff from the bios tries to do it's thing, it fails. I get the following error:
PXE_M0F: Exiting Intel PXE ROM.

Then the bios tries to fallback to other means of booting, and there are none. Anybody got any clues where to start looking for fixes?

Remove the kernel code that powers down the ethernet chip. If that works then poke VIA.

#2 lore raw
Quoted 9 lines · Alan Cox
Quoted 6 lines · Brian Jackson

I'm having a problem with 2.6 (many versions), and my Via Epia M10000 not rebooting correctly. 2.4 works fine. The problem is after the computer restarts, and the pxe stuff from the bios tries to do it's thing, it fails. I get the following error:
PXE_M0F: Exiting Intel PXE ROM.

Quoted 4 lines
Quoted 2 lines

Then the bios tries to fallback to other means of booting, and there are none. Anybody got any clues where to start looking for fixes?

Remove the kernel code that powers down the ethernet chip. If that works

Yay, looks like this bit near line 1950 of via-rhine.c:

	/* Hit power state D3 (sleep) */
	writeb(readb(ioaddr + StickyHW) | 0x03, ioaddr + StickyHW);

I removed that, and it works like a charm now. Thank you very much.

Quoted 1 line

then poke VIA.

Poke them to fix the driver or to fix the bios?

Quoted 2 lines
#3 lore raw
Quoted 7 lines · Brian Jackson
Quoted 1 line

Remove the kernel code that powers down the ethernet chip. If that works

Yay, looks like this bit near line 1950 of via-rhine.c:

	/* Hit power state D3 (sleep) */
	writeb(readb(ioaddr + StickyHW) | 0x03, ioaddr + StickyHW);

I removed that, and it works like a charm now. Thank you very much.

Excellent - its as I thought - the BIOS isn't bringing the chip out of powersave for PXE but assumes it is live. Possibly one for a module option but really it is a BIOS bug, although a very minor and easy to understand how it got missed one

Alan

#4 lore raw
In reply to Brian Jackson

[BCC'ed the people at VNTEK who are on the case]

Quoted 10 lines · Brian Jackson
Quoted 4 lines
Quoted 2 lines

Then the bios tries to fallback to other means of booting, and there are none. Anybody got any clues where to start looking for fixes?

Remove the kernel code that powers down the ethernet chip. If that works

Yay, looks like this bit near line 1950 of via-rhine.c:

	/* Hit power state D3 (sleep) */
	writeb(readb(ioaddr + StickyHW) | 0x03, ioaddr + StickyHW);

I removed that, and it works like a charm now. Thank you very much.

I'm sorry, I should posted a warning earlier. This problem has been reported before.

Quoted 3 lines
Quoted 1 line

then poke VIA.

Poke them to fix the driver or to fix the bios?

BIOS. I can't fix this in the driver. Some hardware requires D3 for WOL.

Roger