Skip to contentSkip to conversation

2.4.18 AND Geode GX1/200Mhz problem

DISCUSSION Pierre Ficheux 6 participants 8 messages Apr 7, 2002 – Apr 13, 2002

lore
Share & tools
lore
#0 lore raw

Hi all,

I have a strange problem with a Geode/GX1 200Mhz based system. My
kernel is compiled with 586 as processor type but the system stops just after the 'Uncompressing Linux...Ok, booting the kernel' message. It's strange as GX1 is claimed to work fine with 2.4.18. The same system works fine with 2.2.18 kernel.

Is there any new patch to test? Any idea?

Thx a lot by advance.

Signature · 4 lines
Pierre FICHEUX -/- CTO OpenWide, France -\- [email protected]
				         http://www.ficheux.com
					 http://www.emulateur-xtel.org
More fun, more freedom, less Micro$oft
#1 lore raw
Quoted 5 lines

I have a strange problem with a Geode/GX1 200Mhz based system. My
kernel is compiled with 586 as processor type but the system stops just after the 'Uncompressing Linux...Ok, booting the kernel' message. It's strange as GX1 is claimed to work fine with 2.4.18. The same system works fine with 2.2.18 kernel.

With 586 and no TSC set it should work fine yes. You might want to plug a serial port in and compile with serial console enabled, see if it gives any clues

#2 lore raw
In reply to Pierre Ficheux
Quoted 10 lines · Alan Cox
Quoted 6 lines

I have a strange problem with a Geode/GX1 200Mhz based system. My
kernel is compiled with 586 as processor type but the system stops just after the 'Uncompressing Linux...Ok, booting the kernel' message. It's strange as GX1 is claimed to work fine with 2.4.18. The same system works fine with 2.2.18 kernel.

With 586 and no TSC set it should work fine yes. You might want to plug a serial port in and compile with serial console enabled, see if it gives any clues

I want to translate this to:

Processor -> 586/K5/5x86/6x86/6x86MX

right?

#3 lore raw
Quoted 5 lines · Peter Wächtler

I want to translate this to:

Processor -> 586/K5/5x86/6x86/6x86MX

right?

Or you can boot with kernel parameter notsc which should have the effect.

Zwane

Signature · 1 line
http://function.linuxpower.ca
#4 lore raw
Quoted 1 line · Zwane Mwaikambo

Or you can boot with kernel parameter notsc which should have the effect.

One might think that, but it isn't true. That'll just stop the kernel from halting immediately when it doesn't find any tsc capability. The system still locks up -- maybe some other parts of the kernel tries to use tsc? I never looked into it as compiling the kernel correctly (sans tsc) works just fine.

On a side note, has anyone seen dd scrolling behavior with video chipse using shared memory? On the geode board I'm playing with, the screen never scrolls correctly. It's probablly something stupid in the 5530's setup that's at fault. (I have ever intention of turning the VGA port off entirely.)

--Ricky

#5 lore raw
Quoted 5 lines

On a side note, has anyone seen dd scrolling behavior with video chipse using shared memory? On the geode board I'm playing with, the screen never scrolls correctly. It's probablly something stupid in the 5530's setup that's at fault. (I have ever intention of turning the VGA port off entirely.)

My 5530 works fine in VESA mode

#6 lore raw
In reply to Ricky Beam
Quoted 5 lines · Ricky Beam

On a side note, has anyone seen dd scrolling behavior with video chipse using shared memory? On the geode board I'm playing with, the screen never scrolls correctly. It's probablly something stupid in the 5530's setup that's at fault. (I have ever intention of turning the VGA port off entirely.)

The VSA CGA/EGA/VGA text emulation seems to be rather awful and has lots of bugs. Compile in the VESA framebuffer support in your Linux kernel and add vga=785 (see linux/Documentation/fb/vesafb.txt) to your kernel command line. It's a choice of:

    Linux kernel frobs the text console -> buggy VSA emulation -> framebuffer

or

Linux kernel frobs the framebuffer directly

Guess which works better? *grin*

/Christer

Signature · 1 line
"Just how much can I get away with and still go to heaven?"
#7 lore raw
In reply to Pierre Ficheux
Quoted 5 lines · Peter Wächtler

I want to translate this to:

Processor -> 586/K5/5x86/6x86/6x86MX

right?

That's right, have tried with 486 too but I got the same problem.

Signature · 4 lines
Pierre FICHEUX -/- CTO OpenWide, France -\- [email protected]
				         http://www.ficheux.com
					 http://www.emulateur-xtel.org
More fun, more freedom, less Micro$oft