Skip to contentSkip to conversation

Illegal instruction with gl

DISCUSSION Wakko Warner 2 participants 5 messages Jan 23, 2004 – Jan 24, 2004

lore
Share & tools
lore
#0 lore raw

I recently upgraded my system board and now gl stuff crashes with illegal instruction.

Here's the trace (space in url to keep off robots)
http:// veg.animx.eu.org/strace.gl.txt

I'm using 2.6.1 SMP PREEMPT on a dual P4 Xeon 2.66ghz (1gb ram)

The board is a supermicro x5da8 (Intel e7505 chipset). The card is a matrox g400 max 32mb dual head. mga, intel_agp and agpgart are loaded.

Modifications to the kernel when I changed the board (old board was an intel chipset, same card with 384mb ram):
Enabled ACPI
Enabled PreEmpt
Changed processor type to p4 (from pIII)
Enabled SMP
Enabled High mem (4GB)
Enable aic79xx (aic7xxx is also enabled but not used) and changed the initial delay from 15000 to 1000

I also enabled new modules for the intel 1000 nic and the intel ICH ac97 sound (both alsa and oss).

These are the only configuration changes. I looked at the X log but saw nothing. The kernel does not complain about anything.

If this is a known problem, please let me know. I'm willing to try any troubleshooting to fix this problem. Any more information available on request.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#1 lore raw

Just to add to the below, I disabled acpi, highmem, smp, and preempt. Same problem. I also tried 2.4.24 kernel, same results.

I modified agpgart on 2.4.24 to use generic intel init which didn't change anything.

This is the first system with an intel E7505 I've used so I don't know if it's a board problem or a kernel problem. Or if it's with the matrox g400.

Here's the kernel messages for agp/drm from 2.6.1 (everything enabled that I disabled):
agpgart: Detected an Intel E7505 Chipset.
agpgart: Maximum main memory to use for agp memory: 941M
agpgart: AGP aperture is 32M @ 0xf8000000
[drm] Initialized mga 3.1.0 20021029 on minor 0
agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
agpgart: Device is in legacy mode, falling back to 2.x
agpgart: Putting AGP V2 device at 0000:00:00.0 into 1x mode
agpgart: Putting AGP V2 device at 0000:01:00.0 into 1x mode

I placed some files in:
http:// veg.animx.eu.org/e7505/
with some extra information (config, full dmesg, lspci -v)

Quoted 38 lines

I recently upgraded my system board and now gl stuff crashes with illegal instruction.

Here's the trace (space in url to keep off robots)
http:// veg.animx.eu.org/strace.gl.txt

I'm using 2.6.1 SMP PREEMPT on a dual P4 Xeon 2.66ghz (1gb ram)

The board is a supermicro x5da8 (Intel e7505 chipset). The card is a matrox g400 max 32mb dual head. mga, intel_agp and agpgart are loaded.

Modifications to the kernel when I changed the board (old board was an intel chipset, same card with 384mb ram):
Enabled ACPI
Enabled PreEmpt
Changed processor type to p4 (from pIII)
Enabled SMP
Enabled High mem (4GB)
Enable aic79xx (aic7xxx is also enabled but not used) and changed the initial delay from 15000 to 1000

I also enabled new modules for the intel 1000 nic and the intel ICH ac97 sound (both alsa and oss).

These are the only configuration changes. I looked at the X log but saw nothing. The kernel does not complain about anything.

If this is a known problem, please let me know. I'm willing to try any troubleshooting to fix this problem. Any more information available on request.

Signature · 6 lines
 Lab tests show that use of micro$oft causes cancer in lab animals
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/
Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#2 lore raw
Quoted 20 lines · Wakko Warner

Just to add to the below, I disabled acpi, highmem, smp, and preempt. Same problem. I also tried 2.4.24 kernel, same results.

I modified agpgart on 2.4.24 to use generic intel init which didn't change anything.

This is the first system with an intel E7505 I've used so I don't know if it's a board problem or a kernel problem. Or if it's with the matrox g400.

Here's the kernel messages for agp/drm from 2.6.1 (everything enabled that I disabled):
agpgart: Detected an Intel E7505 Chipset.
agpgart: Maximum main memory to use for agp memory: 941M
agpgart: AGP aperture is 32M @ 0xf8000000
[drm] Initialized mga 3.1.0 20021029 on minor 0
agpgart: Found an AGP 3.0 compliant device at 0000:00:00.0.
agpgart: Device is in legacy mode, falling back to 2.x
agpgart: Putting AGP V2 device at 0000:00:00.0 into 1x mode
agpgart: Putting AGP V2 device at 0000:01:00.0 into 1x mode

Quoted 3 lines

I placed some files in:
http:// veg.animx.eu.org/e7505/
with some extra information (config, full dmesg, lspci -v)

I see no evidence that this is an agpgart problem. When that does something wrong, you usually end up with either a system lockup, or massive memory corruption. Apps segfaulting would suggest to me that you have a problem with your X GL libraries.

You may have more luck asking the folks at [email protected] about it.

Dave

#3 lore raw
Quoted 6 lines

I see no evidence that this is an agpgart problem. When that does something wrong, you usually end up with either a system lockup, or massive memory corruption. Apps segfaulting would suggest to me that you have a problem with your X GL libraries.

You may have more luck asking the folks at [email protected] about it.

Considering that all I did was change the motherboard, it could be an agp problem. This same card, same kernel, everything worked on another system. I did not change any other software other than the kernel.

The program did not segfault, it received illegal instruction. The only driver that I had to change when I installed this board was the aic79xx. The other system used an addon aic7xxx controller, this has a builtin aic7902 controller which houses /

I'll forward my same message to that list you mentioned.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals
#4 lore raw
Quoted 6 lines
Quoted 6 lines

I see no evidence that this is an agpgart problem. When that does something wrong, you usually end up with either a system lockup, or massive memory corruption. Apps segfaulting would suggest to me that you have a problem with your X GL libraries.

You may have more luck asking the folks at [email protected] about it.

I was using 4.2.1.1 from debian. I just upgraded to 4.3.0 that's in debian experimental and is now working. I really don't know what's changed between versions, but I can't understand why just changing the system board would have caused this. All I know is the fact that the old board used the generic intel setup routines and this one uses 7505 (but I also changed it).

I guess the only thing is this board is too new for X to understand or something.

I'd like to upgrade my video card, but I don't know what to get. I've looked at what 3D is supported by the kernel and looks like I'll be going for an ATI card. Any recommendations or a list of ati radeon cards that work in 3D?

Quoted 10 lines

Considering that all I did was change the motherboard, it could be an agp problem. This same card, same kernel, everything worked on another system. I did not change any other software other than the kernel.

The program did not segfault, it received illegal instruction. The only driver that I had to change when I installed this board was the aic79xx. The other system used an addon aic7xxx controller, this has a builtin aic7902 controller which houses /

I'll forward my same message to that list you mentioned.

Signature · 1 line
 Lab tests show that use of micro$oft causes cancer in lab animals