mtx kernel panic - QID LP PCIe

This forum is for all Parhelia based cards (Millenium P650, P690, P750, QID, Parhelia 256, APVe, ...) that need the mtx driver to run properly.

mtx kernel panic - QID LP PCIe

Postby Sitron » Sun Feb 15, 2009 2:27 pm

Hi folks,

thanks for a great site with unofficial drivers and support (which works! Well, mostly :-))

However, I have a bug to report:

I have a workstation with Debian Lenny, i686, 32 bit. My card is a Matrox QID LP PCIe, with the latest BIOS. I have four 20" 1600x1200 monitors. Kernel is compiled with my own config, but no patches are applied. I have tested the following combos:
2.6.24.7 with the official 1.4.6 driver.
2.6.24.7 with your 1.4.6.1 driver.
2.6.26.2 with your 1.4.6.2 driver.
2.6.27.9 with your 1.4.6.2 driver.

The following log and attachments are made with the 2.6.27.9/1.4.6.2-combo.

I load the mtx-module:
Code: Select all
mtx: module license 'Copyright (c) 2002, 2004, Matrox Graphics Inc.' taints kernel.
[mtx] Unofficial MTX driver v1.4.6.2 by tuxx-home.at
[mtx] Allocated a MTX agp driver structure
mtx 0000:01:00.0: PCI INT A -> GSI 16 (level, low) -> IRQ 16
[mtx] 0x2538(sub:0xffffffff) board found at 01:00.0


Then I start the X.org-server:
Code: Select all
[mtx] Registers at 0xd8000000, size: 8K, flags: 131588, knl_addr: 0xf8b78000
[mtx] Framebuffer at 0xd0000000, size: 128M, flags: 135692, knl_addr: 0x00000000, write-combining: YES
[mtx] Busmastering flags:
[mtx]   Board type detected: PCI
[mtx]   Chipset 0x2538:0x102b was detected
[mtx]   AGP chipset was detected
[mtx]   PCI transfers available for read write
[mtx] Parhelia patches applied: PowerM Cap66Mhz CompBypass
[mtx] Registers at 0xd8000000, size: 8K, flags: 131588, knl_addr: 0xf8b78000
[mtx] Framebuffer at 0xd0000000, size: 128M, flags: 135692, knl_addr: 0x00000000, write-combining: YES
[mtx] Busmastering flags:
[mtx]   Board type detected: PCI
[mtx]   Chipset 0x2538:0x102b was detected
[mtx]   AGP chipset was detected
[mtx]   PCI transfers available for read write
[mtx] Parhelia patches applied: PowerM Cap66Mhz CompBypass
allocation failed: out of vmalloc space - use vmalloc=<size> to increase size.
[mtx] *ERROR* Parhelia OS: Fail to map memory region (0xd1d00000, size:29622272
[mtx] *ERROR* Parhelia OS: Unmapping an invalid memory region (0x00000000, size:29622272)
BUG: unable to handle kernel paging request at feec901c
IP: [<f8d74de8>] :mtx:mtx3966+0x50/0x120
*pde = 36efa067 *pte = 00000000
Oops: 0000 [#1] PREEMPT SMP
Modules linked in: mtx(P) ipv6 fuse 8250_pci 8250_pnp 8250 serial_core intel_agp agpgart evdev

Pid: 4400, comm: Xorg Tainted: P          (2.6.27.9 #1)
EIP: 0060:[<f8d74de8>] EFLAGS: 00013206 CPU: 1
EIP is at mtx3966+0x50/0x120 [mtx]
EAX: 00000001 EBX: feec9004 ECX: f78010c0 EDX: c16dde80
ESI: feec9004 EDI: f8b90004 EBP: fd26b004 ESP: f6e21ee0
DS: 007b ES: 007b FS: 00d8 GS: 0000 SS: 0068
Process Xorg (pid: 4400, ti=f6e20000 task=f7bc4080 task.ti=f6e20000)
Stack: c017fb8e f8b94000 f8b7d004 f887500c f8b76000 f8875000 f8b76038 f8e17ee0
       f8b7d004 f8b90004 f8b76000 f8b76014 f8875000 f76f4194 f8e187ae 00000008
       f71706c0 f715ec5c c015771e f781dcc0 f71706c0 00000000 f7b4d8c0 00000000
Call Trace:
[<c017fb8e>] de_put+0x8/0x37
[<f8e17ee0>] mtx_ctx_cleanup+0x40/0xb0 [mtx]
[<f8e187ae>] mtx_release+0x2e/0x90 [mtx]
[<c015771e>] __fput+0x8a/0x129
[<c0155313>] filp_close+0x4d/0x53
[<c011c66e>] put_files_struct+0x60/0xa6
[<c011da38>] do_exit+0x1a1/0x698
[<c0169461>] mntput_no_expire+0x18/0xed
[<c0155313>] filp_close+0x4d/0x53
[<c011df8a>] do_group_exit+0x5b/0x82
[<c0102bdd>] sysenter_do_call+0x12/0x25
=======================
Code: 04 0f 85 d0 00 00 00 83 ec 0c 57 e8 6b 1b 00 00 83 c4 10 8b 4c 24 04 85 c9 74 50 8b 54 24 04 8b 02 85 c0 74 1a 8b 68 1c 89 ee 90 <8b> 5e 18 8b 47 04 39 83 f8 05 00 00 89 de 74 7c 39 eb 75 ec 8b
EIP: [<f8d74de8>] mtx3966+0x50/0x120 [mtx] SS:ESP 0068:f6e21ee0
---[ end trace 6cad354ea5b28ecf ]---
Fixing recursive fault but reboot is needed!


Attached is the Xorg.0.log and xorg.conf.

I must stress that this only happens when I configure the xorg to use four individual screens. I can get 2+2 working, but then the maximize-window-function maximizes the window to two screens, not one as I want.
Attachments
Xorg.0.log
(66.53 KiB) Downloaded 345 times
xorg.conf
(2.19 KiB) Downloaded 350 times
Sitron
Private
Private
 
Posts: 2
Joined: Sun Feb 15, 2009 1:48 pm

Re: mtx kernel panic - QID LP PCIe

Postby Sitron » Mon Feb 16, 2009 11:15 am

Ok, I think I have messed up. I should be using the option QHI insted. But that does not work either, only getting blank screens. I will post errorlogs soon.
Sitron
Private
Private
 
Posts: 2
Joined: Sun Feb 15, 2009 1:48 pm

Re: mtx kernel panic - QID LP PCIe

Postby tuxx-home.at » Wed Mar 11, 2009 12:21 pm

What do you have the vmalloc kernel command line parameter set to? On my system, it's set to 512M.

You can check that with the following command:

Code: Select all
tuxx@vi-edv003:~$ cat /proc/cmdline
root=/dev/sda1 ro vmalloc=512M vga=0x31b
tuxx@vi-edv003:~$


To modify this value, you have to modify your bootloader configuration which is on recent systems /boot/grub/menu.lst.
Just add a string like "vmalloc=512M" to your kernel command line, reboot your sytem, make sure that this value has been set correctly using the command shown above and the error should be gone.
User avatar
tuxx-home.at
Supreme Commander
Supreme Commander
 
Posts: 2199
Joined: Mon Jan 01, 2007 12:51 pm
Location: Vassach - Austria - Europe


Return to Matrox P-Series Cards

Who is online

Users browsing this forum: No registered users and 1 guest

cron