I am on a laptop too, so there is no switching. ATI manages to fix most of the bugs when they release their new drivers, but then a kernel update throws them off again. They must have like 2 people or some similarly small group of people working on the drivers, judging by the updates we get. Given the small incentive for spending money on Linux, it's understandable.


I can run the 2d drivers fine if I just delete the fglrx module, just means I have to live without hardware acceleration for a while I guess.

From ATIs release notes (http://www2.ati.com/drivers/linux/li...4.13-inst.html):

Try that, and with some luck, you will succeed.

yes thanks
Not this
I've checked, before startx (with kernel 2.6.12.2)

[18:39][vlad@vld36:~]$ /sbin/lspci |grep AGP
00:01.0 PCI bridge: Intel Corp. 82865G/PE/P PCI to AGP Controller (rev 02)
[18:39][vlad@vld36:~]$

[18:39][vlad@vld36:~]$ lsmod|grep agp
intel_agp 20252 1
agpgart 28840 2 fglrx,intel_agp
[18:40][vlad@vld36:~]$

something like this

after startx works fine
but dri disabled

and I found some interesting things:
first
when I am working on 2.6.11.12 kernel (dri works) -

[18:39][vlad@vld36:~]$ lsmod|grep agp
intel_mch_agp 8464 0
intel_agp 20252 1
agpgart 28840 3 fglrx,intel_mch_agp,intel_agp
[18:40][vlad@vld36:~]$

there is 3rd module intel_mch_agp , which is not used as I could understand that
and I repeat - DRI works

[18:40][vlad@vld36:~]$ uname -r
2.6.11.12
[18:48][vlad@vld36:~]$ glxinfo |grep render
direct rendering: Yes
GLX_ATI_render_texture
OpenGL renderer string: RADEON 9200SE DDR Generic
[18:48][vlad@vld36:~]$

another
when I am working on 2.6.12.2 kernel
after startx computer works without dri
when I am trying to stratx again computer freezes sreen goes black and that all

what might it be ?????

(sorry for bad english)

What happens if you try to modprobe intel_mch_agp in 2.6.12.2? I'm completely new to this, but maybe you need that one as well in order to achieve full agp-support?

Yaeh I thought too
but there is not such module in

ls /lib/modules/2.6.12.2/kernel/drivers/char/agp/
agpgart.ko ati-agp.ko efficeon-agp.ko intel-agp.ko

and

ls /lib/modules/2.6.11.12/kernel/drivers/char/agp/
agpgart.ko ati-agp.ko efficeon-agp.ko intel-agp.ko intel-mch-agp.ko

and also I checked kernel .config file
I mean I used old (same) .config file for both kernels

and checked visualy (there are same options in both kernels)

...

What I don't understand is that all I used to need was agpgart. Have they increased the module requirements to use the ATI driver?

Furthermore, when I: lspci | grep AGP I get:

Which Intel motherboard is that? I don't see any that match (as far as I know) in the kernel configuration.

Any suggestions?

just one minute

I found that in 2.6.11.12 kernel I have :
Intel 440LX/BX/GX, I8xx and E7x05 chipset support AGP_INTEL

This option gives you AGP support for the GLX component of XFree86 4.x
on Intel 440LX/BX/GX, 815, 820, 830, 840, 845, 850, 860, 875,
E7205 and E7505 chipsets and full support for the 810, 815, 830M, 845G,
852GM, 855GM, 865G and I915 integrated graphics chipsets.

You should say Y here if you use XFree86 3.3.6 or 4.x and want to
use GLX or DRI, or if you have any Intel integrated graphics
chipsets. If unsure, say Y.

and

Intel i865 chipset support AGP_INTEL_MCH

This option gives you AGP support for the GLX component of XFree86 4.x
on Intel chipsets that support Intel EM64T processors.

You should say Y here if you use XFree86 3.3.6 or 4.x and want to
use GLX or DRI. If unsure, say Y.

and in 2.6.12.2

I only have:

ntel 440LX/BX/GX, I8xx and E7x05 chipset support AGP_INTEL

This option gives you AGP support for the GLX component of XFree86 4.x
on Intel 440LX/BX/GX, 815, 820, 830, 840, 845, 850, 860, 875,
E7205 and E7505 chipsets and full support for the 810, 815, 830M, 845G,
852GM, 855GM, 865G and I915 integrated graphics chipsets.

You should say Y here if you use XFree86 3.3.6 or 4.x and want to
use GLX or DRI, or if you have any Intel integrated graphics
chipsets. If unsure, say Y.

EM64T is if you have an Intel Pentium 4 processor with the 64bit extensions. Do you have such a processor?

Is my Pentium M a 855GM?

No I don't know
No , but module loaded interesting ...

Thanks, Erratio. However, I already fixed the compiler problems, it just seems now that the fglrx module wants to freeze X session attempts. It compiles but it makes the screen go black.

I was a little hasty in posting the last message, or at least should have added a disclaimer. I didn't have time to fully try it out myself, and it didn't end up working when I did...initially it installed without acceleration, and now I have a freezing with a black screen problem as well. Luckily? the problem remained when recompiling for 2.6.11, so it should allow for more controlled troubleshooting...

the dev->slot_name to pci_name(dev) conversion seems to work fine, but the 'inter_module_get_request' seems to be what's causing the problem, since when that was removed I regained acceleration in 2.6.11. I'll play with it a bit more later if I have time, but if anyone knows of a better substitute for 'inter_module_get', that's seems to be what's missing. Rather than replacing them I also just put a '#define inter_module_get(x) inter_module_get_request(x,"")' line in the file, which could be thrown into a preprocessor conditional for a more robust patch.

and just a slight addendum...glancing at the code all the occurences of the dev->slot_name call seem to be in printk() statements, so they shouldn't be causing much of any problems anyway.

Tried and true (at least for me)-

http://web38.austria182.server4free....opic.php?t=188

Hmm well those patches fix compiler issues, but my computer still freezes when it tries to actually use the fglrx module. I wonder what the deal is!