Skip to contentSkip to conversation

2.6.9-rc2-mm4 and nvidia 1.0-6111

DISCUSSION Sid Boyce 6 participants 11 messages Sep 29, 2004 – Sep 30, 2004

lore
Share & tools
lore

Sid Boyce

author
#0 lore raw

The usual patches applied that made it work with -mm3.
Changed all instances of NV_REMAP_PAGE_RANGE to NV_REMAP_PFN_RANGE and nv_remap_page_range to nv_remap_pfn_range in nv.c, nv-linux.h, os-agp.c and os-interface.c as redifined in /usr/src/linux-2.6.9-rc2-mm4/include/linux/mm.h.
The module builds and installs without problems, but on init 5, I get thrown back to vt1, /var/log/XFree86.0.log gives the error:- (--) NVIDIA(0): Linear framebuffer at 0xD8000000
(--) NVIDIA(0): MMIO registers at 0xE1000000
(II) NVIDIA(0): NVIDIA GPU detected as: GeForce FX 5200
(--) NVIDIA(0): VideoBIOS: 04.34.20.42.01
(--) NVIDIA(0): Interlaced video modes are supported on this GPU
(II) NVIDIA(0): Detected AGP rate: 8X
(EE) NVIDIA(0): Failed to allocate config DMA context
(II) UnloadModule: "nvidia"
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found
-------------------------------
Any help appreciated, also posted to nvidia forum.

Regards
Sid.

Signature · 2 lines
Sid Boyce .... Hamradio G3VBV and keen Flyer
=====LINUX ONLY USED HERE=====
#1 lore raw

This is a known issue and it's releated to some changes in Andrew's tree. Not to sound pesimistic but in the future expect the nvidia drivers to be broken for the -mm tree, it is after an the expiremental kernel tree.

matt

Quoted 26 lines · Sid Boyce

The usual patches applied that made it work with -mm3. Changed all instances of NV_REMAP_PAGE_RANGE to NV_REMAP_PFN_RANGE and nv_remap_page_range to nv_remap_pfn_range in nv.c, nv-linux.h, os-agp.c and os-interface.c as redifined in /usr/src/linux-2.6.9-rc2-mm4/include/linux/mm.h.
The module builds and installs without problems, but on init 5, I get thrown back to vt1, /var/log/XFree86.0.log gives the error:- (--) NVIDIA(0): Linear framebuffer at 0xD8000000
(--) NVIDIA(0): MMIO registers at 0xE1000000
(II) NVIDIA(0): NVIDIA GPU detected as: GeForce FX 5200 (--) NVIDIA(0): VideoBIOS: 04.34.20.42.01
(--) NVIDIA(0): Interlaced video modes are supported on this GPU (II) NVIDIA(0): Detected AGP rate: 8X
(EE) NVIDIA(0): Failed to allocate config DMA context
(II) UnloadModule: "nvidia"
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found
-------------------------------
Any help appreciated, also posted to nvidia forum.

Regards
Sid.

#3 lore raw
In reply to Sid Boyce
Quoted 1 line · Sid Boyce

Any help appreciated, also posted to nvidia forum.

I did revert these patches:

convert-references-to-remap_page_range-under-arch-and-documentation-to-remap_pfn_range.patch convert-users-of-remap_page_range-under-drivers-and-net-to-use-remap_pfn_range.patch convert-users-of-remap_page_range-under-include-asm--to-use-remap_pfn_range.patch convert-users-of-remap_page_range-under-sound-to-use-remap_pfn_range.patch for-mm-only-remove-remap_page_range-completely.patch introduce-remap_pfn_range-to-replace-remap_page_range.patch

HTH,
Norberto

#4 lore raw
Quoted 11 lines · Norberto Bensa
Quoted 1 line · Sid Boyce

Any help appreciated, also posted to nvidia forum.

I did revert these patches:

convert-references-to-remap_page_range-under-arch-and-documentation-to-remap_pfn_range.patch convert-users-of-remap_page_range-under-drivers-and-net-to-use-remap_pfn_range.patch convert-users-of-remap_page_range-under-include-asm--to-use-remap_pfn_range.patch convert-users-of-remap_page_range-under-sound-to-use-remap_pfn_range.patch for-mm-only-remove-remap_page_range-completely.patch introduce-remap_pfn_range-to-replace-remap_page_range.patch

Isn't there an nvidia-linux mailing list? This is really OT for LKML.

Lee

#5 lore raw
Quoted 2 lines · Lee Revell

Isn't there an nvidia-linux mailing list? This is really OT for LKML.

I had one for a while where I posted patches but it never gained much momentum. Unless there is a sizable group of people who want this I don't see any need to resurrect it.

#6 lore raw
In reply to Chris Wedgwood
Quoted 8 lines · Chris Wedgwood
Quoted 2 lines · Lee Revell

Isn't there an nvidia-linux mailing list? This is really OT for LKML.

I had one for a while where I posted patches but it never gained much momentum. Unless there is a sizable group of people who want this I don't see any need to resurrect it.

OK, makes sense. With so many people using the driver I guess it's just easiest to deal with nvidia problems on LKML.

Lee

#7 lore raw
Quoted 5 lines · you
Quoted 4 lines · Chris Wedgwood
Quoted 2 lines · Lee Revell

Isn't there an nvidia-linux mailing list? This is really OT for LKML.

Quoted 9 lines
Quoted 3 lines

I had one for a while where I posted patches but it never gained much momentum. Unless there is a sizable group of people who want this I don't see any need to resurrect it.

OK, makes sense. With so many people using the driver I guess it's just easiest to deal with nvidia problems on LKML.

Lee

Just about any out-of-kernel driver using Changed-API-X will be broken, free or non-free. Something more general like linux-drivers or linux-kernel-drivers would probably make more sense.

Sometimes changes in -mm even break in-kernel drivers; it's not really an "NVIDIA problem" as such. I agree with Lee though; it's an unwritten rule that you prefix a subject with [OT] when speaking about something which isn't directly relevant to the kernel.

(By the way, if this breaks outside of -mm patches will appear for stable kernels on http://minion.de/ as did with the 2.5 development tree.)

Signature · 8 lines
Cheers,
Alistair.

personal:   alistair()devzero!co!uk
university: s0348365()sms!ed!ac!uk
student:    CS/AI Undergraduate
contact:    1F2 55 South Clerk Street,
            Edinburgh. EH8 9PP.
#9 lore raw
In reply to Lee Revell
Quoted 2 lines · Lee Revell

OK, makes sense. With so many people using the driver I guess it's just easiest to deal with nvidia problems on LKML.

If there are more than a handful of people I'll restart linux-nvidia (it was [email protected], subscription via majordomo or similar) since the noise on LKML will just piss people off.

author
#2 lore raw
In reply to Matt Heler
Quoted 36 lines · Matt Heler

This is a known issue and it's releated to some changes in Andrew's tree. Not to sound pesimistic but in the future expect the nvidia drivers to be broken for the -mm tree, it is after an the expiremental kernel tree.

matt

Quoted 26 lines · Sid Boyce

The usual patches applied that made it work with -mm3. Changed all instances of NV_REMAP_PAGE_RANGE to NV_REMAP_PFN_RANGE and nv_remap_page_range to nv_remap_pfn_range in nv.c, nv-linux.h, os-agp.c and os-interface.c as redifined in /usr/src/linux-2.6.9-rc2-mm4/include/linux/mm.h.
The module builds and installs without problems, but on init 5, I get thrown back to vt1, /var/log/XFree86.0.log gives the error:- (--) NVIDIA(0): Linear framebuffer at 0xD8000000
(--) NVIDIA(0): MMIO registers at 0xE1000000
(II) NVIDIA(0): NVIDIA GPU detected as: GeForce FX 5200 (--) NVIDIA(0): VideoBIOS: 04.34.20.42.01
(--) NVIDIA(0): Interlaced video modes are supported on this GPU (II) NVIDIA(0): Detected AGP rate: 8X
(EE) NVIDIA(0): Failed to allocate config DMA context
(II) UnloadModule: "nvidia"
(II) UnloadModule: "vgahw"
(II) Unloading /usr/X11R6/lib/modules/libvgahw.a
(EE) Screen(s) found, but none have a usable configuration.

Fatal server error:
no screens found
-------------------------------
Any help appreciated, also posted to nvidia forum.

Regards
Sid.

Thanks, it did seem an easy thing to fix, when it built and inmod, I had high hopes. Regards
Sid.

Signature · 2 lines
Sid Boyce .... Hamradio G3VBV and keen Flyer
=====LINUX ONLY USED HERE=====
author
#8 lore raw
In reply to Alistair John Strachan
Quoted 20 lines · Alistair John Strachan
Quoted 7 lines · you
Quoted 5 lines · Chris Wedgwood
Quoted 2 lines · Lee Revell

Isn't there an nvidia-linux mailing list? This is really OT for LKML.

Quoted 9 lines
Quoted 3 lines

I had one for a while where I posted patches but it never gained much momentum. Unless there is a sizable group of people who want this I don't see any need to resurrect it.

OK, makes sense. With so many people using the driver I guess it's just easiest to deal with nvidia problems on LKML.

Lee

Quoted 6 lines

Just about any out-of-kernel driver using Changed-API-X will be broken, free or non-free. Something more general like linux-drivers or linux-kernel-drivers would probably make more sense.

Sometimes changes in -mm even break in-kernel drivers; it's not really an "NVIDIA problem" as such. I agree with Lee though; it's an unwritten rule that you prefix a subject with [OT] when speaking about something which isn't directly relevant to the kernel.

(By the way, if this breaks outside of -mm patches will appear for stable kernels on http://minion.de/ as did with the 2.5 development tree.)

So, in addition to the patches from http://00f.net/blogs/index.php/2004/09/07/nvidia_kernel_module_and_linux_2_6_9_mm
I've changed NV_REMAP_PAGE_RANGE to NV_REMAP_PFN_RANGE and
remap_page_range to remap_pfn_range in nv-linux.h, nv.c, os-agp.c, and os-interface.c, the missing piece -- see below. I may get around to posting patches to the nvidia forum later today.

I'm not sure if I can divulge the name of my helper in case he gets showered by unwanted email, but a suggested additional change to nv-linux.h has fixed it.

#define NV_REMAP_PFN_RANGE(from, offset, x...) \
     remap_pfn_range(vma, from, ((offset)) >> PAGE_SHIFT), x)

Regards
Sid.

Signature · 2 lines
Sid Boyce .... Hamradio G3VBV and keen Flyer
=====LINUX ONLY USED HERE=====