[KMS] Radeon RS4xx cards miss recognised as AGP capable

Bug #557611 reported by Tormod Volden
20
This bug affects 3 people
Affects Status Importance Assigned to Milestone
X.Org X server
Fix Released
Medium
linux (Ubuntu)
Fix Released
Medium
Unassigned

Bug Description

This was reported directly (and fixed) upstream in https://bugs.freedesktop.org/show_bug.cgi?id=27225 and the fix was merged by Linus at http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=patch;h=f95df9ca6896978108201a77422a1ae2cdc595ec

I believe this is the KMS equivalent to bug 475466 which was fixed in the radeon DDX.

Tags: lucid
Revision history for this message
In , Maarten Fonville (maarten-fonville) wrote :

Created an attachment (id=34296)
Normal failure

Revision history for this message
In , Maarten Fonville (maarten-fonville) wrote :

Created an attachment (id=34297)
Failure with forcing agpmode=-1

Revision history for this message
In , agd5f (agd5f) wrote :

Created an attachment (id=34298)
fix

RS4xx+ IGP chips use an internal gart, however,
some of them have the agp cap bits set in their pci
configs. Make sure to clear the AGP flag as AGP will
not work with them.

Revision history for this message
In , Maarten Fonville (maarten-fonville) wrote :

I will try this fix immediately

and *wow* you are fast :) and (hopefully) a hero ;)

Revision history for this message
In , Maarten Fonville (maarten-fonville) wrote :

I can confirm the patch of Alex does work. So if this patch can gets commited soon into the mainline code it is 'fixed'

Revision history for this message
In , agd5f (agd5f) wrote :

corrected patch sent to Dave.

Revision history for this message
Tormod Volden (tormodvolden) wrote : please cherry-pick fix for Radeon RS4xx cards

This was reported directly (and fixed) upstream in https://bugs.freedesktop.org/show_bug.cgi?id=27225 and the fix was merged by Linus at http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=patch;h=f95df9ca6896978108201a77422a1ae2cdc595ec

I believe this is the KMS equivalent to bug 475466 which was fixed in the radeon DDX.

Revision history for this message
Tormod Volden (tormodvolden) wrote :

Maarten, I took the liberty of subscribing you to this bug report, maybe you can confirm the issue and the fix, and update the description if needed. Thanks in advance.

Revision history for this message
Tormod Volden (tormodvolden) wrote :

For the record, I have verified that the above commit applies cleanly against the lucid kernel tree, but I do not have the hardware to test it.

summary: - please cherry-pick fix for Radeon RS4xx cards
+ [KMS] please cherry-pick fix for Radeon RS4xx cards
Revision history for this message
Maarten Fonville (maarten-fonville) wrote : Re: [KMS] please cherry-pick fix for Radeon RS4xx cards

Tormod, good thing to subscribe me to the bug :)

as for the fix, the fix does work correctly. I tested this on the laptop affected by it by applying the patch manually.
With the note that the first fix that was posted by Alex did had a typo in it (a small but crucial one)
the correct patch was posted on dri-devel and applied to Dave's branch on git.kernel.org.
I have seen Linus also has picked up the patch in the current git, though not in the latest released -rc.

Revision history for this message
Maarten Fonville (maarten-fonville) wrote :

I would like to add btw, that the laptop did still suffer from this bug (which became visible as soon as this bug was resolved): https://bugs.edge.launchpad.net/ubuntu/lucid/+source/linux/+bug/509273?comments=all so for this videocard to work correctly also that patch will have to be applied.

Revision history for this message
Apoorv Parle (apparle) wrote :

So is this bug fixed...?
How do I check it?

Revision history for this message
Maarten Fonville (maarten-fonville) wrote :

There is a fix for this bug. It is in the so called 'upstream'. And it will depend on the people responsible/involved in the Lucid release (Tormod in this case I guess) to actually have this fix in Ubuntu itself.

Revision history for this message
Andy Whitcroft (apw) wrote :

Seems the commit in upstream which is referred to here is this:

commit f95df9ca6896978108201a77422a1ae2cdc595ec
Author: Alex Deucher <email address hidden>
Date: Sun Mar 21 14:02:25 2010 -0400

    drm/radeon/kms: never treat rs4xx as AGP

    RS4xx+ IGP chips use an internal gart, however,
    some of them have the agp cap bits set in their pci
    configs. Make sure to clear the AGP flag as AGP will
    not work with them.

    Should fix fdo bug 27225

    Signed-off-by: Alex Deucher <email address hidden>
    cc: <email address hidden>
    Signed-off-by: Dave Airlie <email address hidden>

Changed in linux (Ubuntu):
status: New → Confirmed
importance: Undecided → Medium
Andy Whitcroft (apw)
summary: - [KMS] please cherry-pick fix for Radeon RS4xx cards
+ [KMS] Radeon RS4xx cards miss recognised as AGP capable
Revision history for this message
Apoorv Parle (apparle) wrote :

Tormod
Waiting for you to fix it in the lucid beta

Revision history for this message
Jeremy Foshee (jeremyfoshee) wrote :

Test Kernels are currently b eing built by Andy Whitcroft. He will make a note in this bug as to where they can be found. These will need to be tested before the fix can be applied to Lucid.

Thanks!

~JFo

Revision history for this message
Andy Whitcroft (apw) wrote :

Ok I have pulled back this fix to Lucid and built some test kernels. If those affected by this issue could test the kernels at the URL below and report back here that would be helpful. Kernels will be here shortly:

    http://people.canonical.com/~apw/lp557611-lucid/

Thanks!

Changed in linux (Ubuntu):
assignee: nobody → Andy Whitcroft (apw)
Revision history for this message
Maarten Fonville (maarten-fonville) wrote :

Sadly I can not test this kernel at the laptop mentioned earlier. Since this is my GF's and she won't be near it for a month (it is far away in another country :P)

Revision history for this message
Apoorv Parle (apparle) wrote :

I tested it and yes, I can see the Desktop effects again
Great job guys

Revision history for this message
Jeremy Foshee (jeremyfoshee) wrote :

Apoorv,
     Thank you for the confirmation of the fix.

Tormod,
     Would you mind testing this and telling us your findings?

Thanks!

~JFo

Revision history for this message
Tormod Volden (tormodvolden) wrote :

Jeremy, I do not have this hardware.

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Will the final version of lucid be fixed or have I to patch it manually? Thank you

Revision history for this message
Tormod Volden (tormodvolden) wrote :

Pondo, please try the test kernel which Andy prepared and tell us if it works.

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Hello; I'd like to help, but my pc is going to be repaired. So I can only test live Cds at the moment (and not with this graphic card). Perhaps i can test it in some days. Thank you!

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Now i had the possibility to test it, but my wireless isn`t working so far and I don't have the access to the router. Sorry that I can't test the kernel.

Revision history for this message
Apoorv Parle (apparle) wrote : Re: [Bug 557611] Re: [KMS] Radeon RS4xx cards miss recognised as AGP capable

Pondo: You can try booting from a USB drive. Enable the persistence and then
install the new kernel and then restart.

When is this patch going to be included in kernel....?

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

My CD-Drive works, but I cant download the kernel with this PC, because I can't connect to the internet. I have to wait till my WLAN will work (Worked ootb in earlier ubuntu versions)

Revision history for this message
Apoorv Parle (apparle) wrote :

I suggested booting from USB so that you could download the kernel from
other computer. Install the USB with persistance(any changes are saved).
Then you could try that kernel and tell us if it works

Revision history for this message
Apoorv Parle (apparle) wrote :

So guys, is this going to be fixed in 10.04?

Revision history for this message
Tormod Volden (tormodvolden) wrote :

It is too late to get this into next week's release, but it can be considered for an SRU (stable release update). Anyway it would normally need confirmation from several users that it really solves the issue. Maarten has tested the patch, but only Apoorv has tested Andy's test kernel.

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Sorry, that I couldn't test it earlier.

With this Kernel (I installed linux-image-2.6.32-20-generic_2.6.32-20.29~lp557611v201004081824_i386.deb, linux-headers-2.6.32-20_2.6.32-20.29~lp557611v201004081824_all.deb and linux-image-2.6.32-20-generic_2.6.32-20.29~lp557611v201004081824_i386.deb). then I started update-grub. After that I rebooted my PC. While booting, PC hangs up with this Error:

"[ 19.004475] [drm:rs400_gart_adjust_size] *ERROR* Forcing to 32M GART size (because of ASIC bug ?)"

Then I only could press the power button and wait for shut down.

Revision history for this message
Apoorv Parle (apparle) wrote :

I also get that line on that kernel but everything goes on well after that.
I never paid attention before, but if, the video memory is taken as only
32MB instead of 256MB then it is not good.

On Fri, Apr 23, 2010 at 12:54 AM, Pondo Sinatra <email address hidden>wrote:

> Sorry, that I couldn't test it earlier.
>
> With this Kernel (I installed linux-
> image-2.6.32-20-generic_2.6.32-20.29~lp557611v201004081824_i386.deb,
> linux-headers-2.6.32-20_2.6.32-20.29~lp557611v201004081824_all.deb and
> linux-
> image-2.6.32-20-generic_2.6.32-20.29~lp557611v201004081824_i386.deb).
> then I started update-grub. After that I rebooted my PC. While booting,
> PC hangs up with this Error:
>
> "[ 19.004475] [drm:rs400_gart_adjust_size] *ERROR* Forcing to 32M
> GART size (because of ASIC bug ?)"
>
> Then I only could press the power button and wait for shut down.
>
> --
> [KMS] Radeon RS4xx cards miss recognised as AGP capable
> https://bugs.launchpad.net/bugs/557611
> You received this bug notification because you are a direct subscriber
> of the bug.
>
> Status in X.Org X server: Unknown
> Status in “linux” package in Ubuntu: Confirmed
>
> Bug description:
> This was reported directly (and fixed) upstream in
> https://bugs.freedesktop.org/show_bug.cgi?id=27225 and the fix was merged
> by Linus at
> http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=patch;h=f95df9ca6896978108201a77422a1ae2cdc595ec
>
> I believe this is the KMS equivalent to bug 475466 which was fixed in the
> radeon DDX.
>
> To unsubscribe from this bug, go to:
> https://bugs.launchpad.net/xorg-server/+bug/557611/+subscribe
>

--
Absence of proof is not proof of absence.
----
apparle

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Sorry, next time I tried it, it worked (with this error line) and the effects are working). But I need the new kernel, because with the new one internet and my scanner are working, with this kernel here not.
Thank you for your work!

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Is this patch (n the folder of this modified kernel) a patch to rum the effects with the normal kernel?
How could I use it?

Revision history for this message
Tormod Volden (tormodvolden) wrote :

Pondo, to use the patch you will have to compile the whole kernel. If Andy is happy with the test coverage here and applies it to the lucid kernel tree, it will appear as a kernel update later in Ubuntu 10.04. Meanwhile maybe you can try the 2.6.34-rc5 kernel from the Mainline Kernel PPA, so that you can have internet, scanner and video effects working at the same time.

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Will there be a solution for the error message in future?
At the moment I'm very busy, but I try to test new kernels, patches, ... in time.
I will start the new Ubuntu installation to use as a productive system around July.

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Hello, is there an newer kernel with the working driver or is there an easy way to modify my Ubuntu (Linux Mint), that the effects are working?
Thank you

Revision history for this message
Maarten Fonville (maarten-fonville) wrote :

Yes,

the current stable 2.6.34 kernel has at least this patch incorporated. For Ubuntu you can download one easily here:
http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.34-lucid/

download headers for 'all' + your own architecture
and download image for your own architecture

And install them of course :)

Revision history for this message
Apoorv Parle (apparle) wrote :

But has the 32MB bug been fixed, else it is useless. Already the radeon
driver performs not so well. So with 32MB memory it would be useless
--
Absence of proof is not proof of absence.
----
apparle

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Hello, Today I tried this Kernel, but my WLAN was instabil and my graphic shows a black screen with an error message: "Glib-Warning **: getpwuidnt ( ): failed due to unknown user id (0)"
I Had to reinstall, because deinstallation of the kernel wasn't a solution. Is it possible to edit some files (or something like that) to have the same effects with the actual kernel?
Thanks for helping

Revision history for this message
Pondo Sinatra (kommissar-keller) wrote :

Today I tried the kernel again; The desktop effects were working fine, but other usb-things didn't work anymore. even "lsusb" wont work in terminal (no output). Can you tell me the workaround to get the effects without changing the kernel?
Tank you

Andy Whitcroft (apw)
Changed in linux (Ubuntu):
assignee: Andy Whitcroft (apw) → nobody
Changed in xorg-server:
importance: Unknown → Medium
status: Unknown → Fix Released
Changed in xorg-server:
importance: Medium → Unknown
Changed in xorg-server:
importance: Unknown → Medium
Revision history for this message
Tormod Volden (tormodvolden) wrote :

Closing as per comment 38, the fix was included in upstream 2.6.34. I checked that it was also included in upstream 2.6.33.20, and in Ubuntu-2.6.32-33.71.

Changed in linux (Ubuntu):
status: Confirmed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.