Arch Linux

pyrrhicvictory

Member

From: Rradhat e UCK-se

Registered: 2007-12-21

Posts: 39

Direct rendering stopped working on Catalyst 8.4-1

I have just finished upgrading my kernel and installing catalyst 8.4-1, however,

Direct rendering has stopped working. In my /var/log/Xorg.0.log I see this (VERY long file, and this bit looked most pertinent to the issue at hand:

(II) fglrx(0): driver needs X.org 7.1.x.y with x.y >= 0.0
(WW) fglrx(0): could not detect X server version (query_status=-3)
(EE) fglrx(0): atiddxDriScreenInit failed, GPS not been initialized. 
(WW) fglrx(0): ***********************************************
(WW) fglrx(0): * DRI initialization failed!                  *
(WW) fglrx(0): * (maybe driver kernel module missing or bad) *
(WW) fglrx(0): * 2D acceleraton available (MMIO)             *
(WW) fglrx(0): * no 3D acceleration available                *
(WW) fglrx(0): ********************************************* *
(II) fglrx(0): FBADPhys: 0x3c000000 FBMappedSize: 0x04000000
(==) fglrx(0): Write-combining range (0xd8000000,0x4000000)
(II) fglrx(0): FBMM initialized for area (0,0)-(1280,8191)
(II) fglrx(0): FBMM auto alloc for area (0,0)-(1280,1024) (front color buffer - assumption)
(II) fglrx(0): Largest offscreen area available: 1280 x 7167
(==) fglrx(0): Backing store disabled
(II) Loading extension FGLRXEXTENSION
(II) Loading extension ATITVOUT
(**) fglrx(0): DPMS enabled
(II) fglrx(0): Using XFree86 Acceleration Architecture (XAA)
    Screen to screen bit blits
    Solid filled rectangles
    Solid Horizontal and Vertical Lines
    Offscreen Pixmaps
    Setting up tile and stipple cache:
        32 128x128 slots
        32 256x256 slots
        16 512x512 slots
(II) fglrx(0): Acceleration enabled
(II) fglrx(0): Direct rendering disabled
(==) fglrx(0): Silken mouse enabled
(==) fglrx(0): Using hardware cursor
(==) RandR enabled
(II) Initializing built-in extension MIT-SHM
(II) Initializing built-in extension XInputExtension
(II) Initializing built-in extension XTEST
(II) Initializing built-in extension XKEYBOARD
(II) Initializing built-in extension XC-APPGROUP
(II) Initializing built-in extension XAccessControlExtension
(II) Initializing built-in extension SECURITY
(II) Initializing built-in extension XINERAMA
(II) Initializing built-in extension XFIXES
(II) Initializing built-in extension XFree86-Bigfont
(II) Initializing built-in extension RENDER
(II) Initializing built-in extension RANDR
(II) Initializing built-in extension COMPOSITE
(II) Initializing built-in extension DAMAGE
(II) Initializing built-in extension XEVIE
(II) AIGLX: Screen 0 is not DRI capable
(II) Loading sub module "GLcore"
(II) LoadModule: "GLcore"
(II) Loading /usr/lib/xorg/modules/extensions//libGLcore.so
(II) Module GLcore: vendor="X.Org Foundation"
    compiled for 1.4.0.90, module version = 1.0.0
    ABI class: X.Org Server Extension, version 0.3
(II) GLX: Initialized MESA-PROXY GL provider for screen 0

I am completely confused as to what it is I can do to repair this issue. I am using arch64, and the card is a Radeon Xpress 200g series. (POS)
If anyone has any insight as to what can be done to alleviate the problem, please do not hesitate to post.

kazuo

Member

From: São Paulo/Brazil

Registered: 2008-03-18

Posts: 413

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

I got the same erro, with the same card. What kernel you are using? I'm using 2.6.25

I got a error on dmesg too:

fglrx: Unknown symbol flush_tlb_page
Symbol init_mm is marked as UNUSED, however this module is using it.
This symbol will go away in the future.
Please evalute if this is the right api to use, and if it really is, submit a report the linux kernel mailinglist together with submitting your code for inclusion.

Some info
http://forums.gentoo.org/viewtopic-p-5039144.html
http://www.phoronix.com/forums/showthread.php?t=7631

I gonna try to solve it tomorrow, now I need to sleep...

Last edited by kazuo (2008-04-20 04:43:07)

SpookyET

Member

Registered: 2008-01-27

Posts: 410

Re: Direct rendering stopped working on Catalyst 8.4-1

People, stop upgrading blindly. The announcement said that it does not work with 2.6.25.

pyrrhicvictory

Member

From: Rradhat e UCK-se

Registered: 2007-12-21

Posts: 39

Re: Direct rendering stopped working on Catalyst 8.4-1

I saw no announcement declaring that it would not work with 2.6.25... This must have appeared after I installed, or was a critical oversight on my part. In any regard, how would I remedy the situation?

kazuo

Member

From: São Paulo/Brazil

Registered: 2008-03-18

Posts: 413

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

People, stop upgrading blindly. The announcement said that it does not work with 2.6.25.

Sorry...

But I can't understand why pacman is forcing the kernel 2.6.25 to be installed,

 sudo pacman -U kernel26-2.6.24.4-1-x86_64.pkg.tar.gz
loading package data...
checking dependencies...
error: failed to prepare transaction (could not satisfy dependencies)
:: catalyst: requires kernel26>=2.6.25

sudo pacman -S catalyst                              
resolving dependencies...
looking for inter-conflicts...

Targets: kernel26-2.6.25-1  catalyst-8.4-1

EDIT:
I think a bug need to be filed for catalyst 8.4, it don't work with x86_64 as assumed on the pkgbuild, the fix only works for 32 bits systems, but as SpookyET stated I'm not gonna do any blindly action...

Last edited by kazuo (2008-04-20 15:27:14)

brain0

Developer

From: Aachen - Germany

Registered: 2005-01-03

Posts: 1,382

Re: Direct rendering stopped working on Catalyst 8.4-1

People, stop upgrading blindly. The announcement said that it does not work with 2.6.25.

Simo fixed it yesterday night.

I think a bug need to be filed for catalyst 8.4, it don't work with x86_64 as assumed on the pkgbuild, the fix only works for 32 bits systems

That is sad. Unfortunately, only two people in the dev team have ATI hardware, none of them on x86_64. So all we can do is find a fix and hope it works? If anybody knows where to find a solution for catalyst on 2.6.25, feel free to post it.

This mess proves again: ATI just sucks. Look what the nvidia guys do for us: http://www.nvnews.net/vbulletin/showthread.php?t=110088

pyrrhicvictory

Member

From: Rradhat e UCK-se

Registered: 2007-12-21

Posts: 39

Re: Direct rendering stopped working on Catalyst 8.4-1

This happens every time there is a new catalyst driver out. The only one that was remotely stable was the 8.3-2 driver (I believe). I am REALLY tired of having to deal with this sort of crap, but its an integrated card.

Cerebral

Forum Fellow

From: Waterloo, ON, CA

Registered: 2005-04-08

Posts: 3,108

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

Yeah, I've got no fix for this yet. The ATI guys are generally a release or two behind in getting newest kernel support.

kazuo

Member

From: São Paulo/Brazil

Registered: 2008-03-18

Posts: 413

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

The links I posted gives a "solution" (or a workaround...) Apply this patch in the kernel sources and compile

--- linux-2.6.25/arch/x86/kernel/init_task.c.orig    2008-04-20 23:28:28.000000000 -0300
+++ linux-2.6.25/arch/x86/kernel/init_task.c    2008-04-20 23:28:40.000000000 -0300
@@ -15,7 +15,7 @@
 static struct signal_struct init_signals = INIT_SIGNALS(init_signals);
 static struct sighand_struct init_sighand = INIT_SIGHAND(init_sighand);
 struct mm_struct init_mm = INIT_MM(init_mm);
-EXPORT_UNUSED_SYMBOL(init_mm); /* will be removed in 2.6.26 */
+EXPORT_SYMBOL(init_mm); /* will be removed in 2.6.26 */
 
 /*
  * Initial thread structure.
--- linux-2.6.25/arch/x86/kernel/smp_64.c.orig    2008-04-20 23:29:10.000000000 -0300
+++ linux-2.6.25/arch/x86/kernel/smp_64.c    2008-04-20 23:29:39.000000000 -0300
@@ -269,6 +269,7 @@
 
     preempt_enable();
 }
+EXPORT_SYMBOL(flush_tlb_page);
 
 static void do_flush_tlb_all(void* info)
 {

Last edited by kazuo (2008-04-21 04:08:19)

brain0

Developer

From: Aachen - Germany

Registered: 2005-01-03

Posts: 1,382

Re: Direct rendering stopped working on Catalyst 8.4-1

Did you test if this actually helps? If yes, that would be a viable workaround for now.

Cerebral

Forum Fellow

From: Waterloo, ON, CA

Registered: 2005-04-08

Posts: 3,108

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

Well, that certainly should help (assuming it's patching the correct files) - all it does is re-export the kernel symbols that fglrx uses.  The kernel maintainers saw fit to remove those exports in 2.6.25, and that's what's causing the problems.

fglrx: Unknown symbol flush_tlb_page
Symbol init_mm is marked as UNUSED, however this module is using it.

fglrx problems can often be fixed by re-exporting symbols that were hidden on a kernel upgrade, as they tend to make use of quite a few that end up disappearing.

kazuo

Member

From: São Paulo/Brazil

Registered: 2008-03-18

Posts: 413

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

Yes it help, actually I think it "solve". But I put the solve inside quotes because I think that the problem need to be solved with ati/amd not in the kernel.

I don't know the reasons for the kernel devs to remove the export of that functions... But Cerebral is correct the update is broken bacause that symbols (init_mm and flush_tlb_page) get un-exported.

And sorry I forgot to say, I tested it, it works without any warning and xv, 3d, all works (well works in the ati-sense)...

Last edited by kazuo (2008-04-21 12:29:21)

neotuli

Lazy Developer

From: London, UK

Registered: 2004-07-06

Posts: 1,204

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

I'm looking into this as well. It's possible a different function was exported that serves the same functionality... or something. I need to sift through lkml and the git repo to figure that out though.

EDIT: 2.6.26 is gonna be a doozie too. The files/functions in question here have all been moved around AGAIN

EDIT2: flush_tlb_page was unexported in 16da2f93054fc379522b93afc71d49751bd8be2b , because it "wasn't used". Its 32 bit brother didn't just happen to get cleaned up like the 64 bit one did. For once... this isn't ATI's fault, and there's actually nothing they can do about it except maybe find a way not to use the function. I'll be sending patches upstream to kernel-land, and I suspect we'll have to patch the -ARCH kernel to make up for this. sigh.

EDIT3: init_mm does not need to be patched. It is just a deprecation warning, and there is no reason to remove that warning. Functionality is not affected.

---

The suggestion box only accepts patches.

pyrrhicvictory

Member

From: Rradhat e UCK-se

Registered: 2007-12-21

Posts: 39

Re: Direct rendering stopped working on Catalyst 8.4-1

The questions are:
A) should I patch?
B) should I wait for a patched kernel (my humble apologies for the inconvenience regarding the faulty driver)?
C) tell ATI to stuff it you know where

I've never had to patch the kernel before, but I suppose there's a first time for everything. If A is the case, can someone point me in the right direction in terms of how it should be done (stock kernel is fine)?

brain0

Developer

From: Aachen - Germany

Registered: 2005-01-03

Posts: 1,382

Re: Direct rendering stopped working on Catalyst 8.4-1

The questions are:
A) should I patch?
B) should I wait for a patched kernel (my humble apologies for the inconvenience regarding the faulty driver)?

There will be another round of kernel-building on Friday or so (assuming I am here on the weekend). Unfortunately, I always compile at least 4 or 5 kernels until I upload one, so it takes all day.

C) tell ATI to stuff it you know where

I've never had to patch the kernel before, but I suppose there's a first time for everything. If A is the case, can someone point me in the right direction in terms of how it should be done (stock kernel is fine)?

I guess Travis or Simo should report this to ATI.

kazuo

Member

From: São Paulo/Brazil

Registered: 2008-03-18

Posts: 413

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

EDIT3: init_mm does not need to be patched. It is just a deprecation warning, and there is no reason to remove that warning. Functionality is not affected.

I patched it because this comment from http://forums.gentoo.org/viewtopic-p-5039144.html

I'm using kernel-2.6.25-rc8 and I couldn't load ati-drivers-8.471.3 even with your patch. The problem was unknown symbol init_mm, which is exported as an unused symbol since 2.6.25. I've tried config option CONFIG_UNUSED_SYMBOLS, but this did not help either, so I changed macro EXPORT_UNUSED_SYMBOL to EXPORT_SYMBOL in file arch/x86/kernel/init_task.c.
Anyway, it will be removed in 2.6.26 completely.

But I will test if it really works without init_mm later today and post back (can't compile a kernel now...).

EDIT. Ok tested, works.

Last edited by kazuo (2008-04-23 11:57:30)

Cerebral

Forum Fellow

From: Waterloo, ON, CA

Registered: 2005-04-08

Posts: 3,108

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

I guess Travis or Simo should report this to ATI.

ATI's actually already aware of it, so no worries there.  I'm hoping their next release will deal with it.

pyrrhicvictory

Member

From: Rradhat e UCK-se

Registered: 2007-12-21

Posts: 39

Re: Direct rendering stopped working on Catalyst 8.4-1

So we wait for the ati driver update, or the kernel patch?

Cerebral

Forum Fellow

From: Waterloo, ON, CA

Registered: 2005-04-08

Posts: 3,108

Website

Re: Direct rendering stopped working on Catalyst 8.4-1

ATI's next release will likely be no sooner than the second week in May, based on previous trends, so it's still a ways off

pyrrhicvictory

Member

From: Rradhat e UCK-se

Registered: 2007-12-21

Posts: 39

Re: Direct rendering stopped working on Catalyst 8.4-1

The latest ATI driver fixes the problem indeed. Thanks.

pedepy

Member

Registered: 2007-02-21

Posts: 198

Re: Direct rendering stopped working on Catalyst 8.4-1

how come I can't find either the latest catalyst or 2.6.25 kernel anywhere in the repos ?! ......

---

chupocabra ... psupsuspsu psu psu