The AMD/ATI Bar & Grill

b4283 · 2009-08-18 09:12:21

there's a 9.8 PKGBUILD written by csslayer in aur/catalyst reply area.
all patches are removed.
i can run quake 3 without any error now.

ZeroCool · 2009-08-18 10:42:37

Could it possible that the packages form csslayer corrupt? I can't unzip.

Last edited by ZeroCool (2009-08-18 14:23:31)

kensai · 2009-08-18 15:09:15

I have just uploaded to AUR, and I'm updating my repo for x86_64 to have tha 9.8 drivers

[catalyst]
Server = http://dev.archlinux.org/~eduardo/repo/x86_64

if someone can send me the i686 packages upload them and email me with the links please.

Cory · 2009-08-18 15:23:13

Dinth wrote:

Install and run dbus as daemon ?

Fixed after installing pacman -S pkgconfig cmake automoc4

Last edited by Cory (2009-08-19 07:25:33)

ZeroCool · 2009-08-18 16:00:40

very big thanks @kensai great job

venky80 · 2009-08-18 23:12:03

venky80 wrote:

AdrenalineJunky wrote:
1) Do you still have delay with kde menu and window resizing
no
2) Does resizing get fixed with xorg-server-maximize-fix
yes
3) Does flash work?
yes
4) How is the performance with 2.6.30? glxgears with 3d and 2d etc
glx gears ~1600 with compositing / ~ 1100 without
fgl_glx gears ~ 375 with compositing / ~ 750 without
on a 3200 igp
5) Are kernel patches still needed
no.
i have mobility radeon HD3670 and I am getting pretty low numbers with glx gears, with composting it is 950fps
fgl_glx gears with composting is only 250 (I thought mobility hd3670 was a better card...)
do u have any optimizations in your xorg.conf

Also mpeg4 (divx/xvid) videos goes black in full screen...it runs for a while then goes blank. I have tried the video in drangon player and mplayer it happens in both.
firefox starts with a dark black box but after that is usable.
I might have to go back to radeon driver

no one is having these issues? I am using godane's repos

lucasdemarchi · 2009-08-19 13:11:58

I got the following warnings when compiling catalyst modules:

Uncompressing ATI Proprietary Linux Driver-8.64...........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................................
make: Entering directory `/usr/src/linux-2.6.30-ARCH'
CC [M] /home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.o
In file included from /home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.c:443:
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/drm_proc.h: In function 'FGLDRM__vma_info':
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/drm_proc.h:497: warning: format '%08lx' expects type 'long unsigned int', but argument 5 has type 'phys_addr_t'
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.c: In function 'firegl_init_devices':
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.c:926: warning: integer constant is too large for 'unsigned long' type
CC [M] /home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.o
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:431:6: warning: #warning "ACPI notification wrapping won't work."
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:464:6: warning: #warning "ACPI notification wrapping won't work."
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:490:6: warning: #warning "ACPI notification wrapping won't work."
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:514:6: warning: #warning "ACPI notification wrapping won't work."

Did anyone get them too?

lucasdemarchi · 2009-08-19 13:40:58

I've just look in the code and:

>/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/drm_proc.h:497: warning: format '%08lx' expects type 'long unsigned int', but argument 5 has type 'phys_addr_t'
It's harmless....

>/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.c: In function 'firegl_init_devices':
>/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/firegl_public.c:926: warning: integer constant is too large for 'unsigned long' type

The following patch applies:
----------- cut
--- build_mod/firegl_public.c.old 2009-08-19 15:33:57.000000000 +0200
+++ build_mod/firegl_public.c 2009-08-19 15:34:27.000000000 +0200
@@ -923,7 +923,7 @@

#ifdef FIREGL_DMA_REMAPPING
//The GART unit of All supported ASICs has 40-bit address range.
- pci_set_dma_mask(pdev, 0xfffffffffful);
+ pci_set_dma_mask(pdev, 0xffffffffffull);
#endif

j++;
------- cut

CC [M] /home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.o
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:431:6: warning: #warning "ACPI notification wrapping won't work."
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:464:6: warning: #warning "ACPI notification wrapping won't work."
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:490:6: warning: #warning "ACPI notification wrapping won't work."
/home/lucas/build/catalyst-9.8/src/archive_files/common/lib/modules/fglrx/build_mod/kcl_acpi.c:514:6: warning: #warning "ACPI notification wrapping won't work."
It's a compilation option...
#if LINUX_VERSION_CODE < KERNEL_VERSION(2,6,29)
..
#else
/* No alternative on 2.6.29 and later for now. The internal ACPI data
* structures are no longer included in the kernel headers. */
#warning "ACPI notification wrapping won't work."
#endif

Don't know what we loose, however. Anybody?

dan.boff · 2009-08-19 21:45:14

i'm using the open source radeon driver. glxgears doesn't seem to work anymore, I only get a black window. it seems that 3d isn't working anymore at all (since the last kernel update maybe, but I don't know exactly ...) also neverball isn't working 'properly' (only the background is drawn), xmoto seems to work just fine though (besides loads of geometry corruption)

does anyone else using the opensource driver have the same issue with glxgears, or could there probably be a configuration problem on my side?

ras0ir · 2009-08-20 07:46:42

dan.boff wrote:

does anyone else using the opensource driver have the same issue with glxgears, or could there probably be a configuration problem on my side?

it works fine here, so it looks like a configuration issue.

dan.boff · 2009-08-20 09:51:28

hm, I update all the git package for my card on a daily basis. xmoto gets more playable everyday, with the last commits there's only some corruption in the background left, while the game itself is totally playable.
anyway glxgears only shows a blank window, I do have about 4000 fps on my ati 4850 though.

according to the xorg log I'm really using the experimental r600 driver

(II) RADEON(0): [dri] Visual configs initialized
drmOpenDevice: node name is /dev/dri/card0
drmOpenDevice: node name is /dev/dri/card0
(II) AIGLX: enabled GLX_texture_from_pixmap with driver support
(II) AIGLX: Loaded and initialized /usr/lib/xorg/modules/dri/r600_dri.so

I reconfigured my xorg.conf with X -configure, but no change with glxgears. Does anybody have any idea what might cause my problem. I really don't get why xmoto should work and glxgears shouldn't, afterall xmoto also uses opengl.

by the way: is it normal that with the experimental driver I get
direct rendering: No (If you want to find out why, try setting LIBGL_DEBUG=verbose)
with glxinfo?

flamelab · 2009-08-20 10:04:21

Try to login without xorg.conf as well (restart X of course)

Last edited by flamelab (2009-08-20 10:04:44)

dan.boff · 2009-08-20 10:44:35

no change without xorg.conf ...

cinan · 2009-08-20 10:59:39

I use fglrx driver. After latest update to 9.8 smooth resizing is working again (also with enabled composite extension in xorg.conf) but everything slows down when I'm running xcompmgr.
Do you have same problem?

PierreR · 2009-08-21 12:08:24

I have a HD 4670. My requirement is to have 1920×1080 hd resolution (16:9) up and running. I don't care about compiz or whatever.

I have tried both open source drivers (ati and radeonhd). According to xrandr none of them seems to be able to display full hd. Is it right ?

I have ended up installing catalyst and catalyst-utils. I think these packages end up downloading twice the ATI driver from the ATI web site. Does it make sense ?

A part from this, it works nicely. except that the screen sometimes flicks on the above right side of the screen. This is specially obvious when I have a blank screen (whatever the color is). And it is a real pain.

FY: I am using the latest catalyst version : 9.8.1 with the arch kernel 2.6.30. Direct rendering works fine according to "glxinfo | grep direct". Following the wiki, I have generated a xorg with aticonfig and did not bother modifying it. I am using xmonad.

Despite the excellent ATI wiki entry, I have found it rather difficult to understand how to get a proper ATI driver working to support full hd. For instance, I am not sure if both ati and radeonhd are supposed to handle it or not.

I have been reading many of the catalyst comments but it is not always easy to make sense of them. Is there a dedicated help page for catalyst and catalyst-utils. I realise that the install procedure is kind of changing at each version and it is probably a real pain to offer a decent support here.

Is there any obvious check I could try ? Is the future brighter with kernel 2.6.31 ?

So this is just a cry for help. In any case, it should be viewed as a critics toward Arch, the Aur package or whatever. On the contrary thanks so much for your work.

lucasdemarchi · 2009-08-21 12:28:16

I have tried both open source drivers (ati and radeonhd). According to xrandr none of them seems to be able to display full hd. Is it right ?

I don't see why you can't set it. Does it give a message related to the virtual size? You can try to put your resolution config directly in xorg.conf

A part from this, it works nicely. except that the screen sometimes flicks on the above right side of the screen. This is specially obvious when I have a blank screen (whatever the color is). And it is a real pain.

With the last catalyst driver I was having some artifacts when xorg-server-catalyst-maximize-fix was installed. With the default xorg-server, I do have the problem of maximizing delay with composite enabled, but no artifacts.

I have ended up installing catalyst and catalyst-utils. I think these packages end up downloading twice the ATI driver from the ATI web site. Does it make sense ?

Yes, if you installed from aur. There's no way to say that a package use the same file of another package, previously downloaded. Why don't you use kensai's repo?

Is there any obvious check I could try ? Is the future brighter with kernel 2.6.31 ?

Open source drivers are getting better and better... I hope in a no-so-distant-future it will support 3D decently. Catalyst does support 3D, but you receive a lot of annoyances as having to re-install on each kernel change and AMD/ATI taking too much time to support newer kernels. For example, as of know there's no support for the to be released kernel 2.6.31.
If you don't mind not having 3D support*, stay with open source drivers.

*I know recent versions of radeonhd and ati are starting to support it, but it's still very experimental.

vacant · 2009-08-21 13:29:30

PierreR wrote:

I think these packages end up downloading twice the ATI driver from the ATI web site. Does it make sense ?

Yes, to download once only, you can build catalyst-utils first (let it download), then after installing catalyst-utils, mv the downloaded driver to the catalyst PKGBUILD directory and proceed to build that.

PierreR · 2009-08-21 13:46:21

lucasdemarchi wrote:

I don't see why you can't set it. Does it give a message related to the virtual size? You can try to put your resolution config directly in xorg.conf

I have only tried it without any xconf (using only hal). xrandr output is not listing the native full hd resolution (1920x1080).

I guess it worths giving it another try and generate a xorg.conf. I don't really need 3D right now.

Is it true that radeonhd would be a better try for the HD 4670 (Radeon 7xx) ?

Thanks

snack · 2009-08-21 13:50:36

PierreR wrote:

Is it true that radeonhd would be a better try for the HD 4670 (Radeon 7xx) ?

Some weeks ago I tried to set up a dual monitor using xrandr, and I wasn't able to make it work properly with radeon, while with radeonhd it worked perfectly. So maybe radeonhd is a better choice to use xrandr (but things may have changed in the meantime... )

AdrenalineJunky · 2009-08-21 15:06:34

radeon/radeonhd tend to be about equivilent, from time to time one will have a feature that the other doesn't have for a while. i'd just pick one and try it, and if it wouldn't work, try the other.

kensai · 2009-08-21 15:29:11

I have uploaded the 9.8 catalyst to my i686 repository.

[catalyst]
Server = http://dev.archlinux.org/~eduardo/repo/x86_64

[catalyst]
Server = http://dev.archlinux.org/~eduardo/repo/i686

connexion2000 · 2009-08-22 12:08:37

Does anyone suffer X hangs using catalyst 9.8?

When there is no user input, screen goes black after some time (it should). But after moving mouse or pressing keyboard, computer doesn't react and monitor doesn't turn on. Why is that?

Dinth · 2009-08-22 13:04:30

Ive got that only once, but i use 3d screensaver. I dont know also if this occurs from 9.8 or occured earlier, as i've used open drivers before.

ChemBro · 2009-08-23 10:02:36

connexion2000 wrote:

When there is no user input, screen goes black after some time (it should). But after moving mouse or pressing keyboard, computer doesn't react and monitor doesn't turn on. Why is that?

Maybe your computer went to sleep and couldn't recover from that?

hunterthomson · 2009-08-23 22:33:01

Ok have installed the new Catalyst form Kensai's repo.

I still need the patched Xserver but all is well and working.

Thank you Kensai

p.s. Regnum is a way cool game

Last edited by hunterthomson (2009-08-23 22:33:54)

Arch Linux

#751 2009-08-18 09:12:21

Re: The AMD/ATI Bar & Grill

#752 2009-08-18 10:42:37

Re: The AMD/ATI Bar & Grill

#753 2009-08-18 15:09:15

Re: The AMD/ATI Bar & Grill

#754 2009-08-18 15:23:13

Re: The AMD/ATI Bar & Grill

#755 2009-08-18 16:00:40

Re: The AMD/ATI Bar & Grill

#756 2009-08-18 23:12:03

Re: The AMD/ATI Bar & Grill

#757 2009-08-19 13:11:58

Re: The AMD/ATI Bar & Grill

#758 2009-08-19 13:40:58

Re: The AMD/ATI Bar & Grill

#759 2009-08-19 21:45:14

Re: The AMD/ATI Bar & Grill

#760 2009-08-20 07:46:42

Re: The AMD/ATI Bar & Grill

#761 2009-08-20 09:51:28

Re: The AMD/ATI Bar & Grill

#762 2009-08-20 10:04:21

Re: The AMD/ATI Bar & Grill

#763 2009-08-20 10:44:35

Re: The AMD/ATI Bar & Grill

#764 2009-08-20 10:59:39

Re: The AMD/ATI Bar & Grill

#765 2009-08-21 12:08:24

Re: The AMD/ATI Bar & Grill

#766 2009-08-21 12:28:16

Re: The AMD/ATI Bar & Grill

#767 2009-08-21 13:29:30

Re: The AMD/ATI Bar & Grill

#768 2009-08-21 13:46:21

Re: The AMD/ATI Bar & Grill

#769 2009-08-21 13:50:36

Re: The AMD/ATI Bar & Grill

#770 2009-08-21 15:06:34

Re: The AMD/ATI Bar & Grill

#771 2009-08-21 15:29:11

Re: The AMD/ATI Bar & Grill

#772 2009-08-22 12:08:37

Re: The AMD/ATI Bar & Grill

#773 2009-08-22 13:04:30

Re: The AMD/ATI Bar & Grill

#774 2009-08-23 10:02:36

Re: The AMD/ATI Bar & Grill

#775 2009-08-23 22:33:01

Re: The AMD/ATI Bar & Grill

Board footer