Ati has released their Catalyst 8.9 drivers. This bug contains ebuild suggestion for the drivers. Tested on x86, radeon 4850. Wine finally works. Ebuild addresses the issue http://bugs.gentoo.org/show_bug.cgi?id=229153. Reproducible: Always Steps to Reproduce:
Created attachment 165871 [details] ebuild suggestion
Thanks for the report and the proposed ebuild, assigning to maintainers
With Catalyst 8.9 + linux 2.6.26 (Gentoo) + Sapphire Radeon HD 2400 XT, PCI-E, 256 Mo, my two previous problems were solved : - Windowed blender works (there are still some screen corruptions when blender is launched, like taskbar which becomes black, but it works) - Digikam OpenGL slideshow works ( http://ati.cchtml.com/show_bug.cgi?id=1250 )
If i might to suggest something. Naming of that ebuild finally overgrown that old one so ebuild can be named ati-drivers-8.9.ebuild so you can use ${P/./-} in src_uri so it will get more automatic. (i was thinking about this since version 8.6 but allways forget to write it :])
(In reply to comment #4) > If i might to suggest something. Naming of that ebuild finally overgrown that > old one so ebuild can be named ati-drivers-8.9.ebuild so you can use ${P/./-} > in src_uri so it will get more automatic. (i was thinking about this since > version 8.6 but allways forget to write it :]) well... that's an idea. But the module version that shows up in dmesg is 8.53.2. So both ways of numbering the ebuild are reasonable. I think, that sticking to "the good old ways" is ok. Maybe a tip for ati on their unofficial bugzilla might do the trick in some time...
Also one "tiny" issue with this release HD3850 and HD4850 does not start with kde4x when composite is enabled...
Thank you for the notice, Please next time attach the proposals as diff please.
(In reply to comment #6) > Also one "tiny" issue with this release HD3850 and HD4850 does not start with > kde4x when composite is enabled... > X server crashes after installing Catalyst 8.532 (this ebuild with an amd64 arch) with message> "Resources temporarily unavailable. X crashed with signal 11". Always reproducible.
(In reply to comment #4) > If i might to suggest something. Naming of that ebuild finally overgrown that > old one so ebuild can be named ati-drivers-8.9.ebuild so you can use ${P/./-} > in src_uri so it will get more automatic. (i was thinking about this since > version 8.6 but allways forget to write it :]) > Do you mean ${PV/./-}? Like this?: SRC_URI="${ATI_URL}/ati-driver-installer-${PV/./-}-x86.x86_64.run" Also, I tried renaming to ati-drivers-8.9.ebuild and setting SRC_URI but it was still listed as a downgrade. Perhaps because 522 > 9? Am I missing something?
Works for me, however: On my HD3200 ati-drivers 8.512, 8.522, and 8.532 all make X stutter when combined with 2.6.25 or 2.6.26. My old Mobility Radeon 9600 works fine with 8.532 and kernel 2.6.26. My HD3200 is working fine now with 8.532 and kernel 2.6.24.
(In reply to comment #9) > (In reply to comment #4) > > If i might to suggest something. Naming of that ebuild finally overgrown that > > old one so ebuild can be named ati-drivers-8.9.ebuild so you can use ${P/./-} > > in src_uri so it will get more automatic. (i was thinking about this since > > version 8.6 but allways forget to write it :]) > > > > Do you mean ${PV/./-}? Like this?: > > SRC_URI="${ATI_URL}/ati-driver-installer-${PV/./-}-x86.x86_64.run" > > Also, I tried renaming to ati-drivers-8.9.ebuild and setting SRC_URI but it was > still listed as a downgrade. Perhaps because 522 > 9? Am I missing something? > Correct it is still downgrade :( i just did not realize that 522 is indeed biger than 9 :(
About that versioning, we had to wait for next year, cause 9.X is allways bigger than 8.X :]
This version is not compatible with xorg 7.4 or xorg-server 1.5, which is why some of you might be experiencing problems. You'll have to wait till next month for that.
I am not that crazy to test it with 7.4 and xorg-server 1.5 it is not working with 1.4.2 and 7.3
Created attachment 166188 [details] Xorg.log after installing ati-drivers-8.532 and rebooting Here is a the Xorg.log after installing ati-drivers-8.532 and rebooting, the last lines surely will indicate what's wrong
(In reply to comment #15) > Created an attachment (id=166188) [edit] > Xorg.log after installing ati-drivers-8.532 and rebooting > > Here is a the Xorg.log after installing ati-drivers-8.532 and rebooting, the > last lines surely will indicate what's wrong > Could you please post your xorg.conf? Does revdep-rebuild show anything that needs rebuilding? Did fglrx module load properly without modprobe errors? Something's also wrong with your DRI... Can someone else please confirm/deny that this ebuild works on amd64/x86_64 architectures?
(In reply to comment #15) > Created an attachment (id=166188) [edit] > Xorg.log after installing ati-drivers-8.532 and rebooting > > Here is a the Xorg.log after installing ati-drivers-8.532 and rebooting, the > last lines surely will indicate what's wrong > w8 a minute. Your log, shows that you've got 8.53.4 kernel module, and 8.52.3 xorg driver... are you sure you did eselect opengl mesa and eselect opengl ati after installing the driver? Is this really the log from after rebooting? I'll check my logs in the evening and post if that are the same version numbers that i have in my Xorg.log.
The crash from FontFileCompleteXLFD() is fixed for me by upgrade pixman to recently released 0.12 or recompile 0.11 with gcc 4.1 Anyway the drivers displays only a black screen and lockups the system randomly on start :( 2.6.26, AMD64, 4GB, HD4850
Created attachment 166289 [details] Xorg.conf as requested Here is the xorg.conf file, I can confirm that the Xorg.log attachment is the one after installing ati-drivers-8.532. No errors at dmesg about fglrx and revdep-rebuild shows that the linking is consistent.
Created attachment 166291 [details] Xorg for ati-drivers-8.522 for comparison Here is the Xorg.log generated with ati-drivers-8.522, hope it is useful.
Created attachment 166293 [details] The confirmed Xorg.log after rebooting and making eselect opengl ati I could confirm that it is certainly the Xorg.log generated after installing ati-drivers-8.532 and making eselect opengl ati as suggested.
Created attachment 166428 [details] Xorg.log on amd64 (working configuration) Works here on amd64 using a Radeon HD 3870.
Created attachment 166430 [details] xorg.conf on amd64 (working configuration)
I'm not that great with ebuilds (so no adjustments attached), but this driver also contains a custom libdri.so from ati. Usually should go into /usr/lib/xorg/modules/extensions/libdri.so (or whatever's appropriate for the system). This driver can crash (especially with dual head systems) without using ati's lib.
This release isn't going to make it into the tree by my hands. It's incompatible with xorg-x11-7.4 (bug #236871) so I'd rather wait for another release. If you want a nice driver in the meantime I recommend the radeonhd live svn ebuild in my overlay. It's niiiice :)
but it supports randr 1.2 with hw accelerated rotation, this is more than all others can do. And it fixes a lot of issues. Furthermore it runs really fine with xorg-server <1.5 so commit it at least masked.
I agree, I don't see how the new driver is any worse than the older ones (better in many cases, it seems). AFAIK, none of the ati-drivers officially support xorg-server 1.5, so why should that be a reason to not let this one in?
(In reply to comment #16) > (In reply to comment #15) > > Can someone else please confirm/deny that this ebuild works on amd64/x86_64 > architectures? > yeah! works well except wget and the ssh connection that asks for a --no-check-certificate, ... I've done it manualy.. with x11-base/xorg-server-1.3.0.0-r6 and an radeon x1200, ... even if it is not list anymore in compatible hardware list...... hope It will not burn my video card... at least I got blender -w working.... xfce integrate composite, xcompmgr, compiz are working too so cool....
(In reply to comment #16) > > Can someone else please confirm/deny that this ebuild works on amd64/x86_64 > architectures? Working okay here. Just play networked UT for an hour with my wife. :D
Okay, you want it you got it. I've no way of testing this ebuild, so if someone can confirm the ati-drivers ebuild in my overlay works, then I'll commit it to the tree. (layman -a je_fro and emerge ati-drivers)
(In reply to comment #30) > Okay, you want it you got it. > I've no way of testing this ebuild, so if someone can confirm the ati-drivers > ebuild in my overlay works, then I'll commit it to the tree. > (layman -a je_fro and emerge ati-drivers) > ebuild seems to works fine here on HD3200
(In reply to comment #15) > Created an attachment (id=166188) [edit] > Xorg.log after installing ati-drivers-8.532 and rebooting > > Here is a the Xorg.log after installing ati-drivers-8.532 and rebooting, the > last lines surely will indicate what's wrong Hi, I've checked your log. Something's not right with you xorg driver, as it's still of an old version. It should be the same as the kernel driver version. I got both 8.53.4.
In cvs.
Thanks to all on this bug! :) je_fro
(In reply to comment #34) > Thanks to all on this bug! > :) > je_fro > Still the same thinhg with the portage ebuild, same error and already reinstalled X.
(In reply to comment #35) > (In reply to comment #34) > > Thanks to all on this bug! > > :) > > je_fro > > > > Still the same thinhg with the portage ebuild, same error and already > reinstalled X. > Recompiled Kernel (2.6.25-gentoo-r7) and ati-drivers started working, tough I did not modify any option, just trying to look for something wrong.