http://bugs.winehq.org/show_bug.cgi?id=28044
Summary: ATI HD 5850 R600g Gallium not recognized Product: Wine Version: 1.3.26 Platform: x86-64 OS/Version: Linux Status: UNCONFIRMED Severity: major Priority: P2 Component: -unknown AssignedTo: wine-bugs@winehq.org ReportedBy: emanuele.emalele@gmail.com
Created an attachment (id=35904) --> (http://bugs.winehq.org/attachment.cgi?id=35904) Exec log
Hi,
when trying to execute something with Wine I obtain a single screen with only corrupted graphic.
By reading the log (
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #1 from Emanuele emanuele.emalele@gmail.com 2011-08-10 07:24:41 CDT --- Created an attachment (id=35905) --> (http://bugs.winehq.org/attachment.cgi?id=35905) OpenGL DRI details
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #2 from Emanuele emanuele.emalele@gmail.com 2011-08-10 07:28:28 CDT --- ...SORRY FOR MISTAKE, DEFECT DESCRIPTION CONTINUE HERE...
By reading the log (Exec_log attached) I've found that my card is scanned as nvidia and of course also the DWORD string I've put in the registry to force recognition are not coupled with a nvidia card.
As per second attach you can read my Opengl DRI configuration.
I set major to gravity because not a single game can be viewable with:
HD5850+Gallium 0.4, R600g driver.
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #3 from Emanuele emanuele.emalele@gmail.com 2011-08-10 09:38:22 CDT --- I made some other discoveries.
I modified the directx.c file to set by default the HW_AMD_VENDOR and by recompiling all Wine sets it up correctly. Anyway in my opinion this is a bug and mine is NOT a fix because it's a forcing for ATI only.
So the bug could come from the fact that GL_VENDOR "X.Org" is not recognized and so NVIDIA is set.
By speaking about registry device id error in the Exec_log, I checked the wined3d_private.h 'supposedly' the GPU database (I do not know if I'm right) just to find that 6899 as device id is for ATI 5870 but it's false because as per lspci -n my HD5850 is reported as 6899.
Anyway as of now I've forced in the registry the value 6898 to adhere to wined3d_private.h for the HD5850 but the value is wrong or not all 5850 board has 6898, of course.
Now my custom Wine build complains only for the GL_VENDOR but sets AMD_VENDOR correctly and it seems to recognize HD5850.
Unfortunately the graphic corruption is still present but I can see a little improvement: something is moving in the screen and I cannot distinguish the objects but I can see more colours.
http://bugs.winehq.org/show_bug.cgi?id=28044
Dmitry Timoshkov dmitry@baikal.ru changed:
What |Removed |Added ---------------------------------------------------------------------------- Severity|major |normal
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #4 from Henri Verbeet hverbeet@gmail.com 2011-08-10 10:07:28 CDT --- (In reply to comment #0)
fixme:d3d_caps:select_card_nvidia_mesa Unknown renderer "Gallium 0.4 on EVERGREEN".
Looks like you're using old 32-bit drivers. The renderer string is supposed to read "Gallium 0.4 on AMD CYPRESS", like your glxinfo. "Gallium 0.4 on EVERGREEN" was only returned by some early versions of the r600g driver.
(In reply to comment #3)
Yes, PCI IDs aren't the same for all HD5850 / HD5870 cards. More importantly, you can't actually distinguish between HD5850 and HD5870 based on PCI ID. This doesn't matter a lot though, since both are reported as "ATI Radeon HD 5800 Series".
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #5 from Emanuele emanuele.emalele@gmail.com 2011-08-10 12:35:27 CDT --- Thank you Henri.
I've tried also with Wine64 but the same error arises. Apart from that I'm using multilib for 32bit compatibility.
The error is referred to X.Org as GL_VENDOR not recognized: is there a way to force it?
By meaning old drivers are you referring to Mesa because my glxinfo seems ok with the proper string.
Should I wait for a fix to Wine or do you think it's a misconfiguration of my system?
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #6 from Henri Verbeet hverbeet@gmail.com 2011-08-10 13:08:59 CDT --- (In reply to comment #5)
The error is referred to X.Org as GL_VENDOR not recognized: is there a way to force it?
I suppose you can try adding some detection to wined3d_guess_gl_vendor(), but it's not going to fix the real problem. The issue seems to be that you have old versions of the 32-bit r600g driver. Creating a WINEDEBUG=+wgl log should be able to tell you for sure. Even if we added detection for those older versions you'd likely have rendering issues.
By meaning old drivers are you referring to Mesa because my glxinfo seems ok with the proper string.
Yes, Mesa. glxinfo doesn't tell you much, because it's a 64-bit program. The WINEDEBUG=+wgl log mentioned above should tell you for sure what Wine is actually using.
Should I wait for a fix to Wine or do you think it's a misconfiguration of my system?
I think it's a misconfiguration.
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #7 from Emanuele emanuele.emalele@gmail.com 2011-08-10 13:29:41 CDT --- Ok the problem is clear.
After your post I checked up the 32bit Mesa and... they are of January (!!!).
You're right about corruption: also with my forcing the graphic doesn't appear clearly.
By googling I've found that it's a Gentoo known problem and the solution is to build up a 32bit chroot environment into the 64bit one.
I'll post a new reply just to confirm that the Gentoo suggested workaround has... worked :)
Thanks for your support.
http://bugs.winehq.org/show_bug.cgi?id=28044
--- Comment #8 from Emanuele emanuele.emalele@gmail.com 2011-08-11 07:41:59 CDT --- Dear Mr. Verbeet,
it seems you've solved my issue.
I followed Gentoo guide, created 32bit system into a 64bit one, launched Wine 1.3.26 and not a single glitch is viewable.
This 'bug' can be closed now because the problem disappeared and with Gallium ATI does render really better than with prop drivers.
Thanks again and for anyone who is using Wine... please use 32bit :)
Regards!
http://bugs.winehq.org/show_bug.cgi?id=28044
Benjamin Hodgetts ben@xnode.org changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|UNCONFIRMED |RESOLVED CC| |ben@xnode.org Resolution| |INVALID
--- Comment #9 from Benjamin Hodgetts ben@xnode.org 2011-08-24 05:43:22 CDT --- Distro / Driver issue.
http://bugs.winehq.org/show_bug.cgi?id=28044
André H. nerv@dawncrow.de changed:
What |Removed |Added ---------------------------------------------------------------------------- Status|RESOLVED |CLOSED CC| |nerv@dawncrow.de
--- Comment #10 from André H. nerv@dawncrow.de 2011-09-08 15:57:47 CDT --- closing invalid
-
wine-bugs＠winehq.org