http://bugs.winehq.org/show_bug.cgi?id=19412
--- Comment #22 from Charles Davis cdavis@mines.edu 2009-07-29 16:40:42 --- (In reply to comment #21)
Created an attachment (id=22689)
--> (http://bugs.winehq.org/attachment.cgi?id=22689) [details]
Test case for the glCompressedTexImage2D bug
The glCompressedTexSubImage2D bug never occured on Macs with Nvidia cards. I only positively reproduced it on Intel GMA X3100 macs, and I think it also occurs on ATI cards, but I have to test there.
Attached is the test case I sent to Apple for this bug. If it runs properly, it should show one of Half Life 2's static backgrounds(pretty dark, because its sRGB'ed). If the driver has the bug, the app shows either a static color or random garbage.
Compile with gcc srgbtexture.c -o srgbtexture -framework opengl -framework glut. There might be a bunch of compile errors due to the way I include the glut headers(or rather, do not include them)
Funny... it crashes and burns on my system with a SIGBUS. When I debugged your test program, I found that the stack was corrupted somehow, and the CPU had jumped to some random address. Note the value in EAX:
(gdb) bt #0 0xffff08ab in ?? () #1 0x00000001 in ?? () (gdb) info reg eax 0xffff088b -63349 ecx 0x0 0 edx 0xfffffc80 -896 ebx 0x11f07b7a 300972922 esp 0xbfffea90 0xbfffea90 ebp 0xbfffea98 0xbfffea98 esi 0x25360 152416 edi 0x15a56ca0 363162784 eip 0xffff08ab 0xffff08ab eflags 0x10282 66178 cs 0x17 23 ss 0x1f 31 ds 0x1f 31 es 0x1f 31 fs 0x0 0 gs 0x37 55
Is this why you could never reproduce this with an nVidia card, or is something else at work? (I think it's the latter; see below.)
The odd part is that Henri's patch works fine as far as I can tell. Maybe wined3d does something that your test program isn't doing (or the other way around).
(And personally, I'd never even consider gaming with Intel's integrated graphics solutions.)