Am 2003.08.19 22:20:58 +0200 schrieb(en) Michel Dänzer:
So here it is at last. :)
http://penguinppc.org/~daenzer/DRI/radeon-pci-roundup.diff
It ended up quite a bit larger than I initially thought it would, but most of it is straightforward (and some less straightforward, like a DDX driver option and 3D driver environment variables) renaming. The main semantic change is the removal of all artifical limits to GART features for PCI GART. Then there's also the new detection of card type from the subsystem ID; I haven't got much of any data about this yet, we'll see how it fares. It seems to work for AGP Mac cards at least. :)
Comments appreciated; if there are no serious objections, I'll probably commit it sometime this week.
I havent looked at it really carefully, but suddenly I found a typo, maybe: radeon_screen.c: + screen->numTexHeaps = RADEON_NR_TEX_HEAPS - 1; should look IMHO: + screen->numTexHeaps = RADEON_NR_TEX_HEAPS;
Is the R200_GART_CLIENT_TEXTURES (was: R200_AGP_CLIENT_TEXTURES) envvar useful? You changed the notes/FIXMEs which got in with the texrect patch. (radeon_texmem.c radeon_texstate.c) So would a RADEON_GART_CLIENT_TEXTURES be useful for the radeon as well? (or for the igp-chips?)
I haven't looked at the patch yet, but this might be correct. The R200 driver doesn't use AGP memory for "normal" textures. The only way to use GART memory on the R200 is by using either GLX_NV_vertex_array (older versions) or GLX_MESA_allocate_memory (as of yesterday).
------------------------------------------------------- This SF.net email is sponsored by Dice.com. Did you know that Dice has over 25,000 tech jobs available today? From careers in IT to Engineering to Tech Sales, Dice has tech jobs from the best hiring companies. http://www.dice.com/index.epl?rel_code4 _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
