summaryrefslogtreecommitdiffstats
path: root/src/gallium
Commit message (Collapse)AuthorAgeFilesLines
* gallium: use new util_fast_exp2(), _log2(), pow() functionsBrian Paul2008-08-223-2/+113
| | | | | New code surrounded with #if FAST_MATH to allow comparing against original code if we need to debug.
* gallium: new u_math.[ch] files for math functionsBrian Paul2008-08-224-0/+206
| | | | So far, optimized/low-precision versions of exp(), exp2(), log2(), pow().
* python/retrace: Interpret texture/surfaces destructions.José Fonseca2008-08-221-2/+8
|
* trace: Don't trace texture/surfaces releases, only destructions.José Fonseca2008-08-221-16/+18
|
* trace: Use a 4 hexadecimal digit suffix.José Fonseca2008-08-221-1/+1
|
* python/retrace: Process the trace call-by-call (instead of reading ↵José Fonseca2008-08-222-27/+23
| | | | everything into memory).
* python/retrace: Support gziped traces.José Fonseca2008-08-212-3/+14
|
* trace: Hack to detect writes to user buffers.José Fonseca2008-08-213-0/+64
| | | | | | It often happens that new data is written directly to the user buffers without mapping/unmapping. This hack marks user buffers and dumps them before passing them to pipe context.
* trace: Split the output stream on windows.José Fonseca2008-08-211-27/+75
| | | | Because windows limits the ammount of memory that can be mapped.
* gallium: test for and cull prims with inf/nan vertices in sp_setup.c code.Brian Paul2008-08-201-11/+34
|
* gallium: fix typo in LINE() macro (replace i+1 with i1 var)Brian Paul2008-08-201-1/+1
| | | | | We were sometimes referencing an invalid vertex. Fixes progs/trivial/line-clip.c test among others.
* python/retrace: Retrace draw_elements and draw_range_elements.José Fonseca2008-08-201-0/+6
|
* python: bindings for draw_range_elements.José Fonseca2008-08-201-0/+10
|
* gallium: fix do_flip bug in sp_surface_copy()Brian Paul2008-08-191-1/+4
| | | | | | | Surfaces are always in y=0=top raster order so the caller should invert the Y coordinate if needed; don't do it in sp_surface_copy(). Fixes a glCopyTexture regression.
* gallium: fix an assertionBrian Paul2008-08-191-1/+1
|
* gallium: WinCE build fixes.José Fonseca2008-08-192-0/+15
|
* python/retrace: Add several missing functions.José Fonseca2008-08-191-9/+48
|
* trace: Fix pipe_clip_state dump.José Fonseca2008-08-191-17/+6
|
* python: Handle null state.José Fonseca2008-08-191-0/+10
|
* python/retrace: Trim null chars.José Fonseca2008-08-191-0/+1
| | | | | They are often left in memory mapped files, and are not part of the XML accept chars anyway.
* trace: Support C++.José Fonseca2008-08-192-0/+18
|
* gallium: Add note about vs_2_0 EXPP mapping.Michal Krol2008-08-191-0/+1
|
* tgsi: Implement LOG opcode for SSE2 codegen.Michal Krol2008-08-192-5/+36
|
* tgsi: Implement EXP opcode for SSE2.Michal Krol2008-08-192-19/+53
|
* xlib: Revert destroying pipe_screen (temporary).José Fonseca2008-08-191-0/+3
|
* tgsi: Fix ARL opcode in SSE2 codegen.Michal Krol2008-08-192-25/+68
|
* tgsi: Workaround debug output buffer size limitations.Michal Krol2008-08-191-3/+18
|
* tgsi: Use NUM_CHANNELS.Michal Krol2008-08-191-4/+4
|
* trace: Explain how to integrate with a state tracker or winsys.José Fonseca2008-08-181-3/+42
|
* trace: Preliminary stream implementation for GDI.José Fonseca2008-08-183-14/+145
|
* trace: Use long longs to ensure covering 64bits integers.José Fonseca2008-08-182-6/+6
|
* trace: Get the trace file from the GALLIUM_TRACE option itself.José Fonseca2008-08-184-13/+12
|
* gallium: Remove TGSI_CC_UN.Michal Krol2008-08-181-6/+5
|
* tgsi: Update condition code vector when storing dest register.Michal Krol2008-08-181-1/+48
|
* tgsi: Saturate modifier obeys ExecMask.Michal Krol2008-08-181-4/+18
| | | | Implement NVIDIA [-1;+1] saturate mode.
* tgsi: Respect condition codes when storing destination register.Michal Krol2008-08-181-8/+113
|
* tgsi: Cosmetic changes.Michal Krol2008-08-181-19/+12
|
* tgsi: Add condition code (CC) register.Michal Krol2008-08-182-7/+35
|
* scons: Fix build.Michal Krol2008-08-181-1/+1
|
* if we can't find a mode, return first. At least we shouldAlan Hourihane2008-08-171-5/+5
| | | | | see the top left portion if we've got larger screens on other CRTC's
* consolidate intel directories.Alan Hourihane2008-08-1751-252/+30
| | | | | | we now have src/gallium/winsys/drm/intel/{common,dri,egl}
* rather than use CRTC 1 (aka LVDS), try and setup all connectors/encodersAlan Hourihane2008-08-171-19/+92
| | | | and get the same output on all.
* xlib: Kill xmesa_surface.José Fonseca2008-08-151-43/+19
| | | | | | | | | | A winsys cannot expect that the surfaces passed display_surface are the surfaces it created, as surface are now in generate texture views created by pipe_screen. Indeed, the fact it was working so far was mere luck. This fixes a weird tiled output when using the trace pipe driver. Winsys' surfaces should die.
* trace: Fix typo.José Fonseca2008-08-151-1/+1
|
* python/retrace: Update frame when PIPE_FLUSH_FRAME spotted.José Fonseca2008-08-151-1/+2
|
* trace: Wrap all textures and surface created by the pipe driver.José Fonseca2008-08-157-49/+378
| | | | | | | That is, Unfortunately, this causes a regression in softpipe, where the output gets tyled.
* trace: Allow multiple screens. Flush after call.José Fonseca2008-08-154-16/+53
|
* pyhon/retrace: Retrace surface contents.José Fonseca2008-08-151-9/+26
| | | | Now capable of replaying trivial/quad-tex-2d
* python/retrace: Keep addresses as strings.José Fonseca2008-08-152-3/+1
| | | | To simplify looking up these in the trace.
* trace: Several fixes.José Fonseca2008-08-152-9/+15
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-18 04:33:32 +0000