summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* fix broken SWZ instructionBrian Paul2005-10-212-12/+13
|
* fix broken KILBrian Paul2005-10-211-2/+4
|
* add the tile bits for offset, and unk bit 31 of sizeDave Airlie2005-10-211-0/+3
|
* restore FASTCALL stuffBrian Paul2005-10-205-7/+25
|
* fix bad loops in TransposeMatrix[fd]()Brian Paul2005-10-201-2/+2
|
* fix another renderbuffer wrapping bugBrian Paul2005-10-201-4/+3
|
* Disable the FASTCALL optimization because it still breaks server-side GLX.Brian Paul2005-10-205-25/+7
|
* Bump driver dates for the crasher fixes yesterday.Eric Anholt2005-10-202-2/+2
|
* descrip.mms syntax correctionJouk Jansen2005-10-201-1/+1
|
* fixes for r300_tex.c from Aapo TahkolaDave Airlie2005-10-201-4/+4
|
* Make SiS a little more like other drivers: Add SIS_DEBUG (only option beingEric Anholt2005-10-194-9/+63
| | | | | "fall" so far) and make the fallback_force option be no_rast like other drivers. Incidentally, makes no_rast not crash (though it still fails to render).
* Remove some long-dead code.Eric Anholt2005-10-194-215/+0
|
* Fix a missed merge for idr's dispatch changes: failure to do driInitExtensionsEric Anholt2005-10-192-1/+12
| | | | in __driCreateNewScreen_20050727 resulted in crashing in dispatch.
* updated for new mm.c, execmem.c filesBrian Paul2005-10-191-0/+6
|
* s/memcpy/_mesa_memcpy/, s/FREE/_mesa_free/Brian Paul2005-10-194-8/+8
|
* Bug #4615: Fix the SiS driver for the renderbuffer changes. Previously, allEric Anholt2005-10-1911-197/+252
| | | | | | | drirenderbuffers pointed at screen offset 0 and NULL. Instead, set up the front buffer at startup but leave the others for later, since sis allocates them on demand rather than using the static method of the other non-sis-descendent drivers. Some basic apps work, but fallbacks may be broken still.
* Fix a missed merge for idr's dispatch changes: failure to do driInitExtensionsEric Anholt2005-10-182-1/+12
| | | | | in __driCreateNewScreen_20050727 resulted in crashing in dispatch. Rendering remains broken.
* move _mesa_exec_malloc/free() up near other malloc/free prototypesBrian Paul2005-10-181-11/+8
|
* Lots of clean-ups. Replace __inline__ with INLINE macro. Use theBrian Paul2005-10-183-50/+134
| | | | | _mesa_calloc(), _mesa_free() functions. Clean-up formatting, doxygen-style comments.
* Add _mesa_exec_malloc() and _mesa_exec_free() for allocatingKeith Whitwell2005-10-1811-7/+122
| | | | | executable memory. Based on Thomas Hellstrom's patch. TODO: glapi.c also needs this, but cannot access this code.
* on r200, don't emit fog coords when fog isn't actually enabled. Fix bug with ↵Roland Scheidegger2005-10-183-4/+7
| | | | vtxfmt not causing a fallback when fog coords are active.
* add GL_EXT_fog_coord support to radeon driver. No vtxfmt code (just uses ↵Roland Scheidegger2005-10-179-57/+215
| | | | fallback) yet. May cause a tcl fallback if fog coord is used together with separate specular lighting. Fog factors are precomputed and then submitted instead of the raw coords (chip limitation, same as on r200).
* fix some obvious typos in radeon_reg.h (only currently unused bits affected)Roland Scheidegger2005-10-161-4/+4
|
* enable cube maps on radeon (#2241 on bugzilla). No vtxfmt code yet (just ↵Roland Scheidegger2005-10-1513-19/+201
| | | | generates vfmt fallback). Code by Andreas Stenglein, some small adjustments by me.
* fix two bugs wrt calculating max possible texture sizes for texture ↵Roland Scheidegger2005-10-151-3/+3
| | | | rectangles/cube maps.
* enable 3rd texture unit on radeon (default still 2 enabled units). Disable ↵Roland Scheidegger2005-10-1317-251/+285
| | | | some multitexcoord codegen stuff noone understands to make it work. Replace most code testing explicitly for unit 0 and 1 with loops instead of adding test for unit 2, smaller/more readable code at the cost of maybe some slight performance hit. (Code provided by Andreas Stenglein, some adjustments by me.)
* make sure texture matrix is updated often enough with texture rectanglesRoland Scheidegger2005-10-131-0/+2
|
* fix potential bug when emitting elts with lots of componentsRoland Scheidegger2005-10-131-1/+2
|
* remove unused _mesa_log2() (bug 4468)Brian Paul2005-10-121-3/+0
|
* change includes (bug 4468)Brian Paul2005-10-121-6/+3
|
* fix GLAPI/extern problem (bug 4757)Brian Paul2005-10-122-3/+3
|
* Minor driver cleanup. Remove unnecessery/unneded ↵Roland Scheidegger2005-10-117-45/+4
| | | | radeon/r200AllocDmaRegionVerts function. No longer compile the radeon_compat stuff which was for drm 1.1 compatibility presumably but unused since about forever into the driver.
* ifdef out some code to make glapi.c linkable in libglx.so (on theIan Romanick2005-10-111-1/+5
| | | | | server). This isn't strictly needed yet, but it doesn't hurt anything either. However, the patches for fd.o bug #2996 require this to build.
* check for PIXMAP in clip routine (SF bug 1323215)Brian Paul2005-10-111-1/+1
|
* fixed constant fogcoord bug (#4729)Brian Paul2005-10-101-2/+8
|
* Add 'get' data for glGetFramebufferAttachmentParameterivEXT enums.Ian Romanick2005-10-102-6/+16
| | | | Remove an extra closing parenthesis in the Linux version of the SWAP_32 macro.
* do not set format of the incoming vertices needed for fog_coord in the ↵Roland Scheidegger2005-10-081-6/+1
| | | | generic state functions. This is done in the vtxfmt/hwtcl/swtcl pathes respectively.
* remove code dealing with drm versions older than 1.3. The driver requires ↵Roland Scheidegger2005-10-081-34/+16
| | | | (and requests) 1.3 since quite some time already to work at all.
* change test for presence of accum bufferBrian Paul2005-10-081-1/+1
|
* Fix some issues with state updates and renderbuffers. Querying GL_RED_BITS,Brian Paul2005-10-084-28/+64
| | | | etc. after calling glRenderBufferStorageEXT gave undefined results.
* MATH_DEBUG changes from bug #4468.Brian Paul2005-10-078-19/+18
|
* Populate arb_fp_instruction negate field correctly.Keith Whitwell2005-10-071-1/+1
|
* Fix a number of point size attenuation problems.Brian Paul2005-10-075-85/+115
| | | | Move size clamping into the rasterization function.
* fix problems with some texgen modes by putting the chip into LOCAL_VIEWER ↵Roland Scheidegger2005-10-051-7/+21
| | | | mode for lighting (potentially causes slight misrenderings with the specular components of lights if the light model is not set for local viewer).
* reverse accidentally commited part of last commit (makes it compile again...).Roland Scheidegger2005-10-051-0/+1
|
* enable point sizes larger than 1 (for aliased points only) by using the hw ↵Roland Scheidegger2005-10-059-25/+59
| | | | point sprite primitive.
* fix byteswapping bug in _mesa_texstore_ycbcr() (Benjamin Herrenschmidt)Brian Paul2005-10-051-7/+7
|
* fix typo in radeon_maos_arrays.c from last commit (code currently unused, ↵Roland Scheidegger2005-10-051-2/+2
| | | | that path is compile-tested only).
* remove unneeded parameter from radeonUploadTexMatrix.Roland Scheidegger2005-10-054-8/+9
|
* Remove the tcl fallback for texture rectangle (by manipulating the texture ↵Roland Scheidegger2005-10-0510-100/+299
| | | | matrix) (tested with texrect). Enable texgen for r/q coordinates (tested with projtex). Fix projected texcoords when an app uses TexCoord3x and the texture matrix to save on vertex size (fixes ut2k3 shadow projectors in tcl mode). From texgenmix, all cases with all texgen or no texgen work, with the exception of texgen enabled for s/t only, this one works with hw tcl, but not with vtxfmt (suspect issues with vtxfmt), the mixed cases do not work (which is expected, and should be rare in practice), with the exception of the first one which hits a tcl fallback.
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-21 12:35:08 +0000