summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* If an application cleared before any state had been emitted, that clear wouldEric Anholt2004-10-022-4/+9
| | | | | | | | happen before any state had been set, causing a hang later on. Fix this by calling r200Flush instead of FIREVERTICES (which checks if any state has been emitted but not flushed, before calling Flush) in r200Clear. While here, add some more debugging info which was useful, and remove an unnecessary save/restore in BackUpAndEmit.
* Convert Rage 128 over to t_vertex.c. While it's slightly slower (10% in ipers)Eric Anholt2004-10-0214-786/+195
| | | | | | | | it's not that big of a deal in more normal apps, and axes a good bit of code. And I assume that t_vertex will only get faster. Removes ~43k from compiled binary. Tested with: quake3, ut, ipers, texcyl, chromium, tuxracer, neverball (kinda)
* Add an option for vertices emitted to be swapped CPU_TO_LE32, to be used by theEric Anholt2004-10-021-0/+7
| | | | r128 code.
* Cut out all of the old texture environment code and rewrote it fromIan Romanick2004-10-024-4759/+390
| | | | | | | | | scratch. There were just too many cut-and-paste errors in the code, and it was too hard to follow. This fixes Mesa bugzilla #1508. #1509 is probably also fixed, but I'd like to do some more testing before I close that particular bug. Additionally, this cuts almost 4,600 lines of code from the driver.
* minor updates and improvementsKarl Schultz2004-10-011-11/+28
|
* Add the glut project to the demos workspace.Karl Schultz2004-10-011-0/+138
|
* Remove glut project from the Mesa workspace. This allows users to buildKarl Schultz2004-10-011-26/+11
| | | | | | this entire workspace by only unzipping MesaLib.zip. The project will be moved to the demos workspace, which is contained in the MesaDemos zip file.
* Add HP shared lib configs.Karl Schultz2004-10-011-0/+2
|
* HP shared lib configKarl Schultz2004-10-012-0/+53
|
* Replace the numerous switch-statements to derive GL_COMBINE post-scaleIan Romanick2004-10-011-419/+208
| | | | | | | | state with a function. This function sets the bits correctly. Did some trivial refactoring on some of the GL_COMBINE code. First pass at replacing classic texture environments with GL_COMBINE. This is controlled by the EXPERIMENTAL_COMBINE_MODE define.
* fix problems in previous versionKarl Schultz2004-10-011-3/+3
|
* Enable builds on AIX 64-bitKarl Schultz2004-10-013-1/+53
|
* Fix mesa solo with doublebuffered apps, the drawable index/stamp weren'tDave Airlie2004-10-011-4/+11
| | | | being updated properly..
* Factored out code for setting texture coordinate wrap bits. Added supportIan Romanick2004-09-302-24/+47
| | | | | | for GL_ARB_texture_mirrored_repeat. Enabled GL_NV_blend_square. It has always actually been supported. Removed redundant EXT versions of ARB extension strings.
* set span.y prior to each span write in draw_depth_pixels, like Karl's other fixBrian Paul2004-09-301-11/+11
|
* Bugzilla #1058: Fix some potential 64bit pointer issues by storing differencesEric Anholt2004-09-301-3/+3
| | | | | | between pointers in appropriate types. Submitted by: Ronny V. Vindenes <[email protected]>
* OK, one more time. Simplify the state-backup system by just storing the fullEric Anholt2004-09-3013-108/+95
| | | | | | | | | state in a ready-to-emit cmdbuf, which avoids the issue Nicolai Haehnle reported where the check() could return differently during backup-and-emit than it should have if it were called at the right time. Move the lit emission before most of the TCL state emission on r200, which fixes neverball issues. Tested with: r100/r200 with neverball, tuxracer, chromium, quake3, ipers
* Fix bug in glDrawPixels when writing color indexed images on color indexedKarl Schultz2004-09-281-2/+2
| | | | | devices. In draw_index_pixels(), use the correct variables for the span Y and span end values.
* Initial revisionNicolai Haehnle2004-09-2828-0/+8258
|
* use GLuint instead of uintDaniel Borca2004-09-281-1/+1
|
* version bumpBrian Paul2004-09-271-1/+1
|
* change include of glx_mangle.h to match other headersBrian Paul2004-09-271-3/+3
|
* remove beos/*.h files from tarball listBrian Paul2004-09-271-1/+0
|
* bump minor version to 2Brian Paul2004-09-271-1/+1
|
* VBO RefCount fix (David Reveman)Brian Paul2004-09-272-0/+10
|
* stencil wrap works, either HW or SW (Ian Romanick)Daniel Borca2004-09-271-4/+1
|
* Bump driver_date for today's fixes.Eric Anholt2004-09-251-1/+1
|
* Can't bail in sisDDLogicOpCode when logicop is off, because it's called withEric Anholt2004-09-251-3/+0
| | | | GL_COPY to turn off logicop. Fixes glean's logicop test.
* Triangle stipple is a fallback, since we don't have code for the stippling.Eric Anholt2004-09-251-1/+1
| | | | Doesn't seem to help with glean's paths test, but I'm pretty sure it's correct.
* Correct a couple of comments.Eric Anholt2004-09-251-2/+2
|
* Fix SiS AGP vertex dispatch by not trying to emit 0 vertices, which would hangEric Anholt2004-09-253-3/+5
| | | | the hardware. Re-enable AGP by default.
* Add sis to the list of DRI drivers, since it had already been converted to theEric Anholt2004-09-251-4/+4
| | | | new interface.
* Refactor the linux-dri config files so that linux-dri-x86 inherits fromEric Anholt2004-09-252-32/+16
| | | | | | | linux-dri. Turn on -O by default (-O2 provides a slight improvement in performance, at a large cost to debuggability). Turn off -fPIC on x86 by default. Turn on -Wall, to catch more stupid mistakes. This could be both done cleaner, and done for more config files, but it works for me for now.
* The previous code would emit a full set of state during the first EmitState onEric Anholt2004-09-2510-150/+177
| | | | | | | | | | | | a new cmdbuf, to ensure that state wasn't lost across UNLOCK/LOCK pairs (in the case of context switching). This was rather inefficient. Instead, after flushing a cmdbuf, mark the state as needing to be saved on unlock. Then, at the beginning of flushing a cmdbuf, if we actually have lost the context, go back and emit a new cmdbuf with the full set of state, before continuing with the cmdbuf flush. Provides a 10-15% improvement in ipers performance in my tests, along with other apps. Tested with: ipers, glxgears, quake3
* Fix texturing in quake3. Some code was left over from the dirty/clean listEric Anholt2004-09-252-6/+3
| | | | setup that now removed atoms from the atomlist on texture deletion.
* Bump DRIVER_DATE for yesterday's changes.Eric Anholt2004-09-251-1/+1
|
* moved to demos/ directoryBrian Paul2004-09-242-577/+0
|
* moved in from the tests/ directory to include in distroBrian Paul2004-09-242-0/+572
|
* updated osmesa infoBrian Paul2004-09-241-12/+9
|
* patches for clearing hw depth buffers from software fallback (Nicolai Haehnle)Brian Paul2004-09-244-10/+117
|
* added (back!) option to use non-packedcolorDaniel Borca2004-09-248-4/+203
|
* -O -Wall warnings cleanups in r200.Eric Anholt2004-09-245-6/+12
|
* Clean up some warnings in the R100 driver with -O -Wall.Eric Anholt2004-09-243-4/+16
|
* Remove an unnecessary calculation of the dest pointer.Eric Anholt2004-09-241-2/+1
|
* don't test for NULL pixels pointer here, do that in the 'store' routines ↵Brian Paul2004-09-231-6/+9
| | | | after validating PBO address
* removed hacked GL tokens, fixed a printfBrian Paul2004-09-231-9/+1
|
* added a comment in BindTextureBrian Paul2004-09-231-0/+4
|
* just some comments and formatting changesBrian Paul2004-09-231-32/+36
|
* fix for Erics new emit state code, rework out the dest pointerDave Airlie2004-09-231-0/+3
| | | | after we check the buffer
* The previous code would emit a full set of state during the first EmitState onEric Anholt2004-09-228-125/+168
| | | | | | | | | | | | | a new cmdbuf, to ensure that state wasn't lost across UNLOCK/LOCK pairs (in the case of context switching). This was rather inefficient. Instead, after flushing a cmdbuf, mark the state as needing to be saved on UNLOCK. Then, at the beginning of flushing a cmdbuf, if we actually have lost the context, go back and emit a new cmdbuf with the full set of state, before continuing with the cmdbuf flush. Also, remove the dirty/clean atom lists, since atoms are emitted in a fixed order these days, and go with a simpler single list. Provides a 14% improvement in ipers performance in my tests, along with other apps.