summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/common
AgeCommit message (Collapse)Author
2005-02-05Fixed a typo in an assertion. (Don't understand how this ever compiled.)Felix Kuehling
Found by Andreas Stenglein.
2005-02-04Added an option texture_heaps that allows selecting which texture heapsFelix Kuehling
will be used. Implemented this option in the Savage driver. On my ProSavageDDR uploads to AGP memory are about 1.5 times as fast as uploads to card memory. On non-IGP hardware the difference may be even bigger. Now mplayer -gl is getting really usable.
2005-02-03Use all texture heaps in a fair way when textures need to be kicked inFelix Kuehling
order to make room for new textures. In particular this fixes texture trashing on the first heap when the second heap is occupied by currently unused textures (observed with Torcs and the Savage driver). Heaps are weighted by their sizes by default but drivers can override these and apply their own weights based on relative texture upload speeds to the respective heaps.
2005-02-02Don't try to kick textures from a NULL tex heap.Felix Kuehling
2005-01-31Bug #2428: #ifdef GLX_DIRECT_RENDERING in DRI drivers is pointless.Adam Jackson
2005-01-26new version check to allow to test for a range of ddx major versions, ↵Roland Scheidegger
instead of just one major version
2005-01-15Use HW_READ_LOCK and HW_READ_UNLOCK in assembler-optimized span readFelix Kuehling
functions.
2005-01-12Simplify usage of drmHash functions and fix bug inKeith Whitwell
__driGarbageCollectDrawables which would get confused while walking the hash values.
2005-01-03Added a boolean option for a floating-point depth buffer.Felix Kuehling
2004-12-14uint*t -> u_int*t changesAlan Hourihane
2004-12-08(Stephane Marchesin, me) add hyperz support to radeon and r200 drivers. Only ↵Roland Scheidegger
fast z clear and z buffer compression are supported for now, hierarchical-z is not. Still problems with multiple apps and z/stencil readback, which is why hyperz is disabled per default. Also add the new point sprite packet drm 1.13 accepts to the sanity code.
2004-12-08Use union type to avoid strict aliasing problems.Alan Hourihane
2004-12-08Check for some header defines before redefining functions. Silences warnings.Alan Hourihane
2004-12-02Fix some warningsAlan Hourihane
2004-11-02Added MMX optimized version of the RGB565 ReadRGBASpan routine.Ian Romanick
2004-11-01Nicolai's sw-clipspan-fixes.patchKeith Whitwell
2004-10-27Bug #1713: Some rare libGL's have __glXFindDRIScreen defined but do notAdam Jackson
export it via glXGetProcAddress. These are not supported anymore, so print an error message to that effect.
2004-10-26As per suggestions my Michel Daenzer, improve the fix for bugzilla #1513.Ian Romanick
2004-10-18Small optimization for big-endian (e.g., PowerPC) systems.Ian Romanick
2004-10-16Only build the MMX/SSE/SSE2 versions for ReadRGBASpan when the pixel formatIan Romanick
matches what they optimized versions actually expect.
2004-10-15Fixed a few places that should have been using GET_SRC_PTR.Ian Romanick
2004-10-14Add support for optimized versions of the code underlying ReadPixelsIan Romanick
(and DrawPixels). The R200, R128, and Unichrome drivers get support in this commit. Other drivers would be easy enough to add for people that have the cards. The DRI (CVS) build will need to be updated to account for the new source files.
2004-10-13Add a license to this file. It was written on VA's watch, so they getKeith Whitwell
the honours.
2004-10-13Initial support for PowerPC specific code in Mesa and DRI drivers. DRIIan Romanick
drivers built on PowerPC systems should now show things like "PowerPC" or "PowerPC/Altivec" in the GL_RENDERER string. The VMX moniker is used for Altivec/Velocity Engine/VMX SIMD additions. I chose this not because I work for IBM but because it's a LOT shorter to type. :)
2004-10-13Refactor the code in driGetRendererString a bit to make it easier toIan Romanick
add support for features of non-x86 CPU types.
2004-10-08Don't compile WriteMonoDepthSpan if HAVE_HW_DEPTH_SPANS is defined -- it won'tEric Anholt
be used in that case, and it wanting WRITE_DEPTH was making r128 die on the undefined symbol.
2004-10-07Add Roland Scheidegger's S3TC patch. This patch does not implement theEric Anholt
(patented) S3TC/DXTC algorithms, but adds an option to dlopen a library module providing functions to do so. Because it uses dlopen, it is only enabled if USE_EXTERNAL_DXTN_LIB=1 is defined (which is only in linux-dri config, so far). It adds support for S3TC to several DRI drivers, and adds a DRI config option to force enabling S3TC even if the software compression/decompression is unavailable. This may allow people to use apps that require S3TC even though they don't have a license to implement the patented material themselves, if those apps use precompressed textures. Ideally we would get permission from the current holder of the patents to implement the algorithm in Mesa, at which point the dlopen mess could go away. Until then, this allows some to run applications they couldn't otherwise, and hopefully will provide us with more push to get the final step of getting that permission done.
2004-09-28use GLuint instead of uintDaniel Borca
2004-09-24patches for clearing hw depth buffers from software fallback (Nicolai Haehnle)Brian Paul
2004-09-07This patch enables GL_ARB_vertex_program and GL_NV_vertex_programDave Airlie
support in the r200 driver. Both extensions can be enabled via options, GL_ARB_vertex_program is on by default, GL_NV_vertex_program off. Option descriptions are in german, english and french. From: Philipp Klaus Krause
2004-08-12Fix dri_util.c build with new headers.Eric Anholt
Submitted by: John Lightsey <john@nixnuts.net>
2004-07-26Fixed hash table allocation to avoid an assertion failure due to aFelix Kuehling
bad config file. Fixed two more typos.
2004-07-10Make GET_PROGRAM_NAME work with uCLibc using the __progname hack.Felix Kuehling
Improved that hack to remove leading directory names.
2004-07-09Fix mesa-solo config to define _GNU_SOURCE so that program_invocation_short_nameJon Smirl
will be correctly defined.
2004-07-04Patch removes _SOLO definition needed for mesa-solo. mesa-soloJon Smirl
uses the NEW_INTERFACE now so _SOLO isn't necessary anymore. Tested with the hardware that I own.
2004-06-29First phase of TLS work. At this point SPARC assembly dispatch isIan Romanick
broken, but it will be fixed "shortly." This is pretty much the same as the patch I sent to the dri-devel list on 22-Jun-2004.
2004-06-26Mesa-solo builds with these changes. There are still more fixups needed toJon Smirl
get individual drivers working again. This converts miniglx to the new dri interface. Thanks to Erdi Chen for the new interface code.
2004-06-19Big-endian texture fixes from Michel Dänzer.Brian Paul
2004-06-15Fix a couple typeos in the tables for driFillInModes.Ian Romanick
2004-06-07Fix up some remaining include issues. Replace Xmalloc / Xfree withIan Romanick
_mesa_malloc / _mesa_free. Replace Bool with GLboolean.
2004-06-06Remove the rest of the need for glxclient.h and remove it from the tree.Ian Romanick
2004-06-03Add a new function, driFillInModes, that is to be used by drivers toIan Romanick
create the set of support __GLcontextModes. This should replace the current driver-specific fill_in_modes functions.
2004-06-03Open/Close FullScreen die. unichrome and savage implemented, code is ifdef'd outJon Smirl
2004-06-02driCheckDriDdxDrmVersion uses a function that is not available toIan Romanick
drivers when DRI_NEW_INTERFACE_ONLY is defined. #ifndef it away in that situation. Add a new function, driCheckDriDdxDrmVersion2, that is passed in the version information that is already supplied to __driCreateNewScreen. Part of the reason that information is supplied to __driCreateNewScreen is so that the driver doesn't have to make those calls to get it! Modify all drivers that support the new interface to use the new function instead of the old. As soon as all drivers support the new interface, driCheckDriDdxDrmVersion can be removed.
2004-06-02Removed need for sarea.h, various touch ups to get rid of type mismatches.Jon Smirl
2004-06-02Require libGL API version 20040602, which actually exportsIan Romanick
__glXGetDrawableInfo. Make sure driCreateNewDrawable sets pdraw->private to NULL for all failure paths.
2004-06-02Make sure USE_NEW_INTERFACE is defined if DRI_NEW_INTERFACE_ONLY isIan Romanick
defined.
2004-06-02Minor changes to make it work better outside X. "Success" is replacedIan Romanick
by explicit 0. Several functions that are not needed with DRI_NEW_INTERFACE_ONLY are blocked with '#ifndef DRI_NEW_INTERFACE_ONLY'. Some further work is still required to make it build without any X include files, though.
2004-06-02Replace use of XID with __DRIid.Ian Romanick
2004-06-02XF86DRIDestroyContext, XF86DRICreateDrawable, andIan Romanick
XF86DRIDestroyDrawable are all called directly from DRI drivers using the new interface. Therefore, prototypes, using available datatypes, must be available in dri_interface.h. Since the prototypes are available there, xf86dri.h is no longer needed for DRI_NEW_INTERFACE_ONLY builds.