summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri/savage
AgeCommit message (Collapse)Author
2005-04-21Fixed TexEnv modes GL_BLEND and GL_ADD on Savage4-based cards. DisabledFelix Kuehling
the use of hardware texture format I8 as it produces garbage at least on ProSavageDDR.
2005-03-24Fixed segfault due to an off-by-one error in SAVAGE_TEX_MAXLEVELS.Felix Kuehling
2005-03-23Initialize texture addresses to something valid.Felix Kuehling
2005-03-14When timestamping textures, need to emit the event with SAVAGE_WAIT_3DFelix Kuehling
set. Otherwise the 3D engine may still be using a texture that is being overwritten. In order to minimize the cost of waiting, timestamp textures only when needed: when a texture image changes, when a different texture is bound to a texture unit or when a texture unit is disabled. (Used to be after flushing every command buffer.)
2005-03-13Updated savage_dri.h from Xorg CVS. This removes definitions that areFelix Kuehling
not part of the DDX->DRI interface and should never have been in this file in the first place.
2005-03-12Removed savagedma.[ch]. They have been unused for a while and were onlyFelix Kuehling
kept around for reference. Now that command DMA is implemented in the DRM module they are no longer needed.
2005-03-10- Flush before updating the timestamp of a texture heap from a bound texture.Felix Kuehling
- Flush before using a new texture or before disabling a texture unit, because savageFlushCmdBuf can only update the timestamp of the last used texture. This fixes corruption in quake2 with single-textured lighting.
2005-03-09If the texture image changed force emitting the texture address, evenFelix Kuehling
if it didn't change. This seems to flush texture cashes. Fixes multi-textured lighting in quake2.
2005-03-06Don't synchronize after each frame. Allow the GPU to be one frameFelix Kuehling
ahead of the CPU for more parallelism of CPU and GPU. Unfortunately there seems to be some broken hardware (like my ProSavageDDR) on which status register updates are delayed sometimes. This leads to very jerky animation if the hardware can buffer more than the current frame. A new option "sync_frames" can be used as a remedy on such broken hardware.
2005-03-06- Fixed two bugs related to DMA buffer handling that showed up with theFelix Kuehling
new DRM version 2.4 and command DMA. - Flush less. - Bumped the driver date.
2005-02-23Don't forget to update a heap's timestamp when a texture image isFelix Kuehling
changed. Other drivers don't need to do this because they're swapping modified textures out of texture memory, which implies a timestamp update.
2005-02-23With Savage DRM version 2.3.x or later use event counters for textureFelix Kuehling
heap aging, similar to the way it's done in the i810 and i855 drivers. This avoids idling the engine on every texture upload.
2005-02-12Fall back properly when GL_COLOR_LOGIC_OP is enabled and != GL_COPY.Felix Kuehling
Add missing per-primitive fallback for polygon stipple.
2005-02-12Offset LOD bias by about 0.3 to match software rendering more closely.Felix Kuehling
Another small tweak to subtile upload (mostly cosmetic).
2005-02-12Make glean's masked clear test pass.Felix Kuehling
2005-02-12Savage3D-based chips seem so use a constant tile stride of 2048 forFelix Kuehling
vertically incomplete tiles, but only if the color depth is 32bpp. Nobody said this was supposed to be logical!
2005-02-09Fixed stride of incomplete tiles.Felix Kuehling
2005-02-05Disable vertex DMA on SuperSavages.Felix Kuehling
2005-02-05- Fixed scissor rectangle not moving with the window (Xorg bug #1731).Felix Kuehling
- Flush buffered primitives before changing scissors. - Require Savage DRM version 2.1.0, so that the broken single-cliprect code can finally rest in peace. - Removed some more dead code.
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.
2005-02-04Made subtile upload a template. This speeds up tiled texture uploads byFelix Kuehling
about a factor 1.3.
2005-02-02Check for heap == NULL before driSetTextureSwapCounterLocation.Felix Kuehling
2005-02-02Fixed flushing of client-side vertex buffers.Felix Kuehling
2005-02-02Fixed some debug output that was unconditionally enabled.Felix Kuehling
2005-01-31Bug #2428: #ifdef GLX_DIRECT_RENDERING in DRI drivers is pointless.Adam Jackson
2005-01-29* Fixed uploading of textures of certain sizes.Felix Kuehling
* When glTexSubImageND is used, track the set of changed tiles in a bit vector and upload only dirty tiles later. This should improve the performance of dynamic light maps and gl movie player plugins. * Renamed debug item "lru" to "tex". Indicate which levels are uploaded completely or partially.
2005-01-24Converted the Savage texture management to use Ian's common texmem code.Felix Kuehling
2005-01-23Fixed initialization of draw destination to front buffer on singleFelix Kuehling
buffered visuals. Also don't change span draw/read buffers in savageDDDrawBuffer. Now glean's polygon offset test works (and PASSes with fixed point depth buffer) with single buffered visuals without winding up in an infinite loop.
2005-01-23My last attempt to fix polygon offsets with the reversed viewport depthFelix Kuehling
range used by the savage driver by negating ctx->MRD broke polygon offsets with software fallbacks. This one adds a REVERSE_DEPTH parameter to t_dd_tritmp.h (defaults to 0) that allows reversing polygon offsets for hardware rendering but not for software fallbacks. For software fallbacks depth values are reversed after polygon offsets have been applied by the depth span functions.
2005-01-22Fixed polygon offset with reversed viewport depth range.Felix Kuehling
2005-01-21Small cleanup:Felix Kuehling
* Remove some unused (mostly empty) functions * Added context parameter to WAIT_IDLE_EMPTY[_LOCKED] for consistency * Added debug messages to WAIT_IDLE_EMPTY[_LOCKED] * Don't flush empty command buffers
2005-01-20* Added options for disabling the fast path (render stage) and vertex DMAFelix Kuehling
* Fixed disabling of the render stage * Added debug output for per-primitive fallbacks * Bumped driver date
2005-01-20Added support for ELTS to the _savage_render_stage. Requires at leastFelix Kuehling
Savage DRM version 2.2.0. Otherwise the render stage is disabled.
2005-01-16Added a fast path for emitting unclipped primitives directly to aFelix Kuehling
vertex buffer. ELTS are not supported yet (missing functionality in the DRM). You need at least Savage DRM version 2.1.3, which fixes a bug that screwed up triangle fans and strips. Moved the texture normalization stage to savagerender.c.
2005-01-15Converted to use spantmp2.h.Felix Kuehling
2005-01-13Make sure DMA buffers are released in savageDestroyContext. UsuallyFelix Kuehling
the DRM would reclaim DMA buffers, but if one process creates and destroys many contexts it can still run out of DMA buffers.
2005-01-10Fixes to cope with the lack of AGP textures. This is all that's neededFelix Kuehling
to support PCI Savages on the Mesa side. Bumped driver date.
2005-01-08* Flush and wait in per-primitive fallback functions. Fixes flickeringFelix Kuehling
stars in glplanet. * Refactored vertex format choosing code. Improved the Savage4 version to choose a format suitable for DMA (size = 32bytes) whenever possible.
2005-01-08Added state debugging option (SAVAGE_DEBUG=state).Felix Kuehling
Fixed: Use savageEmitChangedRegs instead of savageEmitContiguousRegs for emitting Savage4.
2005-01-07Fixed: ptex drawing functions were overridden with standard drawingFelix Kuehling
functions if ANY_RASTER_FLAGS|ANY_FALLBACK_FLAGS were set.
2005-01-05* Fixed handling of scissorsFelix Kuehling
* Only set scissor regs directly if drmMinor < 1 * Don't set texaddr to 0 when a texture unit is disabled. That would trigger the tightened texture state check in the DRM if the texaddr and texdesc registers were not emitted atomically.
2005-01-03Added support for floating point depth buffers on Savage4-basedFelix Kuehling
hardware. By also reversing the depth range this can compensate the loss of accuracy of far objects caused by the projective transformation. Software fallbacks work but are slightly slower since floats in a custom (non IEEE) format have to be encoded and decoded. I havn't done anything about polygon offsets yet. There doesn't seem to be an easy way do get it right except making the offset unit as big as the lowest resolution of depth values. For now float depth is disabled by default but can be enabled through driconf (though I have seen only positive effects so far).
2005-01-02Improved the performance of software fallbacks by not waiting for idleFelix Kuehling
in every single span function. Instead flush and wait in the SpanRenderStart hook and in wrappers around _swrast_Copy/Draw/ReadPixels. Misc. cleanups in savagespan.c while I'm there.
2005-01-01Removed all direct hardware access (MMIO, BCI) from the Savage DRIFelix Kuehling
driver. It uses the new DRM version 2.0.x now, which has just been committed to DRM CVS.
2004-12-23Simplified and optimized _savage_texnorm_stage.Felix Kuehling
2004-12-22Enabled hardware rendering of 1D textures. No need for a software fallback.Felix Kuehling
2004-12-22- Fake projective textures on a single texture unit. A fallback is onlyFelix Kuehling
needed if a second texture unit is enabled. - Also worked around an application bug in Chromium B.S.U.: it sends 3D texture coordinates while only a 2D texture is enabled. This used to trigger a PTEX fallback. Now the 3rd coordinate is just ignored. - Fixed the _savage_texnorm_stage to never normalize homogenous texture coordinates.
2004-12-17Made debugging output controllable via environment variableFelix Kuehling
SAVAGE_DEBUG. Added fallback debugs. Added no_rast option to disable hardware rasterization (everything as software fallback).
2004-12-17Added a TNL pipeline stage that normalizes texture coordinates as aFelix Kuehling
workaround for bad Savage hardware interpolation of big texture coordinates.
2004-12-16Also PUBLICize the config option tableAdam Jackson