summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers
Commit message (Collapse)AuthorAgeFilesLines
* Revert "r600g: Don't negate result of ABS instruction"Dave Airlie2011-02-281-2/+0
| | | | | | This reverts commit b6d40213935da702570eca2c0861bd4b1d7f5254. This actually breaks gears here on my rv670.
* r600g: Process TRUNC with tgis_op2Fabian Bieler2011-02-281-2/+2
| | | | | | TRUNC is neither a scalar instruction nor exclusive to the Trans unit. Signed-off-by: Dave Airlie <[email protected]>
* r600g: Don't negate result of ABS instructionFabian Bieler2011-02-281-0/+2
| | | | Signed-off-by: Dave Airlie <[email protected]>
* i915g: implement cache flushingDaniel Vetter2011-02-276-8/+62
| | | | | | | | | | | With an extremely dumb strategy. But it's the same i915c employs. Also improve the hw_atom code slightly by statically specifying the required batch space. For extremely variably stuff (shaders, constants) it would probably be better to add a new parameter to the hw_atom->validate function. Signed-off-by: Daniel Vetter <[email protected]>
* i915g: buffer validation for blitterDaniel Vetter2011-02-271-0/+11
| | | | Signed-off-by: Daniel Vetter <[email protected]>
* i915g: buffer validation for render stateDaniel Vetter2011-02-273-0/+87
| | | | | | | | Also contains the first few bits for hw state atoms. v2: Implement suggestion by Jakob Bornecrantz. Signed-off-by: Daniel Vetter <[email protected]>
* i915g/winsys: buffer validation supportDaniel Vetter2011-02-271-0/+12
| | | | | | | v2: Add the batch bo to the libdrm validation lost, for otherwise libdrm won't take previously used buffers into account. Signed-off-by: Daniel Vetter <[email protected]>
* i915g: cleanup static state calculation, part 2Daniel Vetter2011-02-273-48/+48
| | | | | | Now also for the DRAW_RECT command Signed-off-by: Daniel Vetter <[email protected]>
* i915g: cleanup static state calculation, part 1Daniel Vetter2011-02-273-89/+103
| | | | | | | Move it to i915_state_static.c This way i915_emit_state.c only emits state and doesn't (re)calculate it. Signed-off-by: Daniel Vetter <[email protected]>
* i915g: make dynamic state emission actually lazyDaniel Vetter2011-02-261-1/+1
| | | | | | Premature semicolon. Signed-off-by: Daniel Vetter <[email protected]>
* i915g: Handle null constants properlyJakob Bornecrantz2011-02-261-3/+6
|
* i915g: fix null deref in draw_rect emissionDaniel Vetter2011-02-261-4/+8
| | | | Signed-off-by: Daniel Vetter <[email protected]>
* i915g: simplify math in constants emissionDaniel Vetter2011-02-261-1/+1
| | | | | | The old code even falls apart for nr == 0 (which is caught earlier, but)! Signed-off-by: Daniel Vetter <[email protected]>
* Check for out of memory when creating fenceAlan Hourihane2011-02-261-0/+3
|
* r600g: explicity set sign bits for RGTCDave Airlie2011-02-251-2/+4
|
* r600g: bc 4/5 or rgtc textures need to be tiled as well.Dave Airlie2011-02-252-10/+10
| | | | | | Make the s3tc upload code more generic. Signed-off-by: Dave Airlie <[email protected]>
* r300g: explicit sign bits on RGTC texturesDave Airlie2011-02-251-2/+4
|
* nvc0: fix PointCoord enable in FP headerChristoph Bumiller2011-02-241-2/+5
|
* nvc0: change TGSI CMP translation to use slctChristoph Bumiller2011-02-243-8/+15
| | | | Saves us the explicit compare instruction needed with selp.
* nvc0: sprite coord enable is per GENERIC, not overall indexChristoph Bumiller2011-02-241-5/+3
|
* nvc0: fix new_value calls using type instead of sizeChristoph Bumiller2011-02-241-3/+3
|
* nvc0: set local memory usage info in shader headerChristoph Bumiller2011-02-246-3/+34
| | | | Before this, l[] access was a no-op.
* nvc0: don't fold loads from local memoryChristoph Bumiller2011-02-241-0/+5
|
* nvc0: presin and preex2 can load from const spaceChristoph Bumiller2011-02-241-2/+2
|
* nvc0: kick out empty live rangesChristoph Bumiller2011-02-241-0/+3
| | | | They affect overlap tests even though they're actually empty.
* nvc0: preemptively insert branch at ENDIFChristoph Bumiller2011-02-242-1/+9
| | | | | | | Might be necessary if a block sneaks in somewhere, like a common block for moves of phi sources after a loop break. This is harmless and normally will be removed before emission.
* nvc0: correct allocation of constrained registersChristoph Bumiller2011-02-241-67/+154
| | | | | | | In linear scan we can't allocate multiple values with different live ranges at the same time to assign them consecutive regs. Maybe we should just switch to graph coloring for all values ...
* nvc0: sync textures with render targets ourselvesChristoph Bumiller2011-02-246-6/+35
| | | | Fixes for example piglit/fbo-flushing and nexuiz' bloom effect.
* nvc0: improve userspace fencingChristoph Bumiller2011-02-246-26/+46
| | | | | | Before, there were situations in which we never checked the fences for completion (some loading screens for example) and thus never released memory.
* nvc0: values for undefined outputs must have file GPRChristoph Bumiller2011-02-241-7/+4
|
* nvc0: multiply polygon offset units by 2Christoph Bumiller2011-02-241-1/+1
| | | | | Wasn't sure if this still was necessary because the piglit test started to fail at some point on nv50 where we already do this.
* nvc0: fix SSGChristoph Bumiller2011-02-241-5/+5
|
* nvc0: don't visit target blocks of a loop break multiple timesChristoph Bumiller2011-02-241-1/+4
|
* nvc0: don't overwrite phi sources at the end of a loopChristoph Bumiller2011-02-241-1/+5
| | | | Except the reference to its own result.
* svga: Ensure rendertargets and textures are always rebound at every command ↵José Fonseca2011-02-244-9/+138
| | | | | | | | buffer start. The svga_update_state() mechanism is inadequate as it will always end up flushing the primitives before processing the SVGA_NEW_COMMAND_BUFFER dirty state flag.
* r600g: EXT_texture_array support.Dave Airlie2011-02-246-9/+39
| | | | | | | | | | | This adds EXT_texture_array support to r600g, it passes the piglit array-texture test but I suspect may not be complete. It currently requires a kernel patch to fix the CS checker to allow these, so you need to use R600_ARRAY_TEXTURE=true for now to enable them. Signed-off-by: Dave Airlie <[email protected]>
* i915g: Lazy emit dynamic stateJakob Bornecrantz2011-02-245-40/+36
|
* i915g: Lazy emit immediate stateJakob Bornecrantz2011-02-245-55/+59
|
* i915g: Disable LIS7 state updates for nowJakob Bornecrantz2011-02-242-1/+5
|
* i915g: Clean up in i915_state_immediateJakob Bornecrantz2011-02-241-5/+1
|
* i915g: Remove outdated commentJakob Bornecrantz2011-02-241-8/+0
|
* i915g: Enable mirror repeat wrap modeJakob Bornecrantz2011-02-243-6/+4
|
* i915g: Always set vbo to flush on flushesJakob Bornecrantz2011-02-241-1/+1
| | | | Reported-by Chris Wilson <[email protected]>
* r600g: Request DWORD aligned vertex buffers.Fabian Bieler2011-02-231-1/+1
| | | | | | The spec says that the offsets in the vertex-fetch instructions need to be byte-aligned and makes no specification with regard to the required alignment of the offset and stride in the vertex resource constant register. However, testing indicates that all three values need to be DWORD aligned.
* r300g: fix missing initializers warningBrian Paul2011-02-221-2/+4
|
* i915g: remove extra semicolonsBrian Paul2011-02-221-3/+3
|
* i915g: s/bool/boolean/ style-fixup in winsysDaniel Vetter2011-02-211-1/+2
| | | | Signed-off-by: Daniel Vetter <[email protected]>
* i915g: Add option to lie about capsJakob Bornecrantz2011-02-213-1/+9
|
* i915g: Move debug fields to screenJakob Bornecrantz2011-02-214-4/+7
|
* i915g: Use debug get once optionsJakob Bornecrantz2011-02-212-3/+9
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-22 13:29:21 +0000