summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2011-12-02uxa: Reset size limits based on AGP sizeChris Wilson
The basis for the constraints are what we can map into the aperture for direct writing with the CPU, so use the size of the mappable region as opposed to the size of the total GTT. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-12-01sna: Reuse the full size of an old handle for ioChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-12-01sna: Move the preservation of the io handle into the common destroy pathChris Wilson
In order to capture and reuse all io buffers. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-30sna: Align pwrite to transfer whole cachelinesChris Wilson
Daniel claims that this is will be faster, or will be once he has completed rewriting pwrite! Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-29sna/gen5: Handle cpu-bo for render targetsChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-29sna/render: Fix check for "migrate whole pixmap"Chris Wilson
The whole pixmap means the sample covers the full width and height, not just either! Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-29sna: Fix assertion around flushing of mmap(PROT_READ)Chris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-29sna/gen6: Set the batch mode prior to checking limits and flushingChris Wilson
If we change contexts, then we will submit the batch obsoleting the earlier resource checks. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-28sna: Always reduce tiling for thin pixmapsChris Wilson
Benchmarking on the current code base, says this is now a win. A reversal of older benchmarks, so expect further tuning. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-28sna: Use the correct pitch when creating an untiled active bufferChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-28sna: Pass the pixmap to sna_replace()Chris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-28sna: Tidy kgem_choose_tiling()Chris Wilson
Reduce the calls to compute the surface size down to one. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-28sna: Don't try to guess when the kernel believes a buffer to be flushedChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna: Eliminate empty glyphs during format conversionChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna: Use 64-bit mov (when available) for bitmap glyphsChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna/gen6+: Only use BLT if the untiled bo will cause per-pixel TLB missesChris Wilson
i.e. only force the BLT if using the sampler is going to be incredibly slow. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna: Only check for all-damaged CPU bo if we may discard the GPU boChris Wilson
The goal of the optimisation is to discard the GPU bo early, so we can skip the extra damage reduction if there is no gpu bo. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna: Discard the gpu-bo is the entire pixmap is written to by the CPU.Chris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna: Disabling tiling if the bo would be smaller than a pageChris Wilson
Once again experiment with untiled smalled buffers. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-27sna: Round up stride to alignment for TLB miss checkingChris Wilson
We want to avoid the condition of reducing the tiling mode (when reusing an active untiled buffer in preference to creating a new) for a wide buffer when doing will force a TLB miss on each sample. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-25sna: Avoid the double application of drawable offsets for tiled spansChris Wilson
As we handle tiled spans indirectly, we need to avoid applying the drawable offsets twice (once in the mi layer generating the spans, and then once more in the tiled rect renderer). Reported-by: Ulrich Müller <ulm@gentoo.org> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43245 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-25sna/gen5+: Prefer using the BLT when either src or dst is untiledChris Wilson
The cost of the TLB miss on every sample far outweighs the impact of the context (and ring) switch. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-25sna/gen5: Use the same prefer_blt_fill() switch for fill-oneChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-25sna/gen2: Wire up NO_COMPOSITE for debuggingChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna/video: Check for Xv presence in the server before initialisingChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna: Downgrade Y-tiling if we must use the BLTChris Wilson
There are many operations, usually the core drawing acceleration, where the BLT is much more preferable than using the CPU. However, the BLT is limited to only using X-tiling, so if we encounter a Y-tiled pixmap target we need to recreate it as X-tiling before proceeding. Hopefully, the pixmap is then kept around and rendered multiple times to amoritize the cost of the copy. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna: Use Y-tiling for source pixmapsChris Wilson
Y-tiling is slightly faster with RENDER operations, so attempt to allocate source-only pixmaps using this tiling mode. Actually using Y-tiling is a delicate balance because it then prevents the use of the BLT. For instance, enabling Y-tiling by default gives a 30% performance improvement on the fish-demo (compositing benchmark) at 2560x1440 on Ironlake but regresses tiger-demo by 2x (spans benchmark). So experiment with this compromise and allow for changing the default tiling. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24src/sna/gen5: Replace the precompiled shadersChris Wilson
Take advantage of a couple of new instructions introduced with Cantiga to reduce the instruction count inside the shaders and improve performance by around 10% in the fish-demo. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna/gen5: Remove the unused SIPChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna: Tweak cancellation of deferred flushChris Wilson
The goal is to keep running until the tick after every stops, irrespective of forced flushes. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna: Always call retire following a mmapChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna/gen5: Use the BLT rather than flush when copyingChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-24sna/gen4,gen5: Fix typo and only emit a flush if the video source is dirtyChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-21uxa: Make the glamor/uxa transition more verboseChris Wilson
And so hopefully make it clearer. In the process we restore the flushing behaviour for UXA back to before the glamor intervention. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-21sna: Correct end-point adjustment for stippling PolySegmentChris Wilson
Reported-by: Roman Jarosz <kedgedev@gmail.com> Reported-by: da_fox@mad.scientist.com Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43134 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-21sna/gen4: Use the special composite rect emission for fill_one()Chris Wilson
In order to workaround a bug in the shaders on gen4, we need to flush the pipeline after every rectangle. The recently introduced fill-one mechanism for gen4, missed this vital step triggering a random hang with an otherwise sane batchbuffer (the missing flush is hard to spot!). Fixes regression from 86f99379ee5 (sna/gen4: Add fill-one). Reported-by: Albert Damen <albrt@gmx.net> Reported-by: Fryderyk Dziarmagowski <fdziarmagowski@gmail.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43083 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-19sna: Micro-optimise sna_push_pixels_solid_blt()Chris Wilson
Unroll the byte reversal as we know the row length is word aligned. Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-19sna: Avoid the NULL damage derefChris Wilson
Fixes x11perf -wdcircle100 -time 1 -repeat 1 -rop GXxor Reported-by: Fryderyk Dziarmagowski <fdziarmagowski@gmail.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=43084 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-19sna/gen7: Apply recent SNB tuning as applicableChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-19sna/gen7: Correct shifts for surface stateChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-18sna/gen7: minor tidy of redundant definesChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-17glamor: Include glamor cflags and libs for atypical buildsChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-17uxa/glamor: Turn on glamor for fill spans and rectsZhigang Gong
This commit only enables two glamor functions for uxa_fill_spans and uxa_poly_fill_rects. Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-17glamor: Initial commit to introduce glamor acceleration.Zhigang Gong
Added one configuration option --enable-glamor to control whether use glamor. Added one new file intel_glamor.c to wrap glamor egl API for intel driver's usage. This commit doesn't really change the driver's control path. It just adds necessary files for glamor and change some configuration. Reviewed-by: Eugeni Dodonov <eugeni.dodonov@intel.com> Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-16uxa/glamor: Introduce additional access modes to wrap glamor accelerationZhigang Gong
Integrate glamor acceleration into UXA framework. Add necessary flushing at the following points: 1. Flush UXA batch buffer before call into glamor. 2. Flush GL operations after return from a glamor function. 3. The point we need to flush UXA batch buffer, we also need to flush GL operations, for example, in intel_flush_callback and couple of places in intel_display.c. Signed-off-by: Zhigang Gong <zhigang.gong@linux.intel.com> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-16Revert "Disable adding normal RTF modes for an eDP"Chris Wilson
This reverts commit 212fa9868767637e8f430485eeb522c99e63fd16. The underlying register programming for eDP is now believed to be fixed as of linux-3.1. References: https://bugs.freedesktop.org/show_bug.cgi?id=38012 Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41070 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-16configure: Version bump for 2.17.0 release2.17.0Chris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-16NEWS: Release notes for 2.17.0Chris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-16sna: The block handler is passed an indirect pointer to the timevalChris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
2011-11-16sna: Correct dependencies for DRI2Chris Wilson
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-10 14:16:54 +0000