summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2010-10-06gallivm: Compute lod as integer whenever possible.José Fonseca
More accurate/faster results for PIPE_TEX_MIPFILTER_NEAREST. Less FP <-> SI conversion overall.
2010-10-06gallivm: Only apply min/max_lod when necessary.José Fonseca
2010-10-06gallivm: don't apply zero lod_biasKeith Whitwell
2010-10-06gallivm: Combined ifloor & fract helper.José Fonseca
The only way to ensure we don't do redundant FP <-> SI conversions.
2010-10-06gallivm: Fast implementation of iround(log2(x))José Fonseca
Not tested yet, but should be correct.
2010-10-06gallivm: Use a faster (and less accurate) log2 in lod computation.José Fonseca
2010-10-06gallivm: Take the type signedness in consideration in round/ceil/floor.José Fonseca
2010-10-06i965: Fix up IF/ELSE/ENDIF for gen6.Eric Anholt
The jump delta is now in the part of the instruction where the destination fields used to be, and the src args are ignored (or not, for the new non-predicated IF that we don't use yet).
2010-10-06i965: Gen6 no longer has the IFF instruction; always use IF.Eric Anholt
2010-10-06i965: Add back gen6 headerless FB writes to the new FS backend.Eric Anholt
It's not that hard to detect when we need the header.
2010-10-06r600g: fix dirty state handlingJerome Glisse
Avoid having object ending up in dead list of dirty object. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-06i965: Also do constant propagation for the second operand of CMP.Eric Anholt
We could do the first operand as well by flipping the comparison, but this covered several CMPs in code I was looking at.
2010-10-06i965: Enable the constant propagation code.Eric Anholt
A debug disable had slipped in.
2010-10-06r600g: avoid segfault due to unintialized list pointerJerome Glisse
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-06llvmpipe: Fix sprite coord perspective interpolation of Q.José Fonseca
Q coordinate's coefficients also need to be multiplied by w, otherwise it will have 1/w, causing problems with TXP.
2010-10-06llvmpipe: Fix perspective interpolation for point sprites.José Fonseca
Once a fragment is generated with LP_INTERP_PERSPECTIVE set for an input, it will do a divide by w for that input. Therefore it's not OK to treat LP_INTERP_PERSPECTIVE as LP_INTERP_LINEAR or vice-versa, even if the attribute is known to not vary. A better strategy would be to take the primitive in consideration when generating the fragment shader key, and therefore avoid the per-fragment perspective divide.
2010-10-06llvmpipe: Dump a few missing shader key flags.José Fonseca
2010-10-06llvmpipe: make debug_fs_variant respect variant->nr_samplersKeith Whitwell
2010-10-06retrace: Handle clear_render_target and clear_depth_stencil.José Fonseca
2010-10-06r600g: add evergreen stencil support.Dave Airlie
this sets the stencil up for evergreen properly.
2010-10-05r600g: userspace fence to avoid kernel call for testing bo busy statusJerome Glisse
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-05st/mesa: replace assertion w/ conditional in framebuffer invalidationBrian Paul
https://bugs.freedesktop.org/show_bug.cgi?id=30632 NOTE: this is a candidate for the 7.9 branch.
2010-10-05r600g: simplify block relocationJerome Glisse
Since flush rework there could be only one relocation per register in a block. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-05r600g: use dirty list to track dirty blocksBas Nieuwenhuizen
Got a speed up by tracking the dirty blocks in a seperate list instead of looping through all blocks. This version should work with block that get their dirty state disabled again and I added a dirty check during the flush as some blocks were already dirty.
2010-10-05nv50: fix always true conditional in shader optimizationNicolas Kaiser
2010-10-05r600g: improve bo flushingJerome Glisse
Flush read cache before writting register. Track flushing inside of a same cs and avoid reflushing same bo if not necessary. Allmost properly force flush if bo rendered too and then use as a texture in same cs (missing pipeline flush dunno if it's needed or not). Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-05r600g: store reloc information in bo structureJerome Glisse
Allow fast lookup of relocation information & id which was a CPU time consumming operation. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-05pb: fix numDelayed accountingDave Airlie
we weren't decreasing when removing from the list.
2010-10-05r600g: avoid unneeded bo waitDave Airlie
if we know the bo has gone not busy, no need to add another bo wait thanks to Andre (taiu) on irc for pointing this out.
2010-10-05r600g: drop use_mem_constant.Dave Airlie
since we plan on using dx10 constant buffers everywhere.
2010-10-05r600g: drop mman allocatorDave Airlie
we don't use this since constant buffers are now being used on all gpus.
2010-10-05r600g: add bo busy backoff.Dave Airlie
When we go to do a lot of bos in one draw like constant bufs we need to avoid bouncing off the busy ioctl, this mitigates by backing off on busy bos for a short amount of times.
2010-10-05pb: don't keep checking buffers after first busyDave Airlie
If we assume busy buffers are added to the list in order its unlikely we'd fine one after the first busy one that isn't busy.
2010-10-05r600g: add bo fenced list.Dave Airlie
this just keeps a list of bos submitted together, and uses them to decide bo busy state for the whole group.
2010-10-04swrast: fix choose_depth_texture_level() to respect mipmap filtering stateBrian Paul
NOTE: this is a candidate for the 7.9 branch.
2010-10-05r300g: fix microtiling for 16-bits-per-channel formatsMarek Olšák
These texture formats (like R16G16B16A16_UNORM) were untested until now because st/mesa doesn't use them. I am testing this with a hacked st/mesa here.
2010-10-04i965: Add support for gen6 FB writes to the new FS.Eric Anholt
This uses message headers for now, since we'll need it for MRT. We can cut out the header later.
2010-10-04i965: In disasm, gen6 fb writes don't put msg reg # in destreg_conditionalmod.Eric Anholt
It instead sensibly appears in the src0 slot.
2010-10-04i965: Add initial folding of constants into operand immediate slots.Eric Anholt
We could try to detect this in expression handling and do it proactively there, but it seems like less logic to do it in one optional pass at the end.
2010-10-04i965: Add trivial dead code elimination in the new FS backend.Eric Anholt
The glsl core should be handling most dead code issues for us, but we generate some things in codegen that may not get used, like the 1/w value or pixel deltas. It seems a lot easier this way than trying to work out up front whether we're going to use those values or not.
2010-10-04i965: Be more conservative on live interval calculation.Eric Anholt
This also means that our intervals now highlight dead code.
2010-10-04r600g: Fix SCons build.Vinson Lee
2010-10-04r600g: remove dead label & fix indentationJerome Glisse
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-04r600g: rename radeon_ws_bo to r600_boJerome Glisse
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-04r600g: use r600_bo for relocation argument, simplify codeJerome Glisse
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-04r600g: allow r600_bo to be a sub allocation of a big boJerome Glisse
Add bo offset everywhere needed if r600_bo is ever a sub bo of a bigger bo. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-04r600g: rename radeon_ws_bo to r600_boJerome Glisse
Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-10-04nvfx: Pair os_malloc_aligned() with os_free_aligned().Krzysztof Smiechowicz
From AROS.
2010-10-04r600g: TODO domain managementDave Airlie
no wonder it was slow, the code is deliberately forcing stuff into GTT, we used to have domain management but it seems to have disappeared.
2010-10-04r600g: fix wwarning in bo_map functionDave Airlie
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-15 00:09:47 +0000