Age | Commit message (Collapse) | Author |
|
Translate vertex shaders independently of fragment shaders.
Previously tried to make fragment shader semantic indexes always start
at zero and exclude holes. This was unnecessary but meant that vertex
shader translation had to be adjusted to take this into account.
Now use a fixed scheme for labelling special FS input semantics
(color, etc), and another fixed scheme for the generics.
With this, vertex shaders can be translated independently of the bound
fragment shader, assuming mesa has done its own job and ensured that
the vertex shader provides at least the inputs the fragment shader is
looking for. The state-tracker didn't attempt to do anything about
this previously, so it shouldn't be needed now.
|
|
One of several cases where we build mesa shaders and then translate
them to tgsi. Many of those cases it's because we're combining two
mesa programs and there are helpers for that, but in this case at
least can go straight to tgsi.
|
|
|
|
|
|
|
|
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Pekka Paalanen <pq@iki.fi>
|
|
Signed-off-by: Francisco Jerez <currojerez@riseup.net>
Signed-off-by: Pekka Paalanen <pq@iki.fi>
|
|
This should fix TXB on G45 and older in the GLSL case.
|
|
New comments should explain some of the confusion about how this message
works.
|
|
|
|
|
|
Fixing this is a prereq for avoiding flagging all state at new
batch time. Eliminating that still causes problems, though (notably
glean logicOp fails on my GM965).
|
|
|
|
|
|
|
|
Saves CPU time, resulting in a 2.5% FPS win on ETQW.
|
|
|
|
This cuts a massive number of waits in ET:QW, which uses a VBO ringbuffer.
Unfortunately it doesn't BufferData when wrapping back to 0, so we can't
be clever with tracking what's been initialized.
|
|
They're the same regardless of execution width for 8, 4x2, and 16.
|
|
This is a 2.9% (+/-.3%) performance win for my GL demo, which hits MAD
sequences for matrix transforms.
|
|
This shaves more instructions off of the VS of my GL demo, but no
performance difference this time at n=6. This also fixes a regression
that was in this path, which is now piglit's glsl-vs-mov-after-deref.
|
|
|
|
Thanks to Chia-I Wu's changes to the extension function
infrastructure, we no longer have to tell the loader which extensions
the driver might enable. This means that intelInitExtensions will
never be called with a NULL context pointer. Remove all the NULL checks.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
|
|
The cube map array index arg is always present.
|
|
Instead of doing math based on the (broken for VBO && offset != 0)
input->count number, just use the BO size. Fixes assertion failure in ETQW.
|
|
|
|
|
|
|
|
This didn't work for quad/quadstrips at all, and for all other primitive types
it only worked when they were unclipped.
Fix up the former in gs stage (could probably do without these changes and
instead set QuadsFollowProvokingVertexConvention to false), and the rest in
clip stage.
|
|
|
|
|
|
|
|
Two-sided stencil and NPOT textures.
|
|
Previously, we'd load linearly from ParameterValues[0] for the constants,
though ParameterValues[1] may not equal ParameterValues[0] + 4. Additionally,
the STATE_VAL type paramters didn't get updated.
Fixes piglit vp-constant-array-huge.vpfp and ET:QW object locations.
Bug #23226.
|
|
Fixes piglit vp-sge-alias test, and the googleearth ground shader. \o/
Bug #22228
(cherry picked from commit 56ab92bad8f1d05bc22b8a8471d5aeb663f220de)
|
|
Fixes piglit arl.vp.
(cherry picked from commit d52d78b4bcd6d4c0578f972c0b8ebac09e632196)
|
|
|
|
|
|
|
|
|
|
Before, we weren't aggressive enough in checking for the start or end
of render-to-texture. In particular, if only the ctx->ReadBuffer had
texture attachments, we were treating that as a render-to-texture case.
This fixes a regression from commit 75bdbdd90b15c8704d87ca195a364ff6a42edbb1
"intel: Don't validate in a texture image used as a render target."
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|