| Age | Commit message (Collapse) | Author |
|---|
|
|
|
The goal here is to avoid regressing performance on ir_to_mesa drivers
for fixed function fragment shaders requiring saturates.
|
|
This should save on the overhead of tree-walking and provide a
convenient place to add more instruction lowering in the future.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
The vector operator collects 2, 3, or 4 scalar components into a
vector. Doing this has several advantages. First, it will make
ud-chain tracking for components of vectors much easier. Second, a
later optimization pass could collect scalars into vectors to allow
generation of SWZ instructions (or similar as operands to other
instructions on R200 and i915). It also enables an easy way to
generate IR for SWZ instructions in the ARB_vertex_program assembler.
|
|
This may grow in the near future.
|
|
The operate just like ir_unop_sin and ir_unop_cos except that they
expect their inputs to be limited to the range [-pi, pi]. Several
GPUs require this limited range for their sine and cosine
instructions, so having these as operations (along with a to-be-written
lowering pass) helps this architectures.
These new operations also matche the semantics of the
GL_ARB_fragment_program SCS instruction. Having these as operations
helps in generating GLSL IR directly from assembly fragment programs.
|
|
Condiation moves with a condition of (a < 0), (a > 0), (a <= 0), or (a
>= 0) can be generated with "a" directly as an operand of the CMP
instruction. This doesn't help much now, but it will help with
assembly shaders that use the CMP instruction.
|
|
|
|
|
|
Some C++ header files were included in an extern "C" block. When building with
Clang, this caused the build to fail due to namespace errors. (GCC did not
report any errors.)
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
http://bugs.freedesktop.org/show_bug.cgi?id=31069
|
|
|
|
It can be tested with if (0) replaced with if (1) to force spilling for all
virtual GRFs. Some simple tests work, but large texturing tests fail.
|
|
Previously _LinkedShaders was a compact array of the linked shaders
for each shader stage. Now it is arranged such that each slot,
indexed by the MESA_SHADER_* defines, refers to a specific shader
stage. As a result, some slots will be NULL. This makes things a
little more complex in the linker, but it simplifies things in other
places.
As a side effect _NumLinkedShaders is removed.
NOTE: This may be a candidate for the 7.9 branch. If there are other
patches that get backported to 7.9 that use _LinkedShader, this patch
should be cherry picked also.
|
|
Also, update ir_to_mesa's "1.30 is unsupported" case to "handle" it.
|
|
|
|
This assertion was added in commit f1c1ee11, but it did not notice
that the array is accessed with 'size-1' instead of 'size'. As a
result, the assertion was off by one. This caused failures in at
least glsl-orangebook-ch06-bump.
|
|
|
|
|
|
Notably missing is choice of registers to spill.
|
|
I'm still not pleased with how builtin uniforms are handled, but as
long as we're relying on the prog_statevar stuff this seems about as
good as it'll get.
|
|
|
|
Another optional ARB_imaging subset extension.
|
|
Fixes: glsl-mat-from-int-ctor-03
|
|
It turns out that most people new to this IR are surprised when an
assignment to (say) 3 components on the LHS takes 4 components on the
RHS. It also makes for quite strange IR output:
(assign (constant bool (1)) (x) (var_ref color) (swiz x (var_ref v) ))
(assign (constant bool (1)) (y) (var_ref color) (swiz yy (var_ref v) ))
(assign (constant bool (1)) (z) (var_ref color) (swiz zzz (var_ref v) ))
But even worse, even we get it wrong, as shown by this line of our
current step(float, vec4):
(assign (constant bool (1)) (w)
(var_ref t)
(expression float b2f (expression bool >=
(swiz w (var_ref x))(var_ref edge))))
where we try to assign a float to the writemasked-out x channel and
don't supply anything for the actual w channel we're writing. Drivers
right now just get lucky since ir_to_mesa spams the float value across
all the source channels of a vec4.
Instead, the RHS will now have a number of components equal to the
number of components actually being written. Hopefully this confuses
everyone less, and it also makes codegen for a scalar target simpler.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
Fixes this GCC warning.
warning: missing initializer for member 'statevar_element::array_indexed'
|
|
|
|
|
|
Include compiler.h for ASSERT symbol.
|
|
Currently GLSL IR forbids any vector comparisons, and defines "ir_binop_equal"
and "ir_binop_nequal" to compare all elements and give a single bool.
This is highly unintuitive and prevents generation of optimal Mesa IR.
Hence, first rename "ir_binop_equal" to "ir_binop_all_equal" and
"ir_binop_nequal" to "ir_binop_any_nequal".
Second, readd "ir_binop_equal" and "ir_binop_nequal" with the same semantics
as less, lequal, etc.
Third, allow all comparisons to acts on vectors.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
|
|
|
|
|
|
This increases the chance that GLSL programs will actually work.
Note that continues and returns are not yet lowered, so linking
will just fail if not supported.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
This allows us to specify different options, especially useful for chips
without unified shaders.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
Changes in v2:
- No longer adds tessellation enums
|
|
Also define it if #version 100 is encountered.
|
|
Bug #30040.
|
|
This should only occur if glCompileShader is called without a prior call
to glShaderSource. An empty source program should be the empty string.
|
|
I swear there was some git option for "don't push things when you've
got uncommitted changes", but I can't find it now.
|
|
Like the constant handling and the handling of other uniforms, we add
the whole thing to the Parameters, avoiding messy, incomplete logic
for adding just the elements of a builting uniform that get used.
This means that a driver that relies only on ParameterValues[] for its
parameters will have an increased parameter load, but drivers
generally don't do that (since they have other params they need to
handle, too).
Fixes glsl-fs-statevar-call (testcase for Ember). Bug #29687.
v2: Continue referencing the STATE_VAR[] file directly when the
uniform will land in STATE_VAR[] formatted exactly as we'd put into a
temporary. When there's array dereferencing, we don't copy-propagate
in Mesa IR (not knowing where the array is in register space), so
smarts here are required or we'll massively increase the temp count.
|
|
At some point this actually triggered, not sure if it still does.
Give a meaningful assert and refuse to smash the stack anyway.
|
|
a temp."
This reverts commit 5ad74779cea07cc6a19a52874cdaef8b018e2f1b.
Sorry, but I had to revert this.
Any commit which needlessly increases the number of temporaries is wrong.
More temporaries mean less shader performance because of reduced parallelism
and therefore less efficient latency hiding. In this case, there is possible
performance degradation of every shader which uses GL state variables.
I cannot accept this.
|
|
The downside of our talloc usage is that we can't really make static
(i.e., not created with new) instances of our IR types. This leads to
a lot of unnecessary dynamic allocation in this patch.
|
|
|
|
The CMP instruction needed to be flipped to properly handle
operand==0.
Fixes fd.o bug 29923.
|
|
Like the constant handling and the handling of other uniforms, we add
the whole thing to the Parameters, avoiding messy, incomplete logic
for adding just the elements of a builting uniform that get used.
This means that a driver that relies only on ParameterValues[] for its
parameters will have an increased parameter load, but drivers
generally don't do that (since they have other params they need to
handle, too).
Fixes glsl-fs-statevar-call (testcase for Ember). Bug #29687.
Regresses glsl-vs-array-04 on 965. Thanks to a slight change in
register allocation, this test of undefined behavior now wraps around
the register space and unexpectedly reads the constant value it's
trying to compare to. The test should probably not look at the
resulting color, since behavior is undefined.
|
|
|
|
|
|
Both i965 and swrast rely on UsesKill to determine whether to do early
depth writes. Fixes glsl-fs-discard-02.
Bug #29835.
|
