| Age | Commit message (Collapse) | Author |
|---|
|
Whereas constant folding evaluates constant expressions at rvalue
nodes, constant propagation tracks constant components of vectors
across execution to replace (possibly swizzled) variable dereferences
with constant values, triggering possible constant folding or reduced
variable liveness.
|
|
This lets us handle arrays much better than trying to work backwards
from assembly.
Fixes fbo-drawbuffers-maxtargets on swrast (i965 needs loop unrolling)
|
|
Support for samplers in general is still incomplete -- anything in a
uniform struct will still be broken. But that doesn't appear to be
any different from master.
Fixes:
glsl-fs-uniform-sampler-array.shader_test
|
|
Fixes glean shaderAPI.
|
|
Fixes a failure in glean shaderAPI.
|
|
Fixes piglit test array_texture.
|
|
Fixes ir_to_mesa handling of unop_log, which used the weird ARB_vp LOG
opcode that doesn't do what we want. This also lets the multiplication
coefficients in there get constant-folded, possibly.
Fixes:
glsl-fs-log
|
|
Optimizations at compile time should generally be done with the goal
of reducing instruction count so that other work, particularly
linking, is less time-consuming if the shader is used multiple times.
However, function inlining increases instruction count for the inlined
function bodies without removing the original function body, since we
don't know if it will be used at link time or not.
Reduces the runtime of linking and executing a Yo Frankie fragment
shader from 0.9 seconds to 0.5 seconds (-45.9%, +/- 2.2%, n=5).
|
|
|
|
|
|
Mixing stderr (_mesa_print_program, _mesa_print_instruction,
_mesa_print_alu) with stdout means that when writing both to a file,
there isn't a consistent ordering between the two.
|
|
While the Mesa IR dumping includes some corresponding GLSL IR for
correlating Mesa IR to GLSL IR, it doesn't completely express it.
This printing includes things like variable declarations and control
flow structure that is hard to read otherwise.
|
|
Replace swizzles on the LHS with additional swizzles on the RHS and a
write mask in the assignment instruction. As part of this add
ir_assignment::set_lhs. Ideally we'd make ir_assignment::lhs private
to prevent erroneous writes, but that would require a lot of code
butchery at this point.
Add ir_assignment constructor that takes an explicit write mask. This
is required for ir_assignment::clone, but it can also be used in other
places. Without this, ir_assignment clones lose their write masks,
and incorrect IR is generated in optimization passes.
Add ir_assignment::whole_variable_written method. This method gets
the variable on the LHS if the whole variable is written or NULL
otherwise. This is different from
ir->lhs->whole_variable_referenced() because the latter has no
knowledge of the write mask stored in the ir_assignment.
Gut all code from ir_to_mesa that handled swizzles on the LHS of
assignments. There is probably some other refactoring that could be
done here, but that can be left for another day.
|
|
Instead of using a linker-assigned location (since samplers don't
actually take up uniform space, being a link-time choice), use the
sampler's varaible pointer as a hash key.
|
|
The new compiler doesn't generate Mesa IR at compile time, and that
compile time code previously wouldn't have reflected the link time
code that actually got used. But do dump the info log of the compile
regardless.
|
|
|
|
Even though the spec says that the limits should be -64/+63, proprietary
drivers support much larger relative offsets and some applications do
depend on this non-standard behavior.
Also program_parse.tab.c has been regenerated.
This fixes the parser error:
ARB_vp: error: relative address offset too large
See also: https://bugs.freedesktop.org/show_bug.cgi?id=28628
4096 * sizeof(vec4) is the maximum size of the constant buffer on NV50.
It is not supposed to be a definite hardware limit, it is for the parser
not to get in the way and let the underlying driver decide whether it can
run the shader or not.
|
|
Fixes:
glsl-uniform-initializer-1
glsl-uniform-initializer-2
glsl-uniform-initializer-3
glsl-uniform-initializer-4
glsl1-GLSL 1.20 uniform array constructor
|
|
|
|
Fixes glsl-uniform-struct.
|
|
This helps makes sure we don't miss any new fields, and makes totally
uninitialized src_regs be PROGRAM_UNDEFINED.
|
|
The symbol_header structure that tracks symbols with a particular name
may have a different (longer) life time than the symbols it tracks.
Not keeping a local copy of the name can lead to use-after-free
errors. For example, the following sequence would trigger such an
error:
char *copy = strdup(name);
_mesa_symbol_table_push_scope(st);
_mesa_symbol_table_add_symbol(st, 0, name, NULL);
_mesa_symbol_table_pop_scope(st);
free(name);
_mesa_symbol_table_find_symbol(st, 0, copy);
With this change, the symbol table keeps a local copy of the name that
has the same life time as the symbol_header for that name. This
resolves some use-after-free errors with built-in functions in the
GLSL compiler.
|
|
This is the option that dumps shader source to files in the current
directory.
|
|
mtypes.h does not use any symbols from compiler.h.
Also add the required headers for files that depended on symbols from
compiler.h but were indirectly including compiler.h through mtypes.h.
|
|
|
|
|
|
|
|
Now drivers, etc. can know which register files are accessed with
indirect addressing. Before we just checked gl_program::NumAddressRegs
but didn't know if that was the constant buffer, temp regs, or what.
The only user of this new field so far will be the gallium state tracker.
|
|
Don't rely on inclusion of other files that already include macros.h.
|
|
It was harmless, but ugly.
|
|
|
|
|
|
|
|
Fixes:
glsl-arb-fragment-coord-conventions
|
|
Fixes:
glsl1-GL state variable reference (diffuse product)
glsl1-GL state variable reference (gl_FrontMaterial.ambient)
glsl1-GL state variable reference (gl_LightSource[0].diffuse)
glsl1-GL state variable reference (point attenuation)
glsl1-GL state variable reference (point size)
glsl1-linear fog
|
|
Fixes:
glsl-uniform-out-of-bounds.
|
|
Fixes:
glsl1-GLSL 1.20 array constructor 1
glsl1-GLSL 1.20 array constructor 2
glsl1-GLSL 1.20 array.length()
glsl1-GLSL 1.20 const array constructor 1
glsl1-GLSL 1.20 const array constructor 2
|
|
This way, we don't need to pass in a parse state, and the context
doesn't grow with the number of passes through optimization.
|
|
This gets glsl-vs-raytrace, glsl-fs-raytrace running on the new
compiler.
|
|
MESA_GLSL=nopt now produces believable output for glsl-fs-raytrace.
|
|
|
|
This cleans up the assembly output of almost all the non-logic tests
glsl-algebraic-*. glsl-algebraic-pow-two needs love (basically,
flattening to a temporary and squaring it).
|
|
void functions have a type of glsl_type::void_type, not a null type.
|
|
|
|
Fixes:
TPPStreamCompiler::assignOperands
|
|
This pulls in multiple i965 driver fixes which will help ensure better
testing coverage during development, and also gets past the conflicts
of the src/mesa/shader -> src/mesa/program move.
Conflicts:
src/mesa/Makefile
src/mesa/main/shaderapi.c
src/mesa/main/shaderobj.h
|
|
Based on the two recent changes to program_lexer.l.
|
|
This avoids two "function defined but not used" warnings. For the yyinput
function we define YY_NO_INPUT which tells flex to simply not generate this
function.
For unput, we add a call to this function, but inside a while(0) so
that it will quiet the warning without actually changing any
functionality.
|
|
Add declarations for two functions generated in the flex ouput. It
would be nicer if flex simply declared these generated functions as
static, but for now we can at least avoid the warning this way.
|
|
|
