android-x86-mesa.git - Androïd/x86 port of Mesa drivers

Age	Commit message (Collapse)	Author
2010-09-15	nv50: put low limit on REG_ALLOC_TEMP and FP_RESULT_COUNT	Christoph Bumiller

2010-09-15	nv50: improve and fix modifier folding optimization	Christoph Bumiller
	Execute before folding loads, because we don't check if it's legal in lower_mods. Ensure that a value's insn pointer is updated when transferring it to a different instruction.
2010-09-15	nv50: consider address register in reload elimination	Christoph Bumiller

2010-09-13	nv50: fix TXP depth comparison value	Christoph Bumiller

2010-09-13	nv50: fix indirect CONST access with large or negative offsets	Christoph Bumiller

2010-09-13	nv50: MOV TEMP[0], -CONST[0] must be float32 negation	Christoph Bumiller

2010-09-13	nv50: interp cannot write flags reg	Christoph Bumiller

2010-09-13	nv50: check for immediates when turning MUL ADD into MAD	Christoph Bumiller

2010-09-13	nv50: handle TGSI EXP and LOG again	Christoph Bumiller

2010-09-12	nv50: match TEMP limit with nv50 ir builder	Christoph Bumiller
	Mesa doesn't respect it anyway, but this makes it assert rather than threads access areas of l[] that don't belong to them.
2010-09-12	nv50: newlines in shader bincode printing	Christoph Bumiller

2010-09-12	nv50: cannot move from local mem to output reg directly	Christoph Bumiller

2010-09-12	nv50: fix size of outputs_written array	Xavier Chantry

2010-09-12	nv50: minor compiler fixes and cleanups	Christoph Bumiller

2010-09-12	nv50: reduce bb_reachable_by runtime from pot to linear	Christoph Bumiller
	As a by-product, remove the memory leak of nv_basic_blocks.
2010-09-09	nv50: fix can_load check for 3rd source	Christoph Bumiller

2010-09-09	nv50: address regs are 16 bit	Christoph Bumiller

2010-09-09	nv50: duplicate interps in load_proj_tex_coords	Christoph Bumiller
	Otherwise we might clobber the origin interpolation result or use the result of the RCP before its definition.
2010-09-09	nv50: create value references with the right type	Christoph Bumiller
	Since atm our OPs aren't typed but instead values are, we need to take care if they're used as different types (e.g. a load makes a value u32 by default). Maybe this should be changed (also to match TGSI), but it should work as well if done properly.
2010-09-09	nv50: use actual loads/stores if TEMPs are accessed indirectly	Christoph Bumiller

2010-09-09	nv50: don't parse again in tgsi_2_nc	Christoph Bumiller

2010-09-09	nv50: prepare for having multiple functions	Christoph Bumiller
	At some point we'll want to support real subroutines instead of just inlining them into the main shader. Since recursive calls are forbidden, we can just save all used registers to a fixed local memory region and restore them on a return, no need for a stack pointer.
2010-09-09	nv50: save tgsi instructions	Christoph Bumiller

2010-09-03	nv50: load address register before using it, not after	Christoph Bumiller

2010-09-02	Merge remote branch 'origin/master' into nv50-compiler	Christoph Bumiller
	Conflicts: src/gallium/drivers/nv50/nv50_program.c
2010-09-02	nv50: fix build-predicate function	Christoph Bumiller

2010-09-02	nv50: fix find_dom_frontier	Christoph Bumiller

2010-09-02	r600g: fix memory/bo leak	Jerome Glisse
	Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-09-02	mesa: fix code generation for ir_unop_sqrt	Brian Paul
	The CMP instruction needed to be flipped to properly handle operand==0. Fixes fd.o bug 29923.
2010-09-02	mesa: fix some printf warnings with casts	Brian Paul

2010-09-02	r600g: fix thinko in shadow code.	Dave Airlie
	spotted by taiu on irc
2010-09-02	r600g: fix logicop, the 3d ROP is the 2D rop shifted twice.	Dave Airlie

2010-09-02	r600g: fix depth texture tests	Dave Airlie

2010-09-01	glsl: Apply implicit conversions to structure constructor parameters.	Kenneth Graunke
	The code for handling implicit conversions should probably get refactored, but for now, this is easy. Fixes piglit test constructor-26.vert.
2010-09-01	glsl: Convert constant record constructor parameters to ir_constants.	Kenneth Graunke
	I'm not sure if this is strictly necessary, but it seems wise.
2010-09-01	glsl: Reject structure constructors that have too many arguments.	Kenneth Graunke
	Fixes piglit test constructor-27.vert.
2010-09-01	glsl2: Remove unnecessary glsl_symbol_table::get_function parameter ↵	Ian Romanick
	return_constructors Now that constructors are not generated as functions or stored in the symbol table, there is no need to flag whether or not constructors should be returned.
2010-09-01	glsl2: Remove unused method glsl_type::generate_constructor	Ian Romanick

2010-09-01	glsl2: Remove unused 'constructor' parameter from glsl_symbol_table::add_type	Ian Romanick

2010-09-01	glsl2: Don't generate constructor functions for structures	Ian Romanick

2010-09-01	glsl2: Emit structure constructors inline	Ian Romanick
	Fixes piglit test cases glsl-[fv]s-all-0[12].
2010-09-02	r600g: add missing vertex fetch formats to the translation table.	Dave Airlie
	fixes at least 2 more piglits.
2010-09-01	r600g: fix binding of same texture to several target slot	Jerome Glisse
	One can bind same texture or sampler to different slot, each slot needs it own state. The solution implemented here is not exactly beautifull or optimal need to think to somethings better. Signed-off-by: Jerome Glisse <jglisse@redhat.com>
2010-09-01	glsl: Fix write mask in matrix-from-matrix constructors.	Kenneth Graunke
	If the matrix being constructed was larger than the source matrix, it would overwrite the lower-right part of the matrix with the wrong values, rather than leaving it as the identity matrix. For example, constructing a mat4 from a mat2 should only use a writemask of "xy" when copying from the source, but was using "xyzw". Fixes the code generated by piglit test constructor-23.vert.
2010-09-01	glsl: Add proper handling for constant matrix-from-matrix constructors.	Kenneth Graunke
	Fixes piglit test case constructor-21.vert and changes constructor-22.vert to give the correct output.
2010-09-01	glsl: Move generate_constructor_(matrix\|vector) to ir_constant ctor.	Kenneth Graunke

2010-09-01	ast_function: Fix check for "too few components".	Kenneth Graunke
	This was triggering even for matrix-from-matrix constructors. It is perfectly legal to construct a mat3 from a mat2 - the rest will be filled in by the identity matrix. Changes piglit test constructor-23.vert from FAIL to PASS, but the generated code is incorrect.
2010-09-01	ast_function: Remove bogus cases from generate_constructor_matrix.	Kenneth Graunke
	There are no integer matrix types, so switching on them is silly.
2010-09-02	r600g: fix incorrect state naming in pipe_sampler vs pipe_sampler_view	Dave Airlie
	fixes problems in valgrind with uninitialised values.
2010-09-01	ir_to_mesa: Load all the STATE_VAR elements of a builtin uniform to a temp.	Eric Anholt
	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.