summaryrefslogtreecommitdiffstats
path: root/src
Commit message (Collapse)AuthorAgeFilesLines
* gallivm: (trivial) fix linear aos sampling of 3d compressed formatsRoland Scheidegger2013-08-221-2/+2
| | | | | | | | block size depth is always 1 even for compressed formats (unless someone invents true 3d compressed formats at least which we can't represent). Nearest (and soa) path had it right. Reviewed-by: Jose Fonseca <[email protected]>
* radeonsi: Fix y/z/w component values of TGSI_SEMANTIC_FOG pixel shader inputsMichel Dänzer2013-08-221-0/+18
| | | | | | | | | They are defined as constant 0.0/0.0/1.0. Three more little piglits. Cc: [email protected] Reviewed-by: Alex Deucher <[email protected]>
* gallium: Support PIPE_FORMAT_R10G10B10A2_UINT.José Fonseca2013-08-223-0/+3
| | | | | | Same as PIPE_FORMAT_B10G10R10A2_UINT but without the swizzling. Reviewed-by: Roland Scheidegger <[email protected]>
* trace: Handle null tokens.José Fonseca2013-08-221-2/+6
| | | | Used for example on stream out without geometry shader.
* ilo: do not need last shader stage for 3DSTATE_SBEChia-I Wu2013-08-224-14/+6
| | | | | | We have set up 3DSTATE_SBE (or 3DSTATE_SF on GEN6) in ilo_shader_select_kernel_routing(). There is no need to pass the last shader stage to the GPE function.
* ilo: fix a potential issue with STATE_SIPChia-I Wu2013-08-221-2/+2
| | | | | Command length is ORed to the wrong place. Since the ORed value is zero, there is no real change.
* ilo: add GEN check to 3DSTATE_CLIPChia-I Wu2013-08-221-0/+2
| | | | Assert that gen6_emit_3DSTATE_CLIP is for GEN 6 and 7.
* build: Add --enable-gallium-osmesa flag.Matt Turner2013-08-212-2/+2
| | | | | | | | The Gallium implementation is apparently not ready for regular consumption, so as much as I hate adding more build-time options, here's another. Acked-by: Brian Paul <[email protected]>
* glsl: Give a warning, not an error, for UBO qualifiers on non-matrices.Ian Romanick2013-08-211-6/+12
| | | | | | Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=59648 Reviewed-by: Matt Turner <[email protected]> Signed-off-by: Ian Romanick <[email protected]>
* glsl: Remove ubo_qualifiers_allowed variable.Matt Turner2013-08-214-11/+2
| | | | | | | No longer used. Reviewed-by: Ian Romanick <[email protected]> Reviewed-by: Anuj Phogat <[email protected]>
* glsl: Drop duplicate error messages.Matt Turner2013-08-211-12/+1
| | | | | | | | This same message is printed in the validate_matrix_layout_for_type function. Reviewed-by: Ian Romanick <[email protected]> Reviewed-by: Anuj Phogat <[email protected]>
* glsl: Rename ubo_qualifiers_valid to ubo_qualifiers_allowed.Matt Turner2013-08-214-6/+6
| | | | | | | | | The variable means that UBO qualifiers are allowed in a particular context (e.g., not allowed in a struct field declaration), rather than a particular set of UBO qualifiers are valid. Reviewed-by: Ian Romanick <[email protected]> Reviewed-by: Anuj Phogat <[email protected]>
* i965/fs: Add code to print out global copy propagation sets.Kenneth Graunke2013-08-211-0/+30
| | | | | | | | | This was invaluable when debugging the global copy propagation algorithm. We may as well commit it in case someone needs to print out the sets in the future. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Paul Berry <[email protected]>
* osmesa: Symlink shared library to LIB_DIRArmin K2013-08-211-0/+1
| | | | | | Cc: 9.2 <[email protected]> Tested-by: Brian Paul <brianp at vmware.com> Reviewed-by: Brian Paul <brianp at vmware.com>
* svga: minor clean-ups in emit_hw_vs_vdecl()Brian Paul2013-08-211-6/+6
|
* gallivm: unify sin and cos implementationRoland Scheidegger2013-08-212-255/+53
| | | | | | | | | | | The (complicated!) math is all identical, there's just minimal differences how sign bit is calculated plus there's an additional subtraction for the argument going into the polynomial for cos. The logic stays 100% the same (with a small exception, sign bit calculation for sin is minimally simplified, applying sign mask after xoring the arguments instead of applying it to each argument). Reviewed-by: Jose Fonseca <[email protected]>
* gallivm: add comment for bogus min/mag filter selection with nearest mip filterRoland Scheidegger2013-08-213-2/+10
| | | | | | | Detected this hunting some other bug, not sure if it really needs fixing but it is definitely wrong. Reviewed-by: Jose Fonseca <[email protected]>
* gallivm: fix rho calculation for 1d caseRoland Scheidegger2013-08-211-1/+1
| | | | | | | Was using wrong (undefined) vector element (the elements are at 0/2 position, not 0/1). Reviewed-by: Jose Fonseca <[email protected]>
* i965/gen7: Set MOCS L3 cacheability for IVB/BYT (v2)Ville Syrjälä2013-08-218-17/+12
| | | | | | | | | | | | | | | | IVB/BYT also has the same L3 cacheability control in MOCS as HSW, so let's make use of it. pts/xonotic and pts/reaction @ 1920x1080 gain ~4% on my IVB GT2. Most other things show less gains/no regressions, except furmark which loses some 10 points. I didn't have a BYT at hand for testing. v2: Don't check (brw->gen == 7) in gen7 functions. (chadv) Signed-off-by: Ville Syrjälä <[email protected]> Reviewed-by: Chad Versace <[email protected]>
* i965/hsw: Populate MOCS for STATE_BASE_ADDRESS (v2)Ville Syrjälä2013-08-212-3/+11
| | | | | | | | | | | Just spotted these unpopulated MOCS fields when comparing the code against BSpec. Set the MOCS to the same as everywhere else in Haswell: L3-cacheable. v2: Annotate state packet fields (chadv). Signed-off-by: Ville Syrjälä <[email protected]> Reviewed-by: Chad Versace <[email protected]>
* glapi/gen: build temporary files in the build directoryMaarten Lankhorst2013-08-211-4/+4
| | | | | | | | | | Writing to the source directory can cause multiple parallel builds from the same source to fail. Create the temporary files in the build directory. Signed-off-by: Maarten Lankhorst <[email protected]> Reviewed-by: Matt Turner <[email protected]> Cc: "9.2" <[email protected]>
* mesa: Never advertise _S3TC compressed formatsIan Romanick2013-08-211-12/+0
| | | | | | | | | | | | The NVIDIA driver doesn't expose them, and piglit's arb_texture_compression-invalid-formats expects them to not be there. This, with the previous commit, fixes piglit arb_texture_compression-invalid-formats. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Anuj Phogat <[email protected]> Cc: "9.2" <[email protected]>
* mesa: Only advertise GL_ETC1_RGB8_OES in ES contextsIan Romanick2013-08-211-1/+9
| | | | | | | | | There is no extension for this format in desktop GL, so an application can't give the format back to glCompressedTexImage2D. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Anuj Phogat <[email protected]> Cc: "9.2" <[email protected]>
* glsl: Track existence of default float precision in GLSL ES fragment shadersIan Romanick2013-08-212-4/+68
| | | | | | | | | | | | | | | | | | | This is required by the spec, and it's a bit tricky because the default precision is scoped. As a result, I'm slightly abusing the symbol table. Fixes piglit no-default-float-precision.frag tests and the piglit default-precision-nested-scope-0[1234].frag tests that are currently on the piglit mailing list for review. On IRC I got confirmation from cwabbot that ARM (Mali T6xx and T400) enforces this requirement and from kusma that NVIDIA (Tegra2) enforces this requirement. We should be safe from regressing shipping applications. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Cc: "9.2" <[email protected]>
* glsl: Merge precision qualifiers tooIan Romanick2013-08-211-0/+3
| | | | | | | | | | | | | We never noticed this before because we previously didn't enfoce GLSL ES fragement shader requirements that precision be defined. There may also have been some interaction here with the addition of GL_ARB_shading_language_420pack, but it doesn't appear to me that it added any new bugs (just perhaps uncovered some old ones). Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Reviewed-by: Matt Turner <[email protected]> Cc: "9.2" <[email protected]>
* glsl: Pass type to is_valid_default_precision_type instead of nameIan Romanick2013-08-211-4/+5
| | | | | | | | This is used by the next patch. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Cc: "9.2" <[email protected]>
* vdpau/decode: Fix comment.Rico Schüller2013-08-211-1/+1
| | | | Reviewed-by: Christian König <[email protected]>
* vl/query: Only support VDP_CHROMA_TYPE_420 for 12 bit formats.Rico Schüller2013-08-211-1/+6
| | | | Reviewed-by: Christian König <[email protected]>
* util: add avx2 and xop detection to cpu detection codeRoland Scheidegger2013-08-203-2/+59
| | | | | | | | | | Going to need this soon (not going to bother with avx2 intrinsics at this time but don't want to do workarounds for true vector shifts if llvm itself can use them just fine and won't need the gazillion instruction emulation). Not really tested other than my cpu returns 0 for these features... (I have no idea if llvm actually would emit avx2/xop instructions neither...) Reviewed-by: Jose Fonseca <[email protected]>
* gallivm: fix bogus aos path detectionRoland Scheidegger2013-08-201-5/+11
| | | | | | | | | Need to check the wrap mode of the actually used coords not a fixed 2. While checking more than necessary would only potentially disable aos and not cause any harm I'm pretty sure for 3d textures it could have caused assertion failures (if s,t coords have simple filter and r not). Reviewed-by: Jose Fonseca <[email protected]>
* gallivm: do clamping of border color correctly for all formatsRoland Scheidegger2013-08-202-46/+256
| | | | | | | | | | | | | | | | | | Turns out it is actually very complicated to figure out what a format really is wrt range, as using channel information for determining unorm/snorm etc. doesn't work for a bunch of cases - namely compressed, subsampled, other. Also while here add clamping for uint/sint as well - d3d10 doesn't actually need this (can only use ld with these formats hence no border) and we could do this outside the shader for GL easily (due to the fixed texture/sampler relation) do it here too just so I can forget about it. v2: move border color clamping out of fetch texel. Also change it to clamp the whole border vector at once (and use vectorized load of border color), which saves a couple of instructions - needs some different handling of mixed signed/unsigned formats so skip the per channel stuff and just derive this from first channel except for special formats. Reviewed-by: Jose Fonseca <[email protected]>
* gallivm: implement better control of per-quad/per-element/scalar lodRoland Scheidegger2013-08-208-55/+149
| | | | | | | | | | | | | | | | There's a new debug value used to disable per-quad lod optimizations in fragment shader (ignored for vs/gs as the results are just too wrong typically). Also trying to detect if a supplied lod value is really a scalar (if it's coming from immediate or constant file) in which case sampler code can use this to stay on per-quad-lod path (in fact for explicit lod could simplify even further and use same lod for both quads in the avx case but this is not implemented yet). Still need to actually implement per-element lod bias (and derivatives), and need to handle per-element lod in size queries. v2: fix comments, prettify. Reviewed-by: Jose Fonseca <[email protected]>
* mesa: use ARRAY_SIZE() macro instead of magic numberBrian Paul2013-08-201-1/+1
| | | | Reviewed-by: Kenneth Graunke <[email protected]>
* build: fix out-of-tree builds in gallium/auxiliaryRoss Burton2013-08-201-0/+4
| | | | | | | | | | The rules were writing files to e.g. util/u_indices_gen.py, but in an out-of-tree build this directory doesn't exist in the build directory. So, create the directories just in case. Cc: [email protected] Reviewed-by: Matt Turner <[email protected]> Signed-off-by: Ross Burton <[email protected]>
* radeonsi: Always pre-load separate VGPRs for centroid vs. center interpolationMichel Dänzer2013-08-201-1/+2
| | | | | | | | | | | | | The LLVM R600 backend currently always uses separate VGPRs for these. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68162 (Centroid interpolation is identical to center interpolation without multisampling, so the shader hardware was only pre-loading one set of interpolation coefficients, and the pixel shader code was using uninitialized values as the centroid interpolation coefficients) Cc: [email protected] Tested-by: Laurent Carlier <[email protected]>
* radeonsi: Fix SPI_BARYC_CNTL register initializationMichel Dänzer2013-08-201-22/+3
| | | | | | | The centroid / center interpolation related bits have different meanings as of SI. Fixes 7 centroid interpolation related piglit tests.
* gallium/osmesa: add same checks to OSMesaMakeCurrent as the other osmesaMaarten Lankhorst2013-08-201-2/+3
| | | | | | | Fixes a opengl crash in wine. Cc: "9.2" <[email protected]> Signed-off-by: Maarten Lankhorst <[email protected]>
* gallium/osmesa: link against static libglapi library too to get the gl exportsMaarten Lankhorst2013-08-201-3/+2
| | | | | | | | | | | | This should fix missing symbols in a osmesa built against shared glapi osmesa build. All opengl exports were missing that are defined in the static glapi, so link against both to fix this. I could swear I've done this before, maybe there was a glitch in the matrix. Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=47824 Cc: "9.2" <[email protected]> Signed-off-by: Maarten Lankhorst <[email protected]>
* i965: Shorten sampler loops in precompile key setup.Kenneth Graunke2013-08-202-2/+4
| | | | | | | | | | | | | | | Now that we have the number of samplers available, we don't need to iterate over all 16. This should be particularly helpful for vertex shaders. v2: Use the correct shader program (caught by Paul Berry). This needs to initialize the exact same set of sampler swizzles as the actual key setup, or else we end up doing recompiles due to some being XYZW and others being 0. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Paul Berry <[email protected]>
* ilo: add ILO_DEBUG=flushChia-I Wu2013-08-2010-12/+29
| | | | | | | | When specified, ilo will print a line similar to cp flushed for render with 949+888 DWords (22.4%) because of frame end for every ilo_cp_flush() call.
* ilo: add ILO_DEBUG=drawChia-I Wu2013-08-205-2/+83
| | | | It can print out pipe_draw_info and the dirty bits set, useful for debugging.
* r600g/sb: Move memsets of member structs to within constructor bodies.Vinson Lee2013-08-192-6/+3
| | | | | | | Silences "Uninitialized pointer field" defects reported by Coverity. Signed-off-by: Vinson Lee <[email protected]> Reviewed-by: Vadim Girlin <[email protected]>
* glsl: Use alignment of container record for its first fieldIan Romanick2013-08-192-2/+28
| | | | | | | | | | | | | | | | | | | | The first field of a record in a UBO has the aligment of the record itself. Fixes piglit vs-struct-pad, fs-struct-pad, and (with the patch posted to the piglit list that extends the test) layout-std140. NOTE: The bit of strangeness with the version of visit_field without the record_type poitner is because that method is pure virtual in the base class. The original implementation of the class did this to ensure derived classes remembered to implement that flavor. Now they can implement either flavor but not both. I don't know a C++ way to enforce that. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Paul Berry <[email protected]> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68195 Cc: "9.2 9.1" [email protected]
* glsl: Add new overload of program_resource_visitor::visit_field methodIan Romanick2013-08-192-10/+48
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | The outer-most record is passed into the visit_field method for the first field. In other words, in the following structure: struct S1 { vec4 v; float f; }; struct S { S1 s1; S1 s2; }; uniform Ubo { S s; }; s.s1.v would get record_type = S (because s1.v is the first non-record field in S), and s.s2.v would get record_type = S1. s.s1.f and s.s2.f would get record_type = NULL becuase they aren't the first field of anything. This new overload isn't used yet, but the next patch will add several uses. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Paul Berry <[email protected]> Cc: "9.2 9.1" [email protected]
* glsl: Disallow embedded structure definitionsIan Romanick2013-08-193-0/+38
| | | | | | | | | | Continue to allow them in GLSL 1.10 because the spec allows it. Generate an error in all other versions because the specs specifically disallow it. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Cc: "9.2" <[email protected]>
* meta: Add default precision qualifier to all fragement shadersIan Romanick2013-08-191-0/+18
| | | | | | Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Cc: "9.2" <[email protected]>
* glsl: Add default precision qualifiers for ES builtinsIan Romanick2013-08-195-0/+6
| | | | | | | | | Once the compiler proplerly checks for default precision qualifiers, these shaders will cease to compile. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]> Cc: "9.2" <[email protected]>
* glsl: Remove extra "types" from error messageIan Romanick2013-08-191-1/+1
| | | | | | | Send it straight to the Department of Redundancy Department. Signed-off-by: Ian Romanick <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* i965: Make the VS binding table as small as possible.Kenneth Graunke2013-08-191-3/+4
| | | | | | | | For some reason, we didn't use this information even though the VS backend has computed it (albeit poorly) for ages. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Paul Berry <[email protected]>
* i965/vs: Rework binding table size calculation.Kenneth Graunke2013-08-195-16/+20
| | | | | | | | | | | | | | | | | | | | | | Unlike the FS, the VS backend already computed the binding table size. However, it did so poorly: after compilation, it looked to see if any pull constants/textures/UBOs were in use, and set num_surfaces to the maximum surface index for that category. If the VS only used a single texture or UBO, this overcounted by quite a bit. The shader time surface was also noted at state upload time (during drawing), not at compile time, which is inefficient. I believe it also had an off by one error. This patch computes it accurately, while also simplifying the code. It also renames num_surfaces to binding_table_size, since num_surfaces wasn't actually the number of surfaces used. For example, a VS that used one UBO and no other surfaces would have set num_surfaces to SURF_INDEX_VS_UBO(1) == 18, rather than 1. A bit of a misnomer there. Signed-off-by: Kenneth Graunke <[email protected]> Reviewed-by: Paul Berry <[email protected]>
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-25 07:29:53 +0000