kc3-lang/angle/src/compiler

src/compiler

Log

Author Commit Date CI Message
Shahbaz Youssefi b712a720 2025-10-10T17:06:44 Translator: Check for recursion during parse GLSL bans static recursion (i.e. it's an error even if recursion could get dead-code-eliminated). Tracking the call graph at parse time lets the translator be precise about this. This change also prepares for a few follow up changes that make CallDAG obsolete. Bug: angleproject:349994211 Change-Id: I66f2cc3e544932f089f8185ccf15b3bd98779d13 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7031487 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 8469b249 2025-10-10T14:12:20 Translator: Reject infinite loop during parse Bug: angleproject:349994211 Change-Id: I38964414b2992319e5f80f18a417210f86d806a6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7029684 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi caf63365 2025-10-15T13:10:37 Translator: Avoid unreachable in presence of shader error When the gl_ClipDistance/gl_CullDistance index is tracked for future validation, the parser expected the index type to be valid, but it may not be if there's a compile error. Bug: angleproject:349994211 Change-Id: I94f39bec34a8ae68ef7d6c3cce256026808c0bc9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7042014 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi ded8ecc0 2025-10-15T15:26:23 Translator: Handle GS and TES output gl_Position like VS The spec declares these special shader variables identically as: ``` out gl_PerVertex { highp vec4 gl_Position; }; ``` Between vertex, geometry and tessellation evaluation shaders. Since the output is nameless, vertex shaders have always treated `gl_Position` as if it's a variable on its own. In this change, TES and GS treat that variable the same way. Additionally, this change removes `gl_Position` as a standalone variable from tessellation _control_ shaders, which must always use `gl_out` instead. Bug: angleproject:349994211 Change-Id: Ib33592fb62af31021a99f76f77c27da214425826 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7042770 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 97b93617 2025-10-15T15:17:15 Translator: Remove gl_out from TES built-ins Tessellation evaluation shaders don't have such a built-in. Like geometry shaders, they only have an `out gl_PerVertex` that is not an array. Bug: angleproject:349994211 Change-Id: I6b333a223b721ca5ceb2d6c5658d802cb66769ad Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7042769 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi d9ff0591 2025-10-15T13:43:51 Translator fuzzer: dump the shader to stderr Bug: chromium:448166385 Change-Id: I67ff0a5b78a488bef2a867abef4b7f1ba72137bf Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7042568 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 255dc673 2025-10-09T00:17:59 Translator: Validate ESSL 100 limitations during parse Validating the shape of the for loop as specifically required by WebGL (per ESSL 100 spec, Appendix A) is more easily done during parse, where the shape of the input GLSL is exactly known. Bug: angleproject:349994211 Change-Id: I05a9b1b6915dbc1f6c74bb9159f6cc85fefc7b67 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7022355 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 84b0cf1a 2025-10-08T15:54:43 Translator: Validate gl_FragData xor gl_FragColor early The translator validates that these two built-ins are not simultaneously used. This validation is moved earlier to the post-parse checks. Bug: angleproject:349994211 Change-Id: I59f1266406c3f82da560f3a18bb88571ab614d75 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7022017 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi 8c15b966 2025-10-08T15:31:25 Translator: Validate MaxFunctionParameters limit during parse Bug: angleproject:349994211 Change-Id: Ic9e4a6d1279cdfdb8d24a029182f3c83a79ef05f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7022016 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi e5b0553a 2025-10-08T15:18:25 Translator: Validate TCS barrier() rules during parse ... instead of doing an AST walk after parse. Bug: angleproject:349994211 Change-Id: I175fecb593ac701208c20f77bb8dd3987c63fa11 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7022014 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi bf63a4ef 2025-10-08T15:13:58 Translator: Track control flow during parse This change adds a stack to track control flow structures. Currently, this replaces the loop/switch nesting level validation and may look unnecessary, but a number of other post-parse validation will be moved to parse-time and build on this stack. Bug: angleproject:349994211 Change-Id: I841ce780f22410055feedb3194c24d84871a89cd Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7022013 Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 8af6c566 2025-10-08T11:25:49 Translator: Add more action routines to glslang.y These will be used during IR generation, but are split in this change because commits with glslang_tab_autogen.cpp take forever to `git cl format`. Bug: angleproject:349994211 Change-Id: Ib0a84710fa63fc33193fd5f485a434981baab61e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7022211 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi eec95a6b 2025-10-02T17:02:14 Translator: Remove NV_shader_framebuffer_fetch This extension is never exposed by ANGLE. AFAICT, it was removed in https://chromium-review.googlesource.com/c/angle/angle/+/688099 and never added again. The current implementation is incorrect even if the extension is exposed; we assume gl_LastFragColorARM is the name, while this extension uses gl_LastFragColor. This change removes mentions of this extension instead of trying to fix it. Bug: angleproject:349994211 Change-Id: I2391a225db95d73579fd71cf261360973e5935f7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7004390 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 074eec2f 2025-10-02T15:31:19 Translator: remove angle_BaseVertex/Instance from built-ins These aren't built-ins, they are just used for emulation. Instead of listing them in the built-in variables, they are now created as uniforms during emulation. This change also gives gl_BaseVertex/Instance their own qualifiers so they aren't considered uniforms. The emulation pass is also optimized to do fewer passes. Bug: angleproject:349994211 Change-Id: I44fd345fd4a2dc4a001eb294952a984ae710d606 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7004388 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi b0ee4239 2025-10-01T16:46:10 Translator: Validate clip/cull distance size during parse ... instead of in a post-parse pass. In this process, these array sizes are determined during parse itself, and so can be applied to the IR earlier. Bug: angleproject:349994211 Change-Id: I542970f9262e346fd7443e48d780e8abfec95315 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7003832 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Matthew Denton fecb8ead 2025-09-04T12:27:48 WGSL: implement inc/dec with generated functions Implements inc/dec with generated WGSL functions that take pointers and perform a post/pre inc/decrement. This works with scalars, vectors, and matrices, both float and int. WGSL supports inc/dec only on integer types, and only as statements (not as expressions). https://www.w3.org/TR/WGSL/#increment-decrement. The regular ++ and -- are used in this specific case. The WGSL outputter records usage of increment/decrement and produces a call to the correct function. A new class is introduced to keep the record of which types need generated inc/dec WGSL functions. Bug: angleproject:42267100 Change-Id: I0e70760ba5bd00f978e216f958216ae3137a146e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6935269 Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 6a997ce2 2025-10-03T00:03:25 Vulkan: SPIR-V Gen: Use unique id instead of pointer Bug: angleproject:349994211 Change-Id: I66d6c3cf58f1e7d850fb19def7aa863774d918a9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7007612 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton 6c4c0055 2025-10-01T15:13:14 WGSL: fix do-while when body has continue If the body of the do while had a "continue" it would skip over the evaluation of the loop condition and potentially loop forever. Change to use the special WGSL "continuing" statement, which goes at the end of a loop body and always executes, even if there is an earlier "continue". Bug: angleproject:42267100 Change-Id: I4ac73e6abcb12e0ff395b83dc5666ac1870724e9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7003772 Reviewed-by: Liza Burakova <liza@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi e1c14f0b 2025-10-02T23:42:01 Translator: Fix gl_LastFragColorARM not being qualified correctly The qualifier adjustment was accidentally done after declaring the variable in some cases. Bug: angleproject:349994211 Change-Id: Iffd7567a119e1cd499c068742edfcb68a6a2336f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7006113 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi d9fa255a 2025-10-02T11:58:06 Translator: Add numerous missing gl_PerVertex validations Bug: angleproject:349994211 Change-Id: Ie4a602dd026a2dab95c9d77e8841e76967fa7f3d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7002885 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 1f3b482f 2025-10-02T12:05:44 Translator: Fix UNREACHABLE() on gl_DepthRange compile error Bug: angleproject:349994211 Change-Id: I1918c8390de3fa4a5c7542b044e91e8d5e944b3a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7002886 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Matthew Denton 4cff5289 2025-08-12T19:31:36 WGSL: append TSymbolUniqueId to overloaded functions dEQP-GLES2.functional.shaders.functions.overloading.* passes after this. So that overloaded functions have unique names in GLSL, append their TSymbol's unique id to the name of the emitted WGSL function. Bug: angleproject:42267100 Change-Id: I25026e71f9e27cb323014cd4102f914226c27d02 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6843927 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Shahbaz Youssefi ae02c329 2025-09-30T00:04:58 Translator: Give gl_DepthRange its own qualifier gl_DepthRange and gl_NumSamples are considered "uniform" by the GLSL spec, but gl_NumSamples gets its own qualifier in the translator. This left gl_DepthRange as the only built-in that had to be detected by its name instead of qualifier. This change adds a qualifier for gl_DepthRange to remove this special-ness. Bug: angleproject:349994211 Change-Id: I2b7b0acf3dcaef22bb4a98e4d35fd2cdfd5c825a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6998990 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi 789693f0 2025-09-30T14:30:59 Translator: Fail compilation if built-in redeclared after qualified If a built-in is qualified then redeclared like so: invariant gl_FragDepth; out float gl_FragDepth; the translator would have to remove the previous qualification statement and apply its effect to the redeclared variable directly. While this is possible, it has always been broken (i.e. there are no users), and is not really useful to support, since the application could have done the following instead: invariant out float gl_FragDepth; The spec does not explicitly call this out as invalid, but this change generates a compile error for the above scenario anyway, as supporting it can get unnecessarily intrusive. Bug: angleproject:349994211 Change-Id: I9ddb181c8808f50aeed5b327a95b3df829206822 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6999640 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton 7b14dd73 2025-08-19T15:57:34 WGSL: RewriteMixedTypeMathExprs RewriteMixedTypeMathExprs: Some mixed-type arithmetic is legal in GLSL but not WGSL. Generate code to perform the arithmetic as specified in GLSL. Example: uvec2 x; uint y; x &= y; Is transformed into: x &= uvec(y); Also, mat2 x; int y; x += y; Is transformed into: x += mat2(float(y), float(y), float(y), float(y)) Bug: angleproject:42267100 Change-Id: I4a0ec1d9806b3331b4b1feff6fbe7c0f212f8120 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6862843 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Shahbaz Youssefi a38f60ea 2025-09-30T13:40:44 Translator: Fail compilation if main() is only prototype Bug: chromium:448166385 Change-Id: Ia6b69839e35f3e0dcdf3e6e918397d3e2fd83fa2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/7000630 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 85631b1d 2025-09-30T14:25:30 Translator fuzzer: Dump the shader before compiling it So that fuzzer reports tell you what the shader is without us needing to debug-print it. Bug: chromium:448166385 Change-Id: If2454093abd73771cbac787c7516259db6be82d2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6999637 Commit-Queue: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi e5b33850 2025-09-26T14:59:20 Translator: Fail earlier if main() is not present. The IR assumes that main() is available on successful parse. Additionally, in preparation for IR removing a check that the AST root is set up, a compile failure to that effect is turned into an assertion. The AST generated before IR is logically only used for validation purposes, so the compiler cannot rely on it to generate an error. Bug: angleproject:349994211 Change-Id: I5d4501b52fed8214213a4ae1dada923decee9a8b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6991529 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi f3405af0 2025-09-25T13:45:08 Translator: Fix gl_in redeclaration This was seemingly untested, and had multiple issues: * GLSL generation would output `gl_in gl_PerVertex { ... } gl_in;` * SPIR-V generation would redeclare `gl_in` despite one already existing. * If sized after declaration, the redeclared `gl_in` was referencing a different variable than the resized one. Bug: angleproject:42264111 Bug: angleproject:349994211 Bug: angleproject:447556589 Change-Id: Id63912b68625bafcc64197807b1c8a86e6f952b5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6985323 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Chris Dalton f95cbef4 2025-09-07T17:50:37 Reland "Add a "noncoherent" layout qualifier to PLS" This is a reland of commit ef9c5ce43110f4f8ba56124bdee927944ce3ddd1 with an invalid assert removed. Original change's description: > Add a "noncoherent" layout qualifier to PLS > > Raster ordering can be expensive. If an app knows the geometry in a draw > does not overlap, it can mark the PLS planes as noncoherent and opt for > using barriers instead. > > Bug: angleproject:40096838 > Change-Id: I2d686c2e64f51940face186637a9fcadcc251bef > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6923300 > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Kenneth Russell <kbr@chromium.org> > Commit-Queue: Chris Dalton <chris@rive.app> Bug: angleproject:40096838 Change-Id: I55077a973063427dfb0d510c87e75ec86ce8bb43 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6991929 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org>
Shahbaz Youssefi 18f28e59 2025-09-25T13:41:45 Translator: Mark redeclared gl_Position/PointSize as built-in When redeclared according to EXT_separate_shader_objects, these variables were given a UserDefined symbol type. While this doesn't affect SPIR-V (with Vulkan being the only backend that supports EXT_separate_shader_objects), it does affect the rest of the AST, including GLSL generation which would output _ugl_Position if support for the extension is enabled. Bug: angleproject:42265086 Bug: angleproject:349994211 Change-Id: I04d088017dbe6e32bfee675f0124c4cc3ffdefc2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6985322 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi 4eb8604f 2025-09-24T00:13:51 Translator: Use EvqNumSamples for gl_NumSamples This qualifier was seemingly intended for this built-in, but EvqUniform was used for it instead; presumably following gl_DepthRange. Bug: angleproject:349994211 Change-Id: I1ec7f76195d790ac2a53c4f2b0f9a136976480d5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6977986 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Shahbaz Youssefi bfdf223a 2025-09-23T11:42:39 Translator: Don't differentiate static read vs write In the end, the only thing that matters is whether a symbol is statically used, so no need to track whether the static use was in a read or a write context. Bug: angleproject:349994211 Change-Id: I4964fc9f76a5f7798f1b5e42ce2ae997dca78063 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6976297 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org>
Shahbaz Youssefi a44693f6 2025-09-17T14:50:37 Vulkan: Fix SPIR-V gen for invariant structs Bug: angleproject:349994211 Change-Id: I7db05d9a6232a577a65f7a337cfd1636bce0c53b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6961859 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Brian Sheedy 44db20f2 2025-09-17T15:57:09 Revert "Add a "noncoherent" layout qualifier to PLS" This reverts commit ef9c5ce43110f4f8ba56124bdee927944ce3ddd1. Reason for revert: Causes issues on Apple silicon crbug.com/445727125 Original change's description: > Add a "noncoherent" layout qualifier to PLS > > Raster ordering can be expensive. If an app knows the geometry in a draw > does not overlap, it can mark the PLS planes as noncoherent and opt for > using barriers instead. > > Bug: angleproject:40096838 > Change-Id: I2d686c2e64f51940face186637a9fcadcc251bef > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6923300 > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Kenneth Russell <kbr@chromium.org> > Commit-Queue: Chris Dalton <chris@rive.app> Bug: angleproject:40096838 No-Presubmit: true No-Tree-Checks: true No-Try: true Change-Id: Ibc24bed27cec406b6e4ffa9aaa8fe3a2110b58ba Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6962955 Reviewed-by: Kenneth Russell <kbr@chromium.org> Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Brian Sheedy <bsheedy@chromium.org>
Shahbaz Youssefi f128872e 2025-09-16T23:37:08 Translator: Compare variables by unique id instead of pointer Bug: angleproject:349994211 Change-Id: I32ab2ffe9a04e196330949484e704913d0f4e41d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6955119 Reviewed-by: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Chris Dalton ef9c5ce4 2025-09-07T17:50:37 Add a "noncoherent" layout qualifier to PLS Raster ordering can be expensive. If an app knows the geometry in a draw does not overlap, it can mark the PLS planes as noncoherent and opt for using barriers instead. Bug: angleproject:40096838 Change-Id: I2d686c2e64f51940face186637a9fcadcc251bef Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6923300 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Commit-Queue: Chris Dalton <chris@rive.app>
Roberto Rodriguez 16771259 2025-09-12T11:10:18 Update main() prototype in shader if main() has been replaced. If the main() function definition has been wrapped/replaced in a shader and a main() function prototype is also present, the prototype needs to be replaced. Otherwise it will continue to reference the replaced main() which can cause issues with shader translation steps. Bug: angleproject:444653099 Change-Id: Ie6ce85cac89e026876a1b6e25cd294f1d8a536c4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6944807 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Kimmo Kinnunen <kkinnunen@apple.com>
Yuxin Hu b8281f5d 2025-09-09T17:15:26 Add condition check in emulate dithering shader code Prior to this change, the emulate dithering shader code will write to all elements of gl_FragData, regardless of whether the dither value is zero or not: gl_FragData[i] = gl_FragData[i] + dither_value. If dither_value is 0, then this doesn't change gl_FragData[i] and it incurs unnecessary shader writes. This change adds a boolean variable that is initialized to false, and it is only set to true if dither value for that output element is changed from zero to non-zero value. This way we are only issuing shader write operations when necessary. Bug: angleproject:425733272 Change-Id: Iacfd81ff9a1d7d0f71ecf8d119033b2500d2869e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6931777 Reviewed-by: Charlie Lao <cclao@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Geoff Lang a0cc64cd 2025-09-10T15:18:17 HLSL: Mark shadow sampler textureLod calls as using gradients TextureLod calls bottom out in the HLSL SampleCmp function which does not take a LOD parameter so we end up dropping it. This causes them to need gradients. Mark them as using gradient so that any loop they are used in is flagged and not unrolled by FXC. Reverted original fix from https://chromium-review.googlesource.com/c/angle/angle/+/6421496 because the wrong number of parameters are passed to the generated comparison sampling function when it is changed to LOD0. Thanks to Mikkel Simonsen at Unity and Calder Young for the tests. Fixed: angleproject:406190894 Change-Id: Ia9255e4336099045366c02f945a3c4de8075427a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6936488 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org>
Matthew Denton a471c005 2025-08-19T17:24:38 WGSL: derivative_uniformity diagnostic should be a warning WGSL requires derivatives to be calculated only if it can statically prove control flow is uniform. GLSL allows dynamically uniform control flow in this case. This CL uses WGSL's diagnostic filter to make derivative_uniformity diagnostics into warnings, to match GLSL semantics. Bug: angleproject:392542001 Change-Id: I98dd55205ad6d3c9d13ca3f94c0e7858ba92d536 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6862844 Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Matthew Denton <mpdenton@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Matthew Denton cfe2c8fe 2025-06-25T13:25:01 WGSL: RewriteMultielementSwizzle WGSL doesn't support assignments to multi-element swizzles. This is used in a lot of shader tests, so temporarily work around this with an AST traverser that splits these assignments into multiple assignments that only assign to single element swizzles. One special case is multiplication-by-a-matrix assignment: vec.xy *= mat; is converted to vec.x = (vec.xy * mat).x; vec.y = (vec.xy * mat).y; Bug: angleproject:392542001 Change-Id: I3f393039aae13eb3f2c5dc5e553f68eb03b6316d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6847280 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Liza Burakova <liza@chromium.org>
Yuxin Hu c1cdc2b5 2025-08-22T15:23:58 Transform SPIRV to use 16-bit float for lower precision uniforms This change adds a ShCompileOption flag transformFloatUniformTo16Bits. The flag is turned on in vulkan backends where VK_KHR_16bit_storage extension is supported, and uniformAndStorageBuffer16BitAccess feature is supported. When the compiler flag is turned on, in the generated SPIRV, float data types in mediump and lowp uniforms are transformed from 32-bit to 16-bit. The 16-bit float uniform data is converted to 32-bit with OpFConvert instruction upon loading in SPIRV, this is to minimize the changes in OutputSPIRV.cpp. The converted variable is decorated with RelaxedPrecision, so that SPIRV compiler should be able to treat the converted variable as 16 bits, and the hardware can still benefit from reduced precision floats. The frontend is also notified such SPIRV shader changes by setting the isFloat16 bit in CollectVariables() step, and the frontend will transform float uniform data from 32-bit to 16-bit before storing the data into memory. That way, the uniform data that SPIRV shader reads matches with the uniform data type transformed in the SPIRV shader. This change also updates some test code to allow relative 2^-10 precision wiggle room for mediump uniform floats. This is valid according to spec: https://developer.arm.com/documentation/102502/0101/Shader-precision Bug: angleproject:405795981 Bug: angleproject:440941211 Change-Id: I05db7f5ef744df513fbad87cfed8aa173890ec26 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6851560 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Tom Sepez eefcdafd 2025-08-29T00:13:27 Fix some old clang-tidy warnings about move constructors -- Remove move of copy-only ImmutableString -- Fix rule-of-five issues with StringPart. Bug: b/364788123 Change-Id: Ief9c2a83df15d2e8856e8f729c1908df65004bfc Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6896659 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Tom Sepez <tsepez@chromium.org>
Tom Sepez 6f0337a0 2025-08-28T23:13:22 Avoid some needless c_str() calls. Remove conversions from string -> char* -> string, as detected some time ago by a clang compiler plugin. Typically, this occurs when passing a c_str() result to a function that expects a string argument. Bug: b/412730353 Change-Id: I1d9c83e9ed5c4900eec266e71f534661f0f3d4d4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6896657 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Tom Sepez <tsepez@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 1ae273ee 2025-08-27T10:47:59 Translator: Output struct members in OutputTree Bug: chromium:438038775 Change-Id: I2a1bed5207b1b2c26f20ee9ebd80fb00ec803bc2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6891072 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi af0386cb 2025-08-27T10:44:53 Reorder structs entirely made of samplers to the end of struct An earlier change [1] moved samplers to the end of structs so that extracting them does not disturb the indices to the non-sampler fields. This change does the same for structs that only contain samplers, because they too get removed from the parent struct, disturbing indices. [1]:https://chromium-review.googlesource.com/c/angle/angle/+/5748789 Bug: chromium:438038775 Change-Id: I7ac2ea13723bf06b4b172b6e70f0dbe85bc45ce3 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6891071 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Cody Northrop 9dc36791 2025-08-28T07:24:21 WGSL: Fix unused variable in translator Test: Build with is_debug = false Bug: angleproject:441701631 Change-Id: Ifa44bb5388788602e76c2ae2005bdd41d2c22dd4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6896036 Auto-Submit: Cody Northrop <cnorthrop@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Tom Sepez a02670d6 2025-08-26T20:41:16 Move unsafe buffers inside header guard macros While this is exactly opposite of what Chromium has chosen to do, there is an issue with clang-format trying to indent preprocessor directives four spaces relative to include guard. This is because Angle's .clang-format file specifies IndentPPDirectives: AfterHash but Chromium's does not. The current placement is sufficient to throw off clang-format's guard detection since the guard macro no longer covers the entire file. Bug: b/436880895 Change-Id: Ic6b99c8cef6213939cdf9b42af8730e1eb423065 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6885892 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org> Auto-Submit: Tom Sepez <tsepez@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton b9cec916 2025-08-12T19:44:18 WGSL: default uniforms gathered in interface block The default uniform struct was being manually output by OutputUniformBlocksAndSamplers(), which did not add the appropriate @align(16) annotations to lay the struct out according to WGSL's uniform address space layout requirements. This CL uses Vulkan's method of gathering the default uniforms into an interface block. The interface block will be output normally by the traverser, including @align() annotations. The variable declaring an instance of the interface block is still output by OutputUniformBlocksAndSamplers() because it needs special @group() and @binding() annotations. Bug: angleproject:376553328 Change-Id: Ib3be7d51ffedefe5ec579a1c9aaf0a535146c694 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6795028 Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Liza Burakova <liza@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Tom Sepez 25390156 2025-08-21T00:13:19 Suppress unsafe buffers on a file-by-file basis in src/ [1 of N] In this CL, we suppress many files but stop short of actually enabling the warning by not removing the line from the unsafe_buffers_paths.txt file. That will happen in a follow-on CL, along with resolving any stragglers missed here. This is mostly a manual change so as to familiarize myself with the kinds of issues faced by the Angle codebase when applying buffer safety warnings. -- Re-generate affected hashes. -- Clang-format applied to all changed files. -- Add a few missing .reserve() calls to vectors as noticed. -- Fix some mismatches between file names and header comments. -- Be more consistent with header comment format (blank lines and trailing //-only lines when a filename comment adjoins license boilerplate). Bug: b/436880895 Change-Id: I3bde5cc2059acbe8345057289214f1a26f1c34aa Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6869022 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu f63751db 2025-08-22T10:34:45 Fix the UniformSortComparator not following strict weak order If one uniform is lowp and the other uniform is mediump, then the compare(a, b) and compare (b, a) both return true. This violates the strict weak order rule where: If comp(a, b) == true then comp(b, a) == false This change fixes the bug. Bug: angleproject:405795981 Change-Id: I47e2668fc0fa87ecea9d3f9bce511c9823a39929 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6876519 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Yuxin Hu 20dca0b9 2025-08-14T13:54:37 Add a new TInterfaceBlock class member to hint ANGLE default Uniform In later change, we will transform floats in ANGLE default uniform struct from 32-bit to 16-bit in SPIRV. The new class member mIsDefaultUniformBlock will help us identify this special struct. This change is no op. Bug: angleproject:405795981 Change-Id: I939caf7491ab5d7de130ba6a2fd8b1202958c375 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6852398 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Sungyong Choi a06cb368 2025-08-12T11:45:43 Fix GCC -Wdangling-pointer warning in NodeStackGuard The previous implementation stored the address of a local member (oldNodeStack) into ConsList::tail, which triggers the warning even though the lifetime is safe within the guard's scope. This change allocates a copy of the previous nodeStack on the heap using std::make_unique, and stores that pointer in tail. This ensures the referenced ConsList object has a lifetime tied to the guard object itself and is not considered a dangling pointer by GCC's static analysis. BUG: angleproject:438226513 Change-Id: Ifea04515db7b1988dff6ec41c60799158b753128 Signed-off-by: Sungyong Choi <sywow.choi@samsung.com> Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6845476 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Yuxin Hu 806c80ec 2025-08-14T13:41:58 Add a new ShaderVariable struct property Add a new ShaderVariable struct property isTransformedFP16FloatUniform. This property will be set by compiler when the corresponding shader variable data type in shader code is transformed from 32 bit to 16 bit. The property value be later passed to LinkedUniform so that the frontend knows if it should transform the data from 32 bit to 16 bit before writing to memory, and if it should transform the data from 16 bit to 32 bit after reading from memory. This change is no op. Bug: angleproject:405795981 Change-Id: Iaae15baa8ab9f7c3295071fa1e1ab04752bd69bb Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6852396 Auto-Submit: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Panfeng Hou 25f5e2cf 2025-08-13T17:03:19 Vulkan: Add support for built-in gl_PrimitiveShadingRateEXT Add support for built-in gl_PrimitiveShadingRateEXT and relative const value. Bug: angleproject:397255625 Change-Id: I25d51f5805697e137b232175ee790808c45850b7 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6845515 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu 1d8279a2 2025-08-13T13:35:02 Do not sort ANGLE internal uniforms or built-in uniforms On Mac, for webGL app on chrome, if the uniforms added by ANGLE are not placed at the top of the ANGLE_UserUniforms struct, the user defined uniforms in the same ANGLE_UserUniforms struct seem to be messed up and not read correctly by shader code. This change fixes it by moving the sortUniforms() function call to the place before we add ANGLE internal uniforms. That way ANGLE internal uniforms will be inserted later and stay on top of ANGLE_UserUniforms struct. Bug: chromium:437678149 Change-Id: I34a0d3dcb98c5fe30ad248381ecb5f7f78252275 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6847281 Auto-Submit: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Matthew Denton 63d8f74c 2025-08-12T19:39:03 WGSL: move DeclareDefaultUniforms() out of SPIR-V WGSL needs the same thing (default uniforms gathered into an interface block). Changes 3 things in the file move: 1. Now capable of adding a named interface block (Vulkan's is nameless) to hold the default uniforms. 2. Removes inactive uniforms 3. Deals with the case where there are no active default uniforms, by not declaring an interface block at all. Bug: angleproject:376553328 Change-Id: Ia25318b4db93068010c3c44414cd00803b783799 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6843929 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Yuxin Hu edb05627 2025-08-12T12:17:36 Fix sort uniform bug Current sort doesn't take care of the case when both uniforms are struct specfiers. Given that struct types are easily broken by sort, make the sort not reordering uniforms if both are structs. Bug: b/437825940 Change-Id: Idda1810ac4234f7e1547735e4e09658ab0a57eed Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6842936 Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton 7e28089b 2025-07-31T18:44:01 WGSL: support bvecs in uniforms WGSL does not allow booleans in the uniform address space. The last CL uses u32s to represent bools in the uniform address space, and this CL uses vecN<u32> to represent bvecN. Bug: angleproject:376553328 Change-Id: Ibf9f1fbf230ca03dd3d4661ba599e5fb1cb26ac4 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6808955 Reviewed-by: Liza Burakova <liza@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Shahbaz Youssefi b2e2766c 2025-08-12T11:08:46 Translator: Reject bool < bool This is not allowed per the GLSL ES spec. Bug: chromium:437983671 Change-Id: I9bd46718a35917dcdf6382cd51253ca2b68d5636 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6842555 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Amirali Abdolrashidi <abdolrashidi@google.com>
Matthew Denton ca3d732d 2025-07-24T16:22:31 WGSL: support bool in uniforms WGSL does not allow booleans in the uniform address space. This CL changes the translator to substitute u32 for bool in the uniform address space, and convert it to bool on use. Arrays of bools are obviously arrays of u32s, and those will need special conversion functions to convert the entire array<u32> to array<bool> if necessary. This also includes the optimization of an array<bool> in a uniform--when indexing into it, only the indexed element will be converted to a native bool, instead of converting the entire array and then indexing. Note that substituting u32 for bool matches std140, so this change requires no changes to layout of uniforms. Also note that WGSL really likes explicit casts, so there's not really a way to avoid inserting explicit casts everywhere when using u32. Bug: angleproject:376553328 Change-Id: I8f72e55c6b401c28ff622622df7a450b7032721f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6785609 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Liza Burakova <liza@chromium.org>
Amirali Abdolrashidi fc0e0397 2025-07-31T13:15:45 Translator: Fix location validation for inout Without this validation, a shader using an inout variable can write into a memory beyond the limit set by MaxDrawBuffers when shader framebuffer fetch is enabled. * Extended the location check in visitSymbol() to include InOut qualifiers (EvqFragmentInOut) as well (in ValidateOutputsTraverser). * Moved RewritePixelLocalStorage() after ValidateOutputs() to avoid issues related to locations not being specified in the InOut nodes from the PLS implementation. * Added a unit test to test that this case results in a compilation error: GLSLTest_ES3.CompileFSWithInoutLocBeyondMaxDrawBuffers Bug: chromium:435139154 Change-Id: I44c729377a2e1d3126c2d279172217b2ab92b57b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6806230 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton a1d5d102 2025-07-22T02:15:43 WGSL: Allow matrices as in/out vars in shaders WGSL only supports scalars and vectors in in/out vars in shaders, matrices will need to be broken into column vectors and then put back together at the beginning (or end) of the shader. Arrays also need to be split, which will be done in another CL. Bug: angleproject:42267100 Change-Id: If1ba28c1b687ae0a3a5a554479f0ff0b5d9df39c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6777201 Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Liza Burakova <liza@chromium.org>
Matthew Denton 6a3ff5a1 2025-07-17T12:41:13 WGSL: extra debug logging for failure cases Bug: angleproject:42267100 Change-Id: I8938526b60af71d15d4327bc10fa172117afef60 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6777200 Reviewed-by: Geoff Lang <geofflang@chromium.org>
Matthew Denton 343e3bdc 2025-07-17T12:38:31 ScalarizeVecAndMatConstructors: also handle scalar constructors In WGSL you cannot "construct" a scalar from a non-scalar (vec or mat). This modifies ScalarizeVecAndMatConstructors to handle scalar constructors as well, indexing and swizzling the argument to select the first element of the nonscalar. This makes dEQP-GLES2.functional.shaders.conversions.vector_to_scalar.* pass (except those that use bool uniforms which are still unsupported). Bug: angleproject:42267100, angleproject:395659799 Change-Id: Ibaa3dc14f36fecb384bfb6e7f226c6d5f377dd5f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6777199 Reviewed-by: Geoff Lang <geofflang@chromium.org>
Geoff Lang b7582680 2025-02-24T18:04:32 Reland: GL: Allow untranslated shaders to pass through on GLES Add an EGL extension EGL_ANGLE_create_context_passthrough_shaders which uses the NULL translator and passes the original shader to the driver. The parser is still used for shader reflection. Always enable the null compiler backend. It has almost no binary size cost and is now potentially used when the null ANGLE backend is not enabled. Bug: angleproject:398857482 Change-Id: Id528189ccbbacb1c444eacb151baadfda9fcc04b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6488609 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Matthew Denton cbff5658 2025-07-16T14:10:47 ScalarizeVecAndMatConstructorArgs: cast for mat constructors In WGSL, matrix constructors also can't handle arguments of differing basic type than the constructor's basic type, so cast when necessary. Makes dEQP-GLES2.functional.shaders.conversions.scalar_to_matrix.* pass for WGSL. dEQP-GLES2.functional.shaders.conversions.matrix_to_matrix.* does not yet pass because WGSL can't handle matrix shader inputs/outputs. Bug: angleproject:42267100, angleproject:395659799 Change-Id: Ia4761190c78ccfe7d748275642361ada93314529 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6777198 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Liza Burakova <liza@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Matthew Denton 2c1a55b3 2025-07-16T13:44:49 WGSL: implement == for vectors by wrapping in all() WGSL's vec comparisons are component-wise, so wrapping the result in a call to all() returns true if all components compare equal, matching the behavior of GLSL's vec comparisons (== and !=). Bug: angleproject:42267100 Change-Id: Icfbfacf72e53096e2567fa89bcd4bf573e457ec1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6777197 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Liza Burakova <liza@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Shahbaz Youssefi e98dec82 2025-07-20T22:16:00 Translator: Remove Sampler2DRectShadow This is a desktop GLSL type. Bug: angleproject:370937467 Change-Id: Ife8b50e57f6e054ccd35df10c2157f8e4a218d37 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6774081 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuxin Hu <yuxinhu@google.com>
Panfeng Hou 251ffe62 2025-07-15T17:37:30 Vulkan: Add support for built-in gl_ShadingRateEXT Add support for built-in gl_ShadingRateEXT of GL_EXT_fragment_shading_rate Bug: angleproject:420310117 Change-Id: Ie11b139a0371b5995f1533a85e02c590cd36109c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6733750 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Yuxin Hu 7cbb77a8 2025-07-14T11:32:04 Fix the uniform reorder bug When sorting uniforms, we should first check if both uniforms are structs, and make sure to place the uniforms with specifier in front of the uniforms without specifier. Bug: chromium:431318231 Change-Id: I8be9501f34073983b7f3d040780b749a7366e3d2 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6733547 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Yuxin Hu <yuxinhu@google.com>
Kimmo Kinnunen 93b659f9 2025-07-04T12:35:29 Remove PoolAllocator push/pop feature PoolAllocator would manage a stack of memory pools upon client calling push() and pop(). This made the code unnecessarily complicated. The feature was only used with nesting of one, to mark the memory unused after a shader compile. Fix by removing the push/pop feature. Instantiate PoolAllocator in places the previous push() was and uninstantiating instead of previous pop(). This removes the feature where the PoolAllocator would hold on to the allocated memory in order to reuse it. This is seen as a progression: the allocator is held by the compiler, the compiler is held by the shader and each shader typically see only one compile. Thus the free pages were just leaking unused until the shader was destroyed. Instead, destructing the PoolAllocator instead of pop() will donate the memory back to platform/OS, where it is likely more useful. To preserve existing Vulkan behavior, add PoolAllocator::reset() which would mark the memory unused but leave them reserved for the PoolAllocator. Removes UB where PageHeader::nextPage would be accessed after ~PageHeader. Bug: angleproject:429513168 Change-Id: I21e58b46e0887380db3a2cab5ce22f0042cfae9e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6701153 Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Kimmo Kinnunen 1443a451 2025-07-04T16:18:30 Regenerate the parsers with current cl format The format tool has changed since last parser generation. This causes overly many changes when downstream regenerates the parsers. The diff formatting fails for flex and bison generated content. Format them in full always. Bug: angleproject:429588208 Change-Id: I7a4b8e966e9502920fc25f5c1c5bf80246f5cf0b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6705451 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com>
Corentin Wallez 5d51bd19 2025-07-01T15:40:10 Translator: Handle arrays of samplers in FlagSamplersWithTexelFetch This was found while running tests for Dawn's "sized binding arrays" (including arrays of textures) with ANGLE + Swiftshader. Adds a test that ASSERTs without the change. Bug: chromium:411573957 Change-Id: Ib7fad0f5c0b4f3aa70e6af5bf5977b82956e1986 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6695013 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Alexey Knyazev 56e9bfff 2025-06-19T00:00:00 GL: Remove obsolete macOS workarounds Bug: angleproject:427600175 Change-Id: I8bfc062b71f1a34de09cbf76b504293ccd081fe0 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6676022 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com>
Alexey Knyazev beae1b4a 2025-06-19T00:00:00 Metal: Remove obsolete workarounds Removed Intel-specific Metal workarounds that are not used on macOS 12 and later. Removed a redundant version check. Bug: angleproject:427600175 Change-Id: I34c7e53108f7e030512c9436ab2b9ae38ad17946 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6672312 Commit-Queue: Alexey Knyazev <lexa.knyazev@gmail.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi f48fcb46 2025-06-19T07:11:48 Revert "Metal: Wraparound int arithmetic for add, sub" This reverts commit 2f6f0514563d81e439e04ba09eee9e8a1d6ca8d1. Reason for revert: Leads to MSL compiler errors: > error: non-const reference cannot bind to vector element Bug: angleproject:382215759 Original change's description: > Metal: Wraparound int arithmetic for add, sub > > Implement wraparound int arithmetic for > int, ivec +, -, +=, -=, ++, --. > > All of the added tests pass also without the changes. > > Bug: angleproject:382215759 > Change-Id: I7605582ab2931e63010c34fba474797f7edd19f9 > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6578729 > Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Reviewed-by: Geoff Lang <geofflang@chromium.org> > Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com> Bug: angleproject:382215759 Bug: chromium:425746892 Change-Id: I831d96c6ee7b8bcdeaa08fbf32d430def87a3c30 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6656001 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Martin Arjovsky a2b84cc0 2025-06-17T09:47:58 Fix bug in translator fuzzer. We were incorrectly checking a type against a spec. bug: angleproject:425580663 Change-Id: I7a57852b143ea26a41414015414d43720e2e53b1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6650788 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Kimmo Kinnunen 94ee620d 2025-05-22T10:07:05 Metal: Allow optimization of simple loops Reimplement the feature to avoid undefined behavior of infinite loops. Add EnsureLoopForwardProgress rewrite pass that inserts a volatile variable access to all loops that it cannot analyze as being finite. Detect loops of form `for (; i <op> x; ++i)` as being finite. The <op> can be any of <,<=,>,>=,==, != operator. The i can be int or uint. The ++i can be -- or ++, -=1, +=1. This assumes that backends using the feature emit signed int arithmetic with defined wraparound semantics. Uses volatile write instead of asm("") due to asm not forcing the behavior in some compiler versions. The volatile variable access is defined in C++ as forward progress, and by inheritance this works in MSL. Later commits may remove injectAsmStatementIntoLoopBodies if ensureLoopForwardProgress is appropriate for all use-cases. Bug: angleproject:418918522 Change-Id: Ic9c29f57044b792195386483208632354d24c854 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6575051 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com>
Matthew Denton 005336e4 2025-06-11T17:17:16 WGPU: Basic texture cubes Allows uploading to texture cube faces and sampling from them in shaders using samplerCube. Bug: angleproject:420782526 Change-Id: I45d4370fcc418f39afb225114d13632a78c7c200 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6593999 Reviewed-by: Liza Burakova <liza@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Yuxin Hu 1e613fec 2025-05-27T16:59:26 Compiler: sort uniforms by precision This change reoreders the AST sequences so that uniform declarations with lower precisions are places in front of uniform declarations with higher precisions. This is to prepare for the upcoming change that converts uniforms declared with mediump or lowp to 16-bit and pack them to 32-bit integeter using packHalf2x16(). Bug: b/405795981 Change-Id: I5e1e293399dc8b51b9a6e83115f95beb0c4a7b1b Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6594255 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Roberto Rodriguez af6313f5 2025-06-02T10:31:27 Check redundant statements for comma op in MSL translator. Bug: angleproject:421891565 Change-Id: I0475476f360f103a3da4df263dae9c604a8bfef5 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6610091 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Kimmo Kinnunen <kkinnunen@apple.com> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com>
Shahbaz Youssefi bc2c85a6 2025-05-27T11:40:33 Translator: Set correct qualifiers on user-declared gl_PerVertex Bug: angleproject:409439499 Change-Id: Ia4bc1335d977f2b138af877e5d698914b182ff52 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6592612 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com>
Kimmo Kinnunen 2f6f0514 2025-05-22T15:59:19 Metal: Wraparound int arithmetic for add, sub Implement wraparound int arithmetic for int, ivec +, -, +=, -=, ++, --. All of the added tests pass also without the changes. Bug: angleproject:382215759 Change-Id: I7605582ab2931e63010c34fba474797f7edd19f9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6578729 Auto-Submit: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com>
Roberto Rodriguez 157a99aa 2025-05-20T23:59:29 Update logic for redeclaring gl_FragDepth. Update ParseContext logic to handle redeclaration of gl_FragDepth. Bug: angleproject:419066563 Change-Id: I774d036922ec75ca151f1c871afc243ca09a3061 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6573532 Reviewed-by: Kimmo Kinnunen <kkinnunen@apple.com> Commit-Queue: Kimmo Kinnunen <kkinnunen@apple.com> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Matthew Denton 96c9f065 2025-05-15T19:22:29 WGPU: Flip y for the default framebuffer Bug: angleproject:389145696 Change-Id: I0d527ad3dc24dbca7e9d914b03edacdc257a568f Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6477137 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org>
Shahbaz Youssefi 671f55d8 2025-04-03T22:54:44 Vulkan: Fix texelFetch(externalSampler) behavior GLES expects YUV decoding to happen with texelFetch when an external sampler is used, but texelFetch's translation (OpImageFetch) does not do such a thing. A transformation is added to replace that with a texture call at the right coordinate. Bug: angleproject:405149439 Change-Id: I3a8d07a6399705ec07718b38085ee4bc1ad2af6c Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6431570 Reviewed-by: Cody Northrop <cnorthrop@google.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Matthew Denton cac1e824 2025-04-29T18:27:36 WGSL: Output driver uniform and UBO structs This is the WGSL half of the change to implement driver uniforms. Driver uniforms are implemented as a UBO and reuse the default set of driver uniforms. User-provided UBOs don't yet have variables outputted for them. This requires moving MSL's ReduceInterfaceBlocks to the tree_ops dir in order to change interface block definitions into struct definitions. Bug: angleproject:389145696 Change-Id: I27f3837b3d115f2ffac66cc545f3b60ca9f01cb6 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6477564 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Matthew Denton <mpdenton@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Matthew Denton 4b69ba93 2025-04-28T16:45:56 WGPU: upload texture bind groups and configured samplers GLSL samplers are split into separate sampler/texture variables in WGSL. Before this CL, the WGSL translator generated shaders that look like: @group(1) @binding(@@@@@@) var ANGLE_sampler_samp2D : sampler; @group(1) @binding(@@@@@@) var ANGLE_texture_samp2D : texture_2d<f32>; @group(1) @binding(@@@@@@) var ANGLE_sampler_sampCube : sampler; @group(1) @binding(@@@@@@) var ANGLE_texture_sampCube : texture_cube<f32>; This CL replaces those with actual binding numbers @group(1) @binding(0) var ANGLE_sampler_samp2D : sampler; @group(1) @binding(1) var ANGLE_texture_samp2D : texture_2d<f32>; ... Such that @binding(n*2) is the WGSL sampler variable corresponding to the n-th GLSL sampler and @binding(n*2+1) is the WGSL texture variable corresponding to the n-th GLSL sampler. This CL then generates binding group layouts matching the above, and uploads textures and configured samplers in bind groups. This makes some of the deqp_gles2 tests 2d texture tests pass, though some fail because they need a flipped y coordinate. Not yet supported: 1. arrays of samplers 2. shadow samplers 3. cube textures Bug: angleproject:389145696 Change-Id: I2ab18ae5ebb4d1289101266bd9451576aa04ce2a Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6382272 Reviewed-by: Liza Burakova <liza@chromium.org> Auto-Submit: Matthew Denton <mpdenton@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 828bec45 2025-04-26T16:12:56 Vulkan: Fix SPIR-V gen for constant constructors ... inside non-constant constructors of different basic type. The SPIR-V generator proactively casts the elements of the inner constructor to the type of the outer constructor, but it was mistakenly using the original inner constructor's type id for the constant instead of the cast type id. Bug: chromium:413775106 Change-Id: I95e8b1dcc1ba36777eb4475042f0d106a6abe109 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6489041 Reviewed-by: Cody Northrop <cnorthrop@google.com> Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com>
Shahbaz Youssefi 555038b8 2025-04-23T12:47:02 Translator Fuzzer: Enable required options Bug: chromium:409603851 Change-Id: I73067fcd3855f6cdcb6eea05b06bafa45d23fb19 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6482917 Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Yuly Novikov <ynovikov@chromium.org> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Shahbaz Youssefi f355e2b3 2025-04-15T18:58:25 Vulkan: Remove preferDriverUniformOverSpecConst This was practically true for every vendor on Android (where rotation matters). For Qualcomm, it was also true due to a bug in version checking and didn't seem to be causing any concerns. Where pre-rotation is supported, it is better to enable this feature to avoid excessive pipeline creation. This change removes the feature and makes sure ANGLE always uses uniforms for rotation instead of spec consts. While technically this may have an adverse effect on platforms that never need pre-rotation, the ability is retained for all vendors since pre-rotation is finding its way into more platforms and would likely eventually be needed everywhere anyway. Bug: angleproject:42265878 Bug: angleproject:42262166 Change-Id: I4b64c04da46db08cfdd44b60789b66d93d8e8b17 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6459025 Reviewed-by: Amirali Abdolrashidi <abdolrashidi@google.com> Reviewed-by: mohan maiya <m.maiya@samsung.com> Auto-Submit: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Cody Northrop <cnorthrop@google.com> Reviewed-by: Cody Northrop <cnorthrop@google.com>
Yuly Novikov 5dd05578 2025-04-14T07:34:57 Revert "GL: Allow untranslated shaders to pass through on GLES" This reverts commit 4e77552b86a89b449ada6d6c18f84285f5812b1d. Reason for revert: breaks ChromeOS and fuzzers Bug: angleproject:398857482 Original change's description: > GL: Allow untranslated shaders to pass through on GLES > > Add an EGL extension EGL_ANGLE_create_context_passthrough_shaders which > uses the NULL translator and passes the original shader to the driver. > The parser is still used for shader reflection. > > Bug: angleproject:398857482 > Change-Id: I7c5fcc318c7e11931f78c08dcbf4764bf77d397d > Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6297527 > Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> > Commit-Queue: Geoff Lang <geofflang@chromium.org> Bug: angleproject:398857482, angleproject:410423936 Bug: chromium:410114655, chromium:410100607, chromium:410121218 Bug: chromium:410052365, chromium:410290507, chromium:410178288 No-Presubmit: true Change-Id: I45b01960637a1cda05d21a7df6d07465f6a8f5e9 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6448984 Bot-Commit: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Auto-Submit: Yuly Novikov <ynovikov@chromium.org> Commit-Queue: Rubber Stamper <rubber-stamper@appspot.gserviceaccount.com> Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Geoff Lang 4e77552b 2025-02-24T18:04:32 GL: Allow untranslated shaders to pass through on GLES Add an EGL extension EGL_ANGLE_create_context_passthrough_shaders which uses the NULL translator and passes the original shader to the driver. The parser is still used for shader reflection. Bug: angleproject:398857482 Change-Id: I7c5fcc318c7e11931f78c08dcbf4764bf77d397d Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6297527 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Shahbaz Youssefi 02ae054c 2025-04-07T13:48:29 Translator: Fix init of inactive output variables Bug: chromium:398401939 Change-Id: I0df494b945b8d0e805a62cf7645d06bf233f36ca Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6438495 Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Cody Northrop d27a1fd0 2025-04-03T19:59:20 Compiler: Allow denorm float values in the lexer This adds an option to preserve denorm values in the lexer, skipping the explicit zero conversions for below range floats. There are applications in the wild that expect to be able to use denorm float values. They are typically immediately converted to integer values, not used in floating point operations. The option is only enabled for Vulkan backends. Test: FloatLexTest, DenormFloatsToIntValues, app traces Bug: b/406827038 Change-Id: Iab5a1a69a540b78ccbce8ea90b532d2d4976e29e Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6432237 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Cody Northrop <cnorthrop@google.com>
Geoff Lang 0e28c030 2025-04-02T17:26:28 ScalarizeVecAndMatConstructorArgs on WGSL shaders. WGSL does not support mixing types in the constructors of vectors and matrices. Use the ScalarizeVecAndMatConstructorArgs traverser to scalarize these constructors and update it to cast to the destination type if the inputs are not in that type already. Bug: angleproject:395659799 Change-Id: I4c270f2b18e79393e3c187916f33ad719bc88a63 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6427067 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Kenneth Russell fd722f9a 2025-03-31T23:23:26 Avoid emitting SampleCmp in discontinuous loops. Apply fix by Mikkel Simonsen from Unity to use texture level 0 when using textureLod with a shadow sampler inside a discontinuous loop. Bug: angleproject:406190894 Change-Id: If02b687f15ec4c3e8bd1b8eece72098d498e2ce1 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6421496 Auto-Submit: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Kenneth Russell <kbr@chromium.org>
Neil Zhang 0d1478f3 2025-03-27T18:40:05 Fix dFdx() dFdy() in frag shader not work as expected When render to FBO0, Y axis will flipping, Angle's compile module will handle this, but only when glsl version > 300, that's a mistake as glsl 100 also have dFdx/dFdy when have GL_OES_standard_derivatives. Bug: angleproject:406819301 Change-Id: I656cfcfdeff1e80db374cfa083578adda6064e69 Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/6407233 Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org> Reviewed-by: Charlie Lao <cclao@google.com> Auto-Submit: Neil Zhang <Neil.Zhang@arm.com> Commit-Queue: Charlie Lao <cclao@google.com>
kmx.io     mail     discord kmxgit v0.4.0