kc3-lang/angle/src/tests/compiler_tests

src/tests/compiler_tests

Log

Author Commit Date CI Message
Austin Eng 153e0224 2018-12-07T13:31:31 Add gl_DrawID emulation for ESSL3 with ANGLE_multi_draw Bug: chromium:890539 Change-Id: I3d9a9429608afb454f076013f3ed844e115c56cc Reviewed-on: https://chromium-review.googlesource.com/c/1368784 Reviewed-by: Kai Ninomiya <kainino@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Austin Eng <enga@chromium.org>
Qin Jiajia 19f2f9eb 2018-12-07T15:10:48 Mark variable to call 'length()' as statically used Bug: angleproject:3008 Change-Id: I8779706c667f49fddb0a741f5bf31c4af54db3e7 Reviewed-on: https://chromium-review.googlesource.com/c/1367404 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jiajia Qin <jiajia.qin@intel.com>
Jamie Madill b980c563 2018-11-27T11:34:27 Reformat all cpp and h files. This applies git cl format --full to all ANGLE sources. Bug: angleproject:2986 Change-Id: Ib504e618c1589332a37e97696cdc3515d739308f Reviewed-on: https://chromium-review.googlesource.com/c/1351367 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Geoff Lang 6aab06e0 2018-11-20T14:04:11 Don't allow unary operators on void types. BUG=900058 Change-Id: Ice8976ddd9cbc9c90162cf7fae2b8f91ddcfcef3 Reviewed-on: https://chromium-review.googlesource.com/c/1344590 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Geoff Lang <geofflang@chromium.org>
Austin Eng 7cf9cd24 2018-10-09T15:27:32 Add flags to shader translator to emulate gl_DrawID Adds support for translating gl_DrawID for implementation of GL_ANGLE_multi_draw. Currently the change only supports and allows emulation of the draw id using a uniform variable named `gl_DrawID`. This uniform is mapped in the translated shader to a hashed name that does not use the gl_ namespace Bug: chromium:890539 Change-Id: I08a246ca911e88e733ccdf22f1ed69dcae948e05 Reviewed-on: https://chromium-review.googlesource.com/c/1271957 Commit-Queue: Austin Eng <enga@chromium.org> Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Yizhou Jiang c4533eae 2018-09-19T15:23:29 Enable ANGLE_texture_multisample in glsl in es 3.0 Enable gsampler2dMS, texelFetch, textureSize in glslang in es 3.0 if ANGLE_texture_multisample is supported. Bug: angleproject:2275 TEST=SamplerMultisampleEXTTest.TextureMultisampleEXTEnabled Change-Id: Ibfa367970db3ae790f3822e57eb50090843dc6db Reviewed-on: https://chromium-review.googlesource.com/c/867521 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Yizhou Jiang <yizhou.jiang@intel.com>
Olli Etuaho 5476e805 2018-10-05T13:23:03 Fix constructing void array zero nodes Correctly sized void arrays can be needed after parsing has recovered from an error and the code is trying to evaluate the constant value of a node. Since now we just have a generic EOpConstruct op instead of different ops for different types, we can simply remove the special handling for void arrays in CreateZeroNode to create the arrays in the correct size. BUG=chromium:890581 TEST=angle_unittests Change-Id: I48d96c9ef1d695cd8583a845fd4bd24a7aaf535c Reviewed-on: https://chromium-review.googlesource.com/c/1264515 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 1dfd8ae2 2018-10-01T15:59:59 Improve test coverage of constant folding Clean up some TODOs in the code as well as add tests to make sure that ANGLE treats expressions indexing into constant arrays as constant expressions in different contexts. This complements the existing tests in ConstantFoldingTest. BUG=angleproject:2298 TEST=angle_unittests Change-Id: I779bf03891f7d06f14d293e69101c05d7dbf57b6 Reviewed-on: https://chromium-review.googlesource.com/1254067 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho a78092cb 2018-09-26T14:16:13 Support ESSL 3.00 EXT_blend_func_extended shaders This adds support for the index layout qualifier that's used in EXT_blend_func_extended to set whether a fragment output should be bound to the primary or secondary blend source color. Output locations are now validated correctly so that two outputs can have the same location as long as they have a different index. Some tests are fixed to allow this. BUG=angleproject:1085 TEST=angle_unittests Change-Id: I1de3ad1406398952287791eca367562bed59d380 Reviewed-on: https://chromium-review.googlesource.com/1245982 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Qin Jiajia 017bda42 2018-09-14T18:11:50 Fix the ASSERT error This change adds shader type checking before entering initializeOutputVariables. BUG: angleproject:2821 Change-Id: Ib931031f2fc187f1f2a1821a09664bbe172a5e90 Reviewed-on: https://chromium-review.googlesource.com/1226229 Commit-Queue: Jiajia Qin <jiajia.qin@intel.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho f0d0408a 2018-08-28T16:02:13 Use OES_texture_storage_multisample_2d_array There's an OES extension for multisample texture arrays, OES_texture_storage_multisample_2d_array. Change references from ANGLE_texture_multisample_array to the native extension in the shader compiler. ANGLE still needs to have robust behavior for out-of-range texel fetches that's not found in the original extension, but this does not need to be spelled out in the extension spec. BUG=angleproject:2775 TEST=angle_unittests Change-Id: Ie80ae767cc92ccaf7389af28789f45547f86978f Reviewed-on: https://chromium-review.googlesource.com/1193266 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 8ca60805 2018-08-23T14:10:02 Add 2D MS array sampler support to compiler This also places textureSize(gsampler2DMS) correctly in the ESSL 3.10 builtins instead of ESSL 3.00 builtins. BUG=angleproject:2775 TEST=angle_unittests Change-Id: Ieb0f7a7424a5558a5569af6d4fcbcc9b12ec9840 Reviewed-on: https://chromium-review.googlesource.com/1186466 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho da41ac69 2018-07-19T16:45:32 Fix decorating ViewID_OVR in HLSL output ViewID_OVR should not be decorated in HLSL output since it is an internal variable. Make sure that DecorateVariableIfNeeded() is used for varyings instead of just Decorate() so that the internalness is checked correctly and ViewID_OVR doesn't get decorated. This avoids possible name conflicts between the internal ViewID_OVR and any user-defined variables named ViewID_OVR. BUG=angleproject:2062 TEST=angle_end2end_tests, angle_unittests Change-Id: I9ed9876d4b2c760e7a11b0b270a2190993e840e5 Reviewed-on: https://chromium-review.googlesource.com/1143398 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho d8b1c5c5 2018-06-20T12:08:46 Return ImmutableString from ArrayString() This makes the compiler a few kilobytes smaller, and prepares getting rid of TString altogether. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I93a003fe27b99bef72f872fa1066e2e108f934c5 Reviewed-on: https://chromium-review.googlesource.com/1107713 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Jamie Madill 4da0d315 2018-06-20T11:46:38 Vulkan: Handle embedded struct uniforms. Also known as nameless structs. Uniform structs without a struct name would not be parsed correctly. This fixes the bug by adding a tree transformation. The transformation gives an internally scoped name to the embedded struct. Bug: angleproject:2665 Change-Id: I43e4dad7d9ad64a40e382066bb136e4f8f719797 Reviewed-on: https://chromium-review.googlesource.com/1101566 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Olli Etuaho c3907efa 2018-06-08T13:03:15 Always use custom float parsing in GLSL We now always use custom parsing code for parsing floats in GLSL shaders. Previously this code was only used in corner cases that stringstream parsing did not handle according to the GLSL spec. This is slightly faster in compiler perftests, and results in a smaller binary as well. Some new test cases are added to make sure that the custom float parsing behaves correctly. BUG=chromium:849245 TEST=angle_unittests, angle_perftests Change-Id: I2a88ec6a8b427016e34519d72bc98216947a4c64 Reviewed-on: https://chromium-review.googlesource.com/1092697 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Jeff Gilbert 79216620 2018-05-29T19:19:54 Use stringstream with locale override. Add test for compiling float literals in locales with comma decimal separators. Handle inexplicable test setlocale failure on Android,Linux. (Require success on other platforms) Skip setting the locale on Android, which is always C locale in C++, but for some reason std::locale::classic isn't implemented as a no-op. Bug: angleproject:2607 Test: angle_unittests Change-Id: I7c97cb56c01335db46f532fb8af3f9a4f2a30564 Reviewed-on: https://chromium-review.googlesource.com/1077789 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 67d2cd07 2018-05-07T10:27:22 Issue a clearer warning about negative modulus operands This also adds unit tests for negative modulus operands. BUG=chromium:839468 TEST=angle_unittests Change-Id: I6ab5959ba4f7045d2bde71d246695ef0983c5608 Reviewed-on: https://chromium-review.googlesource.com/1046055 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho f827123d 2018-04-25T13:08:24 Handle negative float to uint conversion robustly Converting a negative float to uint is undefined in the GLSL ES 3.00.6 spec. However, it improves portability if we don't trigger undefined results in C++ in this case. To do this, we cast negative floats first to signed integer before casting them to unsigned integer. We also issue a warning about an undefined conversion in case a negative float was converted to uint. BUG=chromium:835868 TEST=angle_unittests Change-Id: I9835a739ec80699d420a4f91a3bfa112c9a13604 Reviewed-on: https://chromium-review.googlesource.com/1026681 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho bb52c523 2018-04-06T10:22:22 Invariant declaration doesn't make a variable active Invariant declarations didn't affect static use before, but now they are also skipped in CollectVariables so an invariant declaration is not enough in itself to mark a variable as active. This fixes an assert in CollectVariables. BUG=chromium:829553 TEST=angle_unittests Change-Id: I3e51d2916f091bcc283af136a4abc846ff71447d Reviewed-on: https://chromium-review.googlesource.com/999532 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 5ae64c94 2018-04-06T11:27:03 Fix writing hex values in ImmutableStringBuilder The old code was accidentally using letters offset by 10 when writing out hex values >= 10. Now the letters a to f are used as they should. This is one issue that changed shader output when ImmutableString was introduced, so it is a potential cause for a regression detailed in bug 824062, though this has not been verified. BUG=chromium:824062 TEST=angle_unittests Change-Id: Idb871dffba32a3ab20df0fe17b4b1a98ec00b7fa Reviewed-on: https://chromium-review.googlesource.com/999480 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 4002e92a 2018-04-04T16:55:34 Guard traversers used during parsing against stack overflow Traversers used during parsing can be vulnerable to stack overflow since the AST has not yet been validated for max depth. Make sure to check for traversal depth in traversers used during parsing. We set the maximum traversal depth in ValidateGlobalInitializer and ValidateSwitchStatementList to 256, which matches the default value for validating general AST complexity. The depth check is on regardless of compiler options. In case the traversers go over the maximum traversal depth, they fail validation. BUG=angleproject:2453 TEST=angle_unittests Change-Id: I89ba576e8ef69663ba35d7b9050a6da319f1757c Reviewed-on: https://chromium-review.googlesource.com/995795 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 5fec7ab2 2018-04-04T11:58:33 Identify functions by unique id in BuiltInFunctionEmulator Now that unique ids of all builtins are compile-time constants, we can use them to look up functions in BuiltInFunctionEmulator. This is simpler than using a custom struct with the name and parameters for identifying functions. This requires that we store a reference to a TFunction in those TIntermUnary nodes that were created based on a function. This decreases shader_translator binary size by about 6 KB on Windows. BUG=angleproject:2267 BUG=chromium:823856 TEST=angle_unittests Change-Id: Idd5a00c772c6f26dd36fdbbfbe161d22ab27c2fe Reviewed-on: https://chromium-review.googlesource.com/995372 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho fd643283 2018-04-04T16:12:09 Disallow case statements nested in blocks The GLSL ES 3.00.6 spec is a bit unclear on this, but it does disallow case statements "inside control flow". GLSL ES 3.10 clarifies this and disallows any nesting of case or default labels within other statements. BUG=angleproject:2452 TEST=angle_unittests Change-Id: I289bb39abb5227eab7117638af388b0a57dc5dd8 Reviewed-on: https://chromium-review.googlesource.com/995478 Commit-Queue: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 59c5b897 2018-04-03T11:44:50 Validate gl_FragData and gl_FragColor access after parsing After this simply declaring both variables invariant is not treated as static use. This simplifies ParseContext a bit, but the effect on compiler performance tests seems marginal. BUG=angleproject:2450 TEST=angle_unittests Change-Id: Ib90cb1d2bd1331542d1cd37732f24efb7833036a Reviewed-on: https://chromium-review.googlesource.com/992112 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 94bbed1e 2018-03-20T14:44:53 Collect static use information during parsing We now collect metadata for variables in the symbol table. The metadata is stored in a map using the variable unique id as a key, so we can store the variables themselves as constexpr while still having dynamic metadata. For now we collect whether a variable is statically read or written. This can be used to more accurately determine whether a variable is statically used, but can also enable more optimizations in the future, such as pruning variables that are never read or folding variables that are never written after initialization. The collection is done during parsing, so that nothing is pruned from the AST before the static use is recorded. Static writes are flagged in ParseContext::checkCanBeLValue, as that function is already called for all variables that are written. Static reads are flagged whenever there's an operation that requires a variable to be read. This includes: * Unary and binary math ops * Comma ops * Ternary ops * Assignments * Returning the variable * Passing the variable as an in or inout argument to a function * Using the variable as a constructor argument * Using the variable as an if statement condition * Using the variable as a loop condition or expression * Using the variable as an index * Using the variable as a switch statement init expression In case there are statements that simply refer to a variable without doing operations on it, the variable is being treated as statically read. Examples of such statements: my_var; my_arr[2]; These are a bit of a corner case, but it makes sense to treat them as static use for validation purposes. Collecting correct static use information costs us a bit of compiler performance, but the regression is on the order of just a few percent in the compiler perf tests. BUG=angleproject:2262 TEST=angle_unittests, angle_end2end_tests Change-Id: Ib0d7add7e4a7d11bffeb2a4861eeea982c562234 Reviewed-on: https://chromium-review.googlesource.com/977964 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 89398b65 2018-03-21T17:30:50 Avoid mangled name comparisons of 3-parameter functions The hash values used for looking up built-ins now encode whether the mangled name contains arrays, structs or interface blocks in its parameters list. This is written in the most significant bit of the hash value. We check this bit at the start of built-in lookup and if the bit is set we exit early. After that we know that the lookup name doesn't contain array, struct or interface block parameters. When we find a hash that matches a hash of a built-in function, we now know 3 things: 1) the length of the mangled name matches 2) the open parentheses in the mangled name matches 3) the lookup doesn't contain array, struct or block parameters. Additionally, we have an if statement checking whether the function name matches. Collisions are only possible with functions that 1) have the same name 2) have the same number of parameters With these preconditions we can check beforehand whether collisions are possible for 3-parameter functions. If there are no collisions, we don't need to compare the full mangled name. This is similar to what was already being done with functions that had 0 to 2 parameters. This reduces shader_translator binary size by around 4 KB on Windows. Besides increased complexity, the tradeoff is that an exhaustive search of hash values for possible 3-parameter combinations is costly, so the gen_builtin_functions.py code generation script now takes around one minute to run on a high-end workstation. Due to this, the script now exits early if it detects it has already been run with the same inputs based on a hash value stored in builtin_symbols_hash_autogen.txt. BUG=angleproject:2267 BUG=chromium:823856 TEST=angle_unittests Change-Id: I3ff8c6eb85b90d3c4971ac8d73ee171a07a7e55f Reviewed-on: https://chromium-review.googlesource.com/973372 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho b1de5a7e 2018-03-28T14:07:57 Prevent stack overflow due to recursive swizzle of an l-value Long chains of recursive swizzling could previously cause a stack overflow in checkCanBeLValue. Fold recursive swizzling when it is parsed to prevent this. BUG=angleproject:2439 TEST=angle_unittests Change-Id: I83b4c27442185709f6762d5ec23b93244010da05 Reviewed-on: https://chromium-review.googlesource.com/983593 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 3b678745 2018-03-21T12:59:19 Use a specialized hash function for mangled names The hash values used for looking up built-ins now encode the string length and the location of parentheses as six-bit values, so that we don't need to check for these if the hash matches. This decreases shader_translator binary size on Windows by around 10 KB. BUG=angleproject:2267 BUG=chromium:823856 TEST=angle_unittests Change-Id: If8c28e1c8851750633509ec6273f556e06e91cd1 Reviewed-on: https://chromium-review.googlesource.com/973243 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho e79d0f86 2018-03-19T11:55:14 Rely on hash to check for some mangled name matches When we are looking up a function with only a few parameters, we can optimize the lookup by relying on the information encoded in the hash value. There's often only one list of parameters with the same function name and mangled name length that results in a matching hash, so we don't actually need to compare the full mangled name. We can just compare 1) the hash value of the mangled name 2) the mangled name length 3) the function name to make sure that the mangled name matches the mangled name of the function. This decreases the binary size since we don't need store as many mangled names of built-in functions. Effect on symbol lookup speed is marginal. BUG=angleproject:2267 BUG=chromium:823856 TEST=angle_unittests Change-Id: I3ef41d943209509d4e8e6ece14ebad7e2677abc6 Reviewed-on: https://chromium-review.googlesource.com/973242 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho c26214de 2018-03-16T10:43:11 Move AST utilities to a subdirectory Move AST related utilities to compiler/translator/tree_util. BUG=angleproject:2409 TEST=angle_unittests Change-Id: I7567c2f6f2710292029263257c7ac26e2a144ac8 Reviewed-on: https://chromium-review.googlesource.com/966032 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 40b1c770 2018-03-19T10:29:37 Fix asserts related to writing nameless parameters in HLSL We rely on TSymbol::name() to generate names also for nameless parameters. Fix asserts to reflect this. BUG=chromium:823041 TEST=angle_unittests Change-Id: Ie9b8253a150e79965bf85d8a7f36643ada6c54cc Reviewed-on: https://chromium-review.googlesource.com/968242 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 115b2c4a 2018-03-15T17:46:29 Remove desktop GLSL builtins from symbol lookups Desktop GLSL builtins can be accessed through functions in BuiltIn_autogen.h. They don't need to be included in symbol table lookups. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I8ba188a0d2584353e34159c2732c9e2bd420c168 Reviewed-on: https://chromium-review.googlesource.com/964447 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho b391ec40 2018-03-12T17:04:59 Generate code for looking up built-ins Instead of storing built-ins in a std::unordered_map, we now generate a series of switch statements using the hash value of the look-up string. This works similarly to earlier implementation of looking up unmangled built-ins. Those built-ins that need to be initialized at run-time are stored as member variables of TSymbolTable. This increases compiler init performance significantly, as well as increasing compiler perf test scores around 1-2%. Binary size is larger than before though. BUG=angleproject:2267 TEST=angle_unittests Change-Id: If1dcd36f0d2b30c2ed315cdcf6e831ae9fe70c94 Reviewed-on: https://chromium-review.googlesource.com/960031 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 4121799f 2018-03-15T11:15:33 Clean up switch/case pruning code There was some duplicate switch/case pruning in the code in PruneEmptyCases and RemoveNoOpStatementsFromTheEndOfSwitchStatements. Combine the functionality of both AST transformations into PruneEmptyCases and remove the other transformation. The tests are improved to better cover the full functionality. BUG=angleproject:2402 TEST=angle_unittests, angle_end2end_tests --gtest_filter=*Switch* Change-Id: Ib74b6b9b455769ea15650e9653a9c53635342c49 Reviewed-on: https://chromium-review.googlesource.com/964081 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho d4bd963f 2018-03-08T16:32:44 Don't use TIntermSymbol nodes for function parameters Parameter nodes are not needed - it's simpler to just create a TVariable object for each parameter when the TFunction is initialized. With this change we also store only one object per each parameter type used in built-in functions, instead of one array of TConstParameter entries for each unique parameter sequence. This simplifies code and reduces binary size and compiler memory use. Compiler perf does not seem to be significantly affected. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I2b82400dd594731074309f92a705e75135a4c82c Reviewed-on: https://chromium-review.googlesource.com/955589 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 0b0dcbc1 2018-03-14T11:33:22 Prune empty cases and switch statements The translator already prunes no-ops and unreferenced variables, and this may result in case statements that are followed by nothing. Since the last case statement in as switch statement must always contain a statement to be valid GLSL, the translator must not leave empty case statements in place. They're now being pruned by the PruneEmptyCases AST transformation. This improves on the earlier RemoveEmptySwitchStatements AST transformation that did address empty switch statements but could not remove them if they had a case statement inside. BUG=angleproject:2402 TEST=angle_unittests Change-Id: Ieb9598a744078e45226d8fb7266d877f7835cf0c Reviewed-on: https://chromium-review.googlesource.com/962181 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho e600c0aa 2018-03-02T11:23:29 Use non-human-readable mangled names for types The new mangled name format is as follows: The first character is a hex digit from 0 to F that encodes vector or matrix size. For scalars, structs etc. the character is 0. Then, if it's a struct, the mangled name continues with "{s", followed by mangled names of fields, and ends with "}". If it's an interface block, the mangled name continues with "{i", followed by mangled names of fields, and ends with "}". If it's anything else, the second alphabetic character encodes the basic type. Characters are assigned to basic types in the enumeration order. If it's an array, the mangled name has a suffix [array_size]. This saves a few kilobytes from the binary size. The effect on symbol lookup speed seems mostly marginal. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I26e65dcb48c3478df9a719ffff9c15f2fd12e293 Reviewed-on: https://chromium-review.googlesource.com/945910 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 065aa863 2018-02-22T15:30:27 Generate code for unmangled name lookup Instead of using an std::map at each symbol table level, use the gen_builtin_symbols.py script to build a function to query unmangled names. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I4f1cf1df1f50fe9d909f3249150ee002ee6efb61 Reviewed-on: https://chromium-review.googlesource.com/931885 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho d05f964b 2018-03-05T12:13:26 Last case in switch statement can't be empty in ESSL 3.10 This is based on recent discussion in Khronos, though public specs have not yet been updated to reflect this. BUG=angleproject:2388 TEST=angle_unittests Change-Id: I66a0d03b3c2bb9740772a813b543f8f6c6bb2a28 Reviewed-on: https://chromium-review.googlesource.com/947977 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 92e4e079 2018-03-02T13:34:56 Test shader builtin accessibility from different stages Test that fragment, vertex or compute shader exclusive builtins are not accessible from another shader stage. This will protect against regressions when refactoring the symbol table. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I50f97cbde3f1c5cb4584568e3008f1dab724b769 Reviewed-on: https://chromium-review.googlesource.com/941953 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 28cf920e 2018-03-02T15:54:56 Fix checking texture function offset There was a mistake in one of the texture offset function names being checked. Fix it and add a test making sure that a similar mistake cannot slip in again. BUG=angleproject:2387 TEST=angle_unittests Change-Id: I2dddfbdc2836f9e7e94ca6573a8f3689eecf7eb2 Reviewed-on: https://chromium-review.googlesource.com/946254 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho e765e215 2018-03-01T12:00:48 Test that ESSL 300 standard derivatives compile successfully These functions have the same name in the ESSL 100 extension, creating a possibility for conflicts. This adds a convenient way to test for regressions when refactoring the symbol table. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I9716524a0ea85064ba5f3ee8e7adfbdfc59d257e Reviewed-on: https://chromium-review.googlesource.com/945989 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho d6fef528 2018-03-01T13:41:13 Shorten mangled names of types This will reduce binary size by a few kilobytes once we store mangled names of functions as constexpr strings. This still keeps mangled names human readable. BUG=angleproject:2267 TEST=angle_unittests Change-Id: Ia63980bd511c701a93e2c5e247d24578c7d7f332 Reviewed-on: https://chromium-review.googlesource.com/943101 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 3582c0e2 2018-02-22T14:38:09 Add a test for redefining an unavailable built-in This is going to guard against regressions when symbol table init is optimized. TEST=angle_unittests BUG=angleproject:2267 Change-Id: Iec7b9814f4af215dc0b3e8a39c8cfab004451836 Reviewed-on: https://chromium-review.googlesource.com/931883 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Qin Jiajia 76bf01d7 2018-02-22T14:11:34 Fix that readonly buffer variable can be assigned This change will add memory qualifier checking for binaryNode in case the memory qualifier information is lost. BUG=angleproject:1951 TEST=angle_unittests Change-Id: I3f0cfd7d8a059753cf3c982ee0a977b4b0fd0128 Reviewed-on: https://chromium-review.googlesource.com/929877 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Jiawei Shao e41df655 2018-02-09T14:31:39 ES31: Add shader version check for CS and GS This patch adds the missing check on the shader version for compute and geometry shaders, which can fix a bug that ANGLE GLSL compiler doesn't report a compile error when compiling an empty compute or geometry shader in version 100 or 300. This patch also updates the original compiler tests on the check of shader version. In these tests, the compile errors are all caused by illegal layouts instead of shader versions, which is against the purpose of the tests. This patch also fixes an incorrect case that used an empty compute shader in version 300. BUG=angleproject:1442, angleproject:1941 TEST=angle_unittests Change-Id: Ic26bb8eb312dbc0cec6a879997d0ae7a2e625a0f Reviewed-on: https://chromium-review.googlesource.com/910715 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 7c8567a3 2018-02-20T15:44:07 Always add most extension symbols to symbol table An error will be generated either way if the extension symbols are used incorrectly since each use of an extension function or variable checks whether the extension is enabled. We now also track extension in unmangled built-in names, so that redefining built-ins of extensions that are not enabled can be supported. This includes refactoring the shader extension tests to share a common helper class ShaderExtensionTest. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I9cc5e9bd62fa07796e69256a6a9a493531a62446 Reviewed-on: https://chromium-review.googlesource.com/926526 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 029e8ca7 2018-02-16T14:06:49 Add a constexpr constructor for TFunction Access to TFunction parameters is now handled through two new members: a pointer to a parameter array and a parameter count. There's still also a vector pointer in TFunction for adding function parameters one by one. This is used when parsing user-defined functions. TEST=angle_unittests BUG=angleproject:2267 Change-Id: I86987ae56b7cf37f010d0651e9861789050aec2b Reviewed-on: https://chromium-review.googlesource.com/923987 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho e80825ef 2018-02-16T10:24:53 Simplify built-in function node initialization Built-ins with no math op associated with them now have the op code EOpCallBuiltInFunction set. This makes initializing built-in function nodes simpler, since they can always get the op code from the function symbol. We also no longer look for functions in inner scopes, only from the global scope and from built-in functions. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I55a2642f34bb3c8b8f13183c95fa509ec3b9cfdb Reviewed-on: https://chromium-review.googlesource.com/923724 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho fbb1c792 2018-01-19T16:26:59 Store symbol names as a ImmutableString This will enable compile-time initialization of built-in symbols as well as reducing copying strings. Most of the code that deals with names is changed to use ImmutableString where it makes sense to avoid conversions. The lexer/parser now allocate const char pointers into pool memory instead of allocating TStrings. These are then converted to ImmutableString upon entering TParseContext. BUG=angleproject:2267 TEST=angle_unittests, angle_end2end_tests Change-Id: I244d6271ea1ecf7150d4f89dfa388a7745a1150c Reviewed-on: https://chromium-review.googlesource.com/881561 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 1776fd08 2018-01-31T11:46:52 Make use of CreateBoolNode in traversers This is just refactoring to clean up some duplicate code. A new test is added to make sure this doesn't break UnfoldShortCircuitAST. BUG=angleproject:1490 TEST=angle_unittests Change-Id: I108e646f71ba631cbd5dad2055b64e6a30159742 Reviewed-on: https://chromium-review.googlesource.com/894207 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 68981eb5 2018-01-23T17:46:12 Track parameter qualifiers of functions in call nodes We now add a reference to TFunction to all TIntermAggregate nodes where it is possible, including built-in ops. We also make sure that internal TFunctions added in traversers have correct parameter qualifiers. This makes TLValueTrackingTraverser much simpler. Instead of storing traversed functions or looking up builtin functions from the symbol table, determining which function parameters are out parameters can now be done simply by looking it up from the function symbol associated with the aggregate node. Symbol instances are no longer deleted when a symbol table level goes out of scope, and TFunction destructor no longer clears the parameters. They're all either statically allocated or pool allocated, so this does not result in leaks. TEST=angle_unittests BUG=angleproject:2267 Change-Id: I57e5570da5b5a69a98a8778da3c2dc82b6284738 Reviewed-on: https://chromium-review.googlesource.com/881324 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho b60d30f7 2018-01-16T12:31:06 Make TVariable type immutable This enables using constexpr types for built-in variables and some of the variables created in AST transformations. BUG=angleproject:2267 TEST=angle_unittests Change-Id: Ie85b3c9872a071a7c023ced013b14ad91cff7cee Reviewed-on: https://chromium-review.googlesource.com/868134 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho fa886975 2018-01-18T19:24:54 Fix handling array constructor statements in HLSL output Array constructors that are used as a statement by themselves need to be pruned from the AST before writing HLSL output. This fixes an assert in OutputHLSL. BUG=angleproject:2307 TEST=angle_unittests Change-Id: Ib49461a4be173f3856f5a264ac0af8d818a61798 Reviewed-on: https://chromium-review.googlesource.com/874691 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 2c9cc8b6 2018-01-09T16:13:02 Don't duplicate symbol type information in AST nodes Function prototype nodes and symbol nodes already refer to symbols that have type information, so the type doesn't need to be copied to the TInterm* AST node classes. Now type is only stored in those AST node classes that represent other types of expressions. They use a new TIntermExpression base class for this. Since now we may use the TType from builtin symbols directly instead of copying it, building the mangled names of types in the correct memory pool is also required. The code now realizes the types of built-in variables when they get added to the symbol table. BUG=angleproject:2267 TEST=angle_unittests Change-Id: Ic8d7fc912937cb8abb1e306e58c63bb9c146aae9 Reviewed-on: https://chromium-review.googlesource.com/857005 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho ea22b7a5 2018-01-04T17:09:11 Constant fold array indexing and comparison A virtual function to get the constant value of an AST node is added to TIntermTyped. This way a constant value can be retrieved conveniently from multiple different types of nodes. TIntermSymbol nodes pointing to a const variable can return the value associated with the variable, constructor nodes can build a constant value from their arguments, and indexing nodes can index into a constant array. This enables constant folding operations on constant arrays, while making sure that large amounts of data are not duplicated in the output shader. When folding an operation makes sense, the values of the arguments can be retrieved by using the new TIntermTyped::getConstantValue(). When folding an operation would result in duplicating data, the AST can just be left to be written out as is. For example, if the code contains a constant array of arrays, indexing into individual elements of the inner arrays can be folded, but indexing the top level array is left in place and not replaced with duplicated array literals. Constant folding is supported for indexing and comparisons of arrays. In case constant arrays are only referenced through foldable operations, the variable declarations will be pruned from the AST by the RemoveUnreferencedVariables step. BUG=angleproject:2298 TEST=angle_unittests Change-Id: I5b3be237b7e9fdba56aa9bf0a41b691f4d8f01eb Reviewed-on: https://chromium-review.googlesource.com/850973 Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho ea78d2bb 2018-01-09T12:55:27 Prune unreferenced variables with a constructor initializer Treat aggregate constructors as having no side effects, which enables pruning declarators that have a constructor initializer. Some logic in RemoveUnreferencedVariables is fixed to make this work correctly for structs. The bugs were previously not exposed since constructors were treated as having side effects, but now that those can be pruned the logic needs to be correct. BUG=angleproject:2298 TEST=angle_unittests Change-Id: I6fbe61a9e82065196baa29c200bf556fc21d8962 Reviewed-on: https://chromium-review.googlesource.com/856499 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 765924f0 2018-01-04T12:48:36 Fold ternary and comma ops only after parsing is done In case folding a ternary op or a comma op would change the qualifier of the expression, the folding is deferred to a separate traversal step. After this there are no more cases where the type of a TIntermSymbol node needs to differ from the type of the variable it is referring to. There are still some cases where some parts of TIntermSymbol type are changed while keeping the TVariable type the same though, like when assigning array size to gl_PerVertex nodes or sanitizing qualifiers of struct declarations. BUG=angleproject:2267 TEST=angle_unittests, angle_end2end_tests Change-Id: I1501c8d361f5f765f43ca810d1b7248d9e2c5986 Reviewed-on: https://chromium-review.googlesource.com/850672 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 8b5e8fdb 2017-12-15T14:59:15 Replace remaining usage of TName with TSymbol TName used to contain just a subset of the information stored in TSymbol. It makes more sense to use TSymbol directly instead of converting it to TName. This also improves type safety a bit by making some functions only take in TVariable or TFunction instead of the more generic TName. BUG=angleproject:2267 TEST=angle_unittests Change-Id: Icb46923c25d33ebbbbc06ddc487da25957dda771 Reviewed-on: https://chromium-review.googlesource.com/829143 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 4728bdc8 2017-12-20T17:51:08 Unify looking for symbols with a specific name in the AST Keep only one traverser for looking up symbol nodes by name instead of having two largely identical ones. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I36e906258180e22b7b1353cab79d90266d99fa0e Reviewed-on: https://chromium-review.googlesource.com/836895 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Jiawei Shao 492b5f51 2017-12-13T09:39:27 ES31: Support struct arrays as Geometry Shader inputs This patch adds the support of struct arrays as valid geometry shader user-defined inputs. Struct arrays are accepted as geometry shader inputs to match the vertex shader outputs that are also declared as structs. BUG=angleproject:1941 TEST=angle_unittests Change-Id: I0b5d545b10e9dda576a1c96d7c93ec2450611e9e Reviewed-on: https://chromium-review.googlesource.com/823622 Commit-Queue: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho bed35d76 2017-12-20T16:36:26 Don't query names of empty symbols This makes it possible to return a reference from TSymbol::name() instead of a pointer. This is safer since it completely avoids the possibility of a nullptr dereference. An assert is making sure that the function is not being called for empty symbols. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I44279f65989dbb828322843fc0216ba84d91dedf Reviewed-on: https://chromium-review.googlesource.com/836894 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho b8cb939f 2017-12-20T14:23:19 Fix tracking variables in folded ternary operators The result of folding a ternary operator may be a TIntermSymbol node where the qualifier doesn't match the qualifier of the variable that the node is referring to. Get the qualifier from the variable instead of directly from TIntermSymbol when collecting variables in CollectVariables or when tracking referenced variables in OutputHLSL. BUG=angleproject:2288 TEST=angle_unittests, angle_end2end_tests Change-Id: If294a7fe9dca50f2ebcea3feff887e72a521d395 Reviewed-on: https://chromium-review.googlesource.com/836893 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho f095799b 2017-12-22T11:10:04 Fix accessing the name of a nested struct definition When generating an error message about the struct nesting limit, the code should make sure that the struct definitions are not nested. While nested struct definitions by themselves are also an error, they're not a syntax error so parsing will continue after encountering them. This fixes a regression from commit: Don't allocate name strings for empty symbols. BUG=chromium:797156 TEST=angle_unittests Change-Id: I4149fbe874c0e7ec90e690aec078ccaf7313eab0 Reviewed-on: https://chromium-review.googlesource.com/842643 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho b6af22b5 2017-12-15T14:05:44 Store TVariable* in TIntermSymbol instead of storing id This is an intermediate step to only storing a TVariable * in TIntermSymbol instead of copying the name. This makes it possible to get a constant value out of a TIntermSymbol without doing a symbol table lookup. BUG=angleproject:2267 TEST=angle_unittests Change-Id: Ibff588241a4ad4ac330063296273288b20a072c9 Reviewed-on: https://chromium-review.googlesource.com/829142 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho ae4dbf32 2017-12-08T20:49:00 Don't allocate name strings for empty symbols This removes unnecessary memory allocations. BUG=angleproject:2267 TEST=angle_unittests Change-Id: Ide575ea19ab2f8e9fc93092490f1352efa6024a3 Reviewed-on: https://chromium-review.googlesource.com/817415 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Frank Henigman 0e74e059 2017-12-17T22:17:17 Fix gcc builds. Remove unused function and fix dangling else's that gcc complains about. In the gyp build specify -std=c++14. BUG=none Change-Id: I552942ccc574104a5424a9add00e6e72d6efa28a Reviewed-on: https://chromium-review.googlesource.com/830750 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho a2d98141 2017-12-15T14:18:55 Fix allowing non-constant ternary global initializer Check the qualifier of a node resulting from the folding of a ternary node correctly. The folded node might even be a TIntermConstantUnion with a non-constant qualifier. BUG=angleproject:2285 TEST=angle_unittests Change-Id: I74516e44ce9d78bc54093a5b58d14cf33a57e6e5 Reviewed-on: https://chromium-review.googlesource.com/829138 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 195be942 2017-12-04T23:40:14 Always create TVariables for TIntermSymbol nodes TIntermSymbol nodes are now constructed based on a specific TVariable. This makes sure that all TIntermSymbol nodes that are created to refer to a specific temporary in an AST transform will have consistent data. The TVariable objects are not necessarily added to the symbol table levels - just those variables that can be referred to by their name during parsing need to be reachable through there. In the future this can be taken a step further so that TIntermSymbol nodes just to point to a TVariable instead of duplicating the information. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I4e7bcdb0637cd3b588d3c202ef02f4b7bd7954a1 Reviewed-on: https://chromium-review.googlesource.com/811925 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Jiawei Shao bd924af1 2017-11-16T15:28:04 ES31: Remove support of GL_OES_geometry_shader in compiler This patch intends to remove the support of GL_OES_geometry_shader in ANGLE compiler. To make the implementation simpler, currently we decide to only support "GL_EXT_geometry_shader" as the extension string of geometry shader in ANGLE. This patch also updates all the related shaders in angle_unittests into R"()" format. BUG=angleproject:1941 TEST=angle_unittests Change-Id: Ife9858abeedfb46b02c5c2fb1cda16fa27198511 Reviewed-on: https://chromium-review.googlesource.com/773451 Commit-Queue: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho ebe13bb1 2017-12-11T13:32:44 Remove SH_TRANSLATE_VIEWID_OVR_TO_UNIFORM Instanced multiview is fully implemented and performs much better than broadcasting draw calls on the CPU. There's no need to maintain translating ViewID_OVR to a uniform. BUG=angleproject:2267 TEST=angle_unittests Change-Id: If3a8fa573f37b7926c052aecfcd4271f266751dc Reviewed-on: https://chromium-review.googlesource.com/819417 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho fdb400d7 2017-12-01T15:12:36 Test multiview extension warning Ensure that a warning is generated if the shader tries to set a global layout qualifier. Since a num_views layout qualifier is the only valid global layout qualifier in GLSL ES 1.00, it's sufficient that checkCanUseExtension() that generates the warning is called for "num_views" specifically. Parsing "layout" or "in" storage qualifier may use the simpler isExtensionEnabled() check that doesn't generate a warning. BUG=angleproject:2238 TEST=angle_unittests Change-Id: I19caff7be9cb4f53c581b5f5526b7cade08469d6 Reviewed-on: https://chromium-review.googlesource.com/803577 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 3de2703d 2017-11-30T12:16:47 Fix handling matrix qualifiers on block members Individual block member row_major/column_major layout qualifiers may override the qualifiers set on the block. During parsing, this was already being handled correctly, so that the qualifier is resolved for each block member and recorded for each TField / InterfaceBlockField. Now we always write the qualifiers on a per-member granularity to the output GLSL shaders, so that the native driver gets the correct per-member qualifiers. This replaces earlier behavior where the matrix qualifiers were only written per-block. Also only use qualifiers from individual members in block layout. Since the block-level qualifier information is no longer used after parsing, it is no longer kept in the AST. A dummy value is still set to the InterfaceBlock structs exposed through the ShaderVars interface, since that has existing usage in Chromium that needs to be removed before the field can be removed. Some AMD OpenGL drivers don't seem to handle matrix layout qualifiers correctly, so most of the added tests need to be skipped for AMD GL. On NVIDIA and Intel the tests pass. BUG=angleproject:2271 TEST=angle_unittests, angle_end2end_tests, dEQP-GLES31.functional.program_interface_query.uniform.matrix* Change-Id: I1baa7a633bc2da548743c2190cb72db491b5227a Reviewed-on: https://chromium-review.googlesource.com/800174 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 18841310 2017-11-28T12:48:47 Clean up MatchOutputCodeTest usage It's cleaner to use a generic function to test that a sequence of strings is found in a specific order rather than fetching and comparing string locations in test case code. Also make sure that string occurrences can't overlap when looking for a specific number of occurrences of the same string. TEST=angle_unittests Change-Id: I8ca66c73c7aaa5be8469ded466f51d97a36c801b Reviewed-on: https://chromium-review.googlesource.com/793041 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho ae04e1e4 2017-11-27T16:00:39 Fix scalarizing vec and mat constructor args Scalarizing vec and mat constructor args can generate new statements in the parent block of the constructor. To preserve the correct execution order of expressions, scalarized vector and matrix constructors need to be first moved out from inside loop conditions and sequence operators. This is done whenever the compiler flag to scalarize args is on. BUG=chromium:772653 TEST=angle_unittests Change-Id: Id40f8d848a9d087e186ef2e680c8e4cd440221d9 Reviewed-on: https://chromium-review.googlesource.com/790412 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho f13cadd8 2017-11-28T10:53:09 Fix checking negative index when indexing matrix/vector It's important that the test against the maximum of the valid range is only done if the index is positive, so the sanitized index value is guaranteed to end up in the valid range. This fixes a regression from commit "Add GLSL support for runtime-sized arrays in SSBOs". BUG=chromium:789029 TEST=angle_unittests Change-Id: Ic7125e383a64e46994b072df6d7e642432c521af Reviewed-on: https://chromium-review.googlesource.com/792935 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho ebee5b3b 2017-11-23T12:56:32 Add GLSL support for runtime-sized arrays in SSBOs The GLSL parser now allows a runtime-sized array as the last member in a shader storage block. Clamping indexing against the memory bounds is done by determining the array length at runtime. Runtime-sized arrays are used in dEQP tests for many compute shader tests, so these now work on the OpenGL backend. BUG=angleproject:1951 TEST=angle_unittests, dEQP-GLES31.functional.shaders.linkage.shader_storage_block.* dEQP-GLES31.functional.shaders.builtin_functions.*compute* Change-Id: Ibecca24623ca8e4723af6f0e0421fe9711ea828d Reviewed-on: https://chromium-review.googlesource.com/787976 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 465835d6 2017-09-26T13:34:10 Support arrays of arrays in the API The ShaderVariable class that is used as an interface between the compiler and the rest of the code gets arrays of arrays support. Array of array variables are passed from the compiler just like any other variables. However, when stored in Program state each innermost array constitutes a separate variable. This is done to make the implementation match the GLES specification for program interface query APIs. This will be tested more fully once support for parsing arrays of arrays lands in the compiler. TEST=angle_end2end_tests, angle_unittests BUG=angleproject:2125 Change-Id: I0f7159000f039be92a87a52b3b68cd9a215a21cb Reviewed-on: https://chromium-review.googlesource.com/684742 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 39f74df5 2017-11-20T16:09:57 Remove unreferenced struct types from the AST This expands pruning unreferenced variables so that unreferenced named struct types can also be removed from the AST. Includes a small cleanup in GLSL output so that the output code matching tests can test against clean output. BUG=chromium:786535 TEST=angle_unittests Change-Id: I20974ac99a797e478d82f9203c179d2d58fac268 Reviewed-on: https://chromium-review.googlesource.com/779519 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 3d70ca9c 2017-11-10T16:53:26 Remove unreferenced variables from the AST Unreferenced local and global variables are now pruned from the AST. They will be removed unless their initializer has side effects. The CollectVariables step needs to be run after the pruning, as the pruning may affect which interface variables are statically used. It's also good to gather built-ins that need to be emulated after the pruning, so unnecessary built-in emulation functions are not added to the translator output. This will help handle some dEQP tests for arrays of arrays that have extremely large local arrays that are only used in an array length query. By constant folding the length and pruning unused variables we will avoid adding a large amount of array initialization code to the generated shaders. BUG=angleproject:2166 TEST=angle_unittests, angle_end2end_tests Change-Id: Ic918bfe8f16460bcd6101d73a7a674145f5aeecd Reviewed-on: https://chromium-review.googlesource.com/766434 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho bb27c3a1 2017-11-15T16:32:25 Fix VectorizeVectorScalarArithmetic statement insertion The traverser must avoid inserting two statements to the same position on a single traversal, so it doesn't trigger an assert. BUG=chromium:784078 TEST=angle_unittests Change-Id: I855054e62cc1b1cf4e6bb02af527954151c7d0e7 Reviewed-on: https://chromium-review.googlesource.com/771611 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jiajia Qin a3106c58 2017-11-03T09:39:39 ES31: Add atomic memory functions BUG=angleproject:1442 TEST=angle_unittests, angle_end2end_tests dEQP-GLES31.functional.compute.shared_var.atomic* dEQP-GLES31.functional.compute.basic.shared_atomic_op* dEQP-GLES31.functional.ssbo.atomic.* Change-Id: I82b54fde3a852d3bd917b1e19680baa1c28fce4d Reviewed-on: https://chromium-review.googlesource.com/765061 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 7caa80e7 2017-11-14T15:03:14 Edit tests to reference temporary variables The shader translator is intended to prune unreferenced variables in the future. To maintain test coverage when this is done, most tests that used to have unreferenced GLSL locals and globals are edited to use built-ins or reference the necessary variables instead. BUG=angleproject:2166 TEST=angle_unittests, angle_end2end_tests Change-Id: I3964e7a80d52fc04a95f57e73da6095e433095e8 Reviewed-on: https://chromium-review.googlesource.com/768740 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 7af63727 2017-11-13T15:03:40 Fix nullptr dereference on struct parameter error Function parameter name string does not necessarily exist, so it's better to use the function name as the token in the error message. BUG=chromium:784158 TEST=angle_unittests Change-Id: I8f3b8604fd702bdc9486b8d721a5f60de1ff3fa7 Reviewed-on: https://chromium-review.googlesource.com/765972 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 703671e9 2017-11-08T17:47:18 Unify extension behavior checks Some supportsExtension and isExtensionEnabled checks are now turned into checkCanUseExtension checks. Using checkCanUseExtension is preferable so that warnings are generated correctly when an extension is used and a warn directive is present. isExtensionEnabled is still used in some places where an error message about the extension could be confusing, particularly when a core spec version adds support for something that is also present in an extension. Also make it possible to disable ARB_texture_rectangle extension using an extension directive. ARB_texture_rectangle extension functionality is enabled by default in GLSL when the extension is supported. BUG=angleproject:2238 TEST=angle_unittests Change-Id: I7455293412ff469f54bc7da79df146e7bc127379 Reviewed-on: https://chromium-review.googlesource.com/760737 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jiawei Shao 0e883134 2017-10-26T09:53:50 ES31: Support EXT_geometry_shader in GLSL compiler This patch intends to support "EXT_geometry_shader" as a valid extension string in ANGLE GLSL compiler. We decide to support it because in dEQP-GLES31 all geometry shader related tests are using "EXT_geometry_shader" instead of "OES_geometry_shader". 1. Support new extension string "EXT_geometry_shader" 2. Enable geometry shader layout qualifiers with EXT_geometry_shader 3. Enable geometry shader builtins with EXT_geometry_shader BUG=angleproject:1941 TEST=angle_unittests Change-Id: Iaedd01a9100ccf56243c957db36ff0c983d17060 Reviewed-on: https://chromium-review.googlesource.com/737933 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho ef7fb388 2017-11-07T16:33:49 Size const unsized arrays without an initializer The array size for unsized arrays needs to be sanitized in all cases, since subsequent checks on array indexing depend on all arrays being sized. BUG=chromium:781729 TEST=angle_unittests Change-Id: I3af6c08bb249a19f7c2ef169c877a2b863eb31d3 Reviewed-on: https://chromium-review.googlesource.com/757101 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho d80f2944 2017-11-06T12:44:45 Struct definition not allowed as function parameter type Struct definitions are not allowed as a function parameter type now. This is specified in ESSL 3.00.6 section 12.10. ESSL 3.00.6 section 6.1.1 contradicts this, but that seems like a mistake, it's been fixed in subsequent spec versions. BUG=angleproject:2225 TEST=angle_unittests Change-Id: I6b97d120c440f0c0a45d31bbfaf292fb497160ce Reviewed-on: https://chromium-review.googlesource.com/754606 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho bd3cd506 2017-11-03T15:48:52 Clean up HLSL constructor output Split generating HLSL struct constructors from generating built-in type constructors, as these didn't have much in common. Struct constructors are now only generated when they are needed, as opposed to before, when they were generated on any use of a struct. This changes built-in constructor naming to include "_ctor" and gets rid of having special built-in type names just for constructors. This will make it easier to do changes to constructor output, for example to add constructors for structs in std140 layout. This might be needed to implement SSBOs efficiently. This includes one bug fix for writing out struct declarations for varyings. Also improves const-correctness of accessing structs through TType in general. BUG=angleproject:2218 TEST=angle_unittests, angle_end2end_tests Change-Id: If865fb56f86486b9c4a2c31e016ea16427f4a5fa Reviewed-on: https://chromium-review.googlesource.com/753883 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho b8ee9dd3 2017-10-30T12:43:27 Validate empty statements in switch statements Even an empty statement like ";" is a statement according to the grammar. They should not be allowed in switch statements before the first case statement, but on the other hand a switch statement that has just an empty statement after the last statement is valid. Now the parser creates AST nodes from empty statements so that we can validate switch statements correctly. However, they are pruned shortly after parsing completes in PruneNoOps, so they don't affect further processing of the AST. BUG=angleproject:2181 TEST=angle_unittests Change-Id: I1085056fc34b146142546fc5f2b7f3124b910ab9 Reviewed-on: https://chromium-review.googlesource.com/743621 Reviewed-by: Jamie Madill <jmadill@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 6bb4f501 2017-10-31T14:54:00 Fix MSVS 2015 compilation of tests "GLenum" needs to be prefixed to disambiguate between global GLenum and sh::GLenum. TEST=compile with MSVS 2015 Change-Id: I31a13c8d410bf32c4d5114c03f4b119aae256778 Reviewed-on: https://chromium-review.googlesource.com/746761 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 722bfb51 2017-10-26T17:00:11 Fix detecting duplicate field names in structures Previously field names that were listed in the first declarator list inside a struct declaration were not checked against each other. BUG=angleproject:2204 TEST=angle_unittests Change-Id: Ibf821d45556f6dfe0223dae673644f6795daf4cb Reviewed-on: https://chromium-review.googlesource.com/739825 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho 55bc905f 2017-10-25T17:33:06 Always consider type arrayness for atomic counters Atomic counter arrays may be declared with various different syntax - the array size may be declared as a part of the type or as a part of the declarator. Take this into account when determining whether atomic counter offsets overlap. BUG=angleproject:1729 TEST=angle_unittests Change-Id: I7435ded9401c4c1caab22c22d83fd2ad301df768 Reviewed-on: https://chromium-review.googlesource.com/738140 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Jiawei Shao 65c56dd9 2017-10-13T16:18:57 Fix incorrect hashing on built-in interface block fields This patch intends to fix an error in translating built-in interface block fields. Any field of a built-in interface block should be kept and cannot be hashed. This patch can fix a bug in handling the interface block gl_in when we try to output the translated geometry shader string. BUG=angleproject:1941 TEST=angle_unittest Change-Id: Iebfba4b6a30c8942ed0f66131ad30d12ad96c62a Reviewed-on: https://chromium-review.googlesource.com/719454 Commit-Queue: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Qin Jiajia ca68d988 2017-09-18T16:41:56 ES31: Add std430 support for OpenGL backend BUG=angleproject:1920 TEST=angle_unittests Change-Id: Ie8e171abec053c2c5dca93d6e79db534f74520e7 Reviewed-on: https://chromium-review.googlesource.com/737532 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Olli Etuaho 9cd7163e 2017-10-26T14:43:20 Fix setting array sizes on a constructor Take any array sizes that have been explicitly specified in the shader text into account, and only set the ones that are unsized according to the arguments. BUG=angleproject:2125 TEST=angle_unittests Change-Id: I37d08a86c25f7cd4f3ce5689f2c9fad444e7d5ad Reviewed-on: https://chromium-review.googlesource.com/738141 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Olli Etuaho 454c34cb 2017-10-25T16:35:56 Accept valid geometry shader inputs regardless of syntax Before, only the following style of declarations were accepted: in float f[]; Now also these styles are accepted: in float[] f; in float f[], g[]; BUG=angleproject:2201 TEST=angle_unittests Change-Id: I0af7d355a5e06a67ceef2d6bd69af7e23c180a04 Reviewed-on: https://chromium-review.googlesource.com/738234 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 55bde916 2017-10-25T13:41:13 Parse array specifier with a separate grammar rule This brings the grammar closer to the GLSL ES 3.10 spec. Some corner cases related to handling unsized arrays are fixed. BUG=angleproject:2125 TEST=angle_unittests, angle_end2end_tests Change-Id: I9bcf87b17b97da0e2ec2954d32037c272fde3080 Reviewed-on: https://chromium-review.googlesource.com/738233 Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jiawei Shao 8e4b355b 2017-08-30T14:20:58 ES31: Implement Geometry Shader inputs and outputs This patch intends to implement Geometry Shader inputs and outputs in ANGLE GLSL compiler. 1. Only accept arrays as the inputs of a Geometry Shader. 2. Allow unsized arrays as the inputs of a Geometry Shader after a valid input primitive declaration and assign size to them. 3. Implement Geometry Shader outputs. 4. Allow Geometry Shader inputs and outputs using interpolation qualifiers ('flat', 'smooth', 'centroid'). 5. Allow using 'location' layout qualifier on Geometry Shader inputs and outputs. BUG=angleproject:1941 TEST=angle_unittests Change-Id: Ia7e250277c61f45c8479437b567c2831ff26b433 Reviewed-on: https://chromium-review.googlesource.com/650211 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
kmx.io     mail     discord kmxgit v0.4.0