|
46149c0e
|
2021-09-17T15:38:47
|
|
Translator: Fix pruning functions that declare a struct
When pruning unused functions, the entire declaration of the function is
removed. If the function declares a struct as part of its return value,
the declaration of that struct was lost. This change makes sure that
declaration survives the pruning. For example:
struct S { vec4 v; } unused_func() { ... }
is now replaced with:
struct S { vec4 v; };
This change also makes struct usage validation more stringent.
Bug: chromium:1248753
Change-Id: Idd9a87fd2e785135775cfea62995fd33adaf3c09
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3169691
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
800e82c6
|
2021-08-23T11:05:23
|
|
Translator: Validate precisions
When declaring a variable, a struct field, function parameter etc,
there's a precision necessarily applied to the entity being declared.
AST Validation is added to enforce this. Intermediate nodes derive
their precision from these entities automatically.
Consistency of intermediate nodes is not validated. This is because AST
transformations replace a node with a transformed one, and that may not
have the same precision. Take the following code:
mediump float x = ...;
mediump float y = ...;
... x + y ...
and assume is transformed as such:
highp float driver_uniform;
... (x * driver_uniform) + y ...
The addition was originally done in mediump, but would seemingly need to
be done in highp after transformation. There are a number of options
here:
- Make sure that when nodes are replaced, the precision is unaffected.
This can be intrusive, requiring temp variables.
- Bubble up the new precision
- Accept the discrepancy
ANGLE opts for the last option, which actually respects the original
shader's intended precision for operations, even if some transformation
needs to temporarily evaluate an expression at a higher precision.
Bug: angleproject:4889
Bug: angleproject:6132
Change-Id: Ibcde3a230de159157783b1c6d5ef1cd63ceb4d8f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3114027
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
798c5200
|
2021-07-31T22:17:48
|
|
Translator: Make precise-ness a property of AST nodes
Instead of creating a "set of precise nodes", the nodes themselves are
marked precise. This change is preparation for another change that
overhauls precision propagation through the nodes, and which would also
be storing the promoted precision in the nodes.
Bug: angleproject:4889
Change-Id: Ieec8864be5cd2e95ed6fb4d90b367946d9f33fa1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3065569
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
ebe943e2
|
2021-07-29T11:19:33
|
|
Vulkan: SPIR-V Gen: Fix gl_Clip/CullDistance
These built-ins can be redeclared in the shader. The translator took
these redeclarations (and gl_LastFragData) as UserDefined symbols.
There were a number of hacks such as in name generation and
CollectVariables, to special-case these redeclarations.
This change instead makes sure that these variables continue to be
considered built-ins with the appropriate qualifiers. A number of fixes
are made accordingly:
- Hacks are removed.
- In fragment shaders, ANGLEClipDistance was initialized with
gl_ClipDistance for further use by the shader. The code generation
however mistakenly produced `gl_ClipDistance[0] =
ANGLEClipDistance[0];`, which caused compilation failures by glslang,
but passed the tests accidentally because they expected link failures
(see next item).
- CollectVariables is fixed to correctly collect gl_Clip/CullDistance
in fragment shaders; previously they were collected as output
varyings, and therefore the aforementioned link error was not produced
in the tests after the compilation error was fixed.
Additional fixes:
- The transformation of gl_Clip/CullDistance was always ever done on one
of them due to misplaced breaks in the loop that detected their
presence.
- The transformation of gl_CullDistance was skipped when it was not
redeclared.
Validation is added to make sure these built-ins always have the correct
qualifier even when redeclared. SPIR-V gen support for these variables
have been fixed as well.
Bug: angleproject:4889
Change-Id: Ic8bf2c508357035cb937ef41a28ae22ffc533ebe
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3059921
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
aec5e65c
|
2021-07-28T00:36:12
|
|
Get direct-to-Metal backend to run angle_end2end_tests.
Cherry-pick nameless struct fix from Apple in
https://bugs.webkit.org/show_bug.cgi?id=227482 .
Fix SeparateCompoundStructDeclarations pass to stop generating
multiple declarations; thanks syoussefi@ for advice.
Incorporate additional passes from TranslatorVulkan
(MonomorphizeUnsupportedFunctionsInVulkanGLSL,
RewriteArrayOfArrayOfOpaqueUniforms,
SeparateStructFromUniformDeclarations) needed by RewriteStructSamplers
pass in TranslatorMetalDirect. Fixes many assertion failures in GLSL
tests. Moved these passes out of tree_ops/vulkan. Thanks again to
syoussefi@ for advice and help.
Disable a validation check related to the RewritePipelines pass. Skip
two tests that were failing for other reasons.
With these changes, angle_end2end_tests runs to completion when the
direct-to-Metal backend is turned on. There are still ~1300 failures
of the ~4000 tests which will be investigated next.
Bug: angleproject:5505
Change-Id: Ibca77822543e8e8e8d2a8c862e92cdf74bfa3545
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3058524
Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Kenneth Russell <kbr@chromium.org>
|
|
7c6da493
|
2021-07-26T21:24:25
|
|
Translator: Make sure built-in variables are consistent
Some transformations left the tree in an inconsistent state, for example
if it used gl_SampleMask from ES3.1+OES_sample_variables and the
transformation added a usage of gl_SampleMask from ES3.2. The offending
transformations are fixed and AST validation is improved to catch such
errors.
Bug: angleproject:4889
Change-Id: I9d9ea5bb43a9408dd4c6dc3e89ec20d60dfeff73
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3054613
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
7f199d75
|
2021-07-20T00:11:23
|
|
Vulkan: SPIR-V Gen: Support geometry shaders
Geometry shader built-ins are implemented. A bug in
DeclarePerVertexBlocks is fixed w.r.t to gl_PerVertex arrays and a new
AST validation option is added to catch that sort of error.
Bug: angleproject:4889
Change-Id: I40b2929e450eac8ae840bbf1a1583e6c4c047c88
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3041623
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
0e1c90cf
|
2021-07-21T10:24:26
|
|
Translator: Validate function parameter qualifiers
This change cleans up a confusion in EvqConst and EvqConstReadOnly,
where the former was frequently used instead of the latter for function
parameters.
The change makes the following renames to make the intent of the
relevant qualifiers clearer:
EvqIn -> EvqParamIn
EvqOut -> EvqParamOut
EvqInOut -> EvqParamInOut
EvqConstReadOnly -> EvqParamConst
Bug: angleproject:4889
Change-Id: Idedd32c08a91de069b91b1657d6b783dddece04a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3041622
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
a3df80ea
|
2021-07-19T17:15:13
|
|
Translator: Validate AST consistency w.r.t function calls
This ensures that tree transformations don't create invalid function
calls. Additionally, it's validated that no raw function calls are
issued, as that's deprecated.
Bug: angleproject:4889
Change-Id: Iabbb493dcc0a5654bbb5f930e402bf3492c89979
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/3040119
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
|
|
8b869a95
|
2021-06-13T01:09:27
|
|
Translator: Generate Ops for all built-in functions
EOpCallBuiltInFunction is removed in this change, as well as the
"op": "auto" property in builtin_function_declarations.txt. Instead,
gen_builtin_symbols.py automatically generates Ops for every built-in
function and generates the TOperator enum accordingly.
This simplifies SPIR-V code generation by allowing switches to be used
on operators instead of string comparisons.
Bug: angleproject:4589
Bug: angleproject:4889
Change-Id: Ia351524400b0e12a10a5572e27e9b88c6ec2e61c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2958869
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
64ca25f1
|
2021-06-08T17:15:49
|
|
Vulkan: SPIR-V Gen: Unary operators
The line raster emulation coded uses round() which was previously not
translated. This change implements all operations supported by
TIntermUnary nodes. Bresenham line emulation is now allowed.
Bug: angleproject:4889
Change-Id: I1ffbe3f9bf0299f7efa2c6d1e8f011fefdbb2ddb
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2951624
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
31641b8e
|
2021-06-05T02:17:47
|
|
Declare specialization constants in the AST
A new qualifier (EvqSpecConst) is added so that specialization constants
can be declared in tree.
This enables TVariable references to specialization constants to be
validated, which were in fact invalid as every reference declared a new
variable. That is fixed.
Bug: angleproject:4889
Change-Id: I1711e41cbc5a1260843d2d004d3568bdae11e963
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2941451
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
df977aaf
|
2021-06-08T01:04:32
|
|
Translator: Validate global qualifier declaration's symbol
Global qualifier declarations (invariant, precise) apply to
already-declared variables. This is now validated.
Bug: angleproject:4889
Change-Id: I9889f2d68dcd6a2fc0de2b20a1291899397982d0
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2946114
Reviewed-by: Jonah Ryan-Davis <jonahr@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
daeac238
|
2021-05-08T22:09:38
|
|
Translator: Ensure structs and blocks are uniquely defined
A new AST validation is added to ensure that the same TStructure or
TInterfaceBlock is not redundantly defined. This helps with SPIR-V
generation by allowing the id to be used as key in a hash map that looks
up the corresponding SPIR-V type id.
A bug is fixed where the Vulkan driver uniform declaration created two
identical declarations for ANGLEDepthRangeParams.
A number of other bugs are also fixed in this change, where if a
variable declaration is eliminated (for example due to constant folding,
or inactive interface variable removal) and it contained a struct
specifier, the struct declaration was also removed. OutputGLSLBase had
a hack where structs were declared on first encounter, which was
incorrect as the scope of the declaration could change. Those bugs are
fixed and this hack is removed.
Bug: angleproject:2733
Bug: angleproject:4889
Bug: angleproject:5936
Change-Id: I8e13748c0bf552ae8b052249282769a1f0775603
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2881942
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
c2d1a464
|
2021-04-09T16:53:30
|
|
Translator: Track nameless interface block field indices in type
Fields of nameless interface blocks are stored as TVariables and
referenced through a TIntermSymbols (unlike named blocks which use
EOpIndexDirectInterfaceBlock with a constant field indices). With this
change, the field index is stored in the variable's type which let's the
TField of the TInterfaceBlock be directly accessible (and not need a
search by name).
This will be helpful in translation to SPIR-V as interface block members
are accessed by field index.
Bug: angleproject:4889
Change-Id: If3ab45b1e5f5f9576721dc52e2bdf1161882514f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2818242
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
58eabfd7
|
2021-04-08T23:57:04
|
|
Translator: Validate consistent variable references
Some transformations change variable declarations. This validation
ensures that all references to said variables are appropriately replaced
as well.
Bug: angleproject:2733
Change-Id: I6c2873968eeed4cba66e70069f84eb69a1f77074
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2818140
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
|
|
a7b60f13
|
2021-04-08T23:13:21
|
|
Translator: Validate multi-declarations are separated
Many transformations only look at the first variable in a
TIntermDeclaration. This change validates that once the
SeparateDeclarations transformation has been run, multi-declarations are
not reintroduced.
Bug: angleproject:2733
Change-Id: If5f96689cbdd28817c479dd7c495b91535f87199
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2815565
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|
|
249cb200
|
2019-12-03T11:10:06
|
|
Translator: Rename TIntermInvariantDeclaration to ..GlobalQualifier..
This is to support the upcoming `precise` keyword.
Bug: angleproject:3569
Change-Id: Idbcc8fd6f261757dbbf81b381e7a2dae938d8101
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/1947451
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
|
|
91d469aa
|
2018-08-27T10:47:25
|
|
Initial support for compiler AST validation
This is to be able to catch issues introduced through AST manipulation
earlier than shader compilation time, improving debuggability.
Bug: angleproject:2733
Change-Id: Ic57bc72f2ab438e60f32553d602074f3d72cc4f5
Reviewed-on: https://chromium-review.googlesource.com/c/1199922
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
|