| Age | Commit message (Collapse) | Author |
|---|
|
* Only do scrubbing if needed. When allocating content try to limit size (with scrubbing each token takes up 1k), now it's 16 bytes min size.
* Don't allocate for every call to write on the CallbackWriter - use the m_appendBuffer.
* Don't allocate memory for CallbackWriter use m_appendBuffer.
* Use UnownedStringSlice for suffix output for parsing float/int literals.
Fix typo in invalidFloatingPointLiteralSuffix
* Using memory arena to hold tokens that are not in SourceManager.
* Improve comment on lexing.
* Make UnownedStringSlice allocation simpler on SourceManager.
* Fix error on gcc around UnownedStringSlice - because VC converted string + UnownedStringSlice automatically into a String.
* Fix generateName needing concat string for gcc.
* When constructing a Token in parseAttributeName - because it's a Identifier, we have to set the Name.
* Remove translation through String on getIntrinsicOp
* Make func-cbuffer-param disablable with -exclude compatibility-issue
* Move memory leak in render-test.
* From review - can just use "?:" instead of performing a concat.
|
|
* First step toward supporting use of interfaces as existential types
Traditional generics involve universal quantification. E.g., a declaration like:
```
void drive<T : IVehicle>(T vehicle);
```
indicates for *for all* types `T` that implement the `IVehicle` interface, the `drive()` function is available.
In contrast, whend directly using an interface type like:
```
IVehicle v = ...;
v.doSomething();
```
we only know that there *exists* some concrete type (we could call it `E`) such that `v` refers to a value of type `E`, and `E` implements the `IVehicle` interface. In order to perform an operation like `v.doSomething()` we need to "open" the existential value so that we can look at the concrete type and how it implements the `IVehicle.doSomething` requirement.
This change adds a very explicit representation of existentials to Slang's IR. An operation like `e = makeExistential(v, w)` creates a value of some existential type (interfaces being our only existential types for now), by wrapping a concrete value `v` (the type of `v` can be seen as an implicit operand) and a witness table `w` showing that the type of `v` implements the requirements of the chosen interface type.
In turn, opening of an existential is handled with operations `extractExistential{Value|Type|WitnessTable}` which pull the corresponding piece of information out of a value of existential type (which somewhere in the code had to have been created with `makeExistential`).
The change includes a trivial simplification pass that can detect cases where an `extractExistential*` operation is applied direclty to a `makeExistential` operation, so that there is only one possible result that could be extracted. This allows for simplification of existential types used in trivial ways for local variables (this is mostly so I can check in a functional test, rather than to actually support useful code involving interfaces right now).
The logic in the semantic checking phase of the compiler is comparatively more complex.
When we are about to perform member lookup given an expression like `obj.member` we will first check if `obj` has an existential type, and if it does we will construct a suitable local context in which we extract the value, type, and witness table from the existential (these all become explicit AST expression nodes), and then use the extracted value as the base of the lookup operation.
The nature of existential values is that two different values with the same existential (interface) type could wrap concrete values with differnt types, so that we need to carefully refer only to the extracted type/value/witness-table of specific *values*. We handle this right now by conceptually moving the existential-type value into a local variable (by introducing a `LetExpr` that amounts to `let v = <init> in <body>`) and then require that the extract expressions must refer to the (immutable) variable declaration from which they are extracting a value.
(Eventually we should expand this so that when using an immutable local variable of existential type we just use that variable as-is rather than introduce a new temporary)
A simple test case is included that uses an interface type in an almost trivial way for a local variable; this test can be run and produces the expected results.
A more complex test case that passes an existential into a function is included, but left disabled because a more aggressive simplification approach is required to generate working code from it.
* Add missing file for expected test output
* Fixups for merge from top-of-tree
|
|
* Specialize away resource-type function parameters
Work on #397.
Introduction
------------
Suppose a user writes a function that takes a resource type as a parameter:
```hlsl
float4 getThing(RWStructuredBuffer<float4> buffer, int index)
{
return buffer[index];
}
```
This function creates challenges when generating code for GLSL-based targets, because a global shader parameter of type `RWStructuredBuffer`:
```hlsl
RWStructuredBuffer<float4> gBuffer;
```
translates to a global GLSL `buffer` declaration:
```hlsl
buffer _S0
{
float4 _data[];
} gBuffer;
```
There is no equivalent to that `buffer` declaration that can be used in function parameter position, and it is illegal in GLSL to pass `gBuffer` into a function.
(Aside: yes, we could in principle translate a function parameter like `RWStructuredBuffer<float4> buffer` to `float4 buffer[]`, but that will not in turn generalize to arrays of structured buffers; it is a dead-end strategy)
The solution employed by many shader compilers is to "inline everything" to eliminate the need for parameters of resource types, and then rely on dataflow optimization to eliminate locals of resource types. This strategy can of course lead to an increase in code size, and it also means that call stacks are lost when doing step-through debugging. Another serious issue is that an "early `return`" from a function can turn into the equivalent of a multi-level `break` when inlined, and not all of our targets support multi-level `break`.
The solution implemented in this change works around some, but not all, of the problems with full inlining.
The approach here generates specialized versions of a function like `getThing`, adapted to the actual arguments provided at different call sites.
Thus if we have code like:
```hlsl
RWStructuredBuffer<float4> gA;
RWStructuredBuffer<float4> gB[10];
...
getThing(gA, x);
getThing(gA, y);
getThing(gB[someVal], z);
```
we will generate two specializations of `getThing`: one specialized for the `buffer` parameter being `gA` and the other for `gB`:
```hlsl
float4 getThing_gA(int index) { return gA[index]; }
float4 getThing_gB(int _val, int index) { return gB[_val][index]; }
```
and the call sites will change to match:
```hlsl
getThing_gA(x);
getThing_gA(y);
getThing_gB(someVal, z);
```
Note how in the case where the argument being passed in was obtained by indexing into an array of resources, the callee is specialized to the identity of the global shader parameter (`gB`), and now accepts a new parameter to indicate the array index into it.
While this description motivates the change based on GLSL output, the same basic issue can arise for other targets.
For example, while current HLSL has added the `ConstantBuffer<T>` type, it is not supported on older targets, and it turns out that even dxc does not allow functions to have `ConstantBuffer<T>` parameters.
Longer-term, we will likely need to do even more aggressive specialization both in order to generate SPIR-V output directly, and also to deal with function that have return values or `out` parameters of resource types.
Implementation
--------------
The meat of the change is in `ir-specialize-resources.{h,cpp}`, where we have a pass that looks at all call sites (`IRCall` instructions) in the program, and attempts to replace them with calls to specialized functions, where the specializations are generated on-demand.
The code in this pass is heavily commented, so hopefully it serves to explain itself all right.
After specialization is complete, we may still have functions like the original `getThing` that will produce invalid code when emitted as GLSL, so we need a way to make sure they don't appear in the output.
To date we've had some very ad hoc approaches for ignoring IR constructs that we don't want to affect emitted code, but this change goes ahead and adds a more real dead code elimination (DCE) pass in `ir-dce.{h,cpp}`.
This pass follows a straightforward approach of tagging instructions that are "live" and then propagating liveness through the whole program, before making a single pass to delete anything that isn't live.
When I first added the DCE pass it eliminated *everything* because there were no "roots" for liveness.
I solved this for now by adding a new decoration, `IREntryPointDecoration`, to mark shader entry points in the IR which should always be live (as should anything they depend on).
A secondary problem that arose was that for GLSL ray tracing shaders it is possible for the incoming/outgoing payload or attributes parameters to be unused, but eliminating them as dead would change the signature of a shader an potential break the rules for how ray tracing programs communicate.
I added a very simple `IRDependsOnDecoration` that allows one IR instruction to keep another alive *as if* it used it, without actually using it.
There's also a fixup in the IR dumping logic where I was forgetting to store anything in the mapping from instruction to their names, so that the name of an instruction was getting incremented each time it was referenced.
Testing
-------
There are three different tests added as part of this change:
* The `compute/func-resource-param` test covers the basic `RWStructuredBuffer` case above, which we expect to work fine for D3D11/12, but fail for Vulkan without specialization.
* The `cross-compile/func-resource-param-array` test covers the case where we don't just have one resource, but an array of them. This is not an end-to-end compute test primarily because our `render-test` application doesn't yet handle arrays of resources correctly in its binding logic.
* The `compute/func-cbuffer-param` test covers the case of a function with a `ConstantBuffer<T>` parameter, which requires specialization to become valid for any of our targets.
* fixup: warnings/errors from other compilers
* fixup: typos and cleanup
* fixup: typos
|
|
Before this change, global shader parameters were represented in the IR as just being ordinary global variables.
The only indication that a particular global represented a parameter was when it got a layotu attached to it (as part of back-end processing), and we've had a number of bugs related to layouts being dropped so that what should have been a shader parameter turned into an ordinary global variable in the output.
This change is more strongly motivated by the fact that making shader parameters look like globals means that we cannot easily reason about their value when doing IR transformations.
If we see two `load`s from the same global variable can we assume they yield the same value?
In the general case we cannot, and this means that any transformation that wants to rely on the fact that an input `Texture2D` shader parameter can't actually change over the life of the program needs to do extra work.
The fix here is to introduce a new kind of IR instruction that represents a global shader parameter directly (not a pointer to it as a global would), at which point there isn't even such a notion as a "load" from the parameter, since it represents the value directly.
In several cases logic that used to apply to global variables in case they were shader parameters (by looking for a layout) is now moved to apply to these global parameters.
The biggest source of issues in this change was that switching from pointers to plain values to represent these shader parameters stresses different cases in type legalization. I also had to deal with the case of legalization for GLSL where we actually *do* need global shader parameters that are writable (since varying output goes in the global scope), but in that case I borrowed the use of pointer-like `Out<...>` and `InOut<...>` types to represent that intent, which we were already using for function parameters representing outputs.
A few tests started failing because the changes lead to a slightly different order of code emission, which in some HLSL tests resulted in a function parameter named `s` getting emitted before a global parameter named `s`, leading to the latter getting the name `s_1` instead of `s_0`.
A few SPIR-V tests started failing because the new approach means that we no longer end up performing a load from all varying input parameters at the start of `main` and instead reference the varying inputs directly. The resulting code is more idomatic, but it differed from the baselines for those tests.
|
|
* Move mangled name out of IRGlobalValue
Previously the `IRGlobalValue` type was used as a root for all IR instructions that can have "linkage," in the sense that a definition in one module can satisfy a use in another module.
The mangled symbol name was stored in state directly on each `IRGlobalValue`, which created some complications, and also forced IR instructions that wanted to support linkage to wedge into the hierarchy at that specific point.
This change moves the mangled name out into a decoration: either an `IRImportDecoration` or an `IRExportDecoration`, both of which inherit from `IRLinkageDecoration` which exposes the mangled name.
This change has a few benefits:
* We can now have any kind of instruction be exported/imported, without having to inherit from `IRGlobalValue`. This could potentially let `IRStructType` and `IRWitnessTable` be simplified to just have operand lists instead of dummy chldren as they do today.
* We can now easily have "global values" like functions that explicitly *don't* get linkage, instead of using a null or empty mangled name as a marker.
* We can use the exact opcode on a linkage decoration to distinguish imports from exports, which could be used to more accurately resolve symbols during the linking step.
Other than adding the decorations and making sure that AST->IR lowering adds them, the main changes here are around any code that used `IRGlobalValue`. Variables and parameters of type `IRGlobalValue*` were changed to `IRInst*` easily, so the main challenge was around code that *casts* to `IRGlobalValue*.
In cases where a cast to `IRGlobalValue` also performed a test for the mangled name being non-null/non-empty, we simply switched the code to check for the presence of an `IRLinkageDecoration`, since that is the new way of indicating a value with linakge.
Most of the serious complications arose in `ir.cpp` around the "linking"/target-specialization and generic specialization steps.
The "linking" logic was checking for `IRGlobalValue` to opt into some more complicated cloning logic, and just checking for a linkage decoration here wasn't sufficient since the front-end *does* produce global values without linkage in some cases (e.g., for a function-`static` variable we produce a global variable without linkage). This logic was updated to just check for the cases that used to amount to `IRGlobalValue`s directly by opcode. It might be simpler in the short term to have kept `IRGlobalValue` around to make the existing casts Just Work, but I'm confident that this logic could actually be rewritten for much greater clarity and simplicity and that is the better way forward.
The generic specialization logic was using some really messy code to generate a new mangled name to represent the specialized symbol, and then searching for an existing match for that name.
The original idea there was that an IR module could include "pre-specialized" versions of certain generics to speed up back-end compilation by eliminating the need to specialize in some cases, but this feature has never been implemented so the overhead here is just a waste.
Instead, I moved generic specialization to use a simpler dictionary to map the operands to a `specialize` instruction over to the resulting specialized value.
This allows for some simplifications in the name mangling logic, because it no longer needs to figure out how to produce mangled names from IR instructions representing types/values.
As part of this change I also overhauled the IR emit logic to produce cleaner output by default, borrowing some of the ideas from the logic in `emit.cpp`. IR values are now automatically given names based on their "name hint" decoration, if any, to make the code easier to follow, and I also made it so that types and literals get collapsed into their use sites in a new "simplified" IR dump mode (which is currently the default, with no way to opt into the other mode without tweaking the code). The resulting IR dumps are much nicer to look at, but as a result the one test that involves IR dumping (`ir/string-literal`) doesn't really test what it used to.
One weird issue that came up during testing is that the `transitive-interface` test had previously been producing output that made no sense (that is, the expected output file wasn't really sensible), and somehow these changes were altering its behavior. Changing the test to use `int` values instead of `float` was enough to make the output be what I'd expect, and hand inspection of generating DXBC has me convinced we were compiling the `float` case correctly too. There appears to be some issue around tests with floating-point outputs that we should investigate.
* fixup: C++ declaration order
|
|
* First pass at having an interface to write text to that can be replaced.
Simplifed and made more rigerous the interface used to write formatted strings.
* Added AppContext to simplify setting up and parsing around of streams.
* Added more simplified way to get the std error/out from AppContext.
* Work in progress using dll for tools to speed up testing.
* First pass at ISlangWriter interface.
* Added support for writing VaArgs.
Added NullWriter.
* Use ISlangWriter for output.
* Use ISlangWriter for output - replacing OutputCallback.
Make IRDump go to ISlangWriter
* SlangWriterTargetType -> SlangWriterChannel
Improvements around AppContext
* Shared library working with slang-reflection-test.
* Dll testing working for render-test.
* Include va_list definintion from header.
* Fix errors from clang.
* Fix typo for linux.
* Added -usexes option
* Fix typo.
* Fix arguments problem on linux.
* Fix typo for linux.
* Add windows tool shared library projects.
* Fix warning from x86 win build.
Fix signed warning from slang-test/main.cpp
* First attempt at getting premake to work on travis, and run tests.
* Try moving build out into script.
* Invoke bash scripts so they don't have to be executable.
* Drive configuration/tests from env parameters set by travis
* Try using source to run travis tests.
* Remove the build.linux directory - but doing so will overwrite Makefile.
* Made -fno-delete-null-pointer-checks gcc only.
* Try to fix warning from -fno-delete-null-pointer-checks
* Turn of warnings for unknown switches.
* Try to make premake choose the correct tooling.
* Disabled missing braces warning.
* Disable -Wundefined-var-template on clang.
* -Wunused-function disabled for clang.
* Fix typo due to SlangBool.
* Remove this nullptr tests.
* "-Wno-unused-private-field" for clang.
* Added "-Wno-undefined-bool-conversion"
* Add DominatorList::end fix.
* Split scripts into travis_build.sh travis_test.sh
* Fix gcc/clang template pre-declaration issue around QualType.
* Fix premake to build such that pthread correctly links with slang-glslang
|
|
* Make a test case use IR serialization
* Make all IR instructions usable as parents
This makes it so that every `IRInst` has the list of children that used to be on `IRParentInst` and eliminates `IRParentInst`.
Most places in the code were only checking against `IRParentInst` so that they could know whether there were child instructions to iterate over.
This change bloats the size of every instruction by two pointers, but we hope to be able to eliminate that overhead with a better encoding later.
* Change IR decorations to be instructions.
The main change here is that `IRDecoration` now inherits from `IRInst`, and `IRInst` now has a single linked list that holds both decorations *and* children.
At each point where code used to loop over `getChildren()` on an `IRInst`, I checked whether it made sense to leave the operation as processing just the children, or if it should process both decorations and children.
The thorniest bit was making sure the logic for inserting an instruction into a parent is correct. For the most part, once IR code is built all insertions are explicitly before/after another instruction, so the ordering can't get messed up. The sticking point is any code that does an explicit `insertAtStart` or `insertAtEnd`, but I surveyed those to make sure they are correct in context, and I also made all insertions bottleneck through one routine that does a better job of asserting the preconditions than what was there before. We may still want a "smart" insertion function at some point so that if somebody does `someDecoration->insertAtEnd(someInst)` the decoration intelligently goes to the end of the decoration list, and not the entire decorations-and-children list.
All of the existing decoration types were refactored to provide accessors for their operands, rather than directly exposing fields. In most cases the operands are required to be `IRConstant` nodes of fixed types. Not all of these types need to be kept around in the new approach, but they were left in so that as much existing code as possible can be kept working.
The `IRBuilder` was extended with factory functions to make the various decoration types and attach them.
All the fields in concrete decorations that were using `StringRepresentation` or `Name` pointers are now using IR-level string operands which provide their value as an `UnownedStringSlice`, so logic that was working with those decoration values needed to be updated here and there. I also needed to add the logic to clone string-literal values to the IR cloning pass, since they are now being used in almost every piece of code.
A new type of constant IR instruction for literal pointers was added, to handle the cases where an IR decoration needs an operand that is a raw AST-level pointer. These are even being serialized, although we obviously should not rely on them to round-trip through serialization in the future. Ideally, a follow-on change should add a cleanup pass where we remove any decorations from a module that shouldn't be allowed in the serialized code.
The biggest overall cleanup is in the serialization logic, where a lot of code just disappears because it can process the raw "decorations and children" list as the logical children of an IR instruction. The only special cases left are literals (which seem like they will always need special-casing) and global values (because they have a mangled name, which we plan to move into a decoration).
One other example of a simplification made possible by this change: the `IRNotePatchConstantFunc` instruction was implemented as an instruction only because it couldn't be encoded as a decoration at the time (it needed to have an operand that referenced an IR function).
The IR dumping logic was also updated (which meant a change to the `ir/string-literal` test) to try to make it print out all decorations a bit more systematically now that they are encoded like other instructions. The formatting isn't quite perfect, but it is good enough to be able to read what is going on.
I didn't include updates to the validation logic to ensure that decorations are being added in ways that follow the invariants, but that would be a nice thing to add next.
* fixup: 64-bit issues
* fixup: forward declaration issues
|
|
The `globallycoherent` modifier indicates that resource might be read or written by threads outside of the current thread group, so that any memory barriers that affect it should guarantee coherency at the global memory scope, and not just thread-group scope. The equivalent GLSL modifier appears to be `coherent`.
This change adds the front-end modifier, transforms it into an IR-level decoration during lowering, and then checks for the modifier during code emit.
Note: this logic may not behave correctly when `globallycoherent` is added to a field in a `struct`, since the modifier would then need to be propagated to any variables created during type legalization. Checking up on that is left to future work.
Note: it isn't entirely clear if `globallycoherent` should be treated as a declaration modifier or a type modifier. The point is moot for now because Slang doesn't have any support for type modifiers, but when we get around to that we will need to make a decision.
|
|
* First pass support for early depth stencil.
* Add a simple test to check if output has attributes.
* Use cross compilation to test [earlydepthstencil] on glsl.
* If target is dxil, use dxc to test against.
Add hlsl to test earlydepthstencil against.
* * Added spSessionHasCompileTargetSupport
* Made slang-test use spSessionHasCompileTargetSupport to ignore tests that cannot run
|
|
* Add callable shader support for Vulkan ray tracing
This change extends the previous work to update Vulkan ray tracing support for the finished `GL_NV_ray_tracing` spec.
One of the features missing in the experimental extension that was added to the final spec is "callable shaders," which allow ray tracing shaders to call other shaders as general-purpose subroutines.
Most of the implementation work here mirrors what was done for the `TraceRay()` function to map it to `traceNV()`.
We map the generic `CallShader<P>` function to the non-generic `executeCallableNV`, with a payload identifier that indicates a specific global variable of type `P` (the global variable being generated from a `static` local in `CallShader`). A new modifier is added to identify the payload structure, and the parameter binding/layout logic introduces a new resource kind for callable-shader payload data (where previously the logic had assumed ray and callable payloads should use the same resource kind).
Two test shaders are included: one for the callable shader (`callable.slang`) and one for a ray generation shader that calls it (`callable-caller.slang`). Just for kicks, the payload data type is defined in a shared file so that we can be sure the two agree (trying to emulate what might be good practice, and ensure that ray tracing support works together with other Slang mechanisms).
* Typo fix: assocaited->associated
One instance was found in review, but I went ahead and fixed a bunch since I seem to make this typo a lot.
* Typo fix: defintiion->definition
|
|
* Update version of glslang used
* Update VK raytracing support for final extension spec
A lot of this change is just plain renaming: The `NVX` suffixes become just `NV`, and the extension name changes from `GL_NVX_raytracing` to `GL_NV_ray_tracing`.
The Slang standard library and the GLSL baselines for the tests are consistently updated.
The other detail is that the final spec requires the "payload" identifier in a `traceNV()` call to be a compile-time constant, which means it cannot be defined as a local variable first, as in:
```glsl
int payloadID = 0;
traceNV(..., payloadID); // ERROR
```
In terms of how the original support was implemented, the payload ID is being computed via a special builtin function that maps each global GLSL payload variable to a unique ID. There are a few ways we could try to resolve the problem here:
1. We could aspire to put our equivalent of the `constexpr` modifier on the output of the function, so that the GLSL variable gets declared `const` and thus fits the GLSL rules for a constant expression.
2. We could introduce a pass to replace the payload-location instructions with literal integers.
3. We could use a special-purpose instruction instead of a builtin function call, and have that instruction indicate that it doesn't have side effects (so it can be folded into the call site)
4. We could somehow mark the builtin function as not having side effects.
We choose option (4) simply because it provides a feature that could have other applications. This change adds a `[__readNone]` attribute that can be applied to function declarations to express a promise on the part of the programmer that the given function has no side effects and computes its result strictly from the bits of its input arguments (and not things they point to, etc.). This mirrors an equivalent function attribute in LLVM.
We mark the function that computes a ray payload location with this attribute, and propagate the attribute through the layers of the IR, so that when the emit logic asks if an operation has side effects (to see if it can be folded into the arguments of a subsequent expression), we get an affirmative response.
This change should get all of the features that were present in the experiemntal `NVX` extension working with the final extension spec. It does not address callable shaders, which will come as a subsequent change.
|
|
This code path hadn't been used, and it had a crash due to not inserting the basic blocks it created (for initializing the variable) into the parent function. The fix adds a bit more smarts to the `IRBuilder` to help with inserting basic blocks into the flow of a function.
The actual user issue was around `static const` declarations, and it is clear that the code is incorrectly treating a function local `static const` as if it were just `static`. That will need to be fixed in another change.
|
|
* Premake work in progress for linux.
* Added dump function.
* Remove examples on linux
Small warning fix.
* * Don't build render-test on linux
* Removed work around virtual destructor warning, and just used virtual dtor for simplicity
* Git ignore obj directories
* Fix premake working on windows.
* * Fix sprintf_s functions
* Make generates arg parsing more robust
* Added FloatIntUnion to avoid type punning/strong aliasing issues, and repeated union definitions.
* Work around problems building on linux with getClass claiming a strict aliasing issue.
* Fix for targetBlock appearing potentiall used unintialized to gcc.
* Linux slang link options -fPIC to make dll.
* Add -fPIC to build options on linux.
* Add -ldl for linux on slang.
* Fixes to try and get premake working with .so on linux.
* Make core compile with -fPIC
* Try to fix linux linking with --no-as-needed before -ldl
* Add rpath back.
* Remove render-gl from linux build.
* Re-add location for linux.
* Don't include <malloc.h> except on windows.
* Remove unused line to fix warning on osx.
* Remove ambiguity on OSX for operator <<.
* Fixing ambiguity with operator overloading and Int types for OSX.
* Fix ambiguity around UInt and operator
* Fix ambiguity of UInt conversion for OSX.
* Added UnambiguousInt and UnambiguousUInt to make it easier to work around OSX integer coercion for UInt/Int types.
|
|
* Remove 'register' qualifiers.
These will be illegal come c++17 and give a warning on OSX.
* Add UNREACHABLE_RETURNs to silence compiler warnings.
* Make FileStream::GetPosition() compile on OSX
(w.r.t. the linux build, I believe that strictly-speaking, fpos64_t is specified as an opaque type and the cast to an Int64 is not necessarily well-defined.)
* Avoid an inadvertent trigraph.
|
|
* Fix sampler-less texture functions (#685)
* Fix sampler-less texeture functions
I'm honestly not sure how the original work on this feature in #648 worked at all (probably insufficient testing).
We have these front-end modifiers to indicate that a particular function definition requires a certain GLSL version, or a GLSL extension in order to be used, and they are supposed to be automatically employed by the logic in `emit.cpp` to output `#extension` lines in the output GLSL. However, it turns out that nothing is actually wired up right now, so that adding the modifiers to a declaration is a placebo.
This change propagates the modifiers through as decorations, and then uses them during GLSL code emit, which allows the functions that require `EXT_samplerless_texture_functions` to work.
* fixup: 32-bit warning
* Add serialization support for GLSL extension/version decorations
|
|
This change allows an interface to include `static` methods as requirements, so that types that conform to the interface will need to satisfy the requirement with a `static` method.
The essence of the check is simple: when checking that a method satisfies a requirement, we enforce that both are `static` or both are non-`static`.
Making that simple change and adding a test change broke a few other places in the compiler that this change tries to fix. The main fix is to handle cases where we might look up an "effectively static" member of a type through an instance, and to make sure that we replace the instance-based lookup with type-based lookup. There was already logic along these lines in `lower-to-ir.cpp`, so this change centralizes it in `check.cpp` where it seems to logically belong.
|
|
* Add a warning on missing return, and initial SCCP pass
The user-visible feature added here is a diagnostic for functions with non-`void` return type where control flow might fall off the end. This *sounds* like a trivial diagnostic to add as part of the front-end AST checking, but that can run afoul of really basic stuff like:
```hlsl
int thisFunctionisOkay(int a)
{
while(true)
{
if(a > 10) return a;
a = a*2 + 1;
}
// no return here!
}
```
This function "obviously" doesn't need to have a `return` statement at the end there, but realizing this fact relies on the compiler to understand that the `while(true)` loop can't exit normally, and doesn't contain any `break` statement. One can write "obvious" examples that need more and more complex analysis to rule out.
The answer Slang uses for stuff like this is to do the analysis at the IR level right after initial code generation (this would be before serialization, BTW, so that attached `IRHighLevelDeclDecoration`s can be used).
When lowering the AST to the IR, we always emit a `missingReturn` instruction (a subtype of `IRUnreachable`) at the end of its body if it isn't already terminated. The IR analysis pass to detect missing `return` statements is then as simple as just walking through all the functions in the module and making sure they don't contain `missingReturn` instructions.
For that simple pass to work, we first need to make some effort to remove dead blocks that control flow can never reach. This change adds a very basic initial implementation of Spare Conditional Constant Propagation (SCCP), which is a well-known SSA optimization that combines constant propagation over SSA form with dead code elimination over a CFG to achieve optimizations that are not possible with either optimization along.
For the moment, we don't actually implement any constant *folding* as part of the SCCP pass, so we can eliminate the dead block in a case like the function above (and those in the test case added in this change), but will not catch things like a `while(0 < 1)` loop. Handling more "obvious" cases like that is left for future work.
* fixup: warning on unreachable code
* Handle case where user of an inst isn't in same function/code
The code as assuming any instruction in the SSA work list has to come from the function/code being processed, but this misses the case where an instruction in a generic has a use inside the function that the generic produces.
This change adds code to guard against that case.
|
|
By default, when writing a "method" (aka "member function") in Slang, the `this` parameter is implicitly an `in` parameter. So this:
```hlsl
struct Foo
{
int state;
int getState() { return state; }
void setState(int s) { state = s; }
};
```
is desugared into something like this:
```hlsl
struct Foo { int state };
int Foo_getState(Foo this) { return this.state; }
// BAD:
void Foo_setState(Foo this, int s) { this.state = s; }
```
That "setter" doesn't really do what was intended. It modifies a local copy of type `Foo`, because `in` parameters in HLSL represent by-value copy-in semantics, and are mutable in the body the function. Slang was updated to give a static error on the original code to catch this kind of mistake (so that `this` parameters are unlike ordinary function parameters, and no longer mutable).
Of course, sometimes users *want* a mutable `this` parameter. Rather than make a mutable `this` the default (there are arguments both for and against this), this change adds a new attribute `[mutating]` that can be put on a method (member function) to indicate that its `this` parameter should be an `in out` parameter:
```hlsl
[mutating] void setState(int s) { state = s; }
```
The above will translate to, more or less:
```hlsl
void Foo_setState(inout Foo this, int s) { this.state = s; }
```
One added detail is that `[mutating]` can also be used on interface requirements, with the same semantics. A `[mutating]` requirement can be satisfied with a `[mutating]` or non-`[mutating]` method, while a non-`[mutating]` requirement can't be satisfied with a `[mutating]` method (the call sites would not expect mutation to happen).
The design of `[mutating]` here is heavily influenced by the equivalent `mutating` keyword in Swift.
Notes on the implementation:
* Adding the new attribute was straightforward using the existing support, but I had to change around where attributes get checked in the overall sequencing of static checks, because attributes were being checked *after* function bodies, but with this change I need to look at semantically-checked attributes to determine the mutability of `this`
* The check to restrict it so that `[mutating]` methods cannot satisfy non-`[mutating]` requirements was easy to add, but it points out the fact that there is a huge TODO comment where the actual checking of method *signatures* is supposed to happen. That is a bug waiting to bite users and needs to be fixed!
* While we had special-case logic to detect attempts to modify state accessed through an immutable `this` (e.g., `this.state = s`), that logic didn't trigger when the mutation happened through a function/operator call (e.g., `this.state += s`), so this change factors out the validation logic for that case and calls through to it from both the assignment and `out` argument cases.
* The error message for the special-case check was updated to note that the user could apply `[mutating]` to their function declaration to get rid of the error.
* The semantic checking logic for an explicit `this` expression was already walking up through the scopes (created during parsing) and looking for a scope that represents an outer type declaration that `this` might be referring to. We simply extend it to note when it passes through the scope for a function or similar declaration (`FunctionDeclBase`) and check for the `[mutating]` attribute. If the attribute is seen, it returns a mutable `this` expression, and otherwise leaves it immutable.
* The IR lowering logic then needed to be updated so that when adding an IR-level parameter to represent `this`, it gives it the appropriate "direction" based on the attributes of the function declaration being lowered. The rest of the IR logic works as-is, because it will treat `this` just like an other parameter (whether it is `in` or `inout`).
* This biggest chunk of work was the "implicit `this`" case, because ordinary name lookup may resolve an expression like `state` into `this.state`, so that the `this` expression comes out of "thin air." To handle this case, I extended the structure of the "breadcrumbs" that come along with a lookup result (the breadcrumbs are used for any case where a single identifier like `state` needs to be embellished to a more complex expression as a result of lookup), so that it can identify whether a `Breadcrumb::Kind::This` node comes from a `[mutating]` context or not. Similar to the logic for an explicit `this`, we handle this by noting when we pass through a `FunctionDeclBase` when moving up through scopes, and look for the `[mutating]` attribute on it. The rest of the work was just plumbing the additional state through.
|
|
* Move to newer glslang
* Support cross-compilation of ray tracing shaders to Vulkan
This change allows HLSL shaders authored for DirectX Raytracing (DXR) to be cross-compiled to run with the experimental `GL_NVX_raytracing` extension (aka "VKRay").
* The GLSL extension spec is marked as experimental, so that any shaders written using this support should be ready for breaking changes when the spec is finalized.
* "Callable shaders" are not exposed throug the GLSL extension, so this feature of DXR will not be cross-compiled.
* The experimental Vulkan raytracing extension does not have an equivalent to DXR's "local root signature" concept. This does not visibly impact shader translation (because the local/global root signature mapping is handled outside of the HLSL code), but in practice it means that applications which rely on local root signatures on their DXR path will not be able to use the translation in this change as-is; more work will be needed.
The simplest part of the implementation was to go into the Slang standard library and start adding GLSL translations for the various DXR operations.
In some cases, like mapping `IgnoreHit()` to `ignoreIntersectionNVX()` this is almost trivial.
The various functions to query system-provided values (e.g., `RayTMin()`) were also easy, with the only gotcha being that they map to variables rather than function calls in GLSL, and our handling of `__target_intrinsic` assumes that a bare identifier represents a replacement function name, and not a full expression, so we have to wrap these definitions in parentheses.
The tricky operations are then `TraceRay<P>()` and `ReportHit<A>()`, because these two are generics/templates in HLSL.
GLSL doesn't support generics, even for "standard library" functions, so the raytracing extension implements a slightly complex workaround: the matching operations `traceNVX()` and `reportIntersectionNVX()` pass the payload/attributes argument data via a global variable.
That is, shader code for the GLSL extensions writes to the global variable and then calls the intrinsic function.
The linkage between the call site and the global is established by a modifier keyword (`rayPayloadNVX` and `hitAttributeNVX`, respectively) and in the case of ray payload also uses `location` number to identify which payload global to use (since a single shader can trace rays with multiple payload types).
Our translation strategy in Slang tries to leverage standard language mechanisms instead of special-case logic.
For example, to translate the `ReportHit<A>()` function, we provide both a default declaration that will work for HLSL (where the operation is built-in with the signature given), and a *definition* marked with the `__specialized_for_target(glsl)` modifier.
The GLSL definition declares a function `static` variable that will fill the role of the required global, and then does what the GLSL spec requires: assigns to the global, and then calls the `reportIntersectionNVX` builtin (which we declare as a separate builtin).
Our ordinary lowering process will turn that `static` variable into an ordinary global in the IR, and the `[__vulkanHitAttributes]` attribute on the variable will be emitted as `hitAttributeNVX` in the output.
There is no additional cross-compilation logic in Slang specific to `ReportHit<A>()` - the target-specific definition in the standard library Just Works.
The case for `TraceRay<P>()` is a bit more complicated, simply because the GLSL `traceNVX()` function needs to be passed the `location` for the payload global.
We implement the payload global as a function-`static` variable, with the knowledge that every unique specialization of `TraceRay<P>()` will generate a unique global variable of type `P` to implement our function-`static` variable.
We then add a slightly magical builtin function `__rayPayloadLocation()` that can map such a variable to its generated `location`; the logic for this is implemented in `emit.cpp` and described below.
We also changed the `RayDesc` and `BuiltinTriangleIntersectionAttributes` types from "magic" intrinsic types over to ordinary types (because the GLSL output needs to declare them as ordinary `struct` types).
This ends up removing some cases in the AST and IR type representations.
By itself this change would break HLSL emit, because in that case the types really are intrinsic.
We added a `__target_intrinsic` modifier to these types to make them intrinsic for HLSL, and then updated the downstream passes to handle the notion of target-intrinsic types.
The logic for binding/layout of entry point inputs and outputs was updated so that raytracing stages don't follow the default logic for varying input/output parameters.
This is because the input/output parameters of a raytracing entry point aren't really "varying" in the same sense as those in the rasterization pipeline.
In particular, the SPIR-V model for raytracing input and output treats "ray payload" and "hit attributes" parameters as being in a distinct storage class from `in` or `out` parameters.
We also detect cases where a ray tracing stage declares inputs/outputs that it shouldn't have. This logic could conceivably be extended to other stages (e.g., to give an error on a compute shader with user-defined varying input/output).
The type layout logic added cases for handling raytracing payload and hit-attribute data, but this is currently just a stub implementation that follows the same logic as for varying `in` and `out` parameters (it cannot give meaningful byte sizes/offsets right now).
To my knowledge the GLSL spec doesn't currently specify anything about layout, and I haven't read the DXR spec language carefully enough to know what it says about layout.
A future change should update the layout logic to allow for byte-based layout of ray payloads, etc. so that we can query this information via reflection.
The GLSL legalization logic in `ir.cpp` was updated to factor out the per-entry-point-parameter code into its own function, and then that function was updated to special-case the input/output of a ray-tracing shader.
While for rasterization stages we typically want to take the user-declared input/output and "scalarize" it for use in GLSL (in part to deal with language limitations, and in part to tease system values apart from user-defined input/output), the GLSL spec for raytracing requires payload and hit attribute parameters to be declared as single variables. There is also the issue that even for an `in out` parameter, a ray payload parameter should only turn into a single global, whereas the handling for varying `in out` parameters generates both an `in` and an `out` global for the GLSL case.
Other than the handling of entry point parameters, the GLSL legalization pass doesn't need to do anything special for ray tracing shaders.
The trickiest change in the `emit.cpp` logic is that we now generate `location`s for ray payload arguments (the outgoing from a `TraceRay()` call) on demand during code generation.
This is a bit hacky, and it would be nice to handle it as a separate pass on the IR rather than clutter up the emit logic, but this approach was expedient.
Basically, any of the global variables that got generated from the `static` declarations in the standard library implementation of `TraceRay()` will trigger the logic to assign them a `location`.
The logic for emitting intrinsic operations added a few new `$`-based escape sequences. The `$XP` case handles emitting the location of a generated ray payload variable; this is how we emit the matching location at the site where we call `traceNVX`. The `$XT` case emits the appropriate translation for `RayTCurrent()` in HLSL, because it maps to something different depending on the target stage.
All of the test cases here consist of a pair of an HLSL/Slang shader written to the DXR spec, plus a matching GLSL shader for a baseline.
The GLSL shaders are carefully designed so that when fed into glslang they will produce the same SPIR-V as our cross-compilation process.
This kind of testing is quite fragile, but it seems to be the best we can do until our testing framework code supports *both* DXR and VKRay.
A bunch of the core changes ended up being blocked on issues in the rest of the compiler, so some additional features go implemented or fixed along the way:
The first big wall this work ran into was that the `__specialized_for_target` modifier hasn't actually been working correctly for a while.
It turns out that for the one function that is using it, `saturate()`, we have been outputting the workaround GLSL function in *all* cases (including for HLSL output) rather than only on GLSL targets.
The problem here is that for a generic function with a `__specialized_for_target` modifier or a `__target_intrinsic` modifier, the IR-level decoration will end up attached to the `IRFunc` instruction nested in the `IRGeneric`, but the logic for comparing IR declarations to see which is more specialized (via `getTargetSpecializationLevel()`) was looking only at decorations on the top-level value (the generic).
The quick (hacky) fix here is to make `getTargetSpecializationLevel()` try to look at the return value of a generic rather than the generic itself, so that it can see the decorations that indicate target-specific functions.
A more refined fix would be to attach target-specificity decorations to the outer-most generic (to simplify the "linking" logic).
The only reason not to fold that into the current fix is that the `__target_intrinsic` modifier currently serves double-duty as a marker of target specialization *and* information to drive emit logic. The latter (the emit-related stuff) currently needs to live on the `IRFunc`, and moving it to the generic could easily break a lot of code.
This needs more work in a follow-on fix, but for now target specialization should again be working.
The other big gotcha that the simple "just use the standard library" strategy ran into was that function-`static` variables weren't actually implemented yet, and in particular function-`static` variables inside of generic functions required some careful coding.
The logic in `lower-to-ir.cpp` has this `emitOuterGenerics()` function that is supposed to take a declaration that might be nested inside of zero or more levels of AST generics, and emit corresponding IR generics for all those levels.
This is needed because two different AST functions nested inside a single generic `struct` declaration should turn into distinct `IRFunc`s nested in distinct `IRGeneric`s.
The tricky bit to making that all work is that the same AST-level generic type parameter will then map to *different* IR-level instructions (the parameters of distinct `IRGeneric`s) when lowering each function.
The existing logic handled this in an idiomatic way by making "sub-builders" and "sub-contexts."
This change refactors some of the repeated logic into a `NestedContext` type to help simplify the pattern, and applies it consistently throughout the `lower-to-ir.cpp` file.
Besides that cleanup, the major change is `lowerFunctionStaticVarDecl` which, unsurprisingly, handles lower of function-`static` variables to IR globals.
The careful handling of nested contexts here is needed because if we are in the middle of lowering a generic function, then a `static` variable should turn into its *own* `IRGeneric` wrapping an `IRGlobalVar`. The body of the function should refer to the global variable by specializing the global variable's `IRGeneric` to the parameters of the *functions* `IRGeneric`. This tricky detail is handled by `defaultSpecializeOuterGenerics`.
An additional subtlety not actually required for this raytracing work (and thus not properly tested right now) is handling function-`static` variables with initializers.
These can't just be lowered to globals with initializers, because HLSL follows the C rule that function-`static` variables are initialized when the declaration statement is first executed (and this could be visible in the presence of side-effects).
The lowering strategy here translates any `static` variable with an initializer into *two* globals: one for the actual storage, plus a second `bool` variable to track whether it has been initialized yet.
There are some opportunities to optimize this case, especially for `static const` data, but that will need to wait for future changes.
We've slowly been shifting away from the model where a user thinks of a "profile" as including both a stage and a feature level.
Instead, the user should think about selecting a profile that only describes a feature level (e.g., `sm_6_1`, `glsl_450`, etc.), and then separately specifying a stage (`vertex`, `raygeneration, etc.) for each entry point.
The challenge here is that the command-line processing still only had a single `-profile` switch, and no way to specify the stage.
Adding the `-stage` option was relatively easy, but making it work with the existing validation logic for command-line arguments was tricky, because of the complex model that `slangc` supports for compiling multiple entry points in a single pass.
* In `slang.h` add new reflection parameter categories for ray payloads and hit attributes, as part of entry point input/output signatures.
* A previous change already updated our copy of glslang to one that supports the `GL_NVX_raytracing` extension, so in `slang-glslang.cpp` we just needed to map Slang's `enum` values for the raytracing stage names to their equivalents in the glslang code.
* Moved the logic for looking up a stage by name (`findStageByName()`) out of `check.cpp` and into `compiler.cpp`, with a declaration in `profile.h`
* Added a `$z` suffix to the GLSL translation of `Texture*.SampleLevel()`, to handle cases where the texture element type is not a 4-component vector. Note that this fix should actually be applied to *all* these texture-sampling operations, but I didn't want to add a bunch of changes that are (clearly) not being tested right now.
* The layout logic for entry points was updated to correctly skip producing a `TypeLayout` for an entry point result of type `void`, which meant that the related emit logic now needs to guard against a null value for the result layout.
* In `ir.cpp`, dump decorations on every instruction instead of just selected ones, so that our IR dump output is more complete.
* Added a command-line `-line-directive-mode` option so that we can easily turn off `#line` directives in the output when debugging. Not all cases where plumbed through because the `none` case is realistically the most important.
* Parser was fixed to properly initialize parent links for "scope" declarations used for statements, so that we can walk backwards from a function-scope variable (including a `static`) and see the outer function/generics/etc.
* Added GLSL 460 profile, since it is required for ray tracing. Also updated the logic for computing the "effective" profile to use to recognize that GLSL raytracing stages require GLSL 460.
* Added some conventional ray-tracing shader suffixes to the handling in `slang-test`. This code isn't actually used, but was relevant when I started by copy-pasting some existing VKRay shaders as the starting point for my testing.
* Fixup: typos
|
|
* Change the layout of IROp such that 'main' IROps are 0-x. (#649)
* Removed MANUAL_RANGE instuction types, as no longer needed.
|
|
* * Added support for strings in IR with IRStringLit - with storage of chars after it
* Added kIRDecorationOp_Transitory - can be used for detecting instructions constructed on stack
* Made IRConstant hashing work off type
* Fix comment that is out of date about how an instruction is determines to hold a transitory string.
|
|
* * Remove dispose from IRInst
* Use MemoryArena instead of MemoryPool
* Make all IRInst not require Dtor - by having ref counted array store ptrs that need freeing
* Increase block size - typically compilation is 2Mb of IR space(!)
* Fix issues around StringRepresentation::equal because null has special meaning.
* Don't bother to construct as String to compare StringRepresentation, just used UnownedStringSlice.
* Added fromLiteral support to UnownedStringSlice and use instead of strlen version.
* Use more conventional way to test StringRepresentation against a String.
* Fix gcc/clang template problem with cast.
|
|
Partially fixes #615
There's kind of a mess going on here, and it is difficult to be sure which of the changes here are strictly necessary.
Also, our testing isn't setup to run tests that use `RWTexture2D`, so the only testing I can really run is manual tests using Falcor.
The most basic issue here is that in an earlier change I added `ref` accessors for the subscript operation on various `RW*` types in the standard library, and that included `RWTexture2D` (and the other `RWTexture*` types). The compiler ended up favoring a `ref` accessor over a `set` accessor even when the `set` would suffice, but only the `set` accessor could be lowerd to GLSL/SPIR-V.
This change ends up implementing two different fixes for the same problem:
* Logic has been added to try and favor a `set` accessor over a `ref` accessor in the cases where either could be used (but still require a `ref` accessor to be used when it is really needed)
* The `ref` accessor for `RWTexture*` has been removed, since it turns out that the operations that might have benefited from it (atomics, and component-granularity stores) aren't actually allowed on typed UAVs anyway.
There is a deeper issue here that somebody needs to go through and rationalize our representation and handling of accessors like this, but I'm not going to be able to do that in the time I can put into this PR.
|
|
* Fix typo OuptutTopologyAttribute -> OutputTopologyAttribute
First pass support for handing tesselation shaders - domain and hull.
* Added attribute PatchConstantFuncAttribute
* Added visitHLSLPatchType(HLSLPatchType* type) such that the patch type template parameters are handled
* Added IRNotePatchConstantFunc - such that the patch constant function is referenced within IR
* Added support for outputing typical tesselation attributes (although minimal validation is performed)
* Added findFunctionDeclByName
* Small improvements to diagnostic.
* Improved diagnostics and checking for geometry shader attributes.
* Added diagnostic if patchconstantfunc is not found
Handle assert failure when outputing a domain shader alone and therefore attr->patchConstantFuncDecl is not set.
* Simple script tess.hlsl to test out domain/hull shaders.
* Added url for where hull shader attributes are defined.
* Fix unsigned/signed comparison warning.
* Restore removal of fix in "Improve generic argument inference for builtins (#598)"
* Update tessellation test case to compare against fxc
The test was previously comparing against fixed expected DXBC output, but this caused problems when the test runner tried to execute the test on Linux (where there is no fxc to invoke...), and would also be a potential source of problems down the road if different users run using different builds of fxc.
The simple solution here is to convert the test to compare against fxc output generated on the fly. That test type is already filtered out on non-Windows builds, so it eliminates the portability issue (in a crude way).
I also changed the test to compile both entry points in one compiler invocation, just to streamline things into fewer distinct tests.
* Eliminate unnecessary call to `lowerFuncDecl`
In a very obscure case this could cause a bug, if the patch-constant function had somehow already been lowered (because it was called somewhere else in the code).
The call should not be needed because `ensureDecl` will lower a declaration on-demand if required, so eliminating it causes no problems for code that wouldn't be in that extreme corner case.
|
|
The problem here arose when a complicated l-value was formed like:
```hlsl
struct Foo { float4 a; }
RWStructuredBuffer<Foo> gBuffer;
gBuffer[index].a.xz += whatever;
```
In this case the `gBuffer[index].a.xz` expression is a complex l-value in multiple ways:
* The `gBuffer[index]` subscript could be routed to either a `get` accessor or a `ref` accessor (and maybe also a `set` accessor if we add one to the stdlib definition), and we defer the choice of which to call until as late as possible in codegen today.
* The `_.a` part then becomes a "bound member acess" because we can't actually produce a direct pointer until we've resolved how to implement the subscript operation.
* The `_.xz` part becomes a "swizzled l-value" because there is *no* way to materialize it as a pointer to contiguous storage in the orignal object (the `x` and `z` components of a vector aren't contiguous).
Recent changes to support atomic operations on buffer elements introduced the `ref` accessor on `RWStructuredBuffer`, which made it possible to form a pointer to a buffer element in the IR. This interacted with some code for the "bound member" case that was trying to only introduce a temporary when absolutely necessary, and was doing so by assuming anything with an address didn't need to be moved into a temporary.
The first fix is to clean up that logic in the bound-member case for assignment: always create a temporary, rather than do it conditionally.
The second fix here is more systemic: we add logic to try to coerce the representation of an l-value during codegen into being a simple address, and employ that in cases where we know an address is desired. In a case like the above this helps to get things into the form that is required, so that a swizzled store can be issued.
There is still some potential for cleanup in this logic, but I don't want to introduce more changes than seem necessary to fix the original problem.
|
|
Slang `enum` declarations will always be scoped, e.g.:
```hlsl
enum Color
{
Red,
Green = 2,
Blue,
}
Color c = Color.Red; // Not just `Red`
```
A user can write `enum class` as a placebo for now (to ease sharing of headers with C++).
Slang does not currently support the `::` operator for static member lookup, so it must be `Color.Green` and not `Color::Green`. Support for `::` as an alternate syntax could be added later if there is strong user demand.
An `enum` type can have a declared "tag type" using syntax like C++ `enum class`:
```hlsl
enum MyThings : uint
{
First = 0,
// ...
}
```
The `enum` cases will store their values using that type. An `enum` that doesn't declare a tag type will use the type `int` by default.
Enum cases are assigned values just like in C/C++: cases can have explicit values, but otherwise default to one more than the previous case, or zero for the first case.
All `enum` types will automatically conform to a standard-library `interface` called `__EnumType`, which is used so that basic operators like equality testing can be defined generically for all `enum` types.
This change only adds one operator at first (the `==` comparison), but other should be added later.
An `enum` case needs to be explicitly converted to an integer where needed (e.g., `int(Color.Red)`).
This is implemented by having the main integer types (`int` and `uint`) support built-in initializers that can work for *any* `enum` type (or rather, anything conforming to `__EnumType`).
Eventually these will be restricted so that an `enum` type can only be converted to its associated tag type.
IR code generation completely eliminates `enum` types and their cases.
The `enum` type will be replaced with its tag type, and the cases will be replaced with the tag values.
Currently this could leave some mess in the IR where cast operations are applied between values that actually have the same type.
|
|
* Fix atomic operations on RWBuffer
An earlier change added support for passing true pointers to `__ref` parameters to fix the global `Interlocked*()` functions when applied to `groupshared` variables or `RWStructureBuffer<T>` elements.
That change didn't apply to `RWBuffer<T>` or `RWTexture2D<T>`, etc. because those types had so far only declared `get` and `set` accessors, but not any `ref` accessors (which return a pointer).
The main fixes here are:
* Add `ref` accessors to the subscript oeprations on the `RW*` resource types
* Adjust the logic for emitting calls to subscript accessors so that we don't get quite as eager about invoking a `ref` accessor, and instead try to invoke just a `get` or `set` accessor when these will suffice. This is important for Vulkan cross-compilation, where we don't yet support the semantics of our `ref` accessors.
* Add a test case for atomics on a `RWBuffer`
* Fix up `render-test` so that we can specify a format for a buffer resource, which allows us to use things other than `*StructuredBuffer` and `*ByteAddressBuffer`. The work there is probably not complete; I just did what I could to get the test working.
* A bunch of files got whitespace edits thanks to the fact that I'm using editorconfig and others on the project seemingly arent...
* fixup: remove ifdefed-out code
|
|
Fixes #581
This change adds a new parameter passing mode `__ref` to exist alongisde `in`, `out`, and `inout`.
The `__ref` modifier indicates true by-reference parameter passing (whereas `inout` is copy-in-copy-out).
This is not intended to be something that users interact with directly, but rather a low-level feature that lets us provide a correct signature for the `Interlocked*()` operations in the standard library.
Most of the support for passing what are logically addresses around already exists in the IR, so the majority of the work here is just in introducing the new type `Ref<T>` and then using it appropriately when lowering `__ref` parameters/arguments to the IR.
|
|
* Cleanups around behavior when the compiler fails
* Add another case where we try to `noteInternalErrorLoc()` if an exception in thrown. This one is the in the logic for emitting an IR instruciton. This could be improved by adding another layer at the function level (as a catch-all for instructions with no location), but something is better than nothing.
* Change a bunch of `assert()`s over to `SLANG_ASSERT()`s, so that we can theoretically take more control over them (e.g., make release builds with asserts enabled)
* Some other small cleanups around the assertions we perform.
In the survey I made, I didn't really see many obvious "smoking gun" cases where we could produce a significantly better error message for some of the unimplemented/unexpected paths, other than to actually implement the missing functionality.
* fixup
|
|
|
|
The basic idea here is that when lowering to the IR, the front-end will attach a "name hint" to the IR instruction(s) that represent a given declaration, and then the passes that work on the IR will try to preserve and propagate those names, and then finally the emit logic will use them in place of mangled or unique names when available.
This change does *not* try to deal with the issues that arise when we try to use those variable names in the output without any modification (e.g., handling cases where they might clash with keywords or builtins in the target language). Instead, it tries to establish baseline behavior for propagating through names, so that a later change can concentrate on the issue of using those names exactly when it is legal to do so.
In order to avoid issues around the name "hints" causing problems we take two main steps:
1. We "scrub" each name to reduce it down to the allowed set of identifier characters in C-like languages, and then ensure that it doesn't do things that would be illegal in some downstream languages (e.g., consecutive underscores are not allowed in GLSL) or could clash with Slang's mangled names. This process isn't guaranteed to give distinct results for distinct inputs (it isn't a mangling scheme, after all).
2. We generate a unique ID for each occurence of a given name and always use that as a suffix. This means that even if a name happens to overlap with a keyword (if you somehow have a variable named `do`), we will still add a suffix that makes it not a problem (we'd output `do_0` which is fine).
The logic for generating these names is mostly straightforward. For simple variables, we use their given name directly, while for other declarations we try to form a name that includes their parent declaration (e.g. `SomeType.someMethod`).
Various IR passes need to propagate or preserve this information. The most interesting is type legalization, when we take a variable with an aggregate type and split some of the fields out into their own variables. In that case we generate "dotted" names like `someVar.someTexture` and rely on the emit logic to turn that into `someVar_someTexture`.
During SSA generation, if we are promoting a variable to SSA temporaries, we will try to propagate the name of the variable over to the temporaries (unless they already have a name from some other place). The same applies to block parameters ("phi nodes").
Many of the test changes need their expected output to be updated for this change. Luckily in most cases the output has gotten easier to understand.
|
|
This was a known issue in our IR representation, which was now biting a user. The basic problem is that in code like the following:
```hlsl
RWStructureBuffer<float4> buffer;
...
buffer[index].xz = value;
```
we ideally want to be able to reproduce the original HLSL code exactly, but that requires directly encoding the way that this code writes to two elements of a vector, but not the others.
The currently lowering strategy we had produced IR something like:
```hlsl
float4 tmp = buffer[index];
tmp.xz = value;
buffer[index] = tmp;
```
That transformation might seem valid, but it has some big problems:
* It generates UAV reads that are not needed, which could impact performance
* It performs read-modify-write operations on memory that the programmer didn't explicitly write, which could create data races
The fix here is somewhat obvious: if the "base" of a swizzle operation on a left-hand side resolves to a pointer in our IR, then we can output a "swizzled store" instead of the read-modify-write dance. We currently keep the read-modify-write around since it is potentially needed as a fallback in the general case.
Along the way I also tried to make sure that we handle the case where we have a swizzle of a swizzle on the left-hand side:
```hlsl
buffer[index].xz.y = value;
```
That code should behave the same as `buffer[index].z = value`. I am currently detecting and cleaning up this logic in the lowering path for `SwizzleExpr`, because that is the only place in the lowering logic that "swizzled l-values" currently get created.
|
|
Fixes #527
There were a few problem cases for the IR emit logic. The most obvious, which came up in #527 is that a function body with multiple `return` statements would generate invalid code:
```hlsl
int foo()
{
return 1;
int x = 2;
return x;
}
```
In that case the IR for `foo` would have a single block that has two `return` instructions, which is invalid.
Another case that seems to be arising more often, but that had less obvious consequences was when one arm of an `if` statement ends in a `return`:
```hlsl
if(a)
{
return b;
}
else
{
int c = 0;
}
int d = 0;
```
In that case, the `return` instruction for `return b` would be followed by a branch to the end of the `if` (the `int d = 0;` line), because that would be the normal control flow without the early `return`.
The fix implemented here is to have the IR lowering logic be a bit more careful on two fronts:
1. When emitting a branch, check if the block we are emitting into has already been terminated, and if so just don't emit the branch (since we are logically at an unreachable point in the CFG.
2. Whenever we are about to emit code for a (non-empty) statement, ensure that the current block being build is unterminated. If the current block is terminated, then start a new one.
Case (2) will only matter when there is unreachable code (e.g., in the function `foo()`, the declaration of `x` and the second `return` can never be reached), so I added a warning in that case, and included a test case that triggers the new warning (with a function like `foo()` above).
|
|
* Improve messages when compilation is aborted.
Make sure to include the information from any `Slang::Exception` that was thrown, so that the poor user can at least point us at our own message string from an assertion failure.
This doesn't provide them line-number information in their code or the Slang codebase, so there is still work to be done in making the compiler more friendly about this stuff.
* When aborting compilation, try to note what source location we were working on
This is handled by having exception handlers on the stack at key bottleneck points in semantic checking and IR generation, which can then emit a diagnostic to note what we were working on when things failed.
This is not intended to be an indiciation to the user that their code is at fault for a compiler crash (it is always our fault), but might give them a chance to work around whatever bug is blocking them.
|
|
* There was a simple typo where we were emitting `RaytracingAccelerationStructureType` instead of `RaytracingAccelerationStructure`
* The IR lowering logic was failing to handle types with an `__intrinsic_type` modifier (which maps them to a single IR opcode) that weren't in one of the various special cases. I added a catch-all case to the handling of `DeclRefType`. This notably affected the `RayDesc` type.
* Even if we lower `RayDesc` to an intrinsic type, we still need to lower its *fields* too, and these were getting emitted with mangled names (as would happen for any user-defined fields). The solution I implemented was to allow for fields to have `__target_intrinsic` modifiers in the stdlib, to specify the un-mangled name they should use on each target.
I'm not 100% happy with this solution, because it seems odd to have `RayDesc` be an intrinsic type, but then to also have field keys used in `getField` instructions as if it were an ordinary `struct`. It seems like a better solution would be to have it lower to an IR `struct`, just with an appropriate modifier.
|
|
The basic problem was that the lowering logic was constructing (more or less) `Ptr<@GroupShared X>` instead of `@GroupShared Ptr<X>`.
There were also problems with passes not propagating through rates that should have been (e.g., legalization).
I've added a test case to actually validate `groupshared` support.
|
|
* Introduce an IR-level type system
Up to this point, the Slang IR has used the front-end type system to represent types in the IR.
As a result (but ultimately more importantly) the IR representation of generics and specialization has used AST-level concepts embedded in the IR.
For example, to express the specialization of `vector<T,N>` to a concrete type `float` for `T`, we needed an IR operation that could represent the specialization, with operands that somehow represented the type argument `float`.
The whole thing was very complicated.
The big idea of this change is to introduce a new representation in which types in the IR are just ordinary instructions, so that using them as operands makes sense. The hierarchy of IR types closely mirrors the AST-side hierarchy for now, and that will probably be something we should maintain going forward.
In order to make these changes work, though, I also had to do major overhauls of things like the way substitutions are performed, how we check interface conformances, the way lookup through interface types is done, etc. etc. This is a big change, and unfortunately any attempt to summarize it in the commit message wouldn't do it justice.
* Fix 64-bit build warning
* Fix up some clang warnings/errors
|
|
Fixes #471
|
|
Fixes #61
When lowering from AST to IR, if a call site doesn't supply an argument expression for each of the parameters to the callee, then use the default value expressions (stored as the "initializer" of the parameter decl) for each omitted parameter. This relies on the front-end to have already checked the call site for validity.
Along the way I also cleaned up some of the checking of parameter declarations so that it is more like the checking of ordinary variable declarations (although the code is not yet shared). I also cleaned out some dead cases in the lowering logic for when we don't actually have a declaration available for a callee (these would only matter if we supported functions as first-class values).
I added a simple test case to confirm that call sites both with and without the optional parameter work as expected.
The strategy in this change is extremely simplistic, and might only be appropriate for default parameter value expressions that are compile-time constants (which should be the 99% case). This may require a major overhaul if we decide to handle default parameter values differently (e.g., by generating extra functions to ensure that the separate compilation story is what we want).
Another issue that could change a lot of this logic would be if we start to support by-name parameters at call sites, since we could no longer assume that the argument and parameter lists align one-to-one (with the argument list possibly being shorter). Any work to add more flexible argument passing conventions would need to build a suitable structure to map from arguments to parameters, or vice-versa.
|
|
* Typo
* Add [shader(...)] and clean up some literal handling
* Add supporting for validating the `[shader(...)]` attribute, by checking that its argument is a string literal that names a known shader stage.
* Split the `ConstantExpr` class into distinct subclasses rooted at `LiteralExpr`, so we have `BoolLiteralExpr`, `IntegerLiteralExpr`, `FloatingPointLiteralExpr`, and `StringLiteralExpr`
* Add a `String` type to the stdlib, to be used as the type of a string literal.
This change allows code using `[shader(...)]` to be accepted by the front-end again, but it does nothing about emitting it in final HLSL.
* Allow entry points to be specified via [shader(...)]
Before this change, the compiler would track a list of `EntryPointRequest` objects, based on what the suer specified via API and/or command-line options. Each entry point request would get matched up with an AST `FuncDecl` as part of semantic checking, and then the back end steps (layout, codegen, etc.) would work from that information.
This change makes the compiler modal, in that it can *either* continue to use an explicit list of entry point requests (this is the mode when the list is non-empty), or it can rely on user-supplied attributes on entry point functions to drive codegen (this is the mode when the list is empty).
User-specified `[shader(...)]` attributes are processed at the same place where the association from `EntryPointRequest`s to `FuncDecl`s would otherwise be made, and basically does the same thing in the opposite direction: looks for `FuncDecl`s with the appropriate attribute and synthesizes an `EntryPointRequest` for them.
Subsequent processing should ideally not know where a given `EntryPointRequest` came from, and should handle both methods of specifying the entry points equivalently.
One design choice that might not make immediate sense is that we do *not* process a function as an entry point (applying further validation, etc.) just because it has a `[shader(...)]` modifier, unless we are in the appropriate mode (which in this case is the mode where the user didn't specify their own entry points via API or command line). This is to handle cases where the user wants to explicitly compile only one entry point, so that they (1) don't want us to spend time validating code they don't care about, (2) don't want do get output they don't expect, and (3) might actually be presenting us with code that violates the language rules due to a combination of `#define`s in effect (e.g., they might have a `[shader("vertex")]` function that transitively executes a `discard` because of how the preprocessor was configured, but they don't care because they are compiling a fragment entry point). This decision might be something we revisit over time.
As part of this work, I had to add some logic to pick a "profile version" to use for a combination of a target and stage (because when you specify `[shader("vertex")]` the compiler can't tell if you want `vs_5_0`, `vs_5_1`, etc.). This isn't really complete right now, because something like `-target dxbc` *also* doesn't determine a profile, so there is a bit of a kludge at present. We need to figure out a good long-term plan here, which might involve keeping target format, feature level/version, and pipeline stage as truly orthogonal concepts, rather than conflating them. That would involve more work in the API and command-line layers to de-compose things when the user specifies, e.g., `vs_5_1`, but might make downstream logic easier to manage.
* Emit [shader(...)] attribute on entry point for SM 6.1 and later
This should help ensure that the output from Slang can be compiled with dxc `lib_*` profiles.
* Fix warning
|
|
This commit contains two small bug fixes:
1. In `specializeProgramLayout`, we cannot assume the resourceInfo entries in a varLayout and its corresponding type layout has the same order. Should use `FindResourceLayout`.
2. When generating ir for a switch statement, make sure to remove the breakLabel from the shared context when done. For some reason if a switch statement is being lowered twice, the Dictionary::Add method will complain the statement key already exists.
|
|
The existing code parsed all of the square-bracket `[attributes]` into `HLSLUncheckedAttribute`, and then went on to hand-convert some of them to specialized subclasses of `HLSLAttribute`. When attributes didn't check, they were left as-is, and no error message was issued, because at the time the compiler was focused on accepting arbitrary input.
This change greatly overhauls the handling of `[attributes]`. Attributes are now declared in the stdlib, with declarations like:
```hlsl
__attributeTarget(LoopStmt)
attribute_syntax [unroll(count: int = 0)] : UnrollAttribute;
```
In this syntax, the `unroll` part is giving the attribute name (the `[]` are just for flavor, to make the declaration look like a use site; we could drop it if we don't like the clutter), the `count` is a parameter of the attribute, which we expect to be of type `int`, and which has a default value of `0` if unspecified.
The `: UnrollAttribute` part specifies the meta-level C++ class that will implement this attribute (and corresponds to a class in `modifier-defs.h`). This syntax is similar to our current `syntax` declarations. I'm starting to think we should change it to something like a `__meta_class(UnrollAttribute)` modifier, and then use that uniformly across all cases (e.g., also replacing the curreent `__magic_type(Foo)` syntax).
The `__attributeTarget(LoopStmt)` is a modifier that specifies the meta-level C++ class for syntax that this attribute is allowed to attach to. It is legal to have more than one of these.
Attributes continue to be parsed in an unchecked form, so that we don't tie up semantic analysis and parsing more than necessary. During checking, we look up the attribute name in the current scope, and then replace the unchecked attribute with a more specific one *if* the checking passes.
Checking proceeds in generic and attribute-specific phases. The generic phase includes checking the number of arguments against those specified in the attribute declaration (I don't currently check types, or handle default arguments), and then checking that at least one `__attributeTarget(...)` modifier applies to the syntax node being modified.
The attribute-specific phase then applies to the specialized C++ subclass of `Attribute`, and does the actual checking right now (e.g., that step is responsible for actually type-checking things at present). This can obviously be improved over time.
With this support I went ahead and added declarations for all the HLSL attributes I could find documented on MSDN. I also added a provisional declaration for the `[shader(...)]` attribute that has been added to dxc, but which is not yet documented.
One important detail here is that lookup of attribute names needs to be done carefully, so that we don't let, e.g., local variables shadow an attribute declaration:
```hlsl
int unroll = 5;
// This attribute should *not* get confused by the local variable `unroll`
[unroll] for(...) { .. }
```
The lookup logic already has a notion of a `LookupMask` that can be used to filter declarations out of the result. In this change I surfaced that mask through the main lookup API (rather than requiring a second pass to "refine" lookup results), and made is so that the default lookup mask does *not* include attributes, while an explicit mask can be used to look up *only* attributes.
(An alternatie design we discussed was to follow the approach of C# and have the declaration of an attribute like `[unroll]` actually be `unrollAttribute`, with a suffix. I decided not to follow that approach for now because it seemed like printing good error messages in that case could require us to carefully trim the `Attribute` suffix off of names at times, and using the existing mask behavior seemed simpler.)
To verify that the shadowing behavior is indeed correct, I modified the `loop-unroll.slang` test case.
Smaller notes:
* Removed the `HLSL` prefix from several of the C++ attribute classes
* Made sure to actually validate the modifiers on statements
* Special-cased checking for `ParamDecl` with a null type, because I'm re-using `ParamDecl` for attribute parameters, but can't give a concrete type to some of them right now
* Deleting some old, dead emit-from-AST logic around attributes, rather than try to "fix" code that doesn't run (a more complete scrub of that code is still needed)
* Fixed AST inheritance hierarchy so that a `Modifier` is a `SyntaxNode` rather than a `SyntaxNodeBase`. I have *no* idea why we have both of those, and we need to clean that up soon.
|
|
The main practical change here is that things that used to be `IRValue`s, like literals, are now being expressed as instructions in the global scope.
In order to validate that things are actually being handled correctly, this change introduces an explicit "validation" pass that can be run on the IR to check for different invariants (although it doesn't check many of the important ones right now). I've left the validation pass turned off by default, but with a command-line flag to enable it. We may want to make it be on by default in debug builds, just to keep us honest. The main invariant for the moment is that when on IR instruction is used as an operand to another, it had better come from the same IR module.
Some of the existing passes were violating this rule, in particular when it came to cloning of witness tables related to global generic parameter substitution. Those features can in theory be handled better now by allowing `specialize` instructions at other scopes, but I didn't want to over-complicate this change, so I make just enough fixes to ensure that these steps always clone witness tables they get from the "symbols" on an IR specialization context. In order for this to work when recursively specializing, I had to ensure that the logic for generic specialization had a notion of a "parent" specialization context that it would fall back to to perform cloning when necessary.
This change keeps the logic that was caching and re-using the instructions for literal values within a module, but adds some logic that isn't really being tested right now for picking the right parent instruction to insert a constant instruction into. This logic doesn't trigger right now because all of the cases we are using it on have zero operands (and so they always get "hoisted" to the global scope), but eventually for things like types we want to be able to support instructions with operands (e.g., `vector<float, 4>`) and handle the case where some of those operands come from different scopes (e.g., when nested inside a generic).
The final change here is mostly cosmetic: the `IRBuilder` is now more abstract about where insertion occurs: it tracks a single `IRParentInst` to insert into, and then an optional `IRInst` to insert before. In the common case, that parent is an `IRBlock`, but it could conceivably also be the global scope, or a witness table, etc. Use sites where we used to change those fields directly now use distinct methods `setInsertInto(parent)` and `setInsertBefore(inst)` which capture the two cases we care about. Accessors are also defined to extract the current block (if the current parent is a block), and the current "function" (global value with code, if the current parent is a global value with code, or a block inside one).
With this work in place, it should be possible for a follow-on change to start putting `specialize` instructions at the global scope and thus clean up some of the on-the-fly specialization work. This work should also help with some of the requirements around a distinct IR-level type system and more explicit generics.
|
|
* IR: "everything is an instruction"
This change tries to streamline the representation of the IR in the following ways:
* Every IR value is an instruction (there is no `IRValue` type any more)
* All IR values that can contain other values share a single base (`IRParentInstruction`)
* Dynamic casts to specific IR instruction types can be accomplished with a new `as<Type>(inst)` operation, that uses the IR opcode to implement casts.
The biggest change in terms of number of lines is getting rid of `IRValue`. The diff here could probably be smaller if I'd just done `typedef IRInst IRValue;`.
Along the way I also renamed the `getArg`/`getArgs`/`getArgCount` combination over to `getOperand`/`getOperands`/`getOperandCount` to avoid being confusing when we have something like a `call` instruction where the "arguments" of the call don't line up with the operands of the instruction.
I also tried to clean up the representation of lists of child instructions to try to make it easier to iterate over them with C++ range-based `for` loops. Developers still need to be careful about mutating the contents of a block while iterating over it in this fashion (e.g. if you remove the "current" element, the iteration will end prematurely).
Probably the thorniest change here is that parameters are now just represented as the first N instructions in a block, which means:
* We need to perform a linear search to find the end of the parameter list. This is probably not often a problem, because usually you would be iterating over the parameters anyway, and that will be linear in the number of parameters.
* Algorithms that iterate over a block either need to ignore parameters, treat parameters just like other instructions, or somehow cleave the list into the range of parameters, and the range of "ordinary" instructions (which involves the same linear search above).
* When inserting into a block, we need to be careful not to insert instructions at invalid locations (e.g., insert a temporary before the parameters, or insert a parameter in the middle of the code). I can't pretend that I've handled the details of that here. (This is no different than having to make the same adjustments for phi nodes in a typical SSA representation)
* One possible future-proof approach is to implement a pass that sorts the instructions in a block so that parameters always come first. That would let us implement passes without caring about this detail, and then clean up right before any pass that cares about the relative order of parameters and other instructions.
The current change is missing any work to make literals and other instructions that used to be `IRValue`s properly nest inside of their parent module. Right now these instructions are just left unparented, and may actually end up being shared between distinct modules. Fixing that will need a follow-up change. The biggest challenge there is that it introduces instructions at the global scope that aren't `IRGlobalValue`s.
This change doesn't try to take advantage of any of the new flexibility (e.g., by nesting `specialize` instructions inside of witness tables). The goal is to do exactly what we were doing before, just with a different representation.
* Warning fix
|
|
This allows us to get rid of `IRGlobalValue::dispose()`.
|
|
* Initial work on validating "constexpr"-ness in IR
The underlying issue here is that certain operations in the target shading languages constrain their operands to be compile-time constants. A notable example is the optional texel offset parameter to the `Texture2D.Sample` operation.
When calling these operations in GLSL, the user is required to pass a "constant expression," and any variables in that expression must therefore be marked with the `const` qualifier (and themselves be initialized with constant expressions). Any GLSL output we generate must of course respect these rules.
When calling these operations in HLSL, the user is not so constrained. Instead, they can pass an arbitrary expression, which may involve ordinary variables with no particular markup, and then the compiler is responsible for determining if the actual value after simplification works out to be a constant. In some cases, the requirement that a value be constant might actually trigger things like loop unrolling. Also, it is okay to use a function parameter to determine such a constant expression, as long as the argument turns out to be a constant at all call sites.
The way we have decided to tackle these challenges in Slang is that we we propagate a notion of `constexpr`-ness through the IR. This is currently being tackled in `ir-constexpr.cpp` with a combination of forward and backward iterative dataflow:
* When the operands to an instruction are all `constexpr`, and the opcode is one we believe can be constant-folded, then we infer that the instruction *can* be evaluated as `constexpr`
* When instruction is required to be `constexpr`, then we infer that all of its operands are also required to be `constexpr`.
If this process ever infers that a function parameter is required to be `constexpr`, then we might have to continue propagation at all the call sites to that function.
If after all the propagation is done, there are any cases where an instruction is *required* to be `constexpr`, but it *can't* be `constexpr` (we weren't able to infer `constexpr`-ness for its operands), then we issue an error.
This implementation encodes the idea of `constexpr`-ness in the IR as part of the type system, using a simplified notion of rates. This change adds a `RateQualifiedType` that can represent `@R T`, and then introduces a `ConstExprRate` that can be used for `R`. Many accessors for the type information on IR nodes were updated to distinguish when one wants the "full" type of an IR value (which might include rate information) vs. just the "data" type.
A `constexpr` qualifier was added in the front-end, and is being used to decorate the texel offset parameter for `Texture2D.Sample`. Lowering from AST to IR looks for this qalifier and infers when a function parameter must be typed as `@ConstExpr T` instead of just `T`.
There are lots of limitations and gotchas in the implementation so far:
* The `@ConstExpr` rate is the only one added in this change, but it seems clear that the conceptual `ThreadGroup` rate that was added to represent `groupshared` should probably get folded into the representation.
* I'm not 100% pleased with how many places in the IR I have to special-case for rate-qualified types. At the same type, pulling out rate as a distinct field on `IRValue` would probably require that we pay attention to rate everywhere.
* I've added a test case to show that we can issue errors when users fail to provide a constant expression for the texel offset, but the actual error message isn't great because it doesn't indicate *why* a constant expression was required. Realistically the "initial IR" should contain a few more decorations we can use to relate error conditions back to the original code (even if this is in a side-band structure).
* I've added a test case that is supposed to show that we can back-propagate `constexpr`-ness to local variables, and I've manually confirmed that it works for Vulkan/SPIR-V output, but the level of Vulkan support in `render_test` today means I can't enable the test for check-in.
* While I'm attempting to propagate `@ConstExpr` information from callees to callers, I haven't implemented any logic to specialize callee functions based on values at call sites.
* In a similar vein, there is no handling of control-flow dependence in the current code. If we infer that a phi (block parameter) needs to be `@ConstExpr`, then it isn't actually enough to require that the inputs to the phi (arguments from predecessor blocks) are all `@ConstExpr` because we also need any control-flow decisions that pick which incoming edge we take to be `@ConstExpr` as well.
* As a practical matter, implicit propagation of `@ConstExpr` from a function body to a function parameter should only be allowed for functions that are "local" to a module. Any function that might be accessed from outside of a module should really have had its `@ConstExpr` parameter marked manually, and our pass should validate that they follow their own rules. Right now we have no kind of visibility (`public` vs `private`) system, so I'm kind of ignoring this issue.
While that is a lot of gaps, this is also just enough code to get the Falcor MultiPassPostProcess example working, so I'm inclined to get it checked in.
* Fixup: missing expected output for test
* Fixup: disable test that relies on [unroll] for now
|
|
getSpecailizedMangledName to work properly.
|
|
* Basic IR support for `static const` globals
Our strategy for lowering global *variables* can fall back to putting their initialization into a function, but that isn't really appropriate for global constants (it also isn't appropriate for arrays, but we'll need to deal with that seaprately).
This change adds a distinct case for global constants (rather than treating them as variables), and forces the emission logic to always emit them as a single expression.
Doing this makes assumptions about how the IR for these constants gets emitted (and what optimziations might do to it).
In order to make things work, I had to switch the handling of initializer-list expressions to not be lowered via temporaries and mutation (since that isn't a good fit for reverting to a single expression).
I've added a single test case to ensure that this works in the simplest scenario. My next priority will be to see if this unblocks my work in Falcor.
* Fixup: bug fixes
|
|
* Generate SSA form for IR functions
The basic idea here is simple: in the front-end after we have lowered the AST to initial IR we will apply a set of "mandatory" optimization passes. The first of these is to attempt to translate the all functions into SSA form so that they are amenable to subsequent dataflow optimizations. Eventually, the mandatory optimization passes would include diagnostic passes that make sure variables aren't used when undefined, etc.
Just doing basic SSA generation already cleans up a lot of the messiness in our IR today, because constructs that used to involve many local variables can now be handled via SSA temporaries.
The implementation of SSA generation is in `ir-ssa.cpp`, and it follows the approach of Braun et al.'s "Simple and Efficient Construction of Static Single Assignment Form." I used this instead of the more well-known Cytron et al. algorithm because Braun's algorith mis very simple to code, and does not require auxiliary analyses to generate the dominance frontier.
The main wrinkle in our SSA representation right now is that instead of using ordinary phi nodes, we instead allow basic blocks to have parameters, where predecessor blocks pass in different parameter values. This encodes information equivalent to traditional phi nodes, but has two (small) benefits:
1. There is no fixed relationship between the order of phi operands and predecessor blocks, so we don't have to worry about breaking the phis when we alter the order in which predecessors are stored. This is important for us because predecessors are being stored implicitly.
2. It is easy to operationalize a "branch with arguments" either when lowering to other languages, or when interpreting the IR. A branch with arguments is implemented as a sequence of stores from the arguments to the parameters of the target block (very similar to a call), followed by a jump to the block.
Relevant to the above, this change also adds an interface for enumerating the predecessors or successors of a block in our CFG. Rather than use an auxliary structure, we directly use the information already encoded in the IR:
* The sucessors of a block are the target label operands of its terminator instruction. In our IR this is a contiguous range of `IRUse`s, possible with a stride (to account for the way `switch` interleaves values and blocks).
* The predecessors of a block are a subset of the uses of the block's value. Specifically, they are any uses that are on a terminator instruction, and within the range of values that represent the successor list of that instruction.
One important limitation of the "blocks with arguments" model for handling phis is that it is really only convenient to stash extra arguments on an unconditional terminator instruction. This change works around this prob lem by breaking any "critical edges" - edges between a block with multiple successors and one with multiple predecessors. We assume that "phi" nodes will only ever be needed on a block with multiple predecessors, and because critical edges are broken, each of these predecessors will then have only a single successor, so its branch instruction can handle the extra arguments.
This change introduces a notion of an "undefined" instruction in the IR. This is handled as an instruction rather than a value because I anticipate that we will want to distinguish different undefined values when it comes time to start issuing error messages (those messages will need to point to the variable that was used when undefined).
* Fix expected test output.
Another change was merged that enabled the `glsl-parameter-blocks` test, and its output is affected by our IR optimization work.
|
|
The standard library already has a bunch of these decorations, since they were added to support Slang->Vulkan codegen on the AST-to-AST path. This change makes the IR code generator able to exploit the modifiers so that we pick up a bunch of Vulkan support "for free" in the short term.
The basic change is in `lower-to-ir.cpp` where we copy over any `TargetIntrinsicModifier`s to become `IRTargetIntrinsicDecoration`s with the same information. We then need a bit of logic in `ir.cpp` to make sure we clone them as needed.
The core work of using the modifiers is in `emit.cpp`, where I basically just copy-pasted the existing logic that applied in the AST path (all the AST-related code there is dead, and we should clean it up soon).
The big change that comes with this logic is that when dealing with a member function, the numbering of the argument used in the intrinsic definition string changes, so that `$0` refers to the base object (whereas before the base object was looked up via the base expression of a `MemberExpr` used for the function). This requires a bunch of the definitions in the library to be updated; hopefully I caught them all.
For kicks, I've re-enabled a cross-compilation test just to confirm that we are generating valid SPIR-V for code that performs texture-fetch operations. I don't expect us to keep that test enabled as-is in the long term, though, because it would be much better to instead use render-test to do the same thing. Alas, beefing up the Vulkan support in render-test is an outstanding work item, and I didn't want to pollute this change with more work along those lines.
|
