[clang] [clang][CodeGen][SPIR-V][AMDGPU] Tweak AMDGCNSPIRV ABI to allow for the correct handling of aggregates passed to kernels / functions. (PR #102776)

[Alex Voicu via cfe-commits]

================
@@ -64,6 +66,27 @@ void CommonSPIRABIInfo::setCCs() {
   RuntimeCC = llvm::CallingConv::SPIR_FUNC;
 }
 
+ABIArgInfo SPIRVABIInfo::classifyReturnType(QualType RetTy) const {
+  if (getTarget().getTriple().getVendor() != llvm::Triple::AMD)
+    return DefaultABIInfo::classifyReturnType(RetTy);
+  if (!isAggregateTypeForABI(RetTy) || getRecordArgABI(RetTy, getCXXABI()))
+    return DefaultABIInfo::classifyReturnType(RetTy);
+
+  if (const RecordType *RT = RetTy->getAs<RecordType>()) {
+    const RecordDecl *RD = RT->getDecl();
+    if (RD->hasFlexibleArrayMember())
+      return DefaultABIInfo::classifyReturnType(RetTy);
+  }
+
+  // TODO: The AMDGPU ABI is non-trivial to represent in SPIR-V; in order to
+  // avoid encoding various architecture specific bits here we return everything
+  // as direct to retain type info for things like aggregates, for later perusal
+  // when translating back to LLVM/lowering in the BE. This is also why we
+  // disable flattening as the outcomes can mismatch between SPIR-V and AMDGPU.
+  // This will be revisited / optimised in the future.
----------------
AlexVlx wrote:

> byval doesn't really make sense on a kernel as there is no real caller, but depending on how SPIRV defined its byval attribute, maybe you can codegen LLVM byref into SPIRV byval?

This is a good suggestion, it's one of the things I'm looking at, but it's a bit messy to retain the "this is actually `byref`" (it is doable, but was quite intrusive with tools like the translator), and I admit to not fully debugging / validating it. I just realised that the "add an extension to SPIR-V" bit came off pretty passive-aggressive, I apologise - I think that might just be TheRightWayTM to deal with this, `byref` is common enough to justify it.

https://github.com/llvm/llvm-project/pull/102776