tra added a comment.
Presumably, `__managed__` variables would have to be memory-mapped into the
host address space.
================
Comment at: clang/lib/CodeGen/CGCUDANV.cpp:391
+ SmallVector<SmallVector<llvm::User *, 8>, 8> WorkList;
+ for (auto &&UU : Var->uses()) {
+ WorkList.push_back({UU.getUser()});
----------------
`VarUse` ?
================
Comment at: clang/lib/CodeGen/CGCUDANV.cpp:395
+ while (!WorkList.empty()) {
+ auto &&US = WorkList.pop_back_val();
+ auto *U = US.back();
----------------
`WorkItem`
================
Comment at: clang/lib/CodeGen/CGCUDANV.cpp:552-553
+ llvm::ConstantInt::get(IntTy, Info.Flags.isConstant()),
+ llvm::ConstantInt::get(IntTy, CGM.getLangOpts().HIP &&
+ Info.Flags.isManaged())};
+ Builder.CreateCall(RegisterVar, Args);
----------------
This will always be `false`, because all cases where `Info.Flags.isManaged()`
is true are handled in the other branch of this `if`.
================
Comment at: clang/lib/Sema/SemaDeclAttr.cpp:7728
+ case ParsedAttr::AT_HIPManaged:
+ handleSimpleAttributeWithExclusions<HIPManagedAttr, CUDAGlobalAttr>(S, D,
+ AL);
----------------
The code changes in the patch appear to treat `__managed__` variable as a
superset of a `__device__` var. If that's indeed the case, adding an implicit
`__device__` attribute here would help to simplify the code. This way the
existing code can handle generic `__device__` var functionality without
additional changes, and would use `__managed__` checks for the cases specific
for managed vars only.
================
Comment at: llvm/lib/IR/ReplaceConstant.cpp:20
+namespace llvm {
+Instruction *createReplacementInstr(ConstantExpr *CE, Instruction *Instr) {
+ IRBuilder<NoFolder> Builder(Instr);
----------------
The function could use a descriptive comment.
Despite the comment above, it does seem to do any replacing itself. Presumably
it creates a an instruction that would perform an equivalent calculation at
runtime.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D94814/new/
https://reviews.llvm.org/D94814
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
