[flang-commits] [flang] 2f0b4f4 - [flang][extension] support concatenation with absent optional (#112678)

via flang-commits flang-commits at lists.llvm.org
Thu Oct 17 04:25:14 PDT 2024


Author: jeanPerier
Date: 2024-10-17T13:25:09+02:00
New Revision: 2f0b4f43fc5c1e7587c4d00daa9cc230df2f8a2d

URL: https://github.com/llvm/llvm-project/commit/2f0b4f43fc5c1e7587c4d00daa9cc230df2f8a2d
DIFF: https://github.com/llvm/llvm-project/commit/2f0b4f43fc5c1e7587c4d00daa9cc230df2f8a2d.diff

LOG: [flang][extension] support concatenation with absent optional (#112678)

Fix #112593 by adding support in lowering to concatenation with an
absent optional _assumed length_ dummy argument because:
1. Most compilers seem to support it (most likely by accident).
2. This actually makes the compiler codegen simpler. Codegen was going
out of its way to poke the LLVM optimizer bear by producing an undef
argument for the length.

I insist on the fact that no compiler support this with _explicit
length_ optional arguments and the executable will segfault and I would
discourage users from using that "feature" because runtime checks for
bad optional dereference will kick when used (For instance, "nagfor
-C=present" will produce an executable that abort with an error message
. Flang does not have such runtime check option so far).

Hence, I am not updating the Extensions.md document because this is not
something I think we should advertise.

Added: 
    

Modified: 
    flang/lib/Optimizer/CodeGen/CodeGen.cpp
    flang/test/Fir/optional.fir

Removed: 
    


################################################################################
diff  --git a/flang/lib/Optimizer/CodeGen/CodeGen.cpp b/flang/lib/Optimizer/CodeGen/CodeGen.cpp
index 9b624efa053813..68b8c6613585e6 100644
--- a/flang/lib/Optimizer/CodeGen/CodeGen.cpp
+++ b/flang/lib/Optimizer/CodeGen/CodeGen.cpp
@@ -3373,19 +3373,7 @@ struct AbsentOpConversion : public fir::FIROpConversion<fir::AbsentOp> {
   matchAndRewrite(fir::AbsentOp absent, OpAdaptor,
                   mlir::ConversionPatternRewriter &rewriter) const override {
     mlir::Type ty = convertType(absent.getType());
-    mlir::Location loc = absent.getLoc();
-
-    if (mlir::isa<fir::BoxCharType>(absent.getType())) {
-      auto structTy = mlir::cast<mlir::LLVM::LLVMStructType>(ty);
-      assert(!structTy.isOpaque() && !structTy.getBody().empty());
-      auto undefStruct = rewriter.create<mlir::LLVM::UndefOp>(loc, ty);
-      auto nullField =
-          rewriter.create<mlir::LLVM::ZeroOp>(loc, structTy.getBody()[0]);
-      rewriter.replaceOpWithNewOp<mlir::LLVM::InsertValueOp>(
-          absent, undefStruct, nullField, 0);
-    } else {
-      rewriter.replaceOpWithNewOp<mlir::LLVM::ZeroOp>(absent, ty);
-    }
+    rewriter.replaceOpWithNewOp<mlir::LLVM::ZeroOp>(absent, ty);
     return mlir::success();
   }
 };

diff  --git a/flang/test/Fir/optional.fir b/flang/test/Fir/optional.fir
index 3b350d6fa94195..bded8b5332a300 100644
--- a/flang/test/Fir/optional.fir
+++ b/flang/test/Fir/optional.fir
@@ -47,7 +47,7 @@ func.func @foo3(%arg0: !fir.boxchar<1>) -> i1 {
 // CHECK-LABEL: @bar3
 func.func @bar3() -> i1 {
   %0 = fir.absent !fir.boxchar<1>
-  // CHECK: call i1 @foo3(ptr null, i64 undef)
+  // CHECK: call i1 @foo3(ptr null, i64 0)
   %1 = fir.call @foo3(%0) : (!fir.boxchar<1>) -> i1
   return %1 : i1
 }


        


More information about the flang-commits mailing list