================
@@ -230,3 +230,55 @@
! OFFLOAD-TARGETS-NOT: --offload-targets
! OFFLOAD-TARGETS: "{{[^"]*}}flang{{[^"]*}}" "-fc1" "-triple"
"x86_64-unknown-linux-gnu"
! OFFLOAD-TARGETS-SAME: "--offload-targets=amdgpu-amd-amdhsa"
+
+! Verify the implicit default mapper allocatable members flags are applied to
both host
+! and device phases of the offload compilation when the negative flag is used
and ignored
+! in the positive case as it is the default.
+
+! RUN: %flang -### %s 2>&1 \
+! RUN: -fopenmp -fimplicit-default-mapper-allocatable-members
--offload-arch=gfx90a \
+! RUN: --target=x86_64-unknown-linux-gnu -nogpulib \
+! RUN: | FileCheck %s --check-prefix=ENABLE-IMP-MAPPER
+
+! ENABLE-IMP-MAPPER: "-fc1" "-triple" "x86_64-unknown-linux-gnu"
+! ENABLE-IMP-MAPPER-SAME: "-fopenmp"
+! ENABLE-IMP-MAPPER-SAME: "--offload-targets=amdgpu-amd-amdhsa"
+! ENABLE-IMP-MAPPER-NOT: implicit-default-mapper-allocatable-members
----------------
agozillon wrote:
Thank you, you're correct! Sorry, too used to writing separate cases for
everything.
https://github.com/llvm/llvm-project/pull/228748
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
