[PATCH] D99201: [HIP] Diagnose unaligned atomic for amdgpu

[Yaxun Liu via Phabricator via cfe-commits]

yaxunl updated this revision to Diff 528937.
yaxunl added a comment.
Herald added a subscriber: MaskRay.
Herald added a project: All.

rebase and simplify code


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D99201/new/

https://reviews.llvm.org/D99201

Files:
  clang/lib/Driver/ToolChains/Clang.cpp
  clang/test/Driver/hip-options.hip


Index: clang/test/Driver/hip-options.hip
===================================================================
--- clang/test/Driver/hip-options.hip
+++ clang/test/Driver/hip-options.hip
@@ -138,3 +138,10 @@
 // RUN: %clang -### -nogpuinc -nogpulib -mamdgpu-ieee -mno-amdgpu-ieee 
-ffast-math \
 // RUN:   --cuda-gpu-arch=gfx906  %s 2>&1 | FileCheck 
-check-prefixes=IEEE-OFF-NEG %s
 // IEEE-OFF-NEG-NOT: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} 
"-mamdgpu-ieee"
+
+// Check -Werror=atomic-alignment is passed for amdpu by default.
+
+// RUN: %clang -### -target x86_64-unknown-linux-gnu -nogpuinc -nogpulib \
+// RUN:   --cuda-gpu-arch=gfx906 %s 2>&1 | FileCheck -check-prefix=WARN-ATOMIC 
%s
+// WARN-ATOMIC: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} 
"-Werror=atomic-alignment"
+// WARN-ATOMIC-NOT: clang{{.*}} "-triple" "x86_64-unknown-linux-gnu" {{.*}} 
"-Werror=atomic-alignment"
Index: clang/lib/Driver/ToolChains/Clang.cpp
===================================================================
--- clang/lib/Driver/ToolChains/Clang.cpp
+++ clang/lib/Driver/ToolChains/Clang.cpp
@@ -7209,6 +7209,10 @@
                       options::OPT_mno_unsafe_fp_atomics);
     Args.addOptOutFlag(CmdArgs, options::OPT_mamdgpu_ieee,
                        options::OPT_mno_amdgpu_ieee);
+
+    // AMDGPU does not support atomic lib call. Treat atomic alignment
+    // warnings as errors.
+    CmdArgs.push_back("-Werror=atomic-alignment");
   }
 
   // For all the host OpenMP offloading compile jobs we need to pass the 
targets



Index: clang/test/Driver/hip-options.hip
===================================================================
--- clang/test/Driver/hip-options.hip
+++ clang/test/Driver/hip-options.hip
@@ -138,3 +138,10 @@
 // RUN: %clang -### -nogpuinc -nogpulib -mamdgpu-ieee -mno-amdgpu-ieee -ffast-math \
 // RUN:   --cuda-gpu-arch=gfx906  %s 2>&1 | FileCheck -check-prefixes=IEEE-OFF-NEG %s
 // IEEE-OFF-NEG-NOT: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} "-mamdgpu-ieee"
+
+// Check -Werror=atomic-alignment is passed for amdpu by default.
+
+// RUN: %clang -### -target x86_64-unknown-linux-gnu -nogpuinc -nogpulib \
+// RUN:   --cuda-gpu-arch=gfx906 %s 2>&1 | FileCheck -check-prefix=WARN-ATOMIC %s
+// WARN-ATOMIC: clang{{.*}} "-triple" "amdgcn-amd-amdhsa" {{.*}} "-Werror=atomic-alignment"
+// WARN-ATOMIC-NOT: clang{{.*}} "-triple" "x86_64-unknown-linux-gnu" {{.*}} "-Werror=atomic-alignment"
Index: clang/lib/Driver/ToolChains/Clang.cpp
===================================================================
--- clang/lib/Driver/ToolChains/Clang.cpp
+++ clang/lib/Driver/ToolChains/Clang.cpp
@@ -7209,6 +7209,10 @@
                       options::OPT_mno_unsafe_fp_atomics);
     Args.addOptOutFlag(CmdArgs, options::OPT_mamdgpu_ieee,
                        options::OPT_mno_amdgpu_ieee);
+
+    // AMDGPU does not support atomic lib call. Treat atomic alignment
+    // warnings as errors.
+    CmdArgs.push_back("-Werror=atomic-alignment");
   }
 
   // For all the host OpenMP offloading compile jobs we need to pass the targets

_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits