AMDGPU: Ignore CodeObjectV3 when inlining
This was inhibiting inlining of library functions when clang was
invoking the inliner directly. This is covering a bit of a mess with
subtarget feature handling, and this shouldn't be a subtarget
feature. The behavior is different depending on whether you are using
a -mattr flag in clang, or llc, opt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@353899 91177308-0d34-0410-b5e6-96231b3b80d8
diff --git a/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h b/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
index 5965072..e584006 100644
--- a/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
+++ b/lib/Target/AMDGPU/AMDGPUTargetTransformInfo.h
@@ -84,6 +84,7 @@
AMDGPU::FeatureSGPRInitBug,
AMDGPU::FeatureXNACK,
AMDGPU::FeatureTrapHandler,
+ AMDGPU::FeatureCodeObjectV3,
// Perf-tuning features
AMDGPU::FeatureFastFMAF32,
diff --git a/test/Transforms/Inline/AMDGPU/inline-target-cpu.ll b/test/Transforms/Inline/AMDGPU/inline-target-cpu.ll
index 4258fc5..87330c7 100644
--- a/test/Transforms/Inline/AMDGPU/inline-target-cpu.ll
+++ b/test/Transforms/Inline/AMDGPU/inline-target-cpu.ll
@@ -78,6 +78,17 @@
ret i32 %call
}
+define i32 @enable_codeobjectv3() #10 {
+ ret i32 999
+}
+
+; CHECK-LABEL: @disable_codeobjectv3_call_codeobjectv3(
+; CHECK-NEXT: ret i32 999
+define i32 @disable_codeobjectv3_call_codeobjectv3() #11 {
+ %call = call i32 @enable_codeobjectv3()
+ ret i32 %call
+}
+
attributes #0 = { nounwind }
attributes #1 = { nounwind "target-cpu"="fiji" }
attributes #2 = { nounwind "target-cpu"="fiji" "target-features"="+fp32-denormals" }
@@ -88,3 +99,5 @@
attributes #7 = { nounwind "target-features"="+half-rate-64-ops" }
attributes #8 = { nounwind "target-features"="-load-store-opt" }
attributes #9 = { nounwind "target-features"="+load-store-opt" }
+attributes #10 = { nounwind "target-features"="+code-object-v3" }
+attributes #11 = { nounwind "target-features"="-code-object-v3" }