AMDGPU/SI: Set the code object work group segment size when targeting HSA

Reviewers: arsenm

Subscribers: arsenm, llvm-commits

Differential Revision: http://reviews.llvm.org/D15493

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@255702 91177308-0d34-0410-b5e6-96231b3b80d8

diff --git a/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp b/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
index be0df57e9f0b6a892bab4eca857c0d9f913fbc99..ac8f568da189578f81e2a5f693089a30a393cf4e 100644 (file)
--- a/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
+++ b/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
@@ -620,6 +620,7 @@ void AMDGPUAsmPrinter::EmitAmdKernelCodeT(const MachineFunction &MF,
   header.wavefront_sgpr_count = KernelInfo.NumSGPR;
   header.workitem_vgpr_count = KernelInfo.NumVGPR;
   header.workitem_private_segment_byte_size = KernelInfo.ScratchSize;
+  header.workgroup_group_segment_byte_size = KernelInfo.LDSSize;
 
   AMDGPUTargetStreamer *TS =
       static_cast<AMDGPUTargetStreamer *>(OutStreamer->getTargetStreamer());

diff --git a/test/CodeGen/AMDGPU/private-memory.ll b/test/CodeGen/AMDGPU/private-memory.ll
index 628b8a21b819474d72dd340be38c3f0edfb33724..79778eebd802b7a3d358ed58eb9e3100df19346b 100644 (file)
--- a/test/CodeGen/AMDGPU/private-memory.ll
+++ b/test/CodeGen/AMDGPU/private-memory.ll
@@ -1,5 +1,6 @@
 ; RUN: llc -march=r600 -mcpu=redwood < %s | FileCheck %s -check-prefix=R600 -check-prefix=FUNC
 ; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -march=amdgcn -mcpu=SI < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC
+; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -mtriple=amdgcn--amdhsa -mcpu=kaveri < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC -check-prefix=HSA-PROMOTE
 ; RUN: llc -show-mc-encoding -mattr=-promote-alloca -verify-machineinstrs -march=amdgcn -mcpu=SI < %s | FileCheck %s -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC
 ; RUN: llc -show-mc-encoding -mattr=-promote-alloca -verify-machineinstrs -mtriple=amdgcn-amdhsa -mcpu=kaveri < %s | FileCheck %s -check-prefix=SI-ALLOCA -check-prefix=SI -check-prefix=FUNC -check-prefix=HSA-ALLOCA
 ; RUN: llc -show-mc-encoding -mattr=+promote-alloca -verify-machineinstrs -march=amdgcn -mcpu=tonga < %s | FileCheck %s -check-prefix=SI-PROMOTE -check-prefix=SI -check-prefix=FUNC
@@ -14,6 +15,10 @@ declare i32 @llvm.r600.read.tidig.x() nounwind readnone
 ; R600: LDS_READ
 ; R600: LDS_READ
 
+; HSA-PROMOTE: .amd_kernel_code_t
+; HSA-PROMOTE: workgroup_group_segment_byte_size = 5120
+; HSA-PROMOTE: .end_amd_kernel_code_t
+
 ; SI-PROMOTE: ds_write_b32
 ; SI-PROMOTE: ds_write_b32
 ; SI-PROMOTE: ds_read_b32