Index: llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp =================================================================== --- llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp +++ llvm/trunk/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp @@ -620,6 +620,7 @@ 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(OutStreamer->getTargetStreamer()); Index: llvm/trunk/test/CodeGen/AMDGPU/private-memory.ll =================================================================== --- llvm/trunk/test/CodeGen/AMDGPU/private-memory.ll +++ llvm/trunk/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 @@ ; 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