diff options
Diffstat (limited to 'llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp')
| -rw-r--r-- | llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp | 38 |
1 files changed, 28 insertions, 10 deletions
diff --git a/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp b/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp index 7aa5f1abf65b..078133469549 100644 --- a/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp +++ b/llvm/lib/Target/AMDGPU/MCTargetDesc/AMDGPUTargetStreamer.cpp @@ -17,12 +17,16 @@ #include "Utils/AMDKernelCodeTUtils.h" #include "llvm/BinaryFormat/AMDGPUMetadataVerifier.h" #include "llvm/BinaryFormat/ELF.h" +#include "llvm/MC/MCAssembler.h" #include "llvm/MC/MCContext.h" #include "llvm/MC/MCELFStreamer.h" #include "llvm/MC/MCSectionELF.h" +#include "llvm/MC/MCSubtargetInfo.h" #include "llvm/Support/AMDGPUMetadata.h" #include "llvm/Support/AMDHSAKernelDescriptor.h" +#include "llvm/Support/Casting.h" #include "llvm/Support/FormattedStream.h" +#include "llvm/Support/TargetParser.h" using namespace llvm; using namespace llvm::AMDGPU; @@ -102,6 +106,7 @@ StringRef AMDGPUTargetStreamer::getArchNameFromElfMach(unsigned ElfMach) { case ELF::EF_AMDGPU_MACH_AMDGCN_GFX909: AK = GK_GFX909; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX90A: AK = GK_GFX90A; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX90C: AK = GK_GFX90C; break; + case ELF::EF_AMDGPU_MACH_AMDGCN_GFX940: AK = GK_GFX940; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1010: AK = GK_GFX1010; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1011: AK = GK_GFX1011; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1012: AK = GK_GFX1012; break; @@ -112,6 +117,11 @@ StringRef AMDGPUTargetStreamer::getArchNameFromElfMach(unsigned ElfMach) { case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1033: AK = GK_GFX1033; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1034: AK = GK_GFX1034; break; case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1035: AK = GK_GFX1035; break; + case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1036: AK = GK_GFX1036; break; + case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1100: AK = GK_GFX1100; break; + case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1101: AK = GK_GFX1101; break; + case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1102: AK = GK_GFX1102; break; + case ELF::EF_AMDGPU_MACH_AMDGCN_GFX1103: AK = GK_GFX1103; break; case ELF::EF_AMDGPU_MACH_NONE: AK = GK_NONE; break; } @@ -165,6 +175,7 @@ unsigned AMDGPUTargetStreamer::getElfMach(StringRef GPU) { case GK_GFX909: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX909; case GK_GFX90A: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX90A; case GK_GFX90C: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX90C; + case GK_GFX940: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX940; case GK_GFX1010: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1010; case GK_GFX1011: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1011; case GK_GFX1012: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1012; @@ -175,6 +186,11 @@ unsigned AMDGPUTargetStreamer::getElfMach(StringRef GPU) { case GK_GFX1033: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1033; case GK_GFX1034: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1034; case GK_GFX1035: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1035; + case GK_GFX1036: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1036; + case GK_GFX1100: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1100; + case GK_GFX1101: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1101; + case GK_GFX1102: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1102; + case GK_GFX1103: return ELF::EF_AMDGPU_MACH_AMDGCN_GFX1103; case GK_NONE: return ELF::EF_AMDGPU_MACH_NONE; } @@ -285,7 +301,7 @@ bool AMDGPUTargetAsmStreamer::EmitCodeEnd(const MCSubtargetInfo &STI) { uint32_t Encoded_pad = Encoded_s_code_end; // Instruction cache line size in bytes. - const unsigned Log2CacheLineSize = 6; + const unsigned Log2CacheLineSize = AMDGPU::isGFX11Plus(STI) ? 7 : 6; const unsigned CacheLineSize = 1u << Log2CacheLineSize; // Extra padding amount in bytes to support prefetch mode 3. @@ -439,6 +455,8 @@ void AMDGPUTargetAsmStreamer::EmitAmdhsaKernelDescriptor( PRINT_FIELD(OS, ".amdhsa_forward_progress", KD, compute_pgm_rsrc1, amdhsa::COMPUTE_PGM_RSRC1_FWD_PROGRESS); + PRINT_FIELD(OS, ".amdhsa_shared_vgpr_count", KD, compute_pgm_rsrc3, + amdhsa::COMPUTE_PGM_RSRC3_GFX10_PLUS_SHARED_VGPR_COUNT); } PRINT_FIELD( OS, ".amdhsa_exception_fp_ieee_invalid_op", KD, @@ -515,8 +533,8 @@ void AMDGPUTargetELFStreamer::EmitNote( if (STI.getTargetTriple().getOS() == Triple::AMDHSA) NoteFlags = ELF::SHF_ALLOC; - S.PushSection(); - S.SwitchSection( + S.pushSection(); + S.switchSection( Context.getELFSection(ElfNote::SectionName, ELF::SHT_NOTE, NoteFlags)); S.emitInt32(NameSZ); // namesz S.emitValue(DescSZ, 4); // descz @@ -525,7 +543,7 @@ void AMDGPUTargetELFStreamer::EmitNote( S.emitValueToAlignment(4, 0, 1, 0); // padding 0 EmitDesc(S); // desc S.emitValueToAlignment(4, 0, 1, 0); // padding 0 - S.PopSection(); + S.popSection(); } unsigned AMDGPUTargetELFStreamer::getEFlags() { @@ -691,7 +709,7 @@ AMDGPUTargetELFStreamer::EmitDirectiveHSACodeObjectISAV2(uint32_t Major, OS.emitBytes(VendorName); OS.emitInt8(0); // NULL terminate VendorName OS.emitBytes(ArchName); - OS.emitInt8(0); // NULL terminte ArchName + OS.emitInt8(0); // NULL terminate ArchName }); } @@ -699,9 +717,9 @@ void AMDGPUTargetELFStreamer::EmitAMDKernelCodeT(const amd_kernel_code_t &Header) { MCStreamer &OS = getStreamer(); - OS.PushSection(); + OS.pushSection(); OS.emitBytes(StringRef((const char*)&Header, sizeof(Header))); - OS.PopSection(); + OS.popSection(); } void AMDGPUTargetELFStreamer::EmitAMDGPUSymbolType(StringRef SymbolName, @@ -806,7 +824,7 @@ bool AMDGPUTargetELFStreamer::EmitCodeEnd(const MCSubtargetInfo &STI) { uint32_t Encoded_pad = Encoded_s_code_end; // Instruction cache line size in bytes. - const unsigned Log2CacheLineSize = 6; + const unsigned Log2CacheLineSize = AMDGPU::isGFX11Plus(STI) ? 7 : 6; const unsigned CacheLineSize = 1u << Log2CacheLineSize; // Extra padding amount in bytes to support prefetch mode 3. @@ -818,11 +836,11 @@ bool AMDGPUTargetELFStreamer::EmitCodeEnd(const MCSubtargetInfo &STI) { } MCStreamer &OS = getStreamer(); - OS.PushSection(); + OS.pushSection(); OS.emitValueToAlignment(CacheLineSize, Encoded_pad, 4); for (unsigned I = 0; I < FillSize; I += 4) OS.emitInt32(Encoded_pad); - OS.PopSection(); + OS.popSection(); return true; } |