llvm · Pierre-vh · Sep 21, 2023 · Mar 14, 2023
diff --git a/clang/include/clang/Basic/TargetOptions.h b/clang/include/clang/Basic/TargetOptions.h
@@ -82,7 +82,7 @@ class TargetOptions {
   /// code object version times 100.
   enum CodeObjectVersionKind {
     COV_None,
-    COV_2 = 200,
+    COV_2 = 200, // Unsupported.
     COV_3 = 300,
     COV_4 = 400,
     COV_5 = 500,

diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
@@ -4624,9 +4624,9 @@ defm amdgpu_ieee : BoolOption<"m", "amdgpu-ieee",
 def mcode_object_version_EQ : Joined<["-"], "mcode-object-version=">, Group<m_Group>,
   HelpText<"Specify code object ABI version. Defaults to 4. (AMDGPU only)">,
   Visibility<[ClangOption, CC1Option]>,
-  Values<"none,2,3,4,5">,
+  Values<"none,3,4,5">,
   NormalizedValuesScope<"TargetOptions">,
-  NormalizedValues<["COV_None", "COV_2", "COV_3", "COV_4", "COV_5"]>,
+  NormalizedValues<["COV_None", "COV_3", "COV_4", "COV_5"]>,
   MarshallingInfoEnum<TargetOpts<"CodeObjectVersion">, "COV_4">;
 
 defm cumode : SimpleMFlag<"cumode",

diff --git a/clang/lib/Driver/ToolChains/CommonArgs.cpp b/clang/lib/Driver/ToolChains/CommonArgs.cpp
@@ -2323,7 +2323,7 @@ getAMDGPUCodeObjectArgument(const Driver &D, const llvm::opt::ArgList &Args) {
 
 void tools::checkAMDGPUCodeObjectVersion(const Driver &D,
                                          const llvm::opt::ArgList &Args) {
-  const unsigned MinCodeObjVer = 2;
+  const unsigned MinCodeObjVer = 3;
   const unsigned MaxCodeObjVer = 5;
 
   if (auto *CodeObjArg = getAMDGPUCodeObjectArgument(D, Args)) {

diff --git a/clang/test/CodeGenCUDA/amdgpu-code-object-version.cu b/clang/test/CodeGenCUDA/amdgpu-code-object-version.cu
@@ -3,9 +3,6 @@
 // RUN: %clang_cc1 -fcuda-is-device -triple amdgcn-amd-amdhsa -emit-llvm \
 // RUN:   -o - %s | FileCheck %s -check-prefix=V4
 
-// RUN: %clang_cc1 -fcuda-is-device -triple amdgcn-amd-amdhsa -emit-llvm \
-// RUN:   -mcode-object-version=2 -o - %s | FileCheck -check-prefix=V2 %s
-
 // RUN: %clang_cc1 -fcuda-is-device -triple amdgcn-amd-amdhsa -emit-llvm \
 // RUN:   -mcode-object-version=3 -o - %s | FileCheck -check-prefix=V3 %s
 
@@ -21,7 +18,6 @@
 // RUN: not %clang_cc1 -fcuda-is-device -triple amdgcn-amd-amdhsa -emit-llvm \
 // RUN:   -mcode-object-version=4.1 -o - %s 2>&1| FileCheck %s -check-prefix=INV
 
-// V2: !{{.*}} = !{i32 1, !"amdgpu_code_object_version", i32 200}
 // V3: !{{.*}} = !{i32 1, !"amdgpu_code_object_version", i32 300}
 // V4: !{{.*}} = !{i32 1, !"amdgpu_code_object_version", i32 400}
 // V5: !{{.*}} = !{i32 1, !"amdgpu_code_object_version", i32 500}

diff --git a/clang/test/Driver/hip-code-object-version.hip b/clang/test/Driver/hip-code-object-version.hip
@@ -1,15 +1,5 @@
 // REQUIRES: amdgpu-registered-target
 
-// Check bundle ID for code object v2.
-
-// RUN: not %clang -### --target=x86_64-linux-gnu \
-// RUN:   -mcode-object-version=2 \
-// RUN:   --offload-arch=gfx906 --rocm-path=%S/Inputs/rocm \
-// RUN:   %s 2>&1 | FileCheck -check-prefix=V2 %s
-
-// V2: "-mllvm" "--amdhsa-code-object-version=2"
-// V2: "-targets=host-x86_64-unknown-linux,hip-amdgcn-amd-amdhsa--gfx906"
-
 // Check bundle ID for code object v3.
 
 // RUN: not %clang -### --target=x86_64-linux-gnu \
@@ -61,9 +51,16 @@
 // RUN: not %clang -### --target=x86_64-linux-gnu \
 // RUN:   -mcode-object-version=1 \
 // RUN:   --offload-arch=gfx906 --rocm-path=%S/Inputs/rocm \
-// RUN:   %s 2>&1 | FileCheck -check-prefix=INVALID %s
-// INVALID: error: invalid integral value '1' in '-mcode-object-version=1'
-// INVALID-NOT: error: invalid integral value
+// RUN:   %s 2>&1 | FileCheck -check-prefix=INVALID_1 %s
+// INVALID_1: error: invalid integral value '1' in '-mcode-object-version=1'
+// INVALID_1-NOT: error: invalid integral value
+
+// RUN: not %clang -### --target=x86_64-linux-gnu \
+// RUN:   -mcode-object-version=2 \
+// RUN:   --offload-arch=gfx906 --rocm-path=%S/Inputs/rocm \
+// RUN:   %s 2>&1 | FileCheck -check-prefix=INVALID_2 %s
+// INVALID_2: error: invalid integral value '2' in '-mcode-object-version=2'
+// INVALID_2-NOT: error: invalid integral value
 
 // Check LLVM code object version option --amdhsa-code-object-version
 // is passed to -cc1 and -cc1as, and -mcode-object-version is passed

diff --git a/lld/test/ELF/Inputs/amdgpu-kernel-0.s b/lld/test/ELF/Inputs/amdgpu-kernel-0.s
@@ -1,6 +1,5 @@
 .text
 .globl kernel_0
 .align 64
-.amdgpu_hsa_kernel kernel_0
 kernel_0:
   s_endpgm
diff --git a/lld/test/ELF/Inputs/amdgpu-kernel-1.s b/lld/test/ELF/Inputs/amdgpu-kernel-1.s
@@ -1,6 +1,5 @@
 .text
 .globl kernel_1
 .align 64
-.amdgpu_hsa_kernel kernel_1
 kernel_1:
   s_endpgm
diff --git a/lld/test/ELF/amdgpu-abi-version-err.s b/lld/test/ELF/amdgpu-abi-version-err.s
@@ -1,6 +1,6 @@
 # REQUIRES: amdgpu
-# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx900 -filetype=obj %s -o %t-0.o
-# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx900 --amdhsa-code-object-version=2 -filetype=obj %s -o %t-1.o
+# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx900 --amdhsa-code-object-version=4 -filetype=obj %s -o %t-0.o
+# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx900 --amdhsa-code-object-version=5 -filetype=obj %s -o %t-1.o
 # RUN: not ld.lld -shared %t-0.o %t-1.o -o /dev/null 2>&1 | FileCheck %s
 
 # CHECK: ld.lld: error: incompatible ABI version: {{.*}}-1.o

diff --git a/lld/test/ELF/amdgpu-elf-flags-err.s b/lld/test/ELF/amdgpu-elf-flags-err.s
@@ -1,6 +1,6 @@
 # REQUIRES: amdgpu
-# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx802 --amdhsa-code-object-version=2 -filetype=obj %S/Inputs/amdgpu-kernel-0.s -o %t-0.o
-# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx803 --amdhsa-code-object-version=2 -filetype=obj %S/Inputs/amdgpu-kernel-1.s -o %t-1.o
+# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx802 --amdhsa-code-object-version=4 -filetype=obj %S/Inputs/amdgpu-kernel-0.s -o %t-0.o
+# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx803 --amdhsa-code-object-version=4 -filetype=obj %S/Inputs/amdgpu-kernel-1.s -o %t-1.o
 # RUN: not ld.lld -shared %t-0.o %t-1.o -o /dev/null 2>&1 | FileCheck %s
 
-# CHECK: error: incompatible e_flags: {{.*}}-1.o
+# CHECK: error: incompatible mach: {{.*}}-1.o
diff --git a/lld/test/ELF/amdgpu-elf-flags.s b/lld/test/ELF/amdgpu-elf-flags.s
@@ -1,6 +1,6 @@
 # REQUIRES: amdgpu
-# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx803 --amdhsa-code-object-version=2 -filetype=obj %S/Inputs/amdgpu-kernel-0.s -o %t-0.o
-# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx803 --amdhsa-code-object-version=2 -filetype=obj %S/Inputs/amdgpu-kernel-1.s -o %t-1.o
+# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx803 --amdhsa-code-object-version=4 -filetype=obj %S/Inputs/amdgpu-kernel-0.s -o %t-0.o
+# RUN: llvm-mc -triple amdgcn-amd-amdhsa -mcpu=gfx803 --amdhsa-code-object-version=4 -filetype=obj %S/Inputs/amdgpu-kernel-1.s -o %t-1.o
 # RUN: ld.lld -shared %t-0.o %t-1.o -o %t.so
 # RUN: llvm-readobj --file-headers %t.so | FileCheck --check-prefix=FIRSTLINK %s
 

diff --git a/lld/test/ELF/amdgpu-kernels.s b/lld/test/ELF/amdgpu-kernels.s
diff --git a/llvm/docs/AMDGPUUsage.rst b/llvm/docs/AMDGPUUsage.rst
@@ -1705,8 +1705,7 @@ Code Object V2 Note Records
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. warning::
-  Code object V2 is not the default code object version emitted by
-  this version of LLVM.
+  Code object V2 generation is no longer supported by this version of LLVM.
 
 The AMDGPU backend code object uses the following ELF note record in the
 ``.note`` section when compiling for code object V2.
@@ -2974,8 +2973,7 @@ Code Object V2 Metadata
 +++++++++++++++++++++++
 
 .. warning::
-  Code object V2 is not the default code object version emitted by this version
-  of LLVM.
+  Code object V2 generation is no longer supported by this version of LLVM.
 
 Code object V2 metadata is specified by the ``NT_AMD_HSA_METADATA`` note record
 (see :ref:`amdgpu-note-records-v2`).
@@ -14955,8 +14953,7 @@ Code Object V2 Predefined Symbols
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. warning::
-  Code object V2 is not the default code object version emitted by
-  this version of LLVM.
+  Code object V2 generation is no longer supported by this version of LLVM.
 
 The AMDGPU assembler defines and updates some symbols automatically. These
 symbols do not affect code generation.
@@ -15011,8 +15008,7 @@ Code Object V2 Directives
 ~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. warning::
-  Code object V2 is not the default code object version emitted by
-  this version of LLVM.
+  Code object V2 generation is no longer supported by this version of LLVM.
 
 AMDGPU ABI defines auxiliary data in output code object. In assembly source,
 one can specify them with assembler directives.
@@ -15087,8 +15083,7 @@ Code Object V2 Example Source Code
 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 .. warning::
-  Code Object V2 is not the default code object version emitted by
-  this version of LLVM.
+  Code object V2 generation is no longer supported by this version of LLVM.
 
 Here is an example of a minimal assembly source file, defining one HSA kernel:
 

diff --git a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
@@ -121,35 +121,21 @@ void AMDGPUAsmPrinter::initTargetStreamer(Module &M) {
       TM.getTargetTriple().getOS() != Triple::AMDPAL)
     return;
 
-  if (CodeObjectVersion >= AMDGPU::AMDHSA_COV3)
-    getTargetStreamer()->EmitDirectiveAMDGCNTarget();
+  getTargetStreamer()->EmitDirectiveAMDGCNTarget();
 
   if (TM.getTargetTriple().getOS() == Triple::AMDHSA)
     HSAMetadataStream->begin(M, *getTargetStreamer()->getTargetID());
 
   if (TM.getTargetTriple().getOS() == Triple::AMDPAL)
     getTargetStreamer()->getPALMetadata()->readFromIR(M);
-
-  if (CodeObjectVersion >= AMDGPU::AMDHSA_COV3)
-    return;
-
-  // HSA emits NT_AMD_HSA_CODE_OBJECT_VERSION for code objects v2.
-  if (TM.getTargetTriple().getOS() == Triple::AMDHSA)
-    getTargetStreamer()->EmitDirectiveHSACodeObjectVersion(2, 1);
-
-  // HSA and PAL emit NT_AMD_HSA_ISA_VERSION for code objects v2.
-  IsaVersion Version = getIsaVersion(getGlobalSTI()->getCPU());
-  getTargetStreamer()->EmitDirectiveHSACodeObjectISAV2(
-      Version.Major, Version.Minor, Version.Stepping, "AMD", "AMDGPU");
 }
 
 void AMDGPUAsmPrinter::emitEndOfAsmFile(Module &M) {
   // Init target streamer if it has not yet happened
   if (!IsTargetStreamerInitialized)
     initTargetStreamer(M);
 
-  if (TM.getTargetTriple().getOS() != Triple::AMDHSA ||
-      CodeObjectVersion == AMDGPU::AMDHSA_COV2)
+  if (TM.getTargetTriple().getOS() != Triple::AMDHSA)
     getTargetStreamer()->EmitISAVersion();
 
   // Emit HSA Metadata (NT_AMD_AMDGPU_HSA_METADATA).
@@ -209,7 +195,7 @@ void AMDGPUAsmPrinter::emitFunctionBodyStart() {
   if (!MFI.isEntryFunction())
     return;
 
-  if ((STM.isMesaKernel(F) || CodeObjectVersion == AMDGPU::AMDHSA_COV2) &&
+  if (STM.isMesaKernel(F) &&
       (F.getCallingConv() == CallingConv::AMDGPU_KERNEL ||
        F.getCallingConv() == CallingConv::SPIR_KERNEL)) {
     amd_kernel_code_t KernelCode;
@@ -226,8 +212,7 @@ void AMDGPUAsmPrinter::emitFunctionBodyEnd() {
   if (!MFI.isEntryFunction())
     return;
 
-  if (TM.getTargetTriple().getOS() != Triple::AMDHSA ||
-      CodeObjectVersion == AMDGPU::AMDHSA_COV2)
+  if (TM.getTargetTriple().getOS() != Triple::AMDHSA)
     return;
 
   auto &Streamer = getTargetStreamer()->getStreamer();
@@ -261,8 +246,7 @@ void AMDGPUAsmPrinter::emitFunctionBodyEnd() {
 }
 
 void AMDGPUAsmPrinter::emitFunctionEntryLabel() {
-  if (TM.getTargetTriple().getOS() == Triple::AMDHSA &&
-      CodeObjectVersion >= AMDGPU::AMDHSA_COV3) {
+  if (TM.getTargetTriple().getOS() == Triple::AMDHSA) {
     AsmPrinter::emitFunctionEntryLabel();
     return;
   }
@@ -337,9 +321,6 @@ bool AMDGPUAsmPrinter::doInitialization(Module &M) {
 
   if (TM.getTargetTriple().getOS() == Triple::AMDHSA) {
     switch (CodeObjectVersion) {
-    case AMDGPU::AMDHSA_COV2:
-      HSAMetadataStream.reset(new HSAMD::MetadataStreamerYamlV2());
-      break;
     case AMDGPU::AMDHSA_COV3:
       HSAMetadataStream.reset(new HSAMD::MetadataStreamerMsgPackV3());
       break;