[llvm] 44648cc - [AMDGPU] Always emit lds_size in PAL ELF Metadata 3.0 (#87222)

[via llvm-commits]

Author: Carl Ritson
Date: 2024-05-03T17:01:03+09:00
New Revision: 44648ccb8b7e68a5c259ec857fd4041fdd41c588

URL: https://github.com/llvm/llvm-project/commit/44648ccb8b7e68a5c259ec857fd4041fdd41c588
DIFF: https://github.com/llvm/llvm-project/commit/44648ccb8b7e68a5c259ec857fd4041fdd41c588.diff

LOG: [AMDGPU] Always emit lds_size in PAL ELF Metadata 3.0 (#87222)

Emit lds_size for all shader types in PAL metadata.

Added: 
    

Modified: 
    llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
    llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll

Removed: 
    


################################################################################
diff  --git a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
index 052b231d62a3eb..c090d6133dbad4 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPUAsmPrinter.cpp
@@ -1053,11 +1053,11 @@ static void EmitPALMetadataCommon(AMDGPUPALMetadata *MD,
     MD->setHwStage(CC, ".trap_present",
                    (bool)CurrentProgramInfo.TrapHandlerEnable);
     MD->setHwStage(CC, ".excp_en", CurrentProgramInfo.EXCPEnable);
-
-    MD->setHwStage(CC, ".lds_size",
-                   (unsigned)(CurrentProgramInfo.LdsSize *
-                              getLdsDwGranularity(ST) * sizeof(uint32_t)));
   }
+
+  MD->setHwStage(CC, ".lds_size",
+                 (unsigned)(CurrentProgramInfo.LdsSize *
+                            getLdsDwGranularity(ST) * sizeof(uint32_t)));
 }
 
 // This is the equivalent of EmitProgramInfoSI above, but for when the OS type

diff  --git a/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll b/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll
index cabc11037017e5..6d18f354e65422 100644
--- a/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll
+++ b/llvm/test/CodeGen/AMDGPU/pal-metadata-3.0.ll
@@ -107,10 +107,33 @@
 ; CHECK-NEXT:        .vgpr_limit:     0x100
 ; CHECK-NEXT:        .wavefront_size: 0x40
 ; CHECK-NEXT:        .wgp_mode:       false
+; CHECK-NEXT:      .gs:
+; CHECK-NEXT:        .debug_mode:     false
+; CHECK-NEXT:        .entry_point:    gs_shader
+; CHECK-NEXT:        .ieee_mode:      false
+; CHECK-NEXT:        .lds_size:       0x200
+; CHECK-NEXT:        .mem_ordered:    true
+; CHECK-NEXT:        .scratch_en:     false
+; CHECK-NEXT:        .scratch_memory_size: 0
+; CHECK-NEXT:        .sgpr_count:     0x1
+; CHECK-NEXT:        .vgpr_count:     0x1
+; CHECK-NEXT:        .wgp_mode:       true
+; CHECK-NEXT:      .hs:
+; CHECK-NEXT:        .debug_mode:     false
+; CHECK-NEXT:        .entry_point:    hs_shader
+; CHECK-NEXT:        .ieee_mode:      false
+; CHECK-NEXT:        .lds_size:       0x1000
+; CHECK-NEXT:        .mem_ordered:    true
+; CHECK-NEXT:        .scratch_en:     false
+; CHECK-NEXT:        .scratch_memory_size: 0
+; CHECK-NEXT:        .sgpr_count:     0x1
+; CHECK-NEXT:        .vgpr_count:     0x1
+; CHECK-NEXT:        .wgp_mode:       true
 ; CHECK-NEXT:      .ps:
 ; CHECK-NEXT:        .debug_mode:     false
 ; CHECK-NEXT:        .entry_point:    ps_shader
 ; CHECK-NEXT:        .ieee_mode:      false
+; CHECK-NEXT:        .lds_size:       0
 ; CHECK-NEXT:        .mem_ordered:    true
 ; CHECK-NEXT:        .scratch_en:     false
 ; CHECK-NEXT:        .scratch_memory_size: 0
@@ -145,6 +168,22 @@ define dllexport amdgpu_ps void @ps_shader() #1 {
   ret void
 }
 
+ at LDS.GS = external addrspace(3) global [1 x i32], align 4
+
+define dllexport amdgpu_gs void @gs_shader() #2 {
+  %ptr = getelementptr i32, ptr addrspace(3) @LDS.GS, i32 0
+  store i32 0, ptr addrspace(3) %ptr, align 4
+  ret void
+}
+
+ at LDS.HS = external addrspace(3) global [1024 x i32], align 4
+
+define dllexport amdgpu_hs void @hs_shader() #2 {
+  %ptr = getelementptr i32, ptr addrspace(3) @LDS.HS, i32 0
+  store i32 0, ptr addrspace(3) %ptr, align 4
+  ret void
+}
+
 !amdgpu.pal.metadata.msgpack = !{!0}
 
 ; Function Attrs: nounwind willreturn memory(none)