[PATCH] D43281: [AMDGPU] fixes for lds f32 builtins

[Daniil Fukalov via Phabricator via cfe-commits]

dfukalov created this revision.
dfukalov added reviewers: b-sumner, arsenm.
dfukalov added a project: AMDGPU.
Herald added subscribers: cfe-commits, t-tye, tpr, dstuttard, yaxunl, nhaehnle, wdng, kzhuravl.

1. removed addrspace 3 specifications from builtins description strings since it's not target addrspaces
2. added custom processing for these builtins


Repository:
  rC Clang

https://reviews.llvm.org/D43281

Files:
  include/clang/Basic/BuiltinsAMDGPU.def
  lib/CodeGen/CGBuiltin.cpp
  test/CodeGenOpenCL/builtins-amdgcn-vi.cl


Index: test/CodeGenOpenCL/builtins-amdgcn-vi.cl
===================================================================
--- test/CodeGenOpenCL/builtins-amdgcn-vi.cl
+++ test/CodeGenOpenCL/builtins-amdgcn-vi.cl
@@ -91,18 +91,18 @@
 
 // CHECK-LABEL: @test_ds_fadd
 // CHECK: call float @llvm.amdgcn.ds.fadd(float addrspace(3)* %out, float %src, i32 0, i32 0, i1 false)
-void test_ds_fadd(__attribute__((address_space(3))) float *out, float src) {
+void test_ds_fadd(local float *out, float src) {
   *out = __builtin_amdgcn_ds_fadd(out, src, 0, 0, false);
 }
 
 // CHECK-LABEL: @test_ds_fmin
 // CHECK: call float @llvm.amdgcn.ds.fmin(float addrspace(3)* %out, float %src, i32 0, i32 0, i1 false)
-void test_ds_fmin(__attribute__((address_space(3))) float *out, float src) {
+void test_ds_fmin(local float *out, float src) {
   *out = __builtin_amdgcn_ds_fmin(out, src, 0, 0, false);
 }
 
 // CHECK-LABEL: @test_ds_fmax
 // CHECK: call float @llvm.amdgcn.ds.fmax(float addrspace(3)* %out, float %src, i32 0, i32 0, i1 false)
-void test_ds_fmax(__attribute__((address_space(3))) float *out, float src) {
+void test_ds_fmax(local float *out, float src) {
   *out = __builtin_amdgcn_ds_fmax(out, src, 0, 0, false);
 }
Index: lib/CodeGen/CGBuiltin.cpp
===================================================================
--- lib/CodeGen/CGBuiltin.cpp
+++ lib/CodeGen/CGBuiltin.cpp
@@ -9860,6 +9860,29 @@
     CI->setConvergent();
     return CI;
   }
+  case AMDGPU::BI__builtin_amdgcn_ds_fadd:
+  case AMDGPU::BI__builtin_amdgcn_ds_fmin:
+  case AMDGPU::BI__builtin_amdgcn_ds_fmax: {
+    llvm::SmallVector<llvm::Value *, 5> Args;
+    for (unsigned I = 0; I != 5; ++I)
+      Args.push_back(EmitScalarExpr(E->getArg(I)));
+    Intrinsic::ID ID;
+    switch (BuiltinID) {
+    case AMDGPU::BI__builtin_amdgcn_ds_fadd:
+      ID = Intrinsic::amdgcn_ds_fadd;
+      break;
+    case AMDGPU::BI__builtin_amdgcn_ds_fmin:
+      ID = Intrinsic::amdgcn_ds_fmin;
+      break;
+    case AMDGPU::BI__builtin_amdgcn_ds_fmax:
+      ID = Intrinsic::amdgcn_ds_fmax;
+      break;
+    default:
+      llvm_unreachable("Unknown BuiltinID");
+    }
+    Value *F = CGM.getIntrinsic(ID);
+    return Builder.CreateCall(F, Args);
+  }
 
   // amdgcn workitem
   case AMDGPU::BI__builtin_amdgcn_workitem_id_x:
Index: include/clang/Basic/BuiltinsAMDGPU.def
===================================================================
--- include/clang/Basic/BuiltinsAMDGPU.def
+++ include/clang/Basic/BuiltinsAMDGPU.def
@@ -93,9 +93,9 @@
 BUILTIN(__builtin_amdgcn_readfirstlane, "ii", "nc")
 BUILTIN(__builtin_amdgcn_readlane, "iii", "nc")
 BUILTIN(__builtin_amdgcn_fmed3f, "ffff", "nc")
-BUILTIN(__builtin_amdgcn_ds_fadd, "ff*3fiib", "n")
-BUILTIN(__builtin_amdgcn_ds_fmin, "ff*3fiib", "n")
-BUILTIN(__builtin_amdgcn_ds_fmax, "ff*3fiib", "n")
+BUILTIN(__builtin_amdgcn_ds_fadd, "ff*fiIiIb", "n")
+BUILTIN(__builtin_amdgcn_ds_fmin, "ff*fiIiIb", "n")
+BUILTIN(__builtin_amdgcn_ds_fmax, "ff*fiIiIb", "n")
 
 //===----------------------------------------------------------------------===//
 // VI+ only builtins.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D43281.134185.patch
Type: text/x-patch
Size: 3075 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20180214/df2a0cd9/attachment.bin>