vendor/llvm-project/llvmorg-21-init-19288-gface93e724f4

1 files changed, 40 insertions, 13 deletions

diff --git a/llvm/lib/Target/AMDGPU/AMDGPULowerModuleLDSPass.cpp b/llvm/lib/Target/AMDGPU/AMDGPULowerModuleLDSPass.cpp
index 55497c837ee2..d443f4ea7d5c 100644
--- a/llvm/lib/Target/AMDGPU/AMDGPULowerModuleLDSPass.cpp
+++ b/llvm/lib/Target/AMDGPU/AMDGPULowerModuleLDSPass.cpp
@@ -186,6 +186,7 @@
 #include "llvm/ADT/STLExtras.h"
 #include "llvm/ADT/SetOperations.h"
 #include "llvm/Analysis/CallGraph.h"
+#include "llvm/Analysis/ScopedNoAliasAA.h"
 #include "llvm/CodeGen/TargetPassConfig.h"
 #include "llvm/IR/Constants.h"
 #include "llvm/IR/DerivedTypes.h"
@@ -195,7 +196,6 @@
 #include "llvm/IR/IntrinsicsAMDGPU.h"
 #include "llvm/IR/MDBuilder.h"
 #include "llvm/IR/ReplaceConstant.h"
-#include "llvm/InitializePasses.h"
 #include "llvm/Pass.h"
 #include "llvm/Support/CommandLine.h"
 #include "llvm/Support/Debug.h"
@@ -532,8 +532,7 @@ public:
       auto InsertAt = F->getEntryBlock().getFirstNonPHIOrDbgOrAlloca();
       IRBuilder<> Builder(&*InsertAt);
 
-      It->second =
-          Builder.CreateIntrinsic(Intrinsic::amdgcn_lds_kernel_id, {}, {});
+      It->second = Builder.CreateIntrinsic(Intrinsic::amdgcn_lds_kernel_id, {});
     }
 
     return It->second;
@@ -573,7 +572,7 @@ public:
 
       if (OrderedKernels.size() > UINT32_MAX) {
         // 32 bit keeps it in one SGPR. > 2**32 kernels won't fit on the GPU
-        report_fatal_error("Unimplemented LDS lowering for > 2**32 kernels");
+        reportFatalUsageError("unimplemented LDS lowering for > 2**32 kernels");
       }
 
       for (size_t i = 0; i < OrderedKernels.size(); i++) {
@@ -633,7 +632,8 @@ public:
         if (K.second.size() == 1) {
           KernelAccessVariables.insert(GV);
         } else {
-          report_fatal_error(
+          // FIXME: This should use DiagnosticInfo
+          reportFatalUsageError(
               "cannot lower LDS '" + GV->getName() +
               "' to kernel access as it is reachable from multiple kernels");
         }
@@ -782,7 +782,7 @@ public:
       // backend) difficult to use. This does mean that llvm test cases need
       // to name the kernels.
       if (!Func.hasName()) {
-        report_fatal_error("Anonymous kernels cannot use LDS variables");
+        reportFatalUsageError("anonymous kernels cannot use LDS variables");
       }
 
       std::string VarName =
@@ -878,7 +878,7 @@ public:
         if (KernelsThatIndirectlyAllocateDynamicLDS.contains(func)) {
           assert(isKernelLDS(func));
           if (!func->hasName()) {
-            report_fatal_error("Anonymous kernels cannot use LDS variables");
+            reportFatalUsageError("anonymous kernels cannot use LDS variables");
           }
 
           GlobalVariable *N =
@@ -1016,7 +1016,7 @@ public:
         auto NewGV = uniquifyGVPerKernel(M, GV, F);
         Changed |= (NewGV != GV);
         int BarId = (NumAbsolutes + 1);
-        if (Kernel2BarId.find(F) != Kernel2BarId.end()) {
+        if (Kernel2BarId.contains(F)) {
           BarId = (Kernel2BarId[F] + 1);
         }
         Kernel2BarId[F] = BarId;
@@ -1442,6 +1442,8 @@ private:
     if (!MaxDepth || (A == 1 && !AliasScope))
       return;
 
+    ScopedNoAliasAAResult ScopedNoAlias;
+
     for (User *U : Ptr->users()) {
       if (auto *I = dyn_cast<Instruction>(U)) {
         if (AliasScope && I->mayReadOrWriteMemory()) {
@@ -1451,7 +1453,34 @@ private:
           I->setMetadata(LLVMContext::MD_alias_scope, AS);
 
           MDNode *NA = I->getMetadata(LLVMContext::MD_noalias);
-          NA = (NA ? MDNode::intersect(NA, NoAlias) : NoAlias);
+
+          // Scoped aliases can originate from two different domains.
+          // First domain would be from LDS domain (created by this pass).
+          // All entries (LDS vars) into LDS struct will have same domain.
+
+          // Second domain could be existing scoped aliases that are the
+          // results of noalias params and subsequent optimizations that
+          // may alter thesse sets.
+
+          // We need to be careful how we create new alias sets, and
+          // have right scopes and domains for loads/stores of these new
+          // LDS variables. We intersect NoAlias set if alias sets belong
+          // to the same domain. This is the case if we have memcpy using
+          // LDS variables. Both src and dst of memcpy would belong to
+          // LDS struct, they donot alias.
+          // On the other hand, if one of the domains is LDS and other is
+          // existing domain prior to LDS, we need to have a union of all
+          // these aliases set to preserve existing aliasing information.
+
+          SmallPtrSet<const MDNode *, 16> ExistingDomains, LDSDomains;
+          ScopedNoAlias.collectScopedDomains(NA, ExistingDomains);
+          ScopedNoAlias.collectScopedDomains(NoAlias, LDSDomains);
+          auto Intersection = set_intersection(ExistingDomains, LDSDomains);
+          if (Intersection.empty()) {
+            NA = NA ? MDNode::concatenate(NA, NoAlias) : NoAlias;
+          } else {
+            NA = NA ? MDNode::intersect(NA, NoAlias) : NoAlias;
+          }
           I->setMetadata(LLVMContext::MD_noalias, NA);
         }
       }
@@ -1503,10 +1532,8 @@ public:
   const AMDGPUTargetMachine *TM;
   static char ID;
 
-  AMDGPULowerModuleLDSLegacy(const AMDGPUTargetMachine *TM_ = nullptr)
-      : ModulePass(ID), TM(TM_) {
-    initializeAMDGPULowerModuleLDSLegacyPass(*PassRegistry::getPassRegistry());
-  }
+  AMDGPULowerModuleLDSLegacy(const AMDGPUTargetMachine *TM = nullptr)
+      : ModulePass(ID), TM(TM) {}
 
   void getAnalysisUsage(AnalysisUsage &AU) const override {
     if (!TM)