[AArch64][Falkor] Avoid HW prefetcher tag collisions (step 2)

Summary:
Avoid HW prefetcher instruction tag collisions in loops by inserting
MOVs to change the base address register of strided loads.

Reviewers: t.p.northover, mcrosier

Subscribers: aemerson, rengolin, javed.absar, kristof.beyls, hfinkel, llvm-commits

Differential Revision: https://reviews.llvm.org/D35366

llvm-svn: 308324

llvm/lib/Target/AArch64/AArch64.h

@@ -44,6 +44,7 @@ ModulePass *createAArch64PromoteConstantPass();
 FunctionPass *createAArch64ConditionOptimizerPass();
 FunctionPass *createAArch64A57FPLoadBalancing();
 FunctionPass *createAArch64A53Fix835769();
+FunctionPass *createFalkorHWPFFixPass();
 FunctionPass *createFalkorMarkStridedAccessesPass();
 
 FunctionPass *createAArch64CleanupLocalDynamicTLSPass();
@@ -67,6 +68,7 @@ void initializeAArch64VectorByElementOptPass(PassRegistry&);
 void initializeAArch64PromoteConstantPass(PassRegistry&);
 void initializeAArch64RedundantCopyEliminationPass(PassRegistry&);
 void initializeAArch64StorePairSuppressPass(PassRegistry&);
+void initializeFalkorHWPFFixPass(PassRegistry&);
 void initializeFalkorMarkStridedAccessesLegacyPass(PassRegistry&);
 void initializeLDTLSCleanupPass(PassRegistry&);
 } // end namespace llvm