[AArch64] Don't expand memcmp in strict align mode.

7aecf232 fixed the bug where we would miscompile, but we still generate
a crazy amount of code. Turn off the expansion until someone implements
an appropriate heuristic.

Differential Revision: https://reviews.llvm.org/D77599

diff --git a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
index 4724d6b..e8ba30c 100644 (file)
--- a/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
+++ b/llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp
@@ -629,7 +629,12 @@ int AArch64TTIImpl::getCmpSelInstrCost(unsigned Opcode, Type *ValTy,
 AArch64TTIImpl::TTI::MemCmpExpansionOptions
 AArch64TTIImpl::enableMemCmpExpansion(bool OptSize, bool IsZeroCmp) const {
   TTI::MemCmpExpansionOptions Options;
-  Options.AllowOverlappingLoads = !ST->requiresStrictAlign();
+  if (ST->requiresStrictAlign()) {
+    // TODO: Add cost modeling for strict align. Misaligned loads expand to
+    // a bunch of instructions when strict align is enabled.
+    return Options;
+  }
+  Options.AllowOverlappingLoads = true;
   Options.MaxNumLoads = TLI->getMaxExpandSizeMemcmp(OptSize);
   Options.NumLoadsPerBlock = Options.MaxNumLoads;
   // TODO: Though vector loads usually perform well on AArch64, in some targets

diff --git a/llvm/test/CodeGen/AArch64/bcmp-inline-small.ll b/llvm/test/CodeGen/AArch64/bcmp-inline-small.ll
index a7d0856..12eefa0 100644 (file)
--- a/llvm/test/CodeGen/AArch64/bcmp-inline-small.ll
+++ b/llvm/test/CodeGen/AArch64/bcmp-inline-small.ll
@@ -11,12 +11,12 @@ entry:
   ret i1 %ret
 
 ; CHECK-LABEL: test_b2:
-; CHECK-NOT:   bl bcmp
+; CHECKN-NOT:  bl bcmp
 ; CHECKN:      ldr  x
 ; CHECKN-NEXT: ldr  x
 ; CHECKN-NEXT: ldur x
 ; CHECKN-NEXT: ldur x
-; CHECKS-COUNT-30: ldrb w
+; CHECKS: bl bcmp
 }
 
 define i1 @test_b2_align8(i8* align 8 %s1, i8* align 8 %s2) {
@@ -26,19 +26,13 @@ entry:
   ret i1 %ret
 
 ; CHECK-LABEL: test_b2_align8:
-; CHECK-NOT:   bl bcmp
+; CHECKN-NOT:  bl bcmp
 ; CHECKN:      ldr  x
 ; CHECKN-NEXT: ldr  x
 ; CHECKN-NEXT: ldur x
 ; CHECKN-NEXT: ldur x
-; CHECKS:      ldr  x
-; CHECKS-NEXT: ldr  x
-; CHECKS-NEXT: ldr  w
-; CHECKS-NEXT: ldr  w
-; CHECKS-NEXT: ldrh  w
-; CHECKS-NEXT: ldrh  w
-; CHECKS-NEXT: ldrb  w
-; CHECKS-NEXT: ldrb  w
+; TODO: Four loads should be within the limit, but the heuristic isn't implemented.
+; CHECKS: bl bcmp
 }
 
 define i1 @test_bs(i8* %s1, i8* %s2) optsize {