[llvm] [PowerPC][AIX] Using milicode for memcmp instead of libcall (PR #147093)

[zhijian lin via llvm-commits]

https://github.com/diggerlin created https://github.com/llvm/llvm-project/pull/147093

AIX has "millicode" routines, which are functions loaded at boot time into fixed addresses in kernel memory. This allows them to be customized for the processor. The __memcmp routine is a millicode implementation; we use millicode for the memcmp function instead of a library call to improve performance.

>From 16a43afcf4769d6b2fd06d4502be11f2920ccb5a Mon Sep 17 00:00:00 2001
From: zhijian <zhijian at ca.ibm.com>
Date: Fri, 27 Jun 2025 21:02:33 +0000
Subject: [PATCH 1/2] implement milicode memcmp

---
 llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
index ecd1ff87e7fbc..5bc71bdbccf43 100644
--- a/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
+++ b/llvm/lib/CodeGen/SelectionDAG/SelectionDAGBuilder.cpp
@@ -9065,8 +9065,15 @@ bool SelectionDAGBuilder::visitMemCmpBCmpCall(const CallInst &I) {
 
   // memcmp(S1,S2,2) != 0 -> (*(short*)LHS != *(short*)RHS)  != 0
   // memcmp(S1,S2,4) != 0 -> (*(int*)LHS != *(int*)RHS)  != 0
-  if (!CSize || !isOnlyUsedInZeroEqualityComparison(&I))
+  if (!CSize || !isOnlyUsedInZeroEqualityComparison(&I)) {
+    const Triple& TheTriple = TM.getTargetTriple();
+    if(TheTriple.isOSAIX()) {
+      if (Function *F = I.getCalledFunction()) {
+	F->setName(TheTriple.isArch32Bit() ? "___memcmp" : "___memcmp64");
+      }
+    }
     return false;
+  }
 
   // If the target has a fast compare for the given size, it will return a
   // preferred load type for that size. Require that the load VT is legal and

>From c97352653a6c2cbe3cdfa66adbfd90d5c7b0f4f9 Mon Sep 17 00:00:00 2001
From: zhijian <zhijian at ca.ibm.com>
Date: Fri, 4 Jul 2025 18:02:53 +0000
Subject: [PATCH 2/2] modify test case based on new behavious

---
 llvm/test/CodeGen/PowerPC/memintr32.ll | 2 +-
 llvm/test/CodeGen/PowerPC/memintr64.ll | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/llvm/test/CodeGen/PowerPC/memintr32.ll b/llvm/test/CodeGen/PowerPC/memintr32.ll
index c07a5af17e48a..4f0a9960a546d 100644
--- a/llvm/test/CodeGen/PowerPC/memintr32.ll
+++ b/llvm/test/CodeGen/PowerPC/memintr32.ll
@@ -11,7 +11,7 @@ define i32 @memcmp_test(ptr nocapture noundef readonly %ptr1, ptr nocapture noun
 ; CHECK-AIX-32-P9-NEXT:    mflr r0
 ; CHECK-AIX-32-P9-NEXT:    stwu r1, -64(r1)
 ; CHECK-AIX-32-P9-NEXT:    stw r0, 72(r1)
-; CHECK-AIX-32-P9-NEXT:    bl .memcmp[PR]
+; CHECK-AIX-32-P9-NEXT:    bl .___memcmp[PR]
 ; CHECK-AIX-32-P9-NEXT:    nop
 ; CHECK-AIX-32-P9-NEXT:    addi r1, r1, 64
 ; CHECK-AIX-32-P9-NEXT:    lwz r0, 8(r1)
diff --git a/llvm/test/CodeGen/PowerPC/memintr64.ll b/llvm/test/CodeGen/PowerPC/memintr64.ll
index b3a6650b8f6e6..0b0e556e89b51 100644
--- a/llvm/test/CodeGen/PowerPC/memintr64.ll
+++ b/llvm/test/CodeGen/PowerPC/memintr64.ll
@@ -39,7 +39,7 @@ define noundef i32 @_Z11memcmp_testPKvS0_m(ptr noundef readonly captures(none) %
 ; CHECK-AIX-64-P9-NEXT:    mflr r0
 ; CHECK-AIX-64-P9-NEXT:    stdu r1, -112(r1)
 ; CHECK-AIX-64-P9-NEXT:    std r0, 128(r1)
-; CHECK-AIX-64-P9-NEXT:    bl .memcmp[PR]
+; CHECK-AIX-64-P9-NEXT:    bl .___memcmp64[PR]
 ; CHECK-AIX-64-P9-NEXT:    nop
 ; CHECK-AIX-64-P9-NEXT:    addi r1, r1, 112
 ; CHECK-AIX-64-P9-NEXT:    ld r0, 16(r1)