================
@@ -112,42 +104,46 @@ entry:
define i32 @bcmp_size_2(ptr %s1, ptr %s2) nounwind optsize {
; CHECK-ALIGNED-RV32-LABEL: bcmp_size_2:
; CHECK-ALIGNED-RV32: # %bb.0: # %entry
-; CHECK-ALIGNED-RV32-NEXT: addi sp, sp, -16
-; CHECK-ALIGNED-RV32-NEXT: sw ra, 12(sp) # 4-byte Folded Spill
-; CHECK-ALIGNED-RV32-NEXT: li a2, 2
-; CHECK-ALIGNED-RV32-NEXT: call bcmp
-; CHECK-ALIGNED-RV32-NEXT: lw ra, 12(sp) # 4-byte Folded Reload
-; CHECK-ALIGNED-RV32-NEXT: addi sp, sp, 16
+; CHECK-ALIGNED-RV32-NEXT: lbu a2, 1(a0)
----------------
wangpc-pp wrote:
This seems to be more complicated than what I was thinking. For this case, we
generate i16 loads/compares, and then we expand them since unaligned loads are
illegal. We may expand memcmp to byte loop directly when there is no unaligned
access.
https://github.com/llvm/llvm-project/pull/107548
_______________________________________________
llvm-branch-commits mailing list
llvm-branch-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/llvm-branch-commits
