[llvm] [AArch64][SVE2] Lower read-after-write mask to whilerw (PR #114028)

[Graham Hunter via llvm-commits]

================
@@ -30,6 +30,36 @@ entry:
   ret <vscale x 16 x i1> %active.lane.mask.alias
 }
 
+define <vscale x 16 x i1> @whilerw_8(ptr noalias %a, ptr %b, ptr %c, i32 %n) {
+; CHECK-LABEL: whilerw_8:
+; CHECK:       // %bb.0: // %entry
+; CHECK-NEXT:    whilerw p0.b, x2, x1
+; CHECK-NEXT:    ret
+;
+; CHECK-NOSVE2-LABEL: whilerw_8:
+; CHECK-NOSVE2:       // %bb.0: // %entry
+; CHECK-NOSVE2-NEXT:    subs x8, x2, x1
+; CHECK-NOSVE2-NEXT:    cneg x8, x8, mi
+; CHECK-NOSVE2-NEXT:    cmp x8, #0
+; CHECK-NOSVE2-NEXT:    cset w9, lt
+; CHECK-NOSVE2-NEXT:    whilelo p0.b, xzr, x8
+; CHECK-NOSVE2-NEXT:    sbfx x8, x9, #0, #1
+; CHECK-NOSVE2-NEXT:    whilelo p1.b, xzr, x8
+; CHECK-NOSVE2-NEXT:    sel p0.b, p0, p0.b, p1.b
+; CHECK-NOSVE2-NEXT:    ret
+entry:
+  %b24 = ptrtoint ptr %b to i64
+  %c25 = ptrtoint ptr %c to i64
+  %sub.diff = sub i64 %c25, %b24
+  %0 = tail call i64 @llvm.abs.i64(i64 %sub.diff, i1 false)
+  %neg.compare = icmp slt i64 %0, 0
----------------
huntergr-arm wrote:

This comparison is incorrect -- you shouldn't get a negative value after calling `abs`, and the ref manual indicates it should just be an equality comparison against 0 to combine with the `e < diff` comparison. I'm not sure what to do in the case of the diff being the signed minimum, though it seems unlikely to occur in practice.

I think we got the original code for whilewr wrong too -- should be `sle` against 0 instead of `slt`.

https://github.com/llvm/llvm-project/pull/114028