[Cranelift] fold (or (and ..) (not ..)

cranelift/codegen/src/opts/bitops.isle

@@ -221,3 +221,7 @@
 (rule (simplify (bxor ty (bxor ty x y) y)) x)
 (rule (simplify (bxor ty y (bxor ty x y))) x)
 
+; (x & y) | ~x -> y | ~x
+(rule (simplify (bor ty (band ty x y) (bnot ty x))) (bor ty y (bnot ty x)))
+(rule (simplify (bor ty (bnot ty x) (band ty x y))) (bor ty y (bnot ty x)))
+

cranelift/filetests/filetests/egraph/fold-or-and-not.clif

@@ -0,0 +1,64 @@
+test optimize precise-output
+set opt_level=speed
+target x86_64
+
+function %test1(i8, i8) -> i8 {
+block0(v0: i8, v1: i8):
+    v2 = band v0, v1
+    v3 = bnot v0
+    v4 = bor v2, v3
+    return v4
+}
+
+; function %test1(i8, i8) -> i8 fast {
+; block0(v0: i8, v1: i8):
+;     v3 = bnot v0
+;     v5 = bor v1, v3
+;     return v5
+; }
+
+function %test2(i16, i16) -> i16 {
+block0(v0: i16, v1: i16):
+    v2 = band v0, v1
+    v3 = bnot v0
+    v4 = bor v2, v3
+    return v4
+}
+
+; function %test2(i16, i16) -> i16 fast {
+; block0(v0: i16, v1: i16):
+;     v3 = bnot v0
+;     v5 = bor v1, v3
+;     return v5
+; }
+
+function %test3(i32, i32) -> i32 {
+block0(v0: i32, v1: i32):
+    v2 = band v0, v1
+    v3 = bnot v0
+    v4 = bor v2, v3
+    return v4
+}
+
+; function %test3(i32, i32) -> i32 fast {
+; block0(v0: i32, v1: i32):
+;     v3 = bnot v0
+;     v5 = bor v1, v3
+;     return v5
+; }
+
+function %test4(i64, i64) -> i64 {
+block0(v0: i64, v1: i64):
+    v2 = band v0, v1
+    v3 = bnot v0
+    v4 = bor v2, v3
+    return v4
+}
+
+; function %test4(i64, i64) -> i64 fast {
+; block0(v0: i64, v1: i64):
+;     v3 = bnot v0
+;     v5 = bor v1, v3
+;     return v5
+; }
+

cranelift/filetests/filetests/runtests/bitops.clif

@@ -212,3 +212,13 @@ block0(v0: i64, v1: i64):
 
 ; run: %fold_xor_xor15(0xdead0000, 0x0000beef) == 0xdead0000
 
+function %fold_or_and_not(i32, i32) -> i32 {
+block0(v0: i32, v1: i32):
+    v2 = band v0, v1
+    v3 = bnot v0
+    v4 = bor v2, v3
+    return v4
+}
+
+; run: %fold_or_and_not(0, 0) == 0xffffffff
+