A lightweight WebAssembly runtime that is fast, secure, and standards-compliant
[Cranelift] add simplification rules (#12948)
Hyunbin Kima4fef4299b4e
diff --git a/cranelift/codegen/src/opts/arithmetic.isle b/cranelift/codegen/src/opts/arithmetic.isleindex ba41c5a145..2e0703ef6c 100644--- a/cranelift/codegen/src/opts/arithmetic.isle+++ b/cranelift/codegen/src/opts/arithmetic.isle@@ -386,5 +386,54 @@(rule (simplify (eq (ty_int ty) (iadd cty y x) (iadd cty y x))) (iconst_u ty 1));; (x - y) != x --> y != 0-(rule (simplify (ne cty (isub (ty_int ty) x y) x)) (ne cty y (iconst_u ty 0)))-(rule (simplify (ne cty x (isub (ty_int ty) x y))) (ne cty y (iconst_u ty 0)))+(rule (simplify (ne cty (isub ty x y) x)) (ne cty y (iconst_u ty 0)))+(rule (simplify (ne cty x (isub ty x y))) (ne cty y (iconst_u ty 0)))++;; (x - y) == x --> y == 0+(rule (simplify (eq cty (isub ty x y) x)) (eq cty y (iconst_u ty 0)))+(rule (simplify (eq cty x (isub ty x y))) (eq cty y (iconst_u ty 0)))++;; (x + y) == y --> x == 0+(rule (simplify (eq cty (iadd ty x y) y)) (eq cty x (iconst_u ty 0)))+(rule (simplify (eq cty (iadd ty y x) y)) (eq cty x (iconst_u ty 0)))+(rule (simplify (eq cty y (iadd ty x y))) (eq cty x (iconst_u ty 0)))+(rule (simplify (eq cty y (iadd ty y x))) (eq cty x (iconst_u ty 0)))++;; -x == -y --> x == y+(rule (simplify (eq ty (ineg ty x) (ineg ty y))) (eq ty x y))++;; -((-y) * x) --> x * y+(rule (simplify (ineg ty (imul ty (ineg ty y) x))) (imul ty x y))+(rule (simplify (ineg ty (imul ty x (ineg ty y)))) (imul ty x y))++;; ireduce(xext(a) + xext(b)) --> a + b+(rule (simplify (ireduce ty (iadd cty (sextend cty x) (sextend cty y)))) (iadd ty x y))+(rule (simplify (ireduce ty (iadd cty (sextend cty x) (uextend cty y)))) (iadd ty x y))+(rule (simplify (ireduce ty (iadd cty (uextend cty x) (sextend cty y)))) (iadd ty x y))+(rule (simplify (ireduce ty (iadd cty (uextend cty x) (uextend cty y)))) (iadd ty x y))++;; ireduce(xext(a) - xext(b)) --> a - b+(rule (simplify (ireduce ty (isub cty (sextend cty x) (sextend cty y)))) (isub ty x y))+(rule (simplify (ireduce ty (isub cty (sextend cty x) (uextend cty y)))) (isub ty x y))+(rule (simplify (ireduce ty (isub cty (uextend cty x) (sextend cty y)))) (isub ty x y))+(rule (simplify (ireduce ty (isub cty (uextend cty x) (uextend cty y)))) (isub ty x y))++;; max(x, y) >= x+(rule (simplify (sge ty (smax ty x y) x)) (iconst_u ty 1))+(rule (simplify (sge ty (smax ty y x) x)) (iconst_u ty 1))+(rule (simplify (sle ty x (smax ty x y))) (iconst_u ty 1))+(rule (simplify (sle ty x (smax ty y x))) (iconst_u ty 1))+(rule (simplify (uge ty (umax ty x y) x)) (iconst_u ty 1))+(rule (simplify (uge ty (umax ty y x) x)) (iconst_u ty 1))+(rule (simplify (ule ty x (umax ty x y))) (iconst_u ty 1))+(rule (simplify (ule ty x (umax ty y x))) (iconst_u ty 1))++;; x >= min(x, y)+(rule (simplify (sge ty x (smin ty x y))) (iconst_u ty 1))+(rule (simplify (sge ty x (smin ty y x))) (iconst_u ty 1))+(rule (simplify (sle ty (smin ty x y) x)) (iconst_u ty 1))+(rule (simplify (sle ty (smin ty y x) x)) (iconst_u ty 1))+(rule (simplify (uge ty x (umin ty x y))) (iconst_u ty 1))+(rule (simplify (uge ty x (umin ty y x))) (iconst_u ty 1))+(rule (simplify (ule ty (umin ty x y) x)) (iconst_u ty 1))+(rule (simplify (ule ty (umin ty y x) x)) (iconst_u ty 1))\ No newline at end of filediff --git a/cranelift/codegen/src/opts/bitops.isle b/cranelift/codegen/src/opts/bitops.isleindex 646e03f018..4cba596fed 100644--- a/cranelift/codegen/src/opts/bitops.isle+++ b/cranelift/codegen/src/opts/bitops.isle@@ -696,4 +696,22 @@(rule (simplify (bxor ty (bor ty y (bnot ty x)) (bxor ty x y))) (bor ty x (bnot ty y)))(rule (simplify (bxor ty (bxor ty x y) (bor ty y (bnot ty x)))) (bor ty x (bnot ty y)))(rule (simplify (bxor ty (bor ty y (bnot ty x)) (bxor ty y x))) (bor ty x (bnot ty y)))-(rule (simplify (bxor ty (bxor ty y x) (bor ty y (bnot ty x)))) (bor ty x (bnot ty y)))\ No newline at end of file+(rule (simplify (bxor ty (bxor ty y x) (bor ty y (bnot ty x)))) (bor ty x (bnot ty y)))++;; (x & ~y) == ~y <=> (x | y) == -1+(rule (simplify (eq rty (band cty (bnot cty y) x) (bnot cty y))) (eq rty (bor cty y x) (iconst_s cty -1)))+(rule (simplify (eq rty (band cty x (bnot cty y)) (bnot cty y))) (eq rty (bor cty y x) (iconst_s cty -1)))+(rule (simplify (eq rty (bnot cty y) (band cty (bnot cty y) x))) (eq rty (bor cty y x) (iconst_s cty -1)))+(rule (simplify (eq rty (bnot cty y) (band cty x (bnot cty y)))) (eq rty (bor cty y x) (iconst_s cty -1)))++;; x >=_u (x & y), and (x & y) <=_u x, are always true.+(rule (simplify (uge ty x (band ty x y))) (iconst_u ty 1))+(rule (simplify (uge ty x (band ty y x))) (iconst_u ty 1))+(rule (simplify (ule ty (band ty x y) x)) (iconst_u ty 1))+(rule (simplify (ule ty (band ty y x) x)) (iconst_u ty 1))++;; (x | y) >=_u x, and x <=_u (x | y), are always true.+(rule (simplify (uge ty (bor ty x y) x)) (iconst_u ty 1))+(rule (simplify (uge ty (bor ty y x) x)) (iconst_u ty 1))+(rule (simplify (ule ty x (bor ty x y))) (iconst_u ty 1))+(rule (simplify (ule ty x (bor ty y x))) (iconst_u ty 1))\ No newline at end of filediff --git a/cranelift/filetests/filetests/egraph/arithmetic-precise.clif b/cranelift/filetests/filetests/egraph/arithmetic-precise.clifindex e25f339840..7249b7df8d 100644--- a/cranelift/filetests/filetests/egraph/arithmetic-precise.clif+++ b/cranelift/filetests/filetests/egraph/arithmetic-precise.clif@@ -264,17 +264,171 @@ block0(v0: i32, v1: i32):; return v5; }-;; (x - y) != x --> y != 0 (vector)-function %simplify_vector_icmp_ne_isub_x(i32x4, i32x4) -> i32x4 fast {-block0(v0: i32x4, v1: i32x4):+;; (x - y) == x --> y == 0+function %test_eq_isub_same(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):v2 = isub v0, v1- v3 = icmp ne v2, v0+ v3 = icmp eq v2, v0return v3}-; function %simplify_vector_icmp_ne_isub_x(i32x4, i32x4) -> i32x4 fast {-; block0(v0: i32x4, v1: i32x4):-; v2 = isub v0, v1-; v3 = icmp ne v2, v0+; function %test_eq_isub_same(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v4 = iconst.i32 0+; v5 = icmp eq v1, v4 ; v4 = 0+; return v5+; }++;; (x + y) == y --> x == 0+function %test_eq_iadd_same_rhs(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = iadd v0, v1+ v3 = icmp eq v2, v1+ return v3+}++; function %test_eq_iadd_same_rhs(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v4 = iconst.i32 0+; v5 = icmp eq v0, v4 ; v4 = 0+; return v5+; }++;; -x == -y --> x == y+function %test_eq_ineg_ineg(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = ineg v0+ v3 = ineg v1+ v4 = icmp eq v2, v3+ return v4+}++; function %test_eq_ineg_ineg(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = ineg v0+; v3 = ineg v1+; v4 = icmp eq v2, v3+; return v4+; }++;; -((-y) * x) --> x * y+function %test_ineg_imul_ineg_lhs(i32, i32) -> i32 fast {+block0(v0: i32, v1: i32):+ v2 = ineg v1+ v3 = imul v2, v0+ v4 = ineg v3+ return v4+}++; function %test_ineg_imul_ineg_lhs(i32, i32) -> i32 fast {+; block0(v0: i32, v1: i32):+; v5 = imul v0, v1+; return v5+; }++function %test_ineg_imul_ineg_rhs(i32, i32) -> i32 fast {+block0(v0: i32, v1: i32):+ v2 = ineg v1+ v3 = imul v0, v2+ v4 = ineg v3+ return v4+}++; function %test_ineg_imul_ineg_rhs(i32, i32) -> i32 fast {+; block0(v0: i32, v1: i32):+; v5 = imul v0, v1+; return v5+; }++;; ireduce(sextend(x) + uextend(y)) --> x + y+function %test_ireduce_iadd_extends(i8, i8) -> i8 fast {+block0(v0: i8, v1: i8):+ v2 = sextend.i16 v0+ v3 = uextend.i16 v1+ v4 = iadd v2, v3+ v5 = ireduce.i8 v4+ return v5+}++; function %test_ireduce_iadd_extends(i8, i8) -> i8 fast {+; block0(v0: i8, v1: i8):+; v6 = iadd v0, v1+; return v6+; }++;; ireduce(sextend(x) - uextend(y)) --> x - y+function %test_ireduce_isub_extends(i8, i8) -> i8 fast {+block0(v0: i8, v1: i8):+ v2 = sextend.i16 v0+ v3 = uextend.i16 v1+ v4 = isub v2, v3+ v5 = ireduce.i8 v4+ return v5+}++; function %test_ireduce_isub_extends(i8, i8) -> i8 fast {+; block0(v0: i8, v1: i8):+; v6 = isub v0, v1+; return v6+; }++;; smax(x, y) >= x --> true+function %test_sge_smax_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = smax v0, v1+ v3 = icmp sge v2, v0+ return v3+}++; function %test_sge_smax_x(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = smax v0, v1+; v3 = icmp sge v2, v0+; return v3+; }++;; x <=_u umax(x, y) --> true+function %test_ule_x_umax(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = umax v0, v1+ v3 = icmp ule v0, v2+ return v3+}++; function %test_ule_x_umax(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = umax v0, v1+; v3 = icmp ule v0, v2; return v3; }++;; x >= smin(x, y) --> true+function %test_sge_x_smin(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = smin v0, v1+ v3 = icmp sge v0, v2+ return v3+}++; function %test_sge_x_smin(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = smin v0, v1+; v3 = icmp sge v0, v2+; return v3+; }++;; umin(x, y) <=_u x --> true+function %test_ule_umin_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = umin v0, v1+ v3 = icmp ule v2, v0+ return v3+}++; function %test_ule_umin_x(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = umin v0, v1+; v3 = icmp ule v2, v0+; return v3+; }+diff --git a/cranelift/filetests/filetests/egraph/arithmetic.clif b/cranelift/filetests/filetests/egraph/arithmetic.clifindex b131428a6a..c0f169d7f6 100644--- a/cranelift/filetests/filetests/egraph/arithmetic.clif+++ b/cranelift/filetests/filetests/egraph/arithmetic.clif@@ -363,8 +363,8 @@ block0(v0: i16, v1: i16, v2: i16):return v8}-; check: v12 = iadd v0, v1-; check: v15 = iadd v12, v2+; check: v10 = iadd v0, v1+; check: v15 = iadd v10, v2; check: return v15;; or(x, C) + (-C) --> and(x, ~C)diff --git a/cranelift/filetests/filetests/egraph/extends.clif b/cranelift/filetests/filetests/egraph/extends.clifindex a4b260eebf..81b619b757 100644--- a/cranelift/filetests/filetests/egraph/extends.clif+++ b/cranelift/filetests/filetests/egraph/extends.clif@@ -179,8 +179,8 @@ block0(v0: i16, v1: i16):return v5}-; check: v8 = iadd v0, v1-; check: return v8+; check: v6 = iadd v0, v1+; check: return v6function %extend_bxor_reduce(i64, i64) -> i64 {block0(v0: i64, v1: i64):diff --git a/cranelift/filetests/filetests/egraph/fold-bitops.clif b/cranelift/filetests/filetests/egraph/fold-bitops.clifindex 6a273df125..b5e6d76f80 100644--- a/cranelift/filetests/filetests/egraph/fold-bitops.clif+++ b/cranelift/filetests/filetests/egraph/fold-bitops.clif@@ -494,3 +494,80 @@ block0(v0: i32, v1: i32):; return v7; }+;; (x & ~y) == ~y --> (x | y) == -1+function %test_eq_band_bnot_fullmask(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = bnot v1+ v3 = band v0, v2+ v4 = icmp eq v3, v2+ return v4+}++; function %test_eq_band_bnot_fullmask(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v5 = bor v1, v0+; v6 = iconst.i32 -1+; v7 = icmp eq v5, v6 ; v6 = -1+; return v7+; }++;; x >=_u (x & y) --> true+function %test_uge_x_band(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = band v0, v1+ v3 = icmp uge v0, v2+ return v3+}++; function %test_uge_x_band(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = band v0, v1+; v3 = icmp uge v0, v2+; return v3+; }++;; (x & y) <=_u x --> true+function %test_ule_band_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = band v0, v1+ v3 = icmp ule v2, v0+ return v3+}++; function %test_ule_band_x(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = band v0, v1+; v3 = icmp ule v2, v0+; return v3+; }++;; (x | y) >=_u x --> true+function %test_uge_bor_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = bor v0, v1+ v3 = icmp uge v2, v0+ return v3+}++; function %test_uge_bor_x(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = bor v0, v1+; v3 = icmp uge v2, v0+; return v3+; }++;; x <=_u (x | y) --> true+function %test_ule_x_bor(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = bor v0, v1+ v3 = icmp ule v0, v2+ return v3+}++; function %test_ule_x_bor(i32, i32) -> i8 fast {+; block0(v0: i32, v1: i32):+; v2 = bor v0, v1+; v3 = icmp ule v0, v2+; return v3+; }+diff --git a/cranelift/filetests/filetests/runtests/arithmetic.clif b/cranelift/filetests/filetests/runtests/arithmetic.clifindex e20d731099..ac85426c04 100644--- a/cranelift/filetests/filetests/runtests/arithmetic.clif+++ b/cranelift/filetests/filetests/runtests/arithmetic.clif@@ -821,4 +821,130 @@ block0(v0: i32, v1: i32):; run: %simplify_icmp_ne_isub_x(0, 0) == 0; run: %simplify_icmp_ne_isub_x(1, 0) == 0-; run: %simplify_icmp_ne_isub_x(1, 1) == 1\ No newline at end of file+; run: %simplify_icmp_ne_isub_x(1, 1) == 1++function %test_eq_isub_same(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = isub v0, v1+ v3 = icmp eq v2, v0+ return v3+}++; run: %test_eq_isub_same(0, 0) == 1+; run: %test_eq_isub_same(1, 0) == 1++function %test_eq_iadd_same_rhs(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = iadd v0, v1+ v3 = icmp eq v2, v1+ return v3+}++; run: %test_eq_iadd_same_rhs(0, 0) == 1+; run: %test_eq_iadd_same_rhs(0, 7) == 1+; run: %test_eq_iadd_same_rhs(1, 0) == 0++function %test_eq_ineg_ineg(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = ineg v0+ v3 = ineg v1+ v4 = icmp eq v2, v3+ return v4+}++; run: %test_eq_ineg_ineg(0, 0) == 1+; run: %test_eq_ineg_ineg(1, 1) == 1+; run: %test_eq_ineg_ineg(1, 2) == 0++function %test_ineg_imul_ineg_lhs(i32, i32) -> i32 fast {+block0(v0: i32, v1: i32):+ v2 = ineg v1+ v3 = imul v2, v0+ v4 = ineg v3+ return v4+}++; run: %test_ineg_imul_ineg_lhs(0, 0) == 0+; run: %test_ineg_imul_ineg_lhs(1, 2) == 2+; run: %test_ineg_imul_ineg_lhs(-3, 4) == -12+; run: %test_ineg_imul_ineg_lhs(7, -5) == -35+; run: %test_ineg_imul_ineg_lhs(0x1234, 0x10) == 0x12340++function %test_ineg_imul_ineg_rhs(i32, i32) -> i32 fast {+block0(v0: i32, v1: i32):+ v2 = ineg v1+ v3 = imul v0, v2+ v4 = ineg v3+ return v4+}++; run: %test_ineg_imul_ineg_rhs(0, 0) == 0+; run: %test_ineg_imul_ineg_rhs(1, 2) == 2++function %test_ireduce_iadd_extends(i8, i8) -> i8 fast {+block0(v0: i8, v1: i8):+ v2 = sextend.i16 v0+ v3 = uextend.i16 v1+ v4 = iadd v2, v3+ v5 = ireduce.i8 v4+ return v5+}++; run: %test_ireduce_iadd_extends(0, 0) == 0+; run: %test_ireduce_iadd_extends(1, 2) == 3++function %test_ireduce_isub_extends(i8, i8) -> i8 fast {+block0(v0: i8, v1: i8):+ v2 = sextend.i16 v0+ v3 = uextend.i16 v1+ v4 = isub v2, v3+ v5 = ireduce.i8 v4+ return v5+}++; run: %test_ireduce_isub_extends(0, 0) == 0+; run: %test_ireduce_isub_extends(5, 2) == 3+; run: %test_ireduce_isub_extends(0, 1) == -1+; run: %test_ireduce_isub_extends(-1, 1) == -2+; run: %test_ireduce_isub_extends(-128, 1) == 127++function %test_sge_smax_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = smax v0, v1+ v3 = icmp sge v2, v0+ return v3+}++; run: %test_sge_smax_x(0, 0) == 1+; run: %test_sge_smax_x(1, 2) == 1++function %test_ule_x_umax(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = umax v0, v1+ v3 = icmp ule v0, v2+ return v3+}++; run: %test_ule_x_umax(0, 0) == 1+; run: %test_ule_x_umax(1, 2) == 1++function %test_sge_x_smin(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = smin v0, v1+ v3 = icmp sge v0, v2+ return v3+}++; run: %test_sge_x_smin(0, 0) == 1+; run: %test_sge_x_smin(1, 2) == 1++function %test_ule_umin_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = umin v0, v1+ v3 = icmp ule v2, v0+ return v3+}++; run: %test_ule_umin_x(0, 0) == 1+; run: %test_ule_umin_x(1, 2) == 1+; run: %test_ule_umin_x(2, 1) == 1\ No newline at end of filediff --git a/cranelift/filetests/filetests/runtests/bitops.clif b/cranelift/filetests/filetests/runtests/bitops.clifindex d7ea0a22e4..d049b29da0 100644--- a/cranelift/filetests/filetests/runtests/bitops.clif+++ b/cranelift/filetests/filetests/runtests/bitops.clif@@ -743,3 +743,54 @@ block0(v0: i32, v1: i32):; run: %test_aarch64_extr(1, 0) == 1; run: %test_aarch64_extr(0, 1) == 1++function %test_eq_band_bnot_fullmask(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = bnot v1+ v3 = band v0, v2+ v4 = icmp eq v3, v2+ return v4+}++; run: %test_eq_band_bnot_fullmask(0, 0) == 0+; run: %test_eq_band_bnot_fullmask(-1, 0) == 1++function %test_uge_x_band(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = band v0, v1+ v3 = icmp uge v0, v2+ return v3+}++; run: %test_uge_x_band(0, 0) == 1+; run: %test_uge_x_band(1, 0) == 1++function %test_ule_band_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = band v0, v1+ v3 = icmp ule v2, v0+ return v3+}++; run: %test_ule_band_x(0, 0) == 1+; run: %test_ule_band_x(1, 0) == 1++function %test_uge_bor_x(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = bor v0, v1+ v3 = icmp uge v2, v0+ return v3+}++; run: %test_uge_bor_x(0, 0) == 1+; run: %test_uge_bor_x(1, 0) == 1++function %test_ule_x_bor(i32, i32) -> i8 fast {+block0(v0: i32, v1: i32):+ v2 = bor v0, v1+ v3 = icmp ule v0, v2+ return v3+}++; run: %test_ule_x_bor(0, 0) == 1+; run: %test_ule_x_bor(1, 0) == 1\ No newline at end of file