[OptimizeInstructions] More canonizations for floating points (#5033)

x - C       ->    x + (-C)
min(C, x)   ->    min(x, C)
max(C, x)   ->    max(x, C)

And remove redundant rules

2 files changed, 107 insertions, 30 deletions

diff --git a/test/lit/passes/optimize-instructions.wast b/test/lit/passes/optimize-instructions.wast
index df8037a72..c1835442d 100644
--- a/test/lit/passes/optimize-instructions.wast
+++ b/test/lit/passes/optimize-instructions.wast
@@ -2016,6 +2016,103 @@
     (drop (i32.add (i32.ctz (local.get $x)) (i32.eqz (local.get $y))))
     (drop (i32.add (i32.eqz (local.get $x)) (i32.ctz (local.get $y))))
   )
+  ;; CHECK:      (func $canonicalize-consts-floats (param $x f32) (param $y f64)
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.add
+  ;; CHECK-NEXT:    (local.get $x)
+  ;; CHECK-NEXT:    (f32.const -1)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.add
+  ;; CHECK-NEXT:    (local.get $y)
+  ;; CHECK-NEXT:    (f64.const -1)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.mul
+  ;; CHECK-NEXT:    (local.get $x)
+  ;; CHECK-NEXT:    (f32.const 3.4000000953674316)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.mul
+  ;; CHECK-NEXT:    (local.get $y)
+  ;; CHECK-NEXT:    (f64.const 3.4)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.min
+  ;; CHECK-NEXT:    (local.get $x)
+  ;; CHECK-NEXT:    (f32.const 3.4000000953674316)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.min
+  ;; CHECK-NEXT:    (local.get $y)
+  ;; CHECK-NEXT:    (f64.const 3.4)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.max
+  ;; CHECK-NEXT:    (local.get $x)
+  ;; CHECK-NEXT:    (f32.const 3.4000000953674316)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.max
+  ;; CHECK-NEXT:    (local.get $y)
+  ;; CHECK-NEXT:    (f64.const 3.4)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.const nan:0x400000)
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.const nan:0x8000000000000)
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.const nan:0x400000)
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.const nan:0x8000000000000)
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f32.copysign
+  ;; CHECK-NEXT:    (f32.const 1)
+  ;; CHECK-NEXT:    (local.get $x)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT:  (drop
+  ;; CHECK-NEXT:   (f64.copysign
+  ;; CHECK-NEXT:    (f64.const 1)
+  ;; CHECK-NEXT:    (local.get $y)
+  ;; CHECK-NEXT:   )
+  ;; CHECK-NEXT:  )
+  ;; CHECK-NEXT: )
+  (func $canonicalize-consts-floats (param $x f32) (param $y f64)
+    (drop (f32.sub (local.get $x) (f32.const 1.0)))
+    (drop (f64.sub (local.get $y) (f64.const 1.0)))
+
+    (drop (f32.mul (f32.const 3.4) (local.get $x)))
+    (drop (f64.mul (f64.const 3.4) (local.get $y)))
+
+    (drop (f32.min (f32.const 3.4) (local.get $x)))
+    (drop (f64.min (f64.const 3.4) (local.get $y)))
+
+    (drop (f32.max (f32.const 3.4) (local.get $x)))
+    (drop (f64.max (f64.const 3.4) (local.get $y)))
+
+    (drop (f32.min (f32.const nan) (local.get $x)))
+    (drop (f64.min (f64.const nan) (local.get $y)))
+
+    (drop (f32.max (f32.const nan) (local.get $x)))
+    (drop (f64.max (f64.const nan) (local.get $y)))
+
+    ;; skips
+    (drop (f32.copysign (f32.const 1.0) (local.get $x)))
+    (drop (f64.copysign (f64.const 1.0) (local.get $y)))
+  )
   ;; CHECK:      (func $ne0 (result i32)
   ;; CHECK-NEXT:  (if
   ;; CHECK-NEXT:   (call $ne0)
@@ -10303,15 +10400,15 @@
   )
   ;; CHECK:      (func $const-float-zero (param $fx f32) (param $fy f64)
   ;; CHECK-NEXT:  (drop
-  ;; CHECK-NEXT:   (f32.sub
+  ;; CHECK-NEXT:   (f32.add
   ;; CHECK-NEXT:    (local.get $fx)
-  ;; CHECK-NEXT:    (f32.const 0)
+  ;; CHECK-NEXT:    (f32.const -0)
   ;; CHECK-NEXT:   )
   ;; CHECK-NEXT:  )
   ;; CHECK-NEXT:  (drop
-  ;; CHECK-NEXT:   (f64.sub
+  ;; CHECK-NEXT:   (f64.add
   ;; CHECK-NEXT:    (local.get $fy)
-  ;; CHECK-NEXT:    (f64.const 0)
+  ;; CHECK-NEXT:    (f64.const -0)
   ;; CHECK-NEXT:   )
   ;; CHECK-NEXT:  )
   ;; CHECK-NEXT:  (drop
@@ -10363,9 +10460,9 @@
   ;; CHECK-NEXT:   )
   ;; CHECK-NEXT:  )
   ;; CHECK-NEXT:  (drop
-  ;; CHECK-NEXT:   (f32.sub
+  ;; CHECK-NEXT:   (f32.add
   ;; CHECK-NEXT:    (f32.const -nan:0x34546d)
-  ;; CHECK-NEXT:    (f32.const 0)
+  ;; CHECK-NEXT:    (f32.const -0)
   ;; CHECK-NEXT:   )
   ;; CHECK-NEXT:  )
   ;; CHECK-NEXT: )
diff --git a/test/passes/optimize-instructions_fuzz-exec.txt b/test/passes/optimize-instructions_fuzz-exec.txt
index 16b17cf26..33a3945dc 100644
--- a/test/passes/optimize-instructions_fuzz-exec.txt
+++ b/test/passes/optimize-instructions_fuzz-exec.txt
@@ -73,20 +73,10 @@
    )
   )
   (call $logf32
-   (f32.min
-    (f32.const -nan:0x7fff82)
-    (f32.neg
-     (f32.const -nan:0x7ff622)
-    )
-   )
+   (f32.const nan:0x400000)
   )
   (call $logf32
-   (f32.max
-    (f32.const -nan:0x7fff82)
-    (f32.neg
-     (f32.const -nan:0x7ff622)
-    )
-   )
+   (f32.const nan:0x400000)
   )
  )
  (func $1
@@ -126,20 +116,10 @@
    )
   )
   (call $logf64
-   (f64.min
-    (f64.const -nan:0xfffffffffff82)
-    (f64.neg
-     (f64.const -nan:0xfffffffffa622)
-    )
-   )
+   (f64.const nan:0x8000000000000)
   )
   (call $logf64
-   (f64.max
-    (f64.const -nan:0xfffffffffff82)
-    (f64.neg
-     (f64.const -nan:0xfffffffffa622)
-    )
-   )
+   (f64.const nan:0x8000000000000)
   )
  )
  (func $2