Unary and binary duplicate expression elimination (#3047)

Simplifies patterns in which an expression is applied twice to its operands.

`abs(abs(x))` -> `abs(x)`
`ceil(ceil(x))` -> `ceil(x)`
`floor(floor(x))` -> `floor(x)`
`trunc(trunc(x))` -> `trunc(x)`
`nearest(nearest(x))` -> `nearest(x)`
`eqz(eqz(bool(x)))` -> `bool(x)`
`sext(sext(x))` -> `sext(x)`
`neg(neg(x))` -> `x`

`y - (y - x)` -> `x`
`(x ^ y) ^ y` -> `x`
`(x | y) | y` -> `x | y`
`(x & y) & y` -> `x & y`
`(x % y) % y` -> `x % y`

2 files changed, 490 insertions, 4 deletions

diff --git a/test/passes/optimize-instructions_all-features.txt b/test/passes/optimize-instructions_all-features.txt
index 33e137b9d..dda7a47c0 100644
--- a/test/passes/optimize-instructions_all-features.txt
+++ b/test/passes/optimize-instructions_all-features.txt
@@ -12,6 +12,7 @@
  (type $i32_i64_f32_f64_=>_none (func (param i32 i64 f32 f64)))
  (type $none_=>_anyref (func (result anyref)))
  (type $i32_i32_i32_=>_none (func (param i32 i32 i32)))
+ (type $i32_i32_i32_f64_=>_none (func (param i32 i32 i32 f64)))
  (type $i32_i32_f64_f64_=>_none (func (param i32 i32 f64 f64)))
  (type $i32_i64_f64_i32_=>_none (func (param i32 i64 f64 i32)))
  (type $f32_f64_=>_none (func (param f32 f64)))
@@ -1301,10 +1302,7 @@
   (drop
    (i32.and
     (local.get $0)
-    (i32.and
-     (local.get $0)
-     (i32.const 11)
-    )
+    (i32.const 11)
    )
   )
   (drop
@@ -3786,6 +3784,253 @@
    )
   )
  )
+ (func $duplicate-elimination (param $x i32) (param $y i32) (param $z i32) (param $w f64)
+  (drop
+   (f64.abs
+    (local.get $w)
+   )
+  )
+  (drop
+   (f64.ceil
+    (local.get $w)
+   )
+  )
+  (drop
+   (f64.floor
+    (local.get $w)
+   )
+  )
+  (drop
+   (f64.trunc
+    (local.get $w)
+   )
+  )
+  (drop
+   (f64.nearest
+    (local.get $w)
+   )
+  )
+  (drop
+   (f64.nearest
+    (f64.trunc
+     (local.get $w)
+    )
+   )
+  )
+  (drop
+   (f64.trunc
+    (f64.nearest
+     (local.get $w)
+    )
+   )
+  )
+  (drop
+   (local.get $w)
+  )
+  (drop
+   (f64.neg
+    (local.get $w)
+   )
+  )
+  (drop
+   (local.get $w)
+  )
+  (drop
+   (i32.eqz
+    (i32.eqz
+     (local.get $x)
+    )
+   )
+  )
+  (drop
+   (i32.eqz
+    (local.get $x)
+   )
+  )
+  (drop
+   (i64.eqz
+    (i64.const 1)
+   )
+  )
+  (drop
+   (i32.ne
+    (local.get $x)
+    (i32.const 2)
+   )
+  )
+  (drop
+   (i32.extend8_s
+    (local.get $x)
+   )
+  )
+  (drop
+   (i32.extend16_s
+    (local.get $x)
+   )
+  )
+  (drop
+   (i32.and
+    (local.get $x)
+    (i32.const 1)
+   )
+  )
+  (drop
+   (i32.rem_s
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.rem_u
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (local.get $y)
+  )
+  (drop
+   (local.get $y)
+  )
+  (drop
+   (i32.sub
+    (local.get $y)
+    (i32.sub
+     (local.get $x)
+     (local.get $y)
+    )
+   )
+  )
+  (drop
+   (local.get $y)
+  )
+  (drop
+   (local.get $y)
+  )
+  (drop
+   (local.get $y)
+  )
+  (drop
+   (local.get $y)
+  )
+  (drop
+   (local.get $x)
+  )
+  (drop
+   (i32.and
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.and
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.and
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.and
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.or
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.or
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.or
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.or
+    (local.get $x)
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.or
+    (local.get $z)
+    (i32.or
+     (local.get $x)
+     (local.get $y)
+    )
+   )
+  )
+  (drop
+   (i32.or
+    (local.get $y)
+    (i32.or
+     (local.get $x)
+     (local.get $z)
+    )
+   )
+  )
+  (drop
+   (i32.or
+    (call $ne0)
+    (local.get $x)
+   )
+  )
+  (drop
+   (i32.or
+    (i32.or
+     (call $ne0)
+     (local.get $x)
+    )
+    (call $ne0)
+   )
+  )
+  (drop
+   (i32.or
+    (call $ne0)
+    (local.get $x)
+   )
+  )
+  (drop
+   (i32.or
+    (call $ne0)
+    (i32.or
+     (call $ne0)
+     (local.get $x)
+    )
+   )
+  )
+  (drop
+   (i32.rem_s
+    (i32.rem_s
+     (local.get $y)
+     (local.get $x)
+    )
+    (local.get $y)
+   )
+  )
+  (drop
+   (i32.rem_u
+    (local.get $y)
+    (i32.rem_u
+     (local.get $x)
+     (local.get $y)
+    )
+   )
+  )
+ )
  (func $optimize-bulk-memory-copy (param $dst i32) (param $src i32) (param $sz i32)
   (memory.copy
    (local.get $dst)
diff --git a/test/passes/optimize-instructions_all-features.wast b/test/passes/optimize-instructions_all-features.wast
index c67270a4f..7dc8cffbf 100644
--- a/test/passes/optimize-instructions_all-features.wast
+++ b/test/passes/optimize-instructions_all-features.wast
@@ -4289,6 +4289,247 @@
       )
     ))
   )
+  (func $duplicate-elimination (param $x i32) (param $y i32) (param $z i32) (param $w f64)
+    ;; unary
+    (drop (f64.abs (f64.abs (local.get $w))))
+    (drop (f64.ceil (f64.ceil (local.get $w))))
+    (drop (f64.floor (f64.floor (local.get $w))))
+    (drop (f64.trunc (f64.trunc (local.get $w))))
+    (drop (f64.nearest (f64.nearest (local.get $w))))
+
+    (drop (f64.nearest (f64.trunc (local.get $w)))) ;; skip
+    (drop (f64.trunc (f64.nearest (local.get $w)))) ;; skip
+
+    (drop (f64.neg (f64.neg (local.get $w))))
+    (drop (f64.neg (f64.neg (f64.neg (local.get $w)))))
+    (drop (f64.neg (f64.neg (f64.neg (f64.neg (local.get $w))))))
+
+    (drop (i32.eqz (i32.eqz (local.get $x)))) ;; skip
+    (drop (i32.eqz (i32.eqz (i32.eqz (local.get $x)))))
+    (drop (i32.eqz (i32.eqz (i64.eqz (i64.const 1)))))
+    (drop (i32.eqz (i32.eqz (i32.ne (local.get $x) (i32.const 2)))))
+
+    (drop (i32.extend8_s (i32.extend8_s (local.get $x))))
+    (drop (i32.extend16_s (i32.extend16_s (local.get $x))))
+    (drop (i32.eqz
+      (i32.eqz
+        (i32.and
+          (local.get $x)
+          (i32.const 1)
+        )
+      )
+    ))
+
+    ;; binary
+    ;; ((signed)x % y) % y
+    (drop (i32.rem_s
+      (i32.rem_s
+        (local.get $x)
+        (local.get $y)
+      )
+      (local.get $y)
+    ))
+    ;; ((unsigned)x % y) % y
+    (drop (i32.rem_u
+      (i32.rem_u
+        (local.get $x)
+        (local.get $y)
+      )
+      (local.get $y)
+    ))
+    ;; 0 - (0 - y)
+    (drop (i32.sub
+      (i32.const 0)
+      (i32.sub
+        (i32.const 0)
+        (local.get $y)
+      )
+    ))
+    ;; x - (x - y)
+    (drop (i32.sub
+      (local.get $x)
+      (i32.sub
+        (local.get $x)
+        (local.get $y)
+      )
+    ))
+    ;; y - (x - y)   -   skip
+    (drop (i32.sub
+      (local.get $y)
+      (i32.sub
+        (local.get $x)
+        (local.get $y)
+      )
+    ))
+    ;; x ^ (x ^ y)
+    (drop (i32.xor
+      (local.get $x)
+      (i32.xor
+        (local.get $x)
+        (local.get $y)
+      )
+    ))
+    ;; x ^ (y ^ x)
+    (drop (i32.xor
+      (local.get $x)
+      (i32.xor
+        (local.get $y)
+        (local.get $x)
+      )
+    ))
+    ;; (x ^ y) ^ x
+    (drop (i32.xor
+      (i32.xor
+        (local.get $x)
+        (local.get $y)
+      )
+      (local.get $x)
+    ))
+    ;; (y ^ x) ^ x
+    (drop (i32.xor
+      (i32.xor
+        (local.get $y)
+        (local.get $x)
+      )
+      (local.get $x)
+    ))
+    ;; x ^ (x ^ x)
+    (drop (i32.xor
+      (local.get $x)
+      (i32.xor
+        (local.get $x)
+        (local.get $x)
+      )
+    ))
+    ;; x & (x & y)
+    (drop (i32.and
+      (local.get $x)
+      (i32.and
+        (local.get $x)
+        (local.get $y)
+      )
+    ))
+    ;; x & (y & x)
+    (drop (i32.and
+      (local.get $x)
+      (i32.and
+        (local.get $y)
+        (local.get $x)
+      )
+    ))
+    ;; (x & y) & x
+    (drop (i32.and
+      (i32.and
+        (local.get $x)
+        (local.get $y)
+      )
+      (local.get $x)
+    ))
+    ;; (y & x) & x
+    (drop (i32.and
+      (i32.and
+        (local.get $y)
+        (local.get $x)
+      )
+      (local.get $x)
+    ))
+    ;; x | (x | y)
+    (drop (i32.or
+      (local.get $x)
+      (i32.or
+        (local.get $x)
+        (local.get $y)
+      )
+    ))
+    ;; x | (y | x)
+    (drop (i32.or
+      (local.get $x)
+      (i32.or
+        (local.get $y)
+        (local.get $x)
+      )
+    ))
+    ;; (x | y) | x
+    (drop (i32.or
+      (i32.or
+        (local.get $x)
+        (local.get $y)
+      )
+      (local.get $x)
+    ))
+    ;; (y | x) | x
+    (drop (i32.or
+      (i32.or
+        (local.get $y)
+        (local.get $x)
+      )
+      (local.get $x)
+    ))
+    ;; (y | x) | z   -   skip
+    (drop (i32.or
+      (i32.or
+        (local.get $y)
+        (local.get $x)
+      )
+      (local.get $z)
+    ))
+    ;; (z | x) | y   -   skip
+    (drop (i32.or
+      (i32.or
+        (local.get $z)
+        (local.get $x)
+      )
+      (local.get $y)
+    ))
+    ;; (SE() | x) | x
+    (drop (i32.or
+      (i32.or
+        (call $ne0) ;; side effect
+        (local.get $x)
+      )
+      (local.get $x)
+    ))
+    ;; (x | SE()) | SE()   -   skip
+    (drop (i32.or
+      (i32.or
+        (local.get $x)
+        (call $ne0) ;; side effect
+      )
+      (call $ne0) ;; side effect
+    ))
+    ;; x | (SE() | x)
+    (drop (i32.or
+      (local.get $x)
+      (i32.or
+        (local.get $x)
+        (call $ne0) ;; side effect
+      )
+    ))
+    ;; SE() | (x | SE())   -   skip
+    (drop (i32.or
+      (call $ne0) ;; side effect
+      (i32.or
+        (call $ne0) ;; side effect
+        (local.get $x)
+      )
+    ))
+    ;; (y % x) % y   -   skip
+    (drop (i32.rem_s
+      (i32.rem_s
+        (local.get $y)
+        (local.get $x)
+      )
+      (local.get $y)
+    ))
+    ;; y % (x % y)   -   skip
+    (drop (i32.rem_u
+      (local.get $y)
+      (i32.rem_u
+        (local.get $x)
+        (local.get $y)
+      )
+    ))
+  )
   (func $optimize-bulk-memory-copy (param $dst i32) (param $src i32) (param $sz i32)
     (memory.copy  ;; skip
       (local.get $dst)