cmd/compile: canonicalize comparison argument order

Ensure that any comparison between two values has the same argument
order. This helps ensure that they can be eliminated during the
lowered CSE pass which will be particularly important if we eliminate
the Greater and Geq ops (see #37316).

Example:

  CMP R0, R1
  BLT L1
  CMP R1, R0 // different order, cannot eliminate
  BEQ L2

  CMP R0, R1
  BLT L1
  CMP R0, R1 // same order, can eliminate
  BEQ L2

This does have some drawbacks. Notably comparisons might 'flip'
direction in the assembly output after even small changes to the
code or compiler. It should help make optimizations more reliable
however.

compilecmp master -> HEAD
master (218f4572f5): text/template: make reflect.Value indirections more robust
HEAD (f1661fef3e): cmd/compile: canonicalize comparison argument order
platform: linux/amd64

file      before    after     Δ       %
api       6063927   6068023   +4096   +0.068%
asm       5191757   5183565   -8192   -0.158%
cgo       4893518   4901710   +8192   +0.167%
cover     5330345   5326249   -4096   -0.077%
fix       3417778   3421874   +4096   +0.120%
pprof     14889456  14885360  -4096   -0.028%
test2json 2848138   2844042   -4096   -0.144%
trace     11746239  11733951  -12288  -0.105%
total     132739173 132722789 -16384  -0.012%

Change-Id: I11736b3fe2a4553f6fc65018f475e88217fa22f9
Reviewed-on: https://go-review.googlesource.com/c/go/+/220425
Run-TryBot: Michael Munday <mike.munday@ibm.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>

2 files changed, 7 insertions, 7 deletions

diff --git a/test/codegen/compare_and_branch.go b/test/codegen/compare_and_branch.go
index 33d8d7bd52..23e7810b31 100644
--- a/test/codegen/compare_and_branch.go
+++ b/test/codegen/compare_and_branch.go
@@ -11,7 +11,7 @@ func dummy() {}
 
 // Signed 64-bit compare-and-branch.
 func si64(x, y chan int64) {
-	// s390x:"CGRJ\t[$]4, R[0-9]+, R[0-9]+, "
+	// s390x:"CGRJ\t[$](2|4), R[0-9]+, R[0-9]+, "
 	for <-x < <-y {
 		dummy()
 	}
@@ -47,7 +47,7 @@ func si64x8() {
 
 // Unsigned 64-bit compare-and-branch.
 func ui64(x, y chan uint64) {
-	// s390x:"CLGRJ\t[$]2, R[0-9]+, R[0-9]+, "
+	// s390x:"CLGRJ\t[$](2|4), R[0-9]+, R[0-9]+, "
 	for <-x > <-y {
 		dummy()
 	}
@@ -83,7 +83,7 @@ func ui64x8() {
 
 // Signed 32-bit compare-and-branch.
 func si32(x, y chan int32) {
-	// s390x:"CRJ\t[$]4, R[0-9]+, R[0-9]+, "
+	// s390x:"CRJ\t[$](2|4), R[0-9]+, R[0-9]+, "
 	for <-x < <-y {
 		dummy()
 	}
@@ -119,7 +119,7 @@ func si32x8() {
 
 // Unsigned 32-bit compare-and-branch.
 func ui32(x, y chan uint32) {
-	// s390x:"CLRJ\t[$]2, R[0-9]+, R[0-9]+, "
+	// s390x:"CLRJ\t[$](2|4), R[0-9]+, R[0-9]+, "
 	for <-x > <-y {
 		dummy()
 	}
diff --git a/test/codegen/condmove.go b/test/codegen/condmove.go
index bd3fe59427..00118d1b63 100644
--- a/test/codegen/condmove.go
+++ b/test/codegen/condmove.go
@@ -32,7 +32,7 @@ func cmovuintptr(x, y uintptr) uintptr {
 		x = -y
 	}
 	// amd64:"CMOVQCS"
-	// arm64:"CSEL\tLO"
+	// arm64:"CSEL\t(LO|HI)"
 	// wasm:"Select"
 	return x
 }
@@ -42,7 +42,7 @@ func cmov32bit(x, y uint32) uint32 {
 		x = -y
 	}
 	// amd64:"CMOVLCS"
-	// arm64:"CSEL\tLO"
+	// arm64:"CSEL\t(LO|HI)"
 	// wasm:"Select"
 	return x
 }
@@ -52,7 +52,7 @@ func cmov16bit(x, y uint16) uint16 {
 		x = -y
 	}
 	// amd64:"CMOVWCS"
-	// arm64:"CSEL\tLO"
+	// arm64:"CSEL\t(LO|HI)"
 	// wasm:"Select"
 	return x
 }