all: clean up addition of constants in riscv64 assembly

Use ADD with constants, instead of ADDI. Also use SUB with a positive constant
rather than ADD with a negative constant. The resulting assembly is still the
same.

Change-Id: Ife10bf5ae4122e525f0e7d41b5e463e748236a9c
Reviewed-on: https://go-review.googlesource.com/c/go/+/540136
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: M Zhuo <mzh@golangcn.org>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Mark Ryan <markdryan@rivosinc.com>
Reviewed-by: Heschi Kreinick <heschi@google.com>
Run-TryBot: Joel Sing <joel@sing.id.au>

src/crypto/internal/bigmod/nat_riscv64.s

@@ -80,10 +80,10 @@ loop:
 	MOV	X16, 2*8(X5)	// z[2]
 	MOV	X19, 3*8(X5)	// z[3]
 
-	ADDI	$32, X5
-	ADDI	$32, X7
+	ADD	$32, X5
+	ADD	$32, X7
 
-	ADDI	$-4, X30
+	SUB	$4, X30
 	BNEZ	X30, loop
 
 done:

src/internal/bytealg/compare_riscv64.s

@@ -53,7 +53,7 @@ use_a_len:
 	ADD	$8, X7, X7
 	SUB	X7, X5, X5
 align:
-	ADD	$-1, X7
+	SUB	$1, X7
 	MOVBU	0(X10), X8
 	MOVBU	0(X12), X9
 	BNE	X8, X9, cmp
@@ -79,7 +79,7 @@ compare32:
 	BNE	X17, X18, cmp8b
 	ADD	$32, X10
 	ADD	$32, X12
-	ADD	$-32, X5
+	SUB	$32, X5
 	BGE	X5, X6, compare32
 	BEQZ	X5, cmp_len
 
@@ -95,7 +95,7 @@ compare16:
 	BNE	X17, X18, cmp8b
 	ADD	$16, X10
 	ADD	$16, X12
-	ADD	$-16, X5
+	SUB	$16, X5
 	BEQZ	X5, cmp_len
 
 check8_unaligned:
@@ -128,7 +128,7 @@ compare8_unaligned:
 	BNE	X29, X30, cmp1h
 	ADD	$8, X10
 	ADD	$8, X12
-	ADD	$-8, X5
+	SUB	$8, X5
 	BGE	X5, X6, compare8_unaligned
 	BEQZ	X5, cmp_len
 
@@ -150,7 +150,7 @@ compare4_unaligned:
 	BNE	X19, X20, cmp1d
 	ADD	$4, X10
 	ADD	$4, X12
-	ADD	$-4, X5
+	SUB	$4, X5
 	BGE	X5, X6, compare4_unaligned
 
 compare1:
@@ -160,7 +160,7 @@ compare1:
 	BNE	X8, X9, cmp
 	ADD	$1, X10
 	ADD	$1, X12
-	ADD	$-1, X5
+	SUB	$1, X5
 	JMP	compare1
 
 	// Compare 8 bytes of memory in X15/X16 that are known to differ.

src/internal/bytealg/equal_riscv64.s

@@ -41,7 +41,7 @@ TEXT memequal<>(SB),NOSPLIT|NOFRAME,$0
 	ADD	$8, X9, X9
 	SUB	X9, X12, X12
 align:
-	ADD	$-1, X9
+	SUB	$1, X9
 	MOVBU	0(X10), X19
 	MOVBU	0(X11), X20
 	BNE	X19, X20, not_eq
@@ -67,7 +67,7 @@ loop32:
 	BNE	X16, X17, not_eq
 	ADD	$32, X10
 	ADD	$32, X11
-	ADD	$-32, X12
+	SUB	$32, X12
 	BGE	X12, X9, loop32
 	BEQZ	X12, eq
 
@@ -83,7 +83,7 @@ loop16:
 	BNE	X21, X22, not_eq
 	ADD	$16, X10
 	ADD	$16, X11
-	ADD	$-16, X12
+	SUB	$16, X12
 	BGE	X12, X23, loop16
 	BEQZ	X12, eq
 
@@ -105,7 +105,7 @@ loop4:
 	BNE	X16, X17, not_eq
 	ADD	$4, X10
 	ADD	$4, X11
-	ADD	$-4, X12
+	SUB	$4, X12
 	BGE	X12, X23, loop4
 
 loop1:
@@ -115,7 +115,7 @@ loop1:
 	BNE	X19, X20, not_eq
 	ADD	$1, X10
 	ADD	$1, X11
-	ADD	$-1, X12
+	SUB	$1, X12
 	JMP	loop1
 
 not_eq:

src/internal/bytealg/indexbyte_riscv64.s

@@ -13,7 +13,7 @@ TEXT ·IndexByte<ABIInternal>(SB),NOSPLIT,$0-40
 	AND	$0xff, X13
 	MOV	X10, X12		// store base for later
 	ADD	X10, X11		// end
-	ADD	$-1, X10
+	SUB	$1, X10
 
 loop:
 	ADD	$1, X10
@@ -35,7 +35,7 @@ TEXT ·IndexByteString<ABIInternal>(SB),NOSPLIT,$0-32
 	AND	$0xff, X12
 	MOV	X10, X13		// store base for later
 	ADD	X10, X11		// end
-	ADD	$-1, X10
+	SUB	$1, X10
 
 loop:
 	ADD	$1, X10

src/runtime/asm_riscv64.s

@@ -9,7 +9,7 @@
 // func rt0_go()
 TEXT runtime·rt0_go(SB),NOSPLIT|TOPFRAME,$0
 	// X2 = stack; A0 = argc; A1 = argv
-	ADD	$-24, X2
+	SUB	$24, X2
 	MOV	A0, 8(X2)	// argc
 	MOV	A1, 16(X2)	// argv
 
@@ -57,7 +57,7 @@ nocgo:
 
 	// create a new goroutine to start program
 	MOV	$runtime·mainPC(SB), T0		// entry
-	ADD	$-16, X2
+	SUB	$16, X2
 	MOV	T0, 8(X2)
 	MOV	ZERO, 0(X2)
 	CALL	runtime·newproc(SB)
@@ -161,7 +161,7 @@ TEXT runtime·switchToCrashStack0<ABIInternal>(SB), NOSPLIT, $0-8
 
 	// switch to crashstack
 	MOV	(g_stack+stack_hi)(g), X11
-	ADD	$(-4*8), X11
+	SUB	$(4*8), X11
 	MOV	X11, X2
 
 	// call target function
@@ -219,7 +219,7 @@ TEXT runtime·morestack(SB),NOSPLIT|NOFRAME,$0-0
 	MOV	(g_sched+gobuf_sp)(g), X2
 	// Create a stack frame on g0 to call newstack.
 	MOV	ZERO, -8(X2)	// Zero saved LR in frame
-	ADD	$-8, X2
+	SUB	$8, X2
 	CALL	runtime·newstack(SB)
 
 	// Not reached, but make sure the return PC from the call to newstack
@@ -304,7 +304,7 @@ TEXT runtime·mcall<ABIInternal>(SB), NOSPLIT|NOFRAME, $0-8
 	MOV	0(CTXT), T1			// code pointer
 	MOV	(g_sched+gobuf_sp)(g), X2	// sp = m->g0->sched.sp
 	// we don't need special macro for regabi since arg0(X10) = g
-	ADD	$-16, X2
+	SUB	$16, X2
 	MOV	X10, 8(X2)			// setup g
 	MOV	ZERO, 0(X2)			// clear return address
 	JALR	RA, T1
@@ -366,7 +366,7 @@ TEXT ·asmcgocall(SB),NOSPLIT,$0-20
 	// Now on a scheduling stack (a pthread-created stack).
 g0:
 	// Save room for two of our pointers.
-	ADD	$-16, X2
+	SUB	$16, X2
 	MOV	X9, 0(X2)	// save old g on stack
 	MOV	(g_stack+stack_hi)(X9), X9
 	SUB	X8, X9, X8

src/runtime/memclr_riscv64.s

@@ -23,7 +23,7 @@ TEXT runtime·memclrNoHeapPointers<ABIInternal>(SB),NOSPLIT,$0-16
 	SUB	X5, X9, X5
 	SUB	X5, X11, X11
 align:
-	ADD	$-1, X5
+	SUB	$1, X5
 	MOVB	ZERO, 0(X10)
 	ADD	$1, X10
 	BNEZ	X5, align
@@ -47,7 +47,7 @@ loop64:
 	MOV	ZERO, 48(X10)
 	MOV	ZERO, 56(X10)
 	ADD	$64, X10
-	ADD	$-64, X11
+	SUB	$64, X11
 	BGE	X11, X9, loop64
 	BEQZ	X11, done
 
@@ -60,7 +60,7 @@ zero32:
 	MOV	ZERO, 16(X10)
 	MOV	ZERO, 24(X10)
 	ADD	$32, X10
-	ADD	$-32, X11
+	SUB	$32, X11
 	BEQZ	X11, done
 
 check16:
@@ -70,7 +70,7 @@ zero16:
 	MOV	ZERO, 0(X10)
 	MOV	ZERO, 8(X10)
 	ADD	$16, X10
-	ADD	$-16, X11
+	SUB	$16, X11
 	BEQZ	X11, done
 
 check8:
@@ -79,7 +79,7 @@ check8:
 zero8:
 	MOV	ZERO, 0(X10)
 	ADD	$8, X10
-	ADD	$-8, X11
+	SUB	$8, X11
 	BEQZ	X11, done
 
 check4:
@@ -91,13 +91,13 @@ zero4:
 	MOVB	ZERO, 2(X10)
 	MOVB	ZERO, 3(X10)
 	ADD	$4, X10
-	ADD	$-4, X11
+	SUB	$4, X11
 
 loop1:
 	BEQZ	X11, done
 	MOVB	ZERO, 0(X10)
 	ADD	$1, X10
-	ADD	$-1, X11
+	SUB	$1, X11
 	JMP	loop1
 
 done: