runtime: fix silly frame sizes on arm and arm64

"-8" is not a sensible frame size on arm and we're about to start
rejecting it. Replace it with -4.

Likewise, "-4" is not a sensible frame size on arm64 and we're about
to start rejecting it. Replace it with -8.

Finally, clean up some places we're weirdly inconsistent about using 0
versus -8.

Change-Id: If85e229993d5f7f1f0cfa9852b4e294d053bd784
Reviewed-on: https://go-review.googlesource.com/92038
Run-TryBot: Austin Clements <austin@google.com>
TryBot-Result: Gobot Gobot <gobot@golang.org>
Reviewed-by: Cherry Zhang <cherryyz@google.com>

src/runtime/asm_arm64.s

@@ -300,7 +300,7 @@ TEXT runtime·morestack(SB),NOSPLIT,$-8-0
 	// is still in this function, and not the beginning of the next.
 	UNDEF
 
-TEXT runtime·morestack_noctxt(SB),NOSPLIT,$-4-0
+TEXT runtime·morestack_noctxt(SB),NOSPLIT,$-8-0
 	MOVW	$0, R26
 	B runtime·morestack(SB)
 
@@ -745,7 +745,7 @@ eq:
 	MOVB	R3, ret+16(FP)
 	RET
 
-TEXT runtime·cmpstring(SB),NOSPLIT,$-4-40
+TEXT runtime·cmpstring(SB),NOSPLIT,$-8-40
 	MOVD	s1_base+0(FP), R2
 	MOVD	s1_len+8(FP), R0
 	MOVD	s2_base+16(FP), R3
@@ -753,7 +753,7 @@ TEXT runtime·cmpstring(SB),NOSPLIT,$-4-40
 	ADD	$40, RSP, R7
 	B	runtime·cmpbody<>(SB)
 
-TEXT bytes·Compare(SB),NOSPLIT,$-4-56
+TEXT bytes·Compare(SB),NOSPLIT,$-8-56
 	MOVD	s1+0(FP), R2
 	MOVD	s1+8(FP), R0
 	MOVD	s2+24(FP), R3
@@ -770,7 +770,7 @@ TEXT bytes·Compare(SB),NOSPLIT,$-4-56
 //
 // On exit:
 // R4, R5, and R6 are clobbered
-TEXT runtime·cmpbody<>(SB),NOSPLIT,$-4-0
+TEXT runtime·cmpbody<>(SB),NOSPLIT,$-8-0
 	CMP	R2, R3
 	BEQ	samebytes // same starting pointers; compare lengths
 	CMP	R0, R1

src/runtime/internal/atomic/asm_arm64.s

@@ -29,10 +29,10 @@ ok:
 TEXT runtime∕internal∕atomic·Casuintptr(SB), NOSPLIT, $0-25
 	B	runtime∕internal∕atomic·Cas64(SB)
 
-TEXT runtime∕internal∕atomic·Loaduintptr(SB), NOSPLIT, $-8-16
+TEXT runtime∕internal∕atomic·Loaduintptr(SB), NOSPLIT, $0-16
 	B	runtime∕internal∕atomic·Load64(SB)
 
-TEXT runtime∕internal∕atomic·Loaduint(SB), NOSPLIT, $-8-16
+TEXT runtime∕internal∕atomic·Loaduint(SB), NOSPLIT, $0-16
 	B	runtime∕internal∕atomic·Load64(SB)
 
 TEXT runtime∕internal∕atomic·Storeuintptr(SB), NOSPLIT, $0-16

src/runtime/internal/atomic/atomic_arm64.s

@@ -5,21 +5,21 @@
 #include "textflag.h"
 
 // uint32 runtime∕internal∕atomic·Load(uint32 volatile* addr)
-TEXT ·Load(SB),NOSPLIT,$-8-12
+TEXT ·Load(SB),NOSPLIT,$0-12
 	MOVD	ptr+0(FP), R0
 	LDARW	(R0), R0
 	MOVW	R0, ret+8(FP)
 	RET
 
 // uint64 runtime∕internal∕atomic·Load64(uint64 volatile* addr)
-TEXT ·Load64(SB),NOSPLIT,$-8-16
+TEXT ·Load64(SB),NOSPLIT,$0-16
 	MOVD	ptr+0(FP), R0
 	LDAR	(R0), R0
 	MOVD	R0, ret+8(FP)
 	RET
 
 // void *runtime∕internal∕atomic·Loadp(void *volatile *addr)
-TEXT ·Loadp(SB),NOSPLIT,$-8-16
+TEXT ·Loadp(SB),NOSPLIT,$0-16
 	MOVD	ptr+0(FP), R0
 	LDAR	(R0), R0
 	MOVD	R0, ret+8(FP)

src/runtime/sys_freebsd_arm.s

@@ -74,7 +74,7 @@ TEXT runtime·thr_start(SB),NOSPLIT,$0
 	RET
 
 // Exit the entire program (like C exit)
-TEXT runtime·exit(SB),NOSPLIT,$-8
+TEXT runtime·exit(SB),NOSPLIT,$-4
 	MOVW code+0(FP), R0	// arg 1 exit status
 	MOVW $SYS_exit, R7
 	SWI $0
@@ -99,7 +99,7 @@ storeloop:
 	MOVW.CS	R8, (R8)
 	JMP	0(PC)
 
-TEXT runtime·open(SB),NOSPLIT,$-8
+TEXT runtime·open(SB),NOSPLIT,$-4
 	MOVW name+0(FP), R0	// arg 1 name
 	MOVW mode+4(FP), R1	// arg 2 mode
 	MOVW perm+8(FP), R2	// arg 3 perm
@@ -109,7 +109,7 @@ TEXT runtime·open(SB),NOSPLIT,$-8
 	MOVW	R0, ret+12(FP)
 	RET
 
-TEXT runtime·read(SB),NOSPLIT,$-8
+TEXT runtime·read(SB),NOSPLIT,$-4
 	MOVW fd+0(FP), R0	// arg 1 fd
 	MOVW p+4(FP), R1	// arg 2 buf
 	MOVW n+8(FP), R2	// arg 3 count
@@ -119,7 +119,7 @@ TEXT runtime·read(SB),NOSPLIT,$-8
 	MOVW	R0, ret+12(FP)
 	RET
 
-TEXT runtime·write(SB),NOSPLIT,$-8
+TEXT runtime·write(SB),NOSPLIT,$-4
 	MOVW fd+0(FP), R0	// arg 1 fd
 	MOVW p+4(FP), R1	// arg 2 buf
 	MOVW n+8(FP), R2	// arg 3 count
@@ -129,7 +129,7 @@ TEXT runtime·write(SB),NOSPLIT,$-8
 	MOVW	R0, ret+12(FP)
 	RET
 
-TEXT runtime·closefd(SB),NOSPLIT,$-8
+TEXT runtime·closefd(SB),NOSPLIT,$-4
 	MOVW fd+0(FP), R0	// arg 1 fd
 	MOVW $SYS_close, R7
 	SWI $0
@@ -137,7 +137,7 @@ TEXT runtime·closefd(SB),NOSPLIT,$-8
 	MOVW	R0, ret+4(FP)
 	RET
 
-TEXT runtime·getrlimit(SB),NOSPLIT,$-8
+TEXT runtime·getrlimit(SB),NOSPLIT,$-4
 	MOVW kind+0(FP), R0
 	MOVW limit+4(FP), R1
 	MOVW $SYS_getrlimit, R7
@@ -168,7 +168,7 @@ TEXT runtime·raiseproc(SB),NOSPLIT,$0
 	SWI $0
 	RET
 
-TEXT runtime·setitimer(SB), NOSPLIT, $-8
+TEXT runtime·setitimer(SB), NOSPLIT, $-4
 	MOVW mode+0(FP), R0
 	MOVW new+4(FP), R1
 	MOVW old+8(FP), R2
@@ -216,7 +216,7 @@ TEXT runtime·nanotime(SB), NOSPLIT, $32
 	MOVW R1, ret_hi+4(FP)
 	RET
 
-TEXT runtime·sigaction(SB),NOSPLIT,$-8
+TEXT runtime·sigaction(SB),NOSPLIT,$-4
 	MOVW sig+0(FP), R0		// arg 1 sig
 	MOVW new+4(FP), R1		// arg 2 act
 	MOVW old+8(FP), R2		// arg 3 oact
@@ -283,7 +283,7 @@ TEXT runtime·madvise(SB),NOSPLIT,$0
 	// ignore failure - maybe pages are locked
 	RET
 	
-TEXT runtime·sigaltstack(SB),NOSPLIT,$-8
+TEXT runtime·sigaltstack(SB),NOSPLIT,$-4
 	MOVW new+0(FP), R0
 	MOVW old+4(FP), R1
 	MOVW $SYS_sigaltstack, R7

src/runtime/sys_netbsd_arm.s

@@ -33,7 +33,7 @@ storeloop:
 	MOVW R8, (R8)
 	JMP 0(PC)
 
-TEXT runtime·open(SB),NOSPLIT,$-8
+TEXT runtime·open(SB),NOSPLIT,$-4
 	MOVW name+0(FP), R0
 	MOVW mode+4(FP), R1
 	MOVW perm+8(FP), R2
@@ -42,14 +42,14 @@ TEXT runtime·open(SB),NOSPLIT,$-8
 	MOVW	R0, ret+12(FP)
 	RET
 
-TEXT runtime·closefd(SB),NOSPLIT,$-8
+TEXT runtime·closefd(SB),NOSPLIT,$-4
 	MOVW fd+0(FP), R0
 	SWI $0xa00006
 	MOVW.CS	$-1, R0
 	MOVW	R0, ret+4(FP)
 	RET
 
-TEXT runtime·read(SB),NOSPLIT,$-8
+TEXT runtime·read(SB),NOSPLIT,$-4
 	MOVW fd+0(FP), R0
 	MOVW p+4(FP), R1
 	MOVW n+8(FP), R2