add support for ctors/dtors on arm with modern gcc

a while back, gcc switched from using the old _init/_fini fragments
method for calling ctors and dtors on arm to the __init_array and
__fini_array method. unfortunately, on glibc this depends on ugly
hacks involving making libc.so a linker script and pulling parts of
libc into the main program binary. so I cheat a little bit, and just
write asm to iterate over the init/fini arrays from the _init/_fini
asm. the same approach could be used on any arch it's needed on, but
for now arm is the only one.

crt/arm/crti.s

 .section .init
 .global _init
 _init:
-	push {lr}
+	push {r0,r1,r2,r4,r5,lr}
+
+.weak __fini_array_start
+.weak __fini_array_end
 
 .section .fini
 .global _fini
 _fini:
-	push {lr}
+	push {r4,r5,r6,lr}
+	adr lr, 1f
+	ldr r4, 2f
+	ldr r5, 2f+4
+	add r4, r4, lr
+	add r5, r5, lr
+1:	adr lr, 1b
+	cmp r4, r5
+	beq 3f
+	ldmia r4!, {r3}
+	tst r3,#1
+	moveq pc,r3
+	bx r3
+2:	.word __fini_array_start-1b
+	.word __fini_array_end-1b
+3:	

crt/arm/crtn.s

+.weak __init_array_start
+.weak __init_array_end
+
 .section .init
-	pop {lr}
+	adr lr, 1f
+	ldr r4, 2f
+	ldr r5, 2f+4
+	add r4, r4, lr
+	add r5, r5, lr
+1:	adr lr, 1b
+	cmp r4, r5
+	beq 3f
+	ldmia r4!, {r3}
+	ldm sp, {r0,r1,r2}
+	tst r3,#1
+	moveq pc,r3
+	bx r3
+3:	pop {r0,r1,r2,r4,r5,lr}
 	tst lr,#1
 	moveq pc,lr
 	bx lr
+2:	.word __init_array_start-1b
+	.word __init_array_end-1b
 
 .section .fini
-	pop {lr}
+	pop {r4,r5,r6,lr}
 	tst lr,#1
 	moveq pc,lr
 	bx lr