[ARM] Allow r2 to be passed through the decompressor to the kernel

This is part of a patch from Marc Singer to allow r2 to be
passed to the kernel. Marc's original comments follow:

This revised R2 (atags pointer) patch incorporates comments from Nico
Pitre and Ben Dooks. It modifies the head.S files such that the R2
value set by the bootloader is conveyed to the kernel startup code.
The kernel head.S heuristically validates the pointer. It will set R2
to zero if it believes the pointer is invalid. Presently, it requires
that the ATAGS list reside in the first 16KiB of physical RAM.
Relaxing this contraint may be both desirable as well as tricky.

Signed-off-by: Marc Singer <[e-mail]>
Signed-off-by: Russell King <[e-mail]>

arch/arm/boot/compressed/head.S

···
 		kputc	#'\n'
 		kphex	r5, 8		/* decompressed kernel start */
 		kputc	#'-'
-		kphex	r8, 8		/* decompressed kernel end  */
+		kphex	r9, 8		/* decompressed kernel end  */
 		kputc	#'>'
 		kphex	r4, 8		/* kernel execution address */
 		kputc	#'\n'
···
 		.word	start			@ absolute load/run zImage address
 		.word	_edata			@ zImage end address
 1:		mov	r7, r1			@ save architecture ID
-		mov	r8, #0			@ save r0
+		mov	r8, r2			@ save atags pointer
 
 #ifndef __ARM_ARCH_2__
 		/*
···
 
 		/*
 		 * some architecture specific code can be inserted
-		 * by the linker here, but it should preserve r7 and r8.
+		 * by the linker here, but it should preserve r7, r8, and r9.
 		 */
 
 		.text
···
  * r5     = decompressed kernel start
  * r6     = processor ID
  * r7     = architecture ID
- * r8-r14 = unused
+ * r8     = atags pointer
+ * r9-r14 = corrupted
  */
 		add	r1, r5, r0		@ end of decompressed kernel
 		adr	r2, reloc_start
 		ldr	r3, LC1
 		add	r3, r2, r3
-1:		ldmia	r2!, {r8 - r13}		@ copy relocation code
-		stmia	r1!, {r8 - r13}
-		ldmia	r2!, {r8 - r13}
-		stmia	r1!, {r8 - r13}
+1:		ldmia	r2!, {r9 - r14}		@ copy relocation code
+		stmia	r1!, {r9 - r14}
+		ldmia	r2!, {r9 - r14}
+		stmia	r1!, {r9 - r14}
 		cmp	r2, r3
 		blo	1b
 
···
  *  r4 = kernel execution address
  *  r6 = processor ID
  *  r7 = architecture number
- *  r8 = run-time address of "start"
+ *  r8 = atags pointer
+ *  r9 = run-time address of "start"  (???)
  * On exit,
- *  r1, r2, r3, r8, r9, r12 corrupted
+ *  r1, r2, r3, r9, r10, r12 corrupted
  * This routine must preserve:
- *  r4, r5, r6, r7
+ *  r4, r5, r6, r7, r8
  */
 		.align	5
 cache_on:	mov	r3, #8			@ cache_on function
···
  * bits for the RAM area only.
  */
 		mov	r0, r3
-		mov	r8, r0, lsr #18
-		mov	r8, r8, lsl #18		@ start of RAM
-		add	r9, r8, #0x10000000	@ a reasonable RAM size
+		mov	r9, r0, lsr #18
+		mov	r9, r9, lsl #18		@ start of RAM
+		add	r10, r9, #0x10000000	@ a reasonable RAM size
 		mov	r1, #0x12
 		orr	r1, r1, #3 << 10
 		add	r2, r3, #16384
 1:		cmp	r1, r8			@ if virt > start of RAM
 		orrhs	r1, r1, #0x0c		@ set cacheable, bufferable
-		cmp	r1, r9			@ if virt > end of RAM
+		cmp	r1, r10			@ if virt > end of RAM
 		bichs	r1, r1, #0x0c		@ clear cacheable, bufferable
 		str	r1, [r0], #4		@ 1:1 mapping
 		add	r1, r1, #1048576
···
  * r5     = decompressed kernel start
  * r6     = processor ID
  * r7     = architecture ID
- * r8-r14 = unused
+ * r8     = atags pointer
+ * r9-r14 = corrupted
  */
 		.align	5
-reloc_start:	add	r8, r5, r0
+reloc_start:	add	r9, r5, r0
 		debug_reloc_start
 		mov	r1, r4
 1:
 		.rept	4
-		ldmia	r5!, {r0, r2, r3, r9 - r13}	@ relocate kernel
-		stmia	r1!, {r0, r2, r3, r9 - r13}
+		ldmia	r5!, {r0, r2, r3, r10 - r14}	@ relocate kernel
+		stmia	r1!, {r0, r2, r3, r10 - r14}
 		.endr
 
-		cmp	r5, r8
+		cmp	r5, r9
 		blo	1b
 		debug_reloc_end
 
 call_kernel:	bl	cache_clean_flush
 		bl	cache_off
-		mov	r0, #0
+		mov	r0, #0			@ must be zero
 		mov	r1, r7			@ restore architecture number
+		mov	r2, r8			@ restore atags pointer
 		mov	pc, r4			@ call kernel
 
 /*