start.s

AT91RM9200的完整启动代码：包括loader, boot及U-boot三部分均已编译通过！欢迎下载使用！

	REST_10GPRS(3, r1)
	REST_10GPRS(13, r1)
	REST_8GPRS(23, r1)
	REST_GPR(31, r1)
	lwz	r2,_NIP(r1)	/* Restore environment */
	lwz	r0,_MSR(r1)
	mtspr	SRR0,r2
	mtspr	SRR1,r0
	lwz	r0,GPR0(r1)
	lwz	r2,GPR2(r1)
	lwz	r1,GPR1(r1)
	SYNC
	rfi

/* Cache functions.
*/
	.globl	icache_enable
icache_enable:
	mfspr	r5,HID0		/* turn on the I cache. */
	ori	r5,r5,0x8800	/* Instruction cache only! */
	addis	r6,0,0xFFFF
	ori	r6,r6,0xF7FF
	and	r6,r5,r6	/* clear the invalidate bit */
	sync
	mtspr	HID0,r5
	mtspr	HID0,r6
	isync
	sync
	blr

	.globl	icache_disable
icache_disable:
	mfspr	r5,HID0
	addis	r6,0,0xFFFF
	ori	r6,r6,0x7FFF
	and	r5,r5,r6
	sync
	mtspr	HID0,r5
	isync
	sync
	blr

	.globl	icache_status
icache_status:
	mfspr	r3, HID0
	srwi	r3, r3, 15	/* >>15 & 1=> select bit 16 */
	andi.	r3, r3, 1
	blr

	.globl	dcache_enable
dcache_enable:
	mfspr	r5,HID0		/* turn on the D cache. */
	ori	r5,r5,0x4400	/* Data cache only! */
	mfspr	r4, PVR		/* read PVR */
	srawi	r3, r4, 16	/* shift off the least 16 bits */
	cmpi	0, 0, r3, 0xC	/* Check for Max pvr */
	bne	NotMax
	ori	r5,r5,0x0040	/* setting the DCFA bit, for Max rev 1 errata */
NotMax:
	addis	r6,0,0xFFFF
	ori	r6,r6,0xFBFF
	and	r6,r5,r6	/* clear the invalidate bit */
	sync
	mtspr	HID0,r5
	mtspr	HID0,r6
	isync
	sync
	blr

	.globl	dcache_disable
dcache_disable:
	mfspr	r5,HID0
	addis	r6,0,0xFFFF
	ori	r6,r6,0xBFFF
	and	r5,r5,r6
	sync
	mtspr	HID0,r5
	isync
	sync
	blr

	.globl	dcache_status
dcache_status:
	mfspr	r3, HID0
	srwi	r3, r3, 14	/* >>14 & 1=> select bit 17 */
	andi.	r3, r3, 1
	blr

	.globl	dc_read
dc_read:
/*TODO : who uses this, what should it do?
*/
	blr


	.globl get_pvr
get_pvr:
	mfspr	r3, PVR
	blr


/*------------------------------------------------------------------------------*/

/*
 * void relocate_code (addr_sp, gd, addr_moni)
 *
 * This "function" does not return, instead it continues in RAM
 * after relocating the monitor code.
 *
 * r3 = dest
 * r4 = src
 * r5 = length in bytes
 * r6 = cachelinesize
 */
	.globl	relocate_code
relocate_code:

	mr	r1,  r3		/* Set new stack pointer		*/
	mr	r9,  r4		/* Save copy of Global Data pointer	*/
	mr	r10, r5		/* Save copy of Destination Address	*/

	mr	r3,  r5				/* Destination Address	*/
#ifdef CFG_RAMBOOT
	lis	r4, CFG_SDRAM_BASE@h		/* Source      Address	*/
	ori	r4, r4, CFG_SDRAM_BASE@l
#else
	lis	r4, CFG_MONITOR_BASE@h		/* Source      Address	*/
	ori	r4, r4, CFG_MONITOR_BASE@l
#endif
	lwz	r5, GOT(__init_end)
	sub	r5, r5, r4
	li	r6, CFG_CACHELINE_SIZE		/* Cache Line Size	*/

	/*
	 * Fix GOT pointer:
	 *
	 * New GOT-PTR = (old GOT-PTR - CFG_MONITOR_BASE) + Destination Address
	 *
	 * Offset:
	 */
	sub	r15, r10, r4

	/* First our own GOT */
	add	r14, r14, r15
	/* the the one used by the C code */
	add	r30, r30, r15

	/*
	 * Now relocate code
	 */

	cmplw	cr1,r3,r4
	addi	r0,r5,3
	srwi.	r0,r0,2
	beq	cr1,4f		/* In place copy is not necessary	*/
	beq	7f		/* Protect against 0 count		*/
	mtctr	r0
	bge	cr1,2f

	la	r8,-4(r4)
	la	r7,-4(r3)
1:	lwzu	r0,4(r8)
	stwu	r0,4(r7)
	bdnz	1b
	b	4f

2:	slwi	r0,r0,2
	add	r8,r4,r0
	add	r7,r3,r0
3:	lwzu	r0,-4(r8)
	stwu	r0,-4(r7)
	bdnz	3b

/*
 * Now flush the cache: note that we must start from a cache aligned
 * address. Otherwise we might miss one cache line.
 */
4:	cmpwi	r6,0
	add	r5,r3,r5
	beq	7f		/* Always flush prefetch queue in any case */
	subi	r0,r6,1
	andc	r3,r3,r0
	mr	r4,r3
5:	dcbst	0,r4
	add	r4,r4,r6
	cmplw	r4,r5
	blt	5b
	sync			/* Wait for all dcbst to complete on bus */
	mr	r4,r3
6:	icbi	0,r4
	add	r4,r4,r6
	cmplw	r4,r5
	blt	6b
7:	sync			/* Wait for all icbi to complete on bus	*/
	isync

/*
 * We are done. Do not return, instead branch to second part of board
 * initialization, now running from RAM.
 */

	addi	r0, r10, in_ram - _start + EXC_OFF_SYS_RESET
	mtlr	r0
	blr

in_ram:

	/*
	 * Relocation Function, r14 point to got2+0x8000
	 *
	 * Adjust got2 pointers, no need to check for 0, this code
	 * already puts a few entries in the table.
	 */
	li	r0,__got2_entries@sectoff@l
	la	r3,GOT(_GOT2_TABLE_)
	lwz	r11,GOT(_GOT2_TABLE_)
	mtctr	r0
	sub	r11,r3,r11
	addi	r3,r3,-4
1:	lwzu	r0,4(r3)
	add	r0,r0,r11
	stw	r0,0(r3)
	bdnz	1b

	/*
	 * Now adjust the fixups and the pointers to the fixups
	 * in case we need to move ourselves again.
	 */
2:	li	r0,__fixup_entries@sectoff@l
	lwz	r3,GOT(_FIXUP_TABLE_)
	cmpwi	r0,0
	mtctr	r0
	addi	r3,r3,-4
	beq	4f
3:	lwzu	r4,4(r3)
	lwzux	r0,r4,r11
	add	r0,r0,r11
	stw	r10,0(r3)
	stw	r0,0(r4)
	bdnz	3b
4:
clear_bss:
	/*
	 * Now clear BSS segment
	 */
	lwz	r3,GOT(__bss_start)
	lwz	r4,GOT(_end)

	cmplw	0, r3, r4
	beq	6f

	li	r0, 0
5:
	stw	r0, 0(r3)
	addi	r3, r3, 4
	cmplw	0, r3, r4
	blt	5b
6:

	mr	r3, r9		/* Global Data pointer		*/
	mr	r4, r10		/* Destination Address		*/
	bl	board_init_r

	/*
	 * Copy exception vector code to low memory
	 *
	 * r3: dest_addr
	 * r7: source address, r8: end address, r9: target address
	 */
	.globl	trap_init
trap_init:
	lwz	r7, GOT(_start)
	lwz	r8, GOT(_end_of_vectors)

	li	r9, 0x100		/* reset vector always at 0x100 */

	cmplw	0, r7, r8
	bgelr				/* return if r7>=r8 - just in case */

	mflr	r4			/* save link register		*/
1:
	lwz	r0, 0(r7)
	stw	r0, 0(r9)
	addi	r7, r7, 4
	addi	r9, r9, 4
	cmplw	0, r7, r8
	bne	1b

	/*
	 * relocate `hdlr' and `int_return' entries
	 */
	li	r7, .L_MachineCheck - _start + EXC_OFF_SYS_RESET
	li	r8, Alignment - _start + EXC_OFF_SYS_RESET
2:
	bl	trap_reloc
	addi	r7, r7, 0x100		/* next exception vector	*/
	cmplw	0, r7, r8
	blt	2b

	li	r7, .L_Alignment - _start + EXC_OFF_SYS_RESET
	bl	trap_reloc

	li	r7, .L_ProgramCheck - _start + EXC_OFF_SYS_RESET
	bl	trap_reloc

	li	r7, .L_FPUnavailable - _start + EXC_OFF_SYS_RESET
	li	r8, SystemCall - _start + EXC_OFF_SYS_RESET
3:
	bl	trap_reloc
	addi	r7, r7, 0x100		/* next exception vector	*/
	cmplw	0, r7, r8
	blt	3b

	li	r7, .L_SingleStep - _start + EXC_OFF_SYS_RESET
	li	r8, _end_of_vectors - _start + EXC_OFF_SYS_RESET
4:
	bl	trap_reloc
	addi	r7, r7, 0x100		/* next exception vector	*/
	cmplw	0, r7, r8
	blt	4b

	mtlr	r4			/* restore link register	*/
	blr

	/*
	 * Function: relocate entries for one exception vector
	 */
trap_reloc:
	lwz	r0, 0(r7)		/* hdlr ...			*/
	add	r0, r0, r3		/*  ... += dest_addr		*/
	stw	r0, 0(r7)

	lwz	r0, 4(r7)		/* int_return ...		*/
	add	r0, r0, r3		/*  ... += dest_addr		*/
	stw	r0, 4(r7)

	blr

	/* Setup the BAT registers.
	 */
setup_bats:
	lis	r4, CFG_IBAT0L@h
	ori	r4, r4, CFG_IBAT0L@l
	lis	r3, CFG_IBAT0U@h
	ori	r3, r3, CFG_IBAT0U@l
	mtspr	IBAT0L, r4
	mtspr	IBAT0U, r3
	isync

	lis	r4, CFG_DBAT0L@h
	ori	r4, r4, CFG_DBAT0L@l
	lis	r3, CFG_DBAT0U@h
	ori	r3, r3, CFG_DBAT0U@l
	mtspr	DBAT0L, r4
	mtspr	DBAT0U, r3
	isync

	lis	r4, CFG_IBAT1L@h
	ori	r4, r4, CFG_IBAT1L@l
	lis	r3, CFG_IBAT1U@h
	ori	r3, r3, CFG_IBAT1U@l
	mtspr	IBAT1L, r4
	mtspr	IBAT1U, r3
	isync

	lis	r4, CFG_DBAT1L@h
	ori	r4, r4, CFG_DBAT1L@l
	lis	r3, CFG_DBAT1U@h
	ori	r3, r3, CFG_DBAT1U@l
	mtspr	DBAT1L, r4
	mtspr	DBAT1U, r3
	isync

	lis	r4, CFG_IBAT2L@h
	ori	r4, r4, CFG_IBAT2L@l
	lis	r3, CFG_IBAT2U@h
	ori	r3, r3, CFG_IBAT2U@l
	mtspr	IBAT2L, r4
	mtspr	IBAT2U, r3
	isync

	lis	r4, CFG_DBAT2L@h
	ori	r4, r4, CFG_DBAT2L@l
	lis	r3, CFG_DBAT2U@h
	ori	r3, r3, CFG_DBAT2U@l
	mtspr	DBAT2L, r4
	mtspr	DBAT2U, r3
	isync

	lis	r4, CFG_IBAT3L@h
	ori	r4, r4, CFG_IBAT3L@l
	lis	r3, CFG_IBAT3U@h
	ori	r3, r3, CFG_IBAT3U@l
	mtspr	IBAT3L, r4
	mtspr	IBAT3U, r3
	isync

	lis	r4, CFG_DBAT3L@h
	ori	r4, r4, CFG_DBAT3L@l
	lis	r3, CFG_DBAT3U@h
	ori	r3, r3, CFG_DBAT3U@l
	mtspr	DBAT3L, r4
	mtspr	DBAT3U, r3
	isync

	/* Invalidate TLBs.
	 * -> for (val = 0; val < 0x20000; val+=0x1000)
	 * ->   tlbie(val);
	 */
	lis	r3, 0
	lis	r5, 2

1:
	tlbie	r3
	addi	r3, r3, 0x1000
	cmp	0, 0, r3, r5
	blt	1b

	blr