ARM: Bring back ARMv3 IO and user access code

This partially reverts 357c9c1f07d4546bc3fbc0fd1044d96b114d14ed
(ARM: Remove support for ARMv3 ARM610 and ARM710 CPUs).

Although we only support StrongARM on the RiscPC, we need to keep the
ARMv3 user access code for this platform because the bus does not
understand half-word load/stores.

Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>

4 files changed, 816 insertions, 3 deletions

diff --git a/arch/arm/lib/Makefile b/arch/arm/lib/Makefile
index 2473fd1fd51c..af72969820b4 100644
--- a/arch/arm/lib/Makefile
+++ b/arch/arm/lib/Makefile
@@ -16,13 +16,30 @@ lib-y		:= backtrace.o changebit.o csumipv6.o csumpartial.o   \
 		   call_with_stack.o
 
 mmu-y	:= clear_user.o copy_page.o getuser.o putuser.o
-mmu-y	+= copy_from_user.o copy_to_user.o
+
+# the code in uaccess.S is not preemption safe and
+# probably faster on ARMv3 only
+ifeq ($(CONFIG_PREEMPT),y)
+  mmu-y	+= copy_from_user.o copy_to_user.o
+else
+ifneq ($(CONFIG_CPU_32v3),y)
+  mmu-y	+= copy_from_user.o copy_to_user.o
+else
+  mmu-y	+= uaccess.o
+endif
+endif
 
 # using lib_ here won't override already available weak symbols
 obj-$(CONFIG_UACCESS_WITH_MEMCPY) += uaccess_with_memcpy.o
 
-lib-$(CONFIG_MMU)		+= $(mmu-y)
-lib-y				+= io-readsw-armv4.o io-writesw-armv4.o
+lib-$(CONFIG_MMU) += $(mmu-y)
+
+ifeq ($(CONFIG_CPU_32v3),y)
+  lib-y	+= io-readsw-armv3.o io-writesw-armv3.o
+else
+  lib-y	+= io-readsw-armv4.o io-writesw-armv4.o
+endif
+
 lib-$(CONFIG_ARCH_RPC)		+= ecard.o io-acorn.o floppydma.o
 lib-$(CONFIG_ARCH_SHARK)	+= io-shark.o
 
diff --git a/arch/arm/lib/io-readsw-armv3.S b/arch/arm/lib/io-readsw-armv3.S
new file mode 100644
index 000000000000..88487c8c4f23
--- /dev/null
+++ b/arch/arm/lib/io-readsw-armv3.S
@@ -0,0 +1,106 @@
+/*
+ *  linux/arch/arm/lib/io-readsw-armv3.S
+ *
+ *  Copyright (C) 1995-2000 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+.Linsw_bad_alignment:
+		adr	r0, .Linsw_bad_align_msg
+		mov	r2, lr
+		b	panic
+.Linsw_bad_align_msg:
+		.asciz	"insw: bad buffer alignment (0x%p, lr=0x%08lX)\n"
+		.align
+
+.Linsw_align:	tst	r1, #1
+		bne	.Linsw_bad_alignment
+
+		ldr	r3, [r0]
+		strb	r3, [r1], #1
+		mov	r3, r3, lsr #8
+		strb	r3, [r1], #1
+
+		subs	r2, r2, #1
+		moveq	pc, lr
+
+ENTRY(__raw_readsw)
+		teq	r2, #0		@ do we have to check for the zero len?
+		moveq	pc, lr
+		tst	r1, #3
+		bne	.Linsw_align
+
+.Linsw_aligned:	mov	ip, #0xff
+		orr	ip, ip, ip, lsl #8
+		stmfd	sp!, {r4, r5, r6, lr}
+
+		subs	r2, r2, #8
+		bmi	.Lno_insw_8
+
+.Linsw_8_lp:	ldr	r3, [r0]
+		and	r3, r3, ip
+		ldr	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+
+		ldr	r4, [r0]
+		and	r4, r4, ip
+		ldr	r5, [r0]
+		orr	r4, r4, r5, lsl #16
+
+		ldr	r5, [r0]
+		and	r5, r5, ip
+		ldr	r6, [r0]
+		orr	r5, r5, r6, lsl #16
+
+		ldr	r6, [r0]
+		and	r6, r6, ip
+		ldr	lr, [r0]
+		orr	r6, r6, lr, lsl #16
+
+		stmia	r1!, {r3 - r6}
+
+		subs	r2, r2, #8
+		bpl	.Linsw_8_lp
+
+		tst	r2, #7
+		ldmeqfd	sp!, {r4, r5, r6, pc}
+
+.Lno_insw_8:	tst	r2, #4
+		beq	.Lno_insw_4
+
+		ldr	r3, [r0]
+		and	r3, r3, ip
+		ldr	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+
+		ldr	r4, [r0]
+		and	r4, r4, ip
+		ldr	r5, [r0]
+		orr	r4, r4, r5, lsl #16
+
+		stmia	r1!, {r3, r4}
+
+.Lno_insw_4:	tst	r2, #2
+		beq	.Lno_insw_2
+
+		ldr	r3, [r0]
+		and	r3, r3, ip
+		ldr	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+
+		str	r3, [r1], #4
+
+.Lno_insw_2:	tst	r2, #1
+		ldrne	r3, [r0]
+		strneb	r3, [r1], #1
+		movne	r3, r3, lsr #8
+		strneb	r3, [r1]
+
+		ldmfd	sp!, {r4, r5, r6, pc}
+
+
diff --git a/arch/arm/lib/io-writesw-armv3.S b/arch/arm/lib/io-writesw-armv3.S
new file mode 100644
index 000000000000..49b800419e32
--- /dev/null
+++ b/arch/arm/lib/io-writesw-armv3.S
@@ -0,0 +1,126 @@
+/*
+ *  linux/arch/arm/lib/io-writesw-armv3.S
+ *
+ *  Copyright (C) 1995-2000 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+.Loutsw_bad_alignment:
+		adr	r0, .Loutsw_bad_align_msg
+		mov	r2, lr
+		b	panic
+.Loutsw_bad_align_msg:
+		.asciz	"outsw: bad buffer alignment (0x%p, lr=0x%08lX)\n"
+		.align
+
+.Loutsw_align:	tst	r1, #1
+		bne	.Loutsw_bad_alignment
+
+		add	r1, r1, #2
+
+		ldr	r3, [r1, #-4]
+		mov	r3, r3, lsr #16
+		orr	r3, r3, r3, lsl #16
+		str	r3, [r0]
+		subs	r2, r2, #1
+		moveq	pc, lr
+
+ENTRY(__raw_writesw)
+		teq	r2, #0		@ do we have to check for the zero len?
+		moveq	pc, lr
+		tst	r1, #3
+		bne	.Loutsw_align
+
+		stmfd	sp!, {r4, r5, r6, lr}
+
+		subs	r2, r2, #8
+		bmi	.Lno_outsw_8
+
+.Loutsw_8_lp:	ldmia	r1!, {r3, r4, r5, r6}
+
+		mov	ip, r3, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r3, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+		mov	ip, r4, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r4, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+		mov	ip, r5, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r5, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+		mov	ip, r6, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r6, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+		subs	r2, r2, #8
+		bpl	.Loutsw_8_lp
+
+		tst	r2, #7
+		ldmeqfd	sp!, {r4, r5, r6, pc}
+
+.Lno_outsw_8:	tst	r2, #4
+		beq	.Lno_outsw_4
+
+		ldmia	r1!, {r3, r4}
+
+		mov	ip, r3, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r3, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+		mov	ip, r4, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r4, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+.Lno_outsw_4:	tst	r2, #2
+		beq	.Lno_outsw_2
+
+		ldr	r3, [r1], #4
+
+		mov	ip, r3, lsl #16
+		orr	ip, ip, ip, lsr #16
+		str	ip, [r0]
+
+		mov	ip, r3, lsr #16
+		orr	ip, ip, ip, lsl #16
+		str	ip, [r0]
+
+.Lno_outsw_2:	tst	r2, #1
+
+		ldrne	r3, [r1]
+
+		movne	ip, r3, lsl #16
+		orrne	ip, ip, ip, lsr #16
+		strne	ip, [r0]
+
+		ldmfd	sp!, {r4, r5, r6, pc}
diff --git a/arch/arm/lib/uaccess.S b/arch/arm/lib/uaccess.S
new file mode 100644
index 000000000000..5c908b1cb8ed
--- /dev/null
+++ b/arch/arm/lib/uaccess.S
@@ -0,0 +1,564 @@
+/*
+ *  linux/arch/arm/lib/uaccess.S
+ *
+ *  Copyright (C) 1995, 1996,1997,1998 Russell King
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ *  Routines to block copy data to/from user memory
+ *   These are highly optimised both for the 4k page size
+ *   and for various alignments.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/errno.h>
+#include <asm/domain.h>
+
+		.text
+
+#define PAGE_SHIFT 12
+
+/* Prototype: int __copy_to_user(void *to, const char *from, size_t n)
+ * Purpose  : copy a block to user memory from kernel memory
+ * Params   : to   - user memory
+ *          : from - kernel memory
+ *          : n    - number of bytes to copy
+ * Returns  : Number of bytes NOT copied.
+ */
+
+.Lc2u_dest_not_aligned:
+		rsb	ip, ip, #4
+		cmp	ip, #2
+		ldrb	r3, [r1], #1
+USER(	TUSER(	strb)	r3, [r0], #1)			@ May fault
+		ldrgeb	r3, [r1], #1
+USER(	TUSER(	strgeb) r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #1
+USER(	TUSER(	strgtb) r3, [r0], #1)			@ May fault
+		sub	r2, r2, ip
+		b	.Lc2u_dest_aligned
+
+ENTRY(__copy_to_user)
+		stmfd	sp!, {r2, r4 - r7, lr}
+		cmp	r2, #4
+		blt	.Lc2u_not_enough
+		ands	ip, r0, #3
+		bne	.Lc2u_dest_not_aligned
+.Lc2u_dest_aligned:
+
+		ands	ip, r1, #3
+		bne	.Lc2u_src_not_aligned
+/*
+ * Seeing as there has to be at least 8 bytes to copy, we can
+ * copy one word, and force a user-mode page fault...
+ */
+
+.Lc2u_0fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lc2u_0nowords
+		ldr	r3, [r1], #4
+USER(	TUSER(	str)	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT	@ On each page, use a ld/st??t instruction
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lc2u_0fupi
+/*
+ * ip = max no. of bytes to copy before needing another "strt" insn
+ */
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #32
+		blt	.Lc2u_0rem8lp
+
+.Lc2u_0cpy8lp:	ldmia	r1!, {r3 - r6}
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		ldmia	r1!, {r3 - r6}
+		subs	ip, ip, #32
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		bpl	.Lc2u_0cpy8lp
+
+.Lc2u_0rem8lp:	cmn	ip, #16
+		ldmgeia	r1!, {r3 - r6}
+		stmgeia	r0!, {r3 - r6}			@ Shouldnt fault
+		tst	ip, #8
+		ldmneia	r1!, {r3 - r4}
+		stmneia	r0!, {r3 - r4}			@ Shouldnt fault
+		tst	ip, #4
+		ldrne	r3, [r1], #4
+	TUSER(	strne) r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.Lc2u_0fupi
+.Lc2u_0nowords:	teq	ip, #0
+		beq	.Lc2u_finished
+.Lc2u_nowords:	cmp	ip, #2
+		ldrb	r3, [r1], #1
+USER(	TUSER(	strb)	r3, [r0], #1)			@ May fault
+		ldrgeb	r3, [r1], #1
+USER(	TUSER(	strgeb) r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #1
+USER(	TUSER(	strgtb) r3, [r0], #1)			@ May fault
+		b	.Lc2u_finished
+
+.Lc2u_not_enough:
+		movs	ip, r2
+		bne	.Lc2u_nowords
+.Lc2u_finished:	mov	r0, #0
+		ldmfd	sp!, {r2, r4 - r7, pc}
+
+.Lc2u_src_not_aligned:
+		bic	r1, r1, #3
+		ldr	r7, [r1], #4
+		cmp	ip, #2
+		bgt	.Lc2u_3fupi
+		beq	.Lc2u_2fupi
+.Lc2u_1fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lc2u_1nowords
+		mov	r3, r7, pull #8
+		ldr	r7, [r1], #4
+		orr	r3, r3, r7, push #24
+USER(	TUSER(	str)	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lc2u_1fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.Lc2u_1rem8lp
+
+.Lc2u_1cpy8lp:	mov	r3, r7, pull #8
+		ldmia	r1!, {r4 - r7}
+		subs	ip, ip, #16
+		orr	r3, r3, r4, push #24
+		mov	r4, r4, pull #8
+		orr	r4, r4, r5, push #24
+		mov	r5, r5, pull #8
+		orr	r5, r5, r6, push #24
+		mov	r6, r6, pull #8
+		orr	r6, r6, r7, push #24
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		bpl	.Lc2u_1cpy8lp
+
+.Lc2u_1rem8lp:	tst	ip, #8
+		movne	r3, r7, pull #8
+		ldmneia	r1!, {r4, r7}
+		orrne	r3, r3, r4, push #24
+		movne	r4, r4, pull #8
+		orrne	r4, r4, r7, push #24
+		stmneia	r0!, {r3 - r4}			@ Shouldnt fault
+		tst	ip, #4
+		movne	r3, r7, pull #8
+		ldrne	r7, [r1], #4
+		orrne	r3, r3, r7, push #24
+	TUSER(	strne) r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.Lc2u_1fupi
+.Lc2u_1nowords:	mov	r3, r7, get_byte_1
+		teq	ip, #0
+		beq	.Lc2u_finished
+		cmp	ip, #2
+USER(	TUSER(	strb)	r3, [r0], #1)			@ May fault
+		movge	r3, r7, get_byte_2
+USER(	TUSER(	strgeb) r3, [r0], #1)			@ May fault
+		movgt	r3, r7, get_byte_3
+USER(	TUSER(	strgtb) r3, [r0], #1)			@ May fault
+		b	.Lc2u_finished
+
+.Lc2u_2fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lc2u_2nowords
+		mov	r3, r7, pull #16
+		ldr	r7, [r1], #4
+		orr	r3, r3, r7, push #16
+USER(	TUSER(	str)	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lc2u_2fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.Lc2u_2rem8lp
+
+.Lc2u_2cpy8lp:	mov	r3, r7, pull #16
+		ldmia	r1!, {r4 - r7}
+		subs	ip, ip, #16
+		orr	r3, r3, r4, push #16
+		mov	r4, r4, pull #16
+		orr	r4, r4, r5, push #16
+		mov	r5, r5, pull #16
+		orr	r5, r5, r6, push #16
+		mov	r6, r6, pull #16
+		orr	r6, r6, r7, push #16
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		bpl	.Lc2u_2cpy8lp
+
+.Lc2u_2rem8lp:	tst	ip, #8
+		movne	r3, r7, pull #16
+		ldmneia	r1!, {r4, r7}
+		orrne	r3, r3, r4, push #16
+		movne	r4, r4, pull #16
+		orrne	r4, r4, r7, push #16
+		stmneia	r0!, {r3 - r4}			@ Shouldnt fault
+		tst	ip, #4
+		movne	r3, r7, pull #16
+		ldrne	r7, [r1], #4
+		orrne	r3, r3, r7, push #16
+	TUSER(	strne) r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.Lc2u_2fupi
+.Lc2u_2nowords:	mov	r3, r7, get_byte_2
+		teq	ip, #0
+		beq	.Lc2u_finished
+		cmp	ip, #2
+USER(	TUSER(	strb)	r3, [r0], #1)			@ May fault
+		movge	r3, r7, get_byte_3
+USER(	TUSER(	strgeb) r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #0
+USER(	TUSER(	strgtb) r3, [r0], #1)			@ May fault
+		b	.Lc2u_finished
+
+.Lc2u_3fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lc2u_3nowords
+		mov	r3, r7, pull #24
+		ldr	r7, [r1], #4
+		orr	r3, r3, r7, push #8
+USER(	TUSER(	str)	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lc2u_3fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.Lc2u_3rem8lp
+
+.Lc2u_3cpy8lp:	mov	r3, r7, pull #24
+		ldmia	r1!, {r4 - r7}
+		subs	ip, ip, #16
+		orr	r3, r3, r4, push #8
+		mov	r4, r4, pull #24
+		orr	r4, r4, r5, push #8
+		mov	r5, r5, pull #24
+		orr	r5, r5, r6, push #8
+		mov	r6, r6, pull #24
+		orr	r6, r6, r7, push #8
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		bpl	.Lc2u_3cpy8lp
+
+.Lc2u_3rem8lp:	tst	ip, #8
+		movne	r3, r7, pull #24
+		ldmneia	r1!, {r4, r7}
+		orrne	r3, r3, r4, push #8
+		movne	r4, r4, pull #24
+		orrne	r4, r4, r7, push #8
+		stmneia	r0!, {r3 - r4}			@ Shouldnt fault
+		tst	ip, #4
+		movne	r3, r7, pull #24
+		ldrne	r7, [r1], #4
+		orrne	r3, r3, r7, push #8
+	TUSER(	strne) r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.Lc2u_3fupi
+.Lc2u_3nowords:	mov	r3, r7, get_byte_3
+		teq	ip, #0
+		beq	.Lc2u_finished
+		cmp	ip, #2
+USER(	TUSER(	strb)	r3, [r0], #1)			@ May fault
+		ldrgeb	r3, [r1], #1
+USER(	TUSER(	strgeb) r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #0
+USER(	TUSER(	strgtb) r3, [r0], #1)			@ May fault
+		b	.Lc2u_finished
+ENDPROC(__copy_to_user)
+
+		.pushsection .fixup,"ax"
+		.align	0
+9001:		ldmfd	sp!, {r0, r4 - r7, pc}
+		.popsection
+
+/* Prototype: unsigned long __copy_from_user(void *to,const void *from,unsigned long n);
+ * Purpose  : copy a block from user memory to kernel memory
+ * Params   : to   - kernel memory
+ *          : from - user memory
+ *          : n    - number of bytes to copy
+ * Returns  : Number of bytes NOT copied.
+ */
+.Lcfu_dest_not_aligned:
+		rsb	ip, ip, #4
+		cmp	ip, #2
+USER(	TUSER(	ldrb)	r3, [r1], #1)			@ May fault
+		strb	r3, [r0], #1
+USER(	TUSER(	ldrgeb) r3, [r1], #1)			@ May fault
+		strgeb	r3, [r0], #1
+USER(	TUSER(	ldrgtb) r3, [r1], #1)			@ May fault
+		strgtb	r3, [r0], #1
+		sub	r2, r2, ip
+		b	.Lcfu_dest_aligned
+
+ENTRY(__copy_from_user)
+		stmfd	sp!, {r0, r2, r4 - r7, lr}
+		cmp	r2, #4
+		blt	.Lcfu_not_enough
+		ands	ip, r0, #3
+		bne	.Lcfu_dest_not_aligned
+.Lcfu_dest_aligned:
+		ands	ip, r1, #3
+		bne	.Lcfu_src_not_aligned
+
+/*
+ * Seeing as there has to be at least 8 bytes to copy, we can
+ * copy one word, and force a user-mode page fault...
+ */
+
+.Lcfu_0fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lcfu_0nowords
+USER(	TUSER(	ldr)	r3, [r1], #4)
+		str	r3, [r0], #4
+		mov	ip, r1, lsl #32 - PAGE_SHIFT	@ On each page, use a ld/st??t instruction
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lcfu_0fupi
+/*
+ * ip = max no. of bytes to copy before needing another "strt" insn
+ */
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #32
+		blt	.Lcfu_0rem8lp
+
+.Lcfu_0cpy8lp:	ldmia	r1!, {r3 - r6}			@ Shouldnt fault
+		stmia	r0!, {r3 - r6}
+		ldmia	r1!, {r3 - r6}			@ Shouldnt fault
+		subs	ip, ip, #32
+		stmia	r0!, {r3 - r6}
+		bpl	.Lcfu_0cpy8lp
+
+.Lcfu_0rem8lp:	cmn	ip, #16
+		ldmgeia	r1!, {r3 - r6}			@ Shouldnt fault
+		stmgeia	r0!, {r3 - r6}
+		tst	ip, #8
+		ldmneia	r1!, {r3 - r4}			@ Shouldnt fault
+		stmneia	r0!, {r3 - r4}
+		tst	ip, #4
+	TUSER(	ldrne) r3, [r1], #4			@ Shouldnt fault
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.Lcfu_0fupi
+.Lcfu_0nowords:	teq	ip, #0
+		beq	.Lcfu_finished
+.Lcfu_nowords:	cmp	ip, #2
+USER(	TUSER(	ldrb)	r3, [r1], #1)			@ May fault
+		strb	r3, [r0], #1
+USER(	TUSER(	ldrgeb) r3, [r1], #1)			@ May fault
+		strgeb	r3, [r0], #1
+USER(	TUSER(	ldrgtb) r3, [r1], #1)			@ May fault
+		strgtb	r3, [r0], #1
+		b	.Lcfu_finished
+
+.Lcfu_not_enough:
+		movs	ip, r2
+		bne	.Lcfu_nowords
+.Lcfu_finished:	mov	r0, #0
+		add	sp, sp, #8
+		ldmfd	sp!, {r4 - r7, pc}
+
+.Lcfu_src_not_aligned:
+		bic	r1, r1, #3
+USER(	TUSER(	ldr)	r7, [r1], #4)			@ May fault
+		cmp	ip, #2
+		bgt	.Lcfu_3fupi
+		beq	.Lcfu_2fupi
+.Lcfu_1fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lcfu_1nowords
+		mov	r3, r7, pull #8
+USER(	TUSER(	ldr)	r7, [r1], #4)			@ May fault
+		orr	r3, r3, r7, push #24
+		str	r3, [r0], #4
+		mov	ip, r1, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lcfu_1fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.Lcfu_1rem8lp
+
+.Lcfu_1cpy8lp:	mov	r3, r7, pull #8
+		ldmia	r1!, {r4 - r7}			@ Shouldnt fault
+		subs	ip, ip, #16
+		orr	r3, r3, r4, push #24
+		mov	r4, r4, pull #8
+		orr	r4, r4, r5, push #24
+		mov	r5, r5, pull #8
+		orr	r5, r5, r6, push #24
+		mov	r6, r6, pull #8
+		orr	r6, r6, r7, push #24
+		stmia	r0!, {r3 - r6}
+		bpl	.Lcfu_1cpy8lp
+
+.Lcfu_1rem8lp:	tst	ip, #8
+		movne	r3, r7, pull #8
+		ldmneia	r1!, {r4, r7}			@ Shouldnt fault
+		orrne	r3, r3, r4, push #24
+		movne	r4, r4, pull #8
+		orrne	r4, r4, r7, push #24
+		stmneia	r0!, {r3 - r4}
+		tst	ip, #4
+		movne	r3, r7, pull #8
+USER(	TUSER(	ldrne) r7, [r1], #4)			@ May fault
+		orrne	r3, r3, r7, push #24
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.Lcfu_1fupi
+.Lcfu_1nowords:	mov	r3, r7, get_byte_1
+		teq	ip, #0
+		beq	.Lcfu_finished
+		cmp	ip, #2
+		strb	r3, [r0], #1
+		movge	r3, r7, get_byte_2
+		strgeb	r3, [r0], #1
+		movgt	r3, r7, get_byte_3
+		strgtb	r3, [r0], #1
+		b	.Lcfu_finished
+
+.Lcfu_2fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lcfu_2nowords
+		mov	r3, r7, pull #16
+USER(	TUSER(	ldr)	r7, [r1], #4)			@ May fault
+		orr	r3, r3, r7, push #16
+		str	r3, [r0], #4
+		mov	ip, r1, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lcfu_2fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.Lcfu_2rem8lp
+
+
+.Lcfu_2cpy8lp:	mov	r3, r7, pull #16
+		ldmia	r1!, {r4 - r7}			@ Shouldnt fault
+		subs	ip, ip, #16
+		orr	r3, r3, r4, push #16
+		mov	r4, r4, pull #16
+		orr	r4, r4, r5, push #16
+		mov	r5, r5, pull #16
+		orr	r5, r5, r6, push #16
+		mov	r6, r6, pull #16
+		orr	r6, r6, r7, push #16
+		stmia	r0!, {r3 - r6}
+		bpl	.Lcfu_2cpy8lp
+
+.Lcfu_2rem8lp:	tst	ip, #8
+		movne	r3, r7, pull #16
+		ldmneia	r1!, {r4, r7}			@ Shouldnt fault
+		orrne	r3, r3, r4, push #16
+		movne	r4, r4, pull #16
+		orrne	r4, r4, r7, push #16
+		stmneia	r0!, {r3 - r4}
+		tst	ip, #4
+		movne	r3, r7, pull #16
+USER(	TUSER(	ldrne) r7, [r1], #4)			@ May fault
+		orrne	r3, r3, r7, push #16
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.Lcfu_2fupi
+.Lcfu_2nowords:	mov	r3, r7, get_byte_2
+		teq	ip, #0
+		beq	.Lcfu_finished
+		cmp	ip, #2
+		strb	r3, [r0], #1
+		movge	r3, r7, get_byte_3
+		strgeb	r3, [r0], #1
+USER(	TUSER(	ldrgtb) r3, [r1], #0)			@ May fault
+		strgtb	r3, [r0], #1
+		b	.Lcfu_finished
+
+.Lcfu_3fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.Lcfu_3nowords
+		mov	r3, r7, pull #24
+USER(	TUSER(	ldr)	r7, [r1], #4)			@ May fault
+		orr	r3, r3, r7, push #8
+		str	r3, [r0], #4
+		mov	ip, r1, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.Lcfu_3fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.Lcfu_3rem8lp
+
+.Lcfu_3cpy8lp:	mov	r3, r7, pull #24
+		ldmia	r1!, {r4 - r7}			@ Shouldnt fault
+		orr	r3, r3, r4, push #8
+		mov	r4, r4, pull #24
+		orr	r4, r4, r5, push #8
+		mov	r5, r5, pull #24
+		orr	r5, r5, r6, push #8
+		mov	r6, r6, pull #24
+		orr	r6, r6, r7, push #8
+		stmia	r0!, {r3 - r6}
+		subs	ip, ip, #16
+		bpl	.Lcfu_3cpy8lp
+
+.Lcfu_3rem8lp:	tst	ip, #8
+		movne	r3, r7, pull #24
+		ldmneia	r1!, {r4, r7}			@ Shouldnt fault
+		orrne	r3, r3, r4, push #8
+		movne	r4, r4, pull #24
+		orrne	r4, r4, r7, push #8
+		stmneia	r0!, {r3 - r4}
+		tst	ip, #4
+		movne	r3, r7, pull #24
+USER(	TUSER(	ldrne) r7, [r1], #4)			@ May fault
+		orrne	r3, r3, r7, push #8
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.Lcfu_3fupi
+.Lcfu_3nowords:	mov	r3, r7, get_byte_3
+		teq	ip, #0
+		beq	.Lcfu_finished
+		cmp	ip, #2
+		strb	r3, [r0], #1
+USER(	TUSER(	ldrgeb) r3, [r1], #1)			@ May fault
+		strgeb	r3, [r0], #1
+USER(	TUSER(	ldrgtb) r3, [r1], #1)			@ May fault
+		strgtb	r3, [r0], #1
+		b	.Lcfu_finished
+ENDPROC(__copy_from_user)
+
+		.pushsection .fixup,"ax"
+		.align	0
+		/*
+		 * We took an exception.  r0 contains a pointer to
+		 * the byte not copied.
+		 */
+9001:		ldr	r2, [sp], #4			@ void *to
+		sub	r2, r0, r2			@ bytes copied
+		ldr	r1, [sp], #4			@ unsigned long count
+		subs	r4, r1, r2			@ bytes left to copy
+		movne	r1, r4
+		blne	__memzero
+		mov	r0, r4
+		ldmfd	sp!, {r4 - r7, pc}
+		.popsection
+