10 files changed, 431 insertions, 411 deletions
diff --git a/arch/arm/mm/copypage-feroceon.S b/arch/arm/mm/copypage-feroceon.S
deleted file mode 100644
index 7eb0d320d24..00000000000
--- a/arch/arm/mm/copypage-feroceon.S
+++ /dev/null
@@ -1,95 +0,0 @@
-/*
- *  linux/arch/arm/lib/copypage-feroceon.S
- *
- *  Copyright (C) 2008 Marvell Semiconductors
- *
- * 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.
- *
- * This handles copy_user_page and clear_user_page on Feroceon
- * more optimally than the generic implementations.
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/asm-offsets.h>
-
-	.text
-	.align	5
-
-ENTRY(feroceon_copy_user_page)
-	stmfd	sp!, {r4-r9, lr}
-	mov	ip, #PAGE_SZ
-1:	mov	lr, r1
-	ldmia	r1!, {r2 - r9}
-	pld	[lr, #32]
-	pld	[lr, #64]
-	pld	[lr, #96]
-	pld	[lr, #128]
-	pld	[lr, #160]
-	pld	[lr, #192]
-	pld	[lr, #224]
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	ldmia	r1!, {r2 - r9}
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	stmia	r0, {r2 - r9}
-	subs	ip, ip, #(32 * 8)
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	bne	1b
-	mcr	p15, 0, ip, c7, c10, 4		@ drain WB
-	ldmfd	sp!, {r4-r9, pc}
-
-	.align	5
-
-ENTRY(feroceon_clear_user_page)
-	stmfd	sp!, {r4-r7, lr}
-	mov	r1, #PAGE_SZ/32
-	mov	r2, #0
-	mov	r3, #0
-	mov	r4, #0
-	mov	r5, #0
-	mov	r6, #0
-	mov	r7, #0
-	mov	ip, #0
-	mov	lr, #0
-1:	stmia	r0, {r2-r7, ip, lr}
-	subs	r1, r1, #1
-	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line
-	add	r0, r0, #32
-	bne	1b
-	mcr	p15, 0, r1, c7, c10, 4		@ drain WB
-	ldmfd	sp!, {r4-r7, pc}
-
-	__INITDATA
-
-	.type	feroceon_user_fns, #object
-ENTRY(feroceon_user_fns)
-	.long	feroceon_clear_user_page
-	.long	feroceon_copy_user_page
-	.size	feroceon_user_fns, . - feroceon_user_fns
diff --git a/arch/arm/mm/copypage-feroceon.c b/arch/arm/mm/copypage-feroceon.c
new file mode 100644
index 00000000000..c8347670ab0
--- /dev/null
+++ b/arch/arm/mm/copypage-feroceon.c
@@ -0,0 +1,100 @@
+/*
+ *  linux/arch/arm/mm/copypage-feroceon.S
+ *
+ *  Copyright (C) 2008 Marvell Semiconductors
+ *
+ * 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.
+ *
+ * This handles copy_user_page and clear_user_page on Feroceon
+ * more optimally than the generic implementations.
+ */
+#include <linux/init.h>
+
+#include <asm/page.h>
+
+void __attribute__((naked))
+feroceon_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	asm("\
+	stmfd	sp!, {r4-r9, lr}		\n\
+	mov	ip, %0				\n\
+1:	mov	lr, r1				\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	pld	[lr, #32]			\n\
+	pld	[lr, #64]			\n\
+	pld	[lr, #96]			\n\
+	pld	[lr, #128]			\n\
+	pld	[lr, #160]			\n\
+	pld	[lr, #192]			\n\
+	pld	[lr, #224]			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	ldmia	r1!, {r2 - r9}			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	stmia	r0, {r2 - r9}			\n\
+	subs	ip, ip, #(32 * 8)		\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	bne	1b				\n\
+	mcr	p15, 0, ip, c7, c10, 4		@ drain WB\n\
+	ldmfd	sp!, {r4-r9, pc}"
+	:
+	: "I" (PAGE_SIZE));
+}
+
+void __attribute__((naked))
+feroceon_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm("\
+	stmfd	sp!, {r4-r7, lr}		\n\
+	mov	r1, %0				\n\
+	mov	r2, #0				\n\
+	mov	r3, #0				\n\
+	mov	r4, #0				\n\
+	mov	r5, #0				\n\
+	mov	r6, #0				\n\
+	mov	r7, #0				\n\
+	mov	ip, #0				\n\
+	mov	lr, #0				\n\
+1:	stmia	r0, {r2-r7, ip, lr}		\n\
+	subs	r1, r1, #1			\n\
+	mcr	p15, 0, r0, c7, c14, 1		@ clean and invalidate D line\n\
+	add	r0, r0, #32			\n\
+	bne	1b				\n\
+	mcr	p15, 0, r1, c7, c10, 4		@ drain WB\n\
+	ldmfd	sp!, {r4-r7, pc}"
+	:
+	: "I" (PAGE_SIZE / 32));
+}
+
+struct cpu_user_fns feroceon_user_fns __initdata = {
+	.cpu_clear_user_page	= feroceon_clear_user_page,
+	.cpu_copy_user_page	= feroceon_copy_user_page,
+};
+
diff --git a/arch/arm/mm/copypage-v3.S b/arch/arm/mm/copypage-v3.S
deleted file mode 100644
index 2ee394b11bc..00000000000
--- a/arch/arm/mm/copypage-v3.S
+++ /dev/null
@@ -1,67 +0,0 @@
-/*
- *  linux/arch/arm/lib/copypage.S
- *
- *  Copyright (C) 1995-1999 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.
- *
- *  ASM optimised string functions
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/assembler.h>
-#include <asm/asm-offsets.h>
-
-		.text
-		.align	5
-/*
- * ARMv3 optimised copy_user_page
- *
- * FIXME: do we need to handle cache stuff...
- */
-ENTRY(v3_copy_user_page)
-	stmfd	sp!, {r4, lr}			@	2
-	mov	r2, #PAGE_SZ/64			@	1
-	ldmia	r1!, {r3, r4, ip, lr}		@	4+1
-1:	stmia	r0!, {r3, r4, ip, lr}		@	4
-	ldmia	r1!, {r3, r4, ip, lr}		@	4+1
-	stmia	r0!, {r3, r4, ip, lr}		@	4
-	ldmia	r1!, {r3, r4, ip, lr}		@	4+1
-	stmia	r0!, {r3, r4, ip, lr}		@	4
-	ldmia	r1!, {r3, r4, ip, lr}		@	4
-	subs	r2, r2, #1			@	1
-	stmia	r0!, {r3, r4, ip, lr}		@	4
-	ldmneia	r1!, {r3, r4, ip, lr}		@	4
-	bne	1b				@	1
-	ldmfd	sp!, {r4, pc}			@	3
-
-	.align	5
-/*
- * ARMv3 optimised clear_user_page
- *
- * FIXME: do we need to handle cache stuff...
- */
-ENTRY(v3_clear_user_page)
-	str	lr, [sp, #-4]!
-	mov	r1, #PAGE_SZ/64			@ 1
-	mov	r2, #0				@ 1
-	mov	r3, #0				@ 1
-	mov	ip, #0				@ 1
-	mov	lr, #0				@ 1
-1:	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	subs	r1, r1, #1			@ 1
-	bne	1b				@ 1
-	ldr	pc, [sp], #4
-
-	__INITDATA
-
-	.type	v3_user_fns, #object
-ENTRY(v3_user_fns)
-	.long	v3_clear_user_page
-	.long	v3_copy_user_page
-	.size	v3_user_fns, . - v3_user_fns
diff --git a/arch/arm/mm/copypage-v3.c b/arch/arm/mm/copypage-v3.c
new file mode 100644
index 00000000000..184911089e6
--- /dev/null
+++ b/arch/arm/mm/copypage-v3.c
@@ -0,0 +1,69 @@
+/*
+ *  linux/arch/arm/mm/copypage-v3.c
+ *
+ *  Copyright (C) 1995-1999 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/init.h>
+
+#include <asm/page.h>
+
+/*
+ * ARMv3 optimised copy_user_page
+ *
+ * FIXME: do we need to handle cache stuff...
+ */
+void __attribute__((naked))
+v3_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	asm("\n\
+	stmfd	sp!, {r4, lr}			@	2\n\
+	mov	r2, %2				@	1\n\
+	ldmia	%0!, {r3, r4, ip, lr}		@	4+1\n\
+1:	stmia	%1!, {r3, r4, ip, lr}		@	4\n\
+	ldmia	%0!, {r3, r4, ip, lr}		@	4+1\n\
+	stmia	%1!, {r3, r4, ip, lr}		@	4\n\
+	ldmia	%0!, {r3, r4, ip, lr}		@	4+1\n\
+	stmia	%1!, {r3, r4, ip, lr}		@	4\n\
+	ldmia	%0!, {r3, r4, ip, lr}		@	4\n\
+	subs	r2, r2, #1			@	1\n\
+	stmia	%1!, {r3, r4, ip, lr}		@	4\n\
+	ldmneia	%0!, {r3, r4, ip, lr}		@	4\n\
+	bne	1b				@	1\n\
+	ldmfd	sp!, {r4, pc}			@	3"
+	:
+	: "r" (kfrom), "r" (kto), "I" (PAGE_SIZE / 64));
+}
+
+/*
+ * ARMv3 optimised clear_user_page
+ *
+ * FIXME: do we need to handle cache stuff...
+ */
+void __attribute__((naked)) v3_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm("\n\
+	str	lr, [sp, #-4]!\n\
+	mov	r1, %1				@ 1\n\
+	mov	r2, #0				@ 1\n\
+	mov	r3, #0				@ 1\n\
+	mov	ip, #0				@ 1\n\
+	mov	lr, #0				@ 1\n\
+1:	stmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	%0!, {r2, r3, ip, lr}		@ 4\n\
+	subs	r1, r1, #1			@ 1\n\
+	bne	1b				@ 1\n\
+	ldr	pc, [sp], #4"
+	:
+	: "r" (kaddr), "I" (PAGE_SIZE / 64));
+}
+
+struct cpu_user_fns v3_user_fns __initdata = {
+	.cpu_clear_user_page	= v3_clear_user_page,
+	.cpu_copy_user_page	= v3_copy_user_page,
+};
diff --git a/arch/arm/mm/copypage-v4wb.S b/arch/arm/mm/copypage-v4wb.S
deleted file mode 100644
index 83117354b1c..00000000000
--- a/arch/arm/mm/copypage-v4wb.S
+++ /dev/null
@@ -1,79 +0,0 @@
-/*
- *  linux/arch/arm/lib/copypage.S
- *
- *  Copyright (C) 1995-1999 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.
- *
- *  ASM optimised string functions
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/asm-offsets.h>
-
-	.text
-	.align	5
-/*
- * ARMv4 optimised copy_user_page
- *
- * We flush the destination cache lines just before we write the data into the
- * corresponding address.  Since the Dcache is read-allocate, this removes the
- * Dcache aliasing issue.  The writes will be forwarded to the write buffer,
- * and merged as appropriate.
- *
- * Note: We rely on all ARMv4 processors implementing the "invalidate D line"
- * instruction.  If your processor does not supply this, you have to write your
- * own copy_user_page that does the right thing.
- */
-ENTRY(v4wb_copy_user_page)
-	stmfd	sp!, {r4, lr}			@ 2
-	mov	r2, #PAGE_SZ/64			@ 1
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4
-1:	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4+1
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4
-	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4
-	subs	r2, r2, #1			@ 1
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmneia	r1!, {r3, r4, ip, lr}		@ 4
-	bne	1b				@ 1
-	mcr	p15, 0, r1, c7, c10, 4		@ 1   drain WB
-	ldmfd	 sp!, {r4, pc}			@ 3
-
-	.align	5
-/*
- * ARMv4 optimised clear_user_page
- *
- * Same story as above.
- */
-ENTRY(v4wb_clear_user_page)
-	str	lr, [sp, #-4]!
-	mov	r1, #PAGE_SZ/64			@ 1
-	mov	r2, #0				@ 1
-	mov	r3, #0				@ 1
-	mov	ip, #0				@ 1
-	mov	lr, #0				@ 1
-1:	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	subs	r1, r1, #1			@ 1
-	bne	1b				@ 1
-	mcr	p15, 0, r1, c7, c10, 4		@ 1   drain WB
-	ldr	pc, [sp], #4
-
-	__INITDATA
-
-	.type	v4wb_user_fns, #object
-ENTRY(v4wb_user_fns)
-	.long	v4wb_clear_user_page
-	.long	v4wb_copy_user_page
-	.size	v4wb_user_fns, . - v4wb_user_fns
diff --git a/arch/arm/mm/copypage-v4wb.c b/arch/arm/mm/copypage-v4wb.c
new file mode 100644
index 00000000000..23021082296
--- /dev/null
+++ b/arch/arm/mm/copypage-v4wb.c
@@ -0,0 +1,83 @@
+/*
+ *  linux/arch/arm/mm/copypage-v4wb.c
+ *
+ *  Copyright (C) 1995-1999 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/init.h>
+
+#include <asm/page.h>
+
+/*
+ * ARMv4 optimised copy_user_page
+ *
+ * We flush the destination cache lines just before we write the data into the
+ * corresponding address.  Since the Dcache is read-allocate, this removes the
+ * Dcache aliasing issue.  The writes will be forwarded to the write buffer,
+ * and merged as appropriate.
+ *
+ * Note: We rely on all ARMv4 processors implementing the "invalidate D line"
+ * instruction.  If your processor does not supply this, you have to write your
+ * own copy_user_page that does the right thing.
+ */
+void __attribute__((naked))
+v4wb_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	asm("\
+	stmfd	sp!, {r4, lr}			@ 2\n\
+	mov	r2, %0				@ 1\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4\n\
+1:	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4+1\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4\n\
+	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4\n\
+	subs	r2, r2, #1			@ 1\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmneia	r1!, {r3, r4, ip, lr}		@ 4\n\
+	bne	1b				@ 1\n\
+	mcr	p15, 0, r1, c7, c10, 4		@ 1   drain WB\n\
+	ldmfd	 sp!, {r4, pc}			@ 3"
+	:
+	: "I" (PAGE_SIZE / 64));
+}
+
+/*
+ * ARMv4 optimised clear_user_page
+ *
+ * Same story as above.
+ */
+void __attribute__((naked))
+v4wb_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm("\
+	str	lr, [sp, #-4]!\n\
+	mov	r1, %0				@ 1\n\
+	mov	r2, #0				@ 1\n\
+	mov	r3, #0				@ 1\n\
+	mov	ip, #0				@ 1\n\
+	mov	lr, #0				@ 1\n\
+1:	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	mcr	p15, 0, r0, c7, c6, 1		@ 1   invalidate D line\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	subs	r1, r1, #1			@ 1\n\
+	bne	1b				@ 1\n\
+	mcr	p15, 0, r1, c7, c10, 4		@ 1   drain WB\n\
+	ldr	pc, [sp], #4"
+	:
+	: "I" (PAGE_SIZE / 64));
+}
+
+struct cpu_user_fns v4wb_user_fns __initdata = {
+	.cpu_clear_user_page	= v4wb_clear_user_page,
+	.cpu_copy_user_page	= v4wb_copy_user_page,
+};
diff --git a/arch/arm/mm/copypage-v4wt.S b/arch/arm/mm/copypage-v4wt.S
deleted file mode 100644
index e1f2af28d54..00000000000
--- a/arch/arm/mm/copypage-v4wt.S
+++ /dev/null
@@ -1,73 +0,0 @@
-/*
- *  linux/arch/arm/lib/copypage-v4.S
- *
- *  Copyright (C) 1995-1999 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.
- *
- *  ASM optimised string functions
- *
- *  This is for CPUs with a writethrough cache and 'flush ID cache' is
- *  the only supported cache operation.
- */
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/asm-offsets.h>
-
-	.text
-	.align	5
-/*
- * ARMv4 optimised copy_user_page
- *
- * Since we have writethrough caches, we don't have to worry about
- * dirty data in the cache.  However, we do have to ensure that
- * subsequent reads are up to date.
- */
-ENTRY(v4wt_copy_user_page)
-	stmfd	sp!, {r4, lr}			@ 2
-	mov	r2, #PAGE_SZ/64			@ 1
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4
-1:	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4+1
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmia	r1!, {r3, r4, ip, lr}		@ 4
-	subs	r2, r2, #1			@ 1
-	stmia	r0!, {r3, r4, ip, lr}		@ 4
-	ldmneia	r1!, {r3, r4, ip, lr}		@ 4
-	bne	1b				@ 1
-	mcr	p15, 0, r2, c7, c7, 0		@ flush ID cache
-	ldmfd	sp!, {r4, pc}			@ 3
-
-	.align	5
-/*
- * ARMv4 optimised clear_user_page
- *
- * Same story as above.
- */
-ENTRY(v4wt_clear_user_page)
-	str	lr, [sp, #-4]!
-	mov	r1, #PAGE_SZ/64			@ 1
-	mov	r2, #0				@ 1
-	mov	r3, #0				@ 1
-	mov	ip, #0				@ 1
-	mov	lr, #0				@ 1
-1:	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	stmia	r0!, {r2, r3, ip, lr}		@ 4
-	subs	r1, r1, #1			@ 1
-	bne	1b				@ 1
-	mcr	p15, 0, r2, c7, c7, 0		@ flush ID cache
-	ldr	pc, [sp], #4
-
-	__INITDATA
-
-	.type	v4wt_user_fns, #object
-ENTRY(v4wt_user_fns)
-	.long	v4wt_clear_user_page
-	.long	v4wt_copy_user_page
-	.size	v4wt_user_fns, . - v4wt_user_fns
diff --git a/arch/arm/mm/copypage-v4wt.c b/arch/arm/mm/copypage-v4wt.c
new file mode 100644
index 00000000000..d8ef39503ff
--- /dev/null
+++ b/arch/arm/mm/copypage-v4wt.c
@@ -0,0 +1,77 @@
+/*
+ *  linux/arch/arm/mm/copypage-v4wt.S
+ *
+ *  Copyright (C) 1995-1999 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.
+ *
+ *  This is for CPUs with a writethrough cache and 'flush ID cache' is
+ *  the only supported cache operation.
+ */
+#include <linux/init.h>
+
+#include <asm/page.h>
+
+/*
+ * ARMv4 optimised copy_user_page
+ *
+ * Since we have writethrough caches, we don't have to worry about
+ * dirty data in the cache.  However, we do have to ensure that
+ * subsequent reads are up to date.
+ */
+void __attribute__((naked))
+v4wt_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	asm("\
+	stmfd	sp!, {r4, lr}			@ 2\n\
+	mov	r2, %0				@ 1\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4\n\
+1:	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4+1\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmia	r1!, {r3, r4, ip, lr}		@ 4\n\
+	subs	r2, r2, #1			@ 1\n\
+	stmia	r0!, {r3, r4, ip, lr}		@ 4\n\
+	ldmneia	r1!, {r3, r4, ip, lr}		@ 4\n\
+	bne	1b				@ 1\n\
+	mcr	p15, 0, r2, c7, c7, 0		@ flush ID cache\n\
+	ldmfd	sp!, {r4, pc}			@ 3"
+	:
+	: "I" (PAGE_SIZE / 64));
+}
+
+/*
+ * ARMv4 optimised clear_user_page
+ *
+ * Same story as above.
+ */
+void __attribute__((naked))
+v4wt_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm("\
+	str	lr, [sp, #-4]!\n\
+	mov	r1, %0				@ 1\n\
+	mov	r2, #0				@ 1\n\
+	mov	r3, #0				@ 1\n\
+	mov	ip, #0				@ 1\n\
+	mov	lr, #0				@ 1\n\
+1:	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	stmia	r0!, {r2, r3, ip, lr}		@ 4\n\
+	subs	r1, r1, #1			@ 1\n\
+	bne	1b				@ 1\n\
+	mcr	p15, 0, r2, c7, c7, 0		@ flush ID cache\n\
+	ldr	pc, [sp], #4"
+	:
+	: "I" (PAGE_SIZE / 64));
+}
+
+struct cpu_user_fns v4wt_user_fns __initdata = {
+	.cpu_clear_user_page	= v4wt_clear_user_page,
+	.cpu_copy_user_page	= v4wt_copy_user_page,
+};
diff --git a/arch/arm/mm/copypage-xsc3.S b/arch/arm/mm/copypage-xsc3.S
deleted file mode 100644
index 9a2cb4332b4..00000000000
--- a/arch/arm/mm/copypage-xsc3.S
+++ /dev/null
@@ -1,97 +0,0 @@
-/*
- *  linux/arch/arm/lib/copypage-xsc3.S
- *
- *  Copyright (C) 2004 Intel Corp.
- *
- * 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.
- *
- * Adapted for 3rd gen XScale core, no more mini-dcache
- * Author: Matt Gilbert (matthew.m.gilbert@intel.com)
- */
-
-#include <linux/linkage.h>
-#include <linux/init.h>
-#include <asm/asm-offsets.h>
-
-/*
- * General note:
- *  We don't really want write-allocate cache behaviour for these functions
- *  since that will just eat through 8K of the cache.
- */
-
-	.text
-	.align	5
-/*
- * XSC3 optimised copy_user_page
- *  r0 = destination
- *  r1 = source
- *  r2 = virtual user address of ultimate destination page
- *
- * The source page may have some clean entries in the cache already, but we
- * can safely ignore them - break_cow() will flush them out of the cache
- * if we eventually end up using our copied page.
- *
- */
-ENTRY(xsc3_mc_copy_user_page)
-	stmfd	sp!, {r4, r5, lr}
-	mov	lr, #PAGE_SZ/64-1
-
-	pld	[r1, #0]
-	pld	[r1, #32]
-1:	pld	[r1, #64]
-	pld	[r1, #96]
-
-2:	ldrd	r2, [r1], #8
-	mov	ip, r0
-	ldrd	r4, [r1], #8
-	mcr	p15, 0, ip, c7, c6, 1		@ invalidate
-	strd	r2, [r0], #8
-	ldrd	r2, [r1], #8
-	strd	r4, [r0], #8
-	ldrd	r4, [r1], #8
-	strd	r2, [r0], #8
-	strd	r4, [r0], #8
-	ldrd	r2, [r1], #8
-	mov	ip, r0
-	ldrd	r4, [r1], #8
-	mcr	p15, 0, ip, c7, c6, 1		@ invalidate
-	strd	r2, [r0], #8
-	ldrd	r2, [r1], #8
-	subs	lr, lr, #1
-	strd	r4, [r0], #8
-	ldrd	r4, [r1], #8
-	strd	r2, [r0], #8
-	strd	r4, [r0], #8
-	bgt	1b
-	beq	2b
-
-	ldmfd	sp!, {r4, r5, pc}
-
-	.align	5
-/*
- * XScale optimised clear_user_page
- *  r0 = destination
- *  r1 = virtual user address of ultimate destination page
- */
-ENTRY(xsc3_mc_clear_user_page)
-	mov	r1, #PAGE_SZ/32
-	mov	r2, #0
-	mov	r3, #0
-1:	mcr	p15, 0, r0, c7, c6, 1		@ invalidate line
-	strd	r2, [r0], #8
-	strd	r2, [r0], #8
-	strd	r2, [r0], #8
-	strd	r2, [r0], #8
-	subs	r1, r1, #1
-	bne	1b
-	mov	pc, lr
-
-	__INITDATA
-
-	.type	xsc3_mc_user_fns, #object
-ENTRY(xsc3_mc_user_fns)
-	.long	xsc3_mc_clear_user_page
-	.long	xsc3_mc_copy_user_page
-	.size	xsc3_mc_user_fns, . - xsc3_mc_user_fns
diff --git a/arch/arm/mm/copypage-xsc3.c b/arch/arm/mm/copypage-xsc3.c
new file mode 100644
index 00000000000..51ed502e577
--- /dev/null
+++ b/arch/arm/mm/copypage-xsc3.c
@@ -0,0 +1,102 @@
+/*
+ *  linux/arch/arm/mm/copypage-xsc3.S
+ *
+ *  Copyright (C) 2004 Intel Corp.
+ *
+ * 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.
+ *
+ * Adapted for 3rd gen XScale core, no more mini-dcache
+ * Author: Matt Gilbert (matthew.m.gilbert@intel.com)
+ */
+#include <linux/init.h>
+
+#include <asm/page.h>
+
+/*
+ * General note:
+ *  We don't really want write-allocate cache behaviour for these functions
+ *  since that will just eat through 8K of the cache.
+ */
+
+/*
+ * XSC3 optimised copy_user_page
+ *  r0 = destination
+ *  r1 = source
+ *  r2 = virtual user address of ultimate destination page
+ *
+ * The source page may have some clean entries in the cache already, but we
+ * can safely ignore them - break_cow() will flush them out of the cache
+ * if we eventually end up using our copied page.
+ *
+ */
+void __attribute__((naked))
+xsc3_mc_copy_user_page(void *kto, const void *kfrom, unsigned long vaddr)
+{
+	asm("\
+	stmfd	sp!, {r4, r5, lr}		\n\
+	mov	lr, %0				\n\
+						\n\
+	pld	[r1, #0]			\n\
+	pld	[r1, #32]			\n\
+1:	pld	[r1, #64]			\n\
+	pld	[r1, #96]			\n\
+						\n\
+2:	ldrd	r2, [r1], #8			\n\
+	mov	ip, r0				\n\
+	ldrd	r4, [r1], #8			\n\
+	mcr	p15, 0, ip, c7, c6, 1		@ invalidate\n\
+	strd	r2, [r0], #8			\n\
+	ldrd	r2, [r1], #8			\n\
+	strd	r4, [r0], #8			\n\
+	ldrd	r4, [r1], #8			\n\
+	strd	r2, [r0], #8			\n\
+	strd	r4, [r0], #8			\n\
+	ldrd	r2, [r1], #8			\n\
+	mov	ip, r0				\n\
+	ldrd	r4, [r1], #8			\n\
+	mcr	p15, 0, ip, c7, c6, 1		@ invalidate\n\
+	strd	r2, [r0], #8			\n\
+	ldrd	r2, [r1], #8			\n\
+	subs	lr, lr, #1			\n\
+	strd	r4, [r0], #8			\n\
+	ldrd	r4, [r1], #8			\n\
+	strd	r2, [r0], #8			\n\
+	strd	r4, [r0], #8			\n\
+	bgt	1b				\n\
+	beq	2b				\n\
+						\n\
+	ldmfd	sp!, {r4, r5, pc}"
+	:
+	: "I" (PAGE_SIZE / 64 - 1));
+}
+
+/*
+ * XScale optimised clear_user_page
+ *  r0 = destination
+ *  r1 = virtual user address of ultimate destination page
+ */
+void __attribute__((naked))
+xsc3_mc_clear_user_page(void *kaddr, unsigned long vaddr)
+{
+	asm("\
+	mov	r1, %0				\n\
+	mov	r2, #0				\n\
+	mov	r3, #0				\n\
+1:	mcr	p15, 0, r0, c7, c6, 1		@ invalidate line\n\
+	strd	r2, [r0], #8			\n\
+	strd	r2, [r0], #8			\n\
+	strd	r2, [r0], #8			\n\
+	strd	r2, [r0], #8			\n\
+	subs	r1, r1, #1			\n\
+	bne	1b				\n\
+	mov	pc, lr"
+	:
+	: "I" (PAGE_SIZE / 32));
+}
+
+struct cpu_user_fns xsc3_mc_user_fns __initdata = {
+	.cpu_clear_user_page	= xsc3_mc_clear_user_page,
+	.cpu_copy_user_page	= xsc3_mc_copy_user_page,
+};