diff --git a/arch/arm26/lib/Makefile b/arch/arm26/lib/Makefile
new file mode 100644
index 0000000..6df2b79
--- /dev/null
+++ b/arch/arm26/lib/Makefile
@@ -0,0 +1,26 @@
+#
+# linux/arch/arm26/lib/Makefile
+#
+# Copyright (C) 1995-2000 Russell King
+#
+
+lib-y		:= backtrace.o changebit.o csumipv6.o csumpartial.o   \
+		   csumpartialcopy.o csumpartialcopyuser.o clearbit.o \
+		   copy_page.o delay.o findbit.o memchr.o memcpy.o    \
+		   memset.o memzero.o setbit.o                        \
+		   strchr.o strrchr.o testchangebit.o                 \
+		   testclearbit.o testsetbit.o getuser.o              \
+		   putuser.o ashldi3.o ashrdi3.o lshrdi3.o muldi3.o   \
+		   ucmpdi2.o udivdi3.o lib1funcs.o ecard.o io-acorn.o \
+		   floppydma.o io-readsb.o io-writesb.o io-writesl.o  \
+		   uaccess-kernel.o uaccess-user.o io-readsw.o        \
+		   io-writesw.o io-readsl.o ecard.o io-acorn.o        \
+		   floppydma.o
+
+lib-n		:=
+
+lib-$(CONFIG_VT)+= kbd.o
+
+csumpartialcopy.o: csumpartialcopygeneric.S
+csumpartialcopyuser.o: csumpartialcopygeneric.S
+
diff --git a/arch/arm26/lib/ashldi3.c b/arch/arm26/lib/ashldi3.c
new file mode 100644
index 0000000..130f5a8
--- /dev/null
+++ b/arch/arm26/lib/ashldi3.c
@@ -0,0 +1,61 @@
+/* More subroutines needed by GCC output code on some machines.  */
+/* Compile this one with gcc.  */
+/* Copyright (C) 1989, 92-98, 1999 Free Software Foundation, Inc.
+
+This file is part of GNU CC.
+
+GNU CC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GNU CC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GNU CC; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* support functions required by the kernel. based on code from gcc-2.95.3 */
+/* I Molton	29/07/01 */
+
+#include "gcclib.h"
+
+DItype
+__ashldi3 (DItype u, word_type b)
+{
+  DIunion w;
+  word_type bm;
+  DIunion uu;
+
+  if (b == 0)
+    return u;
+
+  uu.ll = u;
+
+  bm = (sizeof (SItype) * BITS_PER_UNIT) - b;
+  if (bm <= 0)
+    {
+      w.s.low = 0;
+      w.s.high = (USItype)uu.s.low << -bm;
+    }
+  else
+    {
+      USItype carries = (USItype)uu.s.low >> bm;
+      w.s.low = (USItype)uu.s.low << b;
+      w.s.high = ((USItype)uu.s.high << b) | carries;
+    }
+
+  return w.ll;
+}
+
diff --git a/arch/arm26/lib/ashrdi3.c b/arch/arm26/lib/ashrdi3.c
new file mode 100644
index 0000000..71625d2
--- /dev/null
+++ b/arch/arm26/lib/ashrdi3.c
@@ -0,0 +1,61 @@
+/* More subroutines needed by GCC output code on some machines.  */
+/* Compile this one with gcc.  */
+/* Copyright (C) 1989, 92-98, 1999 Free Software Foundation, Inc.
+
+This file is part of GNU CC.
+
+GNU CC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GNU CC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GNU CC; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* support functions required by the kernel. based on code from gcc-2.95.3 */
+/* I Molton     29/07/01 */
+
+#include "gcclib.h"
+
+DItype
+__ashrdi3 (DItype u, word_type b)
+{
+  DIunion w;
+  word_type bm;
+  DIunion uu;
+
+  if (b == 0)
+    return u;
+
+  uu.ll = u;
+
+  bm = (sizeof (SItype) * BITS_PER_UNIT) - b;
+  if (bm <= 0)
+    {
+      /* w.s.high = 1..1 or 0..0 */
+      w.s.high = uu.s.high >> (sizeof (SItype) * BITS_PER_UNIT - 1);
+      w.s.low = uu.s.high >> -bm;
+    }
+  else
+    {
+      USItype carries = (USItype)uu.s.high << bm;
+      w.s.high = uu.s.high >> b;
+      w.s.low = ((USItype)uu.s.low >> b) | carries;
+    }
+
+  return w.ll;
+}
diff --git a/arch/arm26/lib/backtrace.S b/arch/arm26/lib/backtrace.S
new file mode 100644
index 0000000..d793fe4
--- /dev/null
+++ b/arch/arm26/lib/backtrace.S
@@ -0,0 +1,145 @@
+/*
+ *  linux/arch/arm26/lib/backtrace.S
+ *
+ *  Copyright (C) 1995, 1996 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/config.h>
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+		.text
+
+@ fp is 0 or stack frame
+
+#define frame	r4
+#define next	r5
+#define save	r6
+#define mask	r7
+#define offset	r8
+
+ENTRY(__backtrace)
+		mov	r1, #0x10
+		mov	r0, fp
+
+ENTRY(c_backtrace)
+
+#ifdef CONFIG_NO_FRAME_POINTER
+		mov	pc, lr
+#else
+
+		stmfd	sp!, {r4 - r8, lr}	@ Save an extra register so we have a location...
+		mov	mask, #0xfc000003
+		tst	mask, r0
+		movne	r0, #0
+		movs	frame, r0
+1:		moveq	r0, #-2
+		LOADREGS(eqfd, sp!, {r4 - r8, pc})
+
+2:		stmfd	sp!, {pc}		@ calculate offset of PC in STMIA instruction
+		ldr	r0, [sp], #4
+		adr	r1, 2b - 4
+		sub	offset, r0, r1
+
+3:		tst	frame, mask		@ Check for address exceptions...
+		bne	1b
+
+1001:		ldr	next, [frame, #-12]	@ get fp
+1002:		ldr	r2, [frame, #-4]	@ get lr
+1003:		ldr	r3, [frame, #0]		@ get pc
+		sub	save, r3, offset	@ Correct PC for prefetching
+		bic	save, save, mask
+1004:		ldr	r1, [save, #0]		@ get instruction at function
+		mov	r1, r1, lsr #10
+		ldr	r3, .Ldsi+4
+		teq	r1, r3
+		subeq	save, save, #4
+		adr	r0, .Lfe
+		mov	r1, save
+		bic	r2, r2, mask
+		bl	printk			@ print pc and link register
+
+		ldr	r0, [frame, #-8]	@ get sp
+		sub	r0, r0, #4
+1005:		ldr	r1, [save, #4]		@ get instruction at function+4
+		mov	r3, r1, lsr #10
+		ldr	r2, .Ldsi+4
+		teq	r3, r2			@ Check for stmia sp!, {args}
+		addeq	save, save, #4		@ next instruction
+		bleq	.Ldumpstm
+
+		sub	r0, frame, #16
+1006:		ldr	r1, [save, #4]		@ Get 'stmia sp!, {rlist, fp, ip, lr, pc}' instruction
+		mov	r3, r1, lsr #10
+		ldr	r2, .Ldsi
+		teq	r3, r2
+		bleq	.Ldumpstm
+
+		teq	frame, next
+		movne	frame, next
+		teqne	frame, #0
+		bne	3b
+		LOADREGS(fd, sp!, {r4 - r8, pc})
+
+/*
+ * Fixup for LDMDB
+ */
+		.section .fixup,"ax"
+		.align	0
+1007:		ldr	r0, =.Lbad
+		mov	r1, frame
+		bl	printk
+		LOADREGS(fd, sp!, {r4 - r8, pc})
+		.ltorg
+		.previous
+		
+		.section __ex_table,"a"
+		.align	3
+		.long	1001b, 1007b
+		.long	1002b, 1007b
+		.long	1003b, 1007b
+		.long	1004b, 1007b
+		.long	1005b, 1007b
+		.long	1006b, 1007b
+		.previous
+
+#define instr r4
+#define reg   r5
+#define stack r6
+
+.Ldumpstm:	stmfd	sp!, {instr, reg, stack, r7, lr}
+		mov	stack, r0
+		mov	instr, r1
+		mov	reg, #9
+		mov	r7, #0
+1:		mov	r3, #1
+		tst	instr, r3, lsl reg
+		beq	2f
+		add	r7, r7, #1
+		teq	r7, #4
+		moveq	r7, #0
+		moveq	r3, #'\n'
+		movne	r3, #' '
+		ldr	r2, [stack], #-4
+		mov	r1, reg
+		adr	r0, .Lfp
+		bl	printk
+2:		subs	reg, reg, #1
+		bpl	1b
+		teq	r7, #0
+		adrne	r0, .Lcr
+		blne	printk
+		mov	r0, stack
+		LOADREGS(fd, sp!, {instr, reg, stack, r7, pc})
+
+.Lfe:		.asciz	"Function entered at [<%p>] from [<%p>]\n"
+.Lfp:		.asciz	" r%d = %08X%c"
+.Lcr:		.asciz	"\n"
+.Lbad:		.asciz	"Backtrace aborted due to bad frame pointer <%p>\n"
+		.align
+.Ldsi:		.word	0x00e92dd8 >> 2
+		.word	0x00e92d00 >> 2
+
+#endif
diff --git a/arch/arm26/lib/changebit.S b/arch/arm26/lib/changebit.S
new file mode 100644
index 0000000..1b6a077
--- /dev/null
+++ b/arch/arm26/lib/changebit.S
@@ -0,0 +1,28 @@
+/*
+ *  linux/arch/arm26/lib/changebit.S
+ *
+ *  Copyright (C) 1995-1996 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>
+                .text
+
+/* Purpose  : Function to change a bit
+ * Prototype: int change_bit(int bit, void *addr)
+ */
+ENTRY(_change_bit_be)
+		eor	r0, r0, #0x18		@ big endian byte ordering
+ENTRY(_change_bit_le)
+		and	r2, r0, #7
+		mov	r3, #1
+		mov	r3, r3, lsl r2
+		save_and_disable_irqs ip, r2
+		ldrb	r2, [r1, r0, lsr #3]
+		eor	r2, r2, r3
+		strb	r2, [r1, r0, lsr #3]
+		restore_irqs ip
+		RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/clearbit.S b/arch/arm26/lib/clearbit.S
new file mode 100644
index 0000000..0a895b0
--- /dev/null
+++ b/arch/arm26/lib/clearbit.S
@@ -0,0 +1,31 @@
+/*
+ *  linux/arch/arm26/lib/clearbit.S
+ *
+ *  Copyright (C) 1995-1996 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>
+                .text
+
+/*
+ * Purpose  : Function to clear a bit
+ * Prototype: int clear_bit(int bit, void *addr)
+ */
+ENTRY(_clear_bit_be)
+		eor	r0, r0, #0x18		@ big endian byte ordering
+ENTRY(_clear_bit_le)
+		and	r2, r0, #7
+		mov	r3, #1
+		mov	r3, r3, lsl r2
+		save_and_disable_irqs ip, r2
+		ldrb	r2, [r1, r0, lsr #3]
+		bic	r2, r2, r3
+		strb	r2, [r1, r0, lsr #3]
+		restore_irqs ip
+		RETINSTR(mov,pc,lr)
+
+
diff --git a/arch/arm26/lib/copy_page.S b/arch/arm26/lib/copy_page.S
new file mode 100644
index 0000000..2d79ee1
--- /dev/null
+++ b/arch/arm26/lib/copy_page.S
@@ -0,0 +1,62 @@
+/*
+ *  linux/arch/arm26/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 <asm/assembler.h>
+#include <asm/asm_offsets.h>
+
+		.text
+		.align	5
+/*
+ * ARMv3 optimised copy_user_page
+ *
+ * FIXME: rmk do we need to handle cache stuff...
+ * FIXME: im is this right on ARM26?
+ */
+ENTRY(__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
+	LOADREGS(fd, sp!, {r4, pc})		@	3
+
+	.align	5
+/*
+ * ARMv3 optimised clear_user_page
+ *
+ * FIXME: rmk do we need to handle cache stuff...
+ */
+ENTRY(__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
+
+	.section ".init.text", #alloc, #execinstr
+
diff --git a/arch/arm26/lib/csumipv6.S b/arch/arm26/lib/csumipv6.S
new file mode 100644
index 0000000..6283115
--- /dev/null
+++ b/arch/arm26/lib/csumipv6.S
@@ -0,0 +1,32 @@
+/*
+ *  linux/arch/arm26/lib/csumipv6.S
+ *
+ *  Copyright (C) 1995-1998 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.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+		.text
+
+ENTRY(__csum_ipv6_magic)
+		str	lr, [sp, #-4]!
+		adds	ip, r2, r3
+		ldmia	r1, {r1 - r3, lr}
+		adcs	ip, ip, r1
+		adcs	ip, ip, r2
+		adcs	ip, ip, r3
+		adcs	ip, ip, lr
+		ldmia	r0, {r0 - r3}
+		adcs	r0, ip, r0
+		adcs	r0, r0, r1
+		adcs	r0, r0, r2
+		ldr	r2, [sp, #4]
+		adcs	r0, r0, r3
+		adcs	r0, r0, r2
+		adcs	r0, r0, #0
+		LOADREGS(fd, sp!, {pc})
+
diff --git a/arch/arm26/lib/csumpartial.S b/arch/arm26/lib/csumpartial.S
new file mode 100644
index 0000000..e53e710
--- /dev/null
+++ b/arch/arm26/lib/csumpartial.S
@@ -0,0 +1,130 @@
+/*
+ *  linux/arch/arm26/lib/csumpartial.S
+ *
+ *  Copyright (C) 1995-1998 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.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+		.text
+
+/*
+ * Function: __u32 csum_partial(const char *src, int len, __u32 sum)
+ * Params  : r0 = buffer, r1 = len, r2 = checksum
+ * Returns : r0 = new checksum
+ */
+
+buf	.req	r0
+len	.req	r1
+sum	.req	r2
+td0	.req	r3
+td1	.req	r4	@ save before use
+td2	.req	r5	@ save before use
+td3	.req	lr
+
+.zero:		mov	r0, sum
+		add	sp, sp, #4
+		ldr	pc, [sp], #4
+
+		/*
+		 * Handle 0 to 7 bytes, with any alignment of source and
+		 * destination pointers.  Note that when we get here, C = 0
+		 */
+.less8:		teq	len, #0			@ check for zero count
+		beq	.zero
+
+		/* we must have at least one byte. */
+		tst	buf, #1			@ odd address?
+		ldrneb	td0, [buf], #1
+		subne	len, len, #1
+		adcnes	sum, sum, td0, lsl #byte(1)
+
+.less4:		tst	len, #6
+		beq	.less8_byte
+
+		/* we are now half-word aligned */
+
+.less8_wordlp:
+#if __LINUX_ARM_ARCH__ >= 4
+		ldrh	td0, [buf], #2
+		sub	len, len, #2
+#else
+		ldrb	td0, [buf], #1
+		ldrb	td3, [buf], #1
+		sub	len, len, #2
+		orr	td0, td0, td3, lsl #8
+#endif
+		adcs	sum, sum, td0
+		tst	len, #6
+		bne	.less8_wordlp
+
+.less8_byte:	tst	len, #1			@ odd number of bytes
+		ldrneb	td0, [buf], #1		@ include last byte
+		adcnes	sum, sum, td0, lsl #byte(0)	@ update checksum
+
+.done:		adc	r0, sum, #0		@ collect up the last carry
+		ldr	td0, [sp], #4
+		tst	td0, #1			@ check buffer alignment
+		movne	td0, r0, lsl #8		@ rotate checksum by 8 bits
+		orrne	r0, td0, r0, lsr #24
+		ldr	pc, [sp], #4		@ return
+
+.not_aligned:	tst	buf, #1			@ odd address
+		ldrneb	td0, [buf], #1		@ make even
+		subne	len, len, #1
+		adcnes	sum, sum, td0, lsl #byte(1)	@ update checksum
+
+		tst	buf, #2			@ 32-bit aligned?
+#if __LINUX_ARM_ARCH__ >= 4
+		ldrneh	td0, [buf], #2		@ make 32-bit aligned
+		subne	len, len, #2
+#else
+		ldrneb	td0, [buf], #1
+		ldrneb	ip, [buf], #1
+		subne	len, len, #2
+		orrne	td0, td0, ip, lsl #8
+#endif
+		adcnes	sum, sum, td0		@ update checksum
+		mov	pc, lr
+
+ENTRY(csum_partial)
+		stmfd	sp!, {buf, lr}
+		cmp	len, #8			@ Ensure that we have at least
+		blo	.less8			@ 8 bytes to copy.
+
+		adds	sum, sum, #0		@ C = 0
+		tst	buf, #3			@ Test destination alignment
+		blne	.not_aligned		@ aligh destination, return here
+
+1:		bics	ip, len, #31
+		beq	3f
+
+		stmfd	sp!, {r4 - r5}
+2:		ldmia	buf!, {td0, td1, td2, td3}
+		adcs	sum, sum, td0
+		adcs	sum, sum, td1
+		adcs	sum, sum, td2
+		adcs	sum, sum, td3
+		ldmia	buf!, {td0, td1, td2, td3}
+		adcs	sum, sum, td0
+		adcs	sum, sum, td1
+		adcs	sum, sum, td2
+		adcs	sum, sum, td3
+		sub	ip, ip, #32
+		teq	ip, #0
+		bne	2b
+		ldmfd	sp!, {r4 - r5}
+
+3:		tst	len, #0x1c		@ should not change C
+		beq	.less4
+
+4:		ldr	td0, [buf], #4
+		sub	len, len, #4
+		adcs	sum, sum, td0
+		tst	len, #0x1c
+		bne	4b
+		b	.less4
diff --git a/arch/arm26/lib/csumpartialcopy.S b/arch/arm26/lib/csumpartialcopy.S
new file mode 100644
index 0000000..a1c4b5f
--- /dev/null
+++ b/arch/arm26/lib/csumpartialcopy.S
@@ -0,0 +1,52 @@
+/*
+ *  linux/arch/arm26/lib/csumpartialcopy.S
+ *
+ *  Copyright (C) 1995-1998 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.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+		.text
+
+/* Function: __u32 csum_partial_copy_nocheck(const char *src, char *dst, int len, __u32 sum)
+ * Params  : r0 = src, r1 = dst, r2 = len, r3 = checksum
+ * Returns : r0 = new checksum
+ */
+
+		.macro	save_regs
+		stmfd	sp!, {r1, r4 - r8, fp, ip, lr, pc}
+		.endm
+
+		.macro	load_regs,flags
+		LOADREGS(\flags,fp,{r1, r4 - r8, fp, sp, pc})
+		.endm
+
+		.macro	load1b, reg1
+		ldrb	\reg1, [r0], #1
+		.endm
+
+		.macro	load2b, reg1, reg2
+		ldrb	\reg1, [r0], #1
+		ldrb	\reg2, [r0], #1
+		.endm
+
+		.macro	load1l, reg1
+		ldr	\reg1, [r0], #4
+		.endm
+
+		.macro	load2l, reg1, reg2
+		ldr	\reg1, [r0], #4
+		ldr	\reg2, [r0], #4
+		.endm
+
+		.macro	load4l, reg1, reg2, reg3, reg4
+		ldmia	r0!, {\reg1, \reg2, \reg3, \reg4}
+		.endm
+
+#define FN_ENTRY	ENTRY(csum_partial_copy_nocheck)
+
+#include "csumpartialcopygeneric.S"
diff --git a/arch/arm26/lib/csumpartialcopygeneric.S b/arch/arm26/lib/csumpartialcopygeneric.S
new file mode 100644
index 0000000..5249c3a
--- /dev/null
+++ b/arch/arm26/lib/csumpartialcopygeneric.S
@@ -0,0 +1,352 @@
+/*
+ *  linux/arch/arm26/lib/csumpartialcopygeneric.S
+ *
+ *  Copyright (C) 1995-2001 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.
+ *
+ * JMA 01/06/03 Commented out some shl0s; probobly irrelevant to arm26 
+ *
+ */
+
+/*
+ * unsigned int
+ * csum_partial_copy_xxx(const char *src, char *dst, int len, int sum, )
+ *  r0 = src, r1 = dst, r2 = len, r3 = sum
+ *  Returns : r0 = checksum
+ *
+ * Note that 'tst' and 'teq' preserve the carry flag.
+ */
+
+/* Quick hack */
+                .macro  save_regs
+                stmfd   sp!, {r1, r4 - r8, fp, ip, lr, pc}
+                .endm
+
+/* end Quick Hack */
+
+src	.req	r0
+dst	.req	r1
+len	.req	r2
+sum	.req	r3
+
+.zero:		mov	r0, sum
+		load_regs	ea
+
+		/*
+		 * Align an unaligned destination pointer.  We know that
+		 * we have >= 8 bytes here, so we don't need to check
+		 * the length.  Note that the source pointer hasn't been
+		 * aligned yet.
+		 */
+.dst_unaligned:	tst	dst, #1
+		beq	.dst_16bit
+
+		load1b	ip
+		sub	len, len, #1
+		adcs	sum, sum, ip, lsl #byte(1)	@ update checksum
+		strb	ip, [dst], #1
+		tst	dst, #2
+		moveq	pc, lr			@ dst is now 32bit aligned
+
+.dst_16bit:	load2b	r8, ip
+		sub	len, len, #2
+		adcs	sum, sum, r8, lsl #byte(0)
+		strb	r8, [dst], #1
+		adcs	sum, sum, ip, lsl #byte(1)
+		strb	ip, [dst], #1
+		mov	pc, lr			@ dst is now 32bit aligned
+
+		/*
+		 * Handle 0 to 7 bytes, with any alignment of source and
+		 * destination pointers.  Note that when we get here, C = 0
+		 */
+.less8:		teq	len, #0			@ check for zero count
+		beq	.zero
+
+		/* we must have at least one byte. */
+		tst	dst, #1			@ dst 16-bit aligned
+		beq	.less8_aligned
+
+		/* Align dst */
+		load1b	ip
+		sub	len, len, #1
+		adcs	sum, sum, ip, lsl #byte(1)	@ update checksum
+		strb	ip, [dst], #1
+		tst	len, #6
+		beq	.less8_byteonly
+
+1:		load2b	r8, ip
+		sub	len, len, #2
+		adcs	sum, sum, r8, lsl #byte(0)
+		strb	r8, [dst], #1
+		adcs	sum, sum, ip, lsl #byte(1)
+		strb	ip, [dst], #1
+.less8_aligned:	tst	len, #6
+		bne	1b
+.less8_byteonly:
+		tst	len, #1
+		beq	.done
+		load1b	r8
+		adcs	sum, sum, r8, lsl #byte(0)	@ update checksum
+		strb	r8, [dst], #1
+		b	.done
+
+FN_ENTRY
+		mov	ip, sp
+		save_regs
+		sub	fp, ip, #4
+
+		cmp	len, #8			@ Ensure that we have at least
+		blo	.less8			@ 8 bytes to copy.
+
+		adds	sum, sum, #0		@ C = 0
+		tst	dst, #3			@ Test destination alignment
+		blne	.dst_unaligned		@ align destination, return here
+
+		/*
+		 * Ok, the dst pointer is now 32bit aligned, and we know
+		 * that we must have more than 4 bytes to copy.  Note
+		 * that C contains the carry from the dst alignment above.
+		 */
+
+		tst	src, #3			@ Test source alignment
+		bne	.src_not_aligned
+
+		/* Routine for src & dst aligned */
+
+		bics	ip, len, #15
+		beq	2f
+
+1:		load4l	r4, r5, r6, r7
+		stmia	dst!, {r4, r5, r6, r7}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		adcs	sum, sum, r6
+		adcs	sum, sum, r7
+		sub	ip, ip, #16
+		teq	ip, #0
+		bne	1b
+
+2:		ands	ip, len, #12
+		beq	4f
+		tst	ip, #8
+		beq	3f
+		load2l	r4, r5
+		stmia	dst!, {r4, r5}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		tst	ip, #4
+		beq	4f
+
+3:		load1l	r4
+		str	r4, [dst], #4
+		adcs	sum, sum, r4
+
+4:		ands	len, len, #3
+		beq	.done
+		load1l	r4
+		tst	len, #2
+/*		mov	r5, r4, lsr #byte(0)
+FIXME? 0 Shift anyhow!
+*/
+		beq	.exit
+		adcs	sum, sum, r4, push #16
+		strb	r5, [dst], #1
+		mov	r5, r4, lsr #byte(1)
+		strb	r5, [dst], #1
+		mov	r5, r4, lsr #byte(2)
+.exit:		tst	len, #1
+		strneb	r5, [dst], #1
+		andne	r5, r5, #255
+		adcnes	sum, sum, r5, lsl #byte(0)
+
+		/*
+		 * If the dst pointer was not 16-bit aligned, we
+		 * need to rotate the checksum here to get around
+		 * the inefficient byte manipulations in the
+		 * architecture independent code.
+		 */
+.done:		adc	r0, sum, #0
+		ldr	sum, [sp, #0]		@ dst
+		tst	sum, #1
+		movne	sum, r0, lsl #8
+		orrne	r0, sum, r0, lsr #24
+		load_regs	ea
+
+.src_not_aligned:
+		adc	sum, sum, #0		@ include C from dst alignment
+		and	ip, src, #3
+		bic	src, src, #3
+		load1l	r5
+		cmp	ip, #2
+		beq	.src2_aligned
+		bhi	.src3_aligned
+		mov	r4, r5, pull #8		@ C = 0
+		bics	ip, len, #15
+		beq	2f
+1:		load4l	r5, r6, r7, r8
+		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
+		mov	r7, r7, pull #8
+		orr	r7, r7, r8, push #24
+		stmia	dst!, {r4, r5, r6, r7}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		adcs	sum, sum, r6
+		adcs	sum, sum, r7
+		mov	r4, r8, pull #8
+		sub	ip, ip, #16
+		teq	ip, #0
+		bne	1b
+2:		ands	ip, len, #12
+		beq	4f
+		tst	ip, #8
+		beq	3f
+		load2l	r5, r6
+		orr	r4, r4, r5, push #24
+		mov	r5, r5, pull #8
+		orr	r5, r5, r6, push #24
+		stmia	dst!, {r4, r5}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		mov	r4, r6, pull #8
+		tst	ip, #4
+		beq	4f
+3:		load1l	r5
+		orr	r4, r4, r5, push #24
+		str	r4, [dst], #4
+		adcs	sum, sum, r4
+		mov	r4, r5, pull #8
+4:		ands	len, len, #3
+		beq	.done
+/*		mov	r5, r4, lsr #byte(0)
+FIXME? 0 Shift anyhow
+*/
+		tst	len, #2
+		beq	.exit
+		adcs	sum, sum, r4, push #16
+		strb	r5, [dst], #1
+		mov	r5, r4, lsr #byte(1)
+		strb	r5, [dst], #1
+		mov	r5, r4, lsr #byte(2)
+		b	.exit
+
+.src2_aligned:	mov	r4, r5, pull #16
+		adds	sum, sum, #0
+		bics	ip, len, #15
+		beq	2f
+1:		load4l	r5, r6, r7, r8
+		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
+		mov	r7, r7, pull #16
+		orr	r7, r7, r8, push #16
+		stmia	dst!, {r4, r5, r6, r7}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		adcs	sum, sum, r6
+		adcs	sum, sum, r7
+		mov	r4, r8, pull #16
+		sub	ip, ip, #16
+		teq	ip, #0
+		bne	1b
+2:		ands	ip, len, #12
+		beq	4f
+		tst	ip, #8
+		beq	3f
+		load2l	r5, r6
+		orr	r4, r4, r5, push #16
+		mov	r5, r5, pull #16
+		orr	r5, r5, r6, push #16
+		stmia	dst!, {r4, r5}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		mov	r4, r6, pull #16
+		tst	ip, #4
+		beq	4f
+3:		load1l	r5
+		orr	r4, r4, r5, push #16
+		str	r4, [dst], #4
+		adcs	sum, sum, r4
+		mov	r4, r5, pull #16
+4:		ands	len, len, #3
+		beq	.done
+/*		mov	r5, r4, lsr #byte(0)
+FIXME? 0 Shift anyhow
+*/
+		tst	len, #2
+		beq	.exit
+		adcs	sum, sum, r4
+		strb	r5, [dst], #1
+		mov	r5, r4, lsr #byte(1)
+		strb	r5, [dst], #1
+		tst	len, #1
+		beq	.done
+		load1b	r5
+		b	.exit
+
+.src3_aligned:	mov	r4, r5, pull #24
+		adds	sum, sum, #0
+		bics	ip, len, #15
+		beq	2f
+1:		load4l	r5, r6, r7, r8
+		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
+		mov	r7, r7, pull #24
+		orr	r7, r7, r8, push #8
+		stmia	dst!, {r4, r5, r6, r7}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		adcs	sum, sum, r6
+		adcs	sum, sum, r7
+		mov	r4, r8, pull #24
+		sub	ip, ip, #16
+		teq	ip, #0
+		bne	1b
+2:		ands	ip, len, #12
+		beq	4f
+		tst	ip, #8
+		beq	3f
+		load2l	r5, r6
+		orr	r4, r4, r5, push #8
+		mov	r5, r5, pull #24
+		orr	r5, r5, r6, push #8
+		stmia	dst!, {r4, r5}
+		adcs	sum, sum, r4
+		adcs	sum, sum, r5
+		mov	r4, r6, pull #24
+		tst	ip, #4
+		beq	4f
+3:		load1l	r5
+		orr	r4, r4, r5, push #8
+		str	r4, [dst], #4
+		adcs	sum, sum, r4
+		mov	r4, r5, pull #24
+4:		ands	len, len, #3
+		beq	.done
+/*		mov	r5, r4, lsr #byte(0)
+FIXME? 0 Shift anyhow
+*/
+		tst	len, #2
+		beq	.exit
+		strb	r5, [dst], #1
+		adcs	sum, sum, r4
+		load1l	r4
+/*		mov	r5, r4, lsr #byte(0)
+FIXME? 0 Shift anyhow
+*/
+		strb	r5, [dst], #1
+		adcs	sum, sum, r4, push #24
+		mov	r5, r4, lsr #byte(1)
+		b	.exit
diff --git a/arch/arm26/lib/csumpartialcopyuser.S b/arch/arm26/lib/csumpartialcopyuser.S
new file mode 100644
index 0000000..5b82118
--- /dev/null
+++ b/arch/arm26/lib/csumpartialcopyuser.S
@@ -0,0 +1,115 @@
+/*
+ *  linux/arch/arm26/lib/csumpartialcopyuser.S
+ *
+ *  Copyright (C) 1995-1998 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.
+ */
+#include <linux/config.h>
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/errno.h>
+#include <asm/asm_offsets.h>
+
+		.text
+
+		.macro	save_regs
+		stmfd	sp!, {r1 - r2, r4 - r9, fp, ip, lr, pc}
+		mov	r9, sp, lsr #13
+		mov	r9, r9, lsl #13
+		ldr	r9, [r9, #TSK_ADDR_LIMIT]
+		mov	r9, r9, lsr #24
+		.endm
+
+		.macro	load_regs,flags
+		ldm\flags	fp, {r1, r2, r4-r9, fp, sp, pc}^
+		.endm
+
+		.macro	load1b,	reg1
+		tst	r9, #0x01
+9999:		ldreqbt	\reg1, [r0], #1
+		ldrneb	\reg1, [r0], #1
+		.section __ex_table, "a"
+		.align	3
+		.long	9999b, 6001f
+		.previous
+		.endm
+
+		.macro	load2b, reg1, reg2
+		tst	r9, #0x01
+9999:		ldreqbt	\reg1, [r0], #1
+		ldrneb	\reg1, [r0], #1
+9998:		ldreqbt	\reg2, [r0], #1
+		ldrneb	\reg2, [r0], #1
+		.section __ex_table, "a"
+		.long	9999b, 6001f
+		.long	9998b, 6001f
+		.previous
+		.endm
+
+		.macro	load1l, reg1
+		tst	r9, #0x01
+9999:		ldreqt	\reg1, [r0], #4
+		ldrne	\reg1, [r0], #4
+		.section __ex_table, "a"
+		.align	3
+		.long	9999b, 6001f
+		.previous
+		.endm
+
+		.macro	load2l, reg1, reg2
+		tst	r9, #0x01
+		ldmneia	r0!, {\reg1, \reg2}
+9999:		ldreqt	\reg1, [r0], #4
+9998:		ldreqt	\reg2, [r0], #4
+		.section __ex_table, "a"
+		.long	9999b, 6001f
+		.long	9998b, 6001f
+		.previous
+		.endm
+
+		.macro	load4l, reg1, reg2, reg3, reg4
+		tst	r9, #0x01
+		ldmneia	r0!, {\reg1, \reg2, \reg3, \reg4}
+9999:		ldreqt	\reg1, [r0], #4
+9998:		ldreqt	\reg2, [r0], #4
+9997:		ldreqt	\reg3, [r0], #4
+9996:		ldreqt	\reg4, [r0], #4
+		.section __ex_table, "a"
+		.long	9999b, 6001f
+		.long	9998b, 6001f
+		.long	9997b, 6001f
+		.long	9996b, 6001f
+		.previous
+		.endm
+
+/*
+ * unsigned int
+ * csum_partial_copy_from_user(const char *src, char *dst, int len, int sum, int *err_ptr)
+ *  r0 = src, r1 = dst, r2 = len, r3 = sum, [sp] = *err_ptr
+ *  Returns : r0 = checksum, [[sp, #0], #0] = 0 or -EFAULT
+ */
+
+#define FN_ENTRY	ENTRY(csum_partial_copy_from_user)
+
+#include "csumpartialcopygeneric.S"
+
+/*
+ * FIXME: minor buglet here
+ * We don't return the checksum for the data present in the buffer.  To do
+ * so properly, we would have to add in whatever registers were loaded before
+ * the fault, which, with the current asm above is not predictable.
+ */
+		.align	4
+6001:		mov	r4, #-EFAULT
+		ldr	r5, [fp, #4]		@ *err_ptr
+		str	r4, [r5]
+		ldmia	sp, {r1, r2}		@ retrieve dst, len
+		add	r2, r2, r1
+		mov	r0, #0			@ zero the buffer
+6002:		teq	r2, r1
+		strneb	r0, [r1], #1
+		bne	6002b
+		load_regs	ea
diff --git a/arch/arm26/lib/delay.S b/arch/arm26/lib/delay.S
new file mode 100644
index 0000000..66f2b68
--- /dev/null
+++ b/arch/arm26/lib/delay.S
@@ -0,0 +1,57 @@
+/*
+ *  linux/arch/arm26/lib/delay.S
+ *
+ *  Copyright (C) 1995, 1996 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>
+		.text
+
+LC0:		.word	loops_per_jiffy
+
+/*
+ * 0 <= r0 <= 2000
+ */
+ENTRY(udelay)
+		mov	r2,     #0x6800
+		orr	r2, r2, #0x00db
+		mul	r1, r0, r2
+		ldr	r2, LC0
+		ldr	r2, [r2]
+		mov	r1, r1, lsr #11
+		mov	r2, r2, lsr #11
+		mul	r0, r1, r2
+		movs	r0, r0, lsr #6
+		RETINSTR(moveq,pc,lr)
+
+/*
+ * loops = (r0 * 0x10c6 * 100 * loops_per_jiffy) / 2^32
+ *
+ * Oh, if only we had a cycle counter...
+ */
+
+@ Delay routine
+ENTRY(__delay)
+		subs	r0, r0, #1
+#if 0
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+		RETINSTR(movls,pc,lr)
+		subs	r0, r0, #1
+#endif
+		bhi	__delay
+		RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/ecard.S b/arch/arm26/lib/ecard.S
new file mode 100644
index 0000000..b463315
--- /dev/null
+++ b/arch/arm26/lib/ecard.S
@@ -0,0 +1,41 @@
+/*
+ *  linux/arch/arm26/lib/ecard.S
+ *
+ *  Copyright (C) 1995, 1996 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/config.h> /* for CONFIG_CPU_nn */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/hardware.h>
+
+#define CPSR2SPSR(rt)
+
+@ Purpose: call an expansion card loader to read bytes.
+@ Proto  : char read_loader(int offset, char *card_base, char *loader);
+@ Returns: byte read
+
+ENTRY(ecard_loader_read)
+		stmfd	sp!, {r4 - r12, lr}
+		mov	r11, r1
+		mov	r1, r0
+		CPSR2SPSR(r0)
+		mov	lr, pc
+		mov	pc, r2
+		LOADREGS(fd, sp!, {r4 - r12, pc})
+
+@ Purpose: call an expansion card loader to reset the card
+@ Proto  : void read_loader(int card_base, char *loader);
+@ Returns: byte read
+
+ENTRY(ecard_loader_reset)
+		stmfd	sp!, {r4 - r12, lr}
+		mov	r11, r0
+		CPSR2SPSR(r0)
+		mov	lr, pc
+		add	pc, r1, #8
+		LOADREGS(fd, sp!, {r4 - r12, pc})
+
diff --git a/arch/arm26/lib/findbit.S b/arch/arm26/lib/findbit.S
new file mode 100644
index 0000000..26f67cc
--- /dev/null
+++ b/arch/arm26/lib/findbit.S
@@ -0,0 +1,67 @@
+/*
+ *  linux/arch/arm/lib/findbit.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.
+ *
+ * 16th March 2001 - John Ripley <jripley@sonicblue.com>
+ *   Fixed so that "size" is an exclusive not an inclusive quantity.
+ *   All users of these functions expect exclusive sizes, and may
+ *   also call with zero size.
+ * Reworked by rmk.
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+                .text
+
+/*
+ * Purpose  : Find a 'zero' bit
+ * Prototype: int find_first_zero_bit(void *addr, unsigned int maxbit);
+ */
+ENTRY(_find_first_zero_bit_le)
+		teq	r1, #0	
+		beq	3f
+		mov	r2, #0
+1:		ldrb	r3, [r0, r2, lsr #3]
+		eors	r3, r3, #0xff		@ invert bits
+		bne	.found			@ any now set - found zero bit
+		add	r2, r2, #8		@ next bit pointer
+2:		cmp	r2, r1			@ any more?
+		blo	1b
+3:		mov	r0, r1			@ no free bits
+		RETINSTR(mov,pc,lr)
+
+/*
+ * Purpose  : Find next 'zero' bit
+ * Prototype: int find_next_zero_bit(void *addr, unsigned int maxbit, int offset)
+ */
+ENTRY(_find_next_zero_bit_le)
+		teq	r1, #0
+		beq	2b
+		ands	ip, r2, #7
+		beq	1b			@ If new byte, goto old routine
+		ldrb	r3, [r0, r2, lsr #3]
+		eor	r3, r3, #0xff		@ now looking for a 1 bit
+		movs	r3, r3, lsr ip		@ shift off unused bits
+		bne	.found
+		orr	r2, r2, #7		@ if zero, then no bits here
+		add	r2, r2, #1		@ align bit pointer
+		b	2b			@ loop for next bit
+
+/*
+ * One or more bits in the LSB of r3 are assumed to be set.
+ */
+.found:		tst	r3, #0x0f
+		addeq	r2, r2, #4
+		movne	r3, r3, lsl #4
+		tst	r3, #0x30
+		addeq	r2, r2, #2
+		movne	r3, r3, lsl #2
+		tst	r3, #0x40
+		addeq	r2, r2, #1
+		mov	r0, r2
+		RETINSTR(mov,pc,lr)
+
diff --git a/arch/arm26/lib/floppydma.S b/arch/arm26/lib/floppydma.S
new file mode 100644
index 0000000..e99ebbb
--- /dev/null
+++ b/arch/arm26/lib/floppydma.S
@@ -0,0 +1,32 @@
+/*
+ *  linux/arch/arm26/lib/floppydma.S
+ *
+ *  Copyright (C) 1995, 1996 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>
+		.text
+
+		.global	floppy_fiqin_end
+ENTRY(floppy_fiqin_start)
+		subs	r9, r9, #1
+		ldrgtb	r12, [r11, #-4]
+		ldrleb	r12, [r11], #0
+		strb	r12, [r10], #1
+		subs	pc, lr, #4
+floppy_fiqin_end:
+
+		.global	floppy_fiqout_end
+ENTRY(floppy_fiqout_start)
+		subs	r9, r9, #1
+		ldrgeb	r12, [r10], #1
+		movlt	r12, #0
+		strleb	r12, [r11], #0
+		subles	pc, lr, #4
+		strb	r12, [r11, #-4]
+		subs	pc, lr, #4
+floppy_fiqout_end:
diff --git a/arch/arm26/lib/gcclib.h b/arch/arm26/lib/gcclib.h
new file mode 100644
index 0000000..9895e78
--- /dev/null
+++ b/arch/arm26/lib/gcclib.h
@@ -0,0 +1,21 @@
+/* gcclib.h -- definitions for various functions 'borrowed' from gcc-2.95.3 */
+/* I Molton     29/07/01 */
+
+#define BITS_PER_UNIT  8
+#define SI_TYPE_SIZE (sizeof (SItype) * BITS_PER_UNIT)
+
+typedef unsigned int UQItype    __attribute__ ((mode (QI)));
+typedef          int SItype     __attribute__ ((mode (SI)));
+typedef unsigned int USItype    __attribute__ ((mode (SI)));
+typedef          int DItype     __attribute__ ((mode (DI)));
+typedef          int word_type 	__attribute__ ((mode (__word__)));
+typedef unsigned int UDItype    __attribute__ ((mode (DI)));
+
+struct DIstruct {SItype low, high;};
+
+typedef union
+{
+  struct DIstruct s;
+  DItype ll;
+} DIunion;
+
diff --git a/arch/arm26/lib/getuser.S b/arch/arm26/lib/getuser.S
new file mode 100644
index 0000000..e6d59b33
--- /dev/null
+++ b/arch/arm26/lib/getuser.S
@@ -0,0 +1,112 @@
+/*
+ *  linux/arch/arm26/lib/getuser.S
+ *
+ *  Copyright (C) 2001 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.
+ *
+ *  Idea from x86 version, (C) Copyright 1998 Linus Torvalds
+ *
+ * These functions have a non-standard call interface to make them more
+ * efficient, especially as they return an error value in addition to
+ * the "real" return value.
+ *
+ * __get_user_X
+ *
+ * Inputs:	r0 contains the address
+ * Outputs:	r0 is the error code
+ *		r1, r2 contains the zero-extended value
+ *		lr corrupted
+ *
+ * No other registers must be altered.  (see include/asm-arm/uaccess.h
+ * for specific ASM register usage).
+ *
+ * Note that ADDR_LIMIT is either 0 or 0xc0000000.
+ * Note also that it is intended that __get_user_bad is not global.
+ */
+#include <asm/asm_offsets.h>
+#include <asm/thread_info.h>
+#include <asm/errno.h>
+
+        .global __get_user_1
+__get_user_1:
+        bic     r1, sp, #0x1f00
+        bic     r1, r1, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r1, [r1, #TI_ADDR_LIMIT]
+        sub     r1, r1, #1
+        cmp     r0, r1
+        bge     __get_user_bad
+        cmp     r0, #0x02000000
+1:      ldrlsbt r1, [r0]
+        ldrgeb  r1, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+        .global __get_user_2
+__get_user_2:
+        bic     r2, sp, #0x1f00
+        bic     r2, r2, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r2, [r2, #TI_ADDR_LIMIT]
+        sub     r2, r2, #2
+        cmp     r0, r2
+        bge     __get_user_bad
+        cmp     r0, #0x02000000
+2:      ldrlsbt r1, [r0], #1
+3:      ldrlsbt r2, [r0]
+        ldrgeb  r1, [r0], #1
+        ldrgeb  r2, [r0]
+        orr     r1, r1, r2, lsl #8
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+        .global __get_user_4
+__get_user_4:
+        bic     r1, sp, #0x1f00
+        bic     r1, r1, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r1, [r1, #TI_ADDR_LIMIT]
+        sub     r1, r1, #4
+        cmp     r0, r1
+        bge     __get_user_bad
+        cmp     r0, #0x02000000
+4:      ldrlst  r1, [r0]
+        ldrge   r1, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+        .global __get_user_8
+__get_user_8:
+        bic     r2, sp, #0x1f00
+        bic     r2, r2, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r2, [r2, #TI_ADDR_LIMIT]
+        sub     r2, r2, #8
+        cmp     r0, r2
+        bge     __get_user_bad_8
+        cmp     r0, #0x02000000
+5:      ldrlst  r1, [r0], #4
+6:      ldrlst  r2, [r0]
+        ldrge   r1, [r0], #4
+        ldrge   r2, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+__get_user_bad_8:
+        mov     r2, #0
+__get_user_bad:
+        mov     r1, #0
+        mov     r0, #-EFAULT
+        ldmfd   sp!, {pc}^
+
+.section __ex_table, "a"
+	.long	1b, __get_user_bad
+	.long	2b, __get_user_bad
+	.long	3b, __get_user_bad
+	.long	4b, __get_user_bad
+	.long	5b, __get_user_bad_8
+	.long	6b, __get_user_bad_8
+.previous
diff --git a/arch/arm26/lib/io-acorn.S b/arch/arm26/lib/io-acorn.S
new file mode 100644
index 0000000..f6c3e30
--- /dev/null
+++ b/arch/arm26/lib/io-acorn.S
@@ -0,0 +1,71 @@
+/*
+ *  linux/arch/arm26/lib/io-acorn.S
+ *
+ *  Copyright (C) 1995, 1996 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/config.h> /* for CONFIG_CPU_nn */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/hardware.h>
+
+		.text
+		.align
+
+		.equ	diff_pcio_base, PCIO_BASE - IO_BASE
+
+		.macro	outw2	rd
+		mov	r8, \rd, lsl #16
+		orr	r8, r8, r8, lsr #16
+		str	r8, [r3, r0, lsl #2]
+		mov	r8, \rd, lsr #16
+		orr	r8, r8, r8, lsl #16
+		str	r8, [r3, r0, lsl #2]
+		.endm
+
+		.macro	inw2	rd, mask, temp
+		ldr	\rd, [r0]
+		and	\rd, \rd, \mask
+		ldr	\temp, [r0]
+		orr	\rd, \rd, \temp, lsl #16
+		.endm
+
+		.macro	addr	rd
+		tst	\rd, #0x80000000
+		mov	\rd, \rd, lsl #2
+		add	\rd, \rd, #IO_BASE
+		addeq	\rd, \rd, #diff_pcio_base
+		.endm
+
+.iosl_warning:
+		.ascii	"<4>insl/outsl not implemented, called from %08lX\0"
+		.align
+
+/*
+ * These make no sense on Acorn machines.
+ * Print a warning message.
+ */
+ENTRY(insl)
+ENTRY(outsl)
+		adr	r0, .iosl_warning
+		mov	r1, lr
+		b	printk
+
+@ Purpose: write a memc register
+@ Proto  : void memc_write(int register, int value);
+@ Returns: nothing
+
+ENTRY(memc_write)
+		cmp	r0, #7
+		RETINSTR(movgt,pc,lr)
+		mov	r0, r0, lsl #17
+		mov	r1, r1, lsl #15
+		mov	r1, r1, lsr #17
+		orr	r0, r0, r1, lsl #2
+		add	r0, r0, #0x03600000
+		strb	r0, [r0]
+		RETINSTR(mov,pc,lr)
+
diff --git a/arch/arm26/lib/io-readsb.S b/arch/arm26/lib/io-readsb.S
new file mode 100644
index 0000000..4c4d99c
--- /dev/null
+++ b/arch/arm26/lib/io-readsb.S
@@ -0,0 +1,116 @@
+/*
+ *  linux/arch/arm26/lib/io-readsb.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>
+#include <asm/hardware.h>
+
+.insb_align:	rsb	ip, ip, #4
+		cmp	ip, r2
+		movgt	ip, r2
+		cmp	ip, #2
+		ldrb	r3, [r0]
+		strb	r3, [r1], #1
+		ldrgeb	r3, [r0]
+		strgeb	r3, [r1], #1
+		ldrgtb	r3, [r0]
+		strgtb	r3, [r1], #1
+		subs	r2, r2, ip
+		bne	.insb_aligned
+
+ENTRY(__raw_readsb)
+		teq	r2, #0		@ do we have to check for the zero len?
+		moveq	pc, lr
+		ands	ip, r1, #3
+		bne	.insb_align
+
+.insb_aligned:	stmfd	sp!, {r4 - r6, lr}
+
+		subs	r2, r2, #16
+		bmi	.insb_no_16
+
+.insb_16_lp:	ldrb	r3, [r0]
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #8
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #24
+		ldrb	r4, [r0]
+		ldrb	r5, [r0]
+		orr	r4, r4, r5, lsl #8
+		ldrb	r5, [r0]
+		orr	r4, r4, r5, lsl #16
+		ldrb	r5, [r0]
+		orr	r4, r4, r5, lsl #24
+		ldrb	r5, [r0]
+		ldrb	r6, [r0]
+		orr	r5, r5, r6, lsl #8
+		ldrb	r6, [r0]
+		orr	r5, r5, r6, lsl #16
+		ldrb	r6, [r0]
+		orr	r5, r5, r6, lsl #24
+		ldrb	r6, [r0]
+		ldrb	ip, [r0]
+		orr	r6, r6, ip, lsl #8
+		ldrb	ip, [r0]
+		orr	r6, r6, ip, lsl #16
+		ldrb	ip, [r0]
+		orr	r6, r6, ip, lsl #24
+		stmia	r1!, {r3 - r6}
+
+		subs	r2, r2, #16
+		bpl	.insb_16_lp
+
+		tst	r2, #15
+		LOADREGS(eqfd, sp!, {r4 - r6, pc})
+
+.insb_no_16:	tst	r2, #8
+		beq	.insb_no_8
+
+		ldrb	r3, [r0]
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #8
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #24
+		ldrb	r4, [r0]
+		ldrb	r5, [r0]
+		orr	r4, r4, r5, lsl #8
+		ldrb	r5, [r0]
+		orr	r4, r4, r5, lsl #16
+		ldrb	r5, [r0]
+		orr	r4, r4, r5, lsl #24
+		stmia	r1!, {r3, r4}
+
+.insb_no_8:	tst	r2, #4
+		beq	.insb_no_4
+
+		ldrb	r3, [r0]
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #8
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+		ldrb	r4, [r0]
+		orr	r3, r3, r4, lsl #24
+		str	r3, [r1], #4
+
+.insb_no_4:	ands	r2, r2, #3
+		LOADREGS(eqfd, sp!, {r4 - r6, pc})
+
+		cmp	r2, #2
+		ldrb	r3, [r0]
+		strb	r3, [r1], #1
+		ldrgeb	r3, [r0]
+		strgeb	r3, [r1], #1
+		ldrgtb	r3, [r0]
+		strgtb	r3, [r1]
+
+		LOADREGS(fd, sp!, {r4 - r6, pc})
diff --git a/arch/arm26/lib/io-readsl.S b/arch/arm26/lib/io-readsl.S
new file mode 100644
index 0000000..7be208b
--- /dev/null
+++ b/arch/arm26/lib/io-readsl.S
@@ -0,0 +1,78 @@
+/*
+ *  linux/arch/arm26/lib/io-readsl.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>
+#include <asm/hardware.h>
+
+/*
+ * Note that some reads can be aligned on half-word boundaries.
+ */
+ENTRY(__raw_readsl)
+		teq	r2, #0		@ do we have to check for the zero len?
+		moveq	pc, lr
+		ands	ip, r1, #3
+		bne	2f
+
+1:		ldr	r3, [r0]
+		str	r3, [r1], #4
+		subs	r2, r2, #1
+		bne	1b
+		mov	pc, lr
+
+2:		cmp	ip, #2
+		ldr	ip, [r0]
+		blt	4f
+		bgt	6f
+
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+3:		subs	r2, r2, #1
+		ldrne	r3, [r0]
+		orrne	ip, ip, r3, lsl #16
+		strne	ip, [r1], #4
+		movne	ip, r3, lsr #16
+		bne	3b
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+		strb	ip, [r1], #1
+		mov	pc, lr
+
+4:		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+5:		subs	r2, r2, #1
+		ldrne	r3, [r0]
+		orrne	ip, ip, r3, lsl #8
+		strne	ip, [r1], #4
+		movne	ip, r3, lsr #24
+		bne	5b
+		strb	ip, [r1], #1
+		mov	pc, lr
+
+6:		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+7:		subs	r2, r2, #1
+		ldrne	r3, [r0]
+		orrne	ip, ip, r3, lsl #24
+		strne	ip, [r1], #4
+		movne	ip, r3, lsr #8
+		bne	7b
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+		strb	ip, [r1], #1
+		mov	ip, ip, lsr #8
+		strb	ip, [r1], #1
+		mov	pc, lr
+
diff --git a/arch/arm26/lib/io-readsw.S b/arch/arm26/lib/io-readsw.S
new file mode 100644
index 0000000..c65c1f2
--- /dev/null
+++ b/arch/arm26/lib/io-readsw.S
@@ -0,0 +1,107 @@
+/*
+ *  linux/arch/arm26/lib/io-readsw.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>
+#include <asm/hardware.h>
+
+.insw_bad_alignment:
+		adr	r0, .insw_bad_align_msg
+		mov	r2, lr
+		b	panic
+.insw_bad_align_msg:
+		.asciz	"insw: bad buffer alignment (0x%p, lr=0x%08lX)\n"
+		.align
+
+.insw_align:	tst	r1, #1
+		bne	.insw_bad_alignment
+
+		ldr	r3, [r0]
+		strb	r3, [r1], #1
+		mov	r3, r3, lsr #8
+		strb	r3, [r1], #1
+
+		subs	r2, r2, #1
+		RETINSTR(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	.insw_align
+
+.insw_aligned:	mov	ip, #0xff
+		orr	ip, ip, ip, lsl #8
+		stmfd	sp!, {r4, r5, r6, lr}
+
+		subs	r2, r2, #8
+		bmi	.no_insw_8
+
+.insw_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	.insw_8_lp
+
+		tst	r2, #7
+		LOADREGS(eqfd, sp!, {r4, r5, r6, pc})
+
+.no_insw_8:	tst	r2, #4
+		beq	.no_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}
+
+.no_insw_4:	tst	r2, #2
+		beq	.no_insw_2
+
+		ldr	r3, [r0]
+		and	r3, r3, ip
+		ldr	r4, [r0]
+		orr	r3, r3, r4, lsl #16
+
+		str	r3, [r1], #4
+
+.no_insw_2:	tst	r2, #1
+		ldrne	r3, [r0]
+		strneb	r3, [r1], #1
+		movne	r3, r3, lsr #8
+		strneb	r3, [r1]
+
+		LOADREGS(fd, sp!, {r4, r5, r6, pc})
+
+
diff --git a/arch/arm26/lib/io-writesb.S b/arch/arm26/lib/io-writesb.S
new file mode 100644
index 0000000..16251b4
--- /dev/null
+++ b/arch/arm26/lib/io-writesb.S
@@ -0,0 +1,122 @@
+/*
+ *  linux/arch/arm26/lib/io-writesb.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>
+#include <asm/hardware.h>
+
+.outsb_align:	rsb	ip, ip, #4
+		cmp	ip, r2
+		movgt	ip, r2
+		cmp	ip, #2
+		ldrb	r3, [r1], #1
+		strb	r3, [r0]
+		ldrgeb	r3, [r1], #1
+		strgeb	r3, [r0]
+		ldrgtb	r3, [r1], #1
+		strgtb	r3, [r0]
+		subs	r2, r2, ip
+		bne	.outsb_aligned
+
+ENTRY(__raw_writesb)
+		teq	r2, #0		@ do we have to check for the zero len?
+		moveq	pc, lr
+		ands	ip, r1, #3
+		bne	.outsb_align
+
+.outsb_aligned:	stmfd	sp!, {r4 - r6, lr}
+
+		subs	r2, r2, #16
+		bmi	.outsb_no_16
+
+.outsb_16_lp:	ldmia	r1!, {r3 - r6}
+
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+
+		strb	r4, [r0]
+		mov	r4, r4, lsr #8
+		strb	r4, [r0]
+		mov	r4, r4, lsr #8
+		strb	r4, [r0]
+		mov	r4, r4, lsr #8
+		strb	r4, [r0]
+
+		strb	r5, [r0]
+		mov	r5, r5, lsr #8
+		strb	r5, [r0]
+		mov	r5, r5, lsr #8
+		strb	r5, [r0]
+		mov	r5, r5, lsr #8
+		strb	r5, [r0]
+
+		strb	r6, [r0]
+		mov	r6, r6, lsr #8
+		strb	r6, [r0]
+		mov	r6, r6, lsr #8
+		strb	r6, [r0]
+		mov	r6, r6, lsr #8
+		strb	r6, [r0]
+
+		subs	r2, r2, #16
+		bpl	.outsb_16_lp
+
+		tst	r2, #15
+		LOADREGS(eqfd, sp!, {r4 - r6, pc})
+
+.outsb_no_16:	tst	r2, #8
+		beq	.outsb_no_8
+
+		ldmia	r1!, {r3, r4}
+
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+
+		strb	r4, [r0]
+		mov	r4, r4, lsr #8
+		strb	r4, [r0]
+		mov	r4, r4, lsr #8
+		strb	r4, [r0]
+		mov	r4, r4, lsr #8
+		strb	r4, [r0]
+
+.outsb_no_8:	tst	r2, #4
+		beq	.outsb_no_4
+
+		ldr	r3, [r1], #4
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+		mov	r3, r3, lsr #8
+		strb	r3, [r0]
+
+.outsb_no_4:	ands	r2, r2, #3
+		LOADREGS(eqfd, sp!, {r4 - r6, pc})
+
+		cmp	r2, #2
+		ldrb	r3, [r1], #1
+		strb	r3, [r0]
+		ldrgeb	r3, [r1], #1
+		strgeb	r3, [r0]
+		ldrgtb	r3, [r1]
+		strgtb	r3, [r0]
+
+		LOADREGS(fd, sp!, {r4 - r6, pc})
diff --git a/arch/arm26/lib/io-writesl.S b/arch/arm26/lib/io-writesl.S
new file mode 100644
index 0000000..4d6049b
--- /dev/null
+++ b/arch/arm26/lib/io-writesl.S
@@ -0,0 +1,56 @@
+/*
+ *  linux/arch/arm26/lib/io-writesl.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>
+#include <asm/hardware.h>
+
+ENTRY(__raw_writesl)
+		teq	r2, #0		@ do we have to check for the zero len?
+		moveq	pc, lr
+		ands	ip, r1, #3
+		bne	2f
+
+1:		ldr	r3, [r1], #4
+		str	r3, [r0]
+		subs	r2, r2, #1
+		bne	1b
+		mov	pc, lr
+
+2:		bic	r1, r1, #3
+		cmp	ip, #2
+		ldr	r3, [r1], #4
+		bgt	4f
+		blt	5f
+
+3:		mov	ip, r3, lsr #16
+		ldr	r3, [r1], #4
+		orr	ip, ip, r3, lsl #16
+		str	ip, [r0]
+		subs	r2, r2, #1
+		bne	3b
+		mov	pc, lr
+
+4:		mov	ip, r3, lsr #24
+		ldr	r3, [r1], #4
+		orr	ip, ip, r3, lsl #8
+		str	ip, [r0]
+		subs	r2, r2, #1
+		bne	4b
+		mov	pc, lr
+
+5:		mov	ip, r3, lsr #8
+		ldr	r3, [r1], #4
+		orr	ip, ip, r3, lsl #24
+		str	ip, [r0]
+		subs	r2, r2, #1
+		bne	5b
+		mov	pc, lr
+
+
diff --git a/arch/arm26/lib/io-writesw.S b/arch/arm26/lib/io-writesw.S
new file mode 100644
index 0000000..a24f891
--- /dev/null
+++ b/arch/arm26/lib/io-writesw.S
@@ -0,0 +1,127 @@
+/*
+ *  linux/arch/arm26/lib/io-writesw.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>
+#include <asm/hardware.h>
+
+.outsw_bad_alignment:
+		adr	r0, .outsw_bad_align_msg
+		mov	r2, lr
+		b	panic
+.outsw_bad_align_msg:
+		.asciz	"outsw: bad buffer alignment (0x%p, lr=0x%08lX)\n"
+		.align
+
+.outsw_align:	tst	r1, #1
+		bne	.outsw_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
+		RETINSTR(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	.outsw_align
+
+.outsw_aligned:	stmfd	sp!, {r4, r5, r6, lr}
+
+		subs	r2, r2, #8
+		bmi	.no_outsw_8
+
+.outsw_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	.outsw_8_lp
+
+		tst	r2, #7
+		LOADREGS(eqfd, sp!, {r4, r5, r6, pc})
+
+.no_outsw_8:	tst	r2, #4
+		beq	.no_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]
+
+.no_outsw_4:	tst	r2, #2
+		beq	.no_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]
+
+.no_outsw_2:	tst	r2, #1
+
+		ldrne	r3, [r1]
+
+		movne	ip, r3, lsl #16
+		orrne	ip, ip, ip, lsr #16
+		strne	ip, [r0]
+
+		LOADREGS(fd, sp!, {r4, r5, r6, pc})
diff --git a/arch/arm26/lib/kbd.c b/arch/arm26/lib/kbd.c
new file mode 100644
index 0000000..22d2c93
--- /dev/null
+++ b/arch/arm26/lib/kbd.c
@@ -0,0 +1,279 @@
+#include <linux/config.h>
+#include <linux/kd.h>
+//#include <linux/kbd_ll.h>
+#include <linux/kbd_kern.h>
+
+/*
+ * Translation of escaped scancodes to keycodes.
+ * This is now user-settable.
+ * The keycodes 1-88,96-111,119 are fairly standard, and
+ * should probably not be changed - changing might confuse X.
+ * X also interprets scancode 0x5d (KEY_Begin).
+ *
+ * For 1-88 keycode equals scancode.
+ */
+
+#define E0_KPENTER 96
+#define E0_RCTRL   97
+#define E0_KPSLASH 98
+#define E0_PRSCR   99
+#define E0_RALT    100
+#define E0_BREAK   101		/* (control-pause) */
+#define E0_HOME    102
+#define E0_UP      103
+#define E0_PGUP    104
+#define E0_LEFT    105
+#define E0_RIGHT   106
+#define E0_END     107
+#define E0_DOWN    108
+#define E0_PGDN    109
+#define E0_INS     110
+#define E0_DEL     111
+
+/* for USB 106 keyboard */
+#define E0_YEN         124
+#define E0_BACKSLASH   89
+
+
+#define E1_PAUSE   119
+
+/*
+ * The keycodes below are randomly located in 89-95,112-118,120-127.
+ * They could be thrown away (and all occurrences below replaced by 0),
+ * but that would force many users to use the `setkeycodes' utility, where
+ * they needed not before. It does not matter that there are duplicates, as
+ * long as no duplication occurs for any single keyboard.
+ */
+#define SC_LIM 89
+
+#define FOCUS_PF1 85		/* actual code! */
+#define FOCUS_PF2 89
+#define FOCUS_PF3 90
+#define FOCUS_PF4 91
+#define FOCUS_PF5 92
+#define FOCUS_PF6 93
+#define FOCUS_PF7 94
+#define FOCUS_PF8 95
+#define FOCUS_PF9 120
+#define FOCUS_PF10 121
+#define FOCUS_PF11 122
+#define FOCUS_PF12 123
+
+#define JAP_86     124
+/* tfj@olivia.ping.dk:
+ * The four keys are located over the numeric keypad, and are
+ * labelled A1-A4. It's an rc930 keyboard, from
+ * Regnecentralen/RC International, Now ICL.
+ * Scancodes: 59, 5a, 5b, 5c.
+ */
+#define RGN1 124
+#define RGN2 125
+#define RGN3 126
+#define RGN4 127
+
+static unsigned char high_keys[128 - SC_LIM] = {
+	RGN1, RGN2, RGN3, RGN4, 0, 0, 0,	/* 0x59-0x5f */
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x60-0x67 */
+	0, 0, 0, 0, 0, FOCUS_PF11, 0, FOCUS_PF12,	/* 0x68-0x6f */
+	0, 0, 0, FOCUS_PF2, FOCUS_PF9, 0, 0, FOCUS_PF3,	/* 0x70-0x77 */
+	FOCUS_PF4, FOCUS_PF5, FOCUS_PF6, FOCUS_PF7,	/* 0x78-0x7b */
+	FOCUS_PF8, JAP_86, FOCUS_PF10, 0	/* 0x7c-0x7f */
+};
+
+/* BTC */
+#define E0_MACRO   112
+/* LK450 */
+#define E0_F13     113
+#define E0_F14     114
+#define E0_HELP    115
+#define E0_DO      116
+#define E0_F17     117
+#define E0_KPMINPLUS 118
+/*
+ * My OmniKey generates e0 4c for  the "OMNI" key and the
+ * right alt key does nada. [kkoller@nyx10.cs.du.edu]
+ */
+#define E0_OK  124
+/*
+ * New microsoft keyboard is rumoured to have
+ * e0 5b (left window button), e0 5c (right window button),
+ * e0 5d (menu button). [or: LBANNER, RBANNER, RMENU]
+ * [or: Windows_L, Windows_R, TaskMan]
+ */
+#define E0_MSLW        125
+#define E0_MSRW        126
+#define E0_MSTM        127
+
+static unsigned char e0_keys[128] = {
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x00-0x07 */
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x08-0x0f */
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x10-0x17 */
+	0, 0, 0, 0, E0_KPENTER, E0_RCTRL, 0, 0,	/* 0x18-0x1f */
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x20-0x27 */
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x28-0x2f */
+	0, 0, 0, 0, 0, E0_KPSLASH, 0, E0_PRSCR,	/* 0x30-0x37 */
+	E0_RALT, 0, 0, 0, 0, E0_F13, E0_F14, E0_HELP,	/* 0x38-0x3f */
+	E0_DO, E0_F17, 0, 0, 0, 0, E0_BREAK, E0_HOME,	/* 0x40-0x47 */
+	E0_UP, E0_PGUP, 0, E0_LEFT, E0_OK, E0_RIGHT, E0_KPMINPLUS, E0_END,	/* 0x48-0x4f */
+	E0_DOWN, E0_PGDN, E0_INS, E0_DEL, 0, 0, 0, 0,	/* 0x50-0x57 */
+	0, 0, 0, E0_MSLW, E0_MSRW, E0_MSTM, 0, 0,	/* 0x58-0x5f */
+	0, 0, 0, 0, 0, 0, 0, 0,	/* 0x60-0x67 */
+	0, 0, 0, 0, 0, 0, 0, E0_MACRO,	/* 0x68-0x6f */
+	//0, 0, 0, 0, 0, 0, 0, 0,                          /* 0x70-0x77 */
+	0, 0, 0, 0, 0, E0_BACKSLASH, 0, 0,	/* 0x70-0x77 */
+	0, 0, 0, E0_YEN, 0, 0, 0, 0	/* 0x78-0x7f */
+};
+
+static int gen_setkeycode(unsigned int scancode, unsigned int keycode)
+{
+	if (scancode < SC_LIM || scancode > 255 || keycode > 127)
+		return -EINVAL;
+	if (scancode < 128)
+		high_keys[scancode - SC_LIM] = keycode;
+	else
+		e0_keys[scancode - 128] = keycode;
+	return 0;
+}
+
+static int gen_getkeycode(unsigned int scancode)
+{
+	return
+	    (scancode < SC_LIM || scancode > 255) ? -EINVAL :
+	    (scancode <
+	     128) ? high_keys[scancode - SC_LIM] : e0_keys[scancode - 128];
+}
+
+static int
+gen_translate(unsigned char scancode, unsigned char *keycode, char raw_mode)
+{
+	static int prev_scancode;
+
+	/* special prefix scancodes.. */
+	if (scancode == 0xe0 || scancode == 0xe1) {
+		prev_scancode = scancode;
+		return 0;
+	}
+
+	/* 0xFF is sent by a few keyboards, ignore it. 0x00 is error */
+	if (scancode == 0x00 || scancode == 0xff) {
+		prev_scancode = 0;
+		return 0;
+	}
+
+	scancode &= 0x7f;
+
+	if (prev_scancode) {
+		/*
+		 * usually it will be 0xe0, but a Pause key generates
+		 * e1 1d 45 e1 9d c5 when pressed, and nothing when released
+		 */
+		if (prev_scancode != 0xe0) {
+			if (prev_scancode == 0xe1 && scancode == 0x1d) {
+				prev_scancode = 0x100;
+				return 0;
+			}
+				else if (prev_scancode == 0x100
+					 && scancode == 0x45) {
+				*keycode = E1_PAUSE;
+				prev_scancode = 0;
+			} else {
+#ifdef KBD_REPORT_UNKN
+				if (!raw_mode)
+					printk(KERN_INFO
+					       "keyboard: unknown e1 escape sequence\n");
+#endif
+				prev_scancode = 0;
+				return 0;
+			}
+		} else {
+			prev_scancode = 0;
+			/*
+			 *  The keyboard maintains its own internal caps lock and
+			 *  num lock statuses. In caps lock mode E0 AA precedes make
+			 *  code and E0 2A follows break code. In num lock mode,
+			 *  E0 2A precedes make code and E0 AA follows break code.
+			 *  We do our own book-keeping, so we will just ignore these.
+			 */
+			/*
+			 *  For my keyboard there is no caps lock mode, but there are
+			 *  both Shift-L and Shift-R modes. The former mode generates
+			 *  E0 2A / E0 AA pairs, the latter E0 B6 / E0 36 pairs.
+			 *  So, we should also ignore the latter. - aeb@cwi.nl
+			 */
+			if (scancode == 0x2a || scancode == 0x36)
+				return 0;
+
+			if (e0_keys[scancode])
+				*keycode = e0_keys[scancode];
+			else {
+#ifdef KBD_REPORT_UNKN
+				if (!raw_mode)
+					printk(KERN_INFO
+					       "keyboard: unknown scancode e0 %02x\n",
+					       scancode);
+#endif
+				return 0;
+			}
+		}
+	} else if (scancode >= SC_LIM) {
+		/* This happens with the FOCUS 9000 keyboard
+		   Its keys PF1..PF12 are reported to generate
+		   55 73 77 78 79 7a 7b 7c 74 7e 6d 6f
+		   Moreover, unless repeated, they do not generate
+		   key-down events, so we have to zero up_flag below */
+		/* Also, Japanese 86/106 keyboards are reported to
+		   generate 0x73 and 0x7d for \ - and \ | respectively. */
+		/* Also, some Brazilian keyboard is reported to produce
+		   0x73 and 0x7e for \ ? and KP-dot, respectively. */
+
+		*keycode = high_keys[scancode - SC_LIM];
+
+		if (!*keycode) {
+			if (!raw_mode) {
+#ifdef KBD_REPORT_UNKN
+				printk(KERN_INFO
+				       "keyboard: unrecognized scancode (%02x)"
+				       " - ignored\n", scancode);
+#endif
+			}
+			return 0;
+		}
+	} else
+		*keycode = scancode;
+	return 1;
+}
+
+static char gen_unexpected_up(unsigned char keycode)
+{
+	/* unexpected, but this can happen: maybe this was a key release for a
+	   FOCUS 9000 PF key; if we want to see it, we have to clear up_flag */
+	if (keycode >= SC_LIM || keycode == 85)
+		return 0;
+	else
+		return 0200;
+}
+
+/*
+ * These are the default mappings
+ */
+int  (*k_setkeycode)(unsigned int, unsigned int) = gen_setkeycode;
+int  (*k_getkeycode)(unsigned int) = gen_getkeycode;
+int  (*k_translate)(unsigned char, unsigned char *, char) = gen_translate;
+char (*k_unexpected_up)(unsigned char) = gen_unexpected_up;
+void (*k_leds)(unsigned char);
+
+/* Simple translation table for the SysRq keys */
+
+#ifdef CONFIG_MAGIC_SYSRQ
+static unsigned char gen_sysrq_xlate[128] =
+	"\000\0331234567890-=\177\t"	/* 0x00 - 0x0f */
+	"qwertyuiop[]\r\000as"	/* 0x10 - 0x1f */
+	"dfghjkl;'`\000\\zxcv"	/* 0x20 - 0x2f */
+	"bnm,./\000*\000 \000\201\202\203\204\205"	/* 0x30 - 0x3f */
+	"\206\207\210\211\212\000\000789-456+1"	/* 0x40 - 0x4f */
+	"230\177\000\000\213\214\000\000\000\000\000\000\000\000\000\000"	/* 0x50 - 0x5f */
+	"\r\000/";			/* 0x60 - 0x6f */
+
+unsigned char *k_sysrq_xlate = gen_sysrq_xlate;
+int k_sysrq_key = 0x54;
+#endif
diff --git a/arch/arm26/lib/lib1funcs.S b/arch/arm26/lib/lib1funcs.S
new file mode 100644
index 0000000..b8f9518
--- /dev/null
+++ b/arch/arm26/lib/lib1funcs.S
@@ -0,0 +1,314 @@
+@ libgcc1 routines for ARM cpu.
+@ Division routines, written by Richard Earnshaw, (rearnsha@armltd.co.uk)
+
+/* Copyright (C) 1995, 1996, 1998 Free Software Foundation, Inc.
+
+This file is free software; you can redistribute it and/or modify it
+under the terms of the GNU General Public License as published by the
+Free Software Foundation; either version 2, or (at your option) any
+later version.
+
+In addition to the permissions in the GNU General Public License, the
+Free Software Foundation gives you unlimited permission to link the
+compiled version of this file with other programs, and to distribute
+those programs without any restriction coming from the use of this
+file.  (The General Public License restrictions do apply in other
+respects; for example, they cover modification of the file, and
+distribution when not linked into another program.)
+
+This file is distributed in the hope that it will be useful, but
+WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
+General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with this program; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* This code is derived from gcc 2.95.3 */
+/* I Molton     29/07/01 */
+
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+#include <asm/hardware.h>
+#include <linux/config.h>
+
+#define RET	movs
+#define RETc(x)	mov##x##s
+#define RETCOND ^
+
+dividend	.req	r0
+divisor		.req	r1
+result		.req	r2
+overdone        .req    r2
+curbit		.req	r3
+ip		.req	r12
+sp		.req	r13
+lr		.req	r14
+pc		.req	r15
+	
+ENTRY(__udivsi3)
+	cmp	divisor, #0
+	beq	Ldiv0
+	mov	curbit, #1
+	mov	result, #0
+	cmp	dividend, divisor
+	bcc	Lgot_result_udivsi3
+1:
+	@ Unless the divisor is very big, shift it up in multiples of
+	@ four bits, since this is the amount of unwinding in the main
+	@ division loop.  Continue shifting until the divisor is 
+	@ larger than the dividend.
+	cmp	divisor, #0x10000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #4
+	movcc	curbit, curbit, lsl #4
+	bcc	1b
+
+2:
+	@ For very big divisors, we must shift it a bit at a time, or
+	@ we will be in danger of overflowing.
+	cmp	divisor, #0x80000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #1
+	movcc	curbit, curbit, lsl #1
+	bcc	2b
+
+3:
+	@ Test for possible subtractions, and note which bits
+	@ are done in the result.  On the final pass, this may subtract
+	@ too much from the dividend, but the result will be ok, since the
+	@ "bit" will have been shifted out at the bottom.
+	cmp	dividend, divisor
+	subcs	dividend, dividend, divisor
+	orrcs	result, result, curbit
+	cmp	dividend, divisor, lsr #1
+	subcs	dividend, dividend, divisor, lsr #1
+	orrcs	result, result, curbit, lsr #1
+	cmp	dividend, divisor, lsr #2
+	subcs	dividend, dividend, divisor, lsr #2
+	orrcs	result, result, curbit, lsr #2
+	cmp	dividend, divisor, lsr #3
+	subcs	dividend, dividend, divisor, lsr #3
+	orrcs	result, result, curbit, lsr #3
+	cmp	dividend, #0			@ Early termination?
+	movnes	curbit, curbit, lsr #4		@ No, any more bits to do?
+	movne	divisor, divisor, lsr #4
+	bne	3b
+Lgot_result_udivsi3:
+	mov	r0, result
+	RET	pc, lr
+
+Ldiv0:
+	str	lr, [sp, #-4]!
+	bl	__div0
+	mov	r0, #0			@ about as wrong as it could be
+	ldmia	sp!, {pc}RETCOND
+
+/* __umodsi3 ----------------------- */
+
+ENTRY(__umodsi3)
+	cmp	divisor, #0
+	beq	Ldiv0
+	mov	curbit, #1
+	cmp	dividend, divisor
+	RETc(cc)	pc, lr
+1:
+	@ Unless the divisor is very big, shift it up in multiples of
+	@ four bits, since this is the amount of unwinding in the main
+	@ division loop.  Continue shifting until the divisor is 
+	@ larger than the dividend.
+	cmp	divisor, #0x10000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #4
+	movcc	curbit, curbit, lsl #4
+	bcc	1b
+
+2:
+	@ For very big divisors, we must shift it a bit at a time, or
+	@ we will be in danger of overflowing.
+	cmp	divisor, #0x80000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #1
+	movcc	curbit, curbit, lsl #1
+	bcc	2b
+
+3:
+	@ Test for possible subtractions.  On the final pass, this may 
+	@ subtract too much from the dividend, so keep track of which
+	@ subtractions are done, we can fix them up afterwards...
+	mov	overdone, #0
+	cmp	dividend, divisor
+	subcs	dividend, dividend, divisor
+	cmp	dividend, divisor, lsr #1
+	subcs	dividend, dividend, divisor, lsr #1
+	orrcs	overdone, overdone, curbit, ror #1
+	cmp	dividend, divisor, lsr #2
+	subcs	dividend, dividend, divisor, lsr #2
+	orrcs	overdone, overdone, curbit, ror #2
+	cmp	dividend, divisor, lsr #3
+	subcs	dividend, dividend, divisor, lsr #3
+	orrcs	overdone, overdone, curbit, ror #3
+	mov	ip, curbit
+	cmp	dividend, #0			@ Early termination?
+	movnes	curbit, curbit, lsr #4		@ No, any more bits to do?
+	movne	divisor, divisor, lsr #4
+	bne	3b
+
+	@ Any subtractions that we should not have done will be recorded in
+	@ the top three bits of "overdone".  Exactly which were not needed
+	@ are governed by the position of the bit, stored in ip.
+	@ If we terminated early, because dividend became zero,
+	@ then none of the below will match, since the bit in ip will not be
+	@ in the bottom nibble.
+	ands	overdone, overdone, #0xe0000000
+	RETc(eq)	pc, lr				@ No fixups needed
+	tst	overdone, ip, ror #3
+	addne	dividend, dividend, divisor, lsr #3
+	tst	overdone, ip, ror #2
+	addne	dividend, dividend, divisor, lsr #2
+	tst	overdone, ip, ror #1
+	addne	dividend, dividend, divisor, lsr #1
+	RET	pc, lr
+
+ENTRY(__divsi3)
+	eor	ip, dividend, divisor		@ Save the sign of the result.
+	mov	curbit, #1
+	mov	result, #0
+	cmp	divisor, #0
+	rsbmi	divisor, divisor, #0		@ Loops below use unsigned.
+	beq	Ldiv0
+	cmp	dividend, #0
+	rsbmi	dividend, dividend, #0
+	cmp	dividend, divisor
+	bcc	Lgot_result_divsi3
+
+1:
+	@ Unless the divisor is very big, shift it up in multiples of
+	@ four bits, since this is the amount of unwinding in the main
+	@ division loop.  Continue shifting until the divisor is 
+	@ larger than the dividend.
+	cmp	divisor, #0x10000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #4
+	movcc	curbit, curbit, lsl #4
+	bcc	1b
+
+2:
+	@ For very big divisors, we must shift it a bit at a time, or
+	@ we will be in danger of overflowing.
+	cmp	divisor, #0x80000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #1
+	movcc	curbit, curbit, lsl #1
+	bcc	2b
+
+3:
+	@ Test for possible subtractions, and note which bits
+	@ are done in the result.  On the final pass, this may subtract
+	@ too much from the dividend, but the result will be ok, since the
+	@ "bit" will have been shifted out at the bottom.
+	cmp	dividend, divisor
+	subcs	dividend, dividend, divisor
+	orrcs	result, result, curbit
+	cmp	dividend, divisor, lsr #1
+	subcs	dividend, dividend, divisor, lsr #1
+	orrcs	result, result, curbit, lsr #1
+	cmp	dividend, divisor, lsr #2
+	subcs	dividend, dividend, divisor, lsr #2
+	orrcs	result, result, curbit, lsr #2
+	cmp	dividend, divisor, lsr #3
+	subcs	dividend, dividend, divisor, lsr #3
+	orrcs	result, result, curbit, lsr #3
+	cmp	dividend, #0			@ Early termination?
+	movnes	curbit, curbit, lsr #4		@ No, any more bits to do?
+	movne	divisor, divisor, lsr #4
+	bne	3b
+Lgot_result_divsi3:
+	mov	r0, result
+	cmp	ip, #0
+	rsbmi	r0, r0, #0
+	RET	pc, lr
+
+ENTRY(__modsi3)
+	mov	curbit, #1
+	cmp	divisor, #0
+	rsbmi	divisor, divisor, #0		@ Loops below use unsigned.
+	beq	Ldiv0
+	@ Need to save the sign of the dividend, unfortunately, we need
+	@ ip later on; this is faster than pushing lr and using that.
+	str	dividend, [sp, #-4]!
+	cmp	dividend, #0
+	rsbmi	dividend, dividend, #0
+	cmp	dividend, divisor
+	bcc	Lgot_result_modsi3
+
+1:
+	@ Unless the divisor is very big, shift it up in multiples of
+	@ four bits, since this is the amount of unwinding in the main
+	@ division loop.  Continue shifting until the divisor is 
+	@ larger than the dividend.
+	cmp	divisor, #0x10000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #4
+	movcc	curbit, curbit, lsl #4
+	bcc	1b
+
+2:
+	@ For very big divisors, we must shift it a bit at a time, or
+	@ we will be in danger of overflowing.
+	cmp	divisor, #0x80000000
+	cmpcc	divisor, dividend
+	movcc	divisor, divisor, lsl #1
+	movcc	curbit, curbit, lsl #1
+	bcc	2b
+
+3:
+	@ Test for possible subtractions.  On the final pass, this may 
+	@ subtract too much from the dividend, so keep track of which
+	@ subtractions are done, we can fix them up afterwards...
+	mov	overdone, #0
+	cmp	dividend, divisor
+	subcs	dividend, dividend, divisor
+	cmp	dividend, divisor, lsr #1
+	subcs	dividend, dividend, divisor, lsr #1
+	orrcs	overdone, overdone, curbit, ror #1
+	cmp	dividend, divisor, lsr #2
+	subcs	dividend, dividend, divisor, lsr #2
+	orrcs	overdone, overdone, curbit, ror #2
+	cmp	dividend, divisor, lsr #3
+	subcs	dividend, dividend, divisor, lsr #3
+	orrcs	overdone, overdone, curbit, ror #3
+	mov	ip, curbit
+	cmp	dividend, #0			@ Early termination?
+	movnes	curbit, curbit, lsr #4		@ No, any more bits to do?
+	movne	divisor, divisor, lsr #4
+	bne	3b
+
+	@ Any subtractions that we should not have done will be recorded in
+	@ the top three bits of "overdone".  Exactly which were not needed
+	@ are governed by the position of the bit, stored in ip.
+	@ If we terminated early, because dividend became zero,
+	@ then none of the below will match, since the bit in ip will not be
+	@ in the bottom nibble.
+	ands	overdone, overdone, #0xe0000000
+	beq	Lgot_result_modsi3
+	tst	overdone, ip, ror #3
+	addne	dividend, dividend, divisor, lsr #3
+	tst	overdone, ip, ror #2
+	addne	dividend, dividend, divisor, lsr #2
+	tst	overdone, ip, ror #1
+	addne	dividend, dividend, divisor, lsr #1
+Lgot_result_modsi3:
+	ldr	ip, [sp], #4
+	cmp	ip, #0
+	rsbmi	dividend, dividend, #0
+	RET	pc, lr
diff --git a/arch/arm26/lib/longlong.h b/arch/arm26/lib/longlong.h
new file mode 100644
index 0000000..05ec1ab
--- /dev/null
+++ b/arch/arm26/lib/longlong.h
@@ -0,0 +1,184 @@
+/* longlong.h -- based on code from gcc-2.95.3
+
+   definitions for mixed size 32/64 bit arithmetic.
+   Copyright (C) 1991, 92, 94, 95, 96, 1997, 1998 Free Software Foundation, Inc.
+
+   This definition file is free software; you can redistribute it
+   and/or modify it under the terms of the GNU General Public
+   License as published by the Free Software Foundation; either
+   version 2, or (at your option) any later version.
+
+   This definition file is distributed in the hope that it will be
+   useful, but WITHOUT ANY WARRANTY; without even the implied
+   warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+   See the GNU General Public License for more details.
+
+   You should have received a copy of the GNU General Public License
+   along with this program; if not, write to the Free Software
+   Foundation, Inc., 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+/* Borrowed from GCC 2.95.3, I Molton 29/07/01 */
+
+#ifndef SI_TYPE_SIZE
+#define SI_TYPE_SIZE 32
+#endif
+
+#define __BITS4 (SI_TYPE_SIZE / 4)
+#define __ll_B (1L << (SI_TYPE_SIZE / 2))
+#define __ll_lowpart(t) ((USItype) (t) % __ll_B)
+#define __ll_highpart(t) ((USItype) (t) / __ll_B)
+
+/* Define auxiliary asm macros.
+
+   1) umul_ppmm(high_prod, low_prod, multipler, multiplicand)
+   multiplies two USItype integers MULTIPLER and MULTIPLICAND,
+   and generates a two-part USItype product in HIGH_PROD and
+   LOW_PROD.
+
+   2) __umulsidi3(a,b) multiplies two USItype integers A and B,
+   and returns a UDItype product.  This is just a variant of umul_ppmm.
+
+   3) udiv_qrnnd(quotient, remainder, high_numerator, low_numerator,
+   denominator) divides a two-word unsigned integer, composed by the
+   integers HIGH_NUMERATOR and LOW_NUMERATOR, by DENOMINATOR and
+   places the quotient in QUOTIENT and the remainder in REMAINDER.
+   HIGH_NUMERATOR must be less than DENOMINATOR for correct operation.
+   If, in addition, the most significant bit of DENOMINATOR must be 1,
+   then the pre-processor symbol UDIV_NEEDS_NORMALIZATION is defined to 1.
+
+   4) sdiv_qrnnd(quotient, remainder, high_numerator, low_numerator,
+   denominator).  Like udiv_qrnnd but the numbers are signed.  The
+   quotient is rounded towards 0.
+
+   5) count_leading_zeros(count, x) counts the number of zero-bits from
+   the msb to the first non-zero bit.  This is the number of steps X
+   needs to be shifted left to set the msb.  Undefined for X == 0.
+
+   6) add_ssaaaa(high_sum, low_sum, high_addend_1, low_addend_1,
+   high_addend_2, low_addend_2) adds two two-word unsigned integers,
+   composed by HIGH_ADDEND_1 and LOW_ADDEND_1, and HIGH_ADDEND_2 and
+   LOW_ADDEND_2 respectively.  The result is placed in HIGH_SUM and
+   LOW_SUM.  Overflow (i.e. carry out) is not stored anywhere, and is
+   lost.
+
+   7) sub_ddmmss(high_difference, low_difference, high_minuend,
+   low_minuend, high_subtrahend, low_subtrahend) subtracts two
+   two-word unsigned integers, composed by HIGH_MINUEND_1 and
+   LOW_MINUEND_1, and HIGH_SUBTRAHEND_2 and LOW_SUBTRAHEND_2
+   respectively.  The result is placed in HIGH_DIFFERENCE and
+   LOW_DIFFERENCE.  Overflow (i.e. carry out) is not stored anywhere,
+   and is lost.
+
+   If any of these macros are left undefined for a particular CPU,
+   C macros are used.  */
+
+#if defined (__arm__)
+#define add_ssaaaa(sh, sl, ah, al, bh, bl) \
+  __asm__ ("adds	%1, %4, %5					\n\
+	adc	%0, %2, %3"						\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "%r" ((USItype) (ah)),					\
+	     "rI" ((USItype) (bh)),					\
+	     "%r" ((USItype) (al)),					\
+	     "rI" ((USItype) (bl)))
+#define sub_ddmmss(sh, sl, ah, al, bh, bl) \
+  __asm__ ("subs	%1, %4, %5					\n\
+	sbc	%0, %2, %3"						\
+	   : "=r" ((USItype) (sh)),					\
+	     "=&r" ((USItype) (sl))					\
+	   : "r" ((USItype) (ah)),					\
+	     "rI" ((USItype) (bh)),					\
+	     "r" ((USItype) (al)),					\
+	     "rI" ((USItype) (bl)))
+#define umul_ppmm(xh, xl, a, b) \
+{register USItype __t0, __t1, __t2;					\
+  __asm__ ("%@ Inlined umul_ppmm					\n\
+	mov	%2, %5, lsr #16						\n\
+	mov	%0, %6, lsr #16						\n\
+	bic	%3, %5, %2, lsl #16					\n\
+	bic	%4, %6, %0, lsl #16					\n\
+	mul	%1, %3, %4						\n\
+	mul	%4, %2, %4						\n\
+	mul	%3, %0, %3						\n\
+	mul	%0, %2, %0						\n\
+	adds	%3, %4, %3						\n\
+	addcs	%0, %0, #65536						\n\
+	adds	%1, %1, %3, lsl #16					\n\
+	adc	%0, %0, %3, lsr #16"					\
+	   : "=&r" ((USItype) (xh)),					\
+	     "=r" ((USItype) (xl)),					\
+	     "=&r" (__t0), "=&r" (__t1), "=r" (__t2)			\
+	   : "r" ((USItype) (a)),					\
+	     "r" ((USItype) (b)));}
+#define UMUL_TIME 20
+#define UDIV_TIME 100
+#endif /* __arm__ */
+
+#define __umulsidi3(u, v) \
+  ({DIunion __w;							\
+    umul_ppmm (__w.s.high, __w.s.low, u, v);				\
+    __w.ll; })
+
+#define __udiv_qrnnd_c(q, r, n1, n0, d) \
+  do {									\
+    USItype __d1, __d0, __q1, __q0;					\
+    USItype __r1, __r0, __m;						\
+    __d1 = __ll_highpart (d);						\
+    __d0 = __ll_lowpart (d);						\
+									\
+    __r1 = (n1) % __d1;							\
+    __q1 = (n1) / __d1;							\
+    __m = (USItype) __q1 * __d0;					\
+    __r1 = __r1 * __ll_B | __ll_highpart (n0);				\
+    if (__r1 < __m)							\
+      {									\
+	__q1--, __r1 += (d);						\
+	if (__r1 >= (d)) /* i.e. we didn't get carry when adding to __r1 */\
+	  if (__r1 < __m)						\
+	    __q1--, __r1 += (d);					\
+      }									\
+    __r1 -= __m;							\
+									\
+    __r0 = __r1 % __d1;							\
+    __q0 = __r1 / __d1;							\
+    __m = (USItype) __q0 * __d0;					\
+    __r0 = __r0 * __ll_B | __ll_lowpart (n0);				\
+    if (__r0 < __m)							\
+      {									\
+	__q0--, __r0 += (d);						\
+	if (__r0 >= (d))						\
+	  if (__r0 < __m)						\
+	    __q0--, __r0 += (d);					\
+      }									\
+    __r0 -= __m;							\
+									\
+    (q) = (USItype) __q1 * __ll_B | __q0;				\
+    (r) = __r0;								\
+  } while (0)
+
+#define UDIV_NEEDS_NORMALIZATION 1
+#define udiv_qrnnd __udiv_qrnnd_c
+
+extern const UQItype __clz_tab[];
+#define count_leading_zeros(count, x) \
+  do {									\
+    USItype __xr = (x);							\
+    USItype __a;							\
+									\
+    if (SI_TYPE_SIZE <= 32)						\
+      {									\
+	__a = __xr < ((USItype)1<<2*__BITS4)				\
+	  ? (__xr < ((USItype)1<<__BITS4) ? 0 : __BITS4)		\
+	  : (__xr < ((USItype)1<<3*__BITS4) ?  2*__BITS4 : 3*__BITS4);	\
+      }									\
+    else								\
+      {									\
+	for (__a = SI_TYPE_SIZE - 8; __a > 0; __a -= 8)			\
+	  if (((__xr >> __a) & 0xff) != 0)				\
+	    break;							\
+      }									\
+									\
+    (count) = SI_TYPE_SIZE - (__clz_tab[__xr >> __a] + __a);		\
+  } while (0)
diff --git a/arch/arm26/lib/lshrdi3.c b/arch/arm26/lib/lshrdi3.c
new file mode 100644
index 0000000..b666f1b
--- /dev/null
+++ b/arch/arm26/lib/lshrdi3.c
@@ -0,0 +1,61 @@
+/* More subroutines needed by GCC output code on some machines.  */
+/* Compile this one with gcc.  */
+/* Copyright (C) 1989, 92-98, 1999 Free Software Foundation, Inc.
+
+This file is part of GNU CC.
+
+GNU CC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GNU CC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GNU CC; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* support functions required by the kernel. based on code from gcc-2.95.3 */
+/* I Molton     29/07/01 */
+
+#include "gcclib.h"
+
+DItype
+__lshrdi3 (DItype u, word_type b)
+{
+  DIunion w;
+  word_type bm;
+  DIunion uu;
+
+  if (b == 0)
+    return u;
+
+  uu.ll = u;
+
+  bm = (sizeof (SItype) * BITS_PER_UNIT) - b;
+  if (bm <= 0)
+    {
+      w.s.high = 0;
+      w.s.low = (USItype)uu.s.high >> -bm;
+    }
+  else
+    {
+      USItype carries = (USItype)uu.s.high << bm;
+      w.s.high = (USItype)uu.s.high >> b;
+      w.s.low = ((USItype)uu.s.low >> b) | carries;
+    }
+
+  return w.ll;
+}
+
diff --git a/arch/arm26/lib/memchr.S b/arch/arm26/lib/memchr.S
new file mode 100644
index 0000000..34e7c14
--- /dev/null
+++ b/arch/arm26/lib/memchr.S
@@ -0,0 +1,25 @@
+/*
+ *  linux/arch/arm26/lib/memchr.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.
+ *
+ *  ASM optimised string functions
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+	.text
+	.align	5
+ENTRY(memchr)
+1:	subs	r2, r2, #1
+	bmi	2f
+	ldrb	r3, [r0], #1
+	teq	r3, r1
+	bne	1b
+	sub	r0, r0, #1
+2:	movne	r0, #0
+	RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/memcpy.S b/arch/arm26/lib/memcpy.S
new file mode 100644
index 0000000..3f719e4
--- /dev/null
+++ b/arch/arm26/lib/memcpy.S
@@ -0,0 +1,318 @@
+/*
+ *  linux/arch/arm26/lib/memcpy.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 <asm/assembler.h>
+
+		.text
+
+#define ENTER	\
+		mov	ip,sp	;\
+		stmfd	sp!,{r4-r9,fp,ip,lr,pc}	;\
+		sub	fp,ip,#4
+
+#define EXIT	\
+		LOADREGS(ea, fp, {r4 - r9, fp, sp, pc})
+
+#define EXITEQ	\
+		LOADREGS(eqea, fp, {r4 - r9, fp, sp, pc})
+
+/*
+ * Prototype: void memcpy(void *to,const void *from,unsigned long n);
+ * ARM3: cant use memcopy here!!!
+ */
+ENTRY(memcpy)
+ENTRY(memmove)
+		ENTER
+		cmp	r1, r0
+		bcc	19f
+		subs	r2, r2, #4
+		blt	6f
+		ands	ip, r0, #3
+		bne	7f
+		ands	ip, r1, #3
+		bne	8f
+
+1:		subs	r2, r2, #8
+		blt	5f
+		subs	r2, r2, #0x14
+		blt	3f
+2:		ldmia	r1!,{r3 - r9, ip}
+		stmia	r0!,{r3 - r9, ip}
+		subs	r2, r2, #32
+		bge	2b
+		cmn	r2, #16
+		ldmgeia	r1!, {r3 - r6}
+		stmgeia	r0!, {r3 - r6}
+		subge	r2, r2, #0x10
+3:		adds	r2, r2, #0x14
+4:		ldmgeia	r1!, {r3 - r5}
+		stmgeia	r0!, {r3 - r5}
+		subges	r2, r2, #12
+		bge	4b
+5:		adds	r2, r2, #8
+		blt	6f
+		subs	r2, r2, #4
+		ldrlt	r3, [r1], #4
+		ldmgeia	r1!, {r4, r5}
+		strlt	r3, [r0], #4
+		stmgeia	r0!, {r4, r5}
+		subge	r2, r2, #4
+
+6:		adds	r2, r2, #4
+		EXITEQ
+		cmp	r2, #2
+		ldrb	r3, [r1], #1
+		ldrgeb	r4, [r1], #1
+		ldrgtb	r5, [r1], #1
+		strb	r3, [r0], #1
+		strgeb	r4, [r0], #1
+		strgtb	r5, [r0], #1
+		EXIT
+
+7:		rsb	ip, ip, #4
+		cmp	ip, #2
+		ldrb	r3, [r1], #1
+		ldrgeb	r4, [r1], #1
+		ldrgtb	r5, [r1], #1
+		strb	r3, [r0], #1
+		strgeb	r4, [r0], #1
+		strgtb	r5, [r0], #1
+		subs	r2, r2, ip
+		blt	6b
+		ands	ip, r1, #3
+		beq	1b
+
+8:		bic	r1, r1, #3
+		ldr	r7, [r1], #4
+		cmp	ip, #2
+		bgt	15f
+		beq	11f
+		cmp	r2, #12
+		blt	10f
+		sub	r2, r2, #12
+9:		mov	r3, r7, pull #8
+		ldmia	r1!, {r4 - r7}
+		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}
+		subs	r2, r2, #16
+		bge	9b
+		adds	r2, r2, #12
+		blt	100f
+10:		mov	r3, r7, pull #8
+		ldr	r7, [r1], #4
+		subs	r2, r2, #4
+		orr	r3, r3, r7, push #24
+		str	r3, [r0], #4
+		bge	10b
+100:		sub	r1, r1, #3
+		b	6b
+
+11:		cmp	r2, #12
+		blt	13f		/* */
+		sub	r2, r2, #12
+12:		mov	r3, r7, pull #16
+		ldmia	r1!, {r4 - r7}
+		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}
+		subs	r2, r2, #16
+		bge	12b
+		adds	r2, r2, #12
+		blt	14f
+13:		mov	r3, r7, pull #16
+		ldr	r7, [r1], #4
+		subs	r2, r2, #4
+		orr	r3, r3, r7, push #16
+		str	r3, [r0], #4
+		bge	13b
+14:		sub	r1, r1, #2
+		b	6b
+
+15:		cmp	r2, #12
+		blt	17f
+		sub	r2, r2, #12
+16:		mov	r3, r7, pull #24
+		ldmia	r1!, {r4 - r7}
+		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	r2, r2, #16
+		bge	16b
+		adds	r2, r2, #12
+		blt	18f
+17:		mov	r3, r7, pull #24
+		ldr	r7, [r1], #4
+		subs	r2, r2, #4
+		orr	r3, r3, r7, push #8
+		str	r3, [r0], #4
+		bge	17b
+18:		sub	r1, r1, #1
+		b	6b
+
+
+19:		add	r1, r1, r2
+		add	r0, r0, r2
+		subs	r2, r2, #4
+		blt	24f
+		ands	ip, r0, #3
+		bne	25f
+		ands	ip, r1, #3
+		bne	26f
+
+20:		subs	r2, r2, #8
+		blt	23f
+		subs	r2, r2, #0x14
+		blt	22f
+21:		ldmdb	r1!, {r3 - r9, ip}
+		stmdb	r0!, {r3 - r9, ip}
+		subs	r2, r2, #32
+		bge	21b
+22:		cmn	r2, #16
+		ldmgedb	r1!, {r3 - r6}
+		stmgedb	r0!, {r3 - r6}
+		subge	r2, r2, #16
+		adds	r2, r2, #20
+		ldmgedb	r1!, {r3 - r5}
+		stmgedb	r0!, {r3 - r5}
+		subge	r2, r2, #12
+23:		adds	r2, r2, #8
+		blt	24f
+		subs	r2, r2, #4
+		ldrlt	r3, [r1, #-4]!
+		ldmgedb	r1!, {r4, r5}
+		strlt	r3, [r0, #-4]!
+		stmgedb	r0!, {r4, r5}
+		subge	r2, r2, #4
+
+24:		adds	r2, r2, #4
+		EXITEQ
+		cmp	r2, #2
+		ldrb	r3, [r1, #-1]!
+		ldrgeb	r4, [r1, #-1]!
+		ldrgtb	r5, [r1, #-1]!
+		strb	r3, [r0, #-1]!
+		strgeb	r4, [r0, #-1]!
+		strgtb	r5, [r0, #-1]!
+		EXIT
+
+25:		cmp	ip, #2
+		ldrb	r3, [r1, #-1]!
+		ldrgeb	r4, [r1, #-1]!
+		ldrgtb	r5, [r1, #-1]!
+		strb	r3, [r0, #-1]!
+		strgeb	r4, [r0, #-1]!
+		strgtb	r5, [r0, #-1]!
+		subs	r2, r2, ip
+		blt	24b
+		ands	ip, r1, #3
+		beq	20b
+
+26:		bic	r1, r1, #3
+		ldr	r3, [r1], #0
+		cmp	ip, #2
+		blt	34f
+		beq	30f
+		cmp	r2, #12
+		blt	28f
+		sub	r2, r2, #12
+27:		mov	r7, r3, push #8
+		ldmdb	r1!, {r3, r4, r5, r6}
+		orr	r7, r7, r6, pull #24
+		mov	r6, r6, push #8
+		orr	r6, r6, r5, pull #24
+		mov	r5, r5, push #8
+		orr	r5, r5, r4, pull #24
+		mov	r4, r4, push #8
+		orr	r4, r4, r3, pull #24
+		stmdb	r0!, {r4, r5, r6, r7}
+		subs	r2, r2, #16
+		bge	27b
+		adds	r2, r2, #12
+		blt	29f
+28:		mov	ip, r3, push #8
+		ldr	r3, [r1, #-4]!
+		subs	r2, r2, #4
+		orr	ip, ip, r3, pull #24
+		str	ip, [r0, #-4]!
+		bge	28b
+29:		add	r1, r1, #3
+		b	24b
+
+30:		cmp	r2, #12
+		blt	32f
+		sub	r2, r2, #12
+31:		mov	r7, r3, push #16
+		ldmdb	r1!, {r3, r4, r5, r6}
+		orr	r7, r7, r6, pull #16
+		mov	r6, r6, push #16
+		orr	r6, r6, r5, pull #16
+		mov	r5, r5, push #16
+		orr	r5, r5, r4, pull #16
+		mov	r4, r4, push #16
+		orr	r4, r4, r3, pull #16
+		stmdb	r0!, {r4, r5, r6, r7}
+		subs	r2, r2, #16
+		bge	31b
+		adds	r2, r2, #12
+		blt	33f
+32:		mov	ip, r3, push #16
+		ldr	r3, [r1, #-4]!
+		subs	r2, r2, #4
+		orr	ip, ip, r3, pull #16
+		str	ip, [r0, #-4]!
+		bge	32b
+33:		add	r1, r1, #2
+		b	24b
+
+34:		cmp	r2, #12
+		blt	36f
+		sub	r2, r2, #12
+35:		mov	r7, r3, push #24
+		ldmdb	r1!, {r3, r4, r5, r6}
+		orr	r7, r7, r6, pull #8
+		mov	r6, r6, push #24
+		orr	r6, r6, r5, pull #8
+		mov	r5, r5, push #24
+		orr	r5, r5, r4, pull #8
+		mov	r4, r4, push #24
+		orr	r4, r4, r3, pull #8
+		stmdb	r0!, {r4, r5, r6, r7}
+		subs	r2, r2, #16
+		bge	35b
+		adds	r2, r2, #12
+		blt	37f
+36:		mov	ip, r3, push #24
+		ldr	r3, [r1, #-4]!
+		subs	r2, r2, #4
+		orr	ip, ip, r3, pull #8
+		str	ip, [r0, #-4]!
+		bge	36b
+37:		add	r1, r1, #1
+		b	24b
+
+		.align
diff --git a/arch/arm26/lib/memset.S b/arch/arm26/lib/memset.S
new file mode 100644
index 0000000..aedec10
--- /dev/null
+++ b/arch/arm26/lib/memset.S
@@ -0,0 +1,80 @@
+/*
+ *  linux/arch/arm26/lib/memset.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.
+ *
+ *  ASM optimised string functions
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+	.text
+	.align	5
+	.word	0
+
+1:	subs	r2, r2, #4		@ 1 do we have enough
+	blt	5f			@ 1 bytes to align with?
+	cmp	r3, #2			@ 1
+	strltb	r1, [r0], #1		@ 1
+	strleb	r1, [r0], #1		@ 1
+	strb	r1, [r0], #1		@ 1
+	add	r2, r2, r3		@ 1 (r2 = r2 - (4 - r3))
+/*
+ * The pointer is now aligned and the length is adjusted.  Try doing the
+ * memzero again.
+ */
+
+ENTRY(memset)
+	ands	r3, r0, #3		@ 1 unaligned?
+	bne	1b			@ 1
+/*
+ * we know that the pointer in r0 is aligned to a word boundary.
+ */
+	orr	r1, r1, r1, lsl #8
+	orr	r1, r1, r1, lsl #16
+	mov	r3, r1
+	cmp	r2, #16
+	blt	4f
+/*
+ * We need an extra register for this loop - save the return address and
+ * use the LR
+ */
+	str	lr, [sp, #-4]!
+	mov	ip, r1
+	mov	lr, r1
+
+2:	subs	r2, r2, #64
+	stmgeia	r0!, {r1, r3, ip, lr}	@ 64 bytes at a time.
+	stmgeia	r0!, {r1, r3, ip, lr}
+	stmgeia	r0!, {r1, r3, ip, lr}
+	stmgeia	r0!, {r1, r3, ip, lr}
+	bgt	2b
+	LOADREGS(eqfd, sp!, {pc})	@ Now <64 bytes to go.
+/*
+ * No need to correct the count; we're only testing bits from now on
+ */
+	tst	r2, #32
+	stmneia	r0!, {r1, r3, ip, lr}
+	stmneia	r0!, {r1, r3, ip, lr}
+	tst	r2, #16
+	stmneia	r0!, {r1, r3, ip, lr}
+	ldr	lr, [sp], #4
+
+4:	tst	r2, #8
+	stmneia	r0!, {r1, r3}
+	tst	r2, #4
+	strne	r1, [r0], #4
+/*
+ * When we get here, we've got less than 4 bytes to zero.  We
+ * may have an unaligned pointer as well.
+ */
+5:	tst	r2, #2
+	strneb	r1, [r0], #1
+	strneb	r1, [r0], #1
+	tst	r2, #1
+	strneb	r1, [r0], #1
+	RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/memzero.S b/arch/arm26/lib/memzero.S
new file mode 100644
index 0000000..cc5bf68
--- /dev/null
+++ b/arch/arm26/lib/memzero.S
@@ -0,0 +1,80 @@
+/*
+ *  linux/arch/arm26/lib/memzero.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>
+
+	.text
+	.align	5
+	.word	0
+/*
+ * Align the pointer in r0.  r3 contains the number of bytes that we are
+ * mis-aligned by, and r1 is the number of bytes.  If r1 < 4, then we
+ * don't bother; we use byte stores instead.
+ */
+1:	subs	r1, r1, #4		@ 1 do we have enough
+	blt	5f			@ 1 bytes to align with?
+	cmp	r3, #2			@ 1
+	strltb	r2, [r0], #1		@ 1
+	strleb	r2, [r0], #1		@ 1
+	strb	r2, [r0], #1		@ 1
+	add	r1, r1, r3		@ 1 (r1 = r1 - (4 - r3))
+/*
+ * The pointer is now aligned and the length is adjusted.  Try doing the
+ * memzero again.
+ */
+
+ENTRY(__memzero)
+	mov	r2, #0			@ 1
+	ands	r3, r0, #3		@ 1 unaligned?
+	bne	1b			@ 1
+/*
+ * r3 = 0, and we know that the pointer in r0 is aligned to a word boundary.
+ */
+	cmp	r1, #16			@ 1 we can skip this chunk if we
+	blt	4f			@ 1 have < 16 bytes
+/*
+ * We need an extra register for this loop - save the return address and
+ * use the LR
+ */
+	str	lr, [sp, #-4]!		@ 1
+	mov	ip, r2			@ 1
+	mov	lr, r2			@ 1
+
+3:	subs	r1, r1, #64		@ 1 write 32 bytes out per loop
+	stmgeia	r0!, {r2, r3, ip, lr}	@ 4
+	stmgeia	r0!, {r2, r3, ip, lr}	@ 4
+	stmgeia	r0!, {r2, r3, ip, lr}	@ 4
+	stmgeia	r0!, {r2, r3, ip, lr}	@ 4
+	bgt	3b			@ 1
+	LOADREGS(eqfd, sp!, {pc})	@ 1/2 quick exit
+/*
+ * No need to correct the count; we're only testing bits from now on
+ */
+	tst	r1, #32			@ 1
+	stmneia	r0!, {r2, r3, ip, lr}	@ 4
+	stmneia	r0!, {r2, r3, ip, lr}	@ 4
+	tst	r1, #16			@ 1 16 bytes or more?
+	stmneia	r0!, {r2, r3, ip, lr}	@ 4
+	ldr	lr, [sp], #4		@ 1
+
+4:	tst	r1, #8			@ 1 8 bytes or more?
+	stmneia	r0!, {r2, r3}		@ 2
+	tst	r1, #4			@ 1 4 bytes or more?
+	strne	r2, [r0], #4		@ 1
+/*
+ * When we get here, we've got less than 4 bytes to zero.  We
+ * may have an unaligned pointer as well.
+ */
+5:	tst	r1, #2			@ 1 2 bytes or more?
+	strneb	r2, [r0], #1		@ 1
+	strneb	r2, [r0], #1		@ 1
+	tst	r1, #1			@ 1 a byte left over
+	strneb	r2, [r0], #1		@ 1
+	RETINSTR(mov,pc,lr)		@ 1
diff --git a/arch/arm26/lib/muldi3.c b/arch/arm26/lib/muldi3.c
new file mode 100644
index 0000000..44d611b
--- /dev/null
+++ b/arch/arm26/lib/muldi3.c
@@ -0,0 +1,77 @@
+/* More subroutines needed by GCC output code on some machines.  */
+/* Compile this one with gcc.  */
+/* Copyright (C) 1989, 92-98, 1999 Free Software Foundation, Inc.
+
+This file is part of GNU CC.
+
+GNU CC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GNU CC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GNU CC; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* support functions required by the kernel. based on code from gcc-2.95.3 */
+/* I Molton     29/07/01 */
+
+#include "gcclib.h"
+
+#define umul_ppmm(xh, xl, a, b) \
+{register USItype __t0, __t1, __t2;                                     \
+  __asm__ ("%@ Inlined umul_ppmm					\n\
+        mov     %2, %5, lsr #16						\n\
+        mov     %0, %6, lsr #16						\n\
+        bic     %3, %5, %2, lsl #16					\n\
+        bic     %4, %6, %0, lsl #16					\n\
+        mul     %1, %3, %4						\n\
+        mul     %4, %2, %4						\n\
+        mul     %3, %0, %3						\n\
+        mul     %0, %2, %0						\n\
+        adds    %3, %4, %3						\n\
+        addcs   %0, %0, #65536						\n\
+        adds    %1, %1, %3, lsl #16					\n\
+        adc     %0, %0, %3, lsr #16"                                    \
+           : "=&r" ((USItype) (xh)),                                    \
+             "=r" ((USItype) (xl)),                                     \
+             "=&r" (__t0), "=&r" (__t1), "=r" (__t2)                    \
+           : "r" ((USItype) (a)),                                       \
+             "r" ((USItype) (b)));}
+
+
+#define __umulsidi3(u, v) \
+  ({DIunion __w;                                                        \
+    umul_ppmm (__w.s.high, __w.s.low, u, v);                            \
+    __w.ll; })
+
+
+DItype
+__muldi3 (DItype u, DItype v)
+{
+  DIunion w;
+  DIunion uu, vv;
+
+  uu.ll = u,
+  vv.ll = v;
+
+  w.ll = __umulsidi3 (uu.s.low, vv.s.low);
+  w.s.high += ((USItype) uu.s.low * (USItype) vv.s.high
+               + (USItype) uu.s.high * (USItype) vv.s.low);
+
+  return w.ll;
+}
+
diff --git a/arch/arm26/lib/putuser.S b/arch/arm26/lib/putuser.S
new file mode 100644
index 0000000..87588cb
--- /dev/null
+++ b/arch/arm26/lib/putuser.S
@@ -0,0 +1,109 @@
+/*
+ *  linux/arch/arm26/lib/putuser.S
+ *
+ *  Copyright (C) 2001 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.
+ *
+ *  Idea from x86 version, (C) Copyright 1998 Linus Torvalds
+ *
+ * These functions have a non-standard call interface to make
+ * them more efficient, especially as they return an error
+ * value in addition to the "real" return value.
+ *
+ * __put_user_X
+ *
+ * Inputs:	r0 contains the address
+ *		r1, r2 contains the value
+ * Outputs:	r0 is the error code
+ *		lr corrupted
+ *
+ * No other registers must be altered.  (see include/asm-arm/uaccess.h
+ * for specific ASM register usage).
+ *
+ * Note that ADDR_LIMIT is either 0 or 0xc0000000
+ * Note also that it is intended that __put_user_bad is not global.
+ */
+#include <asm/asm_offsets.h>
+#include <asm/thread_info.h>
+#include <asm/errno.h>
+
+        .global __put_user_1
+__put_user_1:
+        bic     r2, sp, #0x1f00
+        bic     r2, r2, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r2, [r2, #TI_ADDR_LIMIT]
+        sub     r2, r2, #1
+        cmp     r0, r2
+        bge     __put_user_bad
+1:      cmp     r0, #0x02000000
+        strlsbt r1, [r0]
+        strgeb  r1, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+        .global __put_user_2
+__put_user_2:
+        bic     r2, sp, #0x1f00
+        bic     r2, r2, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r2, [r2, #TI_ADDR_LIMIT]
+        sub     r2, r2, #2
+        cmp     r0, r2
+        bge     __put_user_bad
+2:      cmp     r0, #0x02000000
+        strlsbt r1, [r0], #1
+        strgeb  r1, [r0], #1
+        mov     r1, r1, lsr #8
+3:      strlsbt r1, [r0]
+        strgeb  r1, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+        .global __put_user_4
+__put_user_4:
+        bic     r2, sp, #0x1f00
+        bic     r2, r2, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     r2, [r2, #TI_ADDR_LIMIT]
+        sub     r2, r2, #4
+        cmp     r0, r2
+4:      bge     __put_user_bad
+        cmp     r0, #0x02000000
+        strlst  r1, [r0]
+        strge   r1, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+        .global __put_user_8
+__put_user_8:
+        bic     ip, sp, #0x1f00
+        bic     ip, ip, #0x00ff
+        str     lr, [sp, #-4]!
+        ldr     ip, [ip, #TI_ADDR_LIMIT]
+        sub     ip, ip, #8
+        cmp     r0, ip
+        bge     __put_user_bad
+        cmp     r0, #0x02000000
+5:      strlst  r1, [r0], #4
+6:      strlst  r2, [r0]
+        strge   r1, [r0], #4
+        strge   r2, [r0]
+        mov     r0, #0
+        ldmfd   sp!, {pc}^
+
+__put_user_bad:
+	mov	r0, #-EFAULT
+	mov	pc, lr
+
+.section __ex_table, "a"
+	.long	1b, __put_user_bad
+	.long	2b, __put_user_bad
+	.long	3b, __put_user_bad
+	.long	4b, __put_user_bad
+	.long	5b, __put_user_bad
+	.long	6b, __put_user_bad
+.previous
diff --git a/arch/arm26/lib/setbit.S b/arch/arm26/lib/setbit.S
new file mode 100644
index 0000000..e180c1a
--- /dev/null
+++ b/arch/arm26/lib/setbit.S
@@ -0,0 +1,29 @@
+/*
+ *  linux/arch/arm26/lib/setbit.S
+ *
+ *  Copyright (C) 1995-1996 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>
+		.text
+
+/*
+ * Purpose  : Function to set a bit
+ * Prototype: int set_bit(int bit, void *addr)
+ */
+ENTRY(_set_bit_be)
+		eor	r0, r0, #0x18		@ big endian byte ordering
+ENTRY(_set_bit_le)
+		and	r2, r0, #7
+		mov	r3, #1
+		mov	r3, r3, lsl r2
+		save_and_disable_irqs ip, r2
+		ldrb	r2, [r1, r0, lsr #3]
+		orr	r2, r2, r3
+		strb	r2, [r1, r0, lsr #3]
+		restore_irqs ip
+		RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/strchr.S b/arch/arm26/lib/strchr.S
new file mode 100644
index 0000000..ecfff21
--- /dev/null
+++ b/arch/arm26/lib/strchr.S
@@ -0,0 +1,25 @@
+/*
+ *  linux/arch/arm26/lib/strchr.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.
+ *
+ *  ASM optimised string functions
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+		.text
+		.align	5
+ENTRY(strchr)
+1:		ldrb	r2, [r0], #1
+		teq	r2, r1
+		teqne	r2, #0
+		bne	1b
+		teq	r2, #0
+		moveq	r0, #0
+		subne	r0, r0, #1
+		RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/strrchr.S b/arch/arm26/lib/strrchr.S
new file mode 100644
index 0000000..db43b28
--- /dev/null
+++ b/arch/arm26/lib/strrchr.S
@@ -0,0 +1,25 @@
+/*
+ *  linux/arch/arm26/lib/strrchr.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.
+ *
+ *  ASM optimised string functions
+ */
+#include <linux/linkage.h>
+#include <asm/assembler.h>
+
+		.text
+		.align	5
+ENTRY(strrchr)
+		mov	r3, #0
+1:		ldrb	r2, [r0], #1
+		teq	r2, r1
+		subeq	r3, r0, #1
+		teq	r2, #0
+		bne	1b
+		mov	r0, r3
+		RETINSTR(mov,pc,lr)
diff --git a/arch/arm26/lib/testchangebit.S b/arch/arm26/lib/testchangebit.S
new file mode 100644
index 0000000..17049a2
--- /dev/null
+++ b/arch/arm26/lib/testchangebit.S
@@ -0,0 +1,29 @@
+/*
+ *  linux/arch/arm26/lib/testchangebit.S
+ *
+ *  Copyright (C) 1995-1996 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>
+                .text
+
+ENTRY(_test_and_change_bit_be)
+		eor	r0, r0, #0x18		@ big endian byte ordering
+ENTRY(_test_and_change_bit_le)
+		add	r1, r1, r0, lsr #3
+		and	r3, r0, #7
+		mov	r0, #1
+		save_and_disable_irqs ip, r2
+		ldrb	r2, [r1]
+		tst	r2, r0, lsl r3
+		eor	r2, r2, r0, lsl r3
+		strb	r2, [r1]
+		restore_irqs ip
+		moveq	r0, #0
+		RETINSTR(mov,pc,lr)
+
+
diff --git a/arch/arm26/lib/testclearbit.S b/arch/arm26/lib/testclearbit.S
new file mode 100644
index 0000000..2506bd7
--- /dev/null
+++ b/arch/arm26/lib/testclearbit.S
@@ -0,0 +1,29 @@
+/*
+ *  linux/arch/arm26/lib/testclearbit.S
+ *
+ *  Copyright (C) 1995-1996 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>
+                .text
+
+ENTRY(_test_and_clear_bit_be)
+		eor	r0, r0, #0x18		@ big endian byte ordering
+ENTRY(_test_and_clear_bit_le)
+		add	r1, r1, r0, lsr #3	@ Get byte offset
+		and	r3, r0, #7		@ Get bit offset
+		mov	r0, #1
+		save_and_disable_irqs ip, r2
+		ldrb	r2, [r1]
+		tst	r2, r0, lsl r3
+		bic	r2, r2, r0, lsl r3
+		strb	r2, [r1]
+		restore_irqs ip
+		moveq	r0, #0
+		RETINSTR(mov,pc,lr)
+
+
diff --git a/arch/arm26/lib/testsetbit.S b/arch/arm26/lib/testsetbit.S
new file mode 100644
index 0000000..f827de6
--- /dev/null
+++ b/arch/arm26/lib/testsetbit.S
@@ -0,0 +1,29 @@
+/*
+ *  linux/arch/arm26/lib/testsetbit.S
+ *
+ *  Copyright (C) 1995-1996 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>
+                .text
+
+ENTRY(_test_and_set_bit_be)
+		eor	r0, r0, #0x18		@ big endian byte ordering
+ENTRY(_test_and_set_bit_le)
+		add	r1, r1, r0, lsr #3	@ Get byte offset
+		and	r3, r0, #7		@ Get bit offset
+		mov	r0, #1
+		save_and_disable_irqs ip, r2
+		ldrb	r2, [r1]
+		tst	r2, r0, lsl r3
+		orr	r2, r2, r0, lsl r3
+		strb	r2, [r1]
+		restore_irqs ip
+		moveq	r0, #0
+		RETINSTR(mov,pc,lr)
+
+
diff --git a/arch/arm26/lib/uaccess-kernel.S b/arch/arm26/lib/uaccess-kernel.S
new file mode 100644
index 0000000..3950a1f
--- /dev/null
+++ b/arch/arm26/lib/uaccess-kernel.S
@@ -0,0 +1,173 @@
+/*
+ *  linux/arch/arm26/lib/uaccess-kernel.S
+ *
+ *  Copyright (C) 1998 Russell King
+ *
+ *  Note!  Some code fragments found in here have a special calling
+ *  convention - they are not APCS compliant!
+ *
+ * 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>
+
+		.text
+
+//FIXME - surely this can be done in C not asm, removing the problem of keeping C and asm in sync? (this is a struct uaccess_t)
+
+		.globl	uaccess_kernel
+uaccess_kernel:
+		.word	uaccess_kernel_put_byte
+		.word	uaccess_kernel_get_byte
+		.word	uaccess_kernel_put_half
+		.word	uaccess_kernel_get_half
+		.word	uaccess_kernel_put_word
+		.word	uaccess_kernel_get_word
+		.word	uaccess_kernel_put_dword
+		.word	uaccess_kernel_copy
+		.word	uaccess_kernel_copy
+		.word	uaccess_kernel_clear
+		.word	uaccess_kernel_strncpy
+		.word	uaccess_kernel_strnlen
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_kernel_put_byte:
+		stmfd	sp!, {lr}
+		strb	r0, [r1]
+		ldmfd	sp!, {pc}^
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_kernel_put_half:
+		stmfd	sp!, {lr}
+		strb	r0, [r1]
+		mov	r0, r0, lsr #8
+		strb	r0, [r1, #1]
+		ldmfd	sp!, {pc}^
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_kernel_put_word:
+		stmfd	sp!, {lr}
+		str	r0, [r1]
+		ldmfd	sp!, {pc}^
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_kernel_put_dword:
+                stmfd   sp!, {lr}
+                str     r0, [r1], #4
+		str	r0, [r1], #0
+                ldmfd   sp!, {pc}^
+
+@ In : r0 = addr, r1 = error
+@ Out: r0 = x, r1 = error
+uaccess_kernel_get_byte:
+		stmfd	sp!, {lr}
+		ldrb	r0, [r0]
+		ldmfd	sp!, {pc}^
+
+@ In : r0 = addr, r1 = error
+@ Out: r0 = x, r1 = error
+uaccess_kernel_get_half:
+		stmfd	sp!, {lr}
+		ldr	r0, [r0]
+		mov	r0, r0, lsl #16
+		mov	r0, r0, lsr #16
+		ldmfd	sp!, {pc}^
+
+@ In : r0 = addr, r1 = error
+@ Out: r0 = x, r1 = error
+uaccess_kernel_get_word:
+		stmfd	sp!, {lr}
+		ldr	r0, [r0]
+		ldmfd	sp!, {pc}^
+
+
+/* Prototype: int uaccess_kernel_copy(void *to, const char *from, size_t n)
+ * Purpose  : copy a block to kernel memory from kernel memory
+ * Params   : to   - kernel memory
+ *          : from - kernel memory
+ *          : n    - number of bytes to copy
+ * Returns  : Number of bytes NOT copied.
+ */
+uaccess_kernel_copy:
+		stmfd	sp!, {lr}
+		bl	memcpy
+		mov	r0, #0
+		ldmfd	sp!, {pc}^
+
+/* Prototype: int uaccess_kernel_clear(void *addr, size_t sz)
+ * Purpose  : clear some kernel memory
+ * Params   : addr - kernel memory address to clear
+ *          : sz   - number of bytes to clear
+ * Returns  : number of bytes NOT cleared
+ */
+uaccess_kernel_clear:
+		stmfd	sp!, {lr}
+		mov	r2, #0
+		cmp	r1, #4
+		blt	2f
+		ands	ip, r0, #3
+		beq	1f
+		cmp	ip, #1
+		strb	r2, [r0], #1
+		strleb	r2, [r0], #1
+		strltb	r2, [r0], #1
+		rsb	ip, ip, #4
+		sub	r1, r1, ip		@  7  6  5  4  3  2  1
+1:		subs	r1, r1, #8		@ -1 -2 -3 -4 -5 -6 -7
+		bmi	2f
+		str	r2, [r0], #4
+		str	r2, [r0], #4
+		b	1b
+2:		adds	r1, r1, #4		@  3  2  1  0 -1 -2 -3
+		strpl	r2, [r0], #4
+		tst	r1, #2			@ 1x 1x 0x 0x 1x 1x 0x
+		strneb	r2, [r0], #1
+		strneb	r2, [r0], #1
+		tst	r1, #1			@ x1 x0 x1 x0 x1 x0 x1
+		strneb	r2, [r0], #1
+		mov	r0, #0
+		ldmfd	sp!, {pc}^
+
+/* Prototype: size_t uaccess_kernel_strncpy(char *dst, char *src, size_t len)
+ * Purpose  : copy a string from kernel memory to kernel memory
+ * Params   : dst - kernel memory destination
+ *          : src - kernel memory source
+ *          : len - maximum length of string
+ * Returns  : number of characters copied
+ */
+uaccess_kernel_strncpy:
+		stmfd	sp!, {lr}
+		mov	ip, r2
+1:		subs	r2, r2, #1
+		bmi	2f
+		ldrb	r3, [r1], #1
+		strb	r3, [r0], #1
+		teq	r3, #0
+		bne	1b
+2:		subs	r0, ip, r2
+		ldmfd	sp!, {pc}^
+
+/* Prototype: int uaccess_kernel_strlen(char *str, long n)
+ * Purpose  : get length of a string in kernel memory
+ * Params   : str - address of string in kernel memory
+ * Returns  : length of string *including terminator*,
+ *            or zero on exception, or n + 1 if too long
+ */
+uaccess_kernel_strnlen:
+		stmfd	sp!, {lr}
+		mov	r2, r0
+1:		ldrb	r1, [r0], #1
+		teq	r1, #0
+		beq	2f
+		subs	r1, r1, #1
+		bne	1b
+		add	r0, r0, #1
+2:		sub	r0, r0, r2
+		ldmfd	sp!, {pc}^
+
diff --git a/arch/arm26/lib/uaccess-user.S b/arch/arm26/lib/uaccess-user.S
new file mode 100644
index 0000000..130b8f2
--- /dev/null
+++ b/arch/arm26/lib/uaccess-user.S
@@ -0,0 +1,718 @@
+/*
+ *  linux/arch/arm26/lib/uaccess-user.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/page.h>
+
+		.text
+
+//FIXME - surely this can be done in C not asm, removing the problem of keeping C and asm in sync? (this is a struct uaccess_t)
+                .globl  uaccess_user
+uaccess_user:
+                .word   uaccess_user_put_byte
+                .word   uaccess_user_get_byte
+                .word   uaccess_user_put_half
+                .word   uaccess_user_get_half
+                .word   uaccess_user_put_word
+                .word   uaccess_user_get_word
+		.word   uaccess_user_put_dword
+                .word   uaccess_user_copy_from_user
+                .word   uaccess_user_copy_to_user
+                .word   uaccess_user_clear_user
+                .word   uaccess_user_strncpy_from_user
+                .word   uaccess_user_strnlen_user
+
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_user_put_byte:
+                stmfd   sp!, {lr}
+USER(           strbt   r0, [r1])
+                ldmfd   sp!, {pc}^
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_user_put_half:
+                stmfd   sp!, {lr}
+USER(           strbt   r0, [r1], #1)
+                mov     r0, r0, lsr #8
+USER(           strbt   r0, [r1])
+                ldmfd   sp!, {pc}^
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_user_put_word:
+                stmfd   sp!, {lr}
+USER(           strt    r0, [r1])
+                ldmfd   sp!, {pc}^
+
+@ In : r0 = x, r1 = addr, r2 = error
+@ Out: r2 = error
+uaccess_user_put_dword:
+                stmfd   sp!, {lr}
+USER(           strt    r0, [r1], #4)
+USER(           strt    r0, [r1], #0)
+                ldmfd   sp!, {pc}^
+
+9001:           mov     r2, #-EFAULT
+                ldmfd   sp!, {pc}^
+
+
+@ In : r0 = addr, r1 = error
+@ Out: r0 = x, r1 = error
+uaccess_user_get_byte:
+                stmfd   sp!, {lr}
+USER(           ldrbt   r0, [r0])
+                ldmfd   sp!, {pc}^
+
+@ In : r0 = addr, r1 = error
+@ Out: r0 = x, r1 = error
+uaccess_user_get_half:
+                stmfd   sp!, {lr}
+USER(           ldrt    r0, [r0])
+                mov     r0, r0, lsl #16
+                mov     r0, r0, lsr #16
+                ldmfd   sp!, {pc}^
+
+@ In : r0 = addr, r1 = error
+@ Out: r0 = x, r1 = error
+uaccess_user_get_word:
+                stmfd   sp!, {lr}
+USER(           ldrt    r0, [r0])
+                ldmfd   sp!, {pc}^
+
+9001:           mov     r1, #-EFAULT
+                ldmfd   sp!, {pc}^
+
+/* Prototype: int uaccess_user_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.
+ */
+
+.c2u_dest_not_aligned:
+		rsb	ip, ip, #4
+		cmp	ip, #2
+		ldrb	r3, [r1], #1
+USER(		strbt	r3, [r0], #1)			@ May fault
+		ldrgeb	r3, [r1], #1
+USER(		strgebt	r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #1
+USER(		strgtbt	r3, [r0], #1)			@ May fault
+		sub	r2, r2, ip
+		b	.c2u_dest_aligned
+
+ENTRY(uaccess_user_copy_to_user)
+		stmfd	sp!, {r2, r4 - r7, lr}
+		cmp	r2, #4
+		blt	.c2u_not_enough
+		ands	ip, r0, #3
+		bne	.c2u_dest_not_aligned
+.c2u_dest_aligned:
+
+		ands	ip, r1, #3
+		bne	.c2u_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...
+ */
+
+.c2u_0fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.c2u_0nowords
+		ldr	r3, [r1], #4
+USER(		strt	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	.c2u_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	.c2u_0rem8lp
+
+.c2u_0cpy8lp:	ldmia	r1!, {r3 - r6}
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		ldmia	r1!, {r3 - r6}
+		stmia	r0!, {r3 - r6}			@ Shouldnt fault
+		subs	ip, ip, #32
+		bpl	.c2u_0cpy8lp
+.c2u_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
+		strnet	r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.c2u_0fupi
+.c2u_0nowords:	teq	ip, #0
+		beq	.c2u_finished
+.c2u_nowords:	cmp	ip, #2
+		ldrb	r3, [r1], #1
+USER(		strbt	r3, [r0], #1)			@ May fault
+		ldrgeb	r3, [r1], #1
+USER(		strgebt	r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #1
+USER(		strgtbt	r3, [r0], #1)			@ May fault
+		b	.c2u_finished
+
+.c2u_not_enough:
+		movs	ip, r2
+		bne	.c2u_nowords
+.c2u_finished:	mov	r0, #0
+		LOADREGS(fd,sp!,{r2, r4 - r7, pc})
+
+.c2u_src_not_aligned:
+		bic	r1, r1, #3
+		ldr	r7, [r1], #4
+		cmp	ip, #2
+		bgt	.c2u_3fupi
+		beq	.c2u_2fupi
+.c2u_1fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.c2u_1nowords
+		mov	r3, r7, pull #8
+		ldr	r7, [r1], #4
+		orr	r3, r3, r7, push #24
+USER(		strt	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.c2u_1fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.c2u_1rem8lp
+
+.c2u_1cpy8lp:	mov	r3, r7, pull #8
+		ldmia	r1!, {r4 - r7}
+		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
+		subs	ip, ip, #16
+		bpl	.c2u_1cpy8lp
+.c2u_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
+		strnet	r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.c2u_1fupi
+.c2u_1nowords:	mov	r3, r7, lsr #byte(1)
+		teq	ip, #0
+		beq	.c2u_finished
+		cmp	ip, #2
+USER(		strbt	r3, [r0], #1)			@ May fault
+		movge	r3, r7, lsr #byte(2)
+USER(		strgebt	r3, [r0], #1)			@ May fault
+		movgt	r3, r7, lsr #byte(3)
+USER(		strgtbt	r3, [r0], #1)			@ May fault
+		b	.c2u_finished
+
+.c2u_2fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.c2u_2nowords
+		mov	r3, r7, pull #16
+		ldr	r7, [r1], #4
+		orr	r3, r3, r7, push #16
+USER(		strt	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.c2u_2fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.c2u_2rem8lp
+
+.c2u_2cpy8lp:	mov	r3, r7, pull #16
+		ldmia	r1!, {r4 - r7}
+		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
+		subs	ip, ip, #16
+		bpl	.c2u_2cpy8lp
+.c2u_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
+		strnet	r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.c2u_2fupi
+.c2u_2nowords:	mov	r3, r7, lsr #byte(2)
+		teq	ip, #0
+		beq	.c2u_finished
+		cmp	ip, #2
+USER(		strbt	r3, [r0], #1)			@ May fault
+		movge	r3, r7, lsr #byte(3)
+USER(		strgebt	r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #0
+USER(		strgtbt	r3, [r0], #1)			@ May fault
+		b	.c2u_finished
+
+.c2u_3fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.c2u_3nowords
+		mov	r3, r7, pull #24
+		ldr	r7, [r1], #4
+		orr	r3, r3, r7, push #8
+USER(		strt	r3, [r0], #4)			@ May fault
+		mov	ip, r0, lsl #32 - PAGE_SHIFT
+		rsb	ip, ip, #0
+		movs	ip, ip, lsr #32 - PAGE_SHIFT
+		beq	.c2u_3fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.c2u_3rem8lp
+
+.c2u_3cpy8lp:	mov	r3, r7, pull #24
+		ldmia	r1!, {r4 - r7}
+		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
+		subs	ip, ip, #16
+		bpl	.c2u_3cpy8lp
+.c2u_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
+		strnet	r3, [r0], #4			@ Shouldnt fault
+		ands	ip, ip, #3
+		beq	.c2u_3fupi
+.c2u_3nowords:	mov	r3, r7, lsr #byte(3)
+		teq	ip, #0
+		beq	.c2u_finished
+		cmp	ip, #2
+USER(		strbt	r3, [r0], #1)			@ May fault
+		ldrgeb	r3, [r1], #1
+USER(		strgebt	r3, [r0], #1)			@ May fault
+		ldrgtb	r3, [r1], #0
+USER(		strgtbt	r3, [r0], #1)			@ May fault
+		b	.c2u_finished
+
+		.section .fixup,"ax"
+		.align	0
+9001:		LOADREGS(fd,sp!, {r0, r4 - r7, pc})
+		.previous
+
+/* Prototype: unsigned long uaccess_user_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.
+ */
+.cfu_dest_not_aligned:
+		rsb	ip, ip, #4
+		cmp	ip, #2
+USER(		ldrbt	r3, [r1], #1)			@ May fault
+		strb	r3, [r0], #1
+USER(		ldrgebt	r3, [r1], #1)			@ May fault
+		strgeb	r3, [r0], #1
+USER(		ldrgtbt	r3, [r1], #1)			@ May fault
+		strgtb	r3, [r0], #1
+		sub	r2, r2, ip
+		b	.cfu_dest_aligned
+
+ENTRY(uaccess_user_copy_from_user)
+		stmfd	sp!, {r0, r2, r4 - r7, lr}
+		cmp	r2, #4
+		blt	.cfu_not_enough
+		ands	ip, r0, #3
+		bne	.cfu_dest_not_aligned
+.cfu_dest_aligned:
+		ands	ip, r1, #3
+		bne	.cfu_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...
+ */
+
+.cfu_0fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.cfu_0nowords
+USER(		ldrt	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	.cfu_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	.cfu_0rem8lp
+
+.cfu_0cpy8lp:	ldmia	r1!, {r3 - r6}			@ Shouldnt fault
+		stmia	r0!, {r3 - r6}
+		ldmia	r1!, {r3 - r6}			@ Shouldnt fault
+		stmia	r0!, {r3 - r6}
+		subs	ip, ip, #32
+		bpl	.cfu_0cpy8lp
+.cfu_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
+		ldrnet	r3, [r1], #4			@ Shouldnt fault
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.cfu_0fupi
+.cfu_0nowords:	teq	ip, #0
+		beq	.cfu_finished
+.cfu_nowords:	cmp	ip, #2
+USER(		ldrbt	r3, [r1], #1)			@ May fault
+		strb	r3, [r0], #1
+USER(		ldrgebt	r3, [r1], #1)			@ May fault
+		strgeb	r3, [r0], #1
+USER(		ldrgtbt	r3, [r1], #1)			@ May fault
+		strgtb	r3, [r0], #1
+		b	.cfu_finished
+
+.cfu_not_enough:
+		movs	ip, r2
+		bne	.cfu_nowords
+.cfu_finished:	mov	r0, #0
+		add	sp, sp, #8
+		LOADREGS(fd,sp!,{r4 - r7, pc})
+
+.cfu_src_not_aligned:
+		bic	r1, r1, #3
+USER(		ldrt	r7, [r1], #4)			@ May fault
+		cmp	ip, #2
+		bgt	.cfu_3fupi
+		beq	.cfu_2fupi
+.cfu_1fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.cfu_1nowords
+		mov	r3, r7, pull #8
+USER(		ldrt	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	.cfu_1fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.cfu_1rem8lp
+
+.cfu_1cpy8lp:	mov	r3, r7, pull #8
+		ldmia	r1!, {r4 - r7}			@ Shouldnt fault
+		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}
+		subs	ip, ip, #16
+		bpl	.cfu_1cpy8lp
+.cfu_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(		ldrnet	r7, [r1], #4)			@ May fault
+		orrne	r3, r3, r7, push #24
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.cfu_1fupi
+.cfu_1nowords:	mov	r3, r7, lsr #byte(1)
+		teq	ip, #0
+		beq	.cfu_finished
+		cmp	ip, #2
+		strb	r3, [r0], #1
+		movge	r3, r7, lsr #byte(2)
+		strgeb	r3, [r0], #1
+		movgt	r3, r7, lsr #byte(3)
+		strgtb	r3, [r0], #1
+		b	.cfu_finished
+
+.cfu_2fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.cfu_2nowords
+		mov	r3, r7, pull #16
+USER(		ldrt	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	.cfu_2fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.cfu_2rem8lp
+
+.cfu_2cpy8lp:	mov	r3, r7, pull #16
+		ldmia	r1!, {r4 - r7}			@ Shouldnt fault
+		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}
+		subs	ip, ip, #16
+		bpl	.cfu_2cpy8lp
+.cfu_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(		ldrnet	r7, [r1], #4)			@ May fault
+		orrne	r3, r3, r7, push #16
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.cfu_2fupi
+.cfu_2nowords:	mov	r3, r7, lsr #byte(2)
+		teq	ip, #0
+		beq	.cfu_finished
+		cmp	ip, #2
+		strb	r3, [r0], #1
+		movge	r3, r7, lsr #byte(3)
+		strgeb	r3, [r0], #1
+USER(		ldrgtbt	r3, [r1], #0)			@ May fault
+		strgtb	r3, [r0], #1
+		b	.cfu_finished
+
+.cfu_3fupi:	subs	r2, r2, #4
+		addmi	ip, r2, #4
+		bmi	.cfu_3nowords
+		mov	r3, r7, pull #24
+USER(		ldrt	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	.cfu_3fupi
+		cmp	r2, ip
+		movlt	ip, r2
+		sub	r2, r2, ip
+		subs	ip, ip, #16
+		blt	.cfu_3rem8lp
+
+.cfu_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	.cfu_3cpy8lp
+.cfu_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(		ldrnet	r7, [r1], #4)			@ May fault
+		orrne	r3, r3, r7, push #8
+		strne	r3, [r0], #4
+		ands	ip, ip, #3
+		beq	.cfu_3fupi
+.cfu_3nowords:	mov	r3, r7, lsr #byte(3)
+		teq	ip, #0
+		beq	.cfu_finished
+		cmp	ip, #2
+		strb	r3, [r0], #1
+USER(		ldrgebt	r3, [r1], #1)			@ May fault
+		strgeb	r3, [r0], #1
+USER(		ldrgtbt	r3, [r1], #1)			@ May fault
+		strgtb	r3, [r0], #1
+		b	.cfu_finished
+
+		.section .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
+		LOADREGS(fd,sp!, {r4 - r7, pc})
+		.previous
+
+/* Prototype: int uaccess_user_clear_user(void *addr, size_t sz)
+ * Purpose  : clear some user memory
+ * Params   : addr - user memory address to clear
+ *          : sz   - number of bytes to clear
+ * Returns  : number of bytes NOT cleared
+ */
+ENTRY(uaccess_user_clear_user)
+		stmfd	sp!, {r1, lr}
+		mov	r2, #0
+		cmp	r1, #4
+		blt	2f
+		ands	ip, r0, #3
+		beq	1f
+		cmp	ip, #2
+USER(		strbt	r2, [r0], #1)
+USER(		strlebt	r2, [r0], #1)
+USER(		strltbt	r2, [r0], #1)
+		rsb	ip, ip, #4
+		sub	r1, r1, ip		@  7  6  5  4  3  2  1
+1:		subs	r1, r1, #8		@ -1 -2 -3 -4 -5 -6 -7
+USER(		strplt	r2, [r0], #4)
+USER(		strplt	r2, [r0], #4)
+		bpl	1b
+		adds	r1, r1, #4		@  3  2  1  0 -1 -2 -3
+USER(		strplt	r2, [r0], #4)
+2:		tst	r1, #2			@ 1x 1x 0x 0x 1x 1x 0x
+USER(		strnebt	r2, [r0], #1)
+USER(		strnebt	r2, [r0], #1)
+		tst	r1, #1			@ x1 x0 x1 x0 x1 x0 x1
+USER(		strnebt	r2, [r0], #1)
+		mov	r0, #0
+		LOADREGS(fd,sp!, {r1, pc})
+
+		.section .fixup,"ax"
+		.align	0
+9001:		LOADREGS(fd,sp!, {r0, pc})
+		.previous
+
+/*
+ * Copy a string from user space to kernel space.
+ *  r0 = dst, r1 = src, r2 = byte length
+ * returns the number of characters copied (strlen of copied string),
+ *  -EFAULT on exception, or "len" if we fill the whole buffer
+ */
+ENTRY(uaccess_user_strncpy_from_user)
+        save_lr
+        mov     ip, r1
+1:      subs    r2, r2, #1
+USER(   ldrplbt r3, [r1], #1)
+        bmi     2f
+        strb    r3, [r0], #1
+        teq     r3, #0
+        bne     1b
+        sub     r1, r1, #1      @ take NUL character out of count
+2:      sub     r0, r1, ip
+        restore_pc
+
+        .section .fixup,"ax"
+        .align  0
+9001:   mov     r3, #0
+        strb    r3, [r0, #0]    @ null terminate
+        mov     r0, #-EFAULT
+        restore_pc
+        .previous
+
+/* Prototype: unsigned long uaccess_user_strnlen_user(const char *str, long n)
+ * Purpose  : get length of a string in user memory
+ * Params   : str - address of string in user memory
+ * Returns  : length of string *including terminator*
+ *            or zero on exception, or n + 1 if too long
+ */
+ENTRY(uaccess_user_strnlen_user)
+        save_lr
+        mov     r2, r0
+1:
+USER(   ldrbt   r3, [r0], #1)
+        teq     r3, #0
+        beq     2f
+        subs    r1, r1, #1
+        bne     1b
+        add     r0, r0, #1
+2:      sub     r0, r0, r2
+        restore_pc
+
+        .section .fixup,"ax"
+        .align  0
+9001:   mov     r0, #0
+        restore_pc
+        .previous
+
diff --git a/arch/arm26/lib/ucmpdi2.c b/arch/arm26/lib/ucmpdi2.c
new file mode 100644
index 0000000..6c6ae63
--- /dev/null
+++ b/arch/arm26/lib/ucmpdi2.c
@@ -0,0 +1,51 @@
+/* More subroutines needed by GCC output code on some machines.  */
+/* Compile this one with gcc.  */
+/* Copyright (C) 1989, 92-98, 1999 Free Software Foundation, Inc.
+
+This file is part of GNU CC.
+
+GNU CC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GNU CC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GNU CC; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* support functions required by the kernel. based on code from gcc-2.95.3 */
+/* I Molton     29/07/01 */
+
+#include "gcclib.h"
+
+word_type
+__ucmpdi2 (DItype a, DItype b)
+{
+  DIunion au, bu;
+
+  au.ll = a, bu.ll = b;
+
+  if ((USItype) au.s.high < (USItype) bu.s.high)
+    return 0;
+  else if ((USItype) au.s.high > (USItype) bu.s.high)
+    return 2;
+  if ((USItype) au.s.low < (USItype) bu.s.low)
+    return 0;
+  else if ((USItype) au.s.low > (USItype) bu.s.low)
+    return 2;
+  return 1;
+}
+
diff --git a/arch/arm26/lib/udivdi3.c b/arch/arm26/lib/udivdi3.c
new file mode 100644
index 0000000..d25195f
--- /dev/null
+++ b/arch/arm26/lib/udivdi3.c
@@ -0,0 +1,242 @@
+/* More subroutines needed by GCC output code on some machines.  */
+/* Compile this one with gcc.  */
+/* Copyright (C) 1989, 92-98, 1999 Free Software Foundation, Inc.
+
+This file is part of GNU CC.
+
+GNU CC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 2, or (at your option)
+any later version.
+
+GNU CC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+You should have received a copy of the GNU General Public License
+along with GNU CC; see the file COPYING.  If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA.  */
+
+/* As a special exception, if you link this library with other files,
+   some of which are compiled with GCC, to produce an executable,
+   this library does not by itself cause the resulting executable
+   to be covered by the GNU General Public License.
+   This exception does not however invalidate any other reasons why
+   the executable file might be covered by the GNU General Public License.
+ */
+/* support functions required by the kernel. based on code from gcc-2.95.3 */
+/* I Molton     29/07/01 */
+
+#include "gcclib.h"
+#include "longlong.h"
+
+static const UQItype __clz_tab[] =
+{
+  0,1,2,2,3,3,3,3,4,4,4,4,4,4,4,4,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,5,
+  6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,6,
+  7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,
+  7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,7,
+  8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
+  8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
+  8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
+  8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,8,
+};
+
+UDItype
+__udivmoddi4 (UDItype n, UDItype d, UDItype *rp)
+{
+  DIunion ww;
+  DIunion nn, dd;
+  DIunion rr;
+  USItype d0, d1, n0, n1, n2;
+  USItype q0, q1;
+  USItype b, bm;
+
+  nn.ll = n;
+  dd.ll = d;
+
+  d0 = dd.s.low;
+  d1 = dd.s.high;
+  n0 = nn.s.low;
+  n1 = nn.s.high;
+
+  if (d1 == 0)
+    {
+      if (d0 > n1)
+        {
+          /* 0q = nn / 0D */
+
+          count_leading_zeros (bm, d0);
+
+          if (bm != 0)
+            {
+              /* Normalize, i.e. make the most significant bit of the
+                 denominator set.  */
+
+              d0 = d0 << bm;
+              n1 = (n1 << bm) | (n0 >> (SI_TYPE_SIZE - bm));
+              n0 = n0 << bm;
+            }
+
+          udiv_qrnnd (q0, n0, n1, n0, d0);
+          q1 = 0;
+
+          /* Remainder in n0 >> bm.  */
+        }
+      else
+        {
+          /* qq = NN / 0d */
+
+          if (d0 == 0)
+            d0 = 1 / d0;        /* Divide intentionally by zero.  */
+
+          count_leading_zeros (bm, d0);
+
+          if (bm == 0)
+            {
+              /* From (n1 >= d0) /\ (the most significant bit of d0 is set),
+                 conclude (the most significant bit of n1 is set) /\ (the
+                 leading quotient digit q1 = 1).
+
+                 This special case is necessary, not an optimization.
+                 (Shifts counts of SI_TYPE_SIZE are undefined.)  */
+
+              n1 -= d0;
+              q1 = 1;
+            }
+          else
+            {
+              /* Normalize.  */
+
+              b = SI_TYPE_SIZE - bm;
+
+              d0 = d0 << bm;
+              n2 = n1 >> b;
+              n1 = (n1 << bm) | (n0 >> b);
+              n0 = n0 << bm;
+
+              udiv_qrnnd (q1, n1, n2, n1, d0);
+            }
+
+          /* n1 != d0...  */
+
+          udiv_qrnnd (q0, n0, n1, n0, d0);
+
+          /* Remainder in n0 >> bm.  */
+        }
+
+      if (rp != 0)
+        {
+          rr.s.low = n0 >> bm;
+          rr.s.high = 0;
+          *rp = rr.ll;
+        }
+    }
+  else
+    {
+      if (d1 > n1)
+        {
+          /* 00 = nn / DD */
+
+          q0 = 0;
+          q1 = 0;
+
+          /* Remainder in n1n0.  */
+          if (rp != 0)
+            {
+              rr.s.low = n0;
+              rr.s.high = n1;
+              *rp = rr.ll;
+            }
+        }
+      else
+        {
+          /* 0q = NN / dd */
+
+          count_leading_zeros (bm, d1);
+          if (bm == 0)
+            {
+              /* From (n1 >= d1) /\ (the most significant bit of d1 is set),
+                 conclude (the most significant bit of n1 is set) /\ (the
+                 quotient digit q0 = 0 or 1).
+
+                 This special case is necessary, not an optimization.  */
+
+              /* The condition on the next line takes advantage of that
+                 n1 >= d1 (true due to program flow).  */
+              if (n1 > d1 || n0 >= d0)
+                {
+                  q0 = 1;
+                  sub_ddmmss (n1, n0, n1, n0, d1, d0);
+                }
+              else
+                q0 = 0;
+
+              q1 = 0;
+
+              if (rp != 0)
+                {
+                  rr.s.low = n0;
+                  rr.s.high = n1;
+                  *rp = rr.ll;
+                }
+            }
+          else
+            {
+              USItype m1, m0;
+              /* Normalize.  */
+
+              b = SI_TYPE_SIZE - bm;
+
+              d1 = (d1 << bm) | (d0 >> b);
+              d0 = d0 << bm;
+              n2 = n1 >> b;
+              n1 = (n1 << bm) | (n0 >> b);
+              n0 = n0 << bm;
+
+              udiv_qrnnd (q0, n1, n2, n1, d1);
+              umul_ppmm (m1, m0, q0, d0);
+
+              if (m1 > n1 || (m1 == n1 && m0 > n0))
+                {
+                  q0--;
+                  sub_ddmmss (m1, m0, m1, m0, d1, d0);
+                }
+
+              q1 = 0;
+
+              /* Remainder in (n1n0 - m1m0) >> bm.  */
+              if (rp != 0)
+                {
+                  sub_ddmmss (n1, n0, n1, n0, m1, m0);
+                  rr.s.low = (n1 << b) | (n0 >> bm);
+                  rr.s.high = n1 >> bm;
+                  *rp = rr.ll;
+                }
+            }
+        }
+    }
+
+  ww.s.low = q0;
+  ww.s.high = q1;
+  return ww.ll;
+}
+
+UDItype
+__udivdi3 (UDItype n, UDItype d)
+{
+  return __udivmoddi4 (n, d, (UDItype *) 0);
+}
+
+UDItype
+__umoddi3 (UDItype u, UDItype v)
+{
+  UDItype w;
+
+  (void) __udivmoddi4 (u ,v, &w);
+
+  return w;
+}
+