blackfin architecture

This adds support for the Analog Devices Blackfin processor architecture, and
currently supports the BF533, BF532, BF531, BF537, BF536, BF534, and BF561
(Dual Core) devices, with a variety of development platforms including those
avaliable from Analog Devices (BF533-EZKit, BF533-STAMP, BF537-STAMP,
BF561-EZKIT), and Bluetechnix!  Tinyboards.

The Blackfin architecture was jointly developed by Intel and Analog Devices
Inc.  (ADI) as the Micro Signal Architecture (MSA) core and introduced it in
December of 2000.  Since then ADI has put this core into its Blackfin
processor family of devices.  The Blackfin core has the advantages of a clean,
orthogonal,RISC-like microprocessor instruction set.  It combines a dual-MAC
(Multiply/Accumulate), state-of-the-art signal processing engine and
single-instruction, multiple-data (SIMD) multimedia capabilities into a single
instruction-set architecture.

The Blackfin architecture, including the instruction set, is described by the
ADSP-BF53x/BF56x Blackfin Processor Programming Reference
http://blackfin.uclinux.org/gf/download/frsrelease/29/2549/Blackfin_PRM.pdf

The Blackfin processor is already supported by major releases of gcc, and
there are binary and source rpms/tarballs for many architectures at:
http://blackfin.uclinux.org/gf/project/toolchain/frs There is complete
documentation, including "getting started" guides available at:
http://docs.blackfin.uclinux.org/ which provides links to the sources and
patches you will need in order to set up a cross-compiling environment for
bfin-linux-uclibc

This patch, as well as the other patches (toolchain, distribution,
uClibc) are actively supported by Analog Devices Inc, at:
http://blackfin.uclinux.org/

We have tested this on LTP, and our test plan (including pass/fails) can
be found at:
http://docs.blackfin.uclinux.org/doku.php?id=testing_the_linux_kernel

[m.kozlowski@tuxland.pl: balance parenthesis in blackfin header files]
Signed-off-by: Bryan Wu <bryan.wu@analog.com>
Signed-off-by: Mariusz Kozlowski <m.kozlowski@tuxland.pl>
Signed-off-by: Aubrey Li <aubrey.li@analog.com>
Signed-off-by: Jie Zhang <jie.zhang@analog.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

24 files changed, 1907 insertions, 0 deletions

diff --git a/arch/blackfin/lib/Makefile b/arch/blackfin/lib/Makefile
new file mode 100644
index 0000000..635288f
--- /dev/null
+++ b/arch/blackfin/lib/Makefile
@@ -0,0 +1,11 @@
+#
+# arch/blackfin/lib/Makefile
+#
+
+lib-y := \
+	ashldi3.o ashrdi3.o lshrdi3.o \
+	muldi3.o divsi3.o udivsi3.o modsi3.o umodsi3.o \
+	checksum.o memcpy.o memset.o memcmp.o memchr.o memmove.o \
+	strcmp.o strcpy.o strncmp.o strncpy.o \
+	umulsi3_highpart.o smulsi3_highpart.o \
+	ins.o outs.o
diff --git a/arch/blackfin/lib/ashldi3.c b/arch/blackfin/lib/ashldi3.c
new file mode 100644
index 0000000..a8c279e
--- /dev/null
+++ b/arch/blackfin/lib/ashldi3.c
@@ -0,0 +1,58 @@
+/*
+ * File:         arch/blackfin/lib/ashldi3.c
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "gcclib.h"
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+DItype __ashldi3(DItype u, word_type b)__attribute__((l1_text));
+#endif
+
+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/blackfin/lib/ashrdi3.c b/arch/blackfin/lib/ashrdi3.c
new file mode 100644
index 0000000..a0d3419
--- /dev/null
+++ b/arch/blackfin/lib/ashrdi3.c
@@ -0,0 +1,59 @@
+/*
+ * File:         arch/blackfin/lib/ashrdi3.c
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include "gcclib.h"
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+DItype __ashrdi3(DItype u, word_type b)__attribute__((l1_text));
+#endif
+
+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/blackfin/lib/checksum.c b/arch/blackfin/lib/checksum.c
new file mode 100644
index 0000000..42768e0
--- /dev/null
+++ b/arch/blackfin/lib/checksum.c
@@ -0,0 +1,140 @@
+/*
+ * File:         arch/blackfin/lib/checksum.c
+ * Based on:     none - original work
+ * Author:
+ *
+ * Created:
+ * Description:  An implementation of the TCP/IP protocol suite for the LINUX
+ *               operating system.  INET is implemented using the  BSD Socket
+ *               interface as the means of communication with the user level.
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <net/checksum.h>
+#include <asm/checksum.h>
+
+#ifdef CONFIG_IP_CHECKSUM_L1
+static unsigned short do_csum(const unsigned char *buff, int len)__attribute__((l1_text));
+#endif
+
+static unsigned short do_csum(const unsigned char *buff, int len)
+{
+	register unsigned long sum = 0;
+	int swappem = 0;
+
+	if (1 & (unsigned long)buff) {
+		sum = *buff << 8;
+		buff++;
+		len--;
+		++swappem;
+	}
+
+	while (len > 1) {
+		sum += *(unsigned short *)buff;
+		buff += 2;
+		len -= 2;
+	}
+
+	if (len > 0)
+		sum += *buff;
+
+	/*  Fold 32-bit sum to 16 bits */
+	while (sum >> 16)
+		sum = (sum & 0xffff) + (sum >> 16);
+
+	if (swappem)
+		sum = ((sum & 0xff00) >> 8) + ((sum & 0x00ff) << 8);
+
+	return sum;
+
+}
+
+/*
+ *	This is a version of ip_compute_csum() optimized for IP headers,
+ *	which always checksum on 4 octet boundaries.
+ */
+unsigned short ip_fast_csum(unsigned char *iph, unsigned int ihl)
+{
+	return ~do_csum(iph, ihl * 4);
+}
+
+/*
+ * computes the checksum of a memory block at buff, length len,
+ * and adds in "sum" (32-bit)
+ *
+ * returns a 32-bit number suitable for feeding into itself
+ * or csum_tcpudp_magic
+ *
+ * this function must be called with even lengths, except
+ * for the last fragment, which may be odd
+ *
+ * it's best to have buff aligned on a 32-bit boundary
+ */
+unsigned int csum_partial(const unsigned char *buff, int len, unsigned int sum)
+{
+	/*
+	 * Just in case we get nasty checksum data...
+	 * Like 0xffff6ec3 in the case of our IPv6 multicast header.
+	 * We fold to begin with, as well as at the end.
+	 */
+	sum = (sum & 0xffff) + (sum >> 16);
+
+	sum += do_csum(buff, len);
+
+	sum = (sum & 0xffff) + (sum >> 16);
+
+	return sum;
+}
+
+/*
+ * this routine is used for miscellaneous IP-like checksums, mainly
+ * in icmp.c
+ */
+unsigned short ip_compute_csum(const unsigned char *buff, int len)
+{
+	return ~do_csum(buff, len);
+}
+
+/*
+ * copy from fs while checksumming, otherwise like csum_partial
+ */
+
+unsigned int
+csum_partial_copy_from_user(const unsigned char *src, unsigned char *dst,
+			    int len, int sum, int *csum_err)
+{
+	if (csum_err)
+		*csum_err = 0;
+	memcpy(dst, src, len);
+	return csum_partial(dst, len, sum);
+}
+
+/*
+ * copy from ds while checksumming, otherwise like csum_partial
+ */
+
+unsigned int csum_partial_copy(const unsigned char *src, unsigned char *dst,
+			       int len, int sum)
+{
+	memcpy(dst, src, len);
+	return csum_partial(dst, len, sum);
+}
diff --git a/arch/blackfin/lib/divsi3.S b/arch/blackfin/lib/divsi3.S
new file mode 100644
index 0000000..3e29861
--- /dev/null
+++ b/arch/blackfin/lib/divsi3.S
@@ -0,0 +1,216 @@
+/*
+ * File:         arch/blackfin/lib/divsi3.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:  16 / 32 bit signed division.
+ *                 Special cases :
+ *                      1)  If(numerator == 0)
+ *                             return 0
+ *                      2)  If(denominator ==0)
+ *                             return positive max = 0x7fffffff
+ *                      3)  If(numerator == denominator)
+ *                             return 1
+ *                      4)  If(denominator ==1)
+ *                             return numerator
+ *                      5)  If(denominator == -1)
+ *                             return -numerator
+ *
+ *                 Operand         : R0 - Numerator   (i)
+ *                                   R1 - Denominator (i)
+ *                                   R0 - Quotient    (o)
+ *                 Registers Used : R2-R7,P0-P2
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+.global   ___divsi3;
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+.section .l1.text
+#else
+.text
+#endif
+
+.align 2;
+___divsi3 :
+
+
+  R3 = R0 ^ R1;
+  R0 = ABS R0;
+
+  CC = V;
+
+  r3 = rot r3 by -1;
+  r1 = abs r1;      /* now both positive, r3.30 means "negate result",
+                    ** r3.31 means overflow, add one to result
+                    */
+  cc = r0 < r1;
+  if cc jump .Lret_zero;
+  r2 = r1 >> 15;
+  cc = r2;
+  if cc jump .Lidents;
+  r2 = r1 << 16;
+  cc = r2 <= r0;
+  if cc jump .Lidents;
+
+  DIVS(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+  DIVQ(R0, R1);
+
+  R0 = R0.L (Z);
+  r1 = r3 >> 31;    /* add overflow issue back in */
+  r0 = r0 + r1;
+  r1 = -r0;
+  cc = bittst(r3, 30);
+  if cc r0 = r1;
+  RTS;
+
+/* Can't use the primitives. Test common identities.
+** If the identity is true, return the value in R2.
+*/
+
+.Lidents:
+  CC = R1 == 0;                   /* check for divide by zero */
+  IF CC JUMP .Lident_return;
+
+  CC = R0 == 0;                   /* check for division of zero */
+  IF CC JUMP .Lzero_return;
+
+  CC = R0 == R1;                  /* check for identical operands */
+  IF CC JUMP .Lident_return;
+
+  CC = R1 == 1;                   /* check for divide by 1 */
+  IF CC JUMP .Lident_return;
+
+  R2.L = ONES R1;
+  R2 = R2.L (Z);
+  CC = R2 == 1;
+  IF CC JUMP .Lpower_of_two;
+
+  /* Identities haven't helped either.
+  ** Perform the full division process.
+  */
+
+  P1 = 31;                        /* Set loop counter   */
+
+  [--SP] = (R7:5);                /* Push registers R5-R7 */
+  R2 = -R1;
+  [--SP] = R2;
+  R2 = R0 << 1;                   /* R2 lsw of dividend  */
+  R6 = R0 ^ R1;                   /* Get sign */
+  R5 = R6 >> 31;                  /* Shift sign to LSB */
+
+  R0 = 0 ;                        /* Clear msw partial remainder */
+  R2 = R2 | R5;                   /* Shift quotient bit */
+  R6 = R0 ^ R1;                   /* Get new quotient bit */
+
+  LSETUP(.Llst,.Llend)  LC0 = P1;   /* Setup loop */
+.Llst:   R7 = R2 >> 31;            /* record copy of carry from R2 */
+        R2 = R2 << 1;             /* Shift 64 bit dividend up by 1 bit */
+        R0 = R0 << 1 || R5 = [SP];
+        R0 = R0 | R7;             /* and add carry */
+        CC = R6 < 0;              /* Check quotient(AQ) */
+                                  /* we might be subtracting divisor (AQ==0) */
+        IF CC R5 = R1;            /* or we might be adding divisor  (AQ==1)*/
+        R0 = R0 + R5;             /* do add or subtract, as indicated by AQ */
+        R6 = R0 ^ R1;             /* Generate next quotient bit */
+        R5 = R6 >> 31;
+                                  /* Assume AQ==1, shift in zero */
+        BITTGL(R5,0);             /* tweak AQ to be what we want to shift in */
+.Llend:  R2 = R2 + R5;             /* and then set shifted-in value to
+                                  ** tweaked AQ.
+                                  */
+  r1 = r3 >> 31;
+  r2 = r2 + r1;
+  cc = bittst(r3,30);
+  r0 = -r2;
+  if !cc r0 = r2;
+  SP += 4;
+  (R7:5)= [SP++];                 /* Pop registers R6-R7 */
+  RTS;
+
+.Lident_return:
+  CC = R1 == 0;                   /* check for divide by zero  => 0x7fffffff */
+  R2 = -1 (X);
+  R2 >>= 1;
+  IF CC JUMP .Ltrue_ident_return;
+
+  CC = R0 == R1;                  /* check for identical operands => 1 */
+  R2 = 1 (Z);
+  IF CC JUMP .Ltrue_ident_return;
+
+  R2 = R0;                        /* assume divide by 1 => numerator */
+  /*FALLTHRU*/
+
+.Ltrue_ident_return:
+  R0 = R2;                        /* Return an identity value */
+  R2 = -R2;
+  CC = bittst(R3,30);
+  IF CC R0 = R2;
+.Lzero_return:
+  RTS;                            /* ...including zero */
+
+.Lpower_of_two:
+  /* Y has a single bit set, which means it's a power of two.
+  ** That means we can perform the division just by shifting
+  ** X to the right the appropriate number of bits
+  */
+
+  /* signbits returns the number of sign bits, minus one.
+  ** 1=>30, 2=>29, ..., 0x40000000=>0. Which means we need
+  ** to shift right n-signbits spaces. It also means 0x80000000
+  ** is a special case, because that *also* gives a signbits of 0
+  */
+
+  R2 = R0 >> 31;
+  CC = R1 < 0;
+  IF CC JUMP .Ltrue_ident_return;
+
+  R1.l = SIGNBITS R1;
+  R1 = R1.L (Z);
+  R1 += -30;
+  R0 = LSHIFT R0 by R1.L;
+  r1 = r3 >> 31;
+  r0 = r0 + r1;
+  R2 = -R0;                       // negate result if necessary
+  CC = bittst(R3,30);
+  IF CC R0 = R2;
+  RTS;
+
+.Lret_zero:
+  R0 = 0;
+  RTS;
diff --git a/arch/blackfin/lib/gcclib.h b/arch/blackfin/lib/gcclib.h
new file mode 100644
index 0000000..9ccd39a
--- /dev/null
+++ b/arch/blackfin/lib/gcclib.h
@@ -0,0 +1,47 @@
+/*
+ * File:         arch/blackfin/lib/gcclib.h
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#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/blackfin/lib/ins.S b/arch/blackfin/lib/ins.S
new file mode 100644
index 0000000..730d2b4
--- /dev/null
+++ b/arch/blackfin/lib/ins.S
@@ -0,0 +1,69 @@
+/*
+ * File:         arch/blackfin/lib/ins.S
+ * Based on:
+ * Author:       Bas Vermeulen <bas@buyways.nl>
+ *
+ * Created:      Tue Mar 22 15:27:24 CEST 2005
+ * Description:  Implementation of ins{bwl} for BlackFin processors using zero overhead loops.
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *               Copyright (C) 2005 Bas Vermeulen, BuyWays BV <bas@buyways.nl>
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+.align 2
+
+ENTRY(_insl)
+	P0 = R0;	/* P0 = port */
+	cli R3;
+	P1 = R1;	/* P1 = address */
+	P2 = R2;	/* P2 = count */
+	SSYNC;
+	LSETUP( .Llong_loop_s, .Llong_loop_e) LC0 = P2;
+.Llong_loop_s: R0 = [P0];
+.Llong_loop_e: [P1++] = R0;
+	sti R3;
+	RTS;
+
+ENTRY(_insw)
+	P0 = R0;	/* P0 = port */
+	cli R3;
+	P1 = R1;	/* P1 = address */
+	P2 = R2;	/* P2 = count */
+	SSYNC;
+	LSETUP( .Lword_loop_s, .Lword_loop_e) LC0 = P2;
+.Lword_loop_s: R0 = W[P0];
+.Lword_loop_e: W[P1++] = R0;
+	sti R3;
+	RTS;
+
+ENTRY(_insb)
+	P0 = R0;	/* P0 = port */
+	cli R3;
+	P1 = R1;	/* P1 = address */
+	P2 = R2;	/* P2 = count */
+	SSYNC;
+	LSETUP( .Lbyte_loop_s, .Lbyte_loop_e) LC0 = P2;
+.Lbyte_loop_s: R0 = B[P0];
+.Lbyte_loop_e: B[P1++] = R0;
+	sti R3;
+	RTS;
diff --git a/arch/blackfin/lib/lshrdi3.c b/arch/blackfin/lib/lshrdi3.c
new file mode 100644
index 0000000..84b9c55
--- /dev/null
+++ b/arch/blackfin/lib/lshrdi3.c
@@ -0,0 +1,72 @@
+/*
+ * File:         arch/blackfin/lib/lshrdi3.c
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#define BITS_PER_UNIT 8
+
+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__)));
+
+struct DIstruct {
+	SItype high, low;
+};
+
+typedef union {
+	struct DIstruct s;
+	DItype ll;
+} DIunion;
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+DItype __lshrdi3(DItype u, word_type b)__attribute__((l1_text));
+#endif
+
+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/blackfin/lib/memchr.S b/arch/blackfin/lib/memchr.S
new file mode 100644
index 0000000..4981222
--- /dev/null
+++ b/arch/blackfin/lib/memchr.S
@@ -0,0 +1,70 @@
+/*
+ * File:         arch/blackfin/lib/memchr.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+/* void *memchr(const void *s, int c, size_t n);
+ * R0 = address (s)
+ * R1 = sought byte (c)
+ * R2 = count (n)
+ *
+ * Returns pointer to located character.
+ */
+
+.text
+
+.align 2
+
+ENTRY(_memchr)
+	P0 = R0;		/* P0 = address */
+	P2 = R2;		/* P2 = count */
+	R1 = R1.B(Z);
+	CC = R2 == 0;
+	IF CC JUMP .Lfailed;
+
+.Lbytes:
+	LSETUP (.Lbyte_loop_s, .Lbyte_loop_e) LC0=P2;
+
+.Lbyte_loop_s:
+	R3 = B[P0++](Z);
+	CC = R3 == R1;
+	IF CC JUMP .Lfound;
+.Lbyte_loop_e:
+	NOP;
+
+.Lfailed:
+	R0=0;
+	RTS;
+
+.Lfound:
+	R0 = P0;
+	R0 += -1;
+	RTS;
+
+.size _memchr,.-_memchr
diff --git a/arch/blackfin/lib/memcmp.S b/arch/blackfin/lib/memcmp.S
new file mode 100644
index 0000000..5b95023
--- /dev/null
+++ b/arch/blackfin/lib/memcmp.S
@@ -0,0 +1,110 @@
+/*
+ * File:         arch/blackfin/lib/memcmp.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+/* int memcmp(const void *s1, const void *s2, size_t n);
+ * R0 = First Address (s1)
+ * R1 = Second Address (s2)
+ * R2 = count (n)
+ *
+ * Favours word aligned data.
+ */
+
+.text
+
+.align 2
+
+ENTRY(_memcmp)
+	I1 = P3;
+	P0 = R0;			/* P0 = s1 address */
+	P3 = R1;			/* P3 = s2 Address  */
+	P2 = R2 ;			/* P2 = count */
+	CC = R2 <= 7(IU);
+	IF CC JUMP .Ltoo_small;
+	I0 = R1;			/* s2 */
+	R1 = R1 | R0;		/* OR addresses together */
+	R1 <<= 30;		/* check bottom two bits */
+	CC =  AZ;			/* AZ set if zero. */
+	IF !CC JUMP .Lbytes ;	/* Jump if addrs not aligned. */
+
+	P1 = P2 >> 2;		/* count = n/4 */
+	R3 =  3;
+	R2 = R2 & R3;		/* remainder */
+	P2 = R2;			/* set remainder */
+
+	LSETUP (.Lquad_loop_s, .Lquad_loop_e) LC0=P1;
+.Lquad_loop_s:
+	MNOP || R0 = [P0++] || R1 = [I0++];
+	CC = R0 == R1;
+	IF !CC JUMP .Lquad_different;
+.Lquad_loop_e:
+	NOP;
+
+	P3 = I0;			/* s2 */
+.Ltoo_small:
+	CC = P2 == 0;		/* Check zero count*/
+	IF CC JUMP .Lfinished;	/* very unlikely*/
+
+.Lbytes:
+	LSETUP (.Lbyte_loop_s, .Lbyte_loop_e) LC0=P2;
+.Lbyte_loop_s:
+	R1 = B[P3++](Z);	/* *s2 */
+	R0 = B[P0++](Z);	/* *s1 */
+	CC = R0 == R1;
+	IF !CC JUMP .Ldifferent;
+.Lbyte_loop_e:
+	NOP;
+
+.Ldifferent:
+	R0 = R0 - R1;
+	P3 = I1;
+	RTS;
+
+.Lquad_different:
+	/* We've read two quads which don't match.
+	 * Can't just compare them, because we're
+	 * a little-endian machine, so the MSBs of
+	 * the regs occur at later addresses in the
+	 * string.
+	 * Arrange to re-read those two quads again,
+	 * byte-by-byte.
+	 */
+	P0 += -4;		/* back up to the start of the */
+	P3 = I0;		/* quads, and increase the*/
+	P2 += 4;		/* remainder count*/
+	P3 += -4;
+	JUMP .Lbytes;
+
+.Lfinished:
+	R0 = 0;
+	P3 = I1;
+	RTS;
+
+.size _memcmp,.-_memcmp
diff --git a/arch/blackfin/lib/memcpy.S b/arch/blackfin/lib/memcpy.S
new file mode 100644
index 0000000..c1e00ef
--- /dev/null
+++ b/arch/blackfin/lib/memcpy.S
@@ -0,0 +1,142 @@
+/*
+ * File:         arch/blackfin/lib/memcpy.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:  internal version of memcpy(), issued by the compiler
+ *               to copy blocks of data around.
+ *               This is really memmove() - it has to be able to deal with
+ *               possible overlaps, because that ambiguity is when the compiler
+ *               gives up and calls a function. We have our own, internal version
+ *               so that we get something we trust, even if the user has redefined
+ *               the normal symbol.
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+/* void *memcpy(void *dest, const void *src, size_t n);
+ * R0 = To Address (dest) (leave unchanged to form result)
+ * R1 = From Address (src)
+ * R2 = count
+ *
+ * Note: Favours word alignment
+ */
+
+#ifdef CONFIG_MEMCPY_L1
+.section .l1.text
+#else
+.text
+#endif
+
+.align 2
+
+ENTRY(_memcpy)
+	CC = R2 <=  0;	/* length not positive? */
+	IF CC JUMP .L_P1L2147483647;	/* Nothing to do */
+
+	P0 = R0 ;	/* dst*/
+	P1 = R1 ;	/* src*/
+	P2 = R2 ;	/* length */
+
+	/* check for overlapping data */
+	CC = R1 < R0;	/* src < dst */
+	IF !CC JUMP .Lno_overlap;
+	R3 = R1 + R2;
+	CC = R0 < R3;	/* and dst < src+len */
+	IF CC JUMP .Lhas_overlap;
+
+.Lno_overlap:
+	/* Check for aligned data.*/
+
+	R3 = R1 | R0;
+	R0 = 0x3;
+	R3 = R3 & R0;
+	CC = R3;	/* low bits set on either address? */
+	IF CC JUMP .Lnot_aligned;
+
+	/* Both addresses are word-aligned, so we can copy
+	at least part of the data using word copies.*/
+	P2 = P2 >> 2;
+	CC = P2 <= 2;
+	IF !CC JUMP .Lmore_than_seven;
+	/* less than eight bytes... */
+	P2 = R2;
+	LSETUP(.Lthree_start, .Lthree_end) LC0=P2;
+	R0 = R1;	/* setup src address for return */
+.Lthree_start:
+	R3 = B[P1++] (X);
+.Lthree_end:
+	B[P0++] = R3;
+
+	RTS;
+
+.Lmore_than_seven:
+	/* There's at least eight bytes to copy. */
+	P2 += -1;	/* because we unroll one iteration */
+	LSETUP(.Lword_loop, .Lword_loop) LC0=P2;
+	R0 = R1;
+	I1 = P1;
+	R3 = [I1++];
+.Lword_loop:
+	MNOP || [P0++] = R3 || R3 = [I1++];
+
+	[P0++] = R3;
+	/* Any remaining bytes to copy? */
+	R3 = 0x3;
+	R3 = R2 & R3;
+	CC = R3 == 0;
+	P1 = I1;	/* in case there's something left, */
+	IF !CC JUMP .Lbytes_left;
+	RTS;
+.Lbytes_left:	P2 = R3;
+.Lnot_aligned:
+	/* From here, we're copying byte-by-byte. */
+	LSETUP (.Lbyte_start, .Lbyte_end) LC0=P2;
+	R0 = R1;	/* Save src address for return */
+.Lbyte_start:
+	R1 = B[P1++] (X);
+.Lbyte_end:
+	B[P0++] = R1;
+
+.L_P1L2147483647:
+	RTS;
+
+.Lhas_overlap:
+	/* Need to reverse the copying, because the
+	 * dst would clobber the src.
+	 * Don't bother to work out alignment for
+	 * the reverse case.
+	 */
+	R0 = R1;	/* save src for later. */
+	P0 = P0 + P2;
+	P0 += -1;
+	P1 = P1 + P2;
+	P1 += -1;
+	LSETUP(.Lover_start, .Lover_end) LC0=P2;
+.Lover_start:
+	R1 = B[P1--] (X);
+.Lover_end:
+	B[P0--] = R1;
+
+	RTS;
diff --git a/arch/blackfin/lib/memmove.S b/arch/blackfin/lib/memmove.S
new file mode 100644
index 0000000..2e5fb7f
--- /dev/null
+++ b/arch/blackfin/lib/memmove.S
@@ -0,0 +1,103 @@
+/*
+ * File:         arch/blackfin/lib/memmove.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+.align 2
+
+/*
+ * C Library function MEMMOVE
+ * R0 = To Address (leave unchanged to form result)
+ * R1 = From Address
+ * R2 = count
+ * Data may overlap
+ */
+
+ENTRY(_memmove)
+	I1 = P3;
+	P0 = R0;                  /* P0 = To address */
+	P3 = R1;                  /* P3 = From Address */
+	P2 = R2;                  /* P2 = count */
+	CC = P2 == 0;             /* Check zero count*/
+	IF CC JUMP .Lfinished;    /* very unlikely */
+
+	CC = R1 < R0 (IU);        /* From < To */
+	IF !CC JUMP .Lno_overlap;
+	R3 = R1 + R2;
+	CC = R0 <= R3 (IU);       /* (From+len) >= To */
+	IF CC JUMP .Loverlap;
+.Lno_overlap:
+	R3 = 11;
+	CC = R2 <= R3;
+	IF CC JUMP .Lbytes;
+	R3 = R1 | R0;             /* OR addresses together */
+	R3 <<= 30;                /* check bottom two bits */
+	CC =  AZ;                 /* AZ set if zero.*/
+	IF !CC JUMP .Lbytes;      /* Jump if addrs not aligned.*/
+
+	I0 = P3;
+	P1 = P2 >> 2;             /* count = n/4 */
+	P1 += -1;
+	R3 =  3;
+	R2 = R2 & R3;             /* remainder */
+	P2 = R2;                  /* set remainder */
+	R1 = [I0++];
+
+	LSETUP (.Lquad_loop, .Lquad_loop) LC0=P1;
+.Lquad_loop: MNOP || [P0++] = R1 || R1 = [I0++];
+	[P0++] = R1;
+
+	CC = P2 == 0;             /* any remaining bytes? */
+	P3 = I0;                  /* Ammend P3 to updated ptr. */
+	IF !CC JUMP .Lbytes;
+	P3 = I1;
+	RTS;
+
+.Lbytes:     LSETUP (.Lbyte2_s, .Lbyte2_e) LC0=P2;
+.Lbyte2_s:   R1 = B[P3++](Z);
+.Lbyte2_e:   B[P0++] = R1;
+
+.Lfinished:  P3 = I1;
+	RTS;
+
+.Loverlap:
+	P2 += -1;
+	P0 = P0 + P2;
+	P3 = P3 + P2;
+	R1 = B[P3--] (Z);
+	CC = P2 == 0;
+	IF CC JUMP .Lno_loop;
+	LSETUP (.Lol_s, .Lol_e) LC0 = P2;
+.Lol_s:    B[P0--] = R1;
+.Lol_e:    R1 = B[P3--] (Z);
+.Lno_loop: B[P0] = R1;
+	P3 = I1;
+	RTS;
+
+.size _memmove,.-_memmove
diff --git a/arch/blackfin/lib/memset.S b/arch/blackfin/lib/memset.S
new file mode 100644
index 0000000..ba6d047
--- /dev/null
+++ b/arch/blackfin/lib/memset.S
@@ -0,0 +1,109 @@
+/*
+ * File:         arch/blackfin/lib/memset.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+.align 2
+
+#ifdef CONFIG_MEMSET_L1
+.section .l1.text
+#else
+.text
+#endif
+
+/*
+ * C Library function MEMSET
+ * R0 = address (leave unchanged to form result)
+ * R1 = filler byte
+ * R2 = count
+ * Favours word aligned data.
+ */
+
+ENTRY(_memset)
+	P0 = R0 ;              /* P0 = address */
+	P2 = R2 ;              /* P2 = count   */
+	R3 = R0 + R2;          /* end          */
+	CC = R2 <= 7(IU);
+	IF CC JUMP  .Ltoo_small;
+	R1 = R1.B (Z);         /* R1 = fill char */
+	R2 =  3;
+	R2 = R0 & R2;          /* addr bottom two bits */
+	CC =  R2 == 0;             /* AZ set if zero.	*/
+	IF !CC JUMP  .Lforce_align ;  /* Jump if addr not aligned. */
+
+.Laligned:
+	P1 = P2 >> 2;          /* count = n/4        */
+	R2 = R1 <<  8;         /* create quad filler */
+	R2.L = R2.L + R1.L(NS);
+	R2.H = R2.L + R1.H(NS);
+	P2 = R3;
+
+	LSETUP (.Lquad_loop , .Lquad_loop) LC0=P1;
+.Lquad_loop:
+	[P0++] = R2;
+
+	CC = P0 == P2;
+	IF !CC JUMP .Lbytes_left;
+	RTS;
+
+.Lbytes_left:
+	R2 = R3;                /* end point */
+	R3 = P0;                /* current position */
+	R2 = R2 - R3;           /* bytes left */
+	P2 = R2;
+
+.Ltoo_small:
+	CC = P2 == 0;           /* Check zero count */
+	IF CC JUMP .Lfinished;    /* Unusual */
+
+.Lbytes:
+	LSETUP (.Lbyte_loop , .Lbyte_loop) LC0=P2;
+.Lbyte_loop:
+	B[P0++] = R1;
+
+.Lfinished:
+	RTS;
+
+.Lforce_align:
+	CC = BITTST (R0, 0);  /* odd byte */
+	R0 = 4;
+	R0 = R0 - R2;
+	P1 = R0;
+	R0 = P0;		    /* Recover return address */
+	IF !CC JUMP .Lskip1;
+	B[P0++] = R1;
+.Lskip1:
+	CC = R2 <= 2;          /* 2 bytes */
+	P2 -= P1;              /* reduce count */
+	IF !CC JUMP .Laligned;
+	B[P0++] = R1;
+	B[P0++] = R1;
+	JUMP .Laligned;
+
+.size _memset,.-_memset
diff --git a/arch/blackfin/lib/modsi3.S b/arch/blackfin/lib/modsi3.S
new file mode 100644
index 0000000..528b8b1
--- /dev/null
+++ b/arch/blackfin/lib/modsi3.S
@@ -0,0 +1,79 @@
+/*
+ * File:         arch/blackfin/lib/modsi3.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:  This program computes 32 bit signed remainder. It calls div32 function
+ *               for quotient estimation.
+ *
+ *               Registers used :
+ *               Numerator/ Denominator in  R0, R1
+ *                 R0  -  returns remainder.
+ *                 R2-R7
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+.global ___modsi3;
+.type ___modsi3, STT_FUNC;
+.extern ___divsi3;
+.type ___divsi3, STT_FUNC;
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+.section .l1.text
+#else
+.text
+#endif
+
+___modsi3:
+
+	CC=R0==0;
+	IF CC JUMP .LRETURN_R0;		/* Return 0, if numerator  == 0 */
+	CC=R1==0;
+	IF CC JUMP .LRETURN_ZERO;		/* Return 0, if denominator == 0 */
+	CC=R0==R1;
+	IF CC JUMP .LRETURN_ZERO;		/* Return 0, if numerator == denominator */
+	CC = R1 == 1;
+	IF CC JUMP .LRETURN_ZERO;		/* Return 0, if denominator ==  1 */
+	CC = R1 == -1;
+	IF CC JUMP .LRETURN_ZERO;		/* Return 0, if denominator == -1 */
+
+	/* Valid input. Use __divsi3() to compute the quotient, and then
+	 * derive the remainder from that. */
+
+	[--SP] = (R7:6);		/* Push  R7 and R6 */
+	[--SP] = RETS;			/* and return address */
+	R7 = R0;			/* Copy of R0 */
+	R6 = R1;			/* Save for later */
+	SP += -12;			/* Should always provide this space */
+	CALL ___divsi3;			/* Compute signed quotient using ___divsi3()*/
+	SP += 12;
+	R0 *= R6;			/* Quotient * divisor */
+	R0 = R7 - R0;			/* Dividend - (quotient * divisor) */
+	RETS = [SP++];			/* Get back return address */
+	(R7:6) = [SP++];		/* Pop registers R7 and R4 */
+	RTS;				/* Store remainder    */
+
+.LRETURN_ZERO:
+	R0 = 0;
+.LRETURN_R0:
+	RTS;
diff --git a/arch/blackfin/lib/muldi3.c b/arch/blackfin/lib/muldi3.c
new file mode 100644
index 0000000..303d0c6
--- /dev/null
+++ b/arch/blackfin/lib/muldi3.c
@@ -0,0 +1,99 @@
+/*
+ * File:         arch/blackfin/lib/muldi3.c
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#ifndef SI_TYPE_SIZE
+#define SI_TYPE_SIZE 32
+#endif
+#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 BITS_PER_UNIT 8
+
+#if !defined(umul_ppmm)
+#define umul_ppmm(w1, w0, u, v)						\
+  do {									\
+    usitype __x0, __x1, __x2, __x3;					\
+    usitype __ul, __vl, __uh, __vh;					\
+									\
+    __ul = __ll_lowpart (u);						\
+    __uh = __ll_highpart (u);						\
+    __vl = __ll_lowpart (v);						\
+    __vh = __ll_highpart (v);						\
+									\
+    __x0 = (usitype) __ul * __vl;					\
+    __x1 = (usitype) __ul * __vh;					\
+    __x2 = (usitype) __uh * __vl;					\
+    __x3 = (usitype) __uh * __vh;					\
+									\
+    __x1 += __ll_highpart (__x0);/* this can't give carry */		\
+    __x1 += __x2;		/* but this indeed can */		\
+    if (__x1 < __x2)		/* did we get it? */			\
+      __x3 += __ll_b;		/* yes, add it in the proper pos. */	\
+									\
+    (w1) = __x3 + __ll_highpart (__x1);					\
+    (w0) = __ll_lowpart (__x1) * __ll_b + __ll_lowpart (__x0);		\
+  } while (0)
+#endif
+
+#if !defined(__umulsidi3)
+#define __umulsidi3(u, v) 						\
+  ({diunion __w;                                                        \
+       umul_ppmm (__w.s.high, __w.s.low, u, v);                         \
+           __w.ll; })
+#endif
+
+typedef unsigned int usitype __attribute__ ((mode(SI)));
+typedef int sitype __attribute__ ((mode(SI)));
+typedef int ditype __attribute__ ((mode(DI)));
+typedef int word_type __attribute__ ((mode(__word__)));
+
+struct distruct {
+	sitype low, high;
+};
+typedef union {
+	struct distruct s;
+	ditype ll;
+} diunion;
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+ditype __muldi3(ditype u, ditype v)__attribute__((l1_text));
+#endif
+
+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/blackfin/lib/outs.S b/arch/blackfin/lib/outs.S
new file mode 100644
index 0000000..f8c876f
--- /dev/null
+++ b/arch/blackfin/lib/outs.S
@@ -0,0 +1,62 @@
+/*
+ * File:         arch/blackfin/lib/outs.S
+ * Based on:
+ * Author:       Bas Vermeulen <bas@buyways.nl>
+ *
+ * Created:      Tue Mar 22 15:27:24 CEST 2005
+ * Description:  Implementation of outs{bwl} for BlackFin processors using zero overhead loops.
+ *
+ * Modified:     Copyright (C) 2005 Bas Vermeulen, BuyWays BV <bas@buyways.nl>
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+.align 2
+
+ENTRY(_outsl)
+	P0 = R0;	/* P0 = port */
+	P1 = R1;	/* P1 = address */
+	P2 = R2;	/* P2 = count */
+
+	LSETUP( .Llong_loop_s, .Llong_loop_e) LC0 = P2;
+.Llong_loop_s: R0 = [P1++];
+.Llong_loop_e: [P0] = R0;
+	RTS;
+
+ENTRY(_outsw)
+	P0 = R0;	/* P0 = port */
+	P1 = R1;	/* P1 = address */
+	P2 = R2;	/* P2 = count */
+
+	LSETUP( .Lword_loop_s, .Lword_loop_e) LC0 = P2;
+.Lword_loop_s: R0 = W[P1++];
+.Lword_loop_e: W[P0] = R0;
+	RTS;
+
+ENTRY(_outsb)
+	P0 = R0;	/* P0 = port */
+	P1 = R1;	/* P1 = address */
+	P2 = R2;	/* P2 = count */
+
+	LSETUP( .Lbyte_loop_s, .Lbyte_loop_e) LC0 = P2;
+.Lbyte_loop_s: R0 = B[P1++];
+.Lbyte_loop_e: B[P0] = R0;
+	RTS;
diff --git a/arch/blackfin/lib/smulsi3_highpart.S b/arch/blackfin/lib/smulsi3_highpart.S
new file mode 100644
index 0000000..10b8f8d
--- /dev/null
+++ b/arch/blackfin/lib/smulsi3_highpart.S
@@ -0,0 +1,30 @@
+.align 2
+.global ___smulsi3_highpart;
+.type ___smulsi3_highpart, STT_FUNC;
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+.section .l1.text
+#else
+.text
+#endif
+
+___smulsi3_highpart:
+	R2 = R1.L * R0.L (FU);
+	R3 = R1.H * R0.L (IS,M);
+	R0 = R0.H * R1.H, R1 = R0.H * R1.L (IS,M);
+
+	R1.L = R2.H + R1.L;
+	cc = ac0;
+	R2 = cc;
+
+	R1.L = R1.L + R3.L;
+	cc = ac0;
+	R1 >>>= 16;
+	R3 >>>= 16;
+	R1 = R1 + R3;
+	R1 = R1 + R2;
+	R2 = cc;
+	R1 = R1 + R2;
+
+	R0 = R0 + R1;
+	RTS;
diff --git a/arch/blackfin/lib/strcmp.c b/arch/blackfin/lib/strcmp.c
new file mode 100644
index 0000000..2ad47c4
--- /dev/null
+++ b/arch/blackfin/lib/strcmp.c
@@ -0,0 +1,11 @@
+#include <linux/types.h>
+
+#define strcmp __inline_strcmp
+#include <asm/string.h>
+#undef strcmp
+
+int strcmp(const char *dest, const char *src)
+{
+	        return __inline_strcmp(dest, src);
+}
+
diff --git a/arch/blackfin/lib/strcpy.c b/arch/blackfin/lib/strcpy.c
new file mode 100644
index 0000000..4dc835a
--- /dev/null
+++ b/arch/blackfin/lib/strcpy.c
@@ -0,0 +1,11 @@
+#include <linux/types.h>
+
+#define strcpy __inline_strcpy
+#include <asm/string.h>
+#undef strcpy
+
+char *strcpy(char *dest, const char *src)
+{
+	        return __inline_strcpy(dest, src);
+}
+
diff --git a/arch/blackfin/lib/strncmp.c b/arch/blackfin/lib/strncmp.c
new file mode 100644
index 0000000..947bcfe
--- /dev/null
+++ b/arch/blackfin/lib/strncmp.c
@@ -0,0 +1,11 @@
+#include <linux/types.h>
+
+#define strncmp __inline_strncmp
+#include <asm/string.h>
+#undef strncmp
+
+int strncmp(const char *cs, const char *ct, size_t count)
+{
+	        return __inline_strncmp(cs, ct, count);
+}
+
diff --git a/arch/blackfin/lib/strncpy.c b/arch/blackfin/lib/strncpy.c
new file mode 100644
index 0000000..77a9b2e
--- /dev/null
+++ b/arch/blackfin/lib/strncpy.c
@@ -0,0 +1,11 @@
+#include <linux/types.h>
+
+#define strncpy __inline_strncpy
+#include <asm/string.h>
+#undef strncpy
+
+char *strncpy(char *dest, const char *src, size_t n)
+{
+	        return __inline_strncpy(dest, src, n);
+}
+
diff --git a/arch/blackfin/lib/udivsi3.S b/arch/blackfin/lib/udivsi3.S
new file mode 100644
index 0000000..d39a129
--- /dev/null
+++ b/arch/blackfin/lib/udivsi3.S
@@ -0,0 +1,298 @@
+/*
+ * File:         arch/blackfin/lib/udivsi3.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#include <linux/linkage.h>
+
+#define CARRY AC0
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+.section .l1.text
+#else
+.text
+#endif
+
+
+ENTRY(___udivsi3)
+
+  CC = R0 < R1 (IU);    /* If X < Y, always return 0 */
+  IF CC JUMP .Lreturn_ident;
+
+  R2 = R1 << 16;
+  CC = R2 <= R0 (IU);
+  IF CC JUMP .Lidents;
+
+  R2 = R0 >> 31;       /* if X is a 31-bit number */
+  R3 = R1 >> 15;       /* and Y is a 15-bit number */
+  R2 = R2 | R3;        /* then it's okay to use the DIVQ builtins (fallthrough to fast)*/
+  CC = R2;
+  IF CC JUMP .Ly_16bit;
+
+/* METHOD 1: FAST DIVQ
+   We know we have a 31-bit dividend, and 15-bit divisor so we can use the
+   simple divq approach (first setting AQ to 0 - implying unsigned division,
+   then 16 DIVQ's).
+*/
+
+  AQ = CC;             /* Clear AQ (CC==0) */
+
+/* ISR States: When dividing two integers (32.0/16.0) using divide primitives,
+   we need to shift the dividend one bit to the left.
+   We have already checked that we have a 31-bit number so we are safe to do
+   that.
+*/
+  R0 <<= 1;
+  DIVQ(R0, R1); // 1
+  DIVQ(R0, R1); // 2
+  DIVQ(R0, R1); // 3
+  DIVQ(R0, R1); // 4
+  DIVQ(R0, R1); // 5
+  DIVQ(R0, R1); // 6
+  DIVQ(R0, R1); // 7
+  DIVQ(R0, R1); // 8
+  DIVQ(R0, R1); // 9
+  DIVQ(R0, R1); // 10
+  DIVQ(R0, R1); // 11
+  DIVQ(R0, R1); // 12
+  DIVQ(R0, R1); // 13
+  DIVQ(R0, R1); // 14
+  DIVQ(R0, R1); // 15
+  DIVQ(R0, R1); // 16
+  R0 = R0.L (Z);
+  RTS;
+
+.Ly_16bit:
+  /* We know that the upper 17 bits of Y might have bits set,
+  ** or that the sign bit of X might have a bit. If Y is a
+  ** 16-bit number, but not bigger, then we can use the builtins
+  ** with a post-divide correction.
+  ** R3 currently holds Y>>15, which means R3's LSB is the
+  ** bit we're interested in.
+  */
+
+  /* According to the ISR, to use the Divide primitives for
+  ** unsigned integer divide, the useable range is 31 bits
+  */
+  CC = ! BITTST(R0, 31);
+
+  /* IF condition is true we can scale our inputs and use the divide primitives,
+  ** with some post-adjustment
+  */
+  R3 += -1;		/* if so, Y is 0x00008nnn */
+  CC &= AZ;
+
+  /* If condition is true we can scale our inputs and use the divide primitives,
+  ** with some post-adjustment
+  */
+  R3 = R1 >> 1;		/* Pre-scaled divisor for primitive case */
+  R2 = R0 >> 16;
+
+  R2 = R3 - R2;		/* shifted divisor < upper 16 bits of dividend */
+  CC &= CARRY;
+  IF CC JUMP .Lshift_and_correct;
+
+  /* Fall through to the identities */
+
+/* METHOD 2: identities and manual calculation
+   We are not able to use the divide primites, but may still catch some special
+   cases.
+*/
+.Lidents:
+  /* Test for common identities. Value to be returned is placed in R2. */
+  CC = R0 == 0;        /* 0/Y => 0 */
+  IF CC JUMP .Lreturn_r0;
+  CC = R0 == R1;       /* X==Y => 1 */
+  IF CC JUMP .Lreturn_ident;
+  CC = R1 == 1;        /* X/1 => X */
+  IF CC JUMP .Lreturn_ident;
+
+  R2.L = ONES R1;
+  R2 = R2.L (Z);
+  CC = R2 == 1;
+  IF CC JUMP .Lpower_of_two;
+
+  [--SP] = (R7:5);                /* Push registers R5-R7 */
+
+  /* Idents don't match. Go for the full operation. */
+
+
+  R6 = 2;                         /* assume we'll shift two */
+  R3 = 1;
+
+  P2 = R1;
+                                  /* If either R0 or R1 have sign set, */
+                                  /* divide them by two, and note it's */
+                                  /* been done. */
+  CC = R1 < 0;
+  R2 = R1 >> 1;
+  IF CC R1 = R2;                  /* Possibly-shifted R1 */
+  IF !CC R6 = R3;                 /* R1 doesn't, so at most 1 shifted */
+
+  P0 = 0;
+  R3 = -R1;
+  [--SP] = R3;
+  R2 = R0 >> 1;
+  R2 = R0 >> 1;
+  CC = R0 < 0;
+  IF CC P0 = R6;                  /* Number of values divided */
+  IF !CC R2 = R0;                 /* Shifted R0 */
+
+                                  /* P0 is 0, 1 (NR/=2) or 2 (NR/=2, DR/=2) */
+
+                                  /* r2 holds Copy dividend  */
+  R3 = 0;                         /* Clear partial remainder */
+  R7 = 0;                         /* Initialise quotient bit */
+
+  P1 = 32;                        /* Set loop counter */
+  LSETUP(.Lulst, .Lulend) LC0 = P1; /* Set loop counter */
+.Lulst:  R6 = R2 >> 31;             /* R6 = sign bit of R2, for carry */
+       R2 = R2 << 1;              /* Shift 64 bit dividend up by 1 bit */
+       R3 = R3 << 1 || R5 = [SP];
+       R3 = R3 | R6;              /* Include any carry */
+       CC = R7 < 0;               /* Check quotient(AQ) */
+                                  /* If AQ==0, we'll sub divisor */
+       IF CC R5 = R1;             /* and if AQ==1, we'll add it. */
+       R3 = R3 + R5;              /* Add/sub divsor to partial remainder */
+       R7 = R3 ^ R1;              /* Generate next quotient bit */
+
+       R5 = R7 >> 31;             /* Get AQ */
+       BITTGL(R5, 0);             /* Invert it, to get what we'll shift */
+.Lulend: R2 = R2 + R5;              /* and "shift" it in. */
+
+  CC = P0 == 0;                   /* Check how many inputs we shifted */
+  IF CC JUMP .Lno_mult;            /* if none... */
+  R6 = R2 << 1;
+  CC = P0 == 1;
+  IF CC R2 = R6;                  /* if 1, Q = Q*2 */
+  IF !CC R1 = P2;                 /* if 2, restore stored divisor */
+
+  R3 = R2;                        /* Copy of R2 */
+  R3 *= R1;                       /* Q * divisor */
+  R5 = R0 - R3;                   /* Z = (dividend - Q * divisor) */
+  CC = R1 <= R5 (IU);             /* Check if divisor <= Z? */
+  R6 = CC;                        /* if yes, R6 = 1 */
+  R2 = R2 + R6;                   /* if yes, add one to quotient(Q) */
+.Lno_mult:
+  SP += 4;
+  (R7:5) = [SP++];                /* Pop registers R5-R7 */
+  R0 = R2;                        /* Store quotient */
+  RTS;
+
+.Lreturn_ident:
+  CC = R0 < R1 (IU);    /* If X < Y, always return 0 */
+  R2 = 0;
+  IF CC JUMP .Ltrue_return_ident;
+  R2 = -1 (X);         /* X/0 => 0xFFFFFFFF */
+  CC = R1 == 0;
+  IF CC JUMP .Ltrue_return_ident;
+  R2 = -R2;            /* R2 now 1 */
+  CC = R0 == R1;       /* X==Y => 1 */
+  IF CC JUMP .Ltrue_return_ident;
+  R2 = R0;             /* X/1 => X */
+  /*FALLTHRU*/
+
+.Ltrue_return_ident:
+  R0 = R2;
+.Lreturn_r0:
+  RTS;
+
+.Lpower_of_two:
+  /* Y has a single bit set, which means it's a power of two.
+  ** That means we can perform the division just by shifting
+  ** X to the right the appropriate number of bits
+  */
+
+  /* signbits returns the number of sign bits, minus one.
+  ** 1=>30, 2=>29, ..., 0x40000000=>0. Which means we need
+  ** to shift right n-signbits spaces. It also means 0x80000000
+  ** is a special case, because that *also* gives a signbits of 0
+  */
+
+  R2 = R0 >> 31;
+  CC = R1 < 0;
+  IF CC JUMP .Ltrue_return_ident;
+
+  R1.l = SIGNBITS R1;
+  R1 = R1.L (Z);
+  R1 += -30;
+  R0 = LSHIFT R0 by R1.L;
+  RTS;
+
+/* METHOD 3: PRESCALE AND USE THE DIVIDE PRIMITIVES WITH SOME POST-CORRECTION
+  Two scaling operations are required to use the divide primitives with a
+  divisor > 0x7FFFF.
+  Firstly (as in method 1) we need to shift the dividend 1 to the left for
+  integer division.
+  Secondly we need to shift both the divisor and dividend 1 to the right so
+  both are in range for the primitives.
+  The left/right shift of the dividend does nothing so we can skip it.
+*/
+.Lshift_and_correct:
+  R2 = R0;
+  // R3 is already R1 >> 1
+  CC=!CC;
+  AQ = CC;                        /* Clear AQ, got here with CC = 0 */
+  DIVQ(R2, R3); // 1
+  DIVQ(R2, R3); // 2
+  DIVQ(R2, R3); // 3
+  DIVQ(R2, R3); // 4
+  DIVQ(R2, R3); // 5
+  DIVQ(R2, R3); // 6
+  DIVQ(R2, R3); // 7
+  DIVQ(R2, R3); // 8
+  DIVQ(R2, R3); // 9
+  DIVQ(R2, R3); // 10
+  DIVQ(R2, R3); // 11
+  DIVQ(R2, R3); // 12
+  DIVQ(R2, R3); // 13
+  DIVQ(R2, R3); // 14
+  DIVQ(R2, R3); // 15
+  DIVQ(R2, R3); // 16
+
+  /* According to the Instruction Set Reference:
+     To divide by a divisor > 0x7FFF,
+     1. prescale and perform divide to obtain quotient (Q) (done above),
+     2. multiply quotient by unscaled divisor (result M)
+     3. subtract the product from the divident to get an error (E = X - M)
+     4. if E < divisor (Y) subtract 1, if E > divisor (Y) add 1, else return quotient (Q)
+   */
+  R3 = R2.L (Z);		/* Q = X' / Y' */
+  R2 = R3;		/* Preserve Q */
+  R2 *= R1;		/* M = Q * Y */
+  R2 = R0 - R2;		/* E = X - M */
+  R0 = R3;		/* Copy Q into result reg */
+
+/* Correction: If result of the multiply is negative, we overflowed
+   and need to correct the result by subtracting 1 from the result.*/
+  R3 = 0xFFFF (Z);
+  R2 = R2 >> 16;		/* E >> 16 */
+  CC = R2 == R3;
+  R3 = 1 ;
+  R1 = R0 - R3;
+  IF CC R0 = R1;
+  RTS;
diff --git a/arch/blackfin/lib/umodsi3.S b/arch/blackfin/lib/umodsi3.S
new file mode 100644
index 0000000..b55ce96
--- /dev/null
+++ b/arch/blackfin/lib/umodsi3.S
@@ -0,0 +1,66 @@
+/*
+ * File:         arch/blackfin/lib/umodsi3.S
+ * Based on:
+ * Author:
+ *
+ * Created:
+ * Description:  libgcc1 routines for Blackfin 5xx
+ *
+ * Modified:
+ *               Copyright 2004-2006 Analog Devices Inc.
+ *
+ * Bugs:         Enter bugs at http://blackfin.uclinux.org/
+ *
+ * This program 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 of the License, or
+ * (at your option) any later version.
+ *
+ * This program 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, see the file COPYING, or write
+ * to the Free Software Foundation, Inc.,
+ * 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+.section .l1.text
+#else
+.text
+#endif
+
+.extern ___udivsi3;
+.globl	___umodsi3
+___umodsi3:
+
+	CC=R0==0;
+	IF CC JUMP .LRETURN_R0;		/* Return 0, if NR == 0 */
+	CC= R1==0;
+	IF CC JUMP .LRETURN_ZERO_VAL;	/* Return 0, if DR == 0 */
+	CC=R0==R1;
+	IF CC JUMP .LRETURN_ZERO_VAL;	/* Return 0, if NR == DR */
+	CC = R1 == 1;
+	IF CC JUMP .LRETURN_ZERO_VAL;	/* Return 0, if  DR == 1 */
+	CC = R0<R1 (IU);
+	IF CC JUMP .LRETURN_R0;		/* Return dividend (R0),IF NR<DR */
+
+	[--SP] = (R7:6);		/* Push registers and */
+	[--SP] = RETS;			/* Return address */
+	R7 = R0;			/* Copy of R0 */
+	R6 = R1;
+	SP += -12;			/* Should always provide this space */
+	CALL ___udivsi3;		/* Compute unsigned quotient using ___udiv32()*/
+	SP += 12;
+	R0 *= R6;			/* Quotient * divisor */
+	R0 = R7 - R0;			/* Dividend - (quotient * divisor) */
+	RETS = [SP++];			/* Pop return address */
+	( R7:6) = [SP++];		/* And registers */
+	RTS;				/* Return remainder */
+.LRETURN_ZERO_VAL:
+	R0 = 0;
+.LRETURN_R0:
+	RTS;
diff --git a/arch/blackfin/lib/umulsi3_highpart.S b/arch/blackfin/lib/umulsi3_highpart.S
new file mode 100644
index 0000000..aac8218
--- /dev/null
+++ b/arch/blackfin/lib/umulsi3_highpart.S
@@ -0,0 +1,23 @@
+.align 2
+.global ___umulsi3_highpart;
+.type ___umulsi3_highpart, STT_FUNC;
+
+#ifdef CONFIG_ARITHMETIC_OPS_L1
+.section .l1.text
+#else
+.text
+#endif
+
+___umulsi3_highpart:
+	R2 = R1.H * R0.H, R3 = R1.L * R0.H (FU);
+	R0 = R1.L * R0.L, R1 = R1.H * R0.L (FU);
+	R0 >>= 16;
+	/* Unsigned multiplication has the nice property that we can
+	   ignore carry on this first addition.  */
+	R0 = R0 + R3;
+	R0 = R0 + R1;
+	cc = ac0;
+	R1 = cc;
+	R1 = PACK(R1.l,R0.h);
+	R0 = R1 + R2;
+	RTS;