Import of Linux/MIPS 1.3.0

1 files changed, 237 insertions, 0 deletions

diff --git a/include/net/checksum.h b/include/net/checksum.h
new file mode 100644
index 000000000..7f9492328
--- /dev/null
+++ b/include/net/checksum.h
@@ -0,0 +1,237 @@
+/*
+ * INET		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.
+ *
+ *		Checksumming functions for IP, TCP, UDP and so on
+ *
+ * Authors:	Jorge Cwik, <jorge@laser.satlink.net>
+ *		Arnt Gulbrandsen, <agulbra@nvg.unit.no>
+ *		Borrows very liberally from tcp.c and ip.c, see those
+ *		files for more names.
+ *
+ *		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.
+ */
+#ifndef _CHECKSUM_H
+#define _CHECKSUM_H
+
+#include <asm/byteorder.h>
+#include "ip.h"
+
+/*
+ *	This is a version of ip_compute_csum() optimized for IP headers,
+ *	which always checksum on 4 octet boundaries.
+ *
+ *	By Jorge Cwik <jorge@laser.satlink.net>, adapted for linux by
+ *	Arnt Gulbrandsen.
+ */
+static inline unsigned short ip_fast_csum(unsigned char * iph,
+					  unsigned int ihl) {
+	unsigned short int sum;
+
+#ifdef __i386__
+	__asm__("
+	    movl (%%esi), %%eax
+	    andl $15, %%ecx
+	    subl $4, %%ecx
+	    jbe 2f
+	    addl 4(%%esi), %%eax
+	    adcl 8(%%esi), %%eax
+	    adcl 12(%%esi), %%eax
+1:	    adcl 16(%%esi), %%eax
+	    lea 4(%%esi), %%esi
+	    decl %%ecx
+	    jne	1b
+	    adcl $0, %%eax
+	    movl %%eax, %%ecx
+	    shrl $16, %%eax
+	    addw %%ecx, %%eax
+	    adcl $0, %%eax
+	    notl %%eax
+	    andl $65535, %%eax
+2:
+	    "
+	: "=a" (sum)
+	: "S" (iph), "c"(ihl)
+	: "ax", "cx", "si");
+#elif defined (__mips__)
+	unsigned long	dummy1, dummy2;
+	/*
+	 * This is optimized for 32-bit MIPS processors.
+	 * I tried it in plain C but the generated code looks to bad to
+	 * use with old first generation MIPS CPUs.
+	 * Using 64-bit code could even further improve these routines.
+	 */
+	__asm__("
+	.set	noreorder
+	.set	noat
+	lw	%0,(%3)
+	subu	%1,4
+	blez	%1,2f
+	sll	%1,%4,2			# delay slot
+	lw	%2,4(%3)
+	addu	%1,%3			# delay slot
+	addu	%0,%2
+	sltu	$1,%0,%2
+	lw	%2,8(%3)
+	addu	%0,$1
+	addu	%0,%2
+	sltu	$1,%0,%2
+	lw	%2,12(%3)
+	addu	%0,$1
+	addu	%0,%2
+	sltu	$1,%0,%2
+	addu	%0,$1
+1:	lw	%2,16(%3)
+	addu	%1,4
+	addu	%0,%2
+	sltu	$1,%0,%2
+	bne	%1,%3,1b
+	addu	%0,$1			# delay slot
+	srl	$1,%0,16
+	addu	%0,$1
+	sltu	$1,%0,$1
+	addu	%0,$1
+	nor	%0,$0,%0
+	andi	%0,0xffff
+2:	.set	at
+	.set	reorder"
+	: "=r" (sum), "=r" (dummy1), "=r" (dummy2)
+	: "r" (iph), "r"(ihl)
+	: "$1");
+#else
+#error Not implemented for this CPU
+#endif
+	return(sum);
+}
+
+
+
+
+/*
+ * computes the checksum of the TCP/UDP pseudo-header
+ * returns a 16-bit checksum, already complemented
+ */
+
+static inline unsigned short int csum_tcpudp_magic(unsigned long saddr,
+						   unsigned long daddr,
+						   unsigned short len,
+						   unsigned short proto,
+						   unsigned int sum) {
+#ifdef __i386__
+    __asm__("
+	addl %2, %0
+	adcl %3, %0
+	adcl %4, %0
+	adcl $0, %0
+	movl %0, %2
+	shrl $16, %2
+	addw %2, %0
+	adcl $0, %0
+	notl %0
+	andl $65535, %0
+	"
+	: "=r" (sum)
+	: "0" (daddr), "S"(saddr), "r"((ntohs(len)<<16)+proto*256), "r"(sum)
+	: "si" );
+#elif defined (__mips__)
+    __asm__("
+	.set	noat
+	addu	%0,%2
+	sltu	$1,%0,%2
+	addu	%0,$1
+	addu	%0,%3
+	sltu	$1,%0,%3
+	addu	%0,$1
+	addu	%0,%4
+	sltu	$1,%0,%4
+	addu	%0,$1
+	srl	$1,%0,16
+	addu	%0,$1
+	sltu	$1,%0,$1
+	addu	%0,$1
+	nor	%0,$0,%0
+	andi	%0,0xffff
+	.set	at"
+	: "=r" (sum)
+	: "0" (daddr), "r"(saddr), "r"((ntohs(len)<<16)+proto*256), "r"(sum)
+	: "$1");
+#else
+#error Not implemented for this CPU
+#endif
+	return((unsigned short)sum);
+}
+
+
+
+/*
+ * 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(unsigned char * buff, int len, unsigned int sum);
+
+
+
+/*
+ * the same as csum_partial, but copies from fs:src while it
+ * checksums
+ *
+ * here even more important to align src and dst on a 32-bit (or even
+ * better 64-bit) boundary
+ */
+
+unsigned int csum_partial_copyffs( char *src, char *dst, int len, int sum);
+
+
+
+
+/*
+ * this routine is used for miscellaneous IP-like checksums, mainly
+ * in icmp.c
+ */
+
+static inline unsigned short ip_compute_csum(unsigned char * buff, int len) {
+    unsigned short int sum;
+
+#ifdef __i386__
+    __asm__("
+	movl %%eax, %%ecx
+	shrl $16, %%ecx
+	addw %%cx, %%ax
+	adcl $0, %%eax
+	notl %%eax
+	andl $65535, %%eax
+	"
+	: "=a"(sum)
+	: "a" (csum_partial(buff, len, 0))
+	: "cx");
+#elif defined (__mips__)
+    __asm__("
+	.set	noat
+	srl	$1,%0,16
+	addu	%0,$1
+	sltu	$1,%0,$1
+	nor	%0,$0,%0
+	andi	%0,0xffff
+	.set	at"
+	: "=r"(sum)
+	: "r" (csum_partial(buff, len, 0))
+	: "$1");
+#else
+#error Not implemented for this CPU
+#endif
+	return(sum);
+}
+
+#endif