Import of Linux/MIPS 2.1.14

79 files changed, 8144 insertions, 5629 deletions

diff --git a/arch/i386/Makefile b/arch/i386/Makefile
index 0b296da44..429b2b2b4 100644
--- a/arch/i386/Makefile
+++ b/arch/i386/Makefile
@@ -13,39 +13,65 @@
 # Copyright (C) 1994 by Linus Torvalds
 #
 
+AS86    =$(CROSS_COMPILE)as86 -0 -a
+AS386   =$(CROSS_COMPILE)as86 -3
+LD86    =$(CROSS_COMPILE)ld86 -0
+
 #
-# Set these to indicate how to link it..
-#
-# -zmagic:   
+# ZIMAGE_OFFSET is the load offset of the compression loader
+# BZIMAGE_OFFSET is the load offset of the high loaded compression loader
 #
-# ZLINKFLAGS   = -Ttext 0x1000
-# LINKFLAGS    = -Ttext 0x100000
+ZIMAGE_OFFSET=0x1000
+BZIMAGE_OFFSET=0x100000
+
 #
-# -qmagic (we need to remove the 32 byte header for bootup purposes)
+# IMAGE_OFFSET is the load offset of the _real_ kernel, soon
+# to be offset by another 0xC0000000...
 #
-ZLINKFLAGS =-qmagic -Ttext 0xfe0
-LINKFLAGS =-qmagic -Ttext 0xfffe0
+IMAGE_OFFSET=0xC0100000
+
+# This is used for ELF - it needs to migrate or be moved.
+LD_RFLAG = -m elf_i386
+
+LD=$(CROSS_COMPILE)ld -m elf_i386
+CPP=$(CC) -E -D__ELF__
+OBJDUMP=$(CROSS_COMPILE)objdump
+OBJDUMP_FLAGS=-k -q 
+ENCAPS=$(CROSS_COMPILE)encaps
+OBJCOPY=$(CROSS_COMPILE)objcopy -O binary -R .note -R .comment -R .stab -R .stabstr
+ZLDFLAGS=-e startup_32 
+LDFLAGS=-e stext
+ZLINKFLAGS =-Ttext $(ZIMAGE_OFFSET) $(ZLDFLAGS)
+BZLINKFLAGS =-Ttext $(BZIMAGE_OFFSET) $(ZLDFLAGS)
+LINKFLAGS =-Ttext $(IMAGE_OFFSET) $(LDFLAGS)
+
 CFLAGS := $(CFLAGS) -pipe
 
+ifdef CONFIG_M386
+CFLAGS := $(CFLAGS) -m386 -DCPU=386
+endif
+
 ifdef CONFIG_M486
-CFLAGS := $(CFLAGS) -m486
-else
+CFLAGS := $(CFLAGS) -m486 -DCPU=486
+endif
+
 ifdef CONFIG_M586
-CFLAGS := $(CFLAGS) -mpentium
-else
-CFLAGS := $(CFLAGS) -m386
+CFLAGS := $(CFLAGS) -m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=586
 endif
+
+ifdef CONFIG_M686
+CFLAGS := $(CFLAGS) -m486 -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=686
+endif
+
+ifdef SMP
+CFLAGS := $(CFLAGS) -D__SMP__
 endif
 
 HEAD := arch/i386/kernel/head.o
 
-SUBDIRS := $(SUBDIRS) arch/i386/kernel arch/i386/mm
+SUBDIRS := $(SUBDIRS) arch/i386/kernel arch/i386/mm arch/i386/lib
 ARCHIVES := arch/i386/kernel/kernel.o arch/i386/mm/mm.o $(ARCHIVES)
-
-ifdef CONFIG_IBCS
-SUBDIRS := $(SUBDIRS) arch/i386/ibcs
-DRIVERS := $(DRIVERS) arch/i386/ibcs/ibcs.o
-endif
+LIBS := $(TOPDIR)/arch/i386/lib/lib.a $(LIBS) $(TOPDIR)/arch/i386/lib/lib.a
 
 ifdef CONFIG_MATH_EMULATION
 SUBDIRS := $(SUBDIRS) arch/i386/math-emu
@@ -63,19 +89,29 @@ MAKEBOOT = $(MAKE) -C arch/$(ARCH)/boot
 zImage: vmlinux
 	@$(MAKEBOOT) zImage
 
+bzImage: vmlinux
+	@$(MAKEBOOT) bzImage
+
 compressed: zImage
 
 zlilo: vmlinux
-	@$(MAKEBOOT) zlilo
+	@$(MAKEBOOT) BOOTIMAGE=zImage zlilo
+
+bzlilo: vmlinux
+	@$(MAKEBOOT) BOOTIMAGE=bzImage zlilo
 
 zdisk: vmlinux
-	@$(MAKEBOOT) zdisk
+	@$(MAKEBOOT) BOOTIMAGE=zImage zdisk
+
+bzdisk: vmlinux
+	@$(MAKEBOOT) BOOTIMAGE=bzImage zdisk
 
 install: vmlinux
 	@$(MAKEBOOT) install
 
 archclean:
 	@$(MAKEBOOT) clean
+	$(MAKE) -C arch/$(ARCH)/kernel clean
 
 archdep:
 	@$(MAKEBOOT) dep
diff --git a/arch/i386/boot/Makefile b/arch/i386/boot/Makefile
index b6a5f1c15..503b0cce6 100644
--- a/arch/i386/boot/Makefile
+++ b/arch/i386/boot/Makefile
@@ -8,23 +8,43 @@
 # Copyright (C) 1994 by Linus Torvalds
 #
 
-AS86    =as86 -0 -a
-LD86    =ld86 -0
+HOSTCFLAGS := $(HOSTCFLAGS) -D__BFD__
+
+ifdef SMP
+HOSTCFLAGS := $(HOSTCFLAGS) -D__SMP__
+endif
 
 zImage: $(CONFIGURE) bootsect setup compressed/vmlinux tools/build
-	tools/build bootsect setup compressed/vmlinux $(ROOT_DEV) > zImage
+	if hash $(ENCAPS) 2> /dev/null; then \
+	  $(OBJDUMP) $(OBJDUMP_FLAGS) -o $(ZIMAGE_OFFSET) compressed/vmlinux > compressed/vmlinux.out; \
+	else \
+	  $(OBJCOPY) compressed/vmlinux compressed/vmlinux.out; \
+	fi
+	tools/build bootsect setup compressed/vmlinux.out $(ROOT_DEV) > zImage
+	sync
+
+bzImage: $(CONFIGURE) bbootsect setup compressed/bvmlinux tools/bbuild
+	if hash $(ENCAPS) 2> /dev/null; then \
+	  $(OBJDUMP) $(OBJDUMP_FLAGS) -o $(BZIMAGE_OFFSET) compressed/bvmlinux > compressed/bvmlinux.out; \
+	else \
+	  $(OBJCOPY) compressed/bvmlinux compressed/bvmlinux.out; \
+	fi
+	tools/bbuild bbootsect setup compressed/bvmlinux.out $(ROOT_DEV) > bzImage
 	sync
 
 compressed/vmlinux: $(TOPDIR)/vmlinux
 	@$(MAKE) -C compressed vmlinux
 
-zdisk: zImage
-	dd bs=8192 if=zImage of=/dev/fd0
+compressed/bvmlinux: $(TOPDIR)/vmlinux
+	@$(MAKE) -C compressed bvmlinux
 
-zlilo: $(CONFIGURE) zImage
+zdisk: $(BOOTIMAGE)
+	dd bs=8192 if=$(BOOTIMAGE) of=/dev/fd0
+
+zlilo: $(CONFIGURE) $(BOOTIMAGE)
 	if [ -f $(INSTALL_PATH)/vmlinuz ]; then mv $(INSTALL_PATH)/vmlinuz $(INSTALL_PATH)/vmlinuz.old; fi
 	if [ -f $(INSTALL_PATH)/System.map ]; then mv $(INSTALL_PATH)/System.map $(INSTALL_PATH)/System.old; fi
-	cat zImage > $(INSTALL_PATH)/vmlinuz
+	cat $(BOOTIMAGE) > $(INSTALL_PATH)/vmlinuz
 	cp $(TOPDIR)/System.map $(INSTALL_PATH)/
 	if [ -x /sbin/lilo ]; then /sbin/lilo; else /etc/lilo/install; fi
 
@@ -32,7 +52,10 @@ install: $(CONFIGURE) zImage
 	sh ./install.sh $(VERSION).$(PATCHLEVEL).$(SUBLEVEL) zImage $(TOPDIR)/System.map "$(INSTALL_PATH)"
 
 tools/build: tools/build.c
-	$(HOSTCC) $(CFLAGS) -o $@ $< -I$(TOPDIR)/include
+	$(HOSTCC) $(HOSTCFLAGS) -o $@ $< -I$(TOPDIR)/include
+
+tools/bbuild: tools/build.c
+	$(HOSTCC) $(HOSTCFLAGS) -D__BIG_KERNEL__ -o $@ $< -I$(TOPDIR)/include
 
 setup: setup.o
 	$(LD86) -s -o $@ $<
@@ -40,7 +63,7 @@ setup: setup.o
 setup.o: setup.s
 	$(AS86) -o $@ $<
 
-setup.s: setup.S $(CONFIGURE) $(TOPDIR)/include/linux/config.h Makefile
+setup.s: setup.S video.S $(CONFIGURE) $(TOPDIR)/include/linux/config.h Makefile
 	$(CPP) -traditional $(SVGA_MODE) $(RAMDISK) $< -o $@
 
 bootsect: bootsect.o
@@ -52,9 +75,20 @@ bootsect.o: bootsect.s
 bootsect.s: bootsect.S $(CONFIGURE) $(TOPDIR)/include/linux/config.h Makefile
 	$(CPP) -traditional $(SVGA_MODE) $(RAMDISK) $< -o $@
 
+bbootsect: bbootsect.o
+	$(LD86) -s -o $@ $<
+
+bbootsect.o: bbootsect.s
+	$(AS86) -o $@ $<
+
+bbootsect.s: bootsect.S $(CONFIGURE) $(TOPDIR)/include/linux/config.h Makefile
+	$(CPP) -D__BIG_KERNEL__ -traditional $(SVGA_MODE) $(RAMDISK) $< -o $@
+
 dep:
 
 clean:
 	rm -f bootsect setup
-	rm -f zImage tools/build
+	rm -f bbootsect
+	rm -f zImage tools/build compressed/vmlinux.out
+	rm -f bzImage tools/bbuild compressed/bvmlinux.out
 	@$(MAKE) -C compressed clean
diff --git a/arch/i386/boot/bootsect.S b/arch/i386/boot/bootsect.S
index f6a0d3158..88a04190d 100644
--- a/arch/i386/boot/bootsect.S
+++ b/arch/i386/boot/bootsect.S
@@ -47,7 +47,7 @@ SWAP_DEV = 0
 #define RAMDISK 0
 #endif 
 #ifndef CONFIG_ROOT_RDONLY
-#define CONFIG_ROOT_RDONLY 0
+#define CONFIG_ROOT_RDONLY 1
 #endif
 
 ! ld86 requires an entry symbol. This may as well be the usual one.
@@ -268,8 +268,14 @@ read_it:
 die:	jne die			! es must be at 64kB boundary
 	xor bx,bx		! bx is starting address within segment
 rp_read:
+#ifdef __BIG_KERNEL__
+#define CALL_HIGHLOAD_KLUDGE .word 0x1eff,0x220 ! call far * bootsect_kludge
+				! NOTE: as86 can't assemble this
+	CALL_HIGHLOAD_KLUDGE	! this is within setup.S
+#else
 	mov ax,es
 	sub ax,#SYSSEG
+#endif
 	cmp ax,syssize		! have we loaded all yet?
 	jbe ok1_read
 	ret
diff --git a/arch/i386/boot/compressed/Makefile b/arch/i386/boot/compressed/Makefile
index b4e3f8482..d7f093e3f 100644
--- a/arch/i386/boot/compressed/Makefile
+++ b/arch/i386/boot/compressed/Makefile
@@ -7,35 +7,57 @@
 HEAD = head.o
 SYSTEM = $(TOPDIR)/vmlinux
 
-OBJECTS = $(HEAD) inflate.o unzip.o misc.o
+OBJECTS = $(HEAD) misc.o
 
 CFLAGS = -O2 -DSTDC_HEADERS
 
-.c.s:
-	$(CC) $(CFLAGS) -S $<
-.s.o:
-	$(AS) -o $*.o $<
-.c.o:
-	$(CC) $(CFLAGS) -c $<
+ifdef SMP
+CFLAGS := $(CFLAGS) -D__SMP__
+endif
+
+TARGET=--target elf32-i386
+INPUT_DATA=input_data
+INPUT_LEN=input_len
 
 all: vmlinux
 
-vmlinux:	piggy.o $(OBJECTS)
-		$(LD) $(ZLINKFLAGS) -o vmlinux $(OBJECTS) piggy.o
+vmlinux: piggy.o $(OBJECTS)
+	$(LD) $(ZLINKFLAGS) -o vmlinux $(OBJECTS) piggy.o
+
+bvmlinux: piggy.o $(OBJECTS)
+	$(LD) $(BZLINKFLAGS) -o bvmlinux $(OBJECTS) piggy.o
+
+ifdef SMP
+
+head.o: head.S $(TOPDIR)/include/linux/tasks.h
+	$(CC) -D__SMP__ -traditional -c head.S
 
-head.o:	head.s
+else
 
-head.s: head.S $(TOPDIR)/include/linux/tasks.h
-	$(CPP) -traditional head.S -o head.s
+head.o: head.S $(TOPDIR)/include/linux/tasks.h
+	$(CC) -traditional -c head.S
 
-piggy.o:	$(SYSTEM) xtract piggyback
-		./xtract $(SYSTEM) | gzip -9 | ./piggyback > piggy.o
+endif
 
-xtract: xtract.c
-	$(HOSTCC) $(CFLAGS) -o xtract xtract.c
 
-piggyback: piggyback.c
-	$(HOSTCC) $(CFLAGS) -o piggyback piggyback.c
+# You cannot compress a file and have the kernel uncompress it, it must
+# be stdin
+piggy.o:	$(SYSTEM)
+	tmppiggy=/tmp/$$$$piggy; \
+	rm -f $$tmppiggy $$tmppiggy.gz $$tmppiggy.lnk; \
+	if hash $(ENCAPS) 2> /dev/null; then \
+	  $(OBJDUMP) $(OBJDUMP_FLAGS) -o $(IMAGE_OFFSET) $(SYSTEM) > $$tmppiggy; \
+	else \
+	  $(OBJCOPY) $(SYSTEM) $$tmppiggy; \
+	fi; \
+	gzip -f -9 < $$tmppiggy > $$tmppiggy.gz; \
+	if hash $(ENCAPS) 2> /dev/null; then \
+	  $(ENCAPS) $(TARGET) piggy.o $$tmppiggy.gz $(INPUT_DATA) $(INPUT_LEN); \
+	else \
+	  echo "SECTIONS { .data : { input_len = .; LONG(input_data_end - input_data) input_data = .; *(.data) input_data_end = .; }}" > $$tmppiggy.lnk; \
+	  $(LD) -m elf_i386 -r -o piggy.o -b binary $$tmppiggy.gz -b elf32-i386 -T $$tmppiggy.lnk; \
+	fi; \
+	rm -f $$tmppiggy $$tmppiggy.gz $$tmppiggy.lnk
 
 clean:
-	rm -f xtract piggyback vmlinux
+	rm -f xtract piggyback vmlinux bvmlinux
diff --git a/arch/i386/boot/compressed/Makefile.debug b/arch/i386/boot/compressed/Makefile.debug
new file mode 100644
index 000000000..83714b505
--- /dev/null
+++ b/arch/i386/boot/compressed/Makefile.debug
@@ -0,0 +1,21 @@
+#
+# linux/arch/i386/boot/compressed/Makefile
+#
+# create a compressed vmlinux image from the original vmlinux
+#
+
+OBJECTS = misc.o 
+
+CFLAGS = -g -O2 -DSTDC_HEADERS -DSTANDALONE_DEBUG -Wall
+
+test-gzip: piggy.o $(OBJECTS)
+	$(CC) -g -o test-gzip $(OBJECTS) piggy.o
+
+clean: 
+	$(RM) inflate.o misc.o test-gzip
+
+inflate.o: inflate.c gzip.h
+
+misc.o: misc.c gzip.h
+
+
diff --git a/arch/i386/boot/compressed/crypt.h b/arch/i386/boot/compressed/crypt.h
deleted file mode 100644
index 2a4c203ca..000000000
--- a/arch/i386/boot/compressed/crypt.h
+++ /dev/null
@@ -1,12 +0,0 @@
-/* crypt.h (dummy version) -- do not perform encryption
- * Hardly worth copyrighting :-)
- */
-
-#ifdef CRYPT
-#  undef CRYPT      /* dummy version */
-#endif
-
-#define RAND_HEAD_LEN  12  /* length of encryption random header */
-
-#define zencode
-#define zdecode
diff --git a/arch/i386/boot/compressed/gzip.h b/arch/i386/boot/compressed/gzip.h
deleted file mode 100644
index 2f738b945..000000000
--- a/arch/i386/boot/compressed/gzip.h
+++ /dev/null
@@ -1,284 +0,0 @@
-/* gzip.h -- common declarations for all gzip modules
- * Copyright (C) 1992-1993 Jean-loup Gailly.
- * This is free software; you can redistribute it and/or modify it under the
- * terms of the GNU General Public License, see the file COPYING.
- */
-
-#if defined(__STDC__) || defined(PROTO)
-#  define OF(args)  args
-#else
-#  define OF(args)  ()
-#endif
-
-#ifdef __STDC__
-   typedef void *voidp;
-#else
-   typedef char *voidp;
-#endif
-
-/* I don't like nested includes, but the string functions are used too often */
-#if defined(HAVE_STRING_H) || defined(STDC_HEADERS)
-#  include <string.h>
-#  define memzero(s, n)     memset ((s), 0, (n))
-#else
-#  include <strings.h>
-#  define strchr            index 
-#  define strrchr           rindex
-#  define memcpy(d, s, n)   bcopy((s), (d), (n)) 
-#  define memcmp(s1, s2, n) bcmp((s1), (s2), (n)) 
-#  define memzero(s, n)     bzero((s), (n))
-#endif
-
-#if !defined(STDC_HEADERS) && defined(HAVE_MEMORY_H)
-#  include <memory.h>
-#endif
-
-#ifndef RETSIGTYPE
-#  define RETSIGTYPE void
-#endif
-
-#define local static
-
-typedef unsigned char  uch;
-typedef unsigned short ush;
-typedef unsigned long  ulg;
-
-/* Return codes from gzip */
-#define OK      0
-#define ERROR   1
-#define WARNING 2
-
-/* Compression methods (see algorithm.doc) */
-#define STORED     0
-#define COMPRESSED 1
-#define PACKED     2
-/* methods 3 to 7 reserved */
-#define DEFLATED   8
-extern int method;         /* compression method */
-
-/* To save memory for 16 bit systems, some arrays are overlayed between
- * the various modules:
- * deflate:  prev+head   window      d_buf  l_buf  outbuf
- * unlzw:    tab_prefix  tab_suffix  stack  inbuf  outbuf
- * inflate:              window             inbuf
- * unpack:               window             inbuf
- * For compression, input is done in window[]. For decompression, output
- * is done in window except for unlzw.
- */
-
-#ifndef	INBUFSIZ
-#  define INBUFSIZ  0x8000  /* input buffer size */
-#endif
-#define INBUF_EXTRA  64     /* required by unlzw() */
-
-#ifndef	OUTBUFSIZ
-#  define OUTBUFSIZ  16384  /* output buffer size */
-#endif
-#define OUTBUF_EXTRA 2048   /* required by unlzw() */
-
-#define DIST_BUFSIZE 0x8000 /* buffer for distances, see trees.c */
-
-#ifdef DYN_ALLOC
-#  define EXTERN(type, array)  extern type * near array
-#  define DECLARE(type, array, size)  type * near array
-#  define ALLOC(type, array, size) { \
-      array = (type*)fcalloc((unsigned)(((size)+1L)/2), 2*sizeof(type)); \
-      if (array == NULL) error("insufficient memory"); \
-   }
-#  define FREE(array) {if (array != NULL) fcfree(array), array=NULL;}
-#else
-#  define EXTERN(type, array)  extern type array[]
-#  define DECLARE(type, array, size)  type array[size]
-#  define ALLOC(type, array, size)
-#  define FREE(array)
-#endif
-
-EXTERN(uch, inbuf);          /* input buffer */
-EXTERN(uch, outbuf);         /* output buffer */
-EXTERN(ush, d_buf);          /* buffer for distances, see trees.c */
-EXTERN(uch, window);         /* Sliding window and suffix table (unlzw) */
-#define tab_suffix window
-#ifndef MAXSEG_64K
-#  define tab_prefix prev    /* hash link (see deflate.c) */
-#  define head (prev+WSIZE)  /* hash head (see deflate.c) */
-   EXTERN(ush, tab_prefix);  /* prefix code (see unlzw.c) */
-#else
-#  define tab_prefix0 prev
-#  define head tab_prefix1
-   EXTERN(ush, tab_prefix0); /* prefix for even codes */
-   EXTERN(ush, tab_prefix1); /* prefix for odd  codes */
-#endif
-
-extern unsigned insize; /* valid bytes in inbuf */
-extern unsigned inptr;  /* index of next byte to be processed in inbuf */
-extern unsigned outcnt; /* bytes in output buffer */
-
-extern long bytes_in;   /* number of input bytes */
-extern long bytes_out;  /* number of output bytes */
-extern long overhead;   /* number of bytes in gzip header */
-
-#define isize bytes_in
-/* for compatibility with old zip sources (to be cleaned) */
-
-extern int  ifd;        /* input file descriptor */
-extern int  ofd;        /* output file descriptor */
-extern char ifname[];   /* input filename or "stdin" */
-extern char ofname[];   /* output filename or "stdout" */
-
-extern ulg time_stamp;  /* original time stamp (modification time) */
-extern long ifile_size; /* input file size, -1 for devices (debug only) */
-
-extern int exit_code;   /* program exit code */
-
-typedef int file_t;     /* Do not use stdio */
-#define NO_FILE  (-1)   /* in memory compression */
-
-
-#define	GZIP_MAGIC     "\037\213" /* Magic header for gzip files, 1F 8B */
-#define	OLD_GZIP_MAGIC "\037\236" /* Magic header for gzip 0.5 = freeze 1.x */
-#define	PKZIP_MAGIC  "PK\003\004" /* Magic header for pkzip files */
-#define	PACK_MAGIC     "\037\036" /* Magic header for packed files */
-
-/* gzip flag byte */
-#define ASCII_FLAG   0x01 /* bit 0 set: file probably ascii text */
-#define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
-#define EXTRA_FIELD  0x04 /* bit 2 set: extra field present */
-#define ORIG_NAME    0x08 /* bit 3 set: original file name present */
-#define COMMENT      0x10 /* bit 4 set: file comment present */
-#define ENCRYPTED    0x20 /* bit 5 set: file is encrypted */
-#define RESERVED     0xC0 /* bit 6,7:   reserved */
-
-/* internal file attribute */
-#define UNKNOWN (-1)
-#define BINARY  0
-#define ASCII   1
-
-#ifndef WSIZE
-#  define WSIZE 0x8000     /* window size--must be a power of two, and */
-#endif                     /*  at least 32K for zip's deflate method */
-
-#define MIN_MATCH  3
-#define MAX_MATCH  258
-/* The minimum and maximum match lengths */
-
-#define MIN_LOOKAHEAD (MAX_MATCH+MIN_MATCH+1)
-/* Minimum amount of lookahead, except at the end of the input file.
- * See deflate.c for comments about the MIN_MATCH+1.
- */
-
-#define MAX_DIST  (WSIZE-MIN_LOOKAHEAD)
-/* In order to simplify the code, particularly on 16 bit machines, match
- * distances are limited to MAX_DIST instead of WSIZE.
- */
-
-extern int decrypt;        /* flag to turn on decryption */
-extern int save_orig_name; /* set if original name must be saved */
-extern int verbose;        /* be verbose (-v) */
-extern int level;          /* compression level */
-extern int test;           /* check .z file integrity */
-extern int to_stdout;      /* output to stdout (-c) */
-
-#define get_byte()  (inptr < insize ? inbuf[inptr++] : fill_inbuf())
-
-/* put_byte is used for the compressed output, put_char for the
- * uncompressed output. However unlzw() uses window for its
- * suffix table instead of its output buffer, so it does not use put_char.
- * (to be cleaned up).
- */
-#define put_byte(c) {outbuf[outcnt++]=(uch)(c); if (outcnt==OUTBUFSIZ)\
-   flush_outbuf();}
-#define put_char(c) {window[outcnt++]=(uch)(c); if (outcnt==WSIZE)\
-   flush_window();}
-
-/* Output a 16 bit value, lsb first */
-#define put_short(w) \
-{ if (outcnt < OUTBUFSIZ-2) { \
-    outbuf[outcnt++] = (uch) ((w) & 0xff); \
-    outbuf[outcnt++] = (uch) ((ush)(w) >> 8); \
-  } else { \
-    put_byte((uch)((w) & 0xff)); \
-    put_byte((uch)((ush)(w) >> 8)); \
-  } \
-}
-
-/* Output a 32 bit value to the bit stream, lsb first */
-#define put_long(n) { \
-    put_short((n) & 0xffff); \
-    put_short(((ulg)(n)) >> 16); \
-}
-
-#define seekable()    0  /* force sequential output */
-#define translate_eol 0  /* no option -a yet */
-
-#define tolow(c)  (isupper(c) ? (c)-'A'+'a' : (c))    /* force to lower case */
-
-/* Macros for getting two-byte and four-byte header values */
-#define SH(p) ((ush)(uch)((p)[0]) | ((ush)(uch)((p)[1]) << 8))
-#define LG(p) ((ulg)(SH(p)) | ((ulg)(SH((p)+2)) << 16))
-
-/* Diagnostic functions */
-#ifdef DEBUG
-#  define Assert(cond,msg) {if(!(cond)) error(msg);}
-#  define Trace(x) fprintf x
-#  define Tracev(x) {if (verbose) fprintf x ;}
-#  define Tracevv(x) {if (verbose>1) fprintf x ;}
-#  define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
-#  define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
-#else
-#  define Assert(cond,msg)
-#  define Trace(x)
-#  define Tracev(x)
-#  define Tracevv(x)
-#  define Tracec(c,x)
-#  define Tracecv(c,x)
-#endif
-
-	/* in zip.c: */
-extern void zip    OF((int in, int out));
-extern int file_read  OF((char *buf,  unsigned size));
-
-	/* in unzip.c */
-extern void unzip        OF((int in, int out));
-extern int check_zipfile OF((int in));
-
-	/* in unpack.c */
-extern void unpack        OF((int in, int out));
-
-	/* in gzip.c */
-RETSIGTYPE abort_gzip   OF((void));
-
-        /* in deflate.c */
-void lm_init OF((int pack_level, ush *flags));
-ulg  deflate OF((void));
-
-        /* in trees.c */
-void ct_init     OF((ush *attr, int *method));
-int  ct_tally    OF((int dist, int lc));
-ulg  flush_block OF((char *buf, ulg stored_len, int eof));
-
-        /* in bits.c */
-void     bi_init    OF((file_t zipfile));
-void     send_bits  OF((int value, int length));
-unsigned bi_reverse OF((unsigned value, int length));
-void     bi_windup  OF((void));
-void     copy_block OF((char *buf, unsigned len, int header));
-extern   int (*read_buf) OF((char *buf, unsigned size));
-
-	/* in util.c: */
-extern ulg  updcrc        OF((uch *s, unsigned n));
-extern void clear_bufs    OF((void));
-extern int  fill_inbuf    OF((void));
-extern void flush_outbuf  OF((void));
-extern void flush_window  OF((void));
-extern char *strlwr       OF((char *s));
-extern char *basename     OF((char *fname));
-extern char *add_envopt   OF((int *argcp, char ***argvp, char *env));
-extern void error         OF((char *m));
-extern void warn          OF((char *a, char *b));
-extern void read_error    OF((void));
-extern void write_error   OF((void));
-extern void display_ratio OF((long num, long den));
-extern voidp xmalloc      OF((unsigned int size));
-
-	/* in inflate.c */
-extern int inflate OF((void));
diff --git a/arch/i386/boot/compressed/head.S b/arch/i386/boot/compressed/head.S
index 6035cea85..6f68a92d4 100644
--- a/arch/i386/boot/compressed/head.S
+++ b/arch/i386/boot/compressed/head.S
@@ -9,18 +9,31 @@
  *
  * NOTE!!! Startup happens at absolute address 0x00001000, which is also where
  * the page directory will exist. The startup code will be overwritten by
- * the page directory.
+ * the page directory. [According to comments etc elsewhere on a compressed
+ * kernel it will end up at 0x1000 + 1Mb I hope so as I assume this. - AC]
+ *
+ * In SMP mode we keep this page safe. Really we ought to shuffle things and
+ * put the trampoline here. - AC. An SMP trampoline enters with %cx holding
+ * the stack base.
  *
  * Page 0 is deliberately kept safe, since System Management Mode code in 
  * laptops may need to access the BIOS data stored there.  This is also
  * useful for future device drivers that either access the BIOS via VM86 
  * mode.
  */
+
+/*
+ * High loaded stuff by Hans Lermen & Werner Almesberger, Feb. 1996
+ */
 .text
 
 #define __ASSEMBLY__
+#include <linux/config.h>
+#include <linux/linkage.h>
 #include <asm/segment.h>
 
+	.globl startup_32
+	
 startup_32:
 	cld
 	cli
@@ -29,7 +42,25 @@ startup_32:
 	mov %ax,%es
 	mov %ax,%fs
 	mov %ax,%gs
-	lss _stack_start,%esp
+#ifdef __SMP__
+	orw %bx,%bx			# What state are we in BX=1 for SMP
+					# 0 for boot
+	jz  2f				# Initial boot
+
+/*
+ *	We are trampolining an SMP processor
+ */
+	mov %ax,%ss
+	xorl %eax,%eax			# Back to 0
+	mov  %cx,%ax			# SP low 16 bits
+	movl %eax,%esp
+	pushl 0				# Clear NT
+	popfl
+	ljmp $(KERNEL_CS), $0x100000    # Into C and sanity
+
+2:
+#endif
+	lss SYMBOL_NAME(stack_start),%esp
 	xorl %eax,%eax
 1:	incl %eax		# check that A20 really IS enabled
 	movl %eax,0x000000	# loop forever if it isn't
@@ -46,8 +77,8 @@ startup_32:
  * Clear BSS
  */
 	xorl %eax,%eax
-	movl $__edata,%edi
-	movl $__end,%ecx
+	movl $ SYMBOL_NAME(_edata),%edi
+	movl $ SYMBOL_NAME(_end),%ecx
 	subl %edi,%ecx
 	cld
 	rep
@@ -55,5 +86,55 @@ startup_32:
 /*
  * Do the decompression, and jump to the new kernel..
  */
-	call _decompress_kernel
+	subl $16,%esp	# place for structure on the stack
+	pushl %esp	# address of structure as first arg
+	call SYMBOL_NAME(decompress_kernel)
+	orl  %eax,%eax 
+	jnz  3f
+	xorl %ebx,%ebx
+	ljmp $(KERNEL_CS), $0x100000
+
+/*
+ * We come here, if we were loaded high.
+ * We need to move the move-in-place routine down to 0x1000
+ * and then start it with the buffer addresses in registers,
+ * which we got from the stack.
+ */
+3:
+	movl $move_routine_start,%esi
+	movl $0x1000,%edi
+	movl $move_routine_end,%ecx
+	subl %esi,%ecx
+	cld
+	rep
+	movsb
+
+	popl %esi	# discard the address
+	popl %esi	# low_buffer_start
+	popl %ecx	# lcount
+	popl %edx	# high_buffer_start
+	popl %eax	# hcount
+	movl $0x100000,%edi
+	cli		# make sure we don't get interrupted
+	ljmp $(KERNEL_CS), $0x1000 # and jump to the move routine
+
+/*
+ * Routine (template) for moving the decompressed kernel in place,
+ * if we were high loaded. This _must_ PIC-code !
+ */
+move_routine_start:
+	rep
+	movsb
+	movl %edx,%esi
+	movl %eax,%ecx	# NOTE: rep movsb won't move if %ecx == 0
+	rep
+	movsb
+	xorl %ebx,%ebx
+/*
+ * Well, the kernel relies on %esp pointing into low mem,
+ * with the decompressor loaded high this is no longer true,
+ * so we set esp here.
+ */
+	mov  $0x90000,%esp
 	ljmp $(KERNEL_CS), $0x100000
+move_routine_end:
diff --git a/arch/i386/boot/compressed/inflate.c b/arch/i386/boot/compressed/inflate.c
deleted file mode 100644
index 848fef6ae..000000000
--- a/arch/i386/boot/compressed/inflate.c
+++ /dev/null
@@ -1,810 +0,0 @@
-#define DEBG(x)
-#define DEBG1(x)
-/* inflate.c -- Not copyrighted 1992 by Mark Adler
-   version c10p1, 10 January 1993 */
-
-/* 
- * Adapted for booting Linux by Hannu Savolainen 1993
- * based on gzip-1.0.3 
- */
-
-#ifndef lint
-static char rcsid[] = "$Id: inflate.c,v 0.10 1993/02/04 13:21:06 jloup Exp $";
-#endif
-
-#include "gzip.h"
-#define slide window
-
-#if defined(STDC_HEADERS) || defined(HAVE_STDLIB_H)
-#  include <sys/types.h>
-#  include <stdlib.h>
-#endif
-
-struct huft {
-  uch e;                /* number of extra bits or operation */
-  uch b;                /* number of bits in this code or subcode */
-  union {
-    ush n;              /* literal, length base, or distance base */
-    struct huft *t;     /* pointer to next level of table */
-  } v;
-};
-
-
-/* Function prototypes */
-int huft_build OF((unsigned *, unsigned, unsigned, ush *, ush *,
-                   struct huft **, int *));
-int huft_free OF((struct huft *));
-int inflate_codes OF((struct huft *, struct huft *, int, int));
-int inflate_stored OF((void));
-int inflate_fixed OF((void));
-int inflate_dynamic OF((void));
-int inflate_block OF((int *));
-int inflate OF((void));
-
-
-#define wp outcnt
-#define flush_output(w) (wp=(w),flush_window())
-
-/* Tables for deflate from PKZIP's appnote.txt. */
-static unsigned border[] = {    /* Order of the bit length code lengths */
-        16, 17, 18, 0, 8, 7, 9, 6, 10, 5, 11, 4, 12, 3, 13, 2, 14, 1, 15};
-static ush cplens[] = {         /* Copy lengths for literal codes 257..285 */
-        3, 4, 5, 6, 7, 8, 9, 10, 11, 13, 15, 17, 19, 23, 27, 31,
-        35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
-        /* note: see note #13 above about the 258 in this list. */
-static ush cplext[] = {         /* Extra bits for literal codes 257..285 */
-        0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 2, 2, 2, 2,
-        3, 3, 3, 3, 4, 4, 4, 4, 5, 5, 5, 5, 0, 99, 99}; /* 99==invalid */
-static ush cpdist[] = {         /* Copy offsets for distance codes 0..29 */
-        1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
-        257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
-        8193, 12289, 16385, 24577};
-static ush cpdext[] = {         /* Extra bits for distance codes */
-        0, 0, 0, 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6,
-        7, 7, 8, 8, 9, 9, 10, 10, 11, 11,
-        12, 12, 13, 13};
-
-
-ulg bb;                         /* bit buffer */
-unsigned bk;                    /* bits in bit buffer */
-
-ush mask_bits[] = {
-    0x0000,
-    0x0001, 0x0003, 0x0007, 0x000f, 0x001f, 0x003f, 0x007f, 0x00ff,
-    0x01ff, 0x03ff, 0x07ff, 0x0fff, 0x1fff, 0x3fff, 0x7fff, 0xffff
-};
-
-#ifdef CRYPT
-  uch cc;
-#  define NEXTBYTE() \
-     (decrypt ? (cc = get_byte(), zdecode(cc), cc) : get_byte())
-#else
-#  define NEXTBYTE()  (uch)get_byte()
-#endif
-#define NEEDBITS(n) {while(k<(n)){b|=((ulg)NEXTBYTE())<<k;k+=8;}}
-#define DUMPBITS(n) {b>>=(n);k-=(n);}
-
-int lbits = 9;          /* bits in base literal/length lookup table */
-int dbits = 6;          /* bits in base distance lookup table */
-
-
-/* If BMAX needs to be larger than 16, then h and x[] should be ulg. */
-#define BMAX 16         /* maximum bit length of any code (16 for explode) */
-#define N_MAX 288       /* maximum number of codes in any set */
-
-
-unsigned hufts;         /* track memory usage */
-
-
-int huft_build(b, n, s, d, e, t, m)
-unsigned *b;            /* code lengths in bits (all assumed <= BMAX) */
-unsigned n;             /* number of codes (assumed <= N_MAX) */
-unsigned s;             /* number of simple-valued codes (0..s-1) */
-ush *d;                 /* list of base values for non-simple codes */
-ush *e;                 /* list of extra bits for non-simple codes */
-struct huft **t;        /* result: starting table */
-int *m;                 /* maximum lookup bits, returns actual */
-/* Given a list of code lengths and a maximum table size, make a set of
-   tables to decode that set of codes.  Return zero on success, one if
-   the given code set is incomplete (the tables are still built in this
-   case), two if the input is invalid (all zero length codes or an
-   oversubscribed set of lengths), and three if not enough memory. */
-{
-  unsigned a;                   /* counter for codes of length k */
-  unsigned c[BMAX+1];           /* bit length count table */
-  unsigned f;                   /* i repeats in table every f entries */
-  int g;                        /* maximum code length */
-  int h;                        /* table level */
-  register unsigned i;          /* counter, current code */
-  register unsigned j;          /* counter */
-  register int k;               /* number of bits in current code */
-  int l;                        /* bits per table (returned in m) */
-  register unsigned *p;         /* pointer into c[], b[], or v[] */
-  register struct huft *q;      /* points to current table */
-  struct huft r;                /* table entry for structure assignment */
-  struct huft *u[BMAX];         /* table stack */
-  unsigned v[N_MAX];            /* values in order of bit length */
-  register int w;               /* bits before this table == (l * h) */
-  unsigned x[BMAX+1];           /* bit offsets, then code stack */
-  unsigned *xp;                 /* pointer into x */
-  int y;                        /* number of dummy codes added */
-  unsigned z;                   /* number of entries in current table */
-
-DEBG("huft1 ");
-
-  /* Generate counts for each bit length */
-  memzero(c, sizeof(c));
-  p = b;  i = n;
-  do {
-    c[*p++]++;                  /* assume all entries <= BMAX */
-  } while (--i);
-  if (c[0] == n)                /* null input--all zero length codes */
-  {
-    *t = (struct huft *)NULL;
-    *m = 0;
-    return 0;
-  }
-
-DEBG("huft2 ");
-
-  /* Find minimum and maximum length, bound *m by those */
-  l = *m;
-  for (j = 1; j <= BMAX; j++)
-    if (c[j])
-      break;
-  k = j;                        /* minimum code length */
-  if ((unsigned)l < j)
-    l = j;
-  for (i = BMAX; i; i--)
-    if (c[i])
-      break;
-  g = i;                        /* maximum code length */
-  if ((unsigned)l > i)
-    l = i;
-  *m = l;
-
-DEBG("huft3 ");
-
-  /* Adjust last length count to fill out codes, if needed */
-  for (y = 1 << j; j < i; j++, y <<= 1)
-    if ((y -= c[j]) < 0)
-      return 2;                 /* bad input: more codes than bits */
-  if ((y -= c[i]) < 0)
-    return 2;
-  c[i] += y;
-
-DEBG("huft4 ");
-
-  /* Generate starting offsets into the value table for each length */
-  x[1] = j = 0;
-  p = c + 1;  xp = x + 2;
-  while (--i) {                 /* note that i == g from above */
-    *xp++ = (j += *p++);
-  }
-
-DEBG("huft5 ");
-
-  /* Make a table of values in order of bit lengths */
-  p = b;  i = 0;
-  do {
-    if ((j = *p++) != 0)
-      v[x[j]++] = i;
-  } while (++i < n);
-
-DEBG("h6 ");
-
-  /* Generate the Huffman codes and for each, make the table entries */
-  x[0] = i = 0;                 /* first Huffman code is zero */
-  p = v;                        /* grab values in bit order */
-  h = -1;                       /* no tables yet--level -1 */
-  w = -l;                       /* bits decoded == (l * h) */
-  u[0] = (struct huft *)NULL;   /* just to keep compilers happy */
-  q = (struct huft *)NULL;      /* ditto */
-  z = 0;                        /* ditto */
-DEBG("h6a ");
-
-  /* go through the bit lengths (k already is bits in shortest code) */
-  for (; k <= g; k++)
-  {
-DEBG("h6b ");
-    a = c[k];
-    while (a--)
-    {
-DEBG("h6b1 ");
-      /* here i is the Huffman code of length k bits for value *p */
-      /* make tables up to required level */
-      while (k > w + l)
-      {
-DEBG1("1 ");
-        h++;
-        w += l;                 /* previous table always l bits */
-
-        /* compute minimum size table less than or equal to l bits */
-        z = (z = g - w) > (unsigned)l ? l : z;  /* upper limit on table size */
-        if ((f = 1 << (j = k - w)) > a + 1)     /* try a k-w bit table */
-        {                       /* too few codes for k-w bit table */
-DEBG1("2 ");
-          f -= a + 1;           /* deduct codes from patterns left */
-          xp = c + k;
-          while (++j < z)       /* try smaller tables up to z bits */
-          {
-            if ((f <<= 1) <= *++xp)
-              break;            /* enough codes to use up j bits */
-            f -= *xp;           /* else deduct codes from patterns */
-          }
-        }
-DEBG1("3 ");
-        z = 1 << j;             /* table entries for j-bit table */
-
-        /* allocate and link in new table */
-        q = (struct huft *)malloc((z + 1)*sizeof(struct huft));
-DEBG1("4 ");
-        hufts += z + 1;         /* track memory usage */
-        *t = q + 1;             /* link to list for huft_free() */
-        *(t = &(q->v.t)) = (struct huft *)NULL;
-        u[h] = ++q;             /* table starts after link */
-
-DEBG1("5 ");
-        /* connect to last table, if there is one */
-        if (h)
-        {
-          x[h] = i;             /* save pattern for backing up */
-          r.b = (uch)l;         /* bits to dump before this table */
-          r.e = (uch)(16 + j);  /* bits in this table */
-          r.v.t = q;            /* pointer to this table */
-          j = i >> (w - l);     /* (get around Turbo C bug) */
-          u[h-1][j] = r;        /* connect to last table */
-        }
-DEBG1("6 ");
-      }
-DEBG("h6c ");
-
-      /* set up table entry in r */
-      r.b = (uch)(k - w);
-      if (p >= v + n)
-        r.e = 99;               /* out of values--invalid code */
-      else if (*p < s)
-      {
-        r.e = (uch)(*p < 256 ? 16 : 15);    /* 256 is end-of-block code */
-        r.v.n = *p++;           /* simple code is just the value */
-      }
-      else
-      {
-        r.e = (uch)e[*p - s];   /* non-simple--look up in lists */
-        r.v.n = d[*p++ - s];
-      }
-DEBG("h6d ");
-
-      /* fill code-like entries with r */
-      f = 1 << (k - w);
-      for (j = i >> w; j < z; j += f)
-        q[j] = r;
-
-      /* backwards increment the k-bit code i */
-      for (j = 1 << (k - 1); i & j; j >>= 1)
-        i ^= j;
-      i ^= j;
-
-      /* backup over finished tables */
-      while ((i & ((1 << w) - 1)) != x[h])
-      {
-        h--;                    /* don't need to update q */
-        w -= l;
-      }
-DEBG("h6e ");
-    }
-DEBG("h6f ");
-  }
-
-DEBG("huft7 ");
-
-  /* Return true (1) if we were given an incomplete table */
-  return y != 0 && g != 1;
-}
-
-
-
-int huft_free(t)
-struct huft *t;         /* table to free */
-/* Free the malloc'ed tables built by huft_build(), which makes a linked
-   list of the tables it made, with the links in a dummy first entry of
-   each table. */
-{
-  register struct huft *p, *q;
-
-
-  /* Go through linked list, freeing from the malloced (t[-1]) address. */
-  p = t;
-  while (p != (struct huft *)NULL)
-  {
-    q = (--p)->v.t;
-    free(p);
-    p = q;
-  } 
-  return 0;
-}
-
-
-int inflate_codes(tl, td, bl, bd)
-struct huft *tl, *td;   /* literal/length and distance decoder tables */
-int bl, bd;             /* number of bits decoded by tl[] and td[] */
-/* inflate (decompress) the codes in a deflated (compressed) block.
-   Return an error code or zero if it all goes ok. */
-{
-  register unsigned e;  /* table entry flag/number of extra bits */
-  unsigned n, d;        /* length and index for copy */
-  unsigned w;           /* current window position */
-  struct huft *t;       /* pointer to table entry */
-  unsigned ml, md;      /* masks for bl and bd bits */
-  register ulg b;       /* bit buffer */
-  register unsigned k;  /* number of bits in bit buffer */
-
-
-  /* make local copies of globals */
-  b = bb;                       /* initialize bit buffer */
-  k = bk;
-  w = wp;                       /* initialize window position */
-
-  /* inflate the coded data */
-  ml = mask_bits[bl];           /* precompute masks for speed */
-  md = mask_bits[bd];
-  for (;;)                      /* do until end of block */
-  {
-    NEEDBITS((unsigned)bl)
-    if ((e = (t = tl + ((unsigned)b & ml))->e) > 16)
-      do {
-        if (e == 99)
-          return 1;
-        DUMPBITS(t->b)
-        e -= 16;
-        NEEDBITS(e)
-      } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
-    DUMPBITS(t->b)
-    if (e == 16)                /* then it's a literal */
-    {
-      slide[w++] = (uch)t->v.n;
-      if (w == WSIZE)
-      {
-        flush_output(w);
-        w = 0;
-      }
-    }
-    else                        /* it's an EOB or a length */
-    {
-      /* exit if end of block */
-      if (e == 15)
-        break;
-
-      /* get length of block to copy */
-      NEEDBITS(e)
-      n = t->v.n + ((unsigned)b & mask_bits[e]);
-      DUMPBITS(e);
-
-      /* decode distance of block to copy */
-      NEEDBITS((unsigned)bd)
-      if ((e = (t = td + ((unsigned)b & md))->e) > 16)
-        do {
-          if (e == 99)
-            return 1;
-          DUMPBITS(t->b)
-          e -= 16;
-          NEEDBITS(e)
-        } while ((e = (t = t->v.t + ((unsigned)b & mask_bits[e]))->e) > 16);
-      DUMPBITS(t->b)
-      NEEDBITS(e)
-      d = w - t->v.n - ((unsigned)b & mask_bits[e]);
-      DUMPBITS(e)
-
-      /* do the copy */
-      do {
-        n -= (e = (e = WSIZE - ((d &= WSIZE-1) > w ? d : w)) > n ? n : e);
-#if !defined(NOMEMCPY) && !defined(DEBUG)
-        if (w - d >= e)         /* (this test assumes unsigned comparison) */
-        {
-          memcpy(slide + w, slide + d, e);
-          w += e;
-          d += e;
-        }
-        else                      /* do it slow to avoid memcpy() overlap */
-#endif /* !NOMEMCPY */
-          do {
-            slide[w++] = slide[d++];
-          } while (--e);
-        if (w == WSIZE)
-        {
-          flush_output(w);
-          w = 0;
-        }
-      } while (n);
-    }
-  }
-
-
-  /* restore the globals from the locals */
-  wp = w;                       /* restore global window pointer */
-  bb = b;                       /* restore global bit buffer */
-  bk = k;
-
-  /* done */
-  return 0;
-}
-
-
-
-int inflate_stored()
-/* "decompress" an inflated type 0 (stored) block. */
-{
-  unsigned n;           /* number of bytes in block */
-  unsigned w;           /* current window position */
-  register ulg b;       /* bit buffer */
-  register unsigned k;  /* number of bits in bit buffer */
-
-DEBG("<stor");
-
-  /* make local copies of globals */
-  b = bb;                       /* initialize bit buffer */
-  k = bk;
-  w = wp;                       /* initialize window position */
-
-
-  /* go to byte boundary */
-  n = k & 7;
-  DUMPBITS(n);
-
-
-  /* get the length and its complement */
-  NEEDBITS(16)
-  n = ((unsigned)b & 0xffff);
-  DUMPBITS(16)
-  NEEDBITS(16)
-  if (n != (unsigned)((~b) & 0xffff))
-    return 1;                   /* error in compressed data */
-  DUMPBITS(16)
-
-
-  /* read and output the compressed data */
-  while (n--)
-  {
-    NEEDBITS(8)
-    slide[w++] = (uch)b;
-    if (w == WSIZE)
-    {
-      flush_output(w);
-      w = 0;
-    }
-    DUMPBITS(8)
-  }
-
-
-  /* restore the globals from the locals */
-  wp = w;                       /* restore global window pointer */
-  bb = b;                       /* restore global bit buffer */
-  bk = k;
-
-  DEBG(">");
-  return 0;
-}
-
-
-
-int inflate_fixed()
-/* decompress an inflated type 1 (fixed Huffman codes) block.  We should
-   either replace this with a custom decoder, or at least precompute the
-   Huffman tables. */
-{
-  int i;                /* temporary variable */
-  struct huft *tl;      /* literal/length code table */
-  struct huft *td;      /* distance code table */
-  int bl;               /* lookup bits for tl */
-  int bd;               /* lookup bits for td */
-  unsigned l[288];      /* length list for huft_build */
-
-DEBG("<fix");
-
-  /* set up literal table */
-  for (i = 0; i < 144; i++)
-    l[i] = 8;
-  for (; i < 256; i++)
-    l[i] = 9;
-  for (; i < 280; i++)
-    l[i] = 7;
-  for (; i < 288; i++)          /* make a complete, but wrong code set */
-    l[i] = 8;
-  bl = 7;
-  if ((i = huft_build(l, 288, 257, cplens, cplext, &tl, &bl)) != 0)
-    return i;
-
-
-  /* set up distance table */
-  for (i = 0; i < 30; i++)      /* make an incomplete code set */
-    l[i] = 5;
-  bd = 5;
-  if ((i = huft_build(l, 30, 0, cpdist, cpdext, &td, &bd)) > 1)
-  {
-    huft_free(tl);
-
-    DEBG(">");
-    return i;
-  }
-
-
-  /* decompress until an end-of-block code */
-  if (inflate_codes(tl, td, bl, bd))
-    return 1;
-
-
-  /* free the decoding tables, return */
-  huft_free(tl);
-  huft_free(td);
-  return 0;
-}
-
-
-
-int inflate_dynamic()
-/* decompress an inflated type 2 (dynamic Huffman codes) block. */
-{
-  int i;                /* temporary variables */
-  unsigned j;
-  unsigned l;           /* last length */
-  unsigned m;           /* mask for bit lengths table */
-  unsigned n;           /* number of lengths to get */
-  struct huft *tl;      /* literal/length code table */
-  struct huft *td;      /* distance code table */
-  int bl;               /* lookup bits for tl */
-  int bd;               /* lookup bits for td */
-  unsigned nb;          /* number of bit length codes */
-  unsigned nl;          /* number of literal/length codes */
-  unsigned nd;          /* number of distance codes */
-#ifdef PKZIP_BUG_WORKAROUND
-  unsigned ll[288+32];  /* literal/length and distance code lengths */
-#else
-  unsigned ll[286+30];  /* literal/length and distance code lengths */
-#endif
-  register ulg b;       /* bit buffer */
-  register unsigned k;  /* number of bits in bit buffer */
-
-DEBG("<dyn");
-
-  /* make local bit buffer */
-  b = bb;
-  k = bk;
-
-
-  /* read in table lengths */
-  NEEDBITS(5)
-  nl = 257 + ((unsigned)b & 0x1f);      /* number of literal/length codes */
-  DUMPBITS(5)
-  NEEDBITS(5)
-  nd = 1 + ((unsigned)b & 0x1f);        /* number of distance codes */
-  DUMPBITS(5)
-  NEEDBITS(4)
-  nb = 4 + ((unsigned)b & 0xf);         /* number of bit length codes */
-  DUMPBITS(4)
-#ifdef PKZIP_BUG_WORKAROUND
-  if (nl > 288 || nd > 32)
-#else
-  if (nl > 286 || nd > 30)
-#endif
-    return 1;                   /* bad lengths */
-
-DEBG("dyn1 ");
-
-  /* read in bit-length-code lengths */
-  for (j = 0; j < nb; j++)
-  {
-    NEEDBITS(3)
-    ll[border[j]] = (unsigned)b & 7;
-    DUMPBITS(3)
-  }
-  for (; j < 19; j++)
-    ll[border[j]] = 0;
-
-DEBG("dyn2 ");
-
-  /* build decoding table for trees--single level, 7 bit lookup */
-  bl = 7;
-  if ((i = huft_build(ll, 19, 19, NULL, NULL, &tl, &bl)) != 0)
-  {
-    if (i == 1)
-      huft_free(tl);
-    return i;                   /* incomplete code set */
-  }
-
-DEBG("dyn3 ");
-
-  /* read in literal and distance code lengths */
-  n = nl + nd;
-  m = mask_bits[bl];
-  i = l = 0;
-  while ((unsigned)i < n)
-  {
-    NEEDBITS((unsigned)bl)
-    j = (td = tl + ((unsigned)b & m))->b;
-    DUMPBITS(j)
-    j = td->v.n;
-    if (j < 16)                 /* length of code in bits (0..15) */
-      ll[i++] = l = j;          /* save last length in l */
-    else if (j == 16)           /* repeat last length 3 to 6 times */
-    {
-      NEEDBITS(2)
-      j = 3 + ((unsigned)b & 3);
-      DUMPBITS(2)
-      if ((unsigned)i + j > n)
-        return 1;
-      while (j--)
-        ll[i++] = l;
-    }
-    else if (j == 17)           /* 3 to 10 zero length codes */
-    {
-      NEEDBITS(3)
-      j = 3 + ((unsigned)b & 7);
-      DUMPBITS(3)
-      if ((unsigned)i + j > n)
-        return 1;
-      while (j--)
-        ll[i++] = 0;
-      l = 0;
-    }
-    else                        /* j == 18: 11 to 138 zero length codes */
-    {
-      NEEDBITS(7)
-      j = 11 + ((unsigned)b & 0x7f);
-      DUMPBITS(7)
-      if ((unsigned)i + j > n)
-        return 1;
-      while (j--)
-        ll[i++] = 0;
-      l = 0;
-    }
-  }
-
-DEBG("dyn4 ");
-
-  /* free decoding table for trees */
-  huft_free(tl);
-
-DEBG("dyn5 ");
-
-  /* restore the global bit buffer */
-  bb = b;
-  bk = k;
-
-DEBG("dyn5a ");
-
-  /* build the decoding tables for literal/length and distance codes */
-  bl = lbits;
-  if ((i = huft_build(ll, nl, 257, cplens, cplext, &tl, &bl)) != 0)
-  {
-DEBG("dyn5b ");
-    if (i == 1) {
-      error(" incomplete literal tree\n");
-      huft_free(tl);
-    }
-    return i;                   /* incomplete code set */
-  }
-DEBG("dyn5c ");
-  bd = dbits;
-  if ((i = huft_build(ll + nl, nd, 0, cpdist, cpdext, &td, &bd)) != 0)
-  {
-DEBG("dyn5d ");
-    if (i == 1) {
-      error(" incomplete distance tree\n");
-#ifdef PKZIP_BUG_WORKAROUND
-      i = 0;
-    }
-#else
-      huft_free(td);
-    }
-    huft_free(tl);
-    return i;                   /* incomplete code set */
-#endif
-  }
-
-DEBG("dyn6 ");
-
-  /* decompress until an end-of-block code */
-  if (inflate_codes(tl, td, bl, bd))
-    return 1;
-
-DEBG("dyn7 ");
-
-  /* free the decoding tables, return */
-  huft_free(tl);
-  huft_free(td);
-
-  DEBG(">");
-  return 0;
-}
-
-
-
-int inflate_block(e)
-int *e;                 /* last block flag */
-/* decompress an inflated block */
-{
-  unsigned t;           /* block type */
-  register ulg b;       /* bit buffer */
-  register unsigned k;  /* number of bits in bit buffer */
-
-  DEBG("<blk");
-
-  /* make local bit buffer */
-  b = bb;
-  k = bk;
-
-
-  /* read in last block bit */
-  NEEDBITS(1)
-  *e = (int)b & 1;
-  DUMPBITS(1)
-
-
-  /* read in block type */
-  NEEDBITS(2)
-  t = (unsigned)b & 3;
-  DUMPBITS(2)
-
-
-  /* restore the global bit buffer */
-  bb = b;
-  bk = k;
-
-  /* inflate that block type */
-  if (t == 2)
-    return inflate_dynamic();
-  if (t == 0)
-    return inflate_stored();
-  if (t == 1)
-    return inflate_fixed();
-
-  DEBG(">");
-
-  /* bad block type */
-  return 2;
-}
-
-
-
-int inflate()
-/* decompress an inflated entry */
-{
-  int e;                /* last block flag */
-  int r;                /* result code */
-  unsigned h;           /* maximum struct huft's malloc'ed */
-
-
-  /* initialize window, bit buffer */
-  wp = 0;
-  bk = 0;
-  bb = 0;
-
-
-  /* decompress until the last block */
-  h = 0;
-  do {
-    hufts = 0;
-    if ((r = inflate_block(&e)) != 0)
-      return r;
-    if (hufts > h)
-      h = hufts;
-  } while (!e);
-
-  /* Undo too much lookahead. The next read will be byte aligned so we
-   * can discard unused bits in the last meaningful byte.
-   */
-  while (bk >= 8) {
-    bk -= 8;
-    inptr--;
-  }
-
-  /* flush out slide */
-  flush_output(wp);
-
-
-  /* return success */
-#ifdef DEBUG
-  fprintf(stderr, "<%u> ", h);
-#endif /* DEBUG */
-  return 0;
-}
diff --git a/arch/i386/boot/compressed/lzw.h b/arch/i386/boot/compressed/lzw.h
deleted file mode 100644
index 4e640f5a2..000000000
--- a/arch/i386/boot/compressed/lzw.h
+++ /dev/null
@@ -1,42 +0,0 @@
-/* lzw.h -- define the lzw functions.
- * Copyright (C) 1992-1993 Jean-loup Gailly.
- * This is free software; you can redistribute it and/or modify it under the
- * terms of the GNU General Public License, see the file COPYING.
- */
-
-#if !defined(OF) && defined(lint)
-#  include "gzip.h"
-#endif
-
-#ifndef BITS
-#  define BITS 16
-#endif
-#define INIT_BITS 9              /* Initial number of bits per code */
-
-#define	LZW_MAGIC  "\037\235"   /* Magic header for lzw files, 1F 9D */
-
-#define BIT_MASK    0x1f /* Mask for 'number of compression bits' */
-/* Mask 0x20 is reserved to mean a fourth header byte, and 0x40 is free.
- * It's a pity that old uncompress does not check bit 0x20. That makes
- * extension of the format actually undesirable because old compress
- * would just crash on the new format instead of giving a meaningful
- * error message. It does check the number of bits, but it's more
- * helpful to say "unsupported format, get a new version" than
- * "can only handle 16 bits".
- */
-
-#define BLOCK_MODE  0x80
-/* Block compression: if table is full and compression rate is dropping,
- * clear the dictionary.
- */
-
-#define LZW_RESERVED 0x60 /* reserved bits */
-
-#define	CLEAR  256       /* flush the dictionary */
-#define FIRST  (CLEAR+1) /* first free entry */
-
-extern int maxbits;      /* max bits per code for LZW */
-extern int block_mode;   /* block compress mode -C compatible with 2.0 */
-
-extern void lzw    OF((int in, int out));
-extern void unlzw  OF((int in, int out));
diff --git a/arch/i386/boot/compressed/misc.c b/arch/i386/boot/compressed/misc.c
index 2623925c7..296517fee 100644
--- a/arch/i386/boot/compressed/misc.c
+++ b/arch/i386/boot/compressed/misc.c
@@ -5,15 +5,75 @@
  * adapted for Linux.
  *
  * malloc by Hannu Savolainen 1993 and Matthias Urlichs 1994
- * puts by Nick Holloway 1993
+ * puts by Nick Holloway 1993, better puts by Martin Mares 1995
+ * High loaded stuff by Hans Lermen & Werner Almesberger, Feb. 1996
  */
 
-#include "gzip.h"
-#include "lzw.h"
+#include <string.h>
 
 #include <asm/segment.h>
+#include <asm/io.h>
 
 /*
+ * gzip declarations
+ */
+
+#define OF(args)  args
+#define STATIC static
+
+#undef memset
+#undef memcpy
+#define memzero(s, n)     memset ((s), 0, (n))
+
+typedef unsigned char  uch;
+typedef unsigned short ush;
+typedef unsigned long  ulg;
+
+#define WSIZE 0x8000		/* Window size must be at least 32k, */
+				/* and a power of two */
+
+static uch *inbuf;	     /* input buffer */
+static uch window[WSIZE];    /* Sliding window buffer */
+
+static unsigned insize = 0;  /* valid bytes in inbuf */
+static unsigned inptr = 0;   /* index of next byte to be processed in inbuf */
+static unsigned outcnt = 0;  /* bytes in output buffer */
+
+/* gzip flag byte */
+#define ASCII_FLAG   0x01 /* bit 0 set: file probably ascii text */
+#define CONTINUATION 0x02 /* bit 1 set: continuation of multi-part gzip file */
+#define EXTRA_FIELD  0x04 /* bit 2 set: extra field present */
+#define ORIG_NAME    0x08 /* bit 3 set: original file name present */
+#define COMMENT      0x10 /* bit 4 set: file comment present */
+#define ENCRYPTED    0x20 /* bit 5 set: file is encrypted */
+#define RESERVED     0xC0 /* bit 6,7:   reserved */
+
+#define get_byte()  (inptr < insize ? inbuf[inptr++] : fill_inbuf())
+		
+/* Diagnostic functions */
+#ifdef DEBUG
+#  define Assert(cond,msg) {if(!(cond)) error(msg);}
+#  define Trace(x) fprintf x
+#  define Tracev(x) {if (verbose) fprintf x ;}
+#  define Tracevv(x) {if (verbose>1) fprintf x ;}
+#  define Tracec(c,x) {if (verbose && (c)) fprintf x ;}
+#  define Tracecv(c,x) {if (verbose>1 && (c)) fprintf x ;}
+#else
+#  define Assert(cond,msg)
+#  define Trace(x)
+#  define Tracev(x)
+#  define Tracevv(x)
+#  define Tracec(c,x)
+#  define Tracecv(c,x)
+#endif
+
+static int  fill_inbuf(void);
+static void flush_window(void);
+static void error(char *m);
+static void gzip_mark(void **);
+static void gzip_release(void **);
+  
+/*
  * These are set up by the setup-routine at boot-time:
  */
 
@@ -24,10 +84,11 @@ struct screen_info {
 	unsigned short orig_video_page;
 	unsigned char  orig_video_mode;
 	unsigned char  orig_video_cols;
-	unsigned short orig_video_ega_ax;
+	unsigned short unused2;
 	unsigned short orig_video_ega_bx;
-	unsigned short orig_video_ega_cx;
+	unsigned short unused3;
 	unsigned char  orig_video_lines;
+	unsigned char  orig_video_isVGA;
 };
 
 /*
@@ -40,82 +101,73 @@ struct screen_info {
 #define ORIG_ROOT_DEV (*(unsigned short *)0x901FC)
 #define AUX_DEVICE_INFO (*(unsigned char *)0x901FF)
 
-#define EOF -1
-
-DECLARE(uch, inbuf, INBUFSIZ);
-DECLARE(uch, outbuf, OUTBUFSIZ+OUTBUF_EXTRA);
-DECLARE(uch, window, WSIZE);
-
-unsigned outcnt;
-unsigned insize;
-unsigned inptr;
-
 extern char input_data[];
 extern int input_len;
 
-int input_ptr;
-
-int method, exit_code, part_nb, last_member;
-int test = 0;
-int force = 0;
-int verbose = 1;
-long bytes_in, bytes_out;
-
-char *output_data;
-unsigned long output_ptr;
-
-extern int end;
-long free_mem_ptr = (long)&end;
-
-int to_stdout = 0;
-int hard_math = 0;
-
-void (*work)(int inf, int outf);
-void makecrc(void);
-
-local int get_method(int);
-
-char *vidmem = (char *)0xb8000;
-int lines, cols;
-
+static long bytes_out = 0;
+static uch *output_data;
+static unsigned long output_ptr = 0;
+
+ 
+static void *malloc(int size);
+static void free(void *where);
+static void error(char *m);
+static void gzip_mark(void **);
+static void gzip_release(void **);
+ 
+#ifndef STANDALONE_DEBUG
 static void puts(const char *);
-
-void *malloc(int size)
+  
+extern int end;
+static long free_mem_ptr = (long)&end;
+static long free_mem_end_ptr = 0x90000;
+ 
+#define INPLACE_MOVE_ROUTINE  0x1000
+#define LOW_BUFFER_START      0x2000
+#define LOW_BUFFER_END       0x90000
+#define LOW_BUFFER_SIZE      ( LOW_BUFFER_END - LOW_BUFFER_START )
+#define HEAP_SIZE             0x2000
+static int high_loaded =0;
+static uch *high_buffer_start /* = (uch *)(((ulg)&end) + HEAP_SIZE)*/;
+
+static char *vidmem = (char *)0xb8000;
+static int vidport;
+static int lines, cols;
+
+#include "../../../../lib/inflate.c"
+
+static void *malloc(int size)
 {
 	void *p;
 
 	if (size <0) error("Malloc error\n");
 	if (free_mem_ptr <= 0) error("Memory error\n");
 
-   while(1) {
 	free_mem_ptr = (free_mem_ptr + 3) & ~3;	/* Align */
 
 	p = (void *)free_mem_ptr;
 	free_mem_ptr += size;
 
-	/*
-  	 * The part of the compressed kernel which has already been expanded
-	 * is no longer needed. Therefore we can reuse it for malloc.
-	 * With bigger kernels, this is necessary.
-	 */
-          
-	if (free_mem_ptr < (long)&end) {
-		if (free_mem_ptr > (long)&input_data[input_ptr])
-			error("\nOut of memory\n");
-
-		return p;
-	}
-	if (free_mem_ptr < 0x90000)
+	if (free_mem_ptr >= free_mem_end_ptr)
+		error("\nOut of memory\n");
+
 	return p;
-	puts("large kernel, low 1M tight...");
-	free_mem_ptr = (long)input_data;
-	}
 }
 
-void free(void *where)
+static void free(void *where)
 {	/* Don't care */
 }
 
+static void gzip_mark(void **ptr)
+{
+	*ptr = (void *) free_mem_ptr;
+}
+
+static void gzip_release(void **ptr)
+{
+	free_mem_ptr = (long) *ptr;
+}
+ 
 static void scroll()
 {
 	int i;
@@ -127,7 +179,7 @@ static void scroll()
 
 static void puts(const char *s)
 {
-	int x,y;
+	int x,y,pos;
 	char c;
 
 	x = SCREEN_INFO.orig_x;
@@ -154,6 +206,12 @@ static void puts(const char *s)
 
 	SCREEN_INFO.orig_x = x;
 	SCREEN_INFO.orig_y = y;
+
+	pos = (x + cols * y) * 2;	/* Update cursor position */
+	outb_p(14, vidport);
+	outb_p(0xff & (pos >> 9), vidport+1);
+	outb_p(15, vidport);
+	outb_p(0xff & (pos >> 1), vidport+1);
 }
 
 __ptr_t memset(__ptr_t s, int c, size_t n)
@@ -172,129 +230,69 @@ __ptr_t memcpy(__ptr_t __dest, __const __ptr_t __src,
 
 	for (i=0;i<__n;i++) d[i] = s[i];
 }
-
-extern ulg crc_32_tab[];   /* crc table, defined below */
-
-/* ===========================================================================
- * Run a set of bytes through the crc shift register.  If s is a NULL
- * pointer, then initialize the crc shift register contents instead.
- * Return the current crc in either case.
- */
-ulg updcrc(s, n)
-    uch *s;                 /* pointer to bytes to pump through */
-    unsigned n;             /* number of bytes in s[] */
-{
-    register ulg c;         /* temporary variable */
-
-    static ulg crc = (ulg)0xffffffffL; /* shift register contents */
-
-    if (s == NULL) {
-	c = 0xffffffffL;
-    } else {
-	c = crc;
-	while (n--) {
-	    c = crc_32_tab[((int)c ^ (*s++)) & 0xff] ^ (c >> 8);
-	}
-    }
-    crc = c;
-    return c ^ 0xffffffffL;       /* (instead of ~c for 64-bit machines) */
-}
-
-/* ===========================================================================
- * Clear input and output buffers
- */
-void clear_bufs()
-{
-    outcnt = 0;
-    insize = inptr = 0;
-    bytes_in = bytes_out = 0L;
-}
+#endif
 
 /* ===========================================================================
  * Fill the input buffer. This is called only when the buffer is empty
  * and at least one byte is really needed.
  */
-int fill_inbuf()
+static int fill_inbuf()
 {
-    int len, i;
-
-    /* Read as much as possible */
-    insize = 0;
-    do {
-	len = INBUFSIZ-insize;
-	if (len > (input_len-input_ptr+1)) len=input_len-input_ptr+1;
-        if (len == 0 || len == EOF) break;
-
-        for (i=0;i<len;i++) inbuf[insize+i] = input_data[input_ptr+i];
-	insize += len;
-	input_ptr += len;
-    } while (insize < INBUFSIZ);
-
-    if (insize == 0) {
-	error("unable to fill buffer\n");
-    }
-    bytes_in += (ulg)insize;
-    inptr = 1;
-    return inbuf[0];
+	if (insize != 0) {
+		error("ran out of input data\n");
+	}
+
+	inbuf = input_data;
+	insize = input_len;
+	inptr = 1;
+	return inbuf[0];
 }
 
 /* ===========================================================================
  * Write the output window window[0..outcnt-1] and update crc and bytes_out.
  * (Used for the decompressed data only.)
  */
-void flush_window()
+static void flush_window_low()
 {
-    if (outcnt == 0) return;
-    updcrc(window, outcnt);
-
-    memcpy(&output_data[output_ptr], (char *)window, outcnt);
-
+    ulg c = crc;         /* temporary variable */
+    unsigned n;
+    uch *in, *out, ch;
+    
+    in = window;
+    out = &output_data[output_ptr]; 
+    for (n = 0; n < outcnt; n++) {
+	    ch = *out++ = *in++;
+	    c = crc_32_tab[((int)c ^ ch) & 0xff] ^ (c >> 8);
+    }
+    crc = c;
     bytes_out += (ulg)outcnt;
     output_ptr += (ulg)outcnt;
     outcnt = 0;
 }
 
-/*
- * Code to compute the CRC-32 table. Borrowed from 
- * gzip-1.0.3/makecrc.c.
- */
-
-ulg crc_32_tab[256];
-
-void
-makecrc(void)
+static void flush_window_high()
 {
-/* Not copyrighted 1990 Mark Adler	*/
-
-  unsigned long c;      /* crc shift register */
-  unsigned long e;      /* polynomial exclusive-or pattern */
-  int i;                /* counter for all possible eight bit values */
-  int k;                /* byte being shifted into crc apparatus */
-
-  /* terms of polynomial defining this crc (except x^32): */
-  static int p[] = {0,1,2,4,5,7,8,10,11,12,16,22,23,26};
-
-  /* Make exclusive-or pattern from polynomial */
-  e = 0;
-  for (i = 0; i < sizeof(p)/sizeof(int); i++)
-    e |= 1L << (31 - p[i]);
-
-  crc_32_tab[0] = 0;
-
-  for (i = 1; i < 256; i++)
-  {
-    c = 0;
-    for (k = i | 256; k != 1; k >>= 1)
-    {
-      c = c & 1 ? (c >> 1) ^ e : c >> 1;
-      if (k & 1)
-        c ^= e;
+    ulg c = crc;         /* temporary variable */
+    unsigned n;
+    uch *in,  ch;
+    in = window;
+    for (n = 0; n < outcnt; n++) {
+	ch = *output_data++ = *in++;
+	if ((ulg)output_data == LOW_BUFFER_END) output_data=high_buffer_start;
+	c = crc_32_tab[((int)c ^ ch) & 0xff] ^ (c >> 8);
     }
-    crc_32_tab[i] = c;
-  }
+    crc = c;
+    bytes_out += (ulg)outcnt;
+    outcnt = 0;
+}
+
+static void flush_window()
+{
+	if (high_loaded) flush_window_high();
+	else flush_window_low();
 }
 
-void error(char *x)
+static void error(char *x)
 {
 	puts("\n\n");
 	puts(x);
@@ -312,107 +310,96 @@ struct {
 	short b;
 	} stack_start = { & user_stack [STACK_SIZE] , KERNEL_DS };
 
-void decompress_kernel()
-{
-	if (SCREEN_INFO.orig_video_mode == 7)
-		vidmem = (char *) 0xb0000;
-	else
-		vidmem = (char *) 0xb8000;
+#ifdef STANDALONE_DEBUG
 
-	lines = SCREEN_INFO.orig_video_lines;
-	cols = SCREEN_INFO.orig_video_cols;
+static void gzip_mark(void **ptr)
+{
+}
 
-	if (EXT_MEM_K < 1024) error("<2M of mem\n");
+static void gzip_release(void **ptr)
+{
+}
 
-	output_data = (char *)0x100000;	/* Points to 1M */
-	output_ptr = 0;
+char output_buffer[1024 * 800];
 
-	exit_code = 0;
-	test = 0;
-	input_ptr = 0;
-	part_nb = 0;
+int
+main(argc, argv)
+	int	argc;
+	char	**argv;
+{
+	output_data = output_buffer;
 
-	clear_bufs();
 	makecrc();
-
 	puts("Uncompressing Linux...");
+	gunzip();
+	puts("done.\n");
+	return 0;
+}
 
-	method = get_method(0);
+#else
 
-	work(0, 0);
+void setup_normal_output_buffer()
+{
+	if (EXT_MEM_K < 1024) error("Less than 2MB of memory.\n");
+	output_data = (char *)0x100000; /* Points to 1M */
+}
 
-	puts("done.\n");
+struct moveparams {
+	uch *low_buffer_start;  int lcount;
+	uch *high_buffer_start; int hcount;
+};
 
-	puts("Now booting the kernel\n");
+void setup_output_buffer_if_we_run_high(struct moveparams *mv)
+{
+	high_buffer_start = (uch *)(((ulg)&end) + HEAP_SIZE);
+	if (EXT_MEM_K < (3*1024)) error("Less than 4MB of memory.\n");
+	mv->low_buffer_start = output_data = (char *)LOW_BUFFER_START;
+	high_loaded = 1;
+	free_mem_end_ptr = (long)high_buffer_start;
+	if ( (0x100000 + LOW_BUFFER_SIZE) > ((ulg)high_buffer_start)) {
+		high_buffer_start = (uch *)(0x100000 + LOW_BUFFER_SIZE);
+		mv->hcount = 0; /* say: we need not to move high_buffer */
+	}
+	else mv->hcount = -1;
+	mv->high_buffer_start = high_buffer_start;
 }
 
-/* ========================================================================
- * Check the magic number of the input file and update ofname if an
- * original name was given and to_stdout is not set.
- * Return the compression method, -1 for error, -2 for warning.
- * Set inptr to the offset of the next byte to be processed.
- * This function may be called repeatedly for an input file consisting
- * of several contiguous gzip'ed members.
- * IN assertions: there is at least one remaining compressed member.
- *   If the member is a zip file, it must be the only one.
- */
-local int get_method(in)
-    int in;        /* input file descriptor */
+void close_output_buffer_if_we_run_high(struct moveparams *mv)
 {
-    uch flags;
-    char magic[2]; /* magic header */
-
-    magic[0] = (char)get_byte();
-    magic[1] = (char)get_byte();
-
-    method = -1;                 /* unknown yet */
-    part_nb++;                   /* number of parts in gzip file */
-    last_member = 0;
-    /* assume multiple members in gzip file except for record oriented I/O */
-
-    if (memcmp(magic, GZIP_MAGIC, 2) == 0
-        || memcmp(magic, OLD_GZIP_MAGIC, 2) == 0) {
-
-	work = unzip;
-	method = (int)get_byte();
-	flags  = (uch)get_byte();
-	if ((flags & ENCRYPTED) != 0)
-	    error("Input is encrypted\n");
-	if ((flags & CONTINUATION) != 0)
-	       error("Multi part input\n");
-	if ((flags & RESERVED) != 0) {
-	    error("Input has invalid flags\n");
-	    exit_code = ERROR;
-	    if (force <= 1) return -1;
-	}
-	(ulg)get_byte();	/* Get timestamp */
-	((ulg)get_byte()) << 8;
-	((ulg)get_byte()) << 16;
-	((ulg)get_byte()) << 24;
-
-	(void)get_byte();  /* Ignore extra flags for the moment */
-	(void)get_byte();  /* Ignore OS type for the moment */
-
-	if ((flags & EXTRA_FIELD) != 0) {
-	    unsigned len = (unsigned)get_byte();
-	    len |= ((unsigned)get_byte())<<8;
-	    while (len--) (void)get_byte();
+	mv->lcount = bytes_out;
+	if (bytes_out > LOW_BUFFER_SIZE) {
+		mv->lcount = LOW_BUFFER_SIZE;
+		if (mv->hcount) mv->hcount = bytes_out - LOW_BUFFER_SIZE;
 	}
+	else mv->hcount = 0;
+}
 
-	/* Get original file name if it was truncated */
-	if ((flags & ORIG_NAME) != 0) {
-	    if (to_stdout || part_nb > 1) {
-		/* Discard the old name */
-		while (get_byte() != 0) /* null */ ;
-	    } else {
-	    } /* to_stdout */
-	} /* orig_name */
-
-	/* Discard file comment if any */
-	if ((flags & COMMENT) != 0) {
-	    while (get_byte() != 0) /* null */ ;
+
+int decompress_kernel(struct moveparams *mv)
+{
+	if (SCREEN_INFO.orig_video_mode == 7) {
+		vidmem = (char *) 0xb0000;
+		vidport = 0x3b4;
+	} else {
+		vidmem = (char *) 0xb8000;
+		vidport = 0x3d4;
 	}
-    } else
-	error("unknown compression method");
-    return method;
+
+	lines = SCREEN_INFO.orig_video_lines;
+	cols = SCREEN_INFO.orig_video_cols;
+
+	if (free_mem_ptr < 0x100000) setup_normal_output_buffer();
+	else setup_output_buffer_if_we_run_high(mv);
+
+	makecrc();
+	puts("Uncompressing Linux...");
+	gunzip();
+	puts("done.\nNow booting the kernel\n");
+	if (high_loaded) close_output_buffer_if_we_run_high(mv);
+	return high_loaded;
 }
+#endif
+
+
+
+
diff --git a/arch/i386/boot/compressed/unzip.c b/arch/i386/boot/compressed/unzip.c
deleted file mode 100644
index d4a6617cd..000000000
--- a/arch/i386/boot/compressed/unzip.c
+++ /dev/null
@@ -1,180 +0,0 @@
-/* unzip.c -- decompress files in gzip or pkzip format.
- * Copyright (C) 1992-1993 Jean-loup Gailly
- *
- * Adapted for Linux booting by Hannu Savolainen 1993
- *
- * This is free software; you can redistribute it and/or modify it under the
- * terms of the GNU General Public License, see the file COPYING.
- *
- * The code in this file is derived from the file funzip.c written
- * and put in the public domain by Mark Adler.
- */
-
-/*
-   This version can extract files in gzip or pkzip format.
-   For the latter, only the first entry is extracted, and it has to be
-   either deflated or stored.
- */
-
-#ifndef lint
-static char rcsid[] = "$Id: unzip.c,v 0.9 1993/02/10 16:07:22 jloup Exp $";
-#endif
-
-#include "gzip.h"
-#include "crypt.h"
-
-#include <stdio.h>
-
-/* PKZIP header definitions */
-#define LOCSIG 0x04034b50L      /* four-byte lead-in (lsb first) */
-#define LOCFLG 6                /* offset of bit flag */
-#define  CRPFLG 1               /*  bit for encrypted entry */
-#define  EXTFLG 8               /*  bit for extended local header */
-#define LOCHOW 8                /* offset of compression method */
-#define LOCTIM 10               /* file mod time (for decryption) */
-#define LOCCRC 14               /* offset of crc */
-#define LOCSIZ 18               /* offset of compressed size */
-#define LOCLEN 22               /* offset of uncompressed length */
-#define LOCFIL 26               /* offset of file name field length */
-#define LOCEXT 28               /* offset of extra field length */
-#define LOCHDR 30               /* size of local header, including sig */
-#define EXTHDR 16               /* size of extended local header, inc sig */
-
-
-/* Globals */
-
-int decrypt;      /* flag to turn on decryption */
-char *key;        /* not used--needed to link crypt.c */
-int pkzip = 0;    /* set for a pkzip file */
-int extended = 0; /* set if extended local header */
-
-/* ===========================================================================
- * Check zip file and advance inptr to the start of the compressed data.
- * Get ofname from the local header if necessary.
- */
-int check_zipfile(in)
-    int in;   /* input file descriptors */
-{
-    uch *h = inbuf + inptr; /* first local header */
-
-    /* ifd = in; */
-
-    /* Check validity of local header, and skip name and extra fields */
-    inptr += LOCHDR + SH(h + LOCFIL) + SH(h + LOCEXT);
-
-    if (inptr > insize || LG(h) != LOCSIG) {
-	error("input not a zip");
-    }
-    method = h[LOCHOW];
-    if (method != STORED && method != DEFLATED) {
-	error("first entry not deflated or stored--can't extract");
-    }
-
-    /* If entry encrypted, decrypt and validate encryption header */
-    if ((decrypt = h[LOCFLG] & CRPFLG) != 0) {
-	error("encrypted file\n");
-	exit_code = ERROR;
-	return -1;
-    }
-
-    /* Save flags for unzip() */
-    extended = (h[LOCFLG] & EXTFLG) != 0;
-    pkzip = 1;
-
-    /* Get ofname and time stamp from local header (to be done) */
-    return 0;
-}
-
-/* ===========================================================================
- * Unzip in to out.  This routine works on both gzip and pkzip files.
- *
- * IN assertions: the buffer inbuf contains already the beginning of
- *   the compressed data, from offsets inptr to insize-1 included.
- *   The magic header has already been checked. The output buffer is cleared.
- */
-void unzip(in, out)
-    int in, out;   /* input and output file descriptors */
-{
-    ulg orig_crc = 0;       /* original crc */
-    ulg orig_len = 0;       /* original uncompressed length */
-    int n;
-    uch buf[EXTHDR];        /* extended local header */
-
-    /* ifd = in;
-    ofd = out; */
-
-    updcrc(NULL, 0);           /* initialize crc */
-
-    if (pkzip && !extended) {  /* crc and length at the end otherwise */
-	orig_crc = LG(inbuf + LOCCRC);
-	orig_len = LG(inbuf + LOCLEN);
-    }
-
-    /* Decompress */
-    if (method == DEFLATED)  {
-
-	int res = inflate();
-
-	if (res == 3) {
-	    error("out of memory");
-	} else if (res != 0) {
-	    error("invalid compressed format");
-	}
-
-    } else if (pkzip && method == STORED) {
-
-	register ulg n = LG(inbuf + LOCLEN);
-
-	if (n != LG(inbuf + LOCSIZ) - (decrypt ? RAND_HEAD_LEN : 0)) {
-
-	    error("length mismatch");
-	}
-	while (n--) {
-	    uch c = (uch)get_byte();
-#ifdef CRYPT
-	    if (decrypt) zdecode(c);
-#endif
-	    if (!test) put_char(c);
-	}
-    } else {
-	error("internal error, invalid method");
-    }
-
-    /* Get the crc and original length */
-    if (!pkzip) {
-        /* crc32  (see algorithm.doc)
-	 * uncompressed input size modulo 2^32
-         */
-	for (n = 0; n < 8; n++) {
-	    buf[n] = (uch)get_byte(); /* may cause an error if EOF */
-	}
-	orig_crc = LG(buf);
-	orig_len = LG(buf+4);
-
-    } else if (extended) {  /* If extended header, check it */
-	/* signature - 4bytes: 0x50 0x4b 0x07 0x08
-	 * CRC-32 value
-         * compressed size 4-bytes
-         * uncompressed size 4-bytes
-	 */
-	for (n = 0; n < EXTHDR; n++) {
-	    buf[n] = (uch)get_byte(); /* may cause an error if EOF */
-	}
-	orig_crc = LG(buf+4);
-	orig_len = LG(buf+12);
-    }
-
-    /* Validate decompression */
-    if (orig_crc != updcrc(outbuf, 0)) {
-	error("crc error");
-    }
-    if (orig_len != bytes_out) {
-	error("length error");
-    }
-
-    /* Check if there are more entries in a pkzip file */
-    if (pkzip && inptr + 4 < insize && LG(inbuf+inptr) == LOCSIG) {
-	    error("zip file has more than one entry");
-    }
-    extended = pkzip = 0; /* for next file */
-}
diff --git a/arch/i386/boot/setup.S b/arch/i386/boot/setup.S
index d4f56ef2a..201a2936b 100644
--- a/arch/i386/boot/setup.S
+++ b/arch/i386/boot/setup.S
@@ -17,23 +17,32 @@
 ! some changes and additional features by Christoph Niemann,
 ! March 1993/June 1994 (Christoph.Niemann@linux.org)
 !
+! add APM BIOS checking by Stephen Rothwell, May 1994
+! (Stephen.Rothwell@pd.necisa.oz.au)
+!
+! High load stuff, initrd support and position independency
+! by Hans Lermen & Werner Almesberger, February 1996
+! <lermen@elserv.ffm.fgan.de>, <almesber@lrc.epfl.ch>
+!
+! Video handling moved to video.S by Martin Mares, March 1996
+! <mj@k332.feld.cvut.cz>
 
 ! NOTE! These had better be the same as in bootsect.s!
 #define __ASSEMBLY__
 #include <linux/config.h>
 #include <asm/segment.h>
-
-#ifndef SVGA_MODE
-#define SVGA_MODE ASK_VGA
-#endif
+#include <linux/version.h>
+#include <linux/compile.h>
 
 ! Signature words to ensure LILO loaded us right
 #define SIG1	0xAA55
 #define SIG2	0x5A5A
 
-INITSEG  = DEF_INITSEG	! we move boot here - out of the way
-SYSSEG   = DEF_SYSSEG	! system loaded at 0x10000 (65536).
-SETUPSEG = DEF_SETUPSEG	! this is the current segment
+INITSEG  = DEF_INITSEG	! 0x9000, we move boot here - out of the way
+SYSSEG   = DEF_SYSSEG	! 0x1000, system loaded at 0x10000 (65536).
+SETUPSEG = DEF_SETUPSEG	! 0x9020, this is the current segment
+			! ... and the former contents of CS
+DELTA_INITSEG = SETUPSEG - INITSEG ! 0x0020
 
 .globl begtext, begdata, begbss, endtext, enddata, endbss
 .text
@@ -46,14 +55,77 @@ begbss:
 
 entry start
 start:
+	jmp	start_of_setup
+! ------------------------ start of header --------------------------------
+!
+! SETUP-header, must start at CS:2 (old 0x9020:2)
+!
+		.ascii	"HdrS"		! Signature for SETUP-header
+		.word	0x0201		! Version number of header format
+					! (must be >= 0x0105
+					! else old loadlin-1.5 will fail)
+realmode_swtch:	.word	0,0		! default_switch,SETUPSEG
+start_sys_seg:	.word	SYSSEG
+		.word	kernel_version	! pointing to kernel version string
+  ! note: above part of header is compatible with loadlin-1.5 (header v1.5),
+  !        must not change it
+
+type_of_loader:	.byte	0		! = 0, old one (LILO, Loadlin,
+					!      Bootlin, SYSLX, bootsect...)
+					! else it is set by the loader:
+					! 0xTV: T=0 for LILO
+					!	T=1 for Loadlin
+					!	T=2 for bootsect-loader
+					!	T=3 for SYSLX
+					!	T=4 for ETHERBOOT
+					!       V = version
+loadflags:	.byte	0	! unused bits =0 (reserved for future development)
+LOADED_HIGH	= 1		! bit within loadflags,
+				! if set, then the kernel is loaded high
+CAN_USE_HEAP	= 0x80		! if set, the loader also has set heap_end_ptr
+				! to tell how much space behind setup.S
+				| can be used for heap purposes.
+				! Only the loader knows what is free!
+setup_move_size: .word  0x8000	! size to move, when we (setup) are not
+				! loaded at 0x90000. We will move ourselves
+				! to 0x90000 then just before jumping into
+				! the kernel. However, only the loader
+				! know how much of data behind us also needs
+				! to be loaded.
+code32_start:	.long	0x1000		! here loaders can put a different
+					! start address for 32-bit code.
+					!   0x1000 = default for zImage
+					! 0x100000 = default for big kernel
+ramdisk_image:	.long	0	! address of loaded ramdisk image
+				! Here the loader (or kernel generator) puts
+				! the 32-bit address were it loaded the image.
+				! This only will be interpreted by the kernel.
+ramdisk_size:	.long	0	! its size in bytes
+bootsect_kludge:
+		.word   bootsect_helper,SETUPSEG
+heap_end_ptr:	.word	modelist+1024	! space from here (exclusive) down to
+				! end of setup code can be used by setup
+				! for local heap purposes.
+! ------------------------ end of header ----------------------------------
+
+start_of_setup:
 ! Bootlin depends on this being done early
 	mov	ax,#0x01500
 	mov	dl,#0x81
 	int	0x13
 
-! Check signature at end of setup
-	mov	ax,#SETUPSEG
+#ifdef SAFE_RESET_DISK_CONTROLLER
+! Reset the disk controller.
+	mov	ax,#0x0000
+	mov	dl,#0x80
+	int	0x13
+#endif
+
+! set DS=CS, we know that SETUPSEG == CS at this point
+	mov	ax,cs		! aka #SETUPSEG
 	mov	ds,ax
+
+! Check signature at end of setup
 	cmp	setup_sig1,#SIG1
 	jne	bad_sig
 	cmp	setup_sig2,#SIG2
@@ -65,13 +137,18 @@ start:
 prtstr:	lodsb
 	and	al,al
 	jz	fin
-	call	prnt1
+	call	prtchr
 	jmp	prtstr
 fin:	ret
 
+! Space printing
+
+prtsp2:	call	prtspc		! Print double space
+prtspc:	mov	al,#0x20	! Print single space (fall-thru!)
+
 ! Part of above routine, this one just prints ascii al
 
-prnt1:	push	ax
+prtchr:	push	ax
 	push	cx
 	xor	bh,bh
 	mov	cx,#0x01
@@ -82,18 +159,18 @@ prnt1:	push	ax
 	ret
 
 beep:	mov	al,#0x07
-	jmp	prnt1
+	jmp	prtchr
 	
 no_sig_mess:	.ascii	"No setup signature found ..."
 		db	0x00
-start_sys_seg:	.word	SYSSEG
 
 good_sig1:
 	jmp	good_sig
 
 ! We now have to find the rest of the setup code/data
 bad_sig:
-	mov	ax,#INITSEG
+	mov	ax,cs		! aka #SETUPSEG
+	sub	ax,#DELTA_INITSEG ! aka #INITSEG
 	mov	ds,ax
 	xor	bh,bh
 	mov	bl,[497]	! get setup sects from boot sector
@@ -108,14 +185,14 @@ bad_sig:
 ! Move rest of setup code/data to here
 	mov	di,#2048	! four sectors loaded by LILO
 	sub	si,si
-	mov	ax,#SETUPSEG
+	mov	ax,cs		! aka #SETUPSEG
 	mov	es,ax
 	mov	ax,#SYSSEG
 	mov	ds,ax
 	rep
 	movsw
 
-	mov	ax,#SETUPSEG
+	mov	ax,cs		! aka #SETUPSEG
 	mov	ds,ax
 	cmp	setup_sig1,#SIG1
 	jne	no_sig
@@ -130,58 +207,53 @@ no_sig_loop:
 	jmp	no_sig_loop
 
 good_sig:
-	mov	ax,#INITSEG
+	mov	ax,cs		! aka #SETUPSEG
+	sub	ax,#DELTA_INITSEG ! aka #INITSEG
 	mov	ds,ax
 
+! check if an old loader tries to load a big-kernel
+	seg cs
+	test	byte ptr loadflags,#LOADED_HIGH ! have we a big kernel ?
+	jz	loader_ok	! NO, no danger even for old loaders
+				! YES, we have a big-kernel
+	seg cs
+	cmp	byte ptr type_of_loader,#0 ! have we one of the new loaders ?
+	jnz	loader_ok	! YES, ok
+				! NO, we have an old loader, must give up
+	push    cs
+	pop	ds
+	lea	si,loader_panic_mess
+	call	prtstr
+	jmp	no_sig_loop
+loader_panic_mess: 
+	.ascii	"Wrong loader, giving up..."
+	db	0
+
+loader_ok:
 ! Get memory size (extended mem, kB)
 
 	mov	ah,#0x88
 	int	0x15
 	mov	[2],ax
 
-! set the keyboard repeat rate to the max
+! Set the keyboard repeat rate to the max
 
 	mov	ax,#0x0305
 	xor	bx,bx		! clear bx
 	int	0x16
 
-! check for EGA/VGA and some config parameters
+! Check for video adapter and its parameters and allow the
+! user to browse video modes.
 
-	mov	ah,#0x12
-	mov	bl,#0x10
-	int	0x10
-	mov	[8],ax
-	mov	[10],bx
-	mov	[12],cx
-	mov	ax,#0x5019
-	cmp	bl,#0x10
-	je	novga
-	mov	ax,#0x1a00	! Added check for EGA/VGA discrimination
-	int	0x10
-	mov	bx,ax
-	mov	ax,#0x5019
-	cmp	bl,#0x1a	! 1a means VGA, anything else EGA or lower
-	jne	novga	
-	call	chsvga
-novga:	mov	[14],ax
-	mov	ah,#0x03	! read cursor pos
-	xor	bh,bh		! clear bh
-	int	0x10		! save it in known place, con_init fetches
-	mov	[0],dx		! it from 0x90000.
-	
-! Get video-card data:
-	
-	mov	ah,#0x0f
-	int	0x10
-	mov	[4],bx		! bh = display page
-	mov	[6],ax		! al = video mode, ah = window width
+	call	video	! NOTE: we need DS pointing to bootsector
 
 ! Get hd0 data
 
 	xor	ax,ax		! clear ax
 	mov	ds,ax
 	lds	si,[4*0x41]
-	mov	ax,#INITSEG
+	mov	ax,cs		! aka #SETUPSEG
+	sub	ax,#DELTA_INITSEG ! aka #INITSEG
 	push	ax
 	mov	es,ax
 	mov	di,#0x0080
@@ -211,7 +283,8 @@ novga:	mov	[14],ax
 	cmp	ah,#3
 	je	is_disk1
 no_disk1:
-	mov	ax,#INITSEG
+	mov	ax,cs		! aka #SETUPSEG
+	sub	ax,#DELTA_INITSEG ! aka #INITSEG
 	mov	es,ax
 	mov	di,#0x0090
 	mov	cx,#0x10
@@ -221,9 +294,10 @@ no_disk1:
 	stosb
 is_disk1:
 
-! check for PS/2 pointing device
+! Check for PS/2 pointing device
 
-	mov	ax,#INITSEG
+	mov	ax,cs		! aka #SETUPSEG
+	sub	ax,#DELTA_INITSEG ! aka #INITSEG
 	mov	ds,ax
 	mov	[0x1ff],#0	! default is no pointing device
 	int	0x11		! int 0x11: equipment determination
@@ -231,23 +305,88 @@ is_disk1:
 	jz	no_psmouse
 	mov	[0x1ff],#0xaa	! device present
 no_psmouse:
-! now we want to move to protected mode ...
 
-	cli			! no interrupts allowed !
-	mov	al,#0x80	! disable NMI for the bootup sequence
-	out	#0x70,al
+#ifdef CONFIG_APM
+! check for APM BIOS
+		! NOTE:	DS is pointing to the bootsector
+		!
+	mov	[64],#0		! version == 0 means no APM BIOS
+
+	mov	ax,#0x05300	! APM BIOS installation check
+	xor	bx,bx
+	int	0x15
+	jc	done_apm_bios	! error -> no APM BIOS
+
+	cmp	bx,#0x0504d	! check for "PM" signature
+	jne	done_apm_bios	! no signature -> no APM BIOS
+
+	mov	[64],ax		! record the APM BIOS version
+	mov	[76],cx		!	and flags
+	and	cx,#0x02	! Is 32 bit supported?
+	je	done_apm_bios	!	no ...
+
+	mov	ax,#0x05304	! Disconnect first just in case
+	xor	bx,bx
+	int	0x15		! ignore return code
+
+	mov	ax,#0x05303	! 32 bit connect
+	xor	bx,bx
+	int	0x15
+	jc	no_32_apm_bios	! error
+
+	mov	[66],ax		! BIOS code segment
+	mov	[68],ebx	! BIOS entry point offset
+	mov	[72],cx		! BIOS 16 bit code segment
+	mov	[74],dx		! BIOS data segment
+	mov	[78],si		! BIOS code segment length
+	mov	[80],di		! BIOS data segment length
+	jmp	done_apm_bios
+
+no_32_apm_bios:
+	and	[76], #0xfffd	! remove 32 bit support bit
+
+done_apm_bios:
+#endif
+
+! Now we want to move to protected mode ...
+
+	seg cs
+	cmp	realmode_swtch,#0
+	jz	rmodeswtch_normal
+	seg cs
+	callf	far * realmode_swtch
+	jmp	rmodeswtch_end
+rmodeswtch_normal:
+        push	cs
+	call	default_switch
+rmodeswtch_end:
+
+! we get the code32 start address and modify the below 'jmpi'
+! (loader may have changed it)
+	seg cs
+	mov	eax,code32_start
+	seg cs
+	mov	code32,eax
 
-! first we move the system to its rightful place
+! Now we move the system to its rightful place
+! ...but we check, if we have a big-kernel.
+! in this case we *must* not move it ...
+	seg cs
+	test	byte ptr loadflags,#LOADED_HIGH
+	jz	do_move0	! we have a normal low loaded zImage
+				! we have a high loaded big kernel
+	jmp	end_move	! ... and we skip moving
 
+do_move0:
 	mov	ax,#0x100	! start of destination segment
+	mov	bp,cs		! aka #SETUPSEG
+	sub	bp,#DELTA_INITSEG ! aka #INITSEG
 	seg cs
 	mov	bx,start_sys_seg	! start of source segment
 	cld			! 'direction'=0, movs moves forward
 do_move:
 	mov	es,ax		! destination segment
 	inc	ah		! instead of add ax,#0x100
-	cmp	ax,#0x9000
-	jz	end_move
 	mov	ds,bx		! source segment
 	add	bx,#0x100
 	sub	di,di
@@ -255,13 +394,58 @@ do_move:
 	mov 	cx,#0x800
 	rep
 	movsw
-	jmp	do_move
+	cmp	bx,bp		! we assume start_sys_seg > 0x200,
+				! so we will perhaps read one page more then
+				! needed, but never overwrite INITSEG because
+				! destination is minimum one page below source
+	jb	do_move
 
 ! then we load the segment descriptors
 
 end_move:
-	mov	ax,#SETUPSEG	! right, forgot this at first. didn't work :-)
+	mov	ax,cs ! aka #SETUPSEG	! right, forgot this at first. didn't work :-)
+	mov	ds,ax
+
+! If we have our code not at 0x90000, we need to move it there now.
+! We also then need to move the params behind it (commandline)
+! Because we would overwrite the code on the current IP, we move
+! it in two steps, jumping high after the first one.
+	mov	ax,cs
+	cmp	ax,#SETUPSEG
+	je	end_move_self
+	cli	! make sure we really have interrupts disabled !
+		! because after this the stack should not be used
+	sub	ax,#DELTA_INITSEG ! aka #INITSEG
+	mov	dx,ss
+	cmp	dx,ax
+	jb	move_self_1
+	add	dx,#INITSEG
+	sub	dx,ax		! this will be SS after the move
+move_self_1:
 	mov	ds,ax
+	mov	ax,#INITSEG	! real INITSEG
+	mov	es,ax
+	seg cs
+	mov	cx,setup_move_size
+	std		! we have to move up, so we use direction down
+			! because the areas may overlap
+	mov	di,cx
+	dec	di
+	mov	si,di
+	sub	cx,#move_self_here+0x200
+	rep
+	movsb
+	jmpi	move_self_here,SETUPSEG ! jump to our final place
+move_self_here:
+	mov	cx,#move_self_here+0x200
+	rep
+	movsb
+	mov	ax,#SETUPSEG
+	mov	ds,ax
+	mov	ss,dx
+			! now we are at the right place
+end_move_self:
+
 	lidt	idt_48		! load idt with 0,0
 	lgdt	gdt_48		! load gdt with whatever appropriate
 
@@ -319,7 +503,7 @@ end_move:
 	mov	al,#0xFB		! mask all irq's but irq2 which
 	out	#0x21,al		! is cascaded
 
-! well, that certainly wasn't fun :-(. Hopefully it works, and we don't
+! Well, that certainly wasn't fun :-(. Hopefully it works, and we don't
 ! need no steenking BIOS anyway (except for the initial loading :-).
 ! The BIOS-routine wants lots of unnecessary data, and it's less
 ! "interesting" anyway. This is how REAL programmers do it.
@@ -327,17 +511,127 @@ end_move:
 ! Well, now's the time to actually move into protected mode. To make
 ! things as simple as possible, we do no register set-up or anything,
 ! we let the gnu-compiled 32-bit programs do that. We just jump to
-! absolute address 0x00000, in 32-bit protected mode.
+! absolute address 0x1000 (or the loader supplied one),
+! in 32-bit protected mode.
 !
 ! Note that the short jump isn't strictly needed, although there are
 ! reasons why it might be a good idea. It won't hurt in any case.
 !
-	xor	ax,ax
-	inc	ax		! protected mode (PE) bit
+	mov	ax,#1		! protected mode (PE) bit
 	lmsw	ax		! This is it!
 	jmp	flush_instr
 flush_instr:
-	jmpi	0x1000,KERNEL_CS	! jmp offset 1000 of segment 0x10 (cs)
+	xor	bx,bx		! Flag to indicate a boot
+
+! NOTE: For high loaded big kernels we need a
+!	jmpi    0x100000,KERNEL_CS
+!
+!	but we yet haven't reloaded the CS register, so the default size 
+!	of the target offset still is 16 bit.
+!       However, using an operant prefix (0x66), the CPU will properly
+!	take our 48 bit far pointer. (INTeL 80386 Programmer's Reference
+!	Manual, Mixing 16-bit and 32-bit code, page 16-6)
+	db	0x66,0xea	! prefix + jmpi-opcode
+code32:	dd	0x1000		! will be set to 0x100000 for big kernels
+	dw	KERNEL_CS
+
+
+kernel_version:	.ascii	UTS_RELEASE
+		.ascii	" ("
+		.ascii	LINUX_COMPILE_BY
+		.ascii	"@"
+		.ascii	LINUX_COMPILE_HOST
+		.ascii	") "
+		.ascii	UTS_VERSION
+		db	0
+
+! This is the default real mode switch routine.
+! to be called just before protected mode transition
+
+default_switch:
+	cli			! no interrupts allowed !
+	mov	al,#0x80	! disable NMI for the bootup sequence
+	out	#0x70,al
+	retf
+
+! This routine only gets called, if we get loaded by the simple
+! bootsect loader _and_ have a bzImage to load.
+! Because there is no place left in the 512 bytes of the boot sector,
+! we must emigrate to code space here.
+!
+bootsect_helper:
+	seg cs
+	cmp	word ptr bootsect_es,#0
+	jnz	bootsect_second
+	seg cs
+	mov	byte ptr type_of_loader,#0x20
+	mov	ax,es
+	shr	ax,#4
+	seg	cs
+	mov	byte ptr bootsect_src_base+2,ah
+	mov	ax,es
+	seg cs
+	mov	bootsect_es,ax
+	sub	ax,#SYSSEG
+	retf			! nothing else to do for now
+bootsect_second:
+	push	cx
+	push	si
+	push	bx
+	test	bx,bx	! 64K full ?
+	jne	bootsect_ex
+	mov	cx,#0x8000	! full 64K move, INT15 moves words
+	push	cs
+	pop	es
+	mov	si,#bootsect_gdt
+	mov	ax,#0x8700
+	int	0x15
+	jc	bootsect_panic	! this, if INT15 fails
+	seg cs
+	mov	es,bootsect_es	! we reset es to always point to 0x10000
+	seg cs
+	inc	byte ptr bootsect_dst_base+2
+bootsect_ex:
+	seg cs
+	mov	ah, byte ptr bootsect_dst_base+2
+	shl	ah,4	! we now have the number of moved frames in ax
+	xor	al,al
+	pop	bx
+	pop	si
+	pop	cx
+	retf
+
+bootsect_gdt:
+	.word	0,0,0,0
+	.word	0,0,0,0
+bootsect_src:
+	.word	0xffff
+bootsect_src_base:
+	.byte	0,0,1			! base = 0x010000
+	.byte	0x93			! typbyte
+	.word	0			! limit16,base24 =0
+bootsect_dst:
+	.word	0xffff
+bootsect_dst_base:
+	.byte	0,0,0x10		! base = 0x100000
+	.byte	0x93			! typbyte
+	.word	0			! limit16,base24 =0
+	.word	0,0,0,0			! BIOS CS
+	.word	0,0,0,0			! BIOS DS
+bootsect_es:
+	.word	0
+
+bootsect_panic:
+	push	cs
+	pop	ds
+	cld
+	lea	si,bootsect_panic_mess
+	call	prtstr
+bootsect_panic_loop:
+	jmp	bootsect_panic_loop
+bootsect_panic_mess:
+	.ascii	"INT15 refuses to access high mem, giving up..."
+	db	0
 
 ! This routine checks that the keyboard command queue is empty
 ! (after emptying the output buffers)
@@ -356,45 +650,6 @@ no_output:
 	test	al,#2		! is input buffer full?
 	jnz	empty_8042	! yes - loop
 	ret
-!
-! Read a key and return the (US-)ascii code in al, scan code in ah
-!
-getkey:
-	xor	ah,ah
-	int	0x16
-	ret
-
-!
-! Read a key with a timeout of 30 seconds. The cmos clock is used to get
-! the time.
-!
-getkt:
-	call	gettime
-	add	al,#30		! wait 30 seconds
-	cmp	al,#60
-	jl	lminute
-	sub	al,#60
-lminute:
-	mov	cl,al
-again:	mov	ah,#0x01
-	int	0x16
-	jnz	getkey		! key pressed, so get it
-	call	gettime
-	cmp	al,cl
-	jne	again
-	mov	al,#0x20	! timeout, return default char `space'
-	ret
-
-!
-! Flush the keyboard buffer
-!
-flush:	mov	ah,#0x01
-	int	0x16
-	jz	empty
-	xor	ah,ah
-	int	0x16
-	jmp	flush
-empty:	ret
 
 !
 ! Read the cmos clock. Return the seconds in al
@@ -413,524 +668,30 @@ gettime:
 	ret
 
 !
-! Delay is needed after doing i/o
+! Delay is needed after doing I/O
 !
 delay:
 	.word	0x00eb			! jmp $+2
 	ret
 
-! Routine trying to recognize type of SVGA-board present (if any)
-! and if it recognize one gives the choices of resolution it offers.
-! If one is found the resolution chosen is given by al,ah (rows,cols).
-
-chsvga:	cld
-	push	ds
-	push	cs
-	mov	ax,[0x01fa]
-	pop	ds
-	mov	modesave,ax
-	mov 	ax,#0xc000
-	mov	es,ax
-	mov	ax,modesave
-	cmp	ax,#NORMAL_VGA
-	je	defvga
-	cmp	ax,#EXTENDED_VGA
-	je	vga50
-	cmp	ax,#ASK_VGA
-	jne	svga
-	lea	si,msg1
-	call	prtstr
-	call	flush
-nokey:	call	getkt
-	cmp	al,#0x0d		! enter ?
-	je	svga			! yes - svga selection
-	cmp	al,#0x20		! space ?
-	je	defvga			! no - repeat
-	call 	beep
-	jmp	nokey
-defvga:	mov	ax,#0x5019
-	pop	ds
-	ret
-/* extended vga mode: 80x50 */
-vga50:
-	mov	ax,#0x1112
-	xor	bl,bl
-	int	0x10		! use 8x8 font set (50 lines on VGA)
-	mov	ax,#0x1200
-	mov	bl,#0x20
-	int	0x10		! use alternate print screen
-	mov	ax,#0x1201
-	mov	bl,#0x34
-	int	0x10		! turn off cursor emulation
-	mov	ah,#0x01
-	mov	cx,#0x0607
-	int	0x10		! turn on cursor (scan lines 6 to 7)
-	pop	ds
-	mov	ax,#0x5032	! return 80x50
-	ret
-/* extended vga mode: 80x28 */
-vga28:
-	pop	ax		! clean the stack
-	mov	ax,#0x1111
-	xor	bl,bl
-	int	0x10		! use 9x14 fontset (28 lines on VGA)
-	mov	ah, #0x01
-	mov	cx,#0x0b0c
-	int	0x10		! turn on cursor (scan lines 11 to 12)
-	pop	ds
-	mov	ax,#0x501c	! return 80x28
-	ret
-/* svga modes */
 !
-!	test for presence of an S3 VGA chip. The algorithm was taken
-!	from the SuperProbe package of XFree86 1.2.1
-!	report bugs to Christoph.Niemann@linux.org
+! Descriptor tables
 !
-svga:   cld
-	mov	cx,#0x0f35	! we store some constants in cl/ch
-	mov	dx,#0x03d4
-	movb	al,#0x38
-	call	inidx
-	mov	bh,al		! store current value of CRT-register 0x38
-	mov	ax,#0x0038
-	call	outidx		! disable writing to special regs
-	movb	al,cl		! check whether we can write special reg 0x35
-	call	inidx
-	movb	bl,al		! save the current value of CRT reg 0x35
-	andb	al,#0xf0	! clear bits 0-3
-	movb	ah,al
-	movb	al,cl		! and write it to CRT reg 0x35
-	call	outidx
-	call	inidx		! now read it back
-	andb	al,ch		! clear the upper 4 bits
-	jz	s3_2		! the first test failed. But we have a
-	movb	ah,bl		! second chance
-	mov	al,cl
-	call	outidx
-	jmp	s3_1		! do the other tests
-s3_2:	mov	ax,cx		! load ah with 0xf and al with 0x35
-	orb	ah,bl		! set the upper 4 bits of ah with the orig value
-	call	outidx		! write ...
-	call	inidx		! ... and reread 
-	andb	al,cl		! turn off the upper 4 bits
-	push	ax
-	movb	ah,bl		! restore old value in register 0x35
-	movb	al,cl
-	call	outidx
-	pop	ax
-	cmp	al,ch		! setting lower 4 bits was successful => bad
-	je	no_s3		! writing is allowed => this is not an S3
-s3_1:	mov	ax,#0x4838	! allow writing to special regs by putting
-	call	outidx		! magic number into CRT-register 0x38
-	movb	al,cl		! check whether we can write special reg 0x35
-	call	inidx
-	movb	bl,al
-	andb	al,#0xf0
-	movb	ah,al
-	movb	al,cl
-	call	outidx
-	call	inidx
-	andb	al,ch
-	jnz	no_s3		! no, we can't write => no S3
-	mov	ax,cx
-	orb	ah,bl
-	call	outidx
-	call	inidx
-	andb	al,ch
-	push	ax
-	movb	ah,bl		! restore old value in register 0x35
-	movb	al,cl
-	call	outidx
-	pop	ax
-	cmp	al,ch
-	jne	no_s31		! writing not possible => no S3
-	movb	al,#0x30
-	call	inidx		! now get the S3 id ...
-	lea	di,idS3
-	mov	cx,#0x10
-	repne
-	scasb
-	je	no_s31
-	lea 	si,dsc_S3	! table of descriptions of video modes for BIOS
-	lea	di,mo_S3	! table of sizes of video modes for my BIOS
-	movb	ah,bh
-	movb	al,#0x38
-	call	outidx		! restore old value of CRT register 0x38
-	br	selmod		! go ask for video mode
-no_s3:	movb	al,#0x35	! restore CRT register 0x35
-	movb	ah,bl
-	call	outidx
-no_s31:	movb	ah,bh
-	movb	al,#0x38
-	call	outidx		! restore old value of CRT register 0x38
-
-	lea 	si,idati		! Check ATI 'clues'
-	mov	di,#0x31
-	mov 	cx,#0x09
-	repe
-	cmpsb
-	jne	noati
-	lea	si,dscati
-	lea	di,moati
-	br	selmod
-noati:	mov	ax,#0x200f		! Check Ahead 'clues'
-	mov	dx,#0x3ce
-	out	dx,ax
-	inc	dx
-	in	al,dx
-	cmp	al,#0x20
-	je	isahed
-	cmp	al,#0x21
-	jne	noahed
-isahed:	lea	si,dscahead
-	lea	di,moahead
-	br	selmod
-noahed:	mov	dx,#0x3c3		! Check Chips & Tech. 'clues'
-	in	al,dx
-	or	al,#0x10
-	out	dx,al
-	mov	dx,#0x104		
-	in	al,dx
-	mov	bl,al
-	mov	dx,#0x3c3
-	in	al,dx
-	and	al,#0xef
-	out	dx,al
-	cmp	bl,[idcandt]
-	jne	nocant
-	lea	si,dsccandt
-	lea	di,mocandt
-	br	selmod
-nocant:	mov	dx,#0x3d4		! Check Cirrus 'clues'
-	mov	al,#0x0c
-	out	dx,al
-	inc	dx
-	in	al,dx
-	mov	bl,al
-	xor	al,al
-	out	dx,al
-	dec	dx
-	mov	al,#0x1f
-	out	dx,al
-	inc	dx
-	in	al,dx
-	mov	bh,al
-	xor	ah,ah
-	shl	al,#4
-	mov	cx,ax
-	mov	al,bh
-	shr	al,#4
-	add	cx,ax
-	shl	cx,#8
-	add	cx,#6
-	mov	ax,cx
-	mov	dx,#0x3c4
-	out	dx,ax
-	inc	dx
-	in	al,dx
-	and	al,al
-	jnz	nocirr
-	mov	al,bh
-	out	dx,al
-	in	al,dx
-	cmp	al,#0x01
-	jne	nocirr
-	call	rst3d4	
-	lea	si,dsccirrus
-	lea	di,mocirrus
-	br	selmod
-rst3d4:	mov	dx,#0x3d4
-	mov	al,bl
-	xor	ah,ah
-	shl	ax,#8
-	add	ax,#0x0c
-	out	dx,ax
-	ret	
-nocirr:	call	rst3d4			! Check Everex 'clues'
-	mov	ax,#0x7000
-	xor	bx,bx
-	int	0x10
-	cmp	al,#0x70
-	jne	noevrx
-	shr	dx,#4
-	cmp	dx,#0x678
-	je	istrid
-	cmp	dx,#0x236
-	je	istrid
-	lea	si,dsceverex
-	lea	di,moeverex
-	br	selmod
-istrid:	lea	cx,ev2tri
-	jmp	cx
-noevrx:	lea	si,idgenoa		! Check Genoa 'clues'
-	xor 	ax,ax
-	seg es
-	mov	al,[0x37]
-	mov	di,ax
-	mov	cx,#0x04
-	dec	si
-	dec	di
-l1:	inc	si
-	inc	di
-	mov	al,(si)
-	test	al,al
-	jz	l2
-	seg es
-	cmp	al,(di)
-l2:	loope 	l1
-	cmp	cx,#0x00
-	jne	nogen
-	lea	si,dscgenoa
-	lea	di,mogenoa
-	br	selmod
-nogen:	cld
-	lea	si,idoakvga
-	mov	di,#0x08
-	mov	cx,#0x08
-	repe
-	cmpsb
-	jne	nooak
-	lea	si,dscoakvga
-	lea	di,mooakvga
-	br	selmod
-nooak:	cld
-	lea	si,idparadise		! Check Paradise 'clues'
-	mov	di,#0x7d
-	mov	cx,#0x04
-	repe
-	cmpsb
-	jne	nopara
-	lea	si,dscparadise
-	lea	di,moparadise
-	br	selmod
-nopara:	mov	dx,#0x3c4		! Check Trident 'clues'
-	mov	al,#0x0e
-	out	dx,al
-	inc	dx
-	in	al,dx
-	xchg	ah,al
-	xor	al,al
-	out	dx,al
-	in	al,dx
-	xchg	al,ah
-	mov	bl,al		! Strange thing ... in the book this wasn't
-	and	bl,#0x02	! necessary but it worked on my card which
-	jz	setb2		! is a trident. Without it the screen goes
-	and	al,#0xfd	! blurred ...
-	jmp	clrb2		!
-setb2:	or	al,#0x02	!
-clrb2:	out	dx,al
-	and	ah,#0x0f
-	cmp	ah,#0x02
-	jne	notrid
-ev2tri:	lea	si,dsctrident
-	lea	di,motrident
-	jmp	selmod
-notrid:	mov	dx,#0x3cd		! Check Tseng 'clues'
-	in	al,dx			! Could things be this simple ! :-)
-	mov	bl,al
-	mov	al,#0x55
-	out	dx,al
-	in	al,dx
-	mov	ah,al
-	mov	al,bl
-	out	dx,al
-	cmp	ah,#0x55
- 	jne	notsen
-	lea	si,dsctseng
-	lea	di,motseng
-	jmp	selmod
-notsen:	mov	dx,#0x3cc		! Check Video7 'clues'
-	in	al,dx
-	mov	dx,#0x3b4
-	and	al,#0x01
-	jz	even7
-	mov	dx,#0x3d4
-even7:	mov	al,#0x0c
-	out	dx,al
-	inc	dx
-	in	al,dx
-	mov	bl,al
-	mov	al,#0x55
-	out	dx,al
-	in	al,dx
-	dec	dx
-	mov	al,#0x1f
-	out	dx,al
-	inc	dx
-	in	al,dx
-	mov	bh,al
-	dec	dx
-	mov	al,#0x0c
-	out	dx,al
-	inc	dx
-	mov	al,bl
-	out	dx,al
-	mov	al,#0x55
-	xor	al,#0xea
-	cmp	al,bh
-	jne	novid7
-	lea	si,dscvideo7
-	lea	di,movideo7
-	jmp	selmod
-novid7:	lea	si,dsunknown
-	lea	di,mounknown
-selmod:	xor	cx,cx
-	mov	cl,(di)
-	mov	ax,modesave
-	cmp	ax,#ASK_VGA
-	je	askmod
-	cmp	ax,#NORMAL_VGA
-	je	askmod
-	cmp	al,cl
-	jl	gotmode
-	push	si
-	lea	si,msg4
-	call	prtstr
-	pop	si
-askmod:	push	si
-	lea	si,msg2
-	call	prtstr
-	pop	si
-	push	si
-	push	cx
-tbl:	pop	bx
-	push	bx
-	mov	al,bl
-	sub	al,cl
-	call	modepr
-	lodsw
-	xchg	al,ah
-	call	dprnt
-	xchg	ah,al
-	push	ax
-	mov	al,#0x78
-	call	prnt1
-	pop	ax
-	call	dprnt
-	push	si
-	lea	si,crlf		! print CR+LF
-	call	prtstr
-	pop	si
-	loop	tbl
-	pop	cx
-	lea	si,msg3
-	call	prtstr
-	pop	si
-	add	cl,#0x30
-	jmp	nonum
-nonumb:	call	beep
-nonum:	call	getkey
-	cmp	al,#0x30	! ascii `0'
-	jb	nonumb
-	cmp	al,#0x3a	! ascii `9'
-	jbe	number
-	cmp	al,#0x61	! ascii `a'
-	jb	nonumb
-	cmp	al,#0x7a	! ascii `z'
-	ja	nonumb
-	sub	al,#0x27
-	cmp	al,cl
-	jae	nonumb
-	sub	al,#0x30
-	jmp	gotmode
-number: cmp	al,cl
-	jae	nonumb
-	sub	al,#0x30
-gotmode:	xor	ah,ah
-	or	al,al
-	beq	vga50
-	push	ax
-	dec	ax
-	beq	vga28
-	add	di,ax
-	mov	al,(di)
-	int 	0x10
-	pop	ax
-	shl	ax,#1
-	add	si,ax
-	lodsw
-	pop	ds
-	ret
-
-! Routine to write al into a VGA-register that is
-! accessed via an index register
-!
-! dx contains the address of the index register
-! al contains the index
-! ah contains the value to write to the data register (dx + 1)
-!
-! no registers are changed
-
-outidx:	out	dx,al
-	push	ax
-	mov	al,ah
-	inc	dx
-	out	dx,al
-	dec	dx
-	pop	ax
-	ret
-inidx:	out	dx,al
-	inc	dx
-	in	al,dx
-	dec	dx
-	ret
-
-! Routine to print a decimal value on screen, the value to be
-! printed is put in al (i.e 0-255). 
-
-dprnt:	push	ax
-	push	cx
-	xor	ah,ah		! Clear ah
-	mov	cl,#0x0a
-	idiv	cl
-	cmp	al,#0x09
-	jbe	lt100
-	call	dprnt
-	jmp	skip10
-lt100:	add	al,#0x30
-	call	prnt1
-skip10:	mov	al,ah
-	add	al,#0x30
-	call	prnt1	
-	pop	cx
-	pop	ax
-	ret
-
-!
-! Routine to print the mode number key on screen. Mode numbers
-! 0-9 print the ascii values `0' to '9', 10-35 are represented by
-! the letters `a' to `z'. This routine prints some spaces around the
-! mode no.
-!
-
-modepr:	push	ax
-	cmp	al,#0x0a
-	jb	digit		! Here is no check for number > 35
-	add	al,#0x27
-digit:	add	al,#0x30
-	mov	modenr, al
-	push 	si
-	lea	si, modestring
-	call	prtstr
-	pop	si
-	pop	ax
-	ret
 
 gdt:
 	.word	0,0,0,0		! dummy
 
 	.word	0,0,0,0		! unused
 
-	.word	0x07FF		! 8Mb - limit=2047 (2048*4096=8Mb)
+	.word	0xFFFF		! 4Gb - (0x100000*0x1000 = 4Gb)
 	.word	0x0000		! base address=0
 	.word	0x9A00		! code read/exec
-	.word	0x00C0		! granularity=4096, 386
+	.word	0x00CF		! granularity=4096, 386 (+5th nibble of limit)
 
-	.word	0x07FF		! 8Mb - limit=2047 (2048*4096=8Mb)
+	.word	0xFFFF		! 4Gb - (0x100000*0x1000 = 4Gb)
 	.word	0x0000		! base address=0
 	.word	0x9200		! data read/write
-	.word	0x00C0		! granularity=4096, 386
+	.word	0x00CF		! granularity=4096, 386 (+5th nibble of limit)
 
 idt_48:
 	.word	0			! idt limit=0
@@ -940,69 +701,26 @@ gdt_48:
 	.word	0x800		! gdt limit=2048, 256 GDT entries
 	.word	512+gdt,0x9	! gdt base = 0X9xxxx
 
-msg1:		.ascii	"Press <RETURN> to see SVGA-modes available, <SPACE> to continue or wait 30 secs."
-		db	0x0d, 0x0a, 0x0a, 0x00
-msg2:		.ascii	"Mode:  COLSxROWS:"
-		db	0x0d, 0x0a, 0x0a, 0x00
-msg3:		db	0x0d, 0x0a
-		.ascii	"Choose mode by pressing the corresponding number or letter."
-crlf:		db	0x0d, 0x0a, 0x00
-msg4:		.ascii	"You passed an undefined mode number to setup. Please choose a new mode."
-		db	0x0d, 0x0a, 0x0a, 0x07, 0x00
-modestring:	.ascii	"   "
-modenr:		db	0x00	! mode number
-		.ascii	":    "
-		db	0x00
-		
-idati:		.ascii	"761295520"
-idcandt:	.byte	0xa5
-idgenoa:	.byte	0x77, 0x00, 0x99, 0x66
-idparadise:	.ascii	"VGA="
-idoakvga:	.ascii  "OAK VGA "
-idS3:		.byte	0x81, 0x82, 0x90, 0x91, 0x92, 0x93, 0x94, 0x95
-		.byte	0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa8, 0xb0
-
-! Manufacturer:	  Numofmodes+2:	Mode:
-! Number of modes is the number of chip-specific svga modes plus the extended
-! modes available on any vga (currently 2)
-
-moati:		.byte	0x06,	0x23, 0x33, 0x22, 0x21
-moahead:	.byte	0x07,	0x22, 0x23, 0x24, 0x2f, 0x34
-mocandt:	.byte	0x04,	0x60, 0x61
-mocirrus:	.byte	0x06,	0x1f, 0x20, 0x22, 0x31
-moeverex:	.byte	0x0c,	0x03, 0x04, 0x07, 0x08, 0x0a, 0x0b, 0x16, 0x18, 0x21, 0x40
-mogenoa:	.byte	0x0c,	0x58, 0x5a, 0x60, 0x61, 0x62, 0x63, 0x64, 0x72, 0x74, 0x78
-moparadise:	.byte	0x04,	0x55, 0x54
-motrident:	.byte	0x09,	0x50, 0x51, 0x52, 0x57, 0x58, 0x59, 0x5a
-motseng:	.byte	0x07,	0x26, 0x2a, 0x23, 0x24, 0x22
-movideo7:	.byte	0x08,	0x40, 0x43, 0x44, 0x41, 0x42, 0x45
-mooakvga:	.byte   0x08,   0x00, 0x07, 0x4e, 0x4f, 0x50, 0x51
-mo_S3:		.byte	0x04,	0x54, 0x55
-mounknown:	.byte	0x02
-
-!			msb = Cols lsb = Rows:
-! The first two modes are standard vga modes available on any vga.
-! mode 0 is 80x50 and mode 1 is 80x28
-
-dscati:		.word	0x5032, 0x501c, 0x8419, 0x842c, 0x641e, 0x6419
-dscahead:	.word	0x5032, 0x501c, 0x842c, 0x8419, 0x841c, 0xa032, 0x5042
-dsccandt:	.word	0x5032, 0x501c, 0x8419, 0x8432
-dsccirrus:	.word	0x5032, 0x501c, 0x8419, 0x842c, 0x841e, 0x6425
-dsceverex:	.word	0x5032, 0x501c, 0x5022, 0x503c, 0x642b, 0x644b, 0x8419, 0x842c, 0x501e, 0x641b, 0xa040, 0x841e
-dscgenoa:	.word	0x5032, 0x501c, 0x5020, 0x642a, 0x8419, 0x841d, 0x8420, 0x842c, 0x843c, 0x503c, 0x5042, 0x644b
-dscparadise:	.word	0x5032, 0x501c, 0x8419, 0x842c
-dsctrident:	.word 	0x5032, 0x501c, 0x501e, 0x502b, 0x503c, 0x8419, 0x841e, 0x842b, 0x843c
-dsctseng:	.word	0x5032, 0x501c, 0x503c, 0x6428, 0x8419, 0x841c, 0x842c
-dscvideo7:	.word	0x5032, 0x501c, 0x502b, 0x503c, 0x643c, 0x8419, 0x842c, 0x841c
-dscoakvga:	.word   0x5032, 0x501c, 0x2819, 0x5019, 0x503c, 0x843c, 0x8419, 0x842b
-dsc_S3:		.word	0x5032, 0x501c, 0x842b, 0x8419
-dsunknown:	.word	0x5032, 0x501c
-modesave:	.word	SVGA_MODE
-
-! This must be last
+!
+! Include video setup & detection code
+!
+
+#include "video.S"
+
+!
+! Setup signature -- must be last
+!
+
 setup_sig1:	.word	SIG1
 setup_sig2:	.word	SIG2
 
+!
+! After this point, there is some free space which is used by the video mode
+! handling code to store the temporary mode table (not used by the kernel).
+!
+
+modelist:
+
 .text
 endtext:
 .data
diff --git a/arch/i386/boot/tools/build.c b/arch/i386/boot/tools/build.c
index 31277a019..a7e9d063f 100644
--- a/arch/i386/boot/tools/build.c
+++ b/arch/i386/boot/tools/build.c
@@ -18,6 +18,8 @@
 
 /*
  * Changes by tytso to allow root device specification
+ * High loaded stuff by Hans Lermen & Werner Almesberger, Feb. 1996
+ * Cross compiling fixes by Gertjan van Wingerde, July 1996
  */
 
 #include <stdio.h>	/* fprintf */
@@ -30,13 +32,20 @@
 #include <fcntl.h>
 #include <linux/a.out.h>
 #include <linux/config.h>
+#include <errno.h>
 
 #define MINIX_HEADER 32
 
 #define N_MAGIC_OFFSET 1024
+#ifndef __BFD__
 static int GCC_HEADER = sizeof(struct exec);
+#endif
 
+#ifdef __BIG_KERNEL__
+#define SYS_SIZE 0xffff
+#else
 #define SYS_SIZE DEF_SYSSIZE
+#endif
 
 #define DEFAULT_MAJOR_ROOT 0
 #define DEFAULT_MINOR_ROOT 0
@@ -48,6 +57,7 @@ static int GCC_HEADER = sizeof(struct exec);
 #define STRINGIFY(x) #x
 
 typedef union {
+	int i;
 	long l;
 	short s[2];
 	char b[4];
@@ -64,6 +74,17 @@ long intel_long(long l)
 	return t.l;
 }
 
+int intel_int(int i)
+{
+	conv t;
+
+	t.b[0] = i & 0xff; i >>= 8;
+        t.b[1] = i & 0xff; i >>= 8;
+        t.b[2] = i & 0xff; i >>= 8;
+        t.b[3] = i & 0xff; i >>= 8;
+        return t.i;
+}
+
 short intel_short(short l)
 {
 	conv t;
@@ -73,7 +94,7 @@ short intel_short(short l)
 	return t.s[0];
 }
 
-void die(char * str)
+void die(const char * str)
 {
 	fprintf(stderr,"%s\n",str);
 	exit(1);
@@ -86,10 +107,12 @@ void usage(void)
 
 int main(int argc, char ** argv)
 {
-	int i,c,id, sz;
-	unsigned long sys_size;
+	int i,c,id,sz,tmp_int;
+	unsigned long sys_size, tmp_long;
 	char buf[1024];
+#ifndef __BFD__
 	struct exec *ex = (struct exec *)buf;
+#endif
 	char major_root, minor_root;
 	struct stat sb;
 	unsigned char setup_sectors;
@@ -167,8 +190,27 @@ int main(int argc, char ** argv)
 	if (((long *) buf)[7] != 0)
 		die("Illegal symbol table in 'setup'");
 	for (i=0 ; (c=read(id,buf,sizeof buf))>0 ; i+=c )
+#ifdef __BIG_KERNEL__
+	{
+		if (!i) {
+			/* Working with memcpy because of alignment constraints
+			   on Sparc - Gertjan */
+			memcpy(&tmp_long, &buf[2], sizeof(long));
+			if (tmp_long != intel_long(0x53726448) )
+				die("Wrong magic in loader header of 'setup'");
+			memcpy(&tmp_int, &buf[6], sizeof(int));
+			if (tmp_int < intel_int(0x200))
+				die("Wrong version of loader header of 'setup'");
+			buf[0x11] = 1; /* LOADED_HIGH */
+			tmp_long = intel_long(0x100000);
+			memcpy(&buf[0x14], &tmp_long, sizeof(long));  /* code32_start */
+		}
+#endif
 		if (write(1,buf,c)!=c)
 			die("Write call failed");
+#ifdef __BIG_KERNEL__
+	}
+#endif
 	if (c != 0)
 		die("read-error on 'setup'");
 	close (id);
@@ -190,6 +232,7 @@ int main(int argc, char ** argv)
 	
 	if ((id=open(argv[3],O_RDONLY,0))<0)
 		die("Unable to open 'system'");
+#ifndef __BFD__
 	if (read(id,buf,GCC_HEADER) != GCC_HEADER)
 		die("Unable to read header of 'system'");
 	if (N_MAGIC(*ex) == ZMAGIC) {
@@ -203,6 +246,14 @@ int main(int argc, char ** argv)
 		ex->a_data /1024,
 		ex->a_bss  /1024);
 	sz = N_SYMOFF(*ex) - GCC_HEADER + 4;
+#else
+	if (fstat (id, &sb)) {
+	  perror ("fstat");
+	  die ("Unable to stat 'system'");
+	}
+	sz = sb.st_size;
+	fprintf (stderr, "System is %d kB\n", sz/1024);
+#endif
 	sys_size = (sz + 15) / 16;
 	if (sys_size > SYS_SIZE)
 		die("System is too big");
diff --git a/arch/i386/boot/video.S b/arch/i386/boot/video.S
new file mode 100644
index 000000000..8507b7081
--- /dev/null
+++ b/arch/i386/boot/video.S
@@ -0,0 +1,1839 @@
+!
+!	Display adapter & video mode setup, version 2.10 (11-Nov-96)
+!
+!	Copyright (C) 1995, 1996 Martin Mares <mj@k332.feld.cvut.cz>
+!	Based on the original setup.S code (C) Linus Torvalds and Mats Anderson
+!
+
+! Enable autodetection of SVGA adapters and modes
+#define CONFIG_VIDEO_SVGA
+
+! Enable autodetection of VESA modes
+#define CONFIG_VIDEO_VESA
+
+! Enable compacting of mode table
+#define CONFIG_VIDEO_COMPACT
+
+! Retain screen contents when switching modes
+#define CONFIG_VIDEO_RETAIN
+
+! Enable local mode list
+#undef CONFIG_VIDEO_LOCAL
+
+! Force 400 scan lines for standard modes (hack to fix bad behaviour
+! of certain broken BIOS'es -- don't use unless needed)
+#undef CONFIG_VIDEO_400_HACK
+
+! A special hack allowing to force specific BIOS mode ID along with specific
+! dimensions. Especially useful for certain X-Window graphics mode hacks
+! (e.g., 800x600 modes on IBM ThinkPad).
+#undef CONFIG_VIDEO_GFX_HACK
+#define VIDEO_GFX_BIOS_AX 0x4f02	/* 800x600 on ThinkPad */
+#define VIDEO_GFX_BIOS_BX 0x0102
+#define VIDEO_GFX_DUMMY_RESOLUTION 0x6425	/* 100x37 */
+
+! This code uses an extended set of video mode numbers. These include:
+! Aliases for standard modes
+!	NORMAL_VGA (-1)
+!	EXTENDED_VGA (-2)
+!	ASK_VGA (-3)
+! Video modes numbered by menu position -- NOT RECOMMENDED because of lack
+! of compatibility when extending the table. These are between 0x00 and 0xff.
+#define VIDEO_FIRST_MENU 0x0000
+! Standard BIOS video modes (BIOS number + 0x0100)
+#define VIDEO_FIRST_BIOS 0x0100
+! VESA BIOS video modes (VESA number + 0x0200)
+#define VIDEO_FIRST_VESA 0x0200
+! Video7 special modes (BIOS number + 0x0900)
+#define VIDEO_FIRST_V7 0x0900
+! Special video modes
+#define VIDEO_FIRST_SPECIAL 0x0f00
+#define VIDEO_80x25 0x0f00
+#define VIDEO_8POINT 0x0f01
+#define VIDEO_80x43 0x0f02
+#define VIDEO_80x28 0x0f03
+#define VIDEO_CURRENT_MODE 0x0f04
+#define VIDEO_80x30 0x0f05
+#define VIDEO_80x34 0x0f06
+#define VIDEO_80x60 0x0f07
+#define VIDEO_GFX_HACK 0x0f08
+#define VIDEO_LAST_SPECIAL 0x0f09
+! Video modes given by resolution
+#define VIDEO_FIRST_RESOLUTION 0x1000
+
+! The "recalculate timings" flag
+#define VIDEO_RECALC 0x8000
+
+! Positions of various video parameters passed to the kernel
+#define PARAM_CURSOR_POS	0
+#define PARAM_VIDEO_PAGE	4
+#define PARAM_VIDEO_MODE	6
+#define PARAM_VIDEO_COLS	7
+#define PARAM_VIDEO_EGA_BX	10
+#define PARAM_VIDEO_LINES	14
+#define PARAM_HAVE_VGA		15
+#define PARAM_FONT_POINTS	16
+
+! Define DO_STORE according to CONFIG_VIDEO_RETAIN
+#ifdef CONFIG_VIDEO_RETAIN
+#define DO_STORE call store_screen
+#else
+#define DO_STORE
+#endif /* CONFIG_VIDEO_RETAIN */
+
+!
+! This is the main entry point called by setup.S
+!
+! Input:
+!	DS pointing to the bootsector
+
+video:	push	ds		! We use different segments
+	push	ds		! FS contains original DS
+	pop	fs
+	push	cs		! DS is equal to CS
+	pop	ds
+	push	cs		! ES is equal to CS
+	pop	es
+	xor	ax,ax
+	mov	gs,ax		! GS is zero
+	cld
+	call	basic_detect	! Basic adapter type testing (EGA/VGA/MDA/CGA)
+#ifdef CONFIG_VIDEO_SELECT
+	seg	fs		! User-selected video mode
+	mov	ax,[0x01fa]
+	cmp	ax,#ASK_VGA	! Bring up the menu
+	jz	vid2
+	call	mode_set	! Set the mode
+	jc	vid1
+	lea	si,badmdt	! Invalid mode ID
+	call	prtstr
+vid2:	call	mode_menu
+vid1:
+#ifdef CONFIG_VIDEO_RETAIN
+	call	restore_screen	! Restore screen contents
+#endif /* CONFIG_VIDEO_RETAIN */
+#endif /* CONFIG_VIDEO_SELECT */
+	call	mode_params	! Store mode parameters
+	pop	ds		! Restore original DS
+	ret
+
+!
+! Detect if we have CGA, MDA, EGA or VGA and pass it to the kernel.
+!
+
+basic_detect:
+	seg	fs		! Default is no VGA
+	movb	[PARAM_HAVE_VGA],#0
+
+	mov	ah,#0x12	! Check EGA/VGA
+	mov	bl,#0x10
+	int	0x10
+	seg	fs
+	mov	[PARAM_VIDEO_EGA_BX],bx	! Used for identification of EGA in the kernel
+	cmp	bl,#0x10	! No, this is a CGA/MDA/HGA card.
+	je	basret
+	incb	[adapter]
+
+	mov	ax,#0x1a00	! Check for EGA/VGA discrimination
+	int	0x10
+	cmp	al,#0x1a	! 1a means VGA, anything else EGA
+	jne	basret
+	seg	fs
+	incb	[PARAM_HAVE_VGA]	! We've detected a VGA
+	incb	[adapter]
+
+basret:	ret
+
+!
+! Store the video mode parameters for later usage by the kernel.
+! This is done by asking the BIOS except for the rows/columns
+! parameters in the default 80x25 mode -- these are set directly,
+! because some very obscure BIOSes supply insane values.
+!
+
+mode_params:
+	mov	ah,#0x03	! Read cursor position
+	xor	bh,bh
+	int	0x10
+	seg	fs
+	mov	[PARAM_CURSOR_POS],dx
+
+	mov	ah,#0x0f	! Read page/mode/width
+	int	0x10
+	seg	fs
+	mov	[PARAM_VIDEO_PAGE],bx
+	seg	fs
+	mov	[PARAM_VIDEO_MODE],ax	! Video mode and screen width
+	cmp	al,#7			! MDA/HGA => segment differs
+	jnz	mopar0
+	mov	[video_segment],#0xb000
+mopar0:	seg	gs			! Font size
+	mov	ax,[0x485]
+	seg	fs
+	mov	[PARAM_FONT_POINTS],ax	! (valid only on EGA/VGA)
+
+	mov	ax,[force_size]	! Forced size?
+	or	ax,ax
+	jz	mopar1
+	seg	fs
+	mov	[PARAM_VIDEO_COLS],ah
+	seg	fs
+	mov	[PARAM_VIDEO_LINES],al
+	ret
+
+mopar1:	mov	al,#25
+	cmpb	[adapter],#0	! If we are on CGA/MDA/HGA, the screen must
+	jz	mopar2		! have 25 lines.
+	seg	gs		! On EGA/VGA, use the EGA+ BIOS variable
+	mov	al,[0x484]	! containing maximal line number.
+	inc	al
+mopar2:	seg	fs
+	movb	[PARAM_VIDEO_LINES],al
+	ret
+
+#ifdef CONFIG_VIDEO_SELECT
+
+!
+! The video mode menu
+!
+
+mode_menu:
+	lea	si,keymsg	! "Return/Space/Timeout" message
+	call	prtstr
+	call	flush
+nokey:	call	getkt
+	cmp	al,#0x0d	! ENTER ?
+	je	listm		! yes - manual mode selection
+	cmp	al,#0x20	! SPACE ?
+	je	defmd1		! no - repeat
+	call 	beep
+	jmp	nokey
+defmd1:	ret			! No mode selected => use the 80x25 default
+
+listm:	call	mode_table	! We need a mode table to be listed
+listm0:	lea	si,name_bann	! Print adapter name
+	call	prtstr
+	mov	si,[card_name]
+	or	si,si
+	jnz	an2
+	mov	al,[adapter]
+	lea	si,old_name
+	or	al,al
+	jz	an1
+	lea	si,ega_name
+	dec	al
+	jz	an1
+	lea	si,vga_name
+	jmp	an1
+an2:	call	prtstr
+	lea	si,svga_name
+an1:	call	prtstr
+	lea	si,listhdr	! Table header
+	call	prtstr
+	mov	dl,#0x30	! DL holds mode number
+	lea	si,modelist
+lm1:	cmp	(si),#ASK_VGA	! End?
+	jz	lm2
+	mov	al,dl		! Menu selection number
+	call	prtchr
+	call	prtsp2
+	lodsw
+	call	prthw		! Mode ID
+	call	prtsp2
+	mov	al,(si+1)
+	call	prtdec		! Rows
+	mov	al,#0x78	! 'x'
+	call	prtchr
+	lodsw
+	call	prtdec		! Columns
+	mov	al,#0x0d	! New line
+	call	prtchr
+	mov	al,#0x0a
+	call	prtchr
+	inc	dl		! Next character
+	cmp	dl,#0x3a
+	jnz	lm1
+	mov	dl,#0x61
+	jmp	lm1
+
+lm2:	lea	si,prompt	! Mode prompt
+	call	prtstr
+	lea	di,edit_buf	! Editor buffer
+lm3:	call	getkey
+	cmp	al,#0x0d	! Enter?
+	jz	lment
+	cmp	al,#0x08	! Backspace?
+	jz	lmbs
+	cmp	al,#0x20	! Printable?
+	jc	lm3
+	cmp	di,#edit_buf+4	! Enough space?
+	jz	lm3
+	stosb
+	call	prtchr
+	jmp	lm3
+
+lmbs:	cmp	di,#edit_buf	! Backspace
+	jz	lm3
+	dec	di
+	mov	al,#0x08
+	call	prtchr
+	call	prtspc
+	mov	al,#0x08
+	call	prtchr
+	jmp	lm3
+
+lment:	movb	(di),#0
+	lea	si,crlft
+	call	prtstr
+	lea	si,edit_buf
+	cmpb	(si),#0		! Empty string => use default mode
+	jz	lmdef
+	cmpb	(si+1),#0	! One character => menu selection
+	jz	mnusel
+	cmp	(si),#0x6373	! "scan" => mode scanning
+	jnz	lmhx
+	cmp	(si+2),#0x6e61
+	jz	lmscan
+lmhx:	xor	bx,bx		! Else => mode ID in hex
+lmhex:	lodsb
+	or	al,al
+	jz	lmuse1
+	sub	al,#0x30
+	jc	lmbad
+	cmp	al,#10
+	jc	lmhx1
+	sub	al,#7
+	and	al,#0xdf
+	cmp	al,#10
+	jc	lmbad
+	cmp	al,#16
+	jnc	lmbad
+lmhx1:	shl	bx,#4
+	or	bl,al
+	jmp	lmhex
+lmuse1:	mov	ax,bx
+	jmp	lmuse
+
+mnusel:	lodsb			! Menu selection
+	xor	ah,ah
+	sub	al,#0x30
+	jc	lmbad
+	cmp	al,#10
+	jc	lmuse
+	cmp	al,#0x61-0x30
+	jc	lmbad
+	sub	al,#0x61-0x30-10
+	cmp	al,#36
+	jnc	lmbad
+lmuse:	call	mode_set
+	jc	lmdef
+lmbad:	lea	si,unknt
+	call	prtstr
+	br	lm2
+
+lmscan:	cmpb	[adapter],#0	! Scanning supported only on EGA/VGA
+	jz	lmbad
+	mov	[mt_end],#0	! Scanning of modes: done as new autodetection
+	movb	[scanning],#1
+	call	mode_table
+	br	listm0
+
+lmdef:	ret
+
+!
+! Additional parts of mode_set... (relative jumps, you know)
+!
+
+setv7:				! Video7 extended modes
+	DO_STORE
+	sub	bh,#VIDEO_FIRST_V7>>8
+	mov	ax,#0x6f05
+	int	0x10
+	stc
+	ret
+
+_setrec:	br	setrec	! Ugly...
+
+!
+! Aliases for backward compatibility.
+!
+
+setalias:
+	mov	ax,#VIDEO_80x25
+	inc	bx
+	jz	mode_set
+	mov	al,#VIDEO_8POINT-VIDEO_FIRST_SPECIAL
+	inc	bx
+	jnz	setbad
+
+	! Fall-thru !
+
+!
+! Setting of user mode (AX=mode ID) => CF=success
+!
+
+mode_set:
+	mov	bx,ax
+	cmp	ah,#0xff
+	jz	setalias
+	test	ah,#VIDEO_RECALC>>8
+	jnz	_setrec
+	cmp	ah,#VIDEO_FIRST_RESOLUTION>>8
+	jnc	setres
+	cmp	ah,#VIDEO_FIRST_SPECIAL>>8
+	jz	setspc
+	cmp	ah,#VIDEO_FIRST_V7>>8
+	jz	setv7
+	cmp	ah,#VIDEO_FIRST_VESA>>8
+	jnc	setvesa
+	or	ah,ah
+	jz	setmenu
+	dec	ah
+	jz	setbios
+setbad:	clc
+	movb	[do_restore],#0	! The screen needn't be restored
+	ret
+
+setvesa:
+	DO_STORE
+	sub	bh,#VIDEO_FIRST_VESA>>8
+	mov	ax,#0x4f02	! VESA BIOS mode set call
+	int	0x10
+	cmp	ax,#0x004f	! AL=4f if implemented, AH=0 if OK
+	jnz	setbad
+	stc
+	ret
+
+setbios:
+	DO_STORE
+	int	0x10		! Standard BIOS mode set call
+	push	bx
+	mov	ah,#0x0f	! Check if really set
+	int	0x10
+	pop	bx
+	cmp	al,bl
+	jnz	setbad
+	stc
+	ret
+
+setspc:	xor	bh,bh		! Set special mode
+	cmp	bl,#VIDEO_LAST_SPECIAL-VIDEO_FIRST_SPECIAL
+	jnc	setbad
+	add	bx,bx
+	.word	0xa7ff, spec_inits	! JMP [BX+spec_inits]
+
+setmenu:
+	or	al,al		! 80x25 is an exception
+	jz	set_80x25
+	push	bx		! Set mode chosen from menu
+	call	mode_table	! Build the mode table
+	pop	ax
+	shl	ax,#2
+	add	si,ax
+	cmp	si,di
+	jnc	setbad
+	mov	ax,(si)		! Fetch mode ID
+_m_s:	jmp	mode_set
+
+setres:
+	push	bx		! Set mode chosen by its resolution
+	call	mode_table
+	pop	bx
+	xchg	bh,bl
+setr1:	lodsw
+	cmp	ax,#ASK_VGA	! End of the list?
+	jz	setbad
+	lodsw
+	cmp	ax,bx
+	jnz	setr1
+	mov	ax,(si-4)	! Fetch mode ID
+	jmp	_m_s
+
+!
+! Recalculate vertical display end registers -- this fixes various
+! inconsistencies of extended modes on many adapters. Called when
+! the VIDEO_RECALC flag is set in the mode ID.
+!
+
+setrec:	sub	ah,#VIDEO_RECALC>>8	! Set the base mode
+	call	mode_set
+	jnc	rct3
+	seg	gs			! Font size in pixels
+	mov	ax,[0x485]
+	seg	gs			! Number of rows
+	mov	bl,[0x484]
+	inc	bl
+	mul	bl			! Number of visible
+	dec	ax			! scan lines - 1
+	mov	dx,#0x3d4
+	mov	bx,ax
+	mov	al,#0x12		! Lower 8 bits
+	mov	ah,bl
+	out	dx,ax
+	mov	al,#0x07		! Bits 8 and 9 in the overflow register
+	call	inidx
+	xchg	ah,al
+	and	ah,#0xbd
+	shr	bh,#1
+	jnc	rct1
+	or	ah,#0x02
+rct1:	shr	bh,#1
+	jnc	rct2
+	or	ah,#0x40
+rct2:	mov	al,#0x07
+	out	dx,ax
+	stc
+rct3:	ret
+
+!
+! Table of routines for setting of the special modes.
+!
+
+spec_inits:
+	.word	set_80x25
+	.word	set_8pixel
+	.word	set_80x43
+	.word	set_80x28
+	.word	set_current
+	.word	set_80x30
+	.word	set_80x34
+	.word	set_80x60
+	.word	set_gfx
+
+!
+! Set the 80x25 mode. If already set, do nothing.
+!
+
+set_80x25:
+	mov	[force_size],#0x5019	! Override possibly broken BIOS vars
+use_80x25:
+#ifdef CONFIG_VIDEO_400_HACK
+	mov	ax,#0x1202	! Force 400 scan lines
+	mov	bl,#0x30
+	int	0x10
+#else
+	mov	ah,#0x0f	! Get current mode ID
+	int	0x10
+	cmp	ax,#0x5007	! Mode 7 (80x25 mono) is the only one available
+	jz	st80		! on CGA/MDA/HGA and is also available on EGAM
+	cmp	ax,#0x5003	! Unknown mode => force 80x25 color
+	jnz	force3
+st80:	cmpb	[adapter],#0	! CGA/MDA/HGA => mode 3/7 is always 80x25
+	jz	set80
+	seg	gs		! This is EGA+ -- beware of 80x50 etc.
+	mov	al,[0x0484]
+	or	al,al		! Some buggy BIOS'es set 0 rows
+	jz	set80
+	cmp	al,#24		! It's hopefully correct
+	jz	set80
+#endif /* CONFIG_VIDEO_400_HACK */
+force3:	DO_STORE
+	mov	ax,#0x0003	! Forced set
+	int	0x10
+set80:	stc
+	ret
+
+!
+! Set the 80x50/80x43 8-pixel mode. Simple BIOS calls.
+!
+
+set_8pixel:
+	DO_STORE
+	call	use_80x25	! The base is 80x25
+set_8pt:
+	mov	ax,#0x1112	! Use 8x8 font
+	xor	bl,bl
+	int	0x10
+	mov	ax,#0x1200	! Use alternate print screen
+	mov	bl,#0x20
+	int	0x10
+	mov	ax,#0x1201	! Turn off cursor emulation
+	mov	bl,#0x34
+	int	0x10
+	mov	ah,#0x01	! Define cursor (scan lines 6 to 7)
+	mov	cx,#0x0607
+	int	0x10
+set_current:
+	stc
+	ret
+
+!
+! Set the 80x28 mode. This mode works on all VGA's, because it's a standard
+! 80x25 mode with 14-point fonts instead of 16-point.
+!
+
+set_80x28:
+	DO_STORE
+	call	use_80x25	! The base is 80x25
+set14:	mov	ax,#0x1111	! Use 9x14 font
+	xor	bl,bl
+	int	0x10
+	mov	ah,#0x01	! Define cursor (scan lines 11 to 12)
+	mov	cx,#0x0b0c
+	int	0x10
+	stc
+	ret
+
+!
+! Set the 80x43 mode. This mode is works on all VGA's.
+! It's a 350-scanline mode with 8-pixel font.
+!
+
+set_80x43:
+	DO_STORE
+	mov	ax,#0x1201	! Set 350 scans
+	mov	bl,#0x30
+	int	0x10
+	mov	ax,#0x0003	! Reset video mode
+	int	0x10
+	jmp	set_8pt		! Use 8-pixel font
+
+!
+! Set the 80x30 mode (all VGA's). 480 scanlines, 16-pixel font.
+!
+
+set_80x30:
+	call	use_80x25	! Start with real 80x25
+	DO_STORE
+	mov	dx,#0x3cc	! Get CRTC port
+	in	al,dx
+	mov	dl,#0xd4
+	ror	al,#1		! Mono or color?
+	jc	set48a
+	mov	dl,#0xb4
+set48a:	mov	ax,#0x0c11	! Vertical sync end (also unlocks CR0-7)
+ 	call	outidx
+	mov	ax,#0x0b06	! Vertical total
+ 	call	outidx
+	mov	ax,#0x3e07	! (Vertical) overflow
+ 	call	outidx
+	mov	ax,#0xea10	! Vertical sync start
+ 	call	outidx
+	mov	ax,#0xdf12	! Vertical display end
+	call	outidx
+	mov	ax,#0xe715	! Vertical blank start
+ 	call	outidx
+	mov	ax,#0x0416	! Vertical blank end
+ 	call	outidx
+	push	dx
+	mov	dl,#0xcc	! Misc output register (read)
+ 	in	al,dx
+ 	mov	dl,#0xc2	! (write)
+ 	and	al,#0x0d	! Preserve clock select bits and color bit
+ 	or	al,#0xe2	! Set correct sync polarity
+ 	out	dx,al
+	pop	dx
+	mov	[force_size],#0x501e
+	stc			! That's all.
+	ret
+
+!
+! Set the 80x34 mode (all VGA's). 480 scans, 14-pixel font.
+!
+
+set_80x34:
+	call	set_80x30	! Set 480 scans
+	call	set14		! And 14-pt font
+	mov	ax,#0xdb12	! VGA vertical display end
+	mov	[force_size],#0x5022
+setvde:	call	outidx
+	stc
+	ret
+
+!
+! Set the 80x60 mode (all VGA's). 480 scans, 8-pixel font.
+!
+
+set_80x60:
+	call	set_80x30	! Set 480 scans
+	call	set_8pt		! And 8-pt font
+	mov	ax,#0xdf12	! VGA vertical display end
+	mov	[force_size],#0x503c
+	jmp	setvde
+
+!
+! Special hack for ThinkPad graphics
+!
+
+set_gfx:
+#ifdef CONFIG_VIDEO_GFX_HACK
+	mov	ax,# VIDEO_GFX_BIOS_AX
+	mov	bx,# VIDEO_GFX_BIOS_BX
+	int	0x10
+	mov	[force_size],# VIDEO_GFX_DUMMY_RESOLUTION
+	stc
+#endif
+	ret
+
+#ifdef CONFIG_VIDEO_RETAIN
+
+!
+! Store screen contents to temporary buffer.
+!
+
+store_screen:
+	cmpb	[do_restore],#0		! Already stored?
+	jnz	stsr
+	testb	[loadflags],#CAN_USE_HEAP	! Have we space for storing?
+	jz	stsr
+	push	ax
+	push	bx
+	push	[force_size]		! Don't force specific size
+	mov	[force_size],#0
+	call	mode_params		! Obtain params of current mode
+	pop	[force_size]
+
+	seg	fs
+	mov	ah,[PARAM_VIDEO_LINES]
+	seg	fs
+	mov	al,[PARAM_VIDEO_COLS]
+	mov	bx,ax			! BX=dimensions
+	mul	ah
+	mov	cx,ax			! CX=number of characters to store
+	add	ax,ax			! Calculate image size
+	add	ax,#modelist+1024+4
+	cmp	ax,[heap_end_ptr]
+	jnc	sts1			! Unfortunately, out of memory
+
+	seg	fs			! Store mode params
+	mov	ax,[PARAM_CURSOR_POS]
+	lea	di,modelist+1024
+	stosw
+	mov	ax,bx
+	stosw
+
+	push	ds			! Store the screen
+	mov	ds,[video_segment]
+	xor	si,si
+	rep
+	movsw
+	pop	ds
+	incb	[do_restore]		! Screen will be restored later
+sts1:	pop	bx
+	pop	ax
+stsr:	ret
+
+!
+! Restore screen contents from temporary buffer.
+!
+
+restore_screen:
+	cmpb	[do_restore],#0		! Has the screen been stored?
+	jz	res1
+	call	mode_params		! Get parameters of current mode
+	seg	fs
+	mov	cl,[PARAM_VIDEO_LINES]
+	seg	fs
+	mov	ch,[PARAM_VIDEO_COLS]
+	lea	si,modelist+1024	! Screen buffer
+	lodsw				! Set cursor position
+	mov	dx,ax
+	cmp	dh,cl
+	jc	res2
+	mov	dh,cl
+	dec	dh
+res2:	cmp	dl,ch
+	jc	res3
+	mov	dl,ch
+	dec	dl
+res3:	mov	ah,#0x02
+	mov	bh,#0x00
+	int	0x10
+	lodsw				! Display size
+	mov	dl,ah			! DL=number of lines
+	mov	ah,#0			! BX=physical length of orig. line
+	mov	bx,ax
+	cmp	dl,cl			! Too many?
+	jc	res4
+	push	ax
+	mov	al,dl
+	sub	al,cl
+	mul	bl
+	add	si,ax
+	add	si,ax
+	pop	ax
+	mov	dl,cl
+res4:	cmp	al,ch			! Too wide?
+	jc	res5
+	mov	al,ch			! AX=width of src. line
+res5:	mov	cl,#0
+	xchg	cl,ch
+	mov	bp,cx			! BP=width of dest. line
+	push	es
+	mov	es,[video_segment]
+	xor	di,di			! Move the data
+	add	bx,bx			! Convert BX and BP to _bytes_
+	add	bp,bp
+res6:	push	si
+	push	di
+	mov	cx,ax
+	rep
+	movsw
+	pop	di
+	pop	si
+	add	di,bp
+	add	si,bx
+	dec	dl
+	jnz	res6
+	pop	es			! Done
+res1:	ret
+
+#endif /* CONFIG_VIDEO_RETAIN */
+
+!
+! Write to indexed VGA register (AL=index, AH=data, DX=index reg. port)
+!
+
+outidx:	out	dx,al
+	push	ax
+	mov	al,ah
+	inc	dx
+	out	dx,al
+	dec	dx
+	pop	ax
+	ret
+
+!
+! Build the table of video modes (stored after the setup.S code at the
+! `modelist' label. Each video mode record looks like:
+!	.word	MODE-ID		(our special mode ID (see above))
+!	.byte	rows		(number of rows)
+!	.byte	columns		(number of columns)
+! Returns address of the end of the table in DI, the end is marked
+! with a ASK_VGA ID.
+!
+
+mode_table:
+	mov	di,[mt_end]	! Already filled?
+	or	di,di
+	jnz	mtab1x
+	lea	di,modelist	! Store standard modes:
+
+	mov	eax,#VIDEO_80x25 + 0x50190000	! The 80x25 mode (ALL)
+	stosd
+	mov	al,[adapter]	! CGA/MDA/HGA -- no more modes
+	or	al,al
+	jz	mtabe
+	dec	al
+	jnz	mtabv
+	mov	eax,#VIDEO_8POINT + 0x502b0000	! The 80x43 EGA mode
+	stosd
+	jmp	mtabe
+mtab1x:	jmp	mtab1
+
+mtabv:	lea	si,vga_modes			! All modes for std VGA
+	mov	cx,#vga_modes_end-vga_modes
+	rep	! I'm unable to use movsw as I don't know how to store a half
+	movsb	! of the expression above to cx without using explicit shr.
+
+	cmpb	[scanning],#0			! Mode scan requested?
+	jz	mscan1
+	call	mode_scan
+mscan1:
+
+#ifdef CONFIG_VIDEO_LOCAL
+	call	local_modes
+#endif /* CONFIG_VIDEO_LOCAL */
+#ifdef CONFIG_VIDEO_VESA
+	call	vesa_modes			! Detect VESA VGA modes
+#endif /* CONFIG_VIDEO_VESA */
+#ifdef CONFIG_VIDEO_SVGA
+	cmpb	[scanning],#0			! Bypass when scanning
+	jnz	mscan2
+	call	svga_modes			! Detect SVGA cards & modes
+mscan2:
+#endif /* CONFIG_VIDEO_SVGA */
+
+mtabe:
+
+#ifdef CONFIG_VIDEO_COMPACT
+	lea	si,modelist	! Compact video mode list if requested.
+	mov	dx,di
+	mov	di,si
+cmt1:	cmp	si,dx		! Scan all modes
+	jz	cmt2
+	lea	bx,modelist	! Find in previous entries
+	mov	cx,(si+2)
+cmt3:	cmp	si,bx
+	jz	cmt4
+	cmp	cx,(bx+2)	! Found => don't copy this entry
+	jz	cmt5
+	add	bx,#4
+	jmp	cmt3
+
+cmt4:	movsd			! Copy entry
+	jmp	cmt1
+
+cmt5:	add	si,#4		! Skip entry
+	jmp	cmt1
+
+cmt2:
+#endif	/* CONFIG_VIDEO_COMPACT */
+
+	mov	(di),#ASK_VGA	! End marker
+	mov	[mt_end],di
+mtab1:	lea	si,modelist	! Returning: SI=mode list, DI=list end
+ret0:	ret
+
+! Modes usable on all standard VGAs
+
+vga_modes:
+	.word	VIDEO_8POINT
+	.word	0x5032		! 80x50
+	.word	VIDEO_80x43
+	.word	0x502b		! 80x43
+	.word	VIDEO_80x28
+	.word	0x501c		! 80x28
+	.word	VIDEO_80x30
+	.word	0x501e		! 80x30
+	.word	VIDEO_80x34
+	.word	0x5022		! 80x34
+	.word	VIDEO_80x60
+	.word	0x503c		! 80x60
+#ifdef CONFIG_VIDEO_GFX_HACK
+	.word	VIDEO_GFX_HACK
+	.word	VIDEO_GFX_DUMMY_RESOLUTION
+#endif
+vga_modes_end:
+
+!
+! Detect VESA modes.
+!
+
+#ifdef CONFIG_VIDEO_VESA
+
+vesa_modes:
+	cmpb	[adapter],#2	! VGA only
+	jnz	ret0
+	mov	bp,di		! BP=original mode table end
+	add	di,#0x200	! Buffer space
+	mov	ax,#0x4f00	! VESA Get card info call
+	int	#0x10
+	mov	di,bp
+	cmp	ax,#0x004f	! Successful?
+	jnz	ret0
+	cmp	(di+0x200),#0x4556
+	jnz	ret0
+	cmp	(di+0x202),#0x4153
+	jnz	ret0
+	mov	[card_name],#vesa_name	! Set name to "VESA VGA"
+	push	gs
+	lgs	si,(di+0x20e)	! GS:SI=mode list
+	mov	cx,#128		! Iteration limit
+vesa1:	seg	gs		! Get next mode in the list
+	lodsw
+	cmp	ax,#0xffff	! End of the table?
+	jz	vesar
+	cmp	ax,#0x0080	! Check validity of mode ID
+	jc	vesa2
+	or	ah,ah		! Valid ID's are 0x0000-0x007f and 0x0100-0x07ff
+	jz	vesan		! [Certain BIOSes erroneously report 0x80-0xff]
+	cmp	ax,#0x0800
+	jnc	vesae
+vesa2:	push	cx
+	mov	cx,ax		! Get mode information structure
+	mov	ax,#0x4f01
+	int	0x10
+	mov	bx,cx		! BX=mode number
+	add	bh,#VIDEO_FIRST_VESA>>8
+	pop	cx
+	cmp	ax,#0x004f
+	jnz	vesan		! Don't report errors (buggy BIOSES :-[ )
+	mov	al,(di)		! Check capabilities. We require
+	and	al,#0x19	! a color text mode.
+	cmp	al,#0x09
+	jnz	vesan
+	cmp	(di+8),#0xb800	! Standard video memory address required
+	jnz	vesan
+	testb	(di),#2		! Mode characteristics supplied?
+	mov	(di),bx		! Store mode number
+	jz	vesa3
+	xor	dx,dx
+	mov	bx,(di+0x12)	! Width
+	or	bh,bh
+	jnz	vesan
+	mov	(di+3),bl
+	mov	ax,(di+0x14)	! Height
+	or	ah,ah
+	jnz	vesan
+	mov	(di+2),al
+	mul	bl
+	cmp	ax,#8193	! Small enough for Linux console driver?
+	jnc	vesan
+	jmp	vesaok
+
+vesa3:	sub	bx,#0x8108	! This mode has no detailed info specified,
+	jc	vesan		! so it must be a standard VESA mode.
+	cmp	bx,#5
+	jnc	vesan
+	mov	ax,(bx+vesa_text_mode_table)
+	mov	(di+2),ax
+vesaok:	add	di,#4		! The mode is valid. Store it.
+vesan:	loop	vesa1		! Next mode. Limit exceeded => error
+vesae:	lea	si,vesaer
+	call	prtstr
+	mov	di,bp		! Discard already found modes.
+vesar:	pop	gs
+	ret
+
+!
+! Dimensions of standard VESA text modes
+!
+
+vesa_text_mode_table:
+	db	60, 80		! 0108
+	db	25, 132		! 0109
+	db	43, 132		! 010A
+	db	50, 132		! 010B
+	db	60, 132		! 010C
+
+#endif	/* CONFIG_VIDEO_VESA */
+
+!
+! Scan for video modes. A bit dirty, but should work.
+!
+
+mode_scan:
+	mov	cx,#0x0100	! Start with mode 0
+scm1:	mov	ah,#0		! Test the mode
+	mov	al,cl
+	int	0x10
+	mov	ah,#0x0f
+	int	0x10
+	cmp	al,cl
+	jnz	scm2		! Mode not set
+	mov	dx,#0x3c0	! Test if it's a text mode
+	mov	al,#0x10		! Mode bits
+	call	inidx
+	and	al,#0x03
+	jnz	scm2
+	mov	dl,#0xce		! Another set of mode bits
+	mov	al,#0x06
+	call	inidx
+	shr	al,#1
+	jc	scm2
+	mov	dl,#0xd4		! Cursor location
+	mov	al,#0x0f
+	call	inidx
+	or	al,al
+	jnz	scm2
+	mov	ax,cx		! OK, store the mode
+	stosw
+	seg	gs		! Number of rows
+	mov	al,[0x484]
+	inc	al
+	stosb
+	seg	gs		! Number of columns
+	mov	ax,[0x44a]
+	stosb
+scm2:	inc	cl
+	jns	scm1
+	mov	ax,#0x0003	! Return back to mode 3
+	int	0x10
+	ret
+
+tstidx:	out	dx,ax		! OUT DX,AX and inidx
+inidx:	out	dx,al		! Read from indexed VGA register
+	inc	dx		! AL=index, DX=index reg port -> AL=data
+	in	al,dx
+	dec	dx
+	ret
+
+!
+! Try to detect type of SVGA card and supply (usually approximate) video
+! mode table for it.
+!
+
+#ifdef CONFIG_VIDEO_SVGA
+
+svga_modes:
+	lea	si,svga_table	! Test all known SVGA adapters
+dosvga:	lodsw
+	mov	bp,ax		! Default mode table
+	or	ax,ax
+	jz	didsv1
+	lodsw			! Pointer to test routine
+	push	si
+	push	di
+	push	es
+	mov	bx,#0xc000
+	mov	es,bx
+	call	ax		! Call test routine
+	pop	es
+	pop	di
+	pop	si
+	or	bp,bp
+	jz	dosvga
+	mov	si,bp		! Found, copy the modes
+	mov	ah,[svga_prefix]
+cpsvga:	lodsb
+	or	al,al
+	jz	didsv
+	stosw
+	movsw
+	jmp	cpsvga
+
+didsv:	mov	[card_name],si	! Store pointer to card name
+didsv1:	ret
+
+!
+! Table of all known SVGA cards. For each card, we store a pointer to
+! a table of video modes supported by the card and a pointer to a routine
+! used for testing of presence of the card. The video mode table is always
+! followed by the name of the card or the chipset.
+!
+
+svga_table:
+	.word	ati_md, ati_test
+	.word	oak_md, oak_test
+	.word	paradise_md, paradise_test
+	.word	realtek_md, realtek_test
+	.word	s3_md, s3_test
+	.word	chips_md, chips_test
+	.word	video7_md, video7_test
+	.word	cirrus5_md, cirrus5_test
+	.word	cirrus6_md, cirrus6_test
+	.word	cirrus1_md, cirrus1_test
+	.word	ahead_md, ahead_test
+	.word	everex_md, everex_test
+	.word	genoa_md, genoa_test
+	.word	trident_md, trident_test
+	.word	tseng_md, tseng_test
+	.word	0
+
+!
+! Test routines and mode tables:
+!
+
+! S3 - The test algorithm was taken from the SuperProbe package
+! for XFree86 1.2.1. Report bugs to Christoph.Niemann@linux.org
+
+s3_test:
+	mov	cx,#0x0f35	! we store some constants in cl/ch
+	mov	dx,#0x03d4
+	movb	al,#0x38
+	call	inidx
+	mov	bh,al		! store current value of CRT-register 0x38
+	mov	ax,#0x0038
+	call	outidx		! disable writing to special regs
+	movb	al,cl		! check whether we can write special reg 0x35
+	call	inidx
+	movb	bl,al		! save the current value of CRT reg 0x35
+	andb	al,#0xf0	! clear bits 0-3
+	movb	ah,al
+	movb	al,cl		! and write it to CRT reg 0x35
+	call	outidx
+	call	inidx		! now read it back
+	andb	al,ch		! clear the upper 4 bits
+	jz	s3_2		! the first test failed. But we have a
+	movb	ah,bl		! second chance
+	mov	al,cl
+	call	outidx
+	jmp	s3_1		! do the other tests
+s3_2:	mov	ax,cx		! load ah with 0xf and al with 0x35
+	orb	ah,bl		! set the upper 4 bits of ah with the orig value
+	call	outidx		! write ...
+	call	inidx		! ... and reread 
+	andb	al,cl		! turn off the upper 4 bits
+	push	ax
+	movb	ah,bl		! restore old value in register 0x35
+	movb	al,cl
+	call	outidx
+	pop	ax
+	cmp	al,ch		! setting lower 4 bits was successful => bad
+	je	no_s3		! writing is allowed => this is not an S3
+s3_1:	mov	ax,#0x4838	! allow writing to special regs by putting
+	call	outidx		! magic number into CRT-register 0x38
+	movb	al,cl		! check whether we can write special reg 0x35
+	call	inidx
+	movb	bl,al
+	andb	al,#0xf0
+	movb	ah,al
+	movb	al,cl
+	call	outidx
+	call	inidx
+	andb	al,ch
+	jnz	no_s3		! no, we can't write => no S3
+	mov	ax,cx
+	orb	ah,bl
+	call	outidx
+	call	inidx
+	andb	al,ch
+	push	ax
+	movb	ah,bl		! restore old value in register 0x35
+	movb	al,cl
+	call	outidx
+	pop	ax
+	cmp	al,ch
+	jne	no_s31		! writing not possible => no S3
+	movb	al,#0x30
+	call	inidx		! now get the S3 id ...
+	lea	di,idS3
+	mov	cx,#0x10
+	repne
+	scasb
+	je	no_s31
+	movb	ah,bh
+	movb	al,#0x38
+	jmp	s3rest
+no_s3:	movb	al,#0x35	! restore CRT register 0x35
+	movb	ah,bl
+	call	outidx
+no_s31:	xor	bp,bp		! Detection failed
+s3rest:	movb	ah,bh
+	movb	al,#0x38	! restore old value of CRT register 0x38
+	br	outidx
+
+idS3:	.byte	0x81, 0x82, 0x90, 0x91, 0x92, 0x93, 0x94, 0x95
+	.byte	0xa0, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa8, 0xb0
+
+s3_md:	.byte	0x54, 0x2b, 0x84
+	.byte	0x55, 0x19, 0x84
+	.byte	0
+	.ascii	"S3"
+	.byte	0
+
+! ATI cards.
+
+ati_test:
+	lea 	si,idati
+	mov	di,#0x31
+	mov 	cx,#0x09
+	repe
+	cmpsb
+	je	atiok
+	xor	bp,bp
+atiok:	ret
+
+idati:	.ascii	"761295520"
+
+ati_md:	.byte	0x23, 0x19, 0x84
+	.byte	0x33, 0x2c, 0x84
+	.byte	0x22, 0x1e, 0x64
+	.byte	0x21, 0x19, 0x64
+	.byte	0x58, 0x21, 0x50
+	.byte	0x5b, 0x1e, 0x50
+	.byte	0
+	.ascii	"ATI"
+	.byte	0
+
+! AHEAD
+
+ahead_test:
+	mov	ax,#0x200f
+	mov	dx,#0x3ce
+	out	dx,ax
+	inc	dx
+	in	al,dx
+	cmp	al,#0x20
+	je	isahed
+	cmp	al,#0x21
+	je	isahed
+	xor	bp,bp
+isahed:	ret
+
+ahead_md:
+	.byte	0x22, 0x2c, 0x84
+	.byte	0x23, 0x19, 0x84
+	.byte	0x24, 0x1c, 0x84
+	.byte	0x2f, 0x32, 0xa0
+	.byte	0x32, 0x22, 0x50
+	.byte	0x34, 0x42, 0x50
+	.byte	0
+	.ascii	"Ahead"
+	.byte	0
+
+! Chips & Tech.
+
+chips_test:
+	mov	dx,#0x3c3
+	in	al,dx
+	or	al,#0x10
+	out	dx,al
+	mov	dx,#0x104		
+	in	al,dx
+	mov	bl,al
+	mov	dx,#0x3c3
+	in	al,dx
+	and	al,#0xef
+	out	dx,al
+	cmp	bl,#0xa5
+	je	cantok
+	xor	bp,bp
+cantok:	ret
+
+chips_md:
+	.byte	0x60, 0x19, 0x84
+	.byte	0x61, 0x32, 0x84
+	.byte	0
+	.ascii	"Chips & Technologies"
+	.byte	0
+
+! Cirrus Logic 5X0
+
+cirrus1_test:
+	mov	dx,#0x3d4
+	mov	al,#0x0c
+	out	dx,al
+	inc	dx
+	in	al,dx
+	mov	bl,al
+	xor	al,al
+	out	dx,al
+	dec	dx
+	mov	al,#0x1f
+	out	dx,al
+	inc	dx
+	in	al,dx
+	mov	bh,al
+	xor	ah,ah
+	shl	al,#4
+	mov	cx,ax
+	mov	al,bh
+	shr	al,#4
+	add	cx,ax
+	shl	cx,#8
+	add	cx,#6
+	mov	ax,cx
+	mov	dx,#0x3c4
+	out	dx,ax
+	inc	dx
+	in	al,dx
+	and	al,al
+	jnz	nocirr
+	mov	al,bh
+	out	dx,al
+	in	al,dx
+	cmp	al,#0x01
+	je	iscirr
+nocirr:	xor	bp,bp
+iscirr: mov	dx,#0x3d4
+	mov	al,bl
+	xor	ah,ah
+	shl	ax,#8
+	add	ax,#0x0c
+	out	dx,ax
+	ret
+
+cirrus1_md:
+	.byte	0x1f, 0x19, 0x84
+	.byte	0x20, 0x2c, 0x84
+	.byte	0x22, 0x1e, 0x84
+	.byte	0x31, 0x25, 0x64
+	.byte	0
+	.ascii	"Cirrus Logic 5X0"
+	.byte	0
+
+! Cirrus Logic 54XX
+
+cirrus5_test:
+	mov	dx,#0x3c4
+	mov	al,#6
+	call	inidx
+	mov	bl,al			! BL=backup
+	mov	ax,#6
+	call	tstidx
+	cmp	al,#0x0f
+	jne	c5fail
+	mov	ax,#0x1206
+	call	tstidx
+	cmp	al,#0x12
+	jne	c5fail
+	mov	al,#0x1e
+	call	inidx
+	mov	bh,al
+	mov	ah,bh
+	and	ah,#0xc0
+	mov	al,#0x1e
+	call	tstidx
+	and	al,#0x3f
+	jne	c5xx
+	mov	al,#0x1e
+	mov	ah,bh
+	or	ah,#0x3f
+	call	tstidx
+	xor	al,#0x3f
+	and	al,#0x3f
+c5xx:	pushf
+	mov	al,#0x1e
+	mov	ah,bh
+	out	dx,ax
+	popf
+	je	c5done
+c5fail:	xor	bp,bp
+c5done:	mov	al,#6
+	mov	ah,bl
+	out	dx,ax
+	ret
+
+cirrus5_md:
+	.byte	0x14, 0x19, 0x84
+	.byte	0x54, 0x2b, 0x84
+	.byte	0
+	.ascii	"Cirrus Logic 54XX"
+	.byte	0
+
+! Cirrus Logic 64XX -- no known extra modes, but must be identified, because
+! it's misidentified by the Ahead test.
+
+cirrus6_test:
+	mov	dx,#0x3ce
+	mov	al,#0x0a
+	call	inidx
+	mov	bl,al			! BL=backup
+	mov	ax,#0xce0a
+	call	tstidx
+	or	al,al
+	jne	c2fail
+	mov	ax,#0xec0a
+	call	tstidx
+	cmp	al,#0x01
+	jne	c2fail
+	mov	al,#0xaa
+	call	inidx	! 4X, 5X, 7X and 8X are valid 64XX chip ID's
+	shr	al,#4
+	sub	al,#4
+	jz	c6done
+	dec	al
+	jz	c6done
+	sub	al,#2
+	jz	c6done
+	dec	al
+	jz	c6done
+c2fail:	xor	bp,bp
+c6done:	mov	al,#0x0a
+	mov	ah,bl
+	out	dx,ax
+	ret
+
+cirrus6_md:
+	.byte	0
+	.ascii	"Cirrus Logic 64XX"
+	.byte	0
+
+! Everex / Trident
+
+everex_test:
+	mov	ax,#0x7000
+	xor	bx,bx
+	int	0x10
+	cmp	al,#0x70
+	jne	noevrx
+	shr	dx,#4
+	cmp	dx,#0x678
+	je	evtrid
+	cmp	dx,#0x236
+	jne	evrxok
+evtrid:	lea	bp,trident_md
+evrxok:	ret
+
+noevrx:	xor	bp,bp
+	ret
+
+everex_md:
+	.byte	0x03, 0x22, 0x50
+	.byte	0x04, 0x3c, 0x50
+	.byte	0x07, 0x2b, 0x64
+	.byte	0x08, 0x4b, 0x64
+	.byte	0x0a, 0x19, 0x84
+	.byte	0x0b, 0x2c, 0x84
+	.byte	0x16, 0x1e, 0x50
+	.byte	0x18, 0x1b, 0x64
+	.byte	0x21, 0x40, 0xa0
+	.byte	0x40, 0x1e, 0x84
+	.byte	0
+	.ascii	"Everex/Trident"
+	.byte	0
+
+! Genoa.
+
+genoa_test:
+	lea	si,idgenoa		! Check Genoa 'clues'
+	xor 	ax,ax
+	seg es
+	mov	al,[0x37]
+	mov	di,ax
+	mov	cx,#0x04
+	dec	si
+	dec	di
+l1:	inc	si
+	inc	di
+	mov	al,(si)
+	test	al,al
+	jz	l2
+	seg es
+	cmp	al,(di)
+l2:	loope 	l1
+	or	cx,cx
+	je	isgen
+	xor	bp,bp
+isgen:	ret
+
+idgenoa: .byte	0x77, 0x00, 0x99, 0x66
+
+genoa_md:
+	.byte	0x58, 0x20, 0x50
+	.byte	0x5a, 0x2a, 0x64
+	.byte	0x60, 0x19, 0x84
+	.byte	0x61, 0x1d, 0x84
+	.byte	0x62, 0x20, 0x84
+	.byte	0x63, 0x2c, 0x84
+	.byte	0x64, 0x3c, 0x84
+	.byte	0x6b, 0x4f, 0x64
+	.byte	0x72, 0x3c, 0x50
+	.byte	0x74, 0x42, 0x50
+	.byte	0x78, 0x4b, 0x64
+	.byte	0
+	.ascii	"Genoa"
+	.byte	0
+
+! OAK
+
+oak_test:
+	lea	si,idoakvga
+	mov	di,#0x08
+	mov	cx,#0x08
+	repe
+	cmpsb
+	je	isoak
+	xor	bp,bp
+isoak:	ret
+
+idoakvga: .ascii  "OAK VGA "
+
+oak_md: .byte	0x4e, 0x3c, 0x50
+	.byte	0x4f, 0x3c, 0x84
+	.byte	0x50, 0x19, 0x84
+	.byte	0x51, 0x2b, 0x84
+	.byte	0
+	.ascii	"OAK"
+	.byte	0
+
+! WD Paradise.
+
+paradise_test:
+	lea	si,idparadise
+	mov	di,#0x7d
+	mov	cx,#0x04
+	repe
+	cmpsb
+	je	ispara
+	xor	bp,bp
+ispara:	ret
+
+idparadise:	.ascii	"VGA="
+
+paradise_md:
+	.byte	0x41, 0x22, 0x50
+	.byte	0x47, 0x1c, 0x84
+	.byte	0x55, 0x19, 0x84
+	.byte	0x54, 0x2c, 0x84
+	.byte	0
+	.ascii	"Paradise"
+	.byte	0
+
+! Trident.
+
+trident_test:
+	mov	dx,#0x3c4
+	mov	al,#0x0e
+	out	dx,al
+	inc	dx
+	in	al,dx
+	xchg	ah,al
+	xor	al,al
+	out	dx,al
+	in	al,dx
+	xchg	al,ah
+	mov	bl,al		! Strange thing ... in the book this wasn't
+	and	bl,#0x02	! necessary but it worked on my card which
+	jz	setb2		! is a trident. Without it the screen goes
+	and	al,#0xfd	! blurred ...
+	jmp	clrb2		!
+setb2:	or	al,#0x02	!
+clrb2:	out	dx,al
+	and	ah,#0x0f
+	cmp	ah,#0x02
+	je	istrid
+	xor	bp,bp
+istrid:	ret
+
+trident_md:
+	.byte	0x50, 0x1e, 0x50
+	.byte	0x51, 0x2b, 0x50
+	.byte	0x52, 0x3c, 0x50
+	.byte	0x57, 0x19, 0x84
+	.byte	0x58, 0x1e, 0x84
+	.byte	0x59, 0x2b, 0x84
+	.byte	0x5a, 0x3c, 0x84
+	.byte	0
+	.ascii	"Trident"
+	.byte	0
+
+! Tseng.
+
+tseng_test:
+	mov	dx,#0x3cd
+	in	al,dx			! Could things be this simple ! :-)
+	mov	bl,al
+	mov	al,#0x55
+	out	dx,al
+	in	al,dx
+	mov	ah,al
+	mov	al,bl
+	out	dx,al
+	cmp	ah,#0x55
+ 	je	istsen
+isnot:	xor	bp,bp
+istsen:	ret
+
+tseng_md:
+	.byte	0x26, 0x3c, 0x50
+	.byte	0x2a, 0x28, 0x64
+	.byte	0x23, 0x19, 0x84
+	.byte	0x24, 0x1c, 0x84
+	.byte	0x22, 0x2c, 0x84
+	.byte	0x21, 0x3c, 0x84
+	.byte	0
+	.ascii	"Tseng"
+	.byte	0
+
+! Video7.
+
+video7_test:
+	mov	dx,#0x3cc
+	in	al,dx
+	mov	dx,#0x3b4
+	and	al,#0x01
+	jz	even7
+	mov	dx,#0x3d4
+even7:	mov	al,#0x0c
+	out	dx,al
+	inc	dx
+	in	al,dx
+	mov	bl,al
+	mov	al,#0x55
+	out	dx,al
+	in	al,dx
+	dec	dx
+	mov	al,#0x1f
+	out	dx,al
+	inc	dx
+	in	al,dx
+	mov	bh,al
+	dec	dx
+	mov	al,#0x0c
+	out	dx,al
+	inc	dx
+	mov	al,bl
+	out	dx,al
+	mov	al,#0x55
+	xor	al,#0xea
+	cmp	al,bh
+	jne	isnot
+	movb	[svga_prefix],#VIDEO_FIRST_V7>>8	! Use special mode switching
+	ret
+
+video7_md:
+	.byte	0x40, 0x2b, 0x50
+	.byte	0x43, 0x3c, 0x50
+	.byte	0x44, 0x3c, 0x64
+	.byte	0x41, 0x19, 0x84
+	.byte	0x42, 0x2c, 0x84
+	.byte	0x45, 0x1c, 0x84
+	.byte	0
+	.ascii	"Video 7"
+	.byte	0
+
+! Realtek VGA
+
+realtek_test:
+	lea	si,idrtvga
+	mov	di,#0x45
+	mov	cx,#0x0b
+	repe
+	cmpsb
+	je	isrt
+	xor	bp,bp
+isrt:	ret
+
+idrtvga:	.ascii	"REALTEK VGA"
+
+realtek_md:
+	.byte	0x1a, 0x3c, 0x50
+	.byte	0x1b, 0x19, 0x84
+	.byte	0x1c, 0x1e, 0x84
+	.byte	0x1d, 0x2b, 0x84
+	.byte	0x1e, 0x3c, 0x84
+	.byte	0
+	.ascii	"REALTEK"
+	.byte	0
+
+#endif	/* CONFIG_VIDEO_SVGA */
+
+!
+! User-defined local mode table (VGA only)
+!
+
+#ifdef CONFIG_VIDEO_LOCAL
+
+local_modes:
+	lea	si,local_mode_table
+locm1:	lodsw
+	or	ax,ax
+	jz	locm2
+	stosw
+	movsw
+	jmp	locm1
+locm2:	ret
+
+! This is the table of local video modes which can be supplied manually
+! by the user. Each entry consists of mode ID (word) and dimensions
+! (byte for column count and another byte for row count). These modes
+! are placed before all SVGA and VESA modes and override them if table
+! compacting is enabled. The table must end with a zero word followed
+! by NUL-terminated video adapter name.
+
+local_mode_table:
+	.word	0x0100		! Example: 40x25
+	.byte	25,40
+	.word	0
+	.ascii	"Local"
+	.byte	0
+
+#endif	/* CONFIG_VIDEO_LOCAL */
+
+!
+! Read a key and return the ASCII code in al, scan code in ah
+!
+
+getkey:	xor	ah,ah
+	int	0x16
+	ret
+
+!
+! Read a key with a timeout of 30 seconds. The hardware clock is used to get
+! the time.
+!
+
+getkt:	call	gettime
+	add	al,#30		! Wait 30 seconds
+	cmp	al,#60
+	jl	lminute
+	sub	al,#60
+lminute:
+	mov	cl,al
+again:	mov	ah,#0x01
+	int	0x16
+	jnz	getkey		! key pressed, so get it
+	call	gettime
+	cmp	al,cl
+	jne	again
+	mov	al,#0x20	! timeout, return default char `space'
+	ret
+
+!
+! Flush the keyboard buffer
+!
+
+flush:	mov	ah,#0x01
+	int	0x16
+	jz	empty
+	xor	ah,ah
+	int	0x16
+	jmp	flush
+empty:	ret
+
+!
+! Print hexadecimal number.
+!
+
+prthw:	push	ax
+	mov	al,ah
+	call	prthb
+	pop	ax
+prthb:	push	ax
+	shr	al,#4
+	call	prthn
+	pop	ax
+	and	al,#0x0f
+prthn:	cmp	al,#0x0a
+	jc	prth1
+	add	al,#0x07
+prth1:	add	al,#0x30
+	br	prtchr
+
+!
+! Print decimal number (AL).
+!
+
+prtdec:	push	ax
+	push	cx
+	xor	ah,ah		! Clear ah
+	mov	cl,#0x0a
+	idiv	cl
+	cmp	al,#0x09
+	jbe	lt100
+	call	prtdec
+	jmp	skip10
+lt100:	add	al,#0x30
+	call	prtchr
+skip10:	mov	al,ah
+	add	al,#0x30
+	call	prtchr	
+	pop	cx
+	pop	ax
+	ret
+
+!
+! VIDEO_SELECT-only variables
+!
+
+mt_end:		.word	0	! End of video mode table if built
+edit_buf:	.space	6	! Line editor buffer
+card_name:	.word	0	! Pointer to adapter name
+scanning:	.byte	0	! Performing mode scan
+do_restore:	.byte	0	! Screen contents altered during mode change
+svga_prefix:	.byte	VIDEO_FIRST_BIOS>>8	! Default prefix for BIOS modes
+
+!
+! Messages:
+!
+
+keymsg:		.ascii	"Press <RETURN> to see video modes available, "
+		.ascii	"<SPACE> to continue or wait 30 secs"
+		db	0x0d, 0x0a, 0
+listhdr:	db	0x0d, 0x0a
+		.ascii	"Mode:    COLSxROWS:"
+crlft:		db	0x0d, 0x0a, 0
+prompt:		db	0x0d, 0x0a
+		.ascii	"Enter mode number: "
+		db	0
+unknt:		.ascii	"Unknown mode ID. Try again."
+		db	0
+badmdt:		.ascii	"You passed an undefined mode number to setup."
+		db	0x0d, 0x0a, 0
+vesaer:		.ascii	"Error: Scanning of VESA modes failed. Please "
+		.ascii	"report to <mj@k332.feld.cvut.cz>."
+		db	0x0d, 0x0a, 0
+old_name:	.ascii	"CGA/MDA/HGA"
+		db	0
+ega_name:	.ascii	"EGA"
+		db	0
+svga_name:	.ascii	" "
+vga_name:	.ascii	"VGA"
+		db	0
+vesa_name:	.ascii	"VESA"
+		db	0
+name_bann:	.ascii	"Video adapter: "
+		db	0
+
+#endif /* CONFIG_VIDEO_SELECT */
+
+!
+! Other variables:
+!
+
+adapter:	.byte	0	! Video adapter: 0=CGA/MDA/HGA,1=EGA,2=VGA
+video_segment:	.word	0xb800	! Video memory segment
+force_size:	.word	0	! Use this size instead of the one in BIOS vars
diff --git a/arch/i386/config.in b/arch/i386/config.in
index bfbf42acb..a5b72a4d0 100644
--- a/arch/i386/config.in
+++ b/arch/i386/config.in
@@ -2,293 +2,113 @@
 # For a description of the syntax of this configuration file,
 # see the Configure script.
 #
+mainmenu_name "Linux Kernel Configuration"
 
-comment 'General setup'
-
-bool 'Kernel math emulation' CONFIG_MATH_EMULATION n
-bool 'Normal floppy disk support' CONFIG_BLK_DEV_FD y
-bool 'Normal (MFM/RLL) disk and IDE disk/cdrom support' CONFIG_ST506 y
-if [ "$CONFIG_ST506" = "y" ]; then
-  comment 'Please see drivers/block/README.ide for help/info on IDE drives'
-  bool '   Use old disk-only driver for primary i/f' CONFIG_BLK_DEV_HD n
-  if [ "$CONFIG_BLK_DEV_HD" = "y" ]; then
-    bool '   Include new IDE driver for secondary i/f support' CONFIG_BLK_DEV_IDE n
-  else
-    bool '   Use new IDE driver for primary/secondary i/f' CONFIG_BLK_DEV_IDE y
-  fi
-  if [ "$CONFIG_BLK_DEV_IDE" = "y" ]; then
-    bool '   Include support for IDE/ATAPI CDROMs' CONFIG_BLK_DEV_IDECD n
-  fi
-fi
-
-bool 'XT harddisk support' CONFIG_BLK_DEV_XD n
-bool 'Networking support' CONFIG_NET y
-bool 'Limit memory to low 16MB' CONFIG_MAX_16M n
-bool 'PCI bios support' CONFIG_PCI y
-if [ "$CONFIG_PCI" = "y" ]; then
- bool '   PCI bridge optimisation (experimental)' CONFIG_PCI_OPTIMIZE y
-fi
-bool 'System V IPC' CONFIG_SYSVIPC y
-bool 'Kernel support for ELF binaries' CONFIG_BINFMT_ELF y
-#bool 'Use -mpentium flag for Pentium-specific optimizations' CONFIG_M586 n
-#if [ "$CONFIG_M586" = "n" ]; then
-bool 'Use -m486 flag for 486-specific optimizations' CONFIG_M486 y
-#fi
+mainmenu_option next_comment
+comment 'Code maturity level options'
+bool 'Prompt for development and/or incomplete code/drivers' CONFIG_EXPERIMENTAL
+endmenu
 
+mainmenu_option next_comment
 comment 'Loadable module support'
-bool 'Set version information on all symbols for modules' CONFIG_MODVERSIONS n
-
-if [ "$CONFIG_NET" = "y" ]; then
-comment 'Networking options'
-bool 'TCP/IP networking' CONFIG_INET y
-if [ "$CONFIG_INET" = "y" ]; then
-bool 'IP: forwarding/gatewaying' CONFIG_IP_FORWARD n
-bool 'IP: multicasting' CONFIG_IP_MULTICAST n
-bool 'IP: firewalling' CONFIG_IP_FIREWALL n
-bool 'IP: accounting' CONFIG_IP_ACCT n
-bool 'IP: tunneling' CONFIG_NET_IPIP n
-if [ "$CONFIG_IP_FORWARD" = "y" -a "$CONFIG_IP_FIREWALL" = "y" ]; then
-  bool 'IP: firewall packet logging' CONFIG_IP_FIREWALL_VERBOSE y
-  bool 'IP: masquerading (ALPHA)' CONFIG_IP_MASQUERADE n
-fi
-comment '(it is safe to leave these untouched)'
-bool 'IP: PC/TCP compatibility mode' CONFIG_INET_PCTCP n
-bool 'IP: Reverse ARP' CONFIG_INET_RARP n
-bool 'IP: Assume subnets are local' CONFIG_INET_SNARL y
-bool 'IP: Disable NAGLE algorithm (normally enabled)' CONFIG_TCP_NAGLE_OFF n
-bool 'IP: Drop source routed frames' CONFIG_IP_NOSR y
-fi
-bool 'The IPX protocol' CONFIG_IPX n
-bool 'Appletalk DDP' CONFIG_ATALK n
-bool 'Amateur Radio AX.25 Level 2' CONFIG_AX25 n
-if [ "$CONFIG_AX25" = "y" ]; then
-  bool 'Amateur Radio NET/ROM' CONFIG_NETROM n
-fi
+bool 'Enable loadable module support' CONFIG_MODULES
+if [ "$CONFIG_MODULES" = "y" ]; then
+  bool 'Set version information on all symbols for modules' CONFIG_MODVERSIONS
+  bool 'Kernel daemon support (e.g. autoload of modules)' CONFIG_KERNELD
 fi
+endmenu
 
-comment 'SCSI support'
-
-bool 'SCSI support?' CONFIG_SCSI y
-
-if [ "$CONFIG_SCSI" = "n" ]; then
-
-comment 'Skipping SCSI configuration options...'
-
-else
-
-comment 'SCSI support type (disk, tape, CDrom)'
-
-bool 'SCSI disk support' CONFIG_BLK_DEV_SD y
-bool 'SCSI tape support' CONFIG_CHR_DEV_ST n
-bool 'SCSI CDROM support' CONFIG_BLK_DEV_SR y
-bool 'SCSI generic support' CONFIG_CHR_DEV_SG n
-
-comment 'Some SCSI devices (e.g. CD jukebox) support multiple LUNs'
-
-bool 'Probe all LUNs on each SCSI device' CONFIG_SCSI_MULTI_LUN n
-
-comment 'SCSI low-level drivers'
+mainmenu_option next_comment
+comment 'General setup'
 
-bool 'Adaptec AHA152X support' CONFIG_SCSI_AHA152X y
-bool 'Adaptec AHA1542 support' CONFIG_SCSI_AHA1542 n
-bool 'Adaptec AHA1740 support' CONFIG_SCSI_AHA1740 y
-bool 'Adaptec AHA274X/284X support' CONFIG_SCSI_AHA274X n
-bool 'BusLogic SCSI support' CONFIG_SCSI_BUSLOGIC n
-bool 'EATA-DMA (DPT,NEC&ATT for ISA,EISA,PCI) support' CONFIG_SCSI_EATA_DMA n
-bool 'UltraStor 14F/34F support' CONFIG_SCSI_U14_34F n
-bool 'Future Domain 16xx SCSI support' CONFIG_SCSI_FUTURE_DOMAIN n
-bool 'Generic NCR5380 SCSI support' CONFIG_SCSI_GENERIC_NCR5380 n
+bool 'Kernel math emulation' CONFIG_MATH_EMULATION
+bool 'Networking support' CONFIG_NET
+bool 'Limit memory to low 16MB' CONFIG_MAX_16M
+bool 'PCI bios support' CONFIG_PCI
 if [ "$CONFIG_PCI" = "y" ]; then
-  bool 'NCR53c7,8xx SCSI support'  CONFIG_SCSI_NCR53C7xx n
+  if [ "$CONFIG_EXPERIMENTAL" = "y" ]; then
+    bool '   PCI bridge optimization (experimental)' CONFIG_PCI_OPTIMIZE
+  fi
 fi
-bool 'Always IN2000 SCSI support (test release)' CONFIG_SCSI_IN2000 n
-bool 'PAS16 SCSI support' CONFIG_SCSI_PAS16 n
-bool 'QLOGIC SCSI support' CONFIG_SCSI_QLOGIC n
-bool 'Seagate ST-02 and Future Domain TMC-8xx SCSI support' CONFIG_SCSI_SEAGATE n
-bool 'Trantor T128/T128F/T228 SCSI support' CONFIG_SCSI_T128 n
-bool 'UltraStor SCSI support' CONFIG_SCSI_ULTRASTOR n
-bool '7000FASST SCSI support' CONFIG_SCSI_7000FASST n
-#bool 'EATA ISA/EISA (DPT PM2011/021/012/022/122/322) support' CONFIG_SCSI_EATA n
-#bool 'SCSI debugging host adapter' CONFIG_SCSI_DEBUG n
+bool 'System V IPC' CONFIG_SYSVIPC
+tristate 'Kernel support for a.out binaries' CONFIG_BINFMT_AOUT
+tristate 'Kernel support for ELF binaries' CONFIG_BINFMT_ELF
+if [ "$CONFIG_EXPERIMENTAL" = "y" ]; then
+  tristate 'Kernel support for JAVA binaries' CONFIG_BINFMT_JAVA
 fi
 
+choice 'Processor type' \
+	"386		CONFIG_M386	\
+	 486		CONFIG_M486	\
+	 Pentium	CONFIG_M586	\
+	 PPro		CONFIG_M686" Pentium
+bool 'Video mode selection support' CONFIG_VIDEO_SELECT
+endmenu
 
-if [ "$CONFIG_NET" = "y" ]; then
+source drivers/block/Config.in
 
-comment 'Network device support'
+if [ "$CONFIG_NET" = "y" ]; then
+  source net/Config.in
+fi
 
-bool 'Network device support?' CONFIG_NETDEVICES y
-if [ "$CONFIG_NETDEVICES" = "n" ]; then
+mainmenu_option next_comment
+comment 'SCSI support'
 
-comment 'Skipping network driver configuration options...'
+tristate 'SCSI support' CONFIG_SCSI
 
-else
-bool 'Dummy net driver support' CONFIG_DUMMY y
-bool 'SLIP (serial line) support' CONFIG_SLIP n
-if [ "$CONFIG_SLIP" = "y" ]; then
-  bool ' CSLIP compressed headers' CONFIG_SLIP_COMPRESSED y
-  bool ' 16 channels instead of 4' SL_SLIP_LOTS n
-fi
-bool 'PPP (point-to-point) support' CONFIG_PPP n
-if [ "$CONFIG_PPP" = "y" ]; then
-  bool ' 16 channels instead of 4' CONFIG_PPP_LOTS n
-fi
-if [ "$CONFIG_AX25" = "y" ]; then
-	bool 'Z8530 SCC kiss emulation driver for AX.25' CONFIG_SCC y
-fi
-bool 'PLIP (parallel port) support' CONFIG_PLIP n
-bool 'EQL (serial line load balancing) support' CONFIG_EQUALIZER n
-bool 'Do you want to be offered ALPHA test drivers' CONFIG_NET_ALPHA n
-bool 'Western Digital/SMC cards' CONFIG_NET_VENDOR_SMC n
-if [ "$CONFIG_NET_VENDOR_SMC" = "y" ]; then
-	bool 'WD80*3 support' CONFIG_WD80x3 n
-	bool 'SMC Ultra support' CONFIG_ULTRA n
-fi
-bool 'AMD LANCE and PCnet (AT1500 and NE2100) support' CONFIG_LANCE n
-bool '3COM cards' CONFIG_NET_VENDOR_3COM y
-if [ "$CONFIG_NET_VENDOR_3COM" = "y" ]; then
-	bool '3c501 support' CONFIG_EL1 n
-	bool '3c503 support' CONFIG_EL2 n
-	if [ "$CONFIG_NET_ALPHA" = "y" ]; then
-		bool '3c505 support' CONFIG_ELPLUS n
-		bool '3c507 support' CONFIG_EL16 n
-	fi
-	bool '3c509/3c579 support' CONFIG_EL3 y
-fi
-bool 'Other ISA cards' CONFIG_NET_ISA n
-if [ "$CONFIG_NET_ISA" = "y" ]; then
-	bool 'Arcnet support' CONFIG_ARCNET n
-	bool 'Cabletron E21xx support' CONFIG_E2100 n
-	bool 'DEPCA support' CONFIG_DEPCA n
-	bool 'EtherWorks 3 support' CONFIG_EWRK3 n
-	if [ "$CONFIG_NET_ALPHA" = "y" ]; then
-		bool 'AT1700 support' CONFIG_AT1700 n
-#		bool 'EtherExpressPro support' CONFIG_EEXPRESS_PRO n
-		bool 'EtherExpress support' CONFIG_EEXPRESS n
-		bool 'NI5210 support' CONFIG_NI52 n
-		bool 'NI6510 support' CONFIG_NI65 n
-		bool 'WaveLAN support' CONFIG_WAVELAN n
-	fi
-	bool 'HP PCLAN+ (27247B and 27252A) support' CONFIG_HPLAN_PLUS n
-	bool 'HP PCLAN (27245 and other 27xxx series) support' CONFIG_HPLAN n
-	bool 'NE2000/NE1000 support' CONFIG_NE2000 y
-	if [ "$CONFIG_AX25" = "y" ]; then
-		bool 'Ottawa PI and PI/2 support' CONFIG_PI y
-	fi
-	bool 'SK_G16 support' CONFIG_SK_G16 n
-fi
-bool 'EISA, VLB, PCI and on board controllers' CONFIG_NET_EISA n
-if [ "$CONFIG_NET_EISA" = "y" ]; then
-	if [ "$CONFIG_NET_ALPHA" = "y" ]; then
-		bool 'Ansel Communications EISA 3200 support' CONFIG_AC3200 n
-	fi
-	bool 'Apricot Xen-II on board ethernet' CONFIG_APRICOT n
-	bool 'DE425, DE434, DE435 support' CONFIG_DE4X5 n
-#	bool 'DEC 21040 PCI support' CONFIG_DEC_ELCP n
-#	bool 'LPL T100V 100Mbs support' CONFIG_LPL_T100 n
-#	bool 'PCnet32 (32 bit VLB and PCI LANCE) support' CONFIG_PCNET32 n
-	bool 'Zenith Z-Note support' CONFIG_ZNET y
-fi
-bool 'Pocket and portable adaptors' CONFIG_NET_POCKET n
-if [ "$CONFIG_NET_POCKET" = "y" ]; then
-	bool 'AT-LAN-TEC/RealTek pocket adaptor support' CONFIG_ATP n
-	bool 'D-Link DE600 pocket adaptor support' CONFIG_DE600 n
-	bool 'D-Link DE620 pocket adaptor support' CONFIG_DE620 n
-#	bool 'Silicom pocket adaptor support' CONFIG_SILICOM_PEA n
-#	bool 'WaveLAN PCMCIA support' CONFIG_WaveLAN n
-#	bool '3 Com 3c589 PCMCIA support' CONFIG_3C589 n
-fi
-bool 'Token Ring driver support' CONFIG_TR n
-if [ "$CONFIG_TR" = "y" ]; then
-	bool 'IBM Tropic chipset based adaptor support' CONFIG_IBMTR y
-fi
-fi
+if [ "$CONFIG_SCSI" != "n" ]; then
+  source drivers/scsi/Config.in
 fi
+endmenu
 
-comment 'CD-ROM drivers (not for SCSI or IDE/ATAPI drives)'
+if [ "$CONFIG_NET" = "y" ]; then
+  mainmenu_option next_comment
+  comment 'Network device support'
 
-bool 'Sony CDU31A/CDU33A CDROM driver support' CONFIG_CDU31A n
-bool 'Mitsumi (not IDE/ATAPI) CDROM driver support' CONFIG_MCD n
-bool 'Matsushita/Panasonic CDROM driver support' CONFIG_SBPCD n
-if [ "$CONFIG_SBPCD" = "y" ]; then
-  bool 'Matsushita/Panasonic second CDROM controller support' CONFIG_SBPCD2 n
-  if [ "$CONFIG_SBPCD2" = "y" ]; then
-    bool 'Matsushita/Panasonic third CDROM controller support' CONFIG_SBPCD3 n
-    if [ "$CONFIG_SBPCD3" = "y" ]; then
-      bool 'Matsushita/Panasonic fourth CDROM controller support' CONFIG_SBPCD4 n
-    fi
+  bool 'Network device support' CONFIG_NETDEVICES
+  if [ "$CONFIG_NETDEVICES" = "y" ]; then
+    source drivers/net/Config.in
   fi
+  endmenu
 fi
-bool 'Aztech/Orchid/Okano/Wearnes (non IDE) CDROM support' CONFIG_AZTCD n
-bool 'Sony CDU535 CDROM driver support' CONFIG_CDU535 n
 
-comment 'Filesystems'
+mainmenu_option next_comment
+comment 'ISDN subsystem'
 
-bool 'Standard (minix) fs support' CONFIG_MINIX_FS y
-bool 'Extended fs support' CONFIG_EXT_FS n
-bool 'Second extended fs support' CONFIG_EXT2_FS y
-bool 'xiafs filesystem support' CONFIG_XIA_FS n
-bool 'msdos fs support' CONFIG_MSDOS_FS y
-if [ "$CONFIG_MSDOS_FS" = "y" ]; then
-#bool 'umsdos: Unix like fs on top of std MSDOS FAT fs' CONFIG_UMSDOS_FS n
-comment 'Umsdos is not supported in 1.3.0: wait for 1.3.1'
+tristate 'ISDN support' CONFIG_ISDN
+if [ "$CONFIG_ISDN" != "n" ]; then
+  source drivers/isdn/Config.in
 fi
-bool '/proc filesystem support' CONFIG_PROC_FS y
-if [ "$CONFIG_INET" = "y" ]; then
-bool 'NFS filesystem support' CONFIG_NFS_FS y
-fi
-if [ "$CONFIG_BLK_DEV_SR" = "y" -o "$CONFIG_CDU31A" = "y" -o "$CONFIG_MCD" = "y" -o "$CONFIG_SBPCD" = "y" -o "$CONFIG_BLK_DEV_IDECD" = "y" -o "$CONFIG_AZTCD" = "y" -o "$CONFIG_CDU535" = "y" ]; then
-	bool 'ISO9660 cdrom filesystem support' CONFIG_ISO9660_FS y
-else
-	bool 'ISO9660 cdrom filesystem support' CONFIG_ISO9660_FS n
-fi
-bool 'OS/2 HPFS filesystem support (read only)' CONFIG_HPFS_FS n
-bool 'System V and Coherent filesystem support' CONFIG_SYSV_FS n
+endmenu
 
-comment 'character devices'
+mainmenu_option next_comment
+comment 'CD-ROM drivers (not for SCSI or IDE/ATAPI drives)'
 
-bool 'Cyclades async mux support' CONFIG_CYCLADES n
-bool 'Parallel printer support' CONFIG_PRINTER n
-bool 'Logitech busmouse support' CONFIG_BUSMOUSE n
-bool 'PS/2 mouse (aka "auxiliary device") support' CONFIG_PSMOUSE n
-if [ "$CONFIG_PSMOUSE" = "y" ]; then
-bool 'C&T 82C710 mouse port support (as on TI Travelmate)' CONFIG_82C710_MOUSE y
+bool 'Support non-SCSI/IDE/ATAPI CDROM drives' CONFIG_CD_NO_IDESCSI
+if [ "$CONFIG_CD_NO_IDESCSI" != "n" ]; then
+  source drivers/cdrom/Config.in
 fi
-bool 'Microsoft busmouse support' CONFIG_MS_BUSMOUSE n
-bool 'ATIXL busmouse support' CONFIG_ATIXL_BUSMOUSE n
-
+endmenu
 
-bool 'QIC-02 tape support' CONFIG_QIC02_TAPE n
-if [ "$CONFIG_QIC02_TAPE" = "y" ]; then
-bool 'Do you want runtime configuration for QIC-02' CONFIG_QIC02_DYNCONF y
-if [ "$CONFIG_QIC02_DYNCONF" != "y" ]; then
+source fs/Config.in
 
-comment '>>> Edit configuration parameters in ./include/linux/tpqic02.h!'
-
-else
-
-comment '>>> Setting runtime QIC-02 configuration is done with qic02conf'
-comment '>>> Which is available from ftp://ftp.funet.fi/pub/OS/Linux/BETA/QIC-02/'
-
-fi
-fi
-
-bool 'QIC-117 tape support' CONFIG_FTAPE n
-if [ "$CONFIG_FTAPE" = "y" ]; then
-int ' number of ftape buffers' NR_FTAPE_BUFFERS 3
-fi
+source drivers/char/Config.in
 
+mainmenu_option next_comment
 comment 'Sound'
 
-bool 'Sound card support' CONFIG_SOUND n
+tristate 'Sound card support' CONFIG_SOUND
+if [ "$CONFIG_SOUND" != "n" ]; then
+  source drivers/sound/Config.in
+fi
+endmenu
 
+mainmenu_option next_comment
 comment 'Kernel hacking'
 
-#bool 'Debug kmalloc/kfree' CONFIG_DEBUG_MALLOC n
-bool 'Kernel profiling support' CONFIG_PROFILE y
+#bool 'Debug kmalloc/kfree' CONFIG_DEBUG_MALLOC
+bool 'Kernel profiling support' CONFIG_PROFILE
 if [ "$CONFIG_PROFILE" = "y" ]; then
   int ' Profile shift count' CONFIG_PROFILE_SHIFT 2
 fi
-if [ "$CONFIG_SCSI" = "y" ]; then
-bool 'Verbose SCSI error reporting (kernel size +=12K)' CONFIG_SCSI_CONSTANTS y
-fi
+endmenu
diff --git a/arch/i386/defconfig b/arch/i386/defconfig
new file mode 100644
index 000000000..0e30c600a
--- /dev/null
+++ b/arch/i386/defconfig
@@ -0,0 +1,176 @@
+#
+# Automatically generated make config: don't edit
+#
+
+#
+# Code maturity level options
+#
+# CONFIG_EXPERIMENTAL is not set
+
+#
+# Loadable module support
+#
+CONFIG_MODULES=y
+# CONFIG_MODVERSIONS is not set
+# CONFIG_KERNELD is not set
+
+#
+# General setup
+#
+# CONFIG_MATH_EMULATION is not set
+CONFIG_NET=y
+# CONFIG_MAX_16M is not set
+CONFIG_PCI=y
+CONFIG_SYSVIPC=y
+CONFIG_BINFMT_AOUT=y
+CONFIG_BINFMT_ELF=y
+# CONFIG_M386 is not set
+# CONFIG_M486 is not set
+CONFIG_M586=y
+# CONFIG_M686 is not set
+# CONFIG_VIDEO_SELECT is not set
+
+#
+# Floppy, IDE, and other block devices
+#
+CONFIG_BLK_DEV_FD=y
+CONFIG_BLK_DEV_IDE=y
+
+#
+# Please see Documentation/ide.txt for help/info on IDE drives
+#
+# CONFIG_BLK_DEV_HD_IDE is not set
+CONFIG_BLK_DEV_IDEDISK=y
+CONFIG_BLK_DEV_IDECD=y
+# CONFIG_BLK_DEV_IDETAPE is not set
+# CONFIG_BLK_DEV_IDEFLOPPY is not set
+CONFIG_BLK_DEV_CMD640=y
+# CONFIG_BLK_DEV_CMD640_ENHANCED is not set
+CONFIG_BLK_DEV_RZ1000=y
+# CONFIG_BLK_DEV_TRITON is not set
+# CONFIG_IDE_CHIPSETS is not set
+
+#
+# Additional Block Devices
+#
+# CONFIG_BLK_DEV_LOOP is not set
+# CONFIG_BLK_DEV_MD is not set
+# CONFIG_BLK_DEV_RAM is not set
+# CONFIG_BLK_DEV_XD is not set
+# CONFIG_BLK_DEV_HD is not set
+
+#
+# Networking options
+#
+# CONFIG_NETLINK is not set
+# CONFIG_FIREWALL is not set
+# CONFIG_NET_ALIAS is not set
+CONFIG_INET=y
+# CONFIG_IP_FORWARD is not set
+# CONFIG_IP_MULTICAST is not set
+# CONFIG_IP_ACCT is not set
+
+#
+# (it is safe to leave these untouched)
+#
+# CONFIG_INET_PCTCP is not set
+# CONFIG_INET_RARP is not set
+# CONFIG_NO_PATH_MTU_DISCOVERY is not set
+CONFIG_IP_NOSR=y
+CONFIG_SKB_LARGE=y
+
+#
+#  
+#
+# CONFIG_IPX is not set
+# CONFIG_ATALK is not set
+# CONFIG_AX25 is not set
+
+#
+# SCSI support
+#
+# CONFIG_SCSI is not set
+
+#
+# Network device support
+#
+CONFIG_NETDEVICES=y
+# CONFIG_ARCNET is not set
+CONFIG_DUMMY=m
+# CONFIG_EQUALIZER is not set
+CONFIG_NET_ETHERNET=y
+CONFIG_NET_VENDOR_3COM=y
+# CONFIG_EL1 is not set
+# CONFIG_EL2 is not set
+CONFIG_EL3=y
+# CONFIG_VORTEX is not set
+# CONFIG_LANCE is not set
+# CONFIG_NET_VENDOR_SMC is not set
+# CONFIG_NET_ISA is not set
+# CONFIG_NET_EISA is not set
+# CONFIG_NET_POCKET is not set
+# CONFIG_FDDI is not set
+# CONFIG_PLIP is not set
+# CONFIG_PPP is not set
+# CONFIG_NET_RADIO is not set
+# CONFIG_SLIP is not set
+# CONFIG_TR is not set
+
+#
+# ISDN subsystem
+#
+# CONFIG_ISDN is not set
+
+#
+# CD-ROM drivers (not for SCSI or IDE/ATAPI drives)
+#
+# CONFIG_CD_NO_IDESCSI is not set
+
+#
+# Filesystems
+#
+# CONFIG_QUOTA is not set
+CONFIG_MINIX_FS=y
+# CONFIG_EXT_FS is not set
+CONFIG_EXT2_FS=y
+# CONFIG_XIA_FS is not set
+CONFIG_FAT_FS=y
+CONFIG_MSDOS_FS=y
+# CONFIG_VFAT_FS is not set
+# CONFIG_UMSDOS_FS is not set
+CONFIG_PROC_FS=y
+CONFIG_NFS_FS=y
+# CONFIG_ROOT_NFS is not set
+# CONFIG_SMB_FS is not set
+CONFIG_ISO9660_FS=y
+# CONFIG_HPFS_FS is not set
+# CONFIG_SYSV_FS is not set
+# CONFIG_AFFS_FS is not set
+# CONFIG_UFS_FS is not set
+
+#
+# Character devices
+#
+CONFIG_SERIAL=y
+# CONFIG_DIGI is not set
+# CONFIG_CYCLADES is not set
+# CONFIG_STALDRV is not set
+# CONFIG_RISCOM8 is not set
+# CONFIG_PRINTER is not set
+# CONFIG_MOUSE is not set
+# CONFIG_UMISC is not set
+# CONFIG_QIC02_TAPE is not set
+# CONFIG_FTAPE is not set
+# CONFIG_APM is not set
+# CONFIG_WATCHDOG is not set
+# CONFIG_RTC is not set
+
+#
+# Sound
+#
+# CONFIG_SOUND is not set
+
+#
+# Kernel hacking
+#
+# CONFIG_PROFILE is not set
diff --git a/arch/i386/ibcs/Makefile b/arch/i386/ibcs/Makefile
deleted file mode 100644
index 7b81ff2bc..000000000
--- a/arch/i386/ibcs/Makefile
+++ /dev/null
@@ -1,37 +0,0 @@
-#
-# Makefile for the iBCS emulator files
-#
-# Note! Dependencies are done automagically by 'make dep', which also
-# removes any old dependencies. DON'T put your own dependencies here
-# unless it's something special (ie not a .c file).
-#
-# Note 2! The CFLAGS definitions are now in the main makefile...
-
-.S.s:
-	$(CPP) -traditional $< -o $*.s
-.c.s:
-	$(CC) $(CFLAGS) -S $<
-.s.o:
-	$(AS) -c -o $*.o $<
-.c.o:
-	$(CC) $(CFLAGS) -c $<
-
-SUBDIRS	= 
-
-OBJS  = emulate.o
-
-ibcs.o: $(OBJS)
-	$(LD) -r -o ibcs.o $(OBJS)
-	sync
-
-dep:
-	$(CPP) -M *.c > .depend
-
-dummy:
-
-#
-# include a dependency file if one exists
-#
-ifeq (.depend,$(wildcard .depend))
-include .depend
-endif
diff --git a/arch/i386/ibcs/binfmt_coff.c b/arch/i386/ibcs/binfmt_coff.c
deleted file mode 100644
index 1a7c760cd..000000000
--- a/arch/i386/ibcs/binfmt_coff.c
+++ /dev/null
@@ -1,784 +0,0 @@
-/*
- * These are the functions used to load COFF IBSC style executables.
- * Information on COFF format may be obtained in either the Intel Binary
- * Compatibility Specification 2 or O'Rilley's book on COFF. The shared
- * libraries are defined only the in the Intel book.
- *
- * This file is based upon code written by Eric Youngdale for the ELF object
- * file format.
- *
- * Author: Al Longyear (longyear@sii.com)
- *
- * Latest Revision:
- *    3 February 1994
- *      Al Longyear (longyear@sii.com)
- *      Cleared first page of bss section using put_fs_byte.
- */
-
-#include <linux/fs.h>
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/mman.h>
-#include <linux/a.out.h>
-#include <linux/errno.h>
-#include <linux/signal.h>
-#include <linux/binfmts.h>
-#include <asm/segment.h>
-#include <linux/string.h>
-#include <linux/fcntl.h>
-#include <linux/ptrace.h>
-#include <linux/coff.h>
-#include <linux/malloc.h>
-
-asmlinkage int sys_exit (int exit_code);
-asmlinkage int sys_close (unsigned fd);
-asmlinkage int sys_open (const char *, int, int);
-asmlinkage int sys_uselib(const char * library);
-
-static int preload_library (struct linux_binprm *exe_bprm,
-			    COFF_SCNHDR * sect,
-			    struct file *fp);
-
-static int load_object (struct linux_binprm *bprm,
-			struct pt_regs *regs,
-			int lib_ok);
-
-/*
- *  Small procedure to test for the proper file alignment.
- */
-
-static inline int
-is_properly_aligned (COFF_SCNHDR *sect)
-{
-    long scnptr = COFF_LONG (sect->s_scnptr);
-    long vaddr  = COFF_LONG (sect->s_vaddr);
-/*
- *  Print the section information if needed
- */
-
-#ifdef COFF_DEBUG
-    printk ("%s, scnptr = %d, vaddr = %d\n",
-	    sect->s_name,
-	    scnptr, vaddr);
-#endif
-
-/*
- *  Return the error code if the section is not properly aligned.
- */
-
-#ifdef COFF_DEBUG
-    if (((vaddr - scnptr) & ~PAGE_MASK) != 0)
-	printk ("bad alignment in %s\n", sect->s_name);
-#endif
-    return ((((vaddr - scnptr) & ~PAGE_MASK) != 0) ? -ENOEXEC : 0);
-}
-
-/*
- *    Clear the bytes in the last page of data.
- */
-
-static
-int clear_memory (unsigned long addr, unsigned long size)
-{
-    int status;
-
-    size = (PAGE_SIZE - (addr & ~PAGE_MASK)) & ~PAGE_MASK;
-    if (size == 0)
-        status = 0;
-    else {
-      
-#ifdef COFF_DEBUG
-        printk ("un-initialized storage in last page %d\n", size);
-#endif
-
-	status = verify_area (VERIFY_WRITE,
-			      (void *) addr, size);
-#ifdef COFF_DEBUG
-	printk ("result from verify_area = %d\n", status);
-#endif
-
-	if (status >= 0)
-	    while (size-- != 0)
-	        put_fs_byte (0, addr++);
-    }
-    return status;
-}
-
-/*
- *  Helper function to process the load operation.
- */
-
-static int
-load_object (struct linux_binprm * bprm, struct pt_regs *regs, int lib_ok)
-{
-    COFF_FILHDR  *coff_hdr = (COFF_FILHDR *) bprm->buf;	/* COFF Header */
-    COFF_SCNHDR  *sect_bufr;	/* Pointer to section table            */
-    COFF_SCNHDR  *text_sect;	/* Pointer to the text section         */
-    COFF_SCNHDR  *data_sect;	/* Pointer to the data section         */
-    COFF_SCNHDR  *bss_sect;	/* Pointer to the bss section          */
-    int text_count;		/* Number of text sections             */
-    int data_count;		/* Number of data sections             */
-    int bss_count;		/* Number of bss sections              */
-    int lib_count;		/* Number of lib sections              */
-    unsigned int start_addr = 0;/* Starting location for program       */
-    int status = 0;		/* Result status register              */
-    int fd = -1;		/* Open file descriptor                */
-    struct file *fp     = NULL;	/* Pointer to the file at "fd"         */
-    short int sections  = 0;	/* Number of sections in the file      */
-    short int aout_size = 0;	/* Size of the a.out header area       */
-    short int flags;		/* Flag bits from the COFF header      */
-
-#ifdef COFF_DEBUG
-    printk ("binfmt_coff entry: %s\n", bprm->filename);
-#endif
-
-/*
- *  Validate the magic value for the object file.
- */
-    do {
-	if (COFF_I386BADMAG (*coff_hdr)) {
-#ifdef COFF_DEBUG
-	    printk ("bad filehdr magic\n");
-#endif
-	    status = -ENOEXEC;
-	    break;
-	}
-/*
- *  The object file should have 32 BIT little endian format. Do not allow
- *  it to have the 16 bit object file flag set as Linux is not able to run
- *  on the 80286/80186/8086.
- */
-	flags = COFF_SHORT (coff_hdr->f_flags);
-	if ((flags & (COFF_F_AR32WR | COFF_F_AR16WR)) != COFF_F_AR32WR) {
-#ifdef COFF_DEBUG
-	    printk ("invalid f_flags bits\n");
-#endif
-	    status = -ENOEXEC;
-	    break;
-	}
-/*
- *  Extract the header information which we need.
- */
-	sections  = COFF_SHORT (coff_hdr->f_nscns);   /* Number of sections */
-	aout_size = COFF_SHORT (coff_hdr->f_opthdr);  /* Size of opt. headr */
-/*
- *  If the file is not executable then reject the execution. This means
- *  that there must not be external references.
- */
-	if ((flags & COFF_F_EXEC) == 0) {
-#ifdef COFF_DEBUG
-	    printk ("not executable bit\n");
-#endif
-	    status = -ENOEXEC;
-	    break;
-	}
-/*
- *  There must be at least one section.
- */
-	if (sections == 0) {
-#ifdef COFF_DEBUG
-	    printk ("no sections\n");
-#endif
-	    status = -ENOEXEC;
-	    break;
-	}
-/*
- *  Do some additional consistency checks.
- *  The system requires mapping for this loader. If you try
- *  to use a file system with no mapping, the format is not valid.
- */
-	if (!bprm->inode->i_op ||
-	    !bprm->inode->i_op->default_file_ops->mmap) {
-#ifdef COFF_DEBUG
-	    printk ("no mmap in fs\n");
-#endif
-	    status = -ENOEXEC;
-	}
-    }
-    while (0);
-/*
- *  Allocate a buffer to hold the entire coff section list.
- */
-    if (status >= 0) {
-	int nbytes = sections * COFF_SCNHSZ;
-
-	sect_bufr = (COFF_SCNHDR *) kmalloc (nbytes, GFP_KERNEL);
-	if (0 == sect_bufr) {
-#ifdef COFF_DEBUG
-	    printk ("kmalloc failed\n");
-#endif
-	    status = -ENOEXEC;
-	}
-/*
- *  Read the section list from the disk file.
- */
-	else {
-	     int old_fs = get_fs ();
-	     set_fs (get_ds ());  /* Make it point to the proper location    */
-	     status = read_exec (bprm->inode,	     /* INODE for file       */
-			    aout_size + COFF_FILHSZ, /* Offset in the file   */
-			    (char *) sect_bufr,      /* Buffer for read      */
-			    nbytes);                 /* Byte count reqd.     */
-	     set_fs (old_fs);	                     /* Restore the selector */
-#ifdef COFF_DEBUG
-	     if (status < 0)
-	        printk ("read aout hdr, status = %d\n", status);
-#endif
-	 }
-    }
-    else
-	sect_bufr = NULL;	/* Errors do not have a section buffer */
-/*
- *  Count the number of sections for the required types and store the location
- *  of the last section for the three primary types.
- */
-    text_count = 0;
-    data_count = 0;
-    bss_count  = 0;
-    lib_count  = 0;
-
-    text_sect = NULL;
-    data_sect = NULL;
-    bss_sect  = NULL;
-/*
- *  Loop through the sections and find the various types
- */
-    if (status >= 0) {
-	int nIndex;
-	COFF_SCNHDR *sect_ptr = sect_bufr;
-
-	for (nIndex = 0; nIndex < sections; ++nIndex) {
-	    long int sect_flags = COFF_LONG (sect_ptr->s_flags);
-
-	    switch (sect_flags) {
-	    case COFF_STYP_TEXT:
-		text_sect = sect_ptr;
-		++text_count;
-		status = is_properly_aligned (sect_ptr);
-		break;
-
-	    case COFF_STYP_DATA:
-		data_sect = sect_ptr;
-		++data_count;
-		status = is_properly_aligned (sect_ptr);
-		break;
-
-	    case COFF_STYP_BSS:
-		bss_sect = sect_ptr;
-		++bss_count;
-		break;
-
-	    case COFF_STYP_LIB:
-#ifdef COFF_DEBUG
-		printk (".lib section found\n");
-#endif
-		++lib_count;
-		break;
-
-	    default:
-		break;
-	    }
-	    sect_ptr = (COFF_SCNHDR *) & ((char *) sect_ptr)[COFF_SCNHSZ];
-	}
-/*
- *  Ensure that there are the required sections. There must be one text
- *  sections and one each of the data and bss sections for an executable.
- *  A library may or may not have a data / bss section.
- */
-	if (text_count != 1) {
-	    status = -ENOEXEC;
-#ifdef COFF_DEBUG
-	    printk ("no text sections\n");
-#endif
-	}
-	else {
-	    if (lib_ok) {
-		if (data_count != 1 || bss_count != 1) {
-		    status = -ENOEXEC;
-#ifdef COFF_DEBUG
-		    printk ("no .data nor .bss sections\n");
-#endif
-		}
-	    }
-	}
-    }
-/*
- *  If there is no additional header then assume the file starts at
- *  the first byte of the text section. This may not be the proper place,
- *  so the best solution is to include the optional header. A shared library
- *  __MUST__ have an optional header to indicate that it is a shared library.
- */
-    if (status >= 0) {
-	if (aout_size == 0) {
-	    if (!lib_ok) {
-		status = -ENOEXEC;
-#ifdef COFF_DEBUG
-		printk ("no header in library\n");
-#endif
-	    }
-	    start_addr = COFF_LONG (text_sect->s_vaddr);
-	}
-/*
- *  There is some header. Ensure that it is sufficient.
- */
-	else {
-	    if (aout_size < COFF_AOUTSZ) {
-		status = -ENOEXEC;
-#ifdef COFF_DEBUG
-		printk ("header too small\n");
-#endif
-	    }
-	    else {
-		COFF_AOUTHDR *aout_hdr =	/* Pointer to a.out header */
-		(COFF_AOUTHDR *) & ((char *) coff_hdr)[COFF_FILHSZ];
-		short int aout_magic = COFF_SHORT (aout_hdr->magic); /* id */
-/*
- *  Validate the magic number in the a.out header. If it is valid then
- *  update the starting symbol location. Do not accept these file formats
- *  when loading a shared library.
- */
-		switch (aout_magic) {
-		case COFF_OMAGIC:
-		case COFF_ZMAGIC:
-		case COFF_STMAGIC:
-		    if (!lib_ok) {
-			status = -ENOEXEC;
-#ifdef COFF_DEBUG
-			printk ("wrong a.out header magic\n");
-#endif
-		    }
-		    start_addr = (unsigned int) COFF_LONG (aout_hdr->entry);
-		    break;
-/*
- *  Magic value for a shared library. This is valid only when loading a
- *  shared library. (There is no need for a start_addr. It won't be used.)
- */
-		case COFF_SHMAGIC:
-		    if (lib_ok) {
-#ifdef COFF_DEBUG
-			printk ("wrong a.out header magic\n");
-#endif
-			status = -ENOEXEC;
-		    }
-		    break;
-
-		default:
-#ifdef COFF_DEBUG
-		    printk ("wrong a.out header magic\n");
-#endif
-		    status = -ENOEXEC;
-		    break;
-		}
-	    }
-	}
-    }
-/*
- *  Fetch a file pointer to the executable.
- */
-    if (status >= 0) {
-	fd = open_inode (bprm->inode, O_RDONLY);
-	if (fd < 0) {
-#ifdef COFF_DEBUG
-	    printk ("can not open inode, result = %d\n", fd);
-#endif
-	    status = fd;
-	}
-	else
-	    fp = current->files->fd[fd];
-    }
-    else
-	fd = -1;		/* Invalidate the open file descriptor */
-/*
- *  Generate the proper values for the text fields
- *
- *  THIS IS THE POINT OF NO RETURN. THE NEW PROCESS WILL TRAP OUT SHOULD
- *  SOMETHING FAIL IN THE LOAD SEQUENCE FROM THIS POINT ONWARD.
- */
-    if (status >= 0) {
-	long text_scnptr = COFF_LONG (text_sect->s_scnptr);
-	long text_size   = COFF_LONG (text_sect->s_size);
-	long text_vaddr  = COFF_LONG (text_sect->s_vaddr);
-
-	long data_scnptr;
-	long data_size;
-	long data_vaddr;
-
-	long bss_size;
-	long bss_vaddr;
-/*
- *  Generate the proper values for the data fields
- */
-	if (data_sect != NULL) {
-	    data_scnptr = COFF_LONG (data_sect->s_scnptr);
-	    data_size   = COFF_LONG (data_sect->s_size);
-	    data_vaddr  = COFF_LONG (data_sect->s_vaddr);
-	}
-	else {
-	    data_scnptr = 0;
-	    data_size   = 0;
-	    data_vaddr  = 0;
-	}
-/*
- *  Generate the proper values for the bss fields
- */
-	if (bss_sect != NULL) {
-	    bss_size  = COFF_LONG (bss_sect->s_size);
-	    bss_vaddr = COFF_LONG (bss_sect->s_vaddr);
-	}
-	else {
-	    bss_size  = 0;
-	    bss_vaddr = 0;
-	}
-/*
- *  Flush the executable from memory. At this point the executable is
- *  committed to being defined or a segmentation violation will occur.
- */
-	if (lib_ok) {
-#ifdef COFF_DEBUG
-	    printk ("flushing executable\n");
-#endif
-	    flush_old_exec (bprm);
-/*
- *  Define the initial locations for the various items in the new process
- */
-	    current->mm->mmap        = NULL;
-	    current->mm->rss         = 0;
-/*
- *  Construct the parameter and environment string table entries.
- */
-	    bprm->p += change_ldt (0, bprm->page);
-	    bprm->p -= MAX_ARG_PAGES*PAGE_SIZE;
-	    bprm->p  = (unsigned long) create_tables ((char *) bprm->p,
-						      bprm->argc,
-						      bprm->envc,
-						      1);
-/*
- *  Do the end processing once the stack has been constructed
- */
-	    current->mm->start_code  = text_vaddr & PAGE_MASK;
-	    current->mm->end_code    = text_vaddr + text_size;
-	    current->mm->end_data    = data_vaddr + data_size;
-	    current->mm->start_brk   =
-	    current->mm->brk         = bss_vaddr + bss_size;
-	    current->suid            =
-	    current->euid            = bprm->e_uid;
-	    current->sgid            =
-	    current->egid            = bprm->e_gid;
-	    current->executable      = bprm->inode; /* Store inode for file  */
-	    ++bprm->inode->i_count;             /* Count the open inode  */
-	    regs->eip                = start_addr;  /* Current EIP register  */
-	    regs->esp                =
-	    current->mm->start_stack = bprm->p;
-	}
-/*
- *   Map the text pages
- */
-
-#ifdef COFF_DEBUG
-	printk (".text: vaddr = %d, size = %d, scnptr = %d\n",
-		 text_vaddr,
-		 text_size,
-		 text_scnptr);
-#endif
-	status = do_mmap (fp,
-			  text_vaddr & PAGE_MASK,
-			  text_size + (text_vaddr & ~PAGE_MASK),
-			  PROT_READ | PROT_EXEC,
-			  MAP_FIXED | MAP_SHARED,
-			  text_scnptr & PAGE_MASK);
-
-	status = (status == (text_vaddr & PAGE_MASK)) ? 0 : -ENOEXEC;
-/*
- *   Map the data pages
- */
-	if (status >= 0 && data_size != 0) {
-#ifdef COFF_DEBUG
-	    printk (".data: vaddr = %d, size = %d, scnptr = %d\n",
-		     data_vaddr,
-		     data_size,
-		     data_scnptr);
-#endif
-	    status = do_mmap (fp,
-			      data_vaddr & PAGE_MASK,
-			      data_size + (data_vaddr & ~PAGE_MASK),
-			      PROT_READ | PROT_WRITE | PROT_EXEC,
-			      MAP_FIXED | MAP_PRIVATE,
-			      data_scnptr & PAGE_MASK);
-
-	    status = (status == (data_vaddr & PAGE_MASK)) ? 0 : -ENOEXEC;
-	}
-/*
- *   Construct the bss data for the process. The bss ranges from the
- *   end of the data (which may not be on a page boundary) to the end
- *   of the bss section. Allocate any necessary pages for the data.
- */
-	if (status >= 0 && bss_size != 0) {
-#ifdef COFF_DEBUG
-	    printk (".bss: vaddr = %d, size = %d\n",
-		     bss_vaddr,
-		     bss_size);
-#endif
-	    zeromap_page_range (PAGE_ALIGN (bss_vaddr),
-				PAGE_ALIGN (bss_size),
-				PAGE_COPY);
-
-	    status = clear_memory (bss_vaddr, bss_size);
-	}
-/*
- *  Load any shared library for the executable.
- */
-	if (status >= 0 && lib_ok && lib_count != 0) {
-	    int nIndex;
-	    COFF_SCNHDR *sect_ptr = sect_bufr;
-/*
- *  Find the library sections. (There should be at least one. It was counted
- *  earlier.) This will eventually recurse to our code and load the shared
- *  library with our own procedures.
- */
-	    for (nIndex = 0; nIndex < sections; ++nIndex) {
-		long int sect_flags = COFF_LONG (sect_ptr->s_flags);
-		if (sect_flags == COFF_STYP_LIB) {
-		    status = preload_library (bprm, sect_ptr, fp);
-		    if (status != 0)
-			break;
-		}
-	    sect_ptr = (COFF_SCNHDR *) &((char *) sect_ptr) [COFF_SCNHSZ];
-	    }
-	}
-/*
- *   Generate any needed trap for this process. If an error occurred then
- *   generate a segmentation violation. If the process is being debugged
- *   then generate the load trap. (Note: If this is a library load then
- *   do not generate the trap here. Pass the error to the caller who
- *   will do it for the process in the outer lay of this procedure call.)
- */
-	if (lib_ok) {
-	    if (status < 0)
-		send_sig (SIGSEGV, current, 0);	/* Generate the error trap  */
-	    else {
-		if (current->flags & PF_PTRACED)
-		    send_sig (SIGTRAP, current, 0);
-	    }
-	    status = 0;		/* We are committed. It can't fail */
-	}
-    }
-/*
- *  Do any cleanup processing
- */
-    if (fd >= 0)
-	sys_close (fd);		/* Close unused code file      */
-
-    if (sect_bufr != NULL)
-	kfree (sect_bufr);	/* Release section list buffer */
-/*
- *  Return the completion status.
- */
-#ifdef COFF_DEBUG
-    printk ("binfmt_coff: result = %d\n", status);
-#endif
-    return (status);
-}
-
-/*
- *  This procedure will load the library listed in the file name given
- *  as the parameter. The result will be non-zero should something fail
- *  to load.
- */
-
-static int
-preload_this_library (struct linux_binprm *exe_bprm, char *lib_name)
-{
-    int   status;
-    int   old_fs = get_fs();
-/*
- *  If debugging then print "we have arrived"
- */
-#ifdef COFF_DEBUG
-    printk ("%s loading shared library %s\n",
-	    exe_bprm->filename,
-	    lib_name);
-#endif
-/*
- *  Change the FS register to the proper kernel address space and attempt
- *  to load the library. The library name is allocated from the kernel
- *  pool.
- */
-    set_fs (get_ds ());
-    status = sys_uselib (lib_name);
-    set_fs (old_fs);
-/*
- *  Return the success/failure to the caller.
- */
-    return (status);
-}
-
-/*
- *  This procedure is called to load a library section. The various
- *  libraries are loaded from the list given in the section data.
- */
-
-static int
-preload_library (struct linux_binprm *exe_bprm,
-		 COFF_SCNHDR * sect, struct file *fp)
-{
-    int status = 0;		/* Completion status                  */
-    long nbytes;		/* Count of bytes in the header area  */
-/*
- *  Fetch the size of the section. There must be enough room for at least
- *  one entry.
- */
-    nbytes = COFF_LONG (sect->s_size);
-    if (nbytes < COFF_SLIBSZ) {
-	status = -ENOEXEC;
-#ifdef COFF_DEBUG
-	printk ("library section too small\n");
-#endif
-    }
-/*
- *  Allocate a buffer to hold the section data
- */
-    else {
-	COFF_SLIBHD *phdr;
-	char *buffer = (char *) kmalloc (nbytes, GFP_KERNEL);
-
-        if (0 == buffer) {
-	    status = -ENOEXEC;
-#ifdef COFF_DEBUG
-	    printk ("kmalloc failed\n");
-#endif
-	}
-	else {
-	    int old_fs   = get_fs ();
-/*
- *  Read the section data from the disk file.
- */
-	    set_fs (get_ds ());   /* Make it point to the proper location    */
-	    status = read_exec (exe_bprm->inode,     /* INODE for file       */
-			    COFF_LONG (sect->s_scnptr), /* Disk location     */
-			    buffer,                     /* Buffer for read   */
-			    nbytes);                    /* Byte count reqd.  */
-	    set_fs (old_fs);                         /* Restore the selector */
-/*
- *  Check the result. The value returned is the byte count actually read.
- */
-	    if (status >= 0 && status != nbytes) {
-#ifdef COFF_DEBUG
-		printk ("read of lib section was short\n");
-#endif
-		status = -ENOEXEC;
-	    }
-	}
-/*
- *  At this point, go through the list of libraries in the data area.
- */
-	phdr = (COFF_SLIBHD *) buffer;
-	while (status >= 0 && nbytes > COFF_SLIBSZ) {
-	    int entry_size  = COFF_LONG (phdr->sl_entsz)   * sizeof (long);
-	    int header_size = COFF_LONG (phdr->sl_pathndx) * sizeof (long);
-/*
- *  Validate the sizes of the various items. I don't trust the linker!!
- */
-	    if ((unsigned) header_size >= (unsigned) nbytes ||
-		entry_size <= 0 ||
-		(unsigned) entry_size <= (unsigned) header_size) {
-		status = -ENOEXEC;
-#ifdef COFF_DEBUG
-		printk ("header count is invalid\n");
-#endif
-	    }
-/*
- *  Load the library. Stop the load process on the first error.
- */
-	    else {
-		status = preload_this_library (exe_bprm,
-					       &((char *) phdr)[header_size]);
-#ifdef COFF_DEBUG
-		printk ("preload_this_library result = %d\n", status);
-#endif
-	    }
-/*
- *  Point to the next library in the section data.
- */
-	    nbytes -= entry_size;
-	    phdr = (COFF_SLIBHD *) &((char *) phdr)[entry_size];
-	}
-/*
- *  Release the space for the library list.
- */
-	if (buffer != NULL)
-	    kfree (buffer);
-    }
-/*
- *  Return the resulting status to the caller.
- */
-    return (status);
-}
-
-/*
- *  This procedure is called by the main load sequence. It will load
- *  the executable and prepare it for execution. It provides the additional
- *  parameters used by the recursive coff loader and tells the loader that
- *  this is the main executable. How simple it is . . . .
- */
-
-int
-load_coff_binary (struct linux_binprm *bprm, struct pt_regs *regs)
-{
-    return (load_object (bprm, regs, 1));
-}
-
-/*
- *   Load the image for any shared library.
- *
- *   This is called when we need to load a library based upon a file name.
- */
-
-int
-load_coff_library (int fd)
-{
-    struct linux_binprm *bprm;  /* Parameters for the load operation   */
-    int    status;              /* Status of the request               */
-/*
- *  Read the first portion of the file.
- */
-    bprm = (struct linux_binprm *) kmalloc (sizeof (struct linux_binprm),
-					    GFP_KERNEL);
-    if (0 == bprm) {
-#ifdef COFF_DEBUG
-	printk ("kmalloc failed\n");
-#endif
-        status = -ENOEXEC;
-    }
-    else {
-        struct file         *file;  /* Pointer to the file table           */
-        struct pt_regs       regs;  /* Register work area                  */
-        int old_fs = get_fs ();     /* Previous FS register value          */
-
-        memset (bprm, '\0', sizeof (struct linux_binprm));
-
-	file           = current->files->fd[fd];
-	bprm->inode    = file->f_inode;   /* The only item _really_ needed */
-	bprm->filename = "";              /* Make it a legal string        */
-/*
- *  Read the section list from the disk file.
- */
-	set_fs (get_ds ());   /* Make it point to the proper location    */
-	status = read_exec (bprm->inode,	 /* INODE for file       */
-			    0L,                  /* Offset in the file   */
-			    bprm->buf,           /* Buffer for read      */
-			    sizeof (bprm->buf)); /* Size of the buffer   */
-	set_fs (old_fs);	                 /* Restore the selector */
-/*
- *  Try to load the library.
- */
-	status = load_object (bprm, &regs, 0);
-/*
- *  Release the work buffer and return the result.
- */
-	kfree (bprm);                 /* Release the buffer area */
-    }
-/*
- *  Return the result of the load operation
- */
-    return (status);
-}
diff --git a/arch/i386/ibcs/binfmt_elf.c b/arch/i386/ibcs/binfmt_elf.c
deleted file mode 100644
index ab6d36915..000000000
--- a/arch/i386/ibcs/binfmt_elf.c
+++ /dev/null
@@ -1,655 +0,0 @@
-/*
- * linux/fs/binfmt_elf.c
- *
- * These are the functions used to load ELF format executables as used
- * on SVr4 machines.  Information on the format may be found in the book
- * "UNIX SYSTEM V RELEASE 4 Programmers Guide: Ansi C and Programming Support
- * Tools".
- *
- * Copyright 1993, 1994: Eric Youngdale (ericy@cais.com).
- */
-#include <linux/fs.h>
-#include <linux/sched.h>
-#include <linux/mm.h>
-#include <linux/mman.h>
-#include <linux/a.out.h>
-#include <linux/errno.h>
-#include <linux/signal.h>
-#include <linux/binfmts.h>
-#include <linux/string.h>
-#include <linux/fcntl.h>
-#include <linux/ptrace.h>
-#include <linux/malloc.h>
-#include <linux/shm.h>
-#include <linux/personality.h>
-
-#include <asm/segment.h>
-
-asmlinkage int sys_exit(int exit_code);
-asmlinkage int sys_close(unsigned fd);
-asmlinkage int sys_open(const char *, int, int);
-asmlinkage int sys_brk(unsigned long);
-
-#define DLINFO_ITEMS 8
-
-#include <linux/elf.h>
-
-/* We need to explicitly zero any fractional pages
-   after the data section (i.e. bss).  This would
-   contain the junk from the file that should not
-   be in memory */
-
-static void padzero(int elf_bss){
-  unsigned int fpnt, nbyte;
-  
-  if(elf_bss & 0xfff) {
-    
-    nbyte = (PAGE_SIZE - (elf_bss & 0xfff)) & 0xfff;
-    if(nbyte){
-      verify_area(VERIFY_WRITE, (void *) elf_bss, nbyte);
-      
-      fpnt = elf_bss;
-      while(fpnt & 0xfff) put_fs_byte(0, fpnt++);
-    };
-  };
-}
-
-unsigned long * create_elf_tables(char * p,int argc,int envc,struct elfhdr * exec, unsigned int load_addr, int ibcs)
-{
-	unsigned long *argv,*envp, *dlinfo;
-	unsigned long * sp;
-	struct vm_area_struct *mpnt;
-
-	mpnt = (struct vm_area_struct *)kmalloc(sizeof(*mpnt), GFP_KERNEL);
-	if (mpnt) {
-		mpnt->vm_task = current;
-		mpnt->vm_start = PAGE_MASK & (unsigned long) p;
-		mpnt->vm_end = TASK_SIZE;
-		mpnt->vm_page_prot = PAGE_COPY;
-		mpnt->vm_flags = VM_STACK_FLAGS;
-		mpnt->vm_ops = NULL;
-		mpnt->vm_inode = NULL;
-		mpnt->vm_offset = 0;
-		mpnt->vm_pte = 0;
-		insert_vm_struct(current, mpnt);
-	}
-	sp = (unsigned long *) (0xfffffffc & (unsigned long) p);
-	if(exec) sp -= DLINFO_ITEMS*2;
-	dlinfo = sp;
-	sp -= envc+1;
-	envp = sp;
-	sp -= argc+1;
-	argv = sp;
-	if (!ibcs) {
-		put_fs_long((unsigned long)envp,--sp);
-		put_fs_long((unsigned long)argv,--sp);
-	}
-
-	/* The constant numbers (0-9) that we are writing here are
-	   described in the header file sys/auxv.h on at least
-	   some versions of SVr4 */
-	if(exec) { /* Put this here for an ELF program interpreter */
-	  struct elf_phdr * eppnt;
-	  eppnt = (struct elf_phdr *) exec->e_phoff;
-	  put_fs_long(3,dlinfo++); put_fs_long(load_addr + exec->e_phoff,dlinfo++);
-	  put_fs_long(4,dlinfo++); put_fs_long(sizeof(struct elf_phdr),dlinfo++);
-	  put_fs_long(5,dlinfo++); put_fs_long(exec->e_phnum,dlinfo++);
-	  put_fs_long(9,dlinfo++); put_fs_long((unsigned long) exec->e_entry,dlinfo++);
-	  put_fs_long(7,dlinfo++); put_fs_long(SHM_RANGE_START,dlinfo++);
-	  put_fs_long(8,dlinfo++); put_fs_long(0,dlinfo++);
-	  put_fs_long(6,dlinfo++); put_fs_long(PAGE_SIZE,dlinfo++);
-	  put_fs_long(0,dlinfo++); put_fs_long(0,dlinfo++);
-	};
-
-	put_fs_long((unsigned long)argc,--sp);
-	current->mm->arg_start = (unsigned long) p;
-	while (argc-->0) {
-		put_fs_long((unsigned long) p,argv++);
-		while (get_fs_byte(p++)) /* nothing */ ;
-	}
-	put_fs_long(0,argv);
-	current->mm->arg_end = current->mm->env_start = (unsigned long) p;
-	while (envc-->0) {
-		put_fs_long((unsigned long) p,envp++);
-		while (get_fs_byte(p++)) /* nothing */ ;
-	}
-	put_fs_long(0,envp);
-	current->mm->env_end = (unsigned long) p;
-	return sp;
-}
-
-
-/* This is much more generalized than the library routine read function,
-   so we keep this separate.  Technically the library read function
-   is only provided so that we can read a.out libraries that have
-   an ELF header */
-
-static unsigned int load_elf_interp(struct elfhdr * interp_elf_ex,
-			     struct inode * interpreter_inode)
-{
-        struct file * file;
-	struct elf_phdr *elf_phdata  =  NULL;
-	struct elf_phdr *eppnt;
-	unsigned int len;
-	unsigned int load_addr;
-	int elf_exec_fileno;
-	int elf_bss;
-	int old_fs, retval;
-	unsigned int last_bss;
-	int error;
-	int i, k;
-	
-	elf_bss = 0;
-	last_bss = 0;
-	error = load_addr = 0;
-	
-	/* First of all, some simple consistency checks */
-	if((interp_elf_ex->e_type != ET_EXEC && 
-	    interp_elf_ex->e_type != ET_DYN) || 
-	   (interp_elf_ex->e_machine != EM_386 && interp_elf_ex->e_machine != EM_486) ||
-	   (!interpreter_inode->i_op || 
-	    !interpreter_inode->i_op->default_file_ops->mmap)){
-		return 0xffffffff;
-	};
-	
-	/* Now read in all of the header information */
-	
-	if(sizeof(struct elf_phdr) * interp_elf_ex->e_phnum > PAGE_SIZE) 
-	    return 0xffffffff;
-	
-	elf_phdata =  (struct elf_phdr *) 
-		kmalloc(sizeof(struct elf_phdr) * interp_elf_ex->e_phnum, GFP_KERNEL);
-	if(!elf_phdata) return 0xffffffff;
-	
-	old_fs = get_fs();
-	set_fs(get_ds());
-	retval = read_exec(interpreter_inode, interp_elf_ex->e_phoff, (char *) elf_phdata,
-			   sizeof(struct elf_phdr) * interp_elf_ex->e_phnum);
-	set_fs(old_fs);
-	
-	elf_exec_fileno = open_inode(interpreter_inode, O_RDONLY);
-	if (elf_exec_fileno < 0) return 0xffffffff;
-	file = current->files->fd[elf_exec_fileno];
-
-	eppnt = elf_phdata;
-	for(i=0; i<interp_elf_ex->e_phnum; i++, eppnt++)
-	  if(eppnt->p_type == PT_LOAD) {
-	    error = do_mmap(file, 
-			    eppnt->p_vaddr & 0xfffff000,
-			    eppnt->p_filesz + (eppnt->p_vaddr & 0xfff),
-			    PROT_READ | PROT_WRITE | PROT_EXEC,
-			    MAP_PRIVATE | (interp_elf_ex->e_type == ET_EXEC ? MAP_FIXED : 0),
-			    eppnt->p_offset & 0xfffff000);
-	    
-	    if(!load_addr && interp_elf_ex->e_type == ET_DYN)
-	      load_addr = error;
-	    k = load_addr + eppnt->p_vaddr + eppnt->p_filesz;
-	    if(k > elf_bss) elf_bss = k;
-	    if(error < 0 && error > -1024) break;  /* Real error */
-	    k = load_addr + eppnt->p_memsz + eppnt->p_vaddr;
-	    if(k > last_bss) last_bss = k;
-	  }
-	
-	/* Now use mmap to map the library into memory. */
-
-	
-	sys_close(elf_exec_fileno);
-	if(error < 0 && error > -1024) {
-	        kfree(elf_phdata);
-		return 0xffffffff;
-	}
-
-	padzero(elf_bss);
-	len = (elf_bss + 0xfff) & 0xfffff000; /* What we have mapped so far */
-
-	/* Map the last of the bss segment */
-	if (last_bss > len)
-	  do_mmap(NULL, len, last_bss-len,
-		  PROT_READ|PROT_WRITE|PROT_EXEC,
-		  MAP_FIXED|MAP_PRIVATE, 0);
-	kfree(elf_phdata);
-
-	return ((unsigned int) interp_elf_ex->e_entry) + load_addr;
-}
-
-static unsigned int load_aout_interp(struct exec * interp_ex,
-			     struct inode * interpreter_inode)
-{
-  int retval;
-  unsigned int elf_entry;
-  
-  current->mm->brk = interp_ex->a_bss +
-    (current->mm->end_data = interp_ex->a_data +
-     (current->mm->end_code = interp_ex->a_text));
-  elf_entry = interp_ex->a_entry;
-  
-  
-  if (N_MAGIC(*interp_ex) == OMAGIC) {
-    do_mmap(NULL, 0, interp_ex->a_text+interp_ex->a_data,
-	    PROT_READ|PROT_WRITE|PROT_EXEC,
-	    MAP_FIXED|MAP_PRIVATE, 0);
-    retval = read_exec(interpreter_inode, 32, (char *) 0, 
-		       interp_ex->a_text+interp_ex->a_data);
-  } else if (N_MAGIC(*interp_ex) == ZMAGIC || N_MAGIC(*interp_ex) == QMAGIC) {
-    do_mmap(NULL, 0, interp_ex->a_text+interp_ex->a_data,
-	    PROT_READ|PROT_WRITE|PROT_EXEC,
-	    MAP_FIXED|MAP_PRIVATE, 0);
-    retval = read_exec(interpreter_inode,
-		       N_TXTOFF(*interp_ex) ,
-		       (char *) N_TXTADDR(*interp_ex),
-		       interp_ex->a_text+interp_ex->a_data);
-  } else
-    retval = -1;
-  
-  if(retval >= 0)
-    do_mmap(NULL, (interp_ex->a_text + interp_ex->a_data + 0xfff) & 
-	    0xfffff000, interp_ex->a_bss,
-	    PROT_READ|PROT_WRITE|PROT_EXEC,
-	    MAP_FIXED|MAP_PRIVATE, 0);
-  if(retval < 0) return 0xffffffff;
-  return elf_entry;
-}
-
-/*
- * These are the functions used to load ELF style executables and shared
- * libraries.  There is no binary dependent code anywhere else.
- */
-
-#define INTERPRETER_NONE 0
-#define INTERPRETER_AOUT 1
-#define INTERPRETER_ELF 2
-
-static int load_elf_binary(struct linux_binprm * bprm, struct pt_regs * regs)
-{
-	struct elfhdr elf_ex;
-	struct elfhdr interp_elf_ex;
-	struct file * file;
-  	struct exec interp_ex;
-	struct inode *interpreter_inode;
-	unsigned int load_addr;
-	unsigned int interpreter_type = INTERPRETER_NONE;
-	int i;
-	int old_fs;
-	int error;
-	struct elf_phdr * elf_ppnt, *elf_phdata;
-	int elf_exec_fileno;
-	unsigned int elf_bss, k, elf_brk;
-	int retval;
-	char * elf_interpreter;
-	unsigned int elf_entry;
-	int status;
-	unsigned int start_code, end_code, end_data;
-	unsigned int elf_stack;
-	char passed_fileno[6];
-	
-	status = 0;
-	load_addr = 0;
-	elf_ex = *((struct elfhdr *) bprm->buf);	  /* exec-header */
-	
-	if (elf_ex.e_ident[0] != 0x7f ||
-	    strncmp(&elf_ex.e_ident[1], "ELF",3) != 0)
-		return  -ENOEXEC;
-	
-	
-	/* First of all, some simple consistency checks */
-	if(elf_ex.e_type != ET_EXEC || 
-	   (elf_ex.e_machine != EM_386 && elf_ex.e_machine != EM_486) ||
-	   (!bprm->inode->i_op || !bprm->inode->i_op->default_file_ops ||
-	    !bprm->inode->i_op->default_file_ops->mmap)){
-		return -ENOEXEC;
-	};
-	
-	/* Now read in all of the header information */
-	
-	elf_phdata = (struct elf_phdr *) kmalloc(elf_ex.e_phentsize * 
-						 elf_ex.e_phnum, GFP_KERNEL);
-	
-	old_fs = get_fs();
-	set_fs(get_ds());
-	retval = read_exec(bprm->inode, elf_ex.e_phoff, (char *) elf_phdata,
-			   elf_ex.e_phentsize * elf_ex.e_phnum);
-	set_fs(old_fs);
-	if (retval < 0) {
-	        kfree (elf_phdata);
-		return retval;
-	}
-	
-	elf_ppnt = elf_phdata;
-	
-	elf_bss = 0;
-	elf_brk = 0;
-	
-	elf_exec_fileno = open_inode(bprm->inode, O_RDONLY);
-
-	if (elf_exec_fileno < 0) {
-	        kfree (elf_phdata);
-		return elf_exec_fileno;
-	}
-	
-	file = current->files->fd[elf_exec_fileno];
-	
-	elf_stack = 0xffffffff;
-	elf_interpreter = NULL;
-	start_code = 0;
-	end_code = 0;
-	end_data = 0;
-	
-	old_fs = get_fs();
-	set_fs(get_ds());
-	
-	for(i=0;i < elf_ex.e_phnum; i++){
-		if(elf_ppnt->p_type == PT_INTERP) {
-			/* This is the program interpreter used for shared libraries - 
-			   for now assume that this is an a.out format binary */
-			
-			elf_interpreter = (char *) kmalloc(elf_ppnt->p_filesz, 
-							   GFP_KERNEL);
-			
-			retval = read_exec(bprm->inode,elf_ppnt->p_offset,elf_interpreter,
-					   elf_ppnt->p_filesz);
-#if 0
-			printk("Using ELF interpreter %s\n", elf_interpreter);
-#endif
-			if(retval >= 0)
-				retval = namei(elf_interpreter, &interpreter_inode);
-			if(retval >= 0)
-				retval = read_exec(interpreter_inode,0,bprm->buf,128);
-			
-			if(retval >= 0){
-				interp_ex = *((struct exec *) bprm->buf);		/* exec-header */
-				interp_elf_ex = *((struct elfhdr *) bprm->buf);	  /* exec-header */
-				
-			};
-			if(retval < 0) {
-			  kfree (elf_phdata);
-			  kfree(elf_interpreter);
-			  return retval;
-			};
-		};
-		elf_ppnt++;
-	};
-	
-	set_fs(old_fs);
-	
-	/* Some simple consistency checks for the interpreter */
-	if(elf_interpreter){
-	        interpreter_type = INTERPRETER_ELF | INTERPRETER_AOUT;
-		if(retval < 0) {
-			kfree(elf_interpreter);
-			kfree(elf_phdata);
-			return -ELIBACC;
-		};
-		/* Now figure out which format our binary is */
-		if((N_MAGIC(interp_ex) != OMAGIC) && 
-		   (N_MAGIC(interp_ex) != ZMAGIC) &&
-		   (N_MAGIC(interp_ex) != QMAGIC)) 
-		  interpreter_type = INTERPRETER_ELF;
-
-		if (interp_elf_ex.e_ident[0] != 0x7f ||
-		    strncmp(&interp_elf_ex.e_ident[1], "ELF",3) != 0)
-		  interpreter_type &= ~INTERPRETER_ELF;
-
-		if(!interpreter_type)
-		  {
-		    kfree(elf_interpreter);
-		    kfree(elf_phdata);
-		    return -ELIBBAD;
-		  };
-	}
-	
-	/* OK, we are done with that, now set up the arg stuff,
-	   and then start this sucker up */
-	
-	if (!bprm->sh_bang) {
-		char * passed_p;
-		
-		if(interpreter_type == INTERPRETER_AOUT) {
-		  sprintf(passed_fileno, "%d", elf_exec_fileno);
-		  passed_p = passed_fileno;
-		
-		  if(elf_interpreter) {
-		    bprm->p = copy_strings(1,&passed_p,bprm->page,bprm->p,2);
-		    bprm->argc++;
-		  };
-		};
-		if (!bprm->p) {
-		        if(elf_interpreter) {
-			      kfree(elf_interpreter);
-			}
-		        kfree (elf_phdata);
-			return -E2BIG;
-		}
-	}
-	
-	/* OK, This is the point of no return */
-	flush_old_exec(bprm);
-
-	current->mm->end_data = 0;
-	current->mm->end_code = 0;
-	current->mm->start_mmap = ELF_START_MMAP;
-	current->mm->mmap = NULL;
-	elf_entry = (unsigned int) elf_ex.e_entry;
-	
-	/* Do this so that we can load the interpreter, if need be.  We will
-	   change some of these later */
-	current->mm->rss = 0;
-	bprm->p += change_ldt(0, bprm->page);
-	current->mm->start_stack = bprm->p;
-	
-	/* Now we do a little grungy work by mmaping the ELF image into
-	   the correct location in memory.  At this point, we assume that
-	   the image should be loaded at fixed address, not at a variable
-	   address. */
-	
-	old_fs = get_fs();
-	set_fs(get_ds());
-	
-	elf_ppnt = elf_phdata;
-	for(i=0;i < elf_ex.e_phnum; i++){
-		
-		if(elf_ppnt->p_type == PT_INTERP) {
-			/* Set these up so that we are able to load the interpreter */
-		  /* Now load the interpreter into user address space */
-		  set_fs(old_fs);
-
-		  if(interpreter_type & 1) elf_entry = 
-		    load_aout_interp(&interp_ex, interpreter_inode);
-
-		  if(interpreter_type & 2) elf_entry = 
-		    load_elf_interp(&interp_elf_ex, interpreter_inode);
-
-		  old_fs = get_fs();
-		  set_fs(get_ds());
-
-		  iput(interpreter_inode);
-		  kfree(elf_interpreter);
-			
-		  if(elf_entry == 0xffffffff) { 
-		    printk("Unable to load interpreter\n");
-		    kfree(elf_phdata);
-		    send_sig(SIGSEGV, current, 0);
-		    return 0;
-		  };
-		};
-		
-		
-		if(elf_ppnt->p_type == PT_LOAD) {
-			error = do_mmap(file,
-					elf_ppnt->p_vaddr & 0xfffff000,
-					elf_ppnt->p_filesz + (elf_ppnt->p_vaddr & 0xfff),
-					PROT_READ | PROT_WRITE | PROT_EXEC,
-					MAP_FIXED | MAP_PRIVATE,
-					elf_ppnt->p_offset & 0xfffff000);
-			
-#ifdef LOW_ELF_STACK
-			if(elf_ppnt->p_vaddr & 0xfffff000 < elf_stack) 
-				elf_stack = elf_ppnt->p_vaddr & 0xfffff000;
-#endif
-			
-			if(!load_addr) 
-			  load_addr = elf_ppnt->p_vaddr - elf_ppnt->p_offset;
-			k = elf_ppnt->p_vaddr;
-			if(k > start_code) start_code = k;
-			k = elf_ppnt->p_vaddr + elf_ppnt->p_filesz;
-			if(k > elf_bss) elf_bss = k;
-			if((elf_ppnt->p_flags | PROT_WRITE) && end_code <  k)
-				end_code = k; 
-			if(end_data < k) end_data = k; 
-			k = elf_ppnt->p_vaddr + elf_ppnt->p_memsz;
-			if(k > elf_brk) elf_brk = k;		     
-		      };
-		elf_ppnt++;
-	};
-	set_fs(old_fs);
-	
-	kfree(elf_phdata);
-	
-	if(interpreter_type != INTERPRETER_AOUT) sys_close(elf_exec_fileno);
-
-       	/* The following 3 lines need a little bit of work if we are loading
-	   an iBCS2 binary.  We should initially load it this way, and if
-	   we get a lcall7, then we should look to see if the iBCS2 execution
-	   profile is present.  If it is, then switch to that, otherwise
-	   bomb. */
-	current->personality = PER_LINUX;
-	current->lcall7 = no_lcall7;
-	current->signal_map = current->signal_invmap = ident_map;
-
-	current->executable = bprm->inode;
-	bprm->inode->i_count++;
-#ifdef LOW_ELF_STACK
-	current->start_stack = p = elf_stack - 4;
-#endif
-	bprm->p -= MAX_ARG_PAGES*PAGE_SIZE;
-	bprm->p = (unsigned long) 
-	  create_elf_tables((char *)bprm->p,
-			bprm->argc,
-			bprm->envc,
-			(interpreter_type == INTERPRETER_ELF ? &elf_ex : NULL),
-			load_addr,    
-			(interpreter_type == INTERPRETER_AOUT ? 0 : 1));
-	if(interpreter_type == INTERPRETER_AOUT)
-	  current->mm->arg_start += strlen(passed_fileno) + 1;
-	current->mm->start_brk = current->mm->brk = elf_brk;
-	current->mm->end_code = end_code;
-	current->mm->start_code = start_code;
-	current->mm->end_data = end_data;
-	current->mm->start_stack = bprm->p;
-	current->suid = current->euid = bprm->e_uid;
-	current->sgid = current->egid = bprm->e_gid;
-
-	/* Calling sys_brk effectively mmaps the pages that we need for the bss and break
-	   sections */
-	current->mm->brk = (elf_bss + 0xfff) & 0xfffff000;
-	sys_brk((elf_brk + 0xfff) & 0xfffff000);
-
-	padzero(elf_bss);
-
-	/* Why this, you ask???  Well SVr4 maps page 0 as read-only,
-	   and some applications "depend" upon this behavior.
-	   Since we do not have the power to recompile these, we
-	   emulate the SVr4 behavior.  Sigh.  */
-	error = do_mmap(NULL, 0, 4096, PROT_READ | PROT_EXEC,
-			MAP_FIXED | MAP_PRIVATE, 0);
-
-	regs->eip = elf_entry;		/* eip, magic happens :-) */
-	regs->esp = bprm->p;			/* stack pointer */
-	if (current->flags & PF_PTRACED)
-		send_sig(SIGTRAP, current, 0);
-	return 0;
-}
-
-/* This is really simpleminded and specialized - we are loading an
-   a.out library that is given an ELF header. */
-
-static int load_elf_library(int fd){
-        struct file * file;
-	struct elfhdr elf_ex;
-	struct elf_phdr *elf_phdata  =  NULL;
-	struct  inode * inode;
-	unsigned int len;
-	int elf_bss;
-	int old_fs, retval;
-	unsigned int bss;
-	int error;
-	int i,j, k;
-	
-	len = 0;
-	file = current->files->fd[fd];
-	inode = file->f_inode;
-	elf_bss = 0;
-	
-	set_fs(KERNEL_DS);
-	if (file->f_op->read(inode, file, (char *) &elf_ex, sizeof(elf_ex)) != sizeof(elf_ex)) {
-		sys_close(fd);
-		return -EACCES;
-	}
-	set_fs(USER_DS);
-	
-	if (elf_ex.e_ident[0] != 0x7f ||
-	    strncmp(&elf_ex.e_ident[1], "ELF",3) != 0)
-		return -ENOEXEC;
-	
-	/* First of all, some simple consistency checks */
-	if(elf_ex.e_type != ET_EXEC || elf_ex.e_phnum > 2 ||
-	   (elf_ex.e_machine != EM_386 && elf_ex.e_machine != EM_486) ||
-	   (!inode->i_op ||
-	    !inode->i_op->default_file_ops->mmap)){
-		return -ENOEXEC;
-	};
-	
-	/* Now read in all of the header information */
-	
-	if(sizeof(struct elf_phdr) * elf_ex.e_phnum > PAGE_SIZE) 
-		return -ENOEXEC;
-	
-	elf_phdata =  (struct elf_phdr *) 
-		kmalloc(sizeof(struct elf_phdr) * elf_ex.e_phnum, GFP_KERNEL);
-	
-	old_fs = get_fs();
-	set_fs(get_ds());
-	retval = read_exec(inode, elf_ex.e_phoff, (char *) elf_phdata,
-			   sizeof(struct elf_phdr) * elf_ex.e_phnum);
-	set_fs(old_fs);
-	
-	j = 0;
-	for(i=0; i<elf_ex.e_phnum; i++)
-		if((elf_phdata + i)->p_type == PT_LOAD) j++;
-	
-	if(j != 1)  {
-		kfree(elf_phdata);
-		return -ENOEXEC;
-	};
-	
-	while(elf_phdata->p_type != PT_LOAD) elf_phdata++;
-	
-	/* Now use mmap to map the library into memory. */
-	error = do_mmap(file,
-			elf_phdata->p_vaddr & 0xfffff000,
-			elf_phdata->p_filesz + (elf_phdata->p_vaddr & 0xfff),
-			PROT_READ | PROT_WRITE | PROT_EXEC,
-			MAP_FIXED | MAP_PRIVATE,
-			elf_phdata->p_offset & 0xfffff000);
-
-	k = elf_phdata->p_vaddr + elf_phdata->p_filesz;
-	if(k > elf_bss) elf_bss = k;
-	
-	sys_close(fd);
-	if (error != elf_phdata->p_vaddr & 0xfffff000) {
-	        kfree(elf_phdata);
-		return error;
-	}
-
-	padzero(elf_bss);
-
-	len = (elf_phdata->p_filesz + elf_phdata->p_vaddr+ 0xfff) & 0xfffff000;
-	bss = elf_phdata->p_memsz + elf_phdata->p_vaddr;
-	if (bss > len)
-	  do_mmap(NULL, len, bss-len,
-		  PROT_READ|PROT_WRITE|PROT_EXEC,
-		  MAP_FIXED|MAP_PRIVATE, 0);
-	kfree(elf_phdata);
-	return 0;
-}
-
-struct linux_binfmt elf_format = { NULL, load_elf_binary, load_elf_library };
diff --git a/arch/i386/ibcs/emulate.c b/arch/i386/ibcs/emulate.c
deleted file mode 100644
index dc3b3d576..000000000
--- a/arch/i386/ibcs/emulate.c
+++ /dev/null
@@ -1,10 +0,0 @@
-/*
- *  linux/abi/emulate.c
- *
- *  Copyright (C) 1993  Linus Torvalds
- */
-
-/*
- * Yes, sir, this file is completely empty, waiting for some real code..
- * I still copyright it, silly me.
- */
diff --git a/arch/i386/kernel/Makefile b/arch/i386/kernel/Makefile
index 701926d26..123dd1207 100644
--- a/arch/i386/kernel/Makefile
+++ b/arch/i386/kernel/Makefile
@@ -7,42 +7,61 @@
 #
 # Note 2! The CFLAGS definitions are now in the main makefile...
 
-.c.s:
-	$(CC) $(CFLAGS) -S $<
-.s.o:
-	$(AS) -o $*.o $<
-.c.o:
-	$(CC) $(CFLAGS) -c $<
-.S.s:
-	$(CPP) -D__ASSEMBLY__ -traditional $< -o $*.s
+#.S.s:
+#	$(CPP) -D__ASSEMBLY__ -traditional $< -o $*.s
+
+ifdef SMP
+.S.o:
+	$(CC) -D__ASSEMBLY__ $(AFLAGS) -traditional -c $< -o $*.o
+else
 .S.o:
 	$(CC) -D__ASSEMBLY__ -traditional -c $< -o $*.o
-
-OBJS  = process.o signal.o entry.o traps.o irq.o vm86.o bios32.o ptrace.o ioport.o ldt.o setup.o
+endif
 
 all: kernel.o head.o
 
-head.o: head.s
+O_TARGET := kernel.o
+O_OBJS   := process.o signal.o entry.o traps.o irq.o vm86.o bios32.o \
+            ptrace.o ioport.o ldt.o setup.o time.o sys_i386.o ksyms.o
 
-head.s: head.S $(TOPDIR)/include/linux/tasks.h
-	$(CPP) -traditional -o $*.s $<
+ifdef SMP
 
-kernel.o: $(OBJS)
-	$(LD) -r -o kernel.o $(OBJS)
-	sync
+O_OBJS += smp.o
 
-dep:
-	$(CPP) -M *.c > .depend
+head.o: head.S $(TOPDIR)/include/linux/tasks.h
+	$(CC) -D__ASSEMBLY__ -D__SMP__ -traditional -c $*.S -o $*.o
 
-modules:
+else
 
-dummy:
+head.o: head.S $(TOPDIR)/include/linux/tasks.h
+	$(CC) -D__ASSEMBLY__ -traditional -c $*.S -o $*.o
 
-#
-# include a dependency file if one exists
-#
-ifeq (.depend,$(wildcard .depend))
-include .depend
 endif
 
+hexify:
+	$(HOSTCC) hexify.c -o hexify
+
+smp.c: trampoline.hex
+
+trampoline.hex: trampoline hexify
+	(dd if=trampoline bs=1 skip=32 | ./hexify >trampoline.hex )
+
+trampoline: trampoline.o trampoline32.o
+	$(LD86) -s -o $@ trampoline.o trampoline32.o
+
+trampoline.o: trampoline.s
+	$(AS86) -o $@ $<
+
+trampoline32.o: trampoline32.s
+	$(AS386) -o $@ $<
+
+trampoline.s: trampoline.S $(CONFIGURE) $(TOPDIR)/include/linux/config.h Makefile
+	$(CPP) -D__SMP__ -traditional $< -o $@
+
+trampoline32.s: trampoline32.S $(CONFIGURE) $(TOPDIR)/include/linux/config.h Makefile
+	$(CPP) -D__SMP__ -traditional $< -o $@
+
+clean:
+	rm -f trampoline hexify
 
+include $(TOPDIR)/Rules.make
diff --git a/arch/i386/kernel/bios32.c b/arch/i386/kernel/bios32.c
index 8cc3d76fa..a1b49a3ae 100644
--- a/arch/i386/kernel/bios32.c
+++ b/arch/i386/kernel/bios32.c
@@ -55,6 +55,7 @@
 #include <linux/bios32.h>
 #include <linux/pci.h>
 
+#include <asm/page.h>
 #include <asm/segment.h>
 
 #define PCIBIOS_PCI_FUNCTION_ID 	0xb1XX
@@ -164,7 +165,7 @@ extern unsigned long check_pcibios(unsigned long memory_start, unsigned long mem
 	int pack;
 
 	if ((pcibios_entry = bios32_service(PCI_SERVICE))) {
-		pci_indirect.address = pcibios_entry;
+		pci_indirect.address = pcibios_entry | PAGE_OFFSET;
 
 		__asm__("lcall (%%edi)\n\t"
 			"jc 1f\n\t"
@@ -363,7 +364,7 @@ int pcibios_write_config_dword (unsigned char bus,
 	return (int) (ret & 0xff00) >> 8;
 }
 
-char *pcibios_strerror (int error)
+const char *pcibios_strerror (int error)
 {
 	static char buf[80];
 
@@ -383,6 +384,12 @@ char *pcibios_strerror (int error)
 		case PCIBIOS_BAD_REGISTER_NUMBER:
 			return "BAD_REGISTER_NUMBER";
 
+                case PCIBIOS_SET_FAILED:          
+			return "SET_FAILED";
+
+                case PCIBIOS_BUFFER_TOO_SMALL:    
+			return "BUFFER_TOO_SMALL";
+
 		default:
 			sprintf (buf, "UNKNOWN RETURN 0x%x", error);
 			return buf;
@@ -411,7 +418,9 @@ unsigned long pcibios_init(unsigned long memory_start, unsigned long memory_end)
 	 *
 	 */
 
-	for (check = (union bios32 *) 0xe0000; check <= (union bios32 *) 0xffff0; ++check) {
+	for (check = (union bios32 *) __va(0xe0000);
+	     check <= (union bios32 *) __va(0xffff0);
+	     ++check) {
 		if (check->fields.signature != BIOS32_SIGNATURE)
 			continue;
 		length = check->fields.length * 16;
@@ -432,17 +441,10 @@ unsigned long pcibios_init(unsigned long memory_start, unsigned long memory_end)
 			if (check->fields.entry >= 0x100000) {
 				printk("pcibios_init: entry in high memory, unable to access\n");
 			} else {
-				bios32_indirect.address = bios32_entry = check->fields.entry;
+				bios32_entry = check->fields.entry;
 				printk ("pcibios_init : BIOS32 Service Directory entry at 0x%lx\n", bios32_entry);
+				bios32_indirect.address = bios32_entry + PAGE_OFFSET;
 			}
-		} else {
-			printk ("pcibios_init : multiple entries, mail drew@colorado.edu\n");
-			/*
-			 * Jeremy Fitzhardinge reports at least one PCI BIOS
-			 * with two different service directories, and as both
-			 * worked for him, we'll just mention the fact, and
-			 * not actually disallow it..
-			 */
 		}
 	}
 #ifdef CONFIG_PCI
@@ -452,6 +454,3 @@ unsigned long pcibios_init(unsigned long memory_start, unsigned long memory_end)
 #endif
 	return memory_start;
 }
-
-
-
diff --git a/arch/i386/kernel/entry.S b/arch/i386/kernel/entry.S
index 1bdf062c8..d421f45f1 100644
--- a/arch/i386/kernel/entry.S
+++ b/arch/i386/kernel/entry.S
@@ -30,18 +30,21 @@
  *	18(%esp) - %eax
  *	1C(%esp) - %ds
  *	20(%esp) - %es
- *      24(%esp) - %fs
- *	28(%esp) - %gs
- *	2C(%esp) - orig_eax
- *	30(%esp) - %eip
- *	34(%esp) - %cs
- *	38(%esp) - %eflags
- *	3C(%esp) - %oldesp
- *	40(%esp) - %oldss
+ *	24(%esp) - orig_eax
+ *	28(%esp) - %eip
+ *	2C(%esp) - %cs
+ *	30(%esp) - %eflags
+ *	34(%esp) - %oldesp
+ *	38(%esp) - %oldss
+ *
+ * "current" is in register %ebx during any slow entries.
  */
 
 #include <linux/sys.h>
+#include <linux/linkage.h>
 #include <asm/segment.h>
+#define ASSEMBLY
+#include <asm/smp.h>
 
 EBX		= 0x00
 ECX		= 0x04
@@ -52,14 +55,12 @@ EBP		= 0x14
 EAX		= 0x18
 DS		= 0x1C
 ES		= 0x20
-FS		= 0x24
-GS		= 0x28
-ORIG_EAX	= 0x2C
-EIP		= 0x30
-CS		= 0x34
-EFLAGS		= 0x38
-OLDESP		= 0x3C
-OLDSS		= 0x40
+ORIG_EAX	= 0x24
+EIP		= 0x28
+CS		= 0x2C
+EFLAGS		= 0x30
+OLDESP		= 0x34
+OLDSS		= 0x38
 
 CF_MASK		= 0x00000001
 IF_MASK		= 0x00000200
@@ -75,26 +76,14 @@ priority	=  8
 signal		= 12
 blocked		= 16
 flags		= 20
-errno		= 24
 dbgreg6		= 52
 dbgreg7		= 56
 exec_domain	= 60
 
 ENOSYS = 38
 
-.globl _system_call,_lcall7
-.globl _device_not_available, _coprocessor_error
-.globl _divide_error,_debug,_nmi,_int3,_overflow,_bounds,_invalid_op
-.globl _double_fault,_coprocessor_segment_overrun
-.globl _invalid_TSS,_segment_not_present,_stack_segment
-.globl _general_protection,_reserved
-.globl _alignment_check,_page_fault
-.globl ret_from_sys_call, _sys_call_table
-
 #define SAVE_ALL \
 	cld; \
-	push %gs; \
-	push %fs; \
 	push %es; \
 	push %ds; \
 	pushl %eax; \
@@ -106,17 +95,123 @@ ENOSYS = 38
 	pushl %ebx; \
 	movl $(KERNEL_DS),%edx; \
 	mov %dx,%ds; \
-	mov %dx,%es; \
-	movl $(USER_DS),%edx; \
-	mov %dx,%fs;
+	mov %dx,%es;
+
+#ifdef	__SMP__
+
+#define GET_PROCESSOR_ID \
+	movl SYMBOL_NAME(apic_reg), %edx; \
+	movl 32(%edx), %eax;\
+	movl %eax,SYMBOL_NAME(apic_retval); \
+	shrl $24,%eax; \
+	andb $0x0F,%al;
+
+/*
+ *	Get the processor ID multiplied by 4
+ */
+
+#define GET_PROCESSOR_OFFSET(x) \
+	movl SYMBOL_NAME(apic_reg), x ; \
+	movl 32( x ), x ; \
+	shrl $22, x ; \
+	andl $0x3C, x ;
+
+/* macro LEAVE_KERNEL decrements kernel_counter and resets kernel_flag and
+   saves processor variables if zero */
+#define	LEAVE_KERNEL \
+	pushfl; \
+	cli; \
+	GET_PROCESSOR_ID \
+	btrl $ SMP_FROM_SYSCALL,SYMBOL_NAME(smp_proc_in_lock)(,%eax,4); \
+	decl SYMBOL_NAME(syscall_count); \
+	decl SYMBOL_NAME(kernel_counter); \
+	jnz 1f;	\
+	movb $(NO_PROC_ID), SYMBOL_NAME(active_kernel_processor); \
+	lock; \
+	btrl $0, SYMBOL_NAME(kernel_flag); \
+1:	popfl;
+
+/* macro ENTER_KERNEL waits for entering the kernel, increments
+   kernel_counter, and reloads the processor variables if necessary
+   uses : %eax, %edx (pushed and popped) 
+
+   Note: We go to great pains to minimise the number of locked operations.
+   We want to spin without locking, and lock when we attempt an update.
+   The pentium has a MESI cache so the spin without lock will exit when
+   another CPU write invalidates our cache, and the lock is avoided when
+   possible so we don't play ping-pong games with the cache line.
+
+*/
+
+#ifndef __SMP_PROF__
+
+#define SMP_PROF_A
+#define SMP_PROF_B 
+
+#else
+
+#define SMP_PROF_A movl $0,SYMBOL_NAME(smp_spins_syscall_cur)(,%eax,4);
+#define SMP_PROF_B incl SYMBOL_NAME(smp_spins_syscall)(,%eax,4); \
+        incl SYMBOL_NAME(smp_spins_syscall_cur)(,%eax,4);
+#endif
+
+#define	ENTER_KERNEL \
+	pushl %eax; \
+	pushl %edx; \
+	pushfl; \
+	cli; \
+	GET_PROCESSOR_ID \
+	btsl $ SMP_FROM_SYSCALL,SYMBOL_NAME(smp_proc_in_lock)(,%eax,4); \
+	SMP_PROF_A \
+1:	lock; \
+	btsl $0, SYMBOL_NAME(kernel_flag); \
+	jnc 3f; \
+	cmpb SYMBOL_NAME(active_kernel_processor), %al; \
+	je 4f; \
+2:	SMP_PROF_B \
+	btl %al, SYMBOL_NAME(smp_invalidate_needed); \
+	jnc 5f; \
+	lock; \
+	btrl %al, SYMBOL_NAME(smp_invalidate_needed); \
+	jnc 5f; \
+	movl %cr3,%edx; \
+	movl %edx,%cr3; \
+5:	btl $0, SYMBOL_NAME(kernel_flag); \
+	jc 2b; \
+	jmp 1b; \
+3:	movb %al, SYMBOL_NAME(active_kernel_processor); \
+4:	incl SYMBOL_NAME(kernel_counter); \
+	incl SYMBOL_NAME(syscall_count); \
+	popfl; \
+	popl %edx; \
+	popl %eax;
+
+
+#define RESTORE_ALL \
+	LEAVE_KERNEL \
+	popl %ebx; \
+	popl %ecx; \
+	popl %edx; \
+	popl %esi; \
+	popl %edi; \
+	popl %ebp; \
+	popl %eax; \
+	pop %ds; \
+	pop %es; \
+	addl $4,%esp; \
+	iret
+
+#define GET_CURRENT \
+	GET_PROCESSOR_OFFSET(%ebx) \
+	movl SYMBOL_NAME(current_set)(%ebx),%ebx
+
+#else
+
+#define GET_CURRENT \
+	movl SYMBOL_NAME(current_set),%ebx
 
 #define RESTORE_ALL \
-	cmpw $(KERNEL_CS),CS(%esp); \
-	je 1f;   \
-	movl _current,%eax; \
-	movl dbgreg7(%eax),%ebx; \
-	movl %ebx,%db7;	\
-1:	popl %ebx; \
+	popl %ebx; \
 	popl %ecx; \
 	popl %edx; \
 	popl %esi; \
@@ -125,16 +220,19 @@ ENOSYS = 38
 	popl %eax; \
 	pop %ds; \
 	pop %es; \
-	pop %fs; \
-	pop %gs; \
 	addl $4,%esp; \
 	iret
+#endif
+
 
-.align 4
-_lcall7:
+ENTRY(lcall7)
 	pushfl			# We get a different stack layout with call gates,
 	pushl %eax		# which has to be cleaned up later..
 	SAVE_ALL
+	GET_CURRENT
+#ifdef __SMP__
+	ENTER_KERNEL
+#endif
 	movl EIP(%esp),%eax	# due to call gates, this is eflags, not eip..
 	movl CS(%esp),%edx	# this is eip..
 	movl EFLAGS(%esp),%ecx	# and this is cs..
@@ -142,7 +240,12 @@ _lcall7:
 	movl %edx,EIP(%esp)	# Now we move them to their "normal" places
 	movl %ecx,CS(%esp)	#
 	movl %esp,%eax
-	movl _current,%edx
+#ifdef __SMP__
+	GET_PROCESSOR_OFFSET(%edx)	# Processor offset into edx
+	movl SYMBOL_NAME(current_set)(,%edx),%edx
+#else
+	movl SYMBOL_NAME(current_set),%edx
+#endif
 	pushl %eax
 	movl exec_domain(%edx),%edx	# Get the execution domain
 	movl 4(%edx),%edx	# Get the lcall7 handler for the domain
@@ -150,398 +253,378 @@ _lcall7:
 	popl %eax
 	jmp ret_from_sys_call
 
-.align 4
+	ALIGN
 handle_bottom_half:
-	pushfl
-	incl _intr_count
-	sti
-	call _do_bottom_half
-	popfl
-	decl _intr_count
+	incl SYMBOL_NAME(intr_count)
+	call SYMBOL_NAME(do_bottom_half)
+	decl SYMBOL_NAME(intr_count)
 	jmp 9f
-.align 4
+	ALIGN
 reschedule:
 	pushl $ret_from_sys_call
-	jmp _schedule
-.align 4
-_system_call:
+	jmp SYMBOL_NAME(schedule)    # test
+
+ENTRY(system_call)
 	pushl %eax			# save orig_eax
 	SAVE_ALL
-	movl $-ENOSYS,EAX(%esp)
+	GET_CURRENT
+#ifdef __SMP__
+	ENTER_KERNEL
+#endif
 	cmpl $(NR_syscalls),%eax
-	jae ret_from_sys_call
-	movl _sys_call_table(,%eax,4),%eax
-	testl %eax,%eax
-	je ret_from_sys_call
-	movl _current,%ebx
-	andl $~CF_MASK,EFLAGS(%esp)	# clear carry - assume no errors
-	movl $0,errno(%ebx)
-	movl %db6,%edx
-	movl %edx,dbgreg6(%ebx)  # save current hardware debugging status
+	jae badsys
 	testb $0x20,flags(%ebx)		# PF_TRACESYS
-	jne 1f
-	call *%eax
-	movl %eax,EAX(%esp)		# save the return value
-	movl errno(%ebx),%edx
-	negl %edx
-	je ret_from_sys_call
-	movl %edx,EAX(%esp)
-	orl $(CF_MASK),EFLAGS(%esp)	# set carry to indicate error
-	jmp ret_from_sys_call
-.align 4
-1:	call _syscall_trace
-	movl ORIG_EAX(%esp),%eax
-	call _sys_call_table(,%eax,4)
+	jne tracesys
+	call SYMBOL_NAME(sys_call_table)(,%eax,4)
 	movl %eax,EAX(%esp)		# save the return value
-	movl _current,%eax
-	movl errno(%eax),%edx
-	negl %edx
-	je 1f
-	movl %edx,EAX(%esp)
-	orl $(CF_MASK),EFLAGS(%esp)	# set carry to indicate error
-1:	call _syscall_trace
-
-	.align 4,0x90
+	ALIGN
+	.globl ret_from_sys_call
 ret_from_sys_call:
-	cmpl $0,_intr_count
-	jne 2f
-9:	movl _bh_mask,%eax
-	andl _bh_active,%eax
-	jne handle_bottom_half
-	movl EFLAGS(%esp),%eax		# check VM86 flag: CS/SS are
-	testl $(VM_MASK),%eax		# different then
+	cmpl $0,SYMBOL_NAME(intr_count)
 	jne 1f
-	cmpw $(KERNEL_CS),CS(%esp)	# was old code segment supervisor ?
-	je 2f
-1:	sti
-	orl $(IF_MASK),%eax		# these just try to make sure
-	andl $~NT_MASK,%eax		# the program doesn't do anything
-	movl %eax,EFLAGS(%esp)		# stupid
-	cmpl $0,_need_resched
-	jne reschedule
-	movl _current,%eax
-	cmpl _task,%eax			# task[0] cannot have signals
-	je 2f
-	cmpl $0,state(%eax)		# state
+9:	movl SYMBOL_NAME(bh_mask),%eax
+	andl SYMBOL_NAME(bh_active),%eax
+	jne handle_bottom_half
+	movl EFLAGS(%esp),%eax		# mix EFLAGS and CS
+	movb CS(%esp),%al
+	testl $(VM_MASK | 3),%eax	# return to VM86 mode or non-supervisor?
+	je 1f
+	cmpl $0,SYMBOL_NAME(need_resched)
 	jne reschedule
-	cmpl $0,counter(%eax)		# counter
-	je reschedule
-	movl blocked(%eax),%ecx
-	movl %ecx,%ebx			# save blocked in %ebx for signal handling
-	notl %ecx
-	andl signal(%eax),%ecx
+	movl blocked(%ebx),%eax
+	movl %eax,%esi			# save blocked in %esi for signal handling
+	notl %eax
+	andl signal(%ebx),%eax
 	jne signal_return
-2:	RESTORE_ALL
-.align 4
+1:	RESTORE_ALL
+	ALIGN
 signal_return:
-	movl %esp,%ecx
-	pushl %ecx
-	testl $(VM_MASK),EFLAGS(%ecx)
+	testl $(VM_MASK),EFLAGS(%esp)
+	pushl %esp
 	jne v86_signal_return
-	pushl %ebx
-	call _do_signal
-	popl %ebx
-	popl %ebx
+	pushl %esi
+	call SYMBOL_NAME(do_signal)
+	addl $8,%esp
 	RESTORE_ALL
-.align 4
+	ALIGN
 v86_signal_return:
-	call _save_v86_state
+	call SYMBOL_NAME(save_v86_state)
 	movl %eax,%esp
 	pushl %eax
-	pushl %ebx
-	call _do_signal
-	popl %ebx
-	popl %ebx
+	pushl %esi
+	call SYMBOL_NAME(do_signal)
+	addl $8,%esp
 	RESTORE_ALL
+	ALIGN
+tracesys:
+	movl $-ENOSYS,EAX(%esp)
+	call SYMBOL_NAME(syscall_trace)
+	movl ORIG_EAX(%esp),%eax
+	call SYMBOL_NAME(sys_call_table)(,%eax,4)
+	movl %eax,EAX(%esp)		# save the return value
+	call SYMBOL_NAME(syscall_trace)
+	jmp ret_from_sys_call
+badsys:
+	movl $-ENOSYS,EAX(%esp)
+	jmp ret_from_sys_call
 
-.align 4
-_divide_error:
+
+ENTRY(divide_error)
 	pushl $0		# no error code
-	pushl $_do_divide_error
-.align 4,0x90
+	pushl $ SYMBOL_NAME(do_divide_error)
+	ALIGN
 error_code:
-	push %fs
-	push %es
 	push %ds
 	pushl %eax
+	xorl %eax,%eax
 	pushl %ebp
 	pushl %edi
 	pushl %esi
 	pushl %edx
+	decl %eax			# eax = -1
 	pushl %ecx
 	pushl %ebx
-	movl $0,%eax
-	movl %eax,%db7			# disable hardware debugging...
+	xorl %ecx,%ecx			# zero ecx
 	cld
-	movl $-1, %eax
+	mov %es,%cx			# get the lower order bits of es
 	xchgl %eax, ORIG_EAX(%esp)	# orig_eax (get the error code. )
-	xorl %ebx,%ebx			# zero ebx
-	mov %gs,%bx			# get the lower order bits of gs
-	xchgl %ebx, GS(%esp)		# get the address and save gs.
+	movl %esp,%edx
+	xchgl %ecx, ES(%esp)		# get the address and save es.
 	pushl %eax			# push the error code
-	lea 4(%esp),%edx
 	pushl %edx
 	movl $(KERNEL_DS),%edx
 	mov %dx,%ds
 	mov %dx,%es
-	movl $(USER_DS),%edx
-	mov %dx,%fs
-	pushl %eax
-	movl _current,%eax
-	movl %db6,%edx
-	movl %edx,dbgreg6(%eax)  # save current hardware debugging status
-	popl %eax
-	call *%ebx
+	GET_CURRENT
+#ifdef __SMP__
+	ENTER_KERNEL
+#endif
+	call *%ecx
 	addl $8,%esp
 	jmp ret_from_sys_call
 
-.align 4
-_coprocessor_error:
+ENTRY(coprocessor_error)
 	pushl $0
-	pushl $_do_coprocessor_error
+	pushl $ SYMBOL_NAME(do_coprocessor_error)
 	jmp error_code
 
-.align 4
-_device_not_available:
+ENTRY(device_not_available)
 	pushl $-1		# mark this as an int
 	SAVE_ALL
+	GET_CURRENT
+#ifdef __SMP__
+	ENTER_KERNEL
+#endif
 	pushl $ret_from_sys_call
 	movl %cr0,%eax
 	testl $0x4,%eax			# EM (math emulation bit)
-	je _math_state_restore
+	je SYMBOL_NAME(math_state_restore)
 	pushl $0		# temporary storage for ORIG_EIP
-	call _math_emulate
+	call  SYMBOL_NAME(math_emulate)
 	addl $4,%esp
 	ret
 
-.align 4
-_debug:
+ENTRY(debug)
 	pushl $0
-	pushl $_do_debug
+	pushl $ SYMBOL_NAME(do_debug)
 	jmp error_code
 
-.align 4
-_nmi:
+ENTRY(nmi)
 	pushl $0
-	pushl $_do_nmi
+	pushl $ SYMBOL_NAME(do_nmi)
 	jmp error_code
 
-.align 4
-_int3:
+ENTRY(int3)
 	pushl $0
-	pushl $_do_int3
+	pushl $ SYMBOL_NAME(do_int3)
 	jmp error_code
 
-.align 4
-_overflow:
+ENTRY(overflow)
 	pushl $0
-	pushl $_do_overflow
+	pushl $ SYMBOL_NAME(do_overflow)
 	jmp error_code
 
-.align 4
-_bounds:
+ENTRY(bounds)
 	pushl $0
-	pushl $_do_bounds
+	pushl $ SYMBOL_NAME(do_bounds)
 	jmp error_code
 
-.align 4
-_invalid_op:
+ENTRY(invalid_op)
 	pushl $0
-	pushl $_do_invalid_op
+	pushl $ SYMBOL_NAME(do_invalid_op)
 	jmp error_code
 
-.align 4
-_coprocessor_segment_overrun:
+ENTRY(coprocessor_segment_overrun)
 	pushl $0
-	pushl $_do_coprocessor_segment_overrun
+	pushl $ SYMBOL_NAME(do_coprocessor_segment_overrun)
 	jmp error_code
 
-.align 4
-_reserved:
+ENTRY(reserved)
 	pushl $0
-	pushl $_do_reserved
+	pushl $ SYMBOL_NAME(do_reserved)
+	jmp error_code
+
+ENTRY(double_fault)
+	pushl $ SYMBOL_NAME(do_double_fault)
 	jmp error_code
 
-.align 4
-_double_fault:
-	pushl $_do_double_fault
+ENTRY(invalid_TSS)
+	pushl $ SYMBOL_NAME(do_invalid_TSS)
 	jmp error_code
 
-.align 4
-_invalid_TSS:
-	pushl $_do_invalid_TSS
+ENTRY(segment_not_present)
+	pushl $ SYMBOL_NAME(do_segment_not_present)
 	jmp error_code
 
-.align 4
-_segment_not_present:
-	pushl $_do_segment_not_present
+ENTRY(stack_segment)
+	pushl $ SYMBOL_NAME(do_stack_segment)
 	jmp error_code
 
-.align 4
-_stack_segment:
-	pushl $_do_stack_segment
+ENTRY(general_protection)
+	pushl $ SYMBOL_NAME(do_general_protection)
 	jmp error_code
 
-.align 4
-_general_protection:
-	pushl $_do_general_protection
+ENTRY(alignment_check)
+	pushl $ SYMBOL_NAME(do_alignment_check)
 	jmp error_code
 
-.align 4
-_alignment_check:
-	pushl $_do_alignment_check
+ENTRY(page_fault)
+	pushl $ SYMBOL_NAME(do_page_fault)
 	jmp error_code
 
-.align 4
-_page_fault:
-	pushl $_do_page_fault
+ENTRY(spurious_interrupt_bug)
+	pushl $0
+	pushl $ SYMBOL_NAME(do_spurious_interrupt_bug)
 	jmp error_code
 
 .data
-.align 4
-_sys_call_table:
-	.long _sys_setup		/* 0 */
-	.long _sys_exit
-	.long _sys_fork
-	.long _sys_read
-	.long _sys_write
-	.long _sys_open			/* 5 */
-	.long _sys_close
-	.long _sys_waitpid
-	.long _sys_creat
-	.long _sys_link
-	.long _sys_unlink		/* 10 */
-	.long _sys_execve
-	.long _sys_chdir
-	.long _sys_time
-	.long _sys_mknod
-	.long _sys_chmod		/* 15 */
-	.long _sys_chown
-	.long _sys_break
-	.long _sys_stat
-	.long _sys_lseek
-	.long _sys_getpid		/* 20 */
-	.long _sys_mount
-	.long _sys_umount
-	.long _sys_setuid
-	.long _sys_getuid
-	.long _sys_stime		/* 25 */
-	.long _sys_ptrace
-	.long _sys_alarm
-	.long _sys_fstat
-	.long _sys_pause
-	.long _sys_utime		/* 30 */
-	.long _sys_stty
-	.long _sys_gtty
-	.long _sys_access
-	.long _sys_nice
-	.long _sys_ftime		/* 35 */
-	.long _sys_sync
-	.long _sys_kill
-	.long _sys_rename
-	.long _sys_mkdir
-	.long _sys_rmdir		/* 40 */
-	.long _sys_dup
-	.long _sys_pipe
-	.long _sys_times
-	.long _sys_prof
-	.long _sys_brk			/* 45 */
-	.long _sys_setgid
-	.long _sys_getgid
-	.long _sys_signal
-	.long _sys_geteuid
-	.long _sys_getegid		/* 50 */
-	.long _sys_acct
-	.long _sys_phys
-	.long _sys_lock
-	.long _sys_ioctl
-	.long _sys_fcntl		/* 55 */
-	.long _sys_mpx
-	.long _sys_setpgid
-	.long _sys_ulimit
-	.long _sys_olduname
-	.long _sys_umask		/* 60 */
-	.long _sys_chroot
-	.long _sys_ustat
-	.long _sys_dup2
-	.long _sys_getppid
-	.long _sys_getpgrp		/* 65 */
-	.long _sys_setsid
-	.long _sys_sigaction
-	.long _sys_sgetmask
-	.long _sys_ssetmask
-	.long _sys_setreuid		/* 70 */
-	.long _sys_setregid
-	.long _sys_sigsuspend
-	.long _sys_sigpending
-	.long _sys_sethostname
-	.long _sys_setrlimit		/* 75 */
-	.long _sys_getrlimit
-	.long _sys_getrusage
-	.long _sys_gettimeofday
-	.long _sys_settimeofday
-	.long _sys_getgroups		/* 80 */
-	.long _sys_setgroups
-	.long _old_select
-	.long _sys_symlink
-	.long _sys_lstat
-	.long _sys_readlink		/* 85 */
-	.long _sys_uselib
-	.long _sys_swapon
-	.long _sys_reboot
-	.long _old_readdir
-	.long _sys_mmap			/* 90 */
-	.long _sys_munmap
-	.long _sys_truncate
-	.long _sys_ftruncate
-	.long _sys_fchmod
-	.long _sys_fchown		/* 95 */
-	.long _sys_getpriority
-	.long _sys_setpriority
-	.long _sys_profil
-	.long _sys_statfs
-	.long _sys_fstatfs		/* 100 */
-	.long _sys_ioperm
-	.long _sys_socketcall
-	.long _sys_syslog
-	.long _sys_setitimer
-	.long _sys_getitimer		/* 105 */
-	.long _sys_newstat
-	.long _sys_newlstat
-	.long _sys_newfstat
-	.long _sys_uname
-	.long _sys_iopl			/* 110 */
-	.long _sys_vhangup
-	.long _sys_idle
-	.long _sys_vm86
-	.long _sys_wait4
-	.long _sys_swapoff		/* 115 */
-	.long _sys_sysinfo
-	.long _sys_ipc
-	.long _sys_fsync
-	.long _sys_sigreturn
-	.long _sys_clone		/* 120 */
-	.long _sys_setdomainname
-	.long _sys_newuname
-	.long _sys_modify_ldt
-	.long _sys_adjtimex
-	.long _sys_mprotect		/* 125 */
-	.long _sys_sigprocmask
-	.long _sys_create_module
-	.long _sys_init_module
-	.long _sys_delete_module
-	.long _sys_get_kernel_syms	/* 130 */
-	.long _sys_quotactl
-	.long _sys_getpgid
-	.long _sys_fchdir
-	.long _sys_bdflush
-	.long _sys_sysfs		/* 135 */
-	.long _sys_personality
-	.long 0				/* for afs_syscall */
-	.long _sys_setfsuid
-	.long _sys_setfsgid
-	.long _sys_llseek		/* 140 */
-	.long _sys_getdents
-	.long _sys_select
-	.long _sys_flock
-	.space (NR_syscalls-140)*4
+ENTRY(sys_call_table)
+	.long SYMBOL_NAME(sys_setup)		/* 0 */
+	.long SYMBOL_NAME(sys_exit)
+	.long SYMBOL_NAME(sys_fork)
+	.long SYMBOL_NAME(sys_read)
+	.long SYMBOL_NAME(sys_write)
+	.long SYMBOL_NAME(sys_open)		/* 5 */
+	.long SYMBOL_NAME(sys_close)
+	.long SYMBOL_NAME(sys_waitpid)
+	.long SYMBOL_NAME(sys_creat)
+	.long SYMBOL_NAME(sys_link)
+	.long SYMBOL_NAME(sys_unlink)		/* 10 */
+	.long SYMBOL_NAME(sys_execve)
+	.long SYMBOL_NAME(sys_chdir)
+	.long SYMBOL_NAME(sys_time)
+	.long SYMBOL_NAME(sys_mknod)
+	.long SYMBOL_NAME(sys_chmod)		/* 15 */
+	.long SYMBOL_NAME(sys_chown)
+	.long SYMBOL_NAME(sys_break)
+	.long SYMBOL_NAME(sys_stat)
+	.long SYMBOL_NAME(sys_lseek)
+	.long SYMBOL_NAME(sys_getpid)		/* 20 */
+	.long SYMBOL_NAME(sys_mount)
+	.long SYMBOL_NAME(sys_umount)
+	.long SYMBOL_NAME(sys_setuid)
+	.long SYMBOL_NAME(sys_getuid)
+	.long SYMBOL_NAME(sys_stime)		/* 25 */
+	.long SYMBOL_NAME(sys_ptrace)
+	.long SYMBOL_NAME(sys_alarm)
+	.long SYMBOL_NAME(sys_fstat)
+	.long SYMBOL_NAME(sys_pause)
+	.long SYMBOL_NAME(sys_utime)		/* 30 */
+	.long SYMBOL_NAME(sys_stty)
+	.long SYMBOL_NAME(sys_gtty)
+	.long SYMBOL_NAME(sys_access)
+	.long SYMBOL_NAME(sys_nice)
+	.long SYMBOL_NAME(sys_ftime)		/* 35 */
+	.long SYMBOL_NAME(sys_sync)
+	.long SYMBOL_NAME(sys_kill)
+	.long SYMBOL_NAME(sys_rename)
+	.long SYMBOL_NAME(sys_mkdir)
+	.long SYMBOL_NAME(sys_rmdir)		/* 40 */
+	.long SYMBOL_NAME(sys_dup)
+	.long SYMBOL_NAME(sys_pipe)
+	.long SYMBOL_NAME(sys_times)
+	.long SYMBOL_NAME(sys_prof)
+	.long SYMBOL_NAME(sys_brk)		/* 45 */
+	.long SYMBOL_NAME(sys_setgid)
+	.long SYMBOL_NAME(sys_getgid)
+	.long SYMBOL_NAME(sys_signal)
+	.long SYMBOL_NAME(sys_geteuid)
+	.long SYMBOL_NAME(sys_getegid)		/* 50 */
+	.long SYMBOL_NAME(sys_acct)
+	.long SYMBOL_NAME(sys_phys)
+	.long SYMBOL_NAME(sys_lock)
+	.long SYMBOL_NAME(sys_ioctl)
+	.long SYMBOL_NAME(sys_fcntl)		/* 55 */
+	.long SYMBOL_NAME(sys_mpx)
+	.long SYMBOL_NAME(sys_setpgid)
+	.long SYMBOL_NAME(sys_ulimit)
+	.long SYMBOL_NAME(sys_olduname)
+	.long SYMBOL_NAME(sys_umask)		/* 60 */
+	.long SYMBOL_NAME(sys_chroot)
+	.long SYMBOL_NAME(sys_ustat)
+	.long SYMBOL_NAME(sys_dup2)
+	.long SYMBOL_NAME(sys_getppid)
+	.long SYMBOL_NAME(sys_getpgrp)		/* 65 */
+	.long SYMBOL_NAME(sys_setsid)
+	.long SYMBOL_NAME(sys_sigaction)
+	.long SYMBOL_NAME(sys_sgetmask)
+	.long SYMBOL_NAME(sys_ssetmask)
+	.long SYMBOL_NAME(sys_setreuid)		/* 70 */
+	.long SYMBOL_NAME(sys_setregid)
+	.long SYMBOL_NAME(sys_sigsuspend)
+	.long SYMBOL_NAME(sys_sigpending)
+	.long SYMBOL_NAME(sys_sethostname)
+	.long SYMBOL_NAME(sys_setrlimit)	/* 75 */
+	.long SYMBOL_NAME(sys_getrlimit)
+	.long SYMBOL_NAME(sys_getrusage)
+	.long SYMBOL_NAME(sys_gettimeofday)
+	.long SYMBOL_NAME(sys_settimeofday)
+	.long SYMBOL_NAME(sys_getgroups)	/* 80 */
+	.long SYMBOL_NAME(sys_setgroups)
+	.long SYMBOL_NAME(old_select)
+	.long SYMBOL_NAME(sys_symlink)
+	.long SYMBOL_NAME(sys_lstat)
+	.long SYMBOL_NAME(sys_readlink)		/* 85 */
+	.long SYMBOL_NAME(sys_uselib)
+	.long SYMBOL_NAME(sys_swapon)
+	.long SYMBOL_NAME(sys_reboot)
+	.long SYMBOL_NAME(old_readdir)
+	.long SYMBOL_NAME(old_mmap)		/* 90 */
+	.long SYMBOL_NAME(sys_munmap)
+	.long SYMBOL_NAME(sys_truncate)
+	.long SYMBOL_NAME(sys_ftruncate)
+	.long SYMBOL_NAME(sys_fchmod)
+	.long SYMBOL_NAME(sys_fchown)		/* 95 */
+	.long SYMBOL_NAME(sys_getpriority)
+	.long SYMBOL_NAME(sys_setpriority)
+	.long SYMBOL_NAME(sys_profil)
+	.long SYMBOL_NAME(sys_statfs)
+	.long SYMBOL_NAME(sys_fstatfs)		/* 100 */
+	.long SYMBOL_NAME(sys_ioperm)
+	.long SYMBOL_NAME(sys_socketcall)
+	.long SYMBOL_NAME(sys_syslog)
+	.long SYMBOL_NAME(sys_setitimer)
+	.long SYMBOL_NAME(sys_getitimer)	/* 105 */
+	.long SYMBOL_NAME(sys_newstat)
+	.long SYMBOL_NAME(sys_newlstat)
+	.long SYMBOL_NAME(sys_newfstat)
+	.long SYMBOL_NAME(sys_uname)
+	.long SYMBOL_NAME(sys_iopl)		/* 110 */
+	.long SYMBOL_NAME(sys_vhangup)
+	.long SYMBOL_NAME(sys_idle)
+	.long SYMBOL_NAME(sys_vm86)
+	.long SYMBOL_NAME(sys_wait4)
+	.long SYMBOL_NAME(sys_swapoff)		/* 115 */
+	.long SYMBOL_NAME(sys_sysinfo)
+	.long SYMBOL_NAME(sys_ipc)
+	.long SYMBOL_NAME(sys_fsync)
+	.long SYMBOL_NAME(sys_sigreturn)
+	.long SYMBOL_NAME(sys_clone)		/* 120 */
+	.long SYMBOL_NAME(sys_setdomainname)
+	.long SYMBOL_NAME(sys_newuname)
+	.long SYMBOL_NAME(sys_modify_ldt)
+	.long SYMBOL_NAME(sys_adjtimex)
+	.long SYMBOL_NAME(sys_mprotect)		/* 125 */
+	.long SYMBOL_NAME(sys_sigprocmask)
+	.long SYMBOL_NAME(sys_create_module)
+	.long SYMBOL_NAME(sys_init_module)
+	.long SYMBOL_NAME(sys_delete_module)
+	.long SYMBOL_NAME(sys_get_kernel_syms)	/* 130 */
+	.long SYMBOL_NAME(sys_quotactl)
+	.long SYMBOL_NAME(sys_getpgid)
+	.long SYMBOL_NAME(sys_fchdir)
+	.long SYMBOL_NAME(sys_bdflush)
+	.long SYMBOL_NAME(sys_sysfs)		/* 135 */
+	.long SYMBOL_NAME(sys_personality)
+	.long SYMBOL_NAME(sys_ni_syscall)	/* for afs_syscall */
+	.long SYMBOL_NAME(sys_setfsuid)
+	.long SYMBOL_NAME(sys_setfsgid)
+	.long SYMBOL_NAME(sys_llseek)		/* 140 */
+	.long SYMBOL_NAME(sys_getdents)
+	.long SYMBOL_NAME(sys_select)
+	.long SYMBOL_NAME(sys_flock)
+	.long SYMBOL_NAME(sys_msync)
+	.long SYMBOL_NAME(sys_readv)		/* 145 */
+	.long SYMBOL_NAME(sys_writev)
+	.long SYMBOL_NAME(sys_getsid)
+	.long SYMBOL_NAME(sys_fdatasync)
+	.long SYMBOL_NAME(sys_sysctl)
+	.long SYMBOL_NAME(sys_mlock)		/* 150 */
+	.long SYMBOL_NAME(sys_munlock)
+	.long SYMBOL_NAME(sys_mlockall)
+	.long SYMBOL_NAME(sys_munlockall)
+	.long SYMBOL_NAME(sys_sched_setparam)
+	.long SYMBOL_NAME(sys_sched_getparam)   /* 155 */
+	.long SYMBOL_NAME(sys_sched_setscheduler)
+	.long SYMBOL_NAME(sys_sched_getscheduler)
+	.long SYMBOL_NAME(sys_sched_yield)
+	.long SYMBOL_NAME(sys_sched_get_priority_max)
+	.long SYMBOL_NAME(sys_sched_get_priority_min)  /* 160 */
+	.long SYMBOL_NAME(sys_sched_rr_get_interval)
+	.long SYMBOL_NAME(sys_nanosleep)
+	.long SYMBOL_NAME(sys_mremap)
+	.long SYMBOL_NAME(sys_setresuid)
+	.long SYMBOL_NAME(sys_getresuid)
+	.rept NR_syscalls-165
+		.long SYMBOL_NAME(sys_ni_syscall)
+	.endr
diff --git a/arch/i386/kernel/head.S b/arch/i386/kernel/head.S
index a3e0df213..fdf38c818 100644
--- a/arch/i386/kernel/head.S
+++ b/arch/i386/kernel/head.S
@@ -9,17 +9,10 @@
  */
 
 .text
-.globl _idt,_gdt,_stext,__stext
-.globl _swapper_pg_dir,_pg0
-.globl _empty_bad_page
-.globl _empty_bad_page_table
-.globl _empty_zero_page
-.globl _floppy_track_buffer
-
-#define __ASSEMBLY__
 #include <linux/tasks.h>
-#include <linux/fd.h>
+#include <linux/linkage.h>
 #include <asm/segment.h>
+#include <linux/config.h>
 
 #define CL_MAGIC_ADDR	0x90020
 #define CL_MAGIC	0xA33F
@@ -27,25 +20,56 @@
 #define CL_OFFSET	0x90022
 
 /*
- * swapper_pg_dir is the main page directory, address 0x00001000 (or at
- * address 0x00101000 for a compressed boot).
+ * swapper_pg_dir is the main page directory, address 0x00101000
  */
-_stext:
-__stext:
+ENTRY(stext)
+ENTRY(_stext)
 startup_32:
+/*
+ * Set segments to known values
+ */
 	cld
 	movl $(KERNEL_DS),%eax
 	mov %ax,%ds
 	mov %ax,%es
 	mov %ax,%fs
 	mov %ax,%gs
+/*
+ * Setup paging (the tables are already set up, just switch them on)
+ */
+	movl $0x101000,%eax
+	movl %eax,%cr3		/* set the page table pointer.. */
+	movl %cr0,%eax
+	orl $0x80000000,%eax
+	movl %eax,%cr0		/* ..and set paging (PG) bit */
+	jmp 1f			/* flush the prefetch-queue */
+1:
+	movl $1f,%eax
+	jmp *%eax		/* make sure eip is relocated */
+1:
+
+#ifdef __SMP__
+	orw  %bx,%bx
+	jz  1f				/* Initial CPU cleans BSS */
+/*
+ *	Set up the stack
+ */
+	mov %ax,%ss
+	xorl %eax,%eax
+	movw %cx, %ax
+	movl %eax,%esp
+	pushl $0
+	popfl
+	jmp checkCPUtype
+1:
 	lss stack_start,%esp
+#endif __SMP__
 /*
  * Clear BSS first so that there are no surprises...
  */
 	xorl %eax,%eax
-	movl $__edata,%edi
-	movl $__end,%ecx
+	movl $ SYMBOL_NAME(_edata),%edi
+	movl $ SYMBOL_NAME(_end),%ecx
 	subl %edi,%ecx
 	cld
 	rep
@@ -73,7 +97,7 @@ startup_32:
  * is for the command line.
  */
 	movl $0x90000,%esi
-	movl $_empty_zero_page,%edi
+	movl $ SYMBOL_NAME(empty_zero_page),%edi
 	movl $512,%ecx
 	cld
 	rep
@@ -84,20 +108,24 @@ startup_32:
 	stosl
 	cmpw $(CL_MAGIC),CL_MAGIC_ADDR
 	jne 1f
-	movl $_empty_zero_page+2048,%edi
+	movl $ SYMBOL_NAME(empty_zero_page)+2048,%edi
 	movzwl CL_OFFSET,%esi
 	addl $(CL_BASE_ADDR),%esi
 	movl $2048,%ecx
 	rep
 	movsb
 1:
+#ifdef __SMP__
+checkCPUtype:
+#endif
+
 /* check if it is 486 or 386. */
 /*
  * XXX - this does a lot of unnecessary setup.  Alignment checks don't
  * apply at our cpl of 0 and the stack ought to be aligned already, and
  * we don't need to preserve eflags.
  */
-	movl $3,_x86
+	movl $3, SYMBOL_NAME(x86)
 	pushfl			# push EFLAGS
 	popl %eax		# get EFLAGS
 	movl %eax,%ecx		# save original EFLAGS
@@ -109,7 +137,7 @@ startup_32:
 	xorl %ecx,%eax		# change in flags
 	andl $0x40000,%eax	# check if AC bit changed
 	je is386
-	movl $4,_x86
+	movl $4,SYMBOL_NAME(x86)
 	movl %ecx,%eax
 	xorl $0x200000,%eax	# check ID flag
 	pushl %eax
@@ -121,24 +149,25 @@ startup_32:
 	je is486
 isnew:	pushl %ecx		# restore original EFLAGS
 	popfl
+	incl SYMBOL_NAME(have_cpuid)	# we have CPUID
 	/* get processor type */
 	movl $1, %eax		# Use the CPUID instruction to 
 	.byte 0x0f, 0xa2	# check the processor type
 	movb %al, %cl		# save reg for future use
 	andb $0x0f,%ah		# mask processor family
-	movb %ah, _x86
+	movb %ah,SYMBOL_NAME(x86)
 	andb $0xf0, %eax	# mask model
 	shrb $4, %al
-	movb %al, _x86_model
+	movb %al,SYMBOL_NAME(x86_model)
 	andb $0x0f, %cl		# mask mask revision
-	movb %cl, _x86_mask
-	movl %edx, _x86_capability
+	movb %cl,SYMBOL_NAME(x86_mask)
+	movl %edx,SYMBOL_NAME(x86_capability)
 	/* get vendor info */
 	xorl %eax, %eax			# call CPUID with 0 -> return vendor ID
 	.byte 0x0f, 0xa2		# CPUID
-	movl %ebx, _x86_vendor_id	# lo 4 chars
-	movl %edx, _x86_vendor_id+4	# next 4 chars
-	movl %ecx, _x86_vendor_id+8	# last 4 chars
+	movl %ebx,SYMBOL_NAME(x86_vendor_id)	# lo 4 chars
+	movl %edx,SYMBOL_NAME(x86_vendor_id)+4	# next 4 chars
+	movl %ecx,SYMBOL_NAME(x86_vendor_id)+8	# last 4 chars
 
 	movl %cr0,%eax		# 486+
 	andl $0x80000011,%eax	# Save PG,PE,ET
@@ -157,7 +186,26 @@ is386:	pushl %ecx		# restore original EFLAGS
 	orl $2,%eax		# set MP
 2:	movl %eax,%cr0
 	call check_x87
-	call setup_paging
+#ifdef __SMP__
+	movb ready,%eax
+	orb %eax,%eax
+	jz 3f	
+#ifdef GAS_KNOWS_CR4
+	movl %cr4,%eax
+	orl $16,%eax
+	movl %eax,%cr4
+#else
+	.byte 0x0f,0x20,0xe0
+	orl $16,%eax
+	.byte 0x0f,0x22,0xe0
+#endif
+	jmp 4f
+#endif
+3:
+#ifdef __SMP__
+	incb ready
+#endif
+4:
 	lgdt gdt_descr
 	lidt idt_descr
 	ljmp $(KERNEL_CS),$1f
@@ -166,23 +214,32 @@ is386:	pushl %ecx		# restore original EFLAGS
 	mov %ax,%es
 	mov %ax,%fs
 	mov %ax,%gs
-	lss stack_start,%esp
+#ifdef __SMP__
+	movl $(KERNEL_DS), %eax
+	mov  %ax,%ss		# Reload the stack pointer (segment only)
+#else
+	lss stack_start,%esp	# Load processor stack
+#endif
 	xorl %eax,%eax
 	lldt %ax
 	pushl %eax		# These are the parameters to main :-)
 	pushl %eax
 	pushl %eax
 	cld			# gcc2 wants the direction flag cleared at all times
-	call _start_kernel
+	call SYMBOL_NAME(start_kernel)
 L6:
 	jmp L6			# main should never return here, but
 				# just in case, we know what happens.
 
+#ifdef __SMP__
+ready:	.byte 0
+#endif
+
 /*
  * We depend on ET to be correct. This checks for 287/387.
  */
 check_x87:
-	movb $0,_hard_math
+	movb $0,SYMBOL_NAME(hard_math)
 	clts
 	fninit
 	fstsw %ax
@@ -192,8 +249,8 @@ check_x87:
 	xorl $4,%eax		/* set EM */
 	movl %eax,%cr0
 	ret
-.align 2
-1:	movb $1,_hard_math
+	ALIGN
+1:	movb $1,SYMBOL_NAME(hard_math)
 	.byte 0xDB,0xE4		/* fsetpm for 287, ignored by 387 */
 	ret
 
@@ -213,7 +270,7 @@ setup_idt:
 	movw %dx,%ax		/* selector = 0x0010 = cs */
 	movw $0x8E00,%dx	/* interrupt gate - dpl=0, present */
 
-	lea _idt,%edi
+	lea SYMBOL_NAME(idt),%edi
 	mov $256,%ecx
 rp_sidt:
 	movl %eax,(%edi)
@@ -225,86 +282,179 @@ rp_sidt:
 
 
 /*
- * Setup_paging
- *
- * This routine sets up paging by setting the page bit
- * in cr0. The page tables are set up, identity-mapping
- * the first 4MB.  The rest are initialized later.
- *
- * (ref: added support for up to 32mb, 17Apr92)  -- Rik Faith
- * (ref: update, 25Sept92)  -- croutons@crunchy.uucp 
- * (ref: 92.10.11 - Linus Torvalds. Corrected 16M limit - no upper memory limit)
- */
-.align 2
-setup_paging:
-	movl $1024*2,%ecx		/* 2 pages - swapper_pg_dir+1 page table */
-	xorl %eax,%eax
-	movl $_swapper_pg_dir,%edi	/* swapper_pg_dir is at 0x1000 */
-	cld;rep;stosl
-/* Identity-map the kernel in low 4MB memory for ease of transition */
-	movl $_pg0+7,_swapper_pg_dir		/* set present bit/user r/w */
-/* But the real place is at 0xC0000000 */
-	movl $_pg0+7,_swapper_pg_dir+3072	/* set present bit/user r/w */
-	movl $_pg0+4092,%edi
-	movl $0x03ff007,%eax		/*  4Mb - 4096 + 7 (r/w user,p) */
-	std
-1:	stosl			/* fill the page backwards - more efficient :-) */
-	subl $0x1000,%eax
-	jge 1b
-	cld
-	movl $_swapper_pg_dir,%eax
-	movl %eax,%cr3			/* cr3 - page directory start */
-	movl %cr0,%eax
-	orl $0x80000000,%eax
-	movl %eax,%cr0		/* set paging (PG) bit */
-	ret			/* this also flushes the prefetch-queue */
-
-/*
  * page 0 is made non-existent, so that kernel NULL pointer references get
- * caught. Thus the swapper page directory has been moved to 0x1000
- *
- * XXX Actually, the swapper page directory is at 0x1000 plus 1 megabyte,
+ * caught. Thus the swapper page directory has been moved to 0x101000
  * with the introduction of the compressed boot code.  Theoretically,
  * the original design of overlaying the startup code with the swapper
  * page directory is still possible --- it would reduce the size of the kernel
  * by 2-3k.  This would be a good thing to do at some point.....
+ *
+ * This is initialized to create a identity-mapping at 0-4M (for bootup
+ * purposes) and another mapping of the 0-4M area at virtual address
+ * 0xC0000000.
  */
 .org 0x1000
-_swapper_pg_dir:
+ENTRY(swapper_pg_dir)
+	.long 0x00102007
+	.fill 767,4,0
+	.long 0x00102007
+	.fill 127,4,0
+
 /*
  * The page tables are initialized to only 4MB here - the final page
- * tables are set up later depending on memory size.
+ * tables are set up later depending on memory size. The "007" at the
+ * end doesn't mean with right to kill, but PRESENT+RW+USER
  */
 .org 0x2000
-_pg0:
+ENTRY(pg0)
+	.long 0x000007,0x001007,0x002007,0x003007,0x004007,0x005007,0x006007,0x007007
+	.long 0x008007,0x009007,0x00a007,0x00b007,0x00c007,0x00d007,0x00e007,0x00f007
+	.long 0x010007,0x011007,0x012007,0x013007,0x014007,0x015007,0x016007,0x017007
+	.long 0x018007,0x019007,0x01a007,0x01b007,0x01c007,0x01d007,0x01e007,0x01f007
+	.long 0x020007,0x021007,0x022007,0x023007,0x024007,0x025007,0x026007,0x027007
+	.long 0x028007,0x029007,0x02a007,0x02b007,0x02c007,0x02d007,0x02e007,0x02f007
+	.long 0x030007,0x031007,0x032007,0x033007,0x034007,0x035007,0x036007,0x037007
+	.long 0x038007,0x039007,0x03a007,0x03b007,0x03c007,0x03d007,0x03e007,0x03f007
+	.long 0x040007,0x041007,0x042007,0x043007,0x044007,0x045007,0x046007,0x047007
+	.long 0x048007,0x049007,0x04a007,0x04b007,0x04c007,0x04d007,0x04e007,0x04f007
+	.long 0x050007,0x051007,0x052007,0x053007,0x054007,0x055007,0x056007,0x057007
+	.long 0x058007,0x059007,0x05a007,0x05b007,0x05c007,0x05d007,0x05e007,0x05f007
+	.long 0x060007,0x061007,0x062007,0x063007,0x064007,0x065007,0x066007,0x067007
+	.long 0x068007,0x069007,0x06a007,0x06b007,0x06c007,0x06d007,0x06e007,0x06f007
+	.long 0x070007,0x071007,0x072007,0x073007,0x074007,0x075007,0x076007,0x077007
+	.long 0x078007,0x079007,0x07a007,0x07b007,0x07c007,0x07d007,0x07e007,0x07f007
+	.long 0x080007,0x081007,0x082007,0x083007,0x084007,0x085007,0x086007,0x087007
+	.long 0x088007,0x089007,0x08a007,0x08b007,0x08c007,0x08d007,0x08e007,0x08f007
+	.long 0x090007,0x091007,0x092007,0x093007,0x094007,0x095007,0x096007,0x097007
+	.long 0x098007,0x099007,0x09a007,0x09b007,0x09c007,0x09d007,0x09e007,0x09f007
+	.long 0x0a0007,0x0a1007,0x0a2007,0x0a3007,0x0a4007,0x0a5007,0x0a6007,0x0a7007
+	.long 0x0a8007,0x0a9007,0x0aa007,0x0ab007,0x0ac007,0x0ad007,0x0ae007,0x0af007
+	.long 0x0b0007,0x0b1007,0x0b2007,0x0b3007,0x0b4007,0x0b5007,0x0b6007,0x0b7007
+	.long 0x0b8007,0x0b9007,0x0ba007,0x0bb007,0x0bc007,0x0bd007,0x0be007,0x0bf007
+	.long 0x0c0007,0x0c1007,0x0c2007,0x0c3007,0x0c4007,0x0c5007,0x0c6007,0x0c7007
+	.long 0x0c8007,0x0c9007,0x0ca007,0x0cb007,0x0cc007,0x0cd007,0x0ce007,0x0cf007
+	.long 0x0d0007,0x0d1007,0x0d2007,0x0d3007,0x0d4007,0x0d5007,0x0d6007,0x0d7007
+	.long 0x0d8007,0x0d9007,0x0da007,0x0db007,0x0dc007,0x0dd007,0x0de007,0x0df007
+	.long 0x0e0007,0x0e1007,0x0e2007,0x0e3007,0x0e4007,0x0e5007,0x0e6007,0x0e7007
+	.long 0x0e8007,0x0e9007,0x0ea007,0x0eb007,0x0ec007,0x0ed007,0x0ee007,0x0ef007
+	.long 0x0f0007,0x0f1007,0x0f2007,0x0f3007,0x0f4007,0x0f5007,0x0f6007,0x0f7007
+	.long 0x0f8007,0x0f9007,0x0fa007,0x0fb007,0x0fc007,0x0fd007,0x0fe007,0x0ff007
+	.long 0x100007,0x101007,0x102007,0x103007,0x104007,0x105007,0x106007,0x107007
+	.long 0x108007,0x109007,0x10a007,0x10b007,0x10c007,0x10d007,0x10e007,0x10f007
+	.long 0x110007,0x111007,0x112007,0x113007,0x114007,0x115007,0x116007,0x117007
+	.long 0x118007,0x119007,0x11a007,0x11b007,0x11c007,0x11d007,0x11e007,0x11f007
+	.long 0x120007,0x121007,0x122007,0x123007,0x124007,0x125007,0x126007,0x127007
+	.long 0x128007,0x129007,0x12a007,0x12b007,0x12c007,0x12d007,0x12e007,0x12f007
+	.long 0x130007,0x131007,0x132007,0x133007,0x134007,0x135007,0x136007,0x137007
+	.long 0x138007,0x139007,0x13a007,0x13b007,0x13c007,0x13d007,0x13e007,0x13f007
+	.long 0x140007,0x141007,0x142007,0x143007,0x144007,0x145007,0x146007,0x147007
+	.long 0x148007,0x149007,0x14a007,0x14b007,0x14c007,0x14d007,0x14e007,0x14f007
+	.long 0x150007,0x151007,0x152007,0x153007,0x154007,0x155007,0x156007,0x157007
+	.long 0x158007,0x159007,0x15a007,0x15b007,0x15c007,0x15d007,0x15e007,0x15f007
+	.long 0x160007,0x161007,0x162007,0x163007,0x164007,0x165007,0x166007,0x167007
+	.long 0x168007,0x169007,0x16a007,0x16b007,0x16c007,0x16d007,0x16e007,0x16f007
+	.long 0x170007,0x171007,0x172007,0x173007,0x174007,0x175007,0x176007,0x177007
+	.long 0x178007,0x179007,0x17a007,0x17b007,0x17c007,0x17d007,0x17e007,0x17f007
+	.long 0x180007,0x181007,0x182007,0x183007,0x184007,0x185007,0x186007,0x187007
+	.long 0x188007,0x189007,0x18a007,0x18b007,0x18c007,0x18d007,0x18e007,0x18f007
+	.long 0x190007,0x191007,0x192007,0x193007,0x194007,0x195007,0x196007,0x197007
+	.long 0x198007,0x199007,0x19a007,0x19b007,0x19c007,0x19d007,0x19e007,0x19f007
+	.long 0x1a0007,0x1a1007,0x1a2007,0x1a3007,0x1a4007,0x1a5007,0x1a6007,0x1a7007
+	.long 0x1a8007,0x1a9007,0x1aa007,0x1ab007,0x1ac007,0x1ad007,0x1ae007,0x1af007
+	.long 0x1b0007,0x1b1007,0x1b2007,0x1b3007,0x1b4007,0x1b5007,0x1b6007,0x1b7007
+	.long 0x1b8007,0x1b9007,0x1ba007,0x1bb007,0x1bc007,0x1bd007,0x1be007,0x1bf007
+	.long 0x1c0007,0x1c1007,0x1c2007,0x1c3007,0x1c4007,0x1c5007,0x1c6007,0x1c7007
+	.long 0x1c8007,0x1c9007,0x1ca007,0x1cb007,0x1cc007,0x1cd007,0x1ce007,0x1cf007
+	.long 0x1d0007,0x1d1007,0x1d2007,0x1d3007,0x1d4007,0x1d5007,0x1d6007,0x1d7007
+	.long 0x1d8007,0x1d9007,0x1da007,0x1db007,0x1dc007,0x1dd007,0x1de007,0x1df007
+	.long 0x1e0007,0x1e1007,0x1e2007,0x1e3007,0x1e4007,0x1e5007,0x1e6007,0x1e7007
+	.long 0x1e8007,0x1e9007,0x1ea007,0x1eb007,0x1ec007,0x1ed007,0x1ee007,0x1ef007
+	.long 0x1f0007,0x1f1007,0x1f2007,0x1f3007,0x1f4007,0x1f5007,0x1f6007,0x1f7007
+	.long 0x1f8007,0x1f9007,0x1fa007,0x1fb007,0x1fc007,0x1fd007,0x1fe007,0x1ff007
+	.long 0x200007,0x201007,0x202007,0x203007,0x204007,0x205007,0x206007,0x207007
+	.long 0x208007,0x209007,0x20a007,0x20b007,0x20c007,0x20d007,0x20e007,0x20f007
+	.long 0x210007,0x211007,0x212007,0x213007,0x214007,0x215007,0x216007,0x217007
+	.long 0x218007,0x219007,0x21a007,0x21b007,0x21c007,0x21d007,0x21e007,0x21f007
+	.long 0x220007,0x221007,0x222007,0x223007,0x224007,0x225007,0x226007,0x227007
+	.long 0x228007,0x229007,0x22a007,0x22b007,0x22c007,0x22d007,0x22e007,0x22f007
+	.long 0x230007,0x231007,0x232007,0x233007,0x234007,0x235007,0x236007,0x237007
+	.long 0x238007,0x239007,0x23a007,0x23b007,0x23c007,0x23d007,0x23e007,0x23f007
+	.long 0x240007,0x241007,0x242007,0x243007,0x244007,0x245007,0x246007,0x247007
+	.long 0x248007,0x249007,0x24a007,0x24b007,0x24c007,0x24d007,0x24e007,0x24f007
+	.long 0x250007,0x251007,0x252007,0x253007,0x254007,0x255007,0x256007,0x257007
+	.long 0x258007,0x259007,0x25a007,0x25b007,0x25c007,0x25d007,0x25e007,0x25f007
+	.long 0x260007,0x261007,0x262007,0x263007,0x264007,0x265007,0x266007,0x267007
+	.long 0x268007,0x269007,0x26a007,0x26b007,0x26c007,0x26d007,0x26e007,0x26f007
+	.long 0x270007,0x271007,0x272007,0x273007,0x274007,0x275007,0x276007,0x277007
+	.long 0x278007,0x279007,0x27a007,0x27b007,0x27c007,0x27d007,0x27e007,0x27f007
+	.long 0x280007,0x281007,0x282007,0x283007,0x284007,0x285007,0x286007,0x287007
+	.long 0x288007,0x289007,0x28a007,0x28b007,0x28c007,0x28d007,0x28e007,0x28f007
+	.long 0x290007,0x291007,0x292007,0x293007,0x294007,0x295007,0x296007,0x297007
+	.long 0x298007,0x299007,0x29a007,0x29b007,0x29c007,0x29d007,0x29e007,0x29f007
+	.long 0x2a0007,0x2a1007,0x2a2007,0x2a3007,0x2a4007,0x2a5007,0x2a6007,0x2a7007
+	.long 0x2a8007,0x2a9007,0x2aa007,0x2ab007,0x2ac007,0x2ad007,0x2ae007,0x2af007
+	.long 0x2b0007,0x2b1007,0x2b2007,0x2b3007,0x2b4007,0x2b5007,0x2b6007,0x2b7007
+	.long 0x2b8007,0x2b9007,0x2ba007,0x2bb007,0x2bc007,0x2bd007,0x2be007,0x2bf007
+	.long 0x2c0007,0x2c1007,0x2c2007,0x2c3007,0x2c4007,0x2c5007,0x2c6007,0x2c7007
+	.long 0x2c8007,0x2c9007,0x2ca007,0x2cb007,0x2cc007,0x2cd007,0x2ce007,0x2cf007
+	.long 0x2d0007,0x2d1007,0x2d2007,0x2d3007,0x2d4007,0x2d5007,0x2d6007,0x2d7007
+	.long 0x2d8007,0x2d9007,0x2da007,0x2db007,0x2dc007,0x2dd007,0x2de007,0x2df007
+	.long 0x2e0007,0x2e1007,0x2e2007,0x2e3007,0x2e4007,0x2e5007,0x2e6007,0x2e7007
+	.long 0x2e8007,0x2e9007,0x2ea007,0x2eb007,0x2ec007,0x2ed007,0x2ee007,0x2ef007
+	.long 0x2f0007,0x2f1007,0x2f2007,0x2f3007,0x2f4007,0x2f5007,0x2f6007,0x2f7007
+	.long 0x2f8007,0x2f9007,0x2fa007,0x2fb007,0x2fc007,0x2fd007,0x2fe007,0x2ff007
+	.long 0x300007,0x301007,0x302007,0x303007,0x304007,0x305007,0x306007,0x307007
+	.long 0x308007,0x309007,0x30a007,0x30b007,0x30c007,0x30d007,0x30e007,0x30f007
+	.long 0x310007,0x311007,0x312007,0x313007,0x314007,0x315007,0x316007,0x317007
+	.long 0x318007,0x319007,0x31a007,0x31b007,0x31c007,0x31d007,0x31e007,0x31f007
+	.long 0x320007,0x321007,0x322007,0x323007,0x324007,0x325007,0x326007,0x327007
+	.long 0x328007,0x329007,0x32a007,0x32b007,0x32c007,0x32d007,0x32e007,0x32f007
+	.long 0x330007,0x331007,0x332007,0x333007,0x334007,0x335007,0x336007,0x337007
+	.long 0x338007,0x339007,0x33a007,0x33b007,0x33c007,0x33d007,0x33e007,0x33f007
+	.long 0x340007,0x341007,0x342007,0x343007,0x344007,0x345007,0x346007,0x347007
+	.long 0x348007,0x349007,0x34a007,0x34b007,0x34c007,0x34d007,0x34e007,0x34f007
+	.long 0x350007,0x351007,0x352007,0x353007,0x354007,0x355007,0x356007,0x357007
+	.long 0x358007,0x359007,0x35a007,0x35b007,0x35c007,0x35d007,0x35e007,0x35f007
+	.long 0x360007,0x361007,0x362007,0x363007,0x364007,0x365007,0x366007,0x367007
+	.long 0x368007,0x369007,0x36a007,0x36b007,0x36c007,0x36d007,0x36e007,0x36f007
+	.long 0x370007,0x371007,0x372007,0x373007,0x374007,0x375007,0x376007,0x377007
+	.long 0x378007,0x379007,0x37a007,0x37b007,0x37c007,0x37d007,0x37e007,0x37f007
+	.long 0x380007,0x381007,0x382007,0x383007,0x384007,0x385007,0x386007,0x387007
+	.long 0x388007,0x389007,0x38a007,0x38b007,0x38c007,0x38d007,0x38e007,0x38f007
+	.long 0x390007,0x391007,0x392007,0x393007,0x394007,0x395007,0x396007,0x397007
+	.long 0x398007,0x399007,0x39a007,0x39b007,0x39c007,0x39d007,0x39e007,0x39f007
+	.long 0x3a0007,0x3a1007,0x3a2007,0x3a3007,0x3a4007,0x3a5007,0x3a6007,0x3a7007
+	.long 0x3a8007,0x3a9007,0x3aa007,0x3ab007,0x3ac007,0x3ad007,0x3ae007,0x3af007
+	.long 0x3b0007,0x3b1007,0x3b2007,0x3b3007,0x3b4007,0x3b5007,0x3b6007,0x3b7007
+	.long 0x3b8007,0x3b9007,0x3ba007,0x3bb007,0x3bc007,0x3bd007,0x3be007,0x3bf007
+	.long 0x3c0007,0x3c1007,0x3c2007,0x3c3007,0x3c4007,0x3c5007,0x3c6007,0x3c7007
+	.long 0x3c8007,0x3c9007,0x3ca007,0x3cb007,0x3cc007,0x3cd007,0x3ce007,0x3cf007
+	.long 0x3d0007,0x3d1007,0x3d2007,0x3d3007,0x3d4007,0x3d5007,0x3d6007,0x3d7007
+	.long 0x3d8007,0x3d9007,0x3da007,0x3db007,0x3dc007,0x3dd007,0x3de007,0x3df007
+	.long 0x3e0007,0x3e1007,0x3e2007,0x3e3007,0x3e4007,0x3e5007,0x3e6007,0x3e7007
+	.long 0x3e8007,0x3e9007,0x3ea007,0x3eb007,0x3ec007,0x3ed007,0x3ee007,0x3ef007
+	.long 0x3f0007,0x3f1007,0x3f2007,0x3f3007,0x3f4007,0x3f5007,0x3f6007,0x3f7007
+	.long 0x3f8007,0x3f9007,0x3fa007,0x3fb007,0x3fc007,0x3fd007,0x3fe007,0x3ff007
 
 .org 0x3000
-_empty_bad_page:
+ENTRY(empty_bad_page)
 
 .org 0x4000
-_empty_bad_page_table:
+ENTRY(empty_bad_page_table)
 
 .org 0x5000
-_empty_zero_page:
+ENTRY(empty_zero_page)
 
 .org 0x6000
-/*
- * floppy_track_buffer is used to buffer one track of floppy data: it
- * has to be separate from the tmp_floppy area, as otherwise a single-
- * sector read/write can mess it up. It can contain one full cylinder (sic) of
- * data (36*2*512 bytes).
- */
-_floppy_track_buffer:
-	.fill 512*2*MAX_BUFFER_SECTORS,1,0
-	
+
 stack_start:
-	.long _init_user_stack+4096
+	.long SYMBOL_NAME(init_user_stack)+4096
 	.long KERNEL_DS
 
 /* This is the default interrupt "handler" :-) */
 int_msg:
 	.asciz "Unknown interrupt\n"
-.align 2
+	ALIGN
 ignore_int:
 	cld
 	pushl %eax
@@ -318,7 +468,7 @@ ignore_int:
 	mov %ax,%es
 	mov %ax,%fs
 	pushl $int_msg
-	call _printk
+	call SYMBOL_NAME(printk)
 	popl %eax
 	pop %fs
 	pop %es
@@ -331,34 +481,41 @@ ignore_int:
 /*
  * The interrupt descriptor table has room for 256 idt's
  */
-.align 4
+	ALIGN
 .word 0
 idt_descr:
 	.word 256*8-1		# idt contains 256 entries
-	.long 0xc0000000+_idt
+	.long SYMBOL_NAME(idt)
 
-.align 4
-_idt:
+ENTRY(idt)
 	.fill 256,8,0		# idt is uninitialized
 
-.align 4
+	ALIGN
 .word 0
 gdt_descr:
+#ifdef CONFIG_APM
+	.word (11+2*NR_TASKS)*8-1
+#else
 	.word (8+2*NR_TASKS)*8-1
-	.long 0xc0000000+_gdt
+#endif
+	.long SYMBOL_NAME(gdt)
 
 /*
  * This gdt setup gives the kernel a 1GB address space at virtual
  * address 0xC0000000 - space enough for expansion, I hope.
  */
-.align 4
-_gdt:
+ENTRY(gdt)
 	.quad 0x0000000000000000	/* NULL descriptor */
 	.quad 0x0000000000000000	/* not used */
-	.quad 0xc0c39a000000ffff	/* 0x10 kernel 1GB code at 0xC0000000 */
-	.quad 0xc0c392000000ffff	/* 0x18 kernel 1GB data at 0xC0000000 */
+	.quad 0x00cf9a000000ffff	/* 0x10 kernel 4GB code at 0x00000000 */
+	.quad 0x00cf92000000ffff	/* 0x18 kernel 4GB data at 0x00000000 */
 	.quad 0x00cbfa000000ffff	/* 0x23 user   3GB code at 0x00000000 */
 	.quad 0x00cbf2000000ffff	/* 0x2b user   3GB data at 0x00000000 */
 	.quad 0x0000000000000000	/* not used */
 	.quad 0x0000000000000000	/* not used */
 	.fill 2*NR_TASKS,8,0		/* space for LDT's and TSS's etc */
+#ifdef CONFIG_APM
+	.quad 0x00c09a0000000000	/* APM CS    code */
+	.quad 0x00809a0000000000	/* APM CS 16 code (16 bit) */
+	.quad 0x00c0920000000000	/* APM DS    data */
+#endif
diff --git a/arch/i386/kernel/hexify.c b/arch/i386/kernel/hexify.c
new file mode 100644
index 000000000..daa331fec
--- /dev/null
+++ b/arch/i386/kernel/hexify.c
@@ -0,0 +1,31 @@
+#include <stdio.h>
+
+
+void main()
+{
+	int c;
+	int comma=0;
+	int count=0;
+	while((c=getchar())!=EOF)
+	{
+		unsigned char x=c;
+		if(comma)
+			printf(",");
+		else
+			comma=1;
+		if(count==8)
+		{
+			count=0;
+			printf("\n");
+		}
+		if(count==0)
+			printf("\t");
+		printf("0x%02X",c);
+		count++;
+	}
+	if(count)
+		printf("\n");
+	exit(0);
+}
+
+		
diff --git a/arch/i386/kernel/ioport.c b/arch/i386/kernel/ioport.c
index c949f70cd..bcefeef9b 100644
--- a/arch/i386/kernel/ioport.c
+++ b/arch/i386/kernel/ioport.c
@@ -75,8 +75,8 @@ unsigned int *stack;
  */
 asmlinkage int sys_iopl(long ebx,long ecx,long edx,
 	     long esi, long edi, long ebp, long eax, long ds,
-	     long es, long fs, long gs, long orig_eax,
-	     long eip,long cs,long eflags,long esp,long ss)
+	     long es, long orig_eax, long eip, long cs,
+	     long eflags, long esp, long ss)
 {
 	unsigned int level = ebx;
 
diff --git a/arch/i386/kernel/irq.c b/arch/i386/kernel/irq.c
index 23085dc6c..e131e5a45 100644
--- a/arch/i386/kernel/irq.c
+++ b/arch/i386/kernel/irq.c
@@ -23,54 +23,68 @@
 #include <linux/ioport.h>
 #include <linux/interrupt.h>
 #include <linux/timex.h>
+#include <linux/malloc.h>
+#include <linux/random.h>
 
 #include <asm/system.h>
 #include <asm/io.h>
 #include <asm/irq.h>
 #include <asm/bitops.h>
+#include <asm/smp.h>
 
 #define CR0_NE 32
 
 static unsigned char cache_21 = 0xff;
 static unsigned char cache_A1 = 0xff;
 
-void disable_irq(unsigned int irq_nr)
+#ifdef __SMP_PROF__
+static unsigned int int_count[NR_CPUS][NR_IRQS] = {{0},};
+#endif
+
+static inline void mask_irq(unsigned int irq_nr)
 {
-	unsigned long flags;
 	unsigned char mask;
 
 	mask = 1 << (irq_nr & 7);
-	save_flags(flags);
 	if (irq_nr < 8) {
-		cli();
 		cache_21 |= mask;
 		outb(cache_21,0x21);
-		restore_flags(flags);
-		return;
+	} else {
+		cache_A1 |= mask;
+		outb(cache_A1,0xA1);
 	}
-	cli();
-	cache_A1 |= mask;
-	outb(cache_A1,0xA1);
-	restore_flags(flags);
 }
 
-void enable_irq(unsigned int irq_nr)
+static inline void unmask_irq(unsigned int irq_nr)
 {
-	unsigned long flags;
 	unsigned char mask;
 
 	mask = ~(1 << (irq_nr & 7));
-	save_flags(flags);
 	if (irq_nr < 8) {
-		cli();
 		cache_21 &= mask;
 		outb(cache_21,0x21);
-		restore_flags(flags);
-		return;
+	} else {
+		cache_A1 &= mask;
+		outb(cache_A1,0xA1);
 	}
+}
+
+void disable_irq(unsigned int irq_nr)
+{
+	unsigned long flags;
+
+	save_flags(flags);
+	cli();
+	mask_irq(irq_nr);
+	restore_flags(flags);
+}
+
+void enable_irq(unsigned int irq_nr)
+{
+	unsigned long flags;
+	save_flags(flags);
 	cli();
-	cache_A1 &= mask;
-	outb(cache_A1,0xA1);
+	unmask_irq(irq_nr);
 	restore_flags(flags);
 }
 
@@ -88,8 +102,16 @@ void enable_irq(unsigned int irq_nr)
  * atomic. The specific handler is chosen depending on the SA_INTERRUPT
  * flag when installing a handler. Finally, one "bad interrupt" handler, that
  * is used when no handler is present.
+ *
+ * The timer interrupt is handled specially to insure that the jiffies
+ * variable is updated at all times.  Specifically, the timer interrupt is
+ * just like the complete handlers except that it is invoked with interrupts
+ * disabled and should never re-enable them.  If other interrupts were
+ * allowed to be processed while the timer interrupt is active, then the
+ * other interrupts would have to avoid using the jiffies variable for delay
+ * and interval timing operations to avoid hanging the system.
  */
-BUILD_IRQ(FIRST,0,0x01)
+BUILD_TIMER_IRQ(FIRST,0,0x01)
 BUILD_IRQ(FIRST,1,0x02)
 BUILD_IRQ(FIRST,2,0x04)
 BUILD_IRQ(FIRST,3,0x08)
@@ -102,19 +124,29 @@ BUILD_IRQ(SECOND,9,0x02)
 BUILD_IRQ(SECOND,10,0x04)
 BUILD_IRQ(SECOND,11,0x08)
 BUILD_IRQ(SECOND,12,0x10)
+#ifdef __SMP__
+BUILD_MSGIRQ(SECOND,13,0x20)
+#else
 BUILD_IRQ(SECOND,13,0x20)
+#endif
 BUILD_IRQ(SECOND,14,0x40)
 BUILD_IRQ(SECOND,15,0x80)
+#ifdef __SMP__
+BUILD_RESCHEDIRQ(16)
+#endif
 
 /*
  * Pointers to the low-level handlers: first the general ones, then the
  * fast ones, then the bad ones.
  */
-static void (*interrupt[16])(void) = {
+static void (*interrupt[17])(void) = {
 	IRQ0_interrupt, IRQ1_interrupt, IRQ2_interrupt, IRQ3_interrupt,
 	IRQ4_interrupt, IRQ5_interrupt, IRQ6_interrupt, IRQ7_interrupt,
 	IRQ8_interrupt, IRQ9_interrupt, IRQ10_interrupt, IRQ11_interrupt,
-	IRQ12_interrupt, IRQ13_interrupt, IRQ14_interrupt, IRQ15_interrupt
+	IRQ12_interrupt, IRQ13_interrupt, IRQ14_interrupt, IRQ15_interrupt	
+#ifdef __SMP__	
+	,IRQ16_interrupt
+#endif
 };
 
 static void (*fast_interrupt[16])(void) = {
@@ -142,39 +174,164 @@ static void (*bad_interrupt[16])(void) = {
 /*
  * Initial irq handlers.
  */
-struct irqaction {
-	void (*handler)(int, struct pt_regs *);
-	unsigned long flags;
-	unsigned long mask;
-	const char *name;
-};
 
-static struct irqaction irq_action[16] = {
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL },
-	{ NULL, 0, 0, NULL }, { NULL, 0, 0, NULL }
+static void no_action(int cpl, void *dev_id, struct pt_regs *regs) { }
+
+#ifdef __SMP__
+
+/*
+ * On SMP boards, irq13 is used for interprocessor interrupts (IPI's).
+ */
+static struct irqaction irq13 = { smp_message_irq, SA_INTERRUPT, 0, "IPI", NULL, NULL };
+
+#else
+
+/*
+ * Note that on a 486, we don't want to do a SIGFPE on an irq13
+ * as the irq is unreliable, and exception 16 works correctly
+ * (ie as explained in the intel literature). On a 386, you
+ * can't use exception 16 due to bad IBM design, so we have to
+ * rely on the less exact irq13.
+ *
+ * Careful.. Not only is IRQ13 unreliable, but it is also
+ * leads to races. IBM designers who came up with it should
+ * be shot.
+ */
+ 
+
+static void math_error_irq(int cpl, void *dev_id, struct pt_regs *regs)
+{
+	outb(0,0xF0);
+	if (ignore_irq13 || !hard_math)
+		return;
+	math_error();
+}
+
+static struct irqaction irq13 = { math_error_irq, 0, 0, "math error", NULL, NULL };
+
+#endif
+
+/*
+ * IRQ2 is cascade interrupt to second interrupt controller
+ */
+static struct irqaction irq2  = { no_action, 0, 0, "cascade", NULL, NULL};
+
+static struct irqaction *irq_action[16] = {
+	NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL,
+	NULL, NULL, NULL, NULL
 };
 
 int get_irq_list(char *buf)
 {
 	int i, len = 0;
-	struct irqaction * action = irq_action;
+	struct irqaction * action;
 
-	for (i = 0 ; i < 16 ; i++, action++) {
-		if (!action->handler)
+	for (i = 0 ; i < 16 ; i++) {
+		action = irq_action[i];
+		if (!action) 
 			continue;
-		len += sprintf(buf+len, "%2d: %8d %c %s\n",
+		len += sprintf(buf+len, "%2d: %10u %c %s",
 			i, kstat.interrupts[i],
 			(action->flags & SA_INTERRUPT) ? '+' : ' ',
 			action->name);
+		for (action=action->next; action; action = action->next) {
+			len += sprintf(buf+len, ",%s %s",
+				(action->flags & SA_INTERRUPT) ? " +" : "",
+				action->name);
+		}
+		len += sprintf(buf+len, "\n");
+	}
+/*
+ *	Linus - should you add NMI counts here ?????
+ */
+#ifdef __SMP_PROF__
+	len+=sprintf(buf+len, "IPI: %8lu received\n",
+		ipi_count);
+#endif		
+	return len;
+}
+
+#ifdef __SMP_PROF__
+
+int get_smp_prof_list(char *buf) {
+	int i,j, len = 0;
+	struct irqaction * action;
+	unsigned long sum_spins = 0;
+	unsigned long sum_spins_syscall = 0;
+	unsigned long sum_spins_sys_idle = 0;
+	unsigned long sum_smp_idle_count = 0;
+
+	for (i=0;i<smp_num_cpus;i++) {
+		int cpunum = cpu_logical_map[i];
+		sum_spins+=smp_spins[cpunum];
+		sum_spins_syscall+=smp_spins_syscall[cpunum];
+		sum_spins_sys_idle+=smp_spins_sys_idle[cpunum];
+		sum_smp_idle_count+=smp_idle_count[cpunum];
+	}
+
+	len += sprintf(buf+len,"CPUS: %10i \n", smp_num_cpus);
+	len += sprintf(buf+len,"            SUM ");
+	for (i=0;i<smp_num_cpus;i++)
+		len += sprintf(buf+len,"        P%1d ",cpu_logical_map[i]);
+	len += sprintf(buf+len,"\n");
+	for (i = 0 ; i < NR_IRQS ; i++) {
+		action = *(i + irq_action);
+		if (!action || !action->handler)
+			continue;
+		len += sprintf(buf+len, "%3d: %10d ",
+			i, kstat.interrupts[i]);
+		for (j=0;j<smp_num_cpus;j++)
+			len+=sprintf(buf+len, "%10d ",
+				int_count[cpu_logical_map[j]][i]);
+		len += sprintf(buf+len, "%c %s\n",
+			(action->flags & SA_INTERRUPT) ? '+' : ' ',
+			action->name);
+		for (action=action->next; action; action = action->next) {
+			len += sprintf(buf+len, ",%s %s",
+				(action->flags & SA_INTERRUPT) ? " +" : "",
+				action->name);
+		}
 	}
+	len+=sprintf(buf+len, "LCK: %10lu",
+		sum_spins);
+
+	for (i=0;i<smp_num_cpus;i++)
+		len+=sprintf(buf+len," %10lu",smp_spins[cpu_logical_map[i]]);
+
+	len +=sprintf(buf+len,"   spins from int\n");
+
+	len+=sprintf(buf+len, "LCK: %10lu",
+		sum_spins_syscall);
+
+	for (i=0;i<smp_num_cpus;i++)
+		len+=sprintf(buf+len," %10lu",smp_spins_syscall[cpu_logical_map[i]]);
+
+	len +=sprintf(buf+len,"   spins from syscall\n");
+
+	len+=sprintf(buf+len, "LCK: %10lu",
+		sum_spins_sys_idle);
+
+	for (i=0;i<smp_num_cpus;i++)
+		len+=sprintf(buf+len," %10lu",smp_spins_sys_idle[cpu_logical_map[i]]);
+
+	len +=sprintf(buf+len,"   spins from sysidle\n");
+	len+=sprintf(buf+len,"IDLE %10lu",sum_smp_idle_count);
+
+	for (i=0;i<smp_num_cpus;i++)
+		len+=sprintf(buf+len," %10lu",smp_idle_count[cpu_logical_map[i]]);
+
+	len +=sprintf(buf+len,"   idle ticks\n");
+
+	len+=sprintf(buf+len, "IPI: %10lu   received\n",
+		ipi_count);
+
 	return len;
 }
+#endif 
+
+
 
 /*
  * do_IRQ handles IRQ's that have been installed without the
@@ -185,10 +342,25 @@ int get_irq_list(char *buf)
  */
 asmlinkage void do_IRQ(int irq, struct pt_regs * regs)
 {
-	struct irqaction * action = irq + irq_action;
+	struct irqaction * action = *(irq + irq_action);
+	int do_random = 0;
+
+#ifdef __SMP__
+	if(smp_threads_ready && active_kernel_processor!=smp_processor_id())
+		panic("IRQ %d: active processor set wrongly(%d not %d).\n", irq, active_kernel_processor, smp_processor_id());
+#endif
 
 	kstat.interrupts[irq]++;
-	action->handler(irq, regs);
+#ifdef __SMP_PROF__
+	int_count[smp_processor_id()][irq]++;
+#endif
+	while (action) {
+		do_random |= action->flags;
+		action->handler(irq, action->dev_id, regs);
+		action = action->next;
+	}
+	if (do_random & SA_SAMPLE_RANDOM)
+		add_interrupt_randomness(irq);
 }
 
 /*
@@ -198,145 +370,159 @@ asmlinkage void do_IRQ(int irq, struct pt_regs * regs)
  */
 asmlinkage void do_fast_IRQ(int irq)
 {
-	struct irqaction * action = irq + irq_action;
+	struct irqaction * action = *(irq + irq_action);
+	int do_random = 0;
+	
+#ifdef __SMP__
+	/* IRQ 13 is allowed - that's a flush tlb */
+	if(smp_threads_ready && active_kernel_processor!=smp_processor_id() && irq!=13)
+		panic("fast_IRQ %d: active processor set wrongly(%d not %d).\n", irq, active_kernel_processor, smp_processor_id());
+#endif
 
 	kstat.interrupts[irq]++;
-	action->handler(irq, NULL);
+#ifdef __SMP_PROF__
+	int_count[smp_processor_id()][irq]++;
+#endif
+	while (action) {
+		do_random |= action->flags;
+		action->handler(irq, action->dev_id, NULL);
+		action = action->next;
+	}
+	if (do_random & SA_SAMPLE_RANDOM)
+		add_interrupt_randomness(irq);
 }
 
-#define SA_PROBE SA_ONESHOT
+int setup_x86_irq(int irq, struct irqaction * new)
+{
+	int shared = 0;
+	struct irqaction *old, **p;
+	unsigned long flags;
+
+	p = irq_action + irq;
+	if ((old = *p) != NULL) {
+		/* Can't share interrupts unless both agree to */
+		if (!(old->flags & new->flags & SA_SHIRQ))
+			return -EBUSY;
+
+		/* Can't share interrupts unless both are same type */
+		if ((old->flags ^ new->flags) & SA_INTERRUPT)
+			return -EBUSY;
+
+		/* add new interrupt at end of irq queue */
+		do {
+			p = &old->next;
+			old = *p;
+		} while (old);
+		shared = 1;
+	}
+
+	if (new->flags & SA_SAMPLE_RANDOM)
+		rand_initialize_irq(irq);
+
+	save_flags(flags);
+	cli();
+	*p = new;
+
+	if (!shared) {
+		if (new->flags & SA_INTERRUPT)
+			set_intr_gate(0x20+irq,fast_interrupt[irq]);
+		else
+			set_intr_gate(0x20+irq,interrupt[irq]);
+		unmask_irq(irq);
+	}
+	restore_flags(flags);
+	return 0;
+}
 
-int request_irq(unsigned int irq, void (*handler)(int, struct pt_regs *),
-	unsigned long irqflags, const char * devname)
+int request_irq(unsigned int irq, 
+		void (*handler)(int, void *, struct pt_regs *),
+		unsigned long irqflags, 
+		const char * devname,
+		void *dev_id)
 {
+	int retval;
 	struct irqaction * action;
-	unsigned long flags;
 
 	if (irq > 15)
 		return -EINVAL;
-	action = irq + irq_action;
-	if (action->handler)
-		return -EBUSY;
 	if (!handler)
 		return -EINVAL;
-	save_flags(flags);
-	cli();
+
+	action = (struct irqaction *)kmalloc(sizeof(struct irqaction), GFP_KERNEL);
+	if (!action)
+		return -ENOMEM;
+
 	action->handler = handler;
 	action->flags = irqflags;
 	action->mask = 0;
 	action->name = devname;
-	if (!(action->flags & SA_PROBE)) { /* SA_ONESHOT is used by probing */
-		if (action->flags & SA_INTERRUPT)
-			set_intr_gate(0x20+irq,fast_interrupt[irq]);
-		else
-			set_intr_gate(0x20+irq,interrupt[irq]);
-	}
-	if (irq < 8) {
-		cache_21 &= ~(1<<irq);
-		outb(cache_21,0x21);
-	} else {
-		cache_21 &= ~(1<<2);
-		cache_A1 &= ~(1<<(irq-8));
-		outb(cache_21,0x21);
-		outb(cache_A1,0xA1);
-	}
-	restore_flags(flags);
-	return 0;
+	action->next = NULL;
+	action->dev_id = dev_id;
+
+	retval = setup_x86_irq(irq, action);
+
+	if (retval)
+		kfree(action);
+	return retval;
 }
 		
-void free_irq(unsigned int irq)
+void free_irq(unsigned int irq, void *dev_id)
 {
-	struct irqaction * action = irq + irq_action;
+	struct irqaction * action, **p;
 	unsigned long flags;
 
 	if (irq > 15) {
 		printk("Trying to free IRQ%d\n",irq);
 		return;
 	}
-	if (!action->handler) {
-		printk("Trying to free free IRQ%d\n",irq);
-		return;
-	}
-	save_flags(flags);
-	cli();
-	if (irq < 8) {
-		cache_21 |= 1 << irq;
-		outb(cache_21,0x21);
-	} else {
-		cache_A1 |= 1 << (irq-8);
-		outb(cache_A1,0xA1);
-	}
-	set_intr_gate(0x20+irq,bad_interrupt[irq]);
-	action->handler = NULL;
-	action->flags = 0;
-	action->mask = 0;
-	action->name = NULL;
-	restore_flags(flags);
-}
+	for (p = irq + irq_action; (action = *p) != NULL; p = &action->next) {
+		if (action->dev_id != dev_id)
+			continue;
 
-/*
- * Note that on a 486, we don't want to do a SIGFPE on a irq13
- * as the irq is unreliable, and exception 16 works correctly
- * (ie as explained in the intel literature). On a 386, you
- * can't use exception 16 due to bad IBM design, so we have to
- * rely on the less exact irq13.
- *
- * Careful.. Not only is IRQ13 unreliable, but it is also
- * leads to races. IBM designers who came up with it should
- * be shot.
- */
-static void math_error_irq(int cpl, struct pt_regs *regs)
-{
-	outb(0,0xF0);
-	if (ignore_irq13 || !hard_math)
+		/* Found it - now free it */
+		save_flags(flags);
+		cli();
+		*p = action->next;
+		if (!irq[irq_action]) {
+			mask_irq(irq);
+			set_intr_gate(0x20+irq,bad_interrupt[irq]);
+		}
+		restore_flags(flags);
+		kfree(action);
 		return;
-	math_error();
+	}
+	printk("Trying to free free IRQ%d\n",irq);
 }
 
-static void no_action(int cpl, struct pt_regs * regs) { }
-
-unsigned int probe_irq_on (void)
+unsigned long probe_irq_on (void)
 {
 	unsigned int i, irqs = 0, irqmask;
 	unsigned long delay;
 
-	/* first, snaffle up any unassigned irqs */
+	/* first, enable any unassigned irqs */
 	for (i = 15; i > 0; i--) {
-		if (!request_irq(i, no_action, SA_PROBE, "probe")) {
+		if (!irq_action[i]) {
 			enable_irq(i);
 			irqs |= (1 << i);
 		}
 	}
 
 	/* wait for spurious interrupts to mask themselves out again */
-	for (delay = jiffies + 2; delay > jiffies; );	/* min 10ms delay */
+	for (delay = jiffies + HZ/10; delay > jiffies; )
+		/* about 100ms delay */;
 
 	/* now filter out any obviously spurious interrupts */
 	irqmask = (((unsigned int)cache_A1)<<8) | (unsigned int)cache_21;
-	for (i = 15; i > 0; i--) {
-		if (irqs & (1 << i) & irqmask) {
-			irqs ^= (1 << i);
-			free_irq(i);
-		}
-	}
-#ifdef DEBUG
-	printk("probe_irq_on:  irqs=0x%04x irqmask=0x%04x\n", irqs, irqmask);
-#endif
-	return irqs;
+	return irqs & ~irqmask;
 }
 
-int probe_irq_off (unsigned int irqs)
+int probe_irq_off (unsigned long irqs)
 {
 	unsigned int i, irqmask;
 
 	irqmask = (((unsigned int)cache_A1)<<8) | (unsigned int)cache_21;
-	for (i = 15; i > 0; i--) {
-		if (irqs & (1 << i)) {
-			free_irq(i);
-		}
-	}
 #ifdef DEBUG
-	printk("probe_irq_off: irqs=0x%04x irqmask=0x%04x\n", irqs, irqmask);
+	printk("probe_irq_off: irqs=0x%04lx irqmask=0x%04x\n", irqs, irqmask);
 #endif
 	irqs &= irqmask;
 	if (!irqs)
@@ -350,6 +536,10 @@ int probe_irq_off (unsigned int irqs)
 void init_IRQ(void)
 {
 	int i;
+	static unsigned char smptrap=0;
+	if(smptrap)
+		return;
+	smptrap=1;
 
 	/* set the clock to 100 Hz */
 	outb_p(0x34,0x43);		/* binary, mode 2, LSB/MSB, ch 0 */
@@ -357,10 +547,13 @@ void init_IRQ(void)
 	outb(LATCH >> 8 , 0x40);	/* MSB */
 	for (i = 0; i < 16 ; i++)
 		set_intr_gate(0x20+i,bad_interrupt[i]);
-	if (request_irq(2, no_action, SA_INTERRUPT, "cascade"))
-		printk("Unable to get IRQ2 for cascade\n");
-	if (request_irq(13,math_error_irq, 0, "math error"))
-		printk("Unable to get IRQ13 for math-error handler\n");
+	/* This bit is a hack because we don't send timer messages to all processors yet */
+	/* It has to be here .. it doesn't work if you put it down the bottom - assembler explodes 8) */
+#ifdef __SMP__	
+	set_intr_gate(0x20+i, interrupt[i]);	/* IRQ '16' - IPI for rescheduling */
+#endif	
 	request_region(0x20,0x20,"pic1");
 	request_region(0xa0,0x20,"pic2");
+	setup_x86_irq(2, &irq2);
+	setup_x86_irq(13, &irq13);
 } 
diff --git a/arch/i386/kernel/ksyms.c b/arch/i386/kernel/ksyms.c
new file mode 100644
index 000000000..bdb6e8005
--- /dev/null
+++ b/arch/i386/kernel/ksyms.c
@@ -0,0 +1,41 @@
+#include <linux/module.h>
+#include <linux/smp.h>
+#include <linux/user.h>
+#include <linux/elfcore.h>
+
+#include <asm/semaphore.h>
+#include <asm/processor.h>
+#include <asm/uaccess.h>
+#include <asm/io.h>
+
+extern void dump_thread(struct pt_regs *, struct user *);
+extern int dump_fpu(elf_fpregset_t *);
+
+static struct symbol_table arch_symbol_table = {
+#include <linux/symtab_begin.h>
+	/* platform dependent support */
+	X(EISA_bus),
+	X(wp_works_ok),
+	X(__verify_write),
+	X(dump_thread),
+	X(dump_fpu),
+	X(ioremap),
+	X(iounmap),
+	XNOVERS(__down_failed),
+	XNOVERS(__up_wakeup),
+#ifdef __SMP__
+	X(apic_reg),		/* Needed internally for the I386 inlines */
+	X(cpu_data),
+	X(syscall_count),
+	X(kernel_flag),
+	X(kernel_counter),
+	X(active_kernel_processor),
+	X(smp_invalidate_needed),
+#endif
+#include <linux/symtab_end.h>
+};
+
+void arch_syms_export(void)
+{
+	register_symtab(&arch_symbol_table);
+}
diff --git a/arch/i386/kernel/ldt.c b/arch/i386/kernel/ldt.c
index bace95f4e..9e2f5e0cd 100644
--- a/arch/i386/kernel/ldt.c
+++ b/arch/i386/kernel/ldt.c
@@ -8,9 +8,11 @@
 #include <linux/sched.h>
 #include <linux/string.h>
 #include <linux/mm.h>
-#include <asm/segment.h>
+#include <linux/vmalloc.h>
+
+#include <asm/uaccess.h>
 #include <asm/system.h>
-#include <linux/ldt.h>
+#include <asm/ldt.h>
 
 static int read_ldt(void * ptr, unsigned long bytecount)
 {
@@ -30,15 +32,39 @@ static int read_ldt(void * ptr, unsigned long bytecount)
 	error = verify_area(VERIFY_WRITE, ptr, size);
 	if (error)
 		return error;
-	memcpy_tofs(ptr, address, size);
+	copy_to_user(ptr, address, size);
 	return size;
 }
 
+static inline int limits_ok(struct modify_ldt_ldt_s *ldt_info)
+{
+	unsigned long base, limit;
+	/* linear address of first and last accessible byte */
+	unsigned long first, last;
+
+	base = ldt_info->base_addr;
+	limit = ldt_info->limit;
+	if (ldt_info->limit_in_pages)
+		limit = limit * PAGE_SIZE + PAGE_SIZE - 1;
+
+	first = base;
+	last = limit + base;
+
+	/* segment grows down? */
+	if (ldt_info->contents == 1) {
+		/* data segment grows down */
+		first = base+limit+1;
+		last = base+65535;
+		if (ldt_info->seg_32bit)
+			last = base-1;
+	}
+	return (last >= first && last < TASK_SIZE);
+}
+
 static int write_ldt(void * ptr, unsigned long bytecount)
 {
 	struct modify_ldt_ldt_s ldt_info;
 	unsigned long *lp;
-	unsigned long base, limit;
 	int error, i;
 
 	if (bytecount != sizeof(ldt_info))
@@ -47,18 +73,12 @@ static int write_ldt(void * ptr, unsigned long bytecount)
 	if (error)
 		return error;
 
-	memcpy_fromfs(&ldt_info, ptr, sizeof(ldt_info));
+	copy_from_user(&ldt_info, ptr, sizeof(ldt_info));
 
 	if (ldt_info.contents == 3 || ldt_info.entry_number >= LDT_ENTRIES)
 		return -EINVAL;
 
-	limit = ldt_info.limit;
-	base = ldt_info.base_addr;
-	if (ldt_info.limit_in_pages)
-		limit *= PAGE_SIZE;
-
-	limit += base;
-	if (limit < base || limit >= 0xC0000000)
+	if (!limits_ok(&ldt_info))
 		return -EINVAL;
 
 	if (!current->ldt) {
diff --git a/arch/i386/kernel/process.c b/arch/i386/kernel/process.c
index a5e8777bf..a26a2c6a3 100644
--- a/arch/i386/kernel/process.c
+++ b/arch/i386/kernel/process.c
@@ -8,6 +8,9 @@
  * This file handles the architecture-dependent parts of process handling..
  */
 
+#define __KERNEL_SYSCALLS__
+#include <stdarg.h>
+
 #include <linux/errno.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
@@ -16,19 +19,32 @@
 #include <linux/unistd.h>
 #include <linux/ptrace.h>
 #include <linux/malloc.h>
-#include <linux/ldt.h>
+#include <linux/vmalloc.h>
 #include <linux/user.h>
 #include <linux/a.out.h>
+#include <linux/interrupt.h>
+#include <linux/config.h>
+#include <linux/unistd.h>
+#include <linux/delay.h>
+#include <linux/smp.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 #include <asm/pgtable.h>
 #include <asm/system.h>
 #include <asm/io.h>
+#include <asm/ldt.h>
 
 asmlinkage void ret_from_sys_call(void) __asm__("ret_from_sys_call");
 
+#ifdef CONFIG_APM
+extern int  apm_do_idle(void);
+extern void apm_do_busy(void);
+#endif
+
 static int hlt_counter=0;
 
+#define HARD_IDLE_TIMEOUT (HZ / 3)
+
 void disable_hlt(void)
 {
 	hlt_counter++;
@@ -39,53 +55,145 @@ void enable_hlt(void)
 	hlt_counter--;
 }
 
-asmlinkage int sys_pipe(unsigned long * fildes)
-{
-	int fd[2];
-	int error;
+#ifndef __SMP__
 
-	error = verify_area(VERIFY_WRITE,fildes,8);
-	if (error)
-		return error;
-	error = do_pipe(fd);
-	if (error)
-		return error;
-	put_fs_long(fd[0],0+fildes);
-	put_fs_long(fd[1],1+fildes);
-	return 0;
+static void hard_idle(void)
+{
+	while (!need_resched) {
+		if (hlt_works_ok && !hlt_counter) {
+#ifdef CONFIG_APM
+				/* If the APM BIOS is not enabled, or there
+				 is an error calling the idle routine, we
+				 should hlt if possible.  We need to check
+				 need_resched again because an interrupt
+				 may have occurred in apm_do_idle(). */
+			start_bh_atomic();
+			if (!apm_do_idle() && !need_resched)
+				__asm__("hlt");
+			end_bh_atomic();
+#else
+			__asm__("hlt");
+#endif
+	        }
+ 		if (need_resched) 
+ 			break;
+		schedule();
+	}
+#ifdef CONFIG_APM
+	apm_do_busy();
+#endif
 }
 
 /*
- * The idle loop on a i386..
+ * The idle loop on a uniprocessor i386..
  */
+ 
 asmlinkage int sys_idle(void)
 {
-	int i;
-	pmd_t * pmd;
+        unsigned long start_idle = 0;
 
 	if (current->pid != 0)
 		return -EPERM;
-
-	/* Map out the low memory: it's no longer needed */
-	pmd = pmd_offset(swapper_pg_dir, 0);
-	for (i = 0 ; i < 768 ; i++)
-		pmd_clear(pmd++);
-
 	/* endless idle loop with no priority at all */
 	current->counter = -100;
-	for (;;) {
-		if (hlt_works_ok && !hlt_counter && !need_resched)
-			__asm__("hlt");
+	for (;;) 
+	{
+		/*
+		 *	We are locked at this point. So we can safely call
+		 *	the APM bios knowing only one CPU at a time will do
+		 *	so.
+		 */
+		if (!start_idle) 
+			start_idle = jiffies;
+		if (jiffies - start_idle > HARD_IDLE_TIMEOUT) 
+		{
+			hard_idle();
+		} 
+		else 
+		{
+			if (hlt_works_ok && !hlt_counter && !need_resched)
+		        	__asm__("hlt");
+		}
+		if (need_resched) 
+			start_idle = 0;
 		schedule();
 	}
 }
 
+#else
+
+/*
+ *	In the SMP world we hlt outside of kernel syscall rather than within
+ *	so as to get the right locking semantics.
+ */
+ 
+asmlinkage int sys_idle(void)
+{
+	if(current->pid != 0)
+		return -EPERM;
+#ifdef __SMP_PROF__
+	smp_spins_sys_idle[smp_processor_id()]+=
+	  smp_spins_syscall_cur[smp_processor_id()];
+#endif
+	current->counter= -100;
+	schedule();
+	return 0;
+}
+
+/*
+ *	This is being executed in task 0 'user space'.
+ */
+
+int cpu_idle(void *unused)
+{
+	while(1)
+	{
+		if(cpu_data[smp_processor_id()].hlt_works_ok && !hlt_counter && !need_resched)
+			__asm("hlt");
+                if(0==(0x7fffffff & smp_process_available)) 
+                	continue;
+                while(0x80000000 & smp_process_available);
+	        cli();
+                while(set_bit(31,&smp_process_available))
+                	while(test_bit(31,&smp_process_available))
+                {
+                	/*
+                	 *	Oops.. This is kind of important in some cases...
+                	 */
+                	if(clear_bit(smp_processor_id(), &smp_invalidate_needed))
+                		local_flush_tlb();
+                }
+                if (0==(0x7fffffff & smp_process_available)){
+                        clear_bit(31,&smp_process_available);
+                        sti();
+                        continue;
+                }
+                smp_process_available--;
+                clear_bit(31,&smp_process_available);
+                sti();
+		idle();
+	}
+}
+
+#endif
+
 /*
  * This routine reboots the machine by asking the keyboard
  * controller to pulse the reset-line low. We try that for a while,
  * and if it doesn't work, we do some other stupid things.
  */
 static long no_idt[2] = {0, 0};
+static int reboot_mode = 0;
+
+void reboot_setup(char *str, int *ints)
+{
+	int mode = 0;
+
+	/* "w" for "warm" reboot (no memory testing etc) */
+	if (str[0] == 'w')
+		mode = 0x1234;
+	reboot_mode = mode;
+}
 
 static inline void kb_wait(void)
 {
@@ -101,15 +209,14 @@ void hard_reset_now(void)
 	int i, j;
 
 	sti();
-/* rebooting needs to touch the page at absolute addr 0 */
-	pg0[0] = 7;
-	*((unsigned short *)0x472) = 0x1234;
+	*((unsigned short *)__va(0x472)) = reboot_mode;
 	for (;;) {
 		for (i=0; i<100; i++) {
 			kb_wait();
 			for(j = 0; j < 100000 ; j++)
 				/* nothing */;
 			outb(0xfe,0x64);	 /* pulse reset low */
+			udelay(100);
 		}
 		__asm__ __volatile__("\tlidt %0": "=m" (no_idt));
 	}
@@ -118,24 +225,27 @@ void hard_reset_now(void)
 void show_regs(struct pt_regs * regs)
 {
 	printk("\n");
-	printk("EIP: %04x:%08lx",0xffff & regs->cs,regs->eip);
-	if (regs->cs & 3)
-		printk(" ESP: %04x:%08lx",0xffff & regs->ss,regs->esp);
+	printk("EIP: %04x:[<%08lx>]",0xffff & regs->xcs,regs->eip);
+	if (regs->xcs & 3)
+		printk(" ESP: %04x:%08lx",0xffff & regs->xss,regs->esp);
 	printk(" EFLAGS: %08lx\n",regs->eflags);
 	printk("EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n",
 		regs->eax,regs->ebx,regs->ecx,regs->edx);
 	printk("ESI: %08lx EDI: %08lx EBP: %08lx",
 		regs->esi, regs->edi, regs->ebp);
-	printk(" DS: %04x ES: %04x FS: %04x GS: %04x\n",
-		0xffff & regs->ds,0xffff & regs->es,
-		0xffff & regs->fs,0xffff & regs->gs);
+	printk(" DS: %04x ES: %04x\n",
+		0xffff & regs->xds,0xffff & regs->xes);
 }
 
 /*
  * Free current thread data structures etc..
  */
+
 void exit_thread(void)
 {
+	/* forget lazy i387 state */
+	if (last_task_used_math == current)
+		last_task_used_math = NULL;
 	/* forget local segments */
 	__asm__ __volatile__("mov %w0,%%fs ; mov %w0,%%gs ; lldt %w0"
 		: /* no outputs */
@@ -166,6 +276,26 @@ void flush_thread(void)
 
 	for (i=0 ; i<8 ; i++)
 		current->debugreg[i] = 0;
+
+	/*
+	 * Forget coprocessor state..
+	 */
+#ifdef __SMP__
+	if (current->flags & PF_USEDFPU) {
+		stts();
+	}
+#else
+	if (last_task_used_math == current) {
+		last_task_used_math = NULL;
+		stts();
+	}
+#endif
+	current->used_math = 0;
+	current->flags &= ~PF_USEDFPU;
+}
+
+void release_thread(struct task_struct *dead_task)
+{
 }
 
 void copy_thread(int nr, unsigned long clone_flags, unsigned long esp,
@@ -186,6 +316,7 @@ void copy_thread(int nr, unsigned long clone_flags, unsigned long esp,
 	childregs = ((struct pt_regs *) (p->kernel_stack_page + PAGE_SIZE)) - 1;
 	p->tss.esp = (unsigned long) childregs;
 	p->tss.eip = (unsigned long) ret_from_sys_call;
+	p->tss.ebx = (unsigned long) p;
 	*childregs = *regs;
 	childregs->eax = 0;
 	childregs->esp = esp;
@@ -210,6 +341,31 @@ void copy_thread(int nr, unsigned long clone_flags, unsigned long esp,
 }
 
 /*
+ * fill in the fpu structure for a core dump..
+ */
+int dump_fpu (struct pt_regs * regs, struct user_i387_struct* fpu)
+{
+	int fpvalid;
+
+/* Flag indicating the math stuff is valid. We don't support this for the
+   soft-float routines yet */
+	if (hard_math) {
+		if ((fpvalid = current->used_math) != 0) {
+			if (last_task_used_math == current)
+				__asm__("clts ; fnsave %0": :"m" (*fpu));
+			else
+				memcpy(fpu,&current->tss.i387.hard,sizeof(*fpu));
+		}
+	} else {
+		/* we should dump the emulator state here, but we need to
+		   convert it into standard 387 format first.. */
+		fpvalid = 0;
+	}
+
+	return fpvalid;
+}
+
+/*
  * fill in the user structure for a core dump..
  */
 void dump_thread(struct pt_regs * regs, struct user * dump)
@@ -220,55 +376,52 @@ void dump_thread(struct pt_regs * regs, struct user * dump)
 	dump->magic = CMAGIC;
 	dump->start_code = 0;
 	dump->start_stack = regs->esp & ~(PAGE_SIZE - 1);
-	dump->u_tsize = ((unsigned long) current->mm->end_code) >> 12;
-	dump->u_dsize = ((unsigned long) (current->mm->brk + (PAGE_SIZE-1))) >> 12;
+	dump->u_tsize = ((unsigned long) current->mm->end_code) >> PAGE_SHIFT;
+	dump->u_dsize = ((unsigned long) (current->mm->brk + (PAGE_SIZE-1))) >> PAGE_SHIFT;
 	dump->u_dsize -= dump->u_tsize;
 	dump->u_ssize = 0;
 	for (i = 0; i < 8; i++)
 		dump->u_debugreg[i] = current->debugreg[i];  
 
 	if (dump->start_stack < TASK_SIZE)
-		dump->u_ssize = ((unsigned long) (TASK_SIZE - dump->start_stack)) >> 12;
-
-	dump->regs = *regs;
-
-/* Flag indicating the math stuff is valid. We don't support this for the
-   soft-float routines yet */
-	if (hard_math) {
-		if ((dump->u_fpvalid = current->used_math) != 0) {
-			if (last_task_used_math == current)
-				__asm__("clts ; fnsave %0": :"m" (dump->i387));
-			else
-				memcpy(&dump->i387,&current->tss.i387.hard,sizeof(dump->i387));
-		}
-	} else {
-		/* we should dump the emulator state here, but we need to
-		   convert it into standard 387 format first.. */
-		dump->u_fpvalid = 0;
-	}
+		dump->u_ssize = ((unsigned long) (TASK_SIZE - dump->start_stack)) >> PAGE_SHIFT;
+
+	dump->regs.ebx = regs->ebx;
+	dump->regs.ecx = regs->ecx;
+	dump->regs.edx = regs->edx;
+	dump->regs.esi = regs->esi;
+	dump->regs.edi = regs->edi;
+	dump->regs.ebp = regs->ebp;
+	dump->regs.eax = regs->eax;
+	dump->regs.ds = regs->xds;
+	dump->regs.es = regs->xes;
+	__asm__("mov %%fs,%0":"=r" (dump->regs.fs));
+	__asm__("mov %%gs,%0":"=r" (dump->regs.gs));
+	dump->regs.orig_eax = regs->orig_eax;
+	dump->regs.eip = regs->eip;
+	dump->regs.cs = regs->xcs;
+	dump->regs.eflags = regs->eflags;
+	dump->regs.esp = regs->esp;
+	dump->regs.ss = regs->xss;
+
+	dump->u_fpvalid = dump_fpu (regs, &dump->i387);
 }
 
 asmlinkage int sys_fork(struct pt_regs regs)
 {
-	return do_fork(COPYVM | SIGCHLD, regs.esp, &regs);
+	return do_fork(SIGCHLD, regs.esp, &regs);
 }
 
 asmlinkage int sys_clone(struct pt_regs regs)
 {
-#ifdef CLONE_ACTUALLY_WORKS_OK
 	unsigned long clone_flags;
 	unsigned long newsp;
 
-	newsp = regs.ebx;
-	clone_flags = regs.ecx;
+	clone_flags = regs.ebx;
+	newsp = regs.ecx;
 	if (!newsp)
 		newsp = regs.esp;
-	if (newsp == regs.esp)
-		clone_flags |= COPYVM;
 	return do_fork(clone_flags, newsp, &regs);
-#else
-	return -ENOSYS;
-#endif
 }
 
 /*
diff --git a/arch/i386/kernel/ptrace.c b/arch/i386/kernel/ptrace.c
index f32035edc..5ea636092 100644
--- a/arch/i386/kernel/ptrace.c
+++ b/arch/i386/kernel/ptrace.c
@@ -11,7 +11,7 @@
 #include <linux/user.h>
 #include <linux/debugreg.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 #include <asm/pgtable.h>
 #include <asm/system.h>
 
@@ -28,10 +28,9 @@
 #define TRAP_FLAG 0x100
 
 /*
- * this is the number to subtract from the top of the stack. To find
- * the local frame.
+ * Offset of eflags on child stack..
  */
-#define MAGICNUMBER 68
+#define EFL_OFFSET ((EFL-2)*4-sizeof(struct pt_regs))
 
 /* change a pid into a task struct. */
 static inline struct task_struct * get_task(int pid)
@@ -83,7 +82,8 @@ static inline int put_stack_long(struct task_struct *task, int offset,
  * and that it is in the task area before calling this: this routine does
  * no checking.
  */
-static unsigned long get_long(struct vm_area_struct * vma, unsigned long addr)
+static unsigned long get_long(struct task_struct * tsk, 
+	struct vm_area_struct * vma, unsigned long addr)
 {
 	pgd_t * pgdir;
 	pmd_t * pgmiddle;
@@ -91,9 +91,9 @@ static unsigned long get_long(struct vm_area_struct * vma, unsigned long addr)
 	unsigned long page;
 
 repeat:
-	pgdir = pgd_offset(vma->vm_task, addr);
+	pgdir = pgd_offset(vma->vm_mm, addr);
 	if (pgd_none(*pgdir)) {
-		do_no_page(vma, addr, 0);
+		do_no_page(tsk, vma, addr, 0);
 		goto repeat;
 	}
 	if (pgd_bad(*pgdir)) {
@@ -103,7 +103,7 @@ repeat:
 	}
 	pgmiddle = pmd_offset(pgdir, addr);
 	if (pmd_none(*pgmiddle)) {
-		do_no_page(vma, addr, 0);
+		do_no_page(tsk, vma, addr, 0);
 		goto repeat;
 	}
 	if (pmd_bad(*pgmiddle)) {
@@ -113,12 +113,12 @@ repeat:
 	}
 	pgtable = pte_offset(pgmiddle, addr);
 	if (!pte_present(*pgtable)) {
-		do_no_page(vma, addr, 0);
+		do_no_page(tsk, vma, addr, 0);
 		goto repeat;
 	}
 	page = pte_page(*pgtable);
 /* this is a hack for non-kernel-mapped video buffers and similar */
-	if (page >= high_memory)
+	if (MAP_NR(page) >= max_mapnr)
 		return 0;
 	page += addr & ~PAGE_MASK;
 	return *(unsigned long *) page;
@@ -133,7 +133,7 @@ repeat:
  * Now keeps R/W state of page so that a text page stays readonly
  * even if a debugger scribbles breakpoints into it.  -M.U-
  */
-static void put_long(struct vm_area_struct * vma, unsigned long addr,
+static void put_long(struct task_struct * tsk, struct vm_area_struct * vma, unsigned long addr,
 	unsigned long data)
 {
 	pgd_t *pgdir;
@@ -142,9 +142,9 @@ static void put_long(struct vm_area_struct * vma, unsigned long addr,
 	unsigned long page;
 
 repeat:
-	pgdir = pgd_offset(vma->vm_task, addr);
+	pgdir = pgd_offset(vma->vm_mm, addr);
 	if (!pgd_present(*pgdir)) {
-		do_no_page(vma, addr, 1);
+		do_no_page(tsk, vma, addr, 1);
 		goto repeat;
 	}
 	if (pgd_bad(*pgdir)) {
@@ -154,7 +154,7 @@ repeat:
 	}
 	pgmiddle = pmd_offset(pgdir, addr);
 	if (pmd_none(*pgmiddle)) {
-		do_no_page(vma, addr, 1);
+		do_no_page(tsk, vma, addr, 1);
 		goto repeat;
 	}
 	if (pmd_bad(*pgmiddle)) {
@@ -164,23 +164,21 @@ repeat:
 	}
 	pgtable = pte_offset(pgmiddle, addr);
 	if (!pte_present(*pgtable)) {
-		do_no_page(vma, addr, 1);
+		do_no_page(tsk, vma, addr, 1);
 		goto repeat;
 	}
 	page = pte_page(*pgtable);
 	if (!pte_write(*pgtable)) {
-		do_wp_page(vma, addr, 1);
+		do_wp_page(tsk, vma, addr, 1);
 		goto repeat;
 	}
 /* this is a hack for non-kernel-mapped video buffers and similar */
-	if (page < high_memory) {
-		page += addr & ~PAGE_MASK;
-		*(unsigned long *) page = data;
-	}
+	if (MAP_NR(page) < max_mapnr)
+		*(unsigned long *) (page + (addr & ~PAGE_MASK)) = data;
 /* we're bypassing pagetables, so we have to set the dirty bit ourselves */
 /* this should also re-instate whatever read-only mode there was before */
-	*pgtable = pte_mkdirty(mk_pte(page, vma->vm_page_prot));
-	invalidate();
+	set_pte(pgtable, pte_mkdirty(mk_pte(page, vma->vm_page_prot)));
+	flush_tlb();
 }
 
 static struct vm_area_struct * find_extend_vma(struct task_struct * tsk, unsigned long addr)
@@ -188,7 +186,7 @@ static struct vm_area_struct * find_extend_vma(struct task_struct * tsk, unsigne
 	struct vm_area_struct * vma;
 
 	addr &= PAGE_MASK;
-	vma = find_vma(tsk,addr);
+	vma = find_vma(tsk->mm,addr);
 	if (!vma)
 		return NULL;
 	if (vma->vm_start <= addr)
@@ -222,8 +220,8 @@ static int read_long(struct task_struct * tsk, unsigned long addr,
 			if (!vma_high || vma_high->vm_start != vma->vm_end)
 				return -EIO;
 		}
-		low = get_long(vma, addr & ~(sizeof(long)-1));
-		high = get_long(vma_high, (addr+sizeof(long)) & ~(sizeof(long)-1));
+		low = get_long(tsk, vma, addr & ~(sizeof(long)-1));
+		high = get_long(tsk, vma_high, (addr+sizeof(long)) & ~(sizeof(long)-1));
 		switch (addr & (sizeof(long)-1)) {
 			case 1:
 				low >>= 8;
@@ -240,7 +238,7 @@ static int read_long(struct task_struct * tsk, unsigned long addr,
 		}
 		*result = low;
 	} else
-		*result = get_long(vma, addr);
+		*result = get_long(tsk, vma, addr);
 	return 0;
 }
 
@@ -264,8 +262,8 @@ static int write_long(struct task_struct * tsk, unsigned long addr,
 			if (!vma_high || vma_high->vm_start != vma->vm_end)
 				return -EIO;
 		}
-		low = get_long(vma, addr & ~(sizeof(long)-1));
-		high = get_long(vma_high, (addr+sizeof(long)) & ~(sizeof(long)-1));
+		low = get_long(tsk, vma, addr & ~(sizeof(long)-1));
+		high = get_long(tsk, vma_high, (addr+sizeof(long)) & ~(sizeof(long)-1));
 		switch (addr & (sizeof(long)-1)) {
 			case 0: /* shouldn't happen, but safety first */
 				low = data;
@@ -289,13 +287,78 @@ static int write_long(struct task_struct * tsk, unsigned long addr,
 				high |= data >> 8;
 				break;
 		}
-		put_long(vma, addr & ~(sizeof(long)-1),low);
-		put_long(vma_high, (addr+sizeof(long)) & ~(sizeof(long)-1),high);
+		put_long(tsk, vma, addr & ~(sizeof(long)-1),low);
+		put_long(tsk, vma_high, (addr+sizeof(long)) & ~(sizeof(long)-1),high);
 	} else
-		put_long(vma, addr, data);
+		put_long(tsk, vma, addr, data);
+	return 0;
+}
+
+static int putreg(struct task_struct *child,
+	unsigned long regno, unsigned long value)
+{
+	switch (regno >> 2) {
+		case ORIG_EAX:
+			return -EIO;
+		case FS:
+			if (value && (value & 3) != 3)
+				return -EIO;
+			child->tss.fs = value;
+			return 0;
+		case GS:
+			if (value && (value & 3) != 3)
+				return -EIO;
+			child->tss.gs = value;
+			return 0;
+		case DS:
+		case ES:
+			if (value && (value & 3) != 3)
+				return -EIO;
+			value &= 0xffff;
+			break;
+		case SS:
+		case CS:
+			if ((value & 3) != 3)
+				return -EIO;
+			value &= 0xffff;
+			break;
+		case EFL:
+			value &= FLAG_MASK;
+			value |= get_stack_long(child, EFL_OFFSET) & ~FLAG_MASK;
+	}
+	if (regno > GS*4)
+		regno -= 2*4;
+	put_stack_long(child, regno - sizeof(struct pt_regs), value);
 	return 0;
 }
 
+static unsigned long getreg(struct task_struct *child,
+	unsigned long regno)
+{
+	unsigned long retval = ~0UL;
+
+	switch (regno >> 2) {
+		case FS:
+			retval = child->tss.fs;
+			break;
+		case GS:
+			retval = child->tss.gs;
+			break;
+		case DS:
+		case ES:
+		case SS:
+		case CS:
+			retval = 0xffff;
+			/* fall through */
+		default:
+			if (regno > GS*4)
+				regno -= 2*4;
+			regno = regno - sizeof(struct pt_regs);
+			retval &= get_stack_long(child, regno);
+	}
+	return retval;
+}
+
 asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 {
 	struct task_struct *child;
@@ -321,8 +384,10 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 			return -EPERM;
 		if ((!child->dumpable ||
 		    (current->uid != child->euid) ||
+		    (current->uid != child->suid) ||
 		    (current->uid != child->uid) ||
 	 	    (current->gid != child->egid) ||
+	 	    (current->gid != child->sgid) ||
 	 	    (current->gid != child->gid)) && !suser())
 			return -EPERM;
 		/* the same process cannot be attached many times */
@@ -358,7 +423,7 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 				return res;
 			res = verify_area(VERIFY_WRITE, (void *) data, sizeof(long));
 			if (!res)
-				put_fs_long(tmp,(unsigned long *) data);
+				put_user(tmp,(unsigned long *) data);
 			return res;
 		}
 
@@ -375,22 +440,15 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 			if (res)
 				return res;
 			tmp = 0;  /* Default return condition */
-			if(addr < 17*sizeof(long)) {
-			  addr = addr >> 2; /* temporary hack. */
-
-			  tmp = get_stack_long(child, sizeof(long)*addr - MAGICNUMBER);
-			  if (addr == DS || addr == ES ||
-			      addr == FS || addr == GS ||
-			      addr == CS || addr == SS)
-			    tmp &= 0xffff;
-			};
+			if(addr < 17*sizeof(long))
+				tmp = getreg(child, addr);
 			if(addr >= (long) &dummy->u_debugreg[0] &&
 			   addr <= (long) &dummy->u_debugreg[7]){
 				addr -= (long) &dummy->u_debugreg[0];
 				addr = addr >> 2;
 				tmp = child->debugreg[addr];
 			};
-			put_fs_long(tmp,(unsigned long *) data);
+			put_user(tmp,(unsigned long *) data);
 			return 0;
 		}
 
@@ -404,35 +462,14 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 			    addr > sizeof(struct user) - 3)
 				return -EIO;
 
-			addr = addr >> 2; /* temporary hack. */
-
-			if (addr == ORIG_EAX)
-				return -EIO;
-			if (addr == DS || addr == ES ||
-			    addr == FS || addr == GS ||
-			    addr == CS || addr == SS) {
-			    	data &= 0xffff;
-			    	if (data && (data & 3) != 3)
-					return -EIO;
-			}
-			if (addr == EFL) {   /* flags. */
-				data &= FLAG_MASK;
-				data |= get_stack_long(child, EFL*sizeof(long)-MAGICNUMBER)  & ~FLAG_MASK;
-			}
-		  /* Do not allow the user to set the debug register for kernel
-		     address space */
-		  if(addr < 17){
-			  if (put_stack_long(child, sizeof(long)*addr-MAGICNUMBER, data))
-				return -EIO;
-			return 0;
-			};
+			if (addr < 17*sizeof(long))
+				return putreg(child, addr, data);
 
 		  /* We need to be very careful here.  We implicitly
 		     want to modify a portion of the task_struct, and we
 		     have to be selective about what portions we allow someone
 		     to modify. */
 
-		  addr = addr << 2;  /* Convert back again */
 		  if(addr >= (long) &dummy->u_debugreg[0] &&
 		     addr <= (long) &dummy->u_debugreg[7]){
 
@@ -466,10 +503,10 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 			else
 				child->flags &= ~PF_TRACESYS;
 			child->exit_code = data;
-			child->state = TASK_RUNNING;
+			wake_up_process(child);
 	/* make sure the single step bit is not set. */
-			tmp = get_stack_long(child, sizeof(long)*EFL-MAGICNUMBER) & ~TRAP_FLAG;
-			put_stack_long(child, sizeof(long)*EFL-MAGICNUMBER,tmp);
+			tmp = get_stack_long(child, EFL_OFFSET) & ~TRAP_FLAG;
+			put_stack_long(child, EFL_OFFSET,tmp);
 			return 0;
 		}
 
@@ -481,11 +518,13 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 		case PTRACE_KILL: {
 			long tmp;
 
-			child->state = TASK_RUNNING;
+			if (child->state == TASK_ZOMBIE)	/* already dead */
+				return 0;
+			wake_up_process(child);
 			child->exit_code = SIGKILL;
 	/* make sure the single step bit is not set. */
-			tmp = get_stack_long(child, sizeof(long)*EFL-MAGICNUMBER) & ~TRAP_FLAG;
-			put_stack_long(child, sizeof(long)*EFL-MAGICNUMBER,tmp);
+			tmp = get_stack_long(child, EFL_OFFSET) & ~TRAP_FLAG;
+			put_stack_long(child, EFL_OFFSET, tmp);
 			return 0;
 		}
 
@@ -495,9 +534,9 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 			if ((unsigned long) data > NSIG)
 				return -EIO;
 			child->flags &= ~PF_TRACESYS;
-			tmp = get_stack_long(child, sizeof(long)*EFL-MAGICNUMBER) | TRAP_FLAG;
-			put_stack_long(child, sizeof(long)*EFL-MAGICNUMBER,tmp);
-			child->state = TASK_RUNNING;
+			tmp = get_stack_long(child, EFL_OFFSET) | TRAP_FLAG;
+			put_stack_long(child, EFL_OFFSET, tmp);
+			wake_up_process(child);
 			child->exit_code = data;
 	/* give it a chance to run. */
 			return 0;
@@ -509,14 +548,14 @@ asmlinkage int sys_ptrace(long request, long pid, long addr, long data)
 			if ((unsigned long) data > NSIG)
 				return -EIO;
 			child->flags &= ~(PF_PTRACED|PF_TRACESYS);
-			child->state = TASK_RUNNING;
+			wake_up_process(child);
 			child->exit_code = data;
 			REMOVE_LINKS(child);
 			child->p_pptr = child->p_opptr;
 			SET_LINKS(child);
 			/* make sure the single step bit is not set. */
-			tmp = get_stack_long(child, sizeof(long)*EFL-MAGICNUMBER) & ~TRAP_FLAG;
-			put_stack_long(child, sizeof(long)*EFL-MAGICNUMBER,tmp);
+			tmp = get_stack_long(child, EFL_OFFSET) & ~TRAP_FLAG;
+			put_stack_long(child, EFL_OFFSET, tmp);
 			return 0;
 		}
 
diff --git a/arch/i386/kernel/setup.c b/arch/i386/kernel/setup.c
index 5cc4c5d7d..3dae50628 100644
--- a/arch/i386/kernel/setup.c
+++ b/arch/i386/kernel/setup.c
@@ -5,7 +5,7 @@
  */
 
 /*
- * This file handles the architecture-dependent parts of process handling..
+ * This file handles the architecture-dependent parts of initialization
  */
 
 #include <linux/errno.h>
@@ -16,30 +16,37 @@
 #include <linux/unistd.h>
 #include <linux/ptrace.h>
 #include <linux/malloc.h>
-#include <linux/ldt.h>
 #include <linux/user.h>
 #include <linux/a.out.h>
 #include <linux/tty.h>
 #include <linux/ioport.h>
 #include <linux/delay.h>
-
-#include <asm/segment.h>
+#include <linux/config.h>
+#ifdef CONFIG_APM
+#include <linux/apm_bios.h>
+#endif
+#ifdef CONFIG_BLK_DEV_RAM
+#include <linux/blk.h>
+#endif
+#include <asm/uaccess.h>
 #include <asm/system.h>
+#include <asm/smp.h>
 
 /*
  * Tell us the machine setup..
  */
-char hard_math = 0;		/* set by boot/head.S */
-char x86 = 0;			/* set by boot/head.S to 3 or 4 */
-char x86_model = 0;		/* set by boot/head.S */
-char x86_mask = 0;		/* set by boot/head.S */
-int x86_capability = 0;		/* set by boot/head.S */
+char hard_math = 0;		/* set by kernel/head.S */
+char x86 = 0;			/* set by kernel/head.S to 3..6 */
+char x86_model = 0;		/* set by kernel/head.S */
+char x86_mask = 0;		/* set by kernel/head.S */
+int x86_capability = 0;		/* set by kernel/head.S */
 int fdiv_bug = 0;		/* set if Pentium(TM) with FP bug */
+int have_cpuid = 0;             /* set if CPUID instruction works */
 
-char x86_vendor_id[13] = "Unknown";
+char x86_vendor_id[13] = "unknown";
 
 char ignore_irq13 = 0;		/* set if exception 16 works */
-char wp_works_ok = 0;		/* set if paging hardware honours WP */ 
+char wp_works_ok = -1;		/* set if paging hardware honours WP */ 
 char hlt_works_ok = 1;		/* set if the "hlt" instruction works */
 
 /*
@@ -52,11 +59,20 @@ int EISA_bus = 0;
  */
 struct drive_info_struct { char dummy[32]; } drive_info;
 struct screen_info screen_info;
+#ifdef CONFIG_APM
+struct apm_bios_info apm_bios_info;
+#endif
 
 unsigned char aux_device_present;
-extern int ramdisk_size;
+
+#ifdef CONFIG_BLK_DEV_RAM
+extern int rd_doload;		/* 1 = load ramdisk, 0 = don't load */
+extern int rd_prompt;		/* 1 = prompt for ramdisk, 0 = don't prompt */
+extern int rd_image_start;	/* starting block # of image */
+#endif
+
 extern int root_mountflags;
-extern int etext, edata, end;
+extern int _etext, _edata, _end;
 
 extern char empty_zero_page[PAGE_SIZE];
 
@@ -65,16 +81,28 @@ extern char empty_zero_page[PAGE_SIZE];
  */
 #define PARAM	empty_zero_page
 #define EXT_MEM_K (*(unsigned short *) (PARAM+2))
+#ifdef CONFIG_APM
+#define APM_BIOS_INFO (*(struct apm_bios_info *) (PARAM+64))
+#endif
 #define DRIVE_INFO (*(struct drive_info_struct *) (PARAM+0x80))
 #define SCREEN_INFO (*(struct screen_info *) (PARAM+0))
 #define MOUNT_ROOT_RDONLY (*(unsigned short *) (PARAM+0x1F2))
-#define RAMDISK_SIZE (*(unsigned short *) (PARAM+0x1F8))
+#define RAMDISK_FLAGS (*(unsigned short *) (PARAM+0x1F8))
 #define ORIG_ROOT_DEV (*(unsigned short *) (PARAM+0x1FC))
 #define AUX_DEVICE_INFO (*(unsigned char *) (PARAM+0x1FF))
+#define LOADER_TYPE (*(unsigned char *) (PARAM+0x210))
+#define KERNEL_START (*(unsigned long *) (PARAM+0x214))
+#define INITRD_START (*(unsigned long *) (PARAM+0x218))
+#define INITRD_SIZE (*(unsigned long *) (PARAM+0x21c))
 #define COMMAND_LINE ((char *) (PARAM+2048))
 #define COMMAND_LINE_SIZE 256
 
+#define RAMDISK_IMAGE_START_MASK  	0x07FF
+#define RAMDISK_PROMPT_FLAG		0x8000
+#define RAMDISK_LOAD_FLAG		0x4000	
+
 static char command_line[COMMAND_LINE_SIZE] = { 0, };
+       char saved_command_line[COMMAND_LINE_SIZE];
 
 void setup_arch(char **cmdline_p,
 	unsigned long * memory_start_p, unsigned long * memory_end_p)
@@ -82,35 +110,64 @@ void setup_arch(char **cmdline_p,
 	unsigned long memory_start, memory_end;
 	char c = ' ', *to = command_line, *from = COMMAND_LINE;
 	int len = 0;
+	static unsigned char smptrap=0;
+
+	if(smptrap==1)
+	{
+		return;
+	}
+	smptrap=1;
 
- 	ROOT_DEV = ORIG_ROOT_DEV;
+ 	ROOT_DEV = to_kdev_t(ORIG_ROOT_DEV);
  	drive_info = DRIVE_INFO;
  	screen_info = SCREEN_INFO;
+#ifdef CONFIG_APM
+	apm_bios_info = APM_BIOS_INFO;
+#endif
 	aux_device_present = AUX_DEVICE_INFO;
 	memory_end = (1<<20) + (EXT_MEM_K<<10);
 	memory_end &= PAGE_MASK;
-	ramdisk_size = RAMDISK_SIZE;
+#ifdef CONFIG_BLK_DEV_RAM
+	rd_image_start = RAMDISK_FLAGS & RAMDISK_IMAGE_START_MASK;
+	rd_prompt = ((RAMDISK_FLAGS & RAMDISK_PROMPT_FLAG) != 0);
+	rd_doload = ((RAMDISK_FLAGS & RAMDISK_LOAD_FLAG) != 0);
+#endif
 #ifdef CONFIG_MAX_16M
 	if (memory_end > 16*1024*1024)
 		memory_end = 16*1024*1024;
 #endif
-	if (MOUNT_ROOT_RDONLY)
-		root_mountflags |= MS_RDONLY;
-	memory_start = (unsigned long) &end;
-	init_task.mm->start_code = TASK_SIZE;
-	init_task.mm->end_code = TASK_SIZE + (unsigned long) &etext;
-	init_task.mm->end_data = TASK_SIZE + (unsigned long) &edata;
-	init_task.mm->brk = TASK_SIZE + (unsigned long) &end;
+	if (!MOUNT_ROOT_RDONLY)
+		root_mountflags &= ~MS_RDONLY;
+	memory_start = (unsigned long) &_end;
+	init_task.mm->start_code = PAGE_OFFSET;
+	init_task.mm->end_code = (unsigned long) &_etext;
+	init_task.mm->end_data = (unsigned long) &_edata;
+	init_task.mm->brk = (unsigned long) &_end;
+
+	/* Save unparsed command line copy for /proc/cmdline */
+	memcpy(saved_command_line, COMMAND_LINE, COMMAND_LINE_SIZE);
+	saved_command_line[COMMAND_LINE_SIZE-1] = '\0';
 
 	for (;;) {
-		if (c == ' ' && *(unsigned long *)from == *(unsigned long *)"mem=") {
-			memory_end = simple_strtoul(from+4, &from, 0);
-			if ( *from == 'K' || *from == 'k' ) {
-				memory_end = memory_end << 10;
-				from++;
-			} else if ( *from == 'M' || *from == 'm' ) {
-				memory_end = memory_end << 20;
-				from++;
+		/*
+		 * "mem=nopentium" disables the 4MB page tables.
+		 * "mem=XXX[kKmM]" overrides the BIOS-reported
+		 * memory size
+		 */
+		if (c == ' ' && *(const unsigned long *)from == *(const unsigned long *)"mem=") {
+			if (to != command_line) to--;
+			if (!memcmp(from+4, "nopentium", 9)) {
+				from += 9+4;
+				x86_capability &= ~8;
+			} else {
+				memory_end = simple_strtoul(from+4, &from, 0);
+				if ( *from == 'K' || *from == 'k' ) {
+					memory_end = memory_end << 10;
+					from++;
+				} else if ( *from == 'M' || *from == 'm' ) {
+					memory_end = memory_end << 20;
+					from++;
+				}
 			}
 		}
 		c = *(from++);
@@ -122,60 +179,159 @@ void setup_arch(char **cmdline_p,
 	}
 	*to = '\0';
 	*cmdline_p = command_line;
+	memory_end += PAGE_OFFSET;
 	*memory_start_p = memory_start;
 	*memory_end_p = memory_end;
+
+#ifdef CONFIG_BLK_DEV_INITRD
+	if (LOADER_TYPE) {
+		initrd_start = INITRD_START ? INITRD_START + PAGE_OFFSET : 0;
+		initrd_end = initrd_start+INITRD_SIZE;
+		if (initrd_end > memory_end) {
+			printk("initrd extends beyond end of memory "
+			    "(0x%08lx > 0x%08lx)\ndisabling initrd\n",
+			    initrd_end,memory_end);
+			initrd_start = 0;
+		}
+	}
+#endif
+
 	/* request io space for devices used on all i[345]86 PC'S */
 	request_region(0x00,0x20,"dma1");
 	request_region(0x40,0x20,"timer");
-	request_region(0x70,0x10,"rtc");
 	request_region(0x80,0x20,"dma page reg");
 	request_region(0xc0,0x20,"dma2");
-	request_region(0xf0,0x2,"npu");
-	request_region(0xf8,0x8,"npu");
+	request_region(0xf0,0x10,"npu");
+}
+
+static const char * i486model(unsigned int nr)
+{
+	static const char *model[] = {
+		"0","DX","SX","DX/2","4","SX/2","6","DX/2-WB","DX/4","DX/4-WB",
+		"10","11","12","13","Am5x86-WT","Am5x86-WB"
+	};
+	if (nr < sizeof(model)/sizeof(char *))
+		return model[nr];
+	return NULL;
+}
+
+static const char * i586model(unsigned int nr)
+{
+	static const char *model[] = {
+		"0", "Pentium 60/66","Pentium 75+","OverDrive PODP5V83"
+	};
+	if (nr < sizeof(model)/sizeof(char *))
+		return model[nr];
+	return NULL;
+}
+
+static const char * i686model(unsigned int nr)
+{
+	static const char *model[] = {
+		"PPro A-step", "Pentium Pro"
+	};
+	if (nr < sizeof(model)/sizeof(char *))
+		return model[nr];
+	return NULL;
+}
+
+static const char * getmodel(int x86, int model)
+{
+        const char *p = NULL;
+        static char nbuf[12];
+	switch (x86) {
+		case 4:
+			p = i486model(model);
+			break;
+		case 5:
+			p = i586model(model);
+			break;
+		case 6:
+			p = i686model(model);
+			break;
+	}
+        if (p)
+                return p;
+
+        sprintf(nbuf, "%d", model);
+        return nbuf;
 }
 
 int get_cpuinfo(char * buffer)
 {
-	char *model[2][9]={{"DX","SX","DX/2","4","SX/2","6",
-				"DX/2-WB","DX/4"},
-			{"Pentium 60/66","Pentium 90/100","3",
-				"4","5","6","7","8"}};
-	char mask[2];
-	mask[0] = x86_mask+'@';
-	mask[1] = '\0';
-	return sprintf(buffer,"cpu\t\t: %c86\n"
-			      "model\t\t: %s\n"
-			      "mask\t\t: %s\n"
-			      "vid\t\t: %s\n"
-			      "fdiv_bug\t: %s\n"
-			      "math\t\t: %s\n"
-			      "hlt\t\t: %s\n"
-			      "wp\t\t: %s\n"
-			      "Integrated NPU\t: %s\n"
-			      "Enhanced VM86\t: %s\n"
-			      "IO Breakpoints\t: %s\n"
-			      "4MB Pages\t: %s\n"
-			      "TS Counters\t: %s\n"
-			      "Pentium MSR\t: %s\n"
-			      "Mach. Ch. Exep.\t: %s\n"
-			      "CMPXCHGB8B\t: %s\n"
-		              "BogoMips\t: %lu.%02lu\n",
-			      x86+'0', 
-			      x86_model ? model[x86-4][x86_model-1] : "Unknown",
-			      x86_mask ? mask : "Unknown",
-			      x86_vendor_id,
-			      fdiv_bug ? "yes" : "no",
-			      hard_math ? "yes" : "no",
-			      hlt_works_ok ? "yes" : "no",
-			      wp_works_ok ? "yes" : "no",
-			      x86_capability & 1 ? "yes" : "no",
-			      x86_capability & 2 ? "yes" : "no",
-			      x86_capability & 4 ? "yes" : "no",
-			      x86_capability & 8 ? "yes" : "no",
-			      x86_capability & 16 ? "yes" : "no",
-			      x86_capability & 32 ? "yes" : "no",
-			      x86_capability & 128 ? "yes" : "no",
-			      x86_capability & 256 ? "yes" : "no",
-		              loops_per_sec/500000, (loops_per_sec/5000) % 100
-			      );
+        int i, len = 0;
+        static const char *x86_cap_flags[] = {
+                "fpu", "vme", "de", "pse", "tsc", "msr", "pae", "mce",
+                "cx8", "apic", "10", "11", "mtrr", "pge", "mca", "cmov",
+                "16", "17", "18", "19", "20", "21", "22", "mmx",
+                "24", "25", "26", "27", "28", "29", "30", "31"
+        };
+        
+#ifdef __SMP__
+        int n;
+
+#define CD(X)		(cpu_data[n].X)
+/* SMP has the wrong name for loops_per_sec */
+#define loops_per_sec	udelay_val
+#define CPUN n
+
+        for ( n = 0 ; n < 32 ; n++ ) {
+                if ( cpu_present_map & (1<<n) ) {
+                        if (len) buffer[len++] = '\n'; 
+
+#else
+#define CD(X) (X)
+#define CPUN 0
+#endif
+
+                        len += sprintf(buffer+len,"processor\t: %d\n"
+                                       "cpu\t\t: %c86\n"
+                                       "model\t\t: %s\n"
+                                       "vendor_id\t: %s\n",
+                                       CPUN,
+                                       CD(x86)+'0',
+                                       CD(have_cpuid) ? 
+                                         getmodel(CD(x86), CD(x86_model)) :
+                                         "unknown",
+                                       CD(x86_vendor_id));
+        
+                        if (CD(x86_mask))
+                                len += sprintf(buffer+len,
+                                               "stepping\t: %d\n",
+                                               CD(x86_mask));
+                        else
+                                len += sprintf(buffer+len, 
+                                               "stepping\t: unknown\n");
+        
+                        len += sprintf(buffer+len,
+                                       "fdiv_bug\t: %s\n"
+                                       "hlt_bug\t\t: %s\n"
+                                       "fpu\t\t: %s\n"
+                                       "fpu_exception\t: %s\n"
+                                       "cpuid\t\t: %s\n"
+                                       "wp\t\t: %s\n"
+                                       "flags\t\t:",
+                                       CD(fdiv_bug) ? "yes" : "no",
+                                       CD(hlt_works_ok) ? "no" : "yes",
+                                       CD(hard_math) ? "yes" : "no",
+                                       (CD(hard_math) && ignore_irq13)
+                                         ? "yes" : "no",
+                                       CD(have_cpuid) ? "yes" : "no",
+                                       CD(wp_works_ok) ? "yes" : "no");
+        
+                        for ( i = 0 ; i < 32 ; i++ ) {
+                                if ( CD(x86_capability) & (1 << i) ) {
+                                        len += sprintf(buffer+len, " %s",
+                                                       x86_cap_flags[i]);
+                                }
+                        }
+                        len += sprintf(buffer+len,
+                                       "\nbogomips\t: %lu.%02lu\n",
+                                       CD(loops_per_sec+2500)/500000,
+                                       (CD(loops_per_sec+2500)/5000) % 100);
+#ifdef __SMP__
+                }
+        }
+#endif
+        return len;
 }
diff --git a/arch/i386/kernel/signal.c b/arch/i386/kernel/signal.c
index 3db1a6985..c53a40e16 100644
--- a/arch/i386/kernel/signal.c
+++ b/arch/i386/kernel/signal.c
@@ -4,6 +4,8 @@
  *  Copyright (C) 1991, 1992  Linus Torvalds
  */
 
+#include <linux/config.h>
+
 #include <linux/sched.h>
 #include <linux/mm.h>
 #include <linux/kernel.h>
@@ -13,13 +15,14 @@
 #include <linux/ptrace.h>
 #include <linux/unistd.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 
 #define _S(nr) (1<<((nr)-1))
 
 #define _BLOCKABLE (~(_S(SIGKILL) | _S(SIGSTOP)))
 
 asmlinkage int sys_waitpid(pid_t pid,unsigned long * stat_addr, int options);
+asmlinkage int do_signal(unsigned long oldmask, struct pt_regs * regs);
 
 /*
  * atomically swap in the new signal mask, and wait for a signal.
@@ -40,28 +43,67 @@ asmlinkage int sys_sigsuspend(int restart, unsigned long oldmask, unsigned long
 	}
 }
 
+static inline void restore_i387_hard(struct _fpstate *buf)
+{
+#ifdef __SMP__
+	if (current->flags & PF_USEDFPU) {
+		stts();
+	}
+#else
+	if (current == last_task_used_math) {
+		last_task_used_math = NULL;
+		stts();
+	}
+#endif
+	current->used_math = 1;
+	current->flags &= ~PF_USEDFPU;
+	copy_from_user(&current->tss.i387.hard, buf, sizeof(*buf));
+}
+
+static void restore_i387(struct _fpstate *buf)
+{
+#ifndef CONFIG_MATH_EMULATION
+	restore_i387_hard(buf);
+#else
+	if (hard_math) {
+		restore_i387_hard(buf);
+		return;
+	}
+	restore_i387_soft(buf);
+#endif	
+}
+	
+
 /*
  * This sets regs->esp even though we don't actually use sigstacks yet..
  */
 asmlinkage int sys_sigreturn(unsigned long __unused)
 {
-#define COPY(x) regs->x = context.x
-#define COPY_SEG(x) \
-if ((context.x & 0xfffc) && (context.x & 3) != 3) goto badframe; COPY(x);
-#define COPY_SEG_STRICT(x) \
-if (!(context.x & 0xfffc) || (context.x & 3) != 3) goto badframe; COPY(x);
-	struct sigcontext_struct context;
+#define COPY(x) regs->x = context->x
+#define COPY_SEG(seg) \
+{ unsigned int tmp = context->seg; \
+if ((tmp & 0xfffc) && (tmp & 3) != 3) goto badframe; \
+regs->x##seg = tmp; }
+#define COPY_SEG_STRICT(seg) \
+{ unsigned int tmp = context->seg; \
+if ((tmp & 0xfffc) && (tmp & 3) != 3) goto badframe; \
+regs->x##seg = tmp; }
+#define GET_SEG(seg) \
+{ unsigned int tmp = context->seg; \
+if ((tmp & 0xfffc) && (tmp & 3) != 3) goto badframe; \
+__asm__("mov %w0,%%" #seg: :"r" (tmp)); }
+	struct sigcontext * context;
 	struct pt_regs * regs;
 
 	regs = (struct pt_regs *) &__unused;
-	if (verify_area(VERIFY_READ, (void *) regs->esp, sizeof(context)))
+	context = (struct sigcontext *) regs->esp;
+	if (verify_area(VERIFY_READ, context, sizeof(*context)))
 		goto badframe;
-	memcpy_fromfs(&context,(void *) regs->esp, sizeof(context));
-	current->blocked = context.oldmask & _BLOCKABLE;
+	current->blocked = context->oldmask & _BLOCKABLE;
 	COPY_SEG(ds);
 	COPY_SEG(es);
-	COPY_SEG(fs);
-	COPY_SEG(gs);
+	GET_SEG(fs);
+	GET_SEG(gs);
 	COPY_SEG_STRICT(ss);
 	COPY_SEG_STRICT(cs);
 	COPY(eip);
@@ -70,66 +112,164 @@ if (!(context.x & 0xfffc) || (context.x & 3) != 3) goto badframe; COPY(x);
 	COPY(esp); COPY(ebp);
 	COPY(edi); COPY(esi);
 	regs->eflags &= ~0x40DD5;
-	regs->eflags |= context.eflags & 0x40DD5;
+	regs->eflags |= context->eflags & 0x40DD5;
 	regs->orig_eax = -1;		/* disable syscall checks */
-	return context.eax;
+	if (context->fpstate) {
+		struct _fpstate * buf = context->fpstate;
+		if (verify_area(VERIFY_READ, buf, sizeof(*buf)))
+			goto badframe;
+		restore_i387(buf);
+	}
+	return context->eax;
 badframe:
 	do_exit(SIGSEGV);
 }
 
+static inline struct _fpstate * save_i387_hard(struct _fpstate * buf)
+{
+#ifdef __SMP__
+	if (current->flags & PF_USEDFPU) {
+		__asm__ __volatile__("fnsave %0":"=m" (current->tss.i387.hard));
+		stts();
+		current->flags &= ~PF_USEDFPU;
+	}
+#else
+	if (current == last_task_used_math) {
+		__asm__ __volatile__("fnsave %0":"=m" (current->tss.i387.hard));
+		last_task_used_math = NULL;
+		__asm__ __volatile__("fwait");	/* not needed on 486+ */
+		stts();
+	}
+#endif
+	current->tss.i387.hard.status = current->tss.i387.hard.swd;
+	copy_to_user(buf, &current->tss.i387.hard, sizeof(*buf));
+	current->used_math = 0;
+	return buf;
+}
+
+static struct _fpstate * save_i387(struct _fpstate * buf)
+{
+	if (!current->used_math)
+		return NULL;
+
+#ifndef CONFIG_MATH_EMULATION
+	return save_i387_hard(buf);
+#else
+	if (hard_math)
+		return save_i387_hard(buf);
+	return save_i387_soft(buf);
+#endif
+}
+
 /*
  * Set up a signal frame... Make the stack look the way iBCS2 expects
  * it to look.
  */
-void setup_frame(struct sigaction * sa, unsigned long ** fp, unsigned long eip,
-	struct pt_regs * regs, int signr, unsigned long oldmask)
+static void setup_frame(struct sigaction * sa,
+	struct pt_regs * regs, int signr,
+	unsigned long oldmask)
 {
 	unsigned long * frame;
 
-#define __CODE ((unsigned long)(frame+24))
-#define CODE(x) ((unsigned long *) ((x)+__CODE))
-	frame = *fp;
-	if (regs->ss != USER_DS)
+	frame = (unsigned long *) regs->esp;
+	if ((regs->xss & 0xffff) != USER_DS && sa->sa_restorer)
 		frame = (unsigned long *) sa->sa_restorer;
-	frame -= 32;
-	if (verify_area(VERIFY_WRITE,frame,32*4))
+	frame -= 64;
+	if (verify_area(VERIFY_WRITE,frame,64*4))
 		do_exit(SIGSEGV);
+
 /* set up the "normal" stack seen by the signal handler (iBCS2) */
-	put_fs_long(__CODE,frame);
+#define __CODE ((unsigned long)(frame+24))
+#define CODE(x) ((unsigned long *) ((x)+__CODE))
+	if (put_user(__CODE,frame))
+		do_exit(SIGSEGV);
 	if (current->exec_domain && current->exec_domain->signal_invmap)
-		put_fs_long(current->exec_domain->signal_invmap[signr], frame+1);
+		put_user(current->exec_domain->signal_invmap[signr], frame+1);
 	else
-		put_fs_long(signr, frame+1);
-	put_fs_long(regs->gs, frame+2);
-	put_fs_long(regs->fs, frame+3);
-	put_fs_long(regs->es, frame+4);
-	put_fs_long(regs->ds, frame+5);
-	put_fs_long(regs->edi, frame+6);
-	put_fs_long(regs->esi, frame+7);
-	put_fs_long(regs->ebp, frame+8);
-	put_fs_long((long)*fp, frame+9);
-	put_fs_long(regs->ebx, frame+10);
-	put_fs_long(regs->edx, frame+11);
-	put_fs_long(regs->ecx, frame+12);
-	put_fs_long(regs->eax, frame+13);
-	put_fs_long(current->tss.trap_no, frame+14);
-	put_fs_long(current->tss.error_code, frame+15);
-	put_fs_long(eip, frame+16);
-	put_fs_long(regs->cs, frame+17);
-	put_fs_long(regs->eflags, frame+18);
-	put_fs_long(regs->esp, frame+19);
-	put_fs_long(regs->ss, frame+20);
-	put_fs_long(0,frame+21);		/* 387 state pointer - not implemented*/
+		put_user(signr, frame+1);
+	{
+		unsigned int tmp = 0;
+#define PUT_SEG(seg, mem) \
+__asm__("mov %%" #seg",%w0":"=r" (tmp):"0" (tmp)); put_user(tmp,mem);
+		PUT_SEG(gs, frame+2);
+		PUT_SEG(fs, frame+3);
+	}
+	put_user(regs->xes, frame+4);
+	put_user(regs->xds, frame+5);
+	put_user(regs->edi, frame+6);
+	put_user(regs->esi, frame+7);
+	put_user(regs->ebp, frame+8);
+	put_user(regs->esp, frame+9);
+	put_user(regs->ebx, frame+10);
+	put_user(regs->edx, frame+11);
+	put_user(regs->ecx, frame+12);
+	put_user(regs->eax, frame+13);
+	put_user(current->tss.trap_no, frame+14);
+	put_user(current->tss.error_code, frame+15);
+	put_user(regs->eip, frame+16);
+	put_user(regs->xcs, frame+17);
+	put_user(regs->eflags, frame+18);
+	put_user(regs->esp, frame+19);
+	put_user(regs->xss, frame+20);
+	put_user((unsigned long) save_i387((struct _fpstate *)(frame+32)),frame+21);
 /* non-iBCS2 extensions.. */
-	put_fs_long(oldmask, frame+22);
-	put_fs_long(current->tss.cr2, frame+23);
+	put_user(oldmask, frame+22);
+	put_user(current->tss.cr2, frame+23);
 /* set up the return code... */
-	put_fs_long(0x0000b858, CODE(0));	/* popl %eax ; movl $,%eax */
-	put_fs_long(0x80cd0000, CODE(4));	/* int $0x80 */
-	put_fs_long(__NR_sigreturn, CODE(2));
-	*fp = frame;
+	put_user(0x0000b858, CODE(0));	/* popl %eax ; movl $,%eax */
+	put_user(0x80cd0000, CODE(4));	/* int $0x80 */
+	put_user(__NR_sigreturn, CODE(2));
 #undef __CODE
 #undef CODE
+
+	/* Set up registers for signal handler */
+	regs->esp = (unsigned long) frame;
+	regs->eip = (unsigned long) sa->sa_handler;
+	{
+		unsigned long seg = USER_DS;
+		__asm__("mov %w0,%%fs ; mov %w0,%%gs":"=r" (seg) :"0" (seg));
+		set_fs(seg);
+		regs->xds = seg;
+		regs->xes = seg;
+		regs->xss = seg;
+		regs->xcs = USER_CS;
+	}
+	regs->eflags &= ~TF_MASK;
+}
+
+/*
+ * OK, we're invoking a handler
+ */	
+static void handle_signal(unsigned long signr, struct sigaction *sa,
+	unsigned long oldmask, struct pt_regs * regs)
+{
+	/* are we from a system call? */
+	if (regs->orig_eax >= 0) {
+		/* If so, check system call restarting.. */
+		switch (regs->eax) {
+			case -ERESTARTNOHAND:
+				regs->eax = -EINTR;
+				break;
+
+			case -ERESTARTSYS:
+				if (!(sa->sa_flags & SA_RESTART)) {
+					regs->eax = -EINTR;
+					break;
+				}
+			/* fallthrough */
+			case -ERESTARTNOINTR:
+				regs->eax = regs->orig_eax;
+				regs->eip -= 2;
+		}
+	}
+
+	/* set up the stack frame */
+	setup_frame(sa, regs, signr, oldmask);
+
+	if (sa->sa_flags & SA_ONESHOT)
+		sa->sa_handler = NULL;
+	if (!(sa->sa_flags & SA_NOMASK))
+		current->blocked |= (sa->sa_mask | _S(signr)) & _BLOCKABLE;
 }
 
 /*
@@ -144,18 +284,21 @@ void setup_frame(struct sigaction * sa, unsigned long ** fp, unsigned long eip,
 asmlinkage int do_signal(unsigned long oldmask, struct pt_regs * regs)
 {
 	unsigned long mask = ~current->blocked;
-	unsigned long handler_signal = 0;
-	unsigned long *frame = NULL;
-	unsigned long eip = 0;
 	unsigned long signr;
 	struct sigaction * sa;
 
 	while ((signr = current->signal & mask)) {
+		/*
+		 *	This stops gcc flipping out. Otherwise the assembler
+		 *	including volatiles for the inline function to get
+		 *	current combined with this gets it confused.
+		 */
+		struct task_struct *t=current;
 		__asm__("bsf %3,%1\n\t"
 			"btrl %1,%0"
-			:"=m" (current->signal),"=r" (signr)
-			:"0" (current->signal), "1" (signr));
-		sa = current->sigaction + signr;
+			:"=m" (t->signal),"=r" (signr)
+			:"0" (t->signal), "1" (signr));
+		sa = current->sig->action + signr;
 		signr++;
 		if ((current->flags & PF_PTRACED) && signr != SIGKILL) {
 			current->exit_code = signr;
@@ -171,7 +314,7 @@ asmlinkage int do_signal(unsigned long oldmask, struct pt_regs * regs)
 				current->signal |= _S(signr);
 				continue;
 			}
-			sa = current->sigaction + signr - 1;
+			sa = current->sig->action + signr - 1;
 		}
 		if (sa->sa_handler == SIG_IGN) {
 			if (signr != SIGCHLD)
@@ -188,12 +331,15 @@ asmlinkage int do_signal(unsigned long oldmask, struct pt_regs * regs)
 			case SIGCONT: case SIGCHLD: case SIGWINCH:
 				continue;
 
-			case SIGSTOP: case SIGTSTP: case SIGTTIN: case SIGTTOU:
+			case SIGTSTP: case SIGTTIN: case SIGTTOU:
+				if (is_orphaned_pgrp(current->pgrp))
+					continue;
+			case SIGSTOP:
 				if (current->flags & PF_PTRACED)
 					continue;
 				current->state = TASK_STOPPED;
 				current->exit_code = signr;
-				if (!(current->p_pptr->sigaction[SIGCHLD-1].sa_flags & 
+				if (!(current->p_pptr->sig->action[SIGCHLD-1].sa_flags & 
 						SA_NOCLDSTOP))
 					notify_parent(current);
 				schedule();
@@ -208,53 +354,23 @@ asmlinkage int do_signal(unsigned long oldmask, struct pt_regs * regs)
 				/* fall through */
 			default:
 				current->signal |= _S(signr & 0x7f);
+				current->flags |= PF_SIGNALED;
 				do_exit(signr);
 			}
 		}
-		/*
-		 * OK, we're invoking a handler
-		 */
-		if (regs->orig_eax >= 0) {
-			if (regs->eax == -ERESTARTNOHAND ||
-			   (regs->eax == -ERESTARTSYS && !(sa->sa_flags & SA_RESTART)))
-				regs->eax = -EINTR;
-		}
-		handler_signal |= 1 << (signr-1);
-		mask &= ~sa->sa_mask;
+		handle_signal(signr, sa, oldmask, regs);
+		return 1;
 	}
-	if (regs->orig_eax >= 0 &&
-	    (regs->eax == -ERESTARTNOHAND ||
-	     regs->eax == -ERESTARTSYS ||
-	     regs->eax == -ERESTARTNOINTR)) {
-		regs->eax = regs->orig_eax;
-		regs->eip -= 2;
-	}
-	if (!handler_signal)		/* no handler will be called - return 0 */
-		return 0;
-	eip = regs->eip;
-	frame = (unsigned long *) regs->esp;
-	signr = 1;
-	sa = current->sigaction;
-	for (mask = 1 ; mask ; sa++,signr++,mask += mask) {
-		if (mask > handler_signal)
-			break;
-		if (!(mask & handler_signal))
-			continue;
-		setup_frame(sa,&frame,eip,regs,signr,oldmask);
-		eip = (unsigned long) sa->sa_handler;
-		if (sa->sa_flags & SA_ONESHOT)
-			sa->sa_handler = NULL;
-/* force a supervisor-mode page-in of the signal handler to reduce races */
-		__asm__("testb $0,%%fs:%0": :"m" (*(char *) eip));
-		regs->cs = USER_CS; regs->ss = USER_DS;
-		regs->ds = USER_DS; regs->es = USER_DS;
-		regs->gs = USER_DS; regs->fs = USER_DS;
-		current->blocked |= sa->sa_mask;
-		oldmask |= sa->sa_mask;
+
+	/* Did we come from a system call? */
+	if (regs->orig_eax >= 0) {
+		/* Restart the system call - no handlers present */
+		if (regs->eax == -ERESTARTNOHAND ||
+		    regs->eax == -ERESTARTSYS ||
+		    regs->eax == -ERESTARTNOINTR) {
+			regs->eax = regs->orig_eax;
+			regs->eip -= 2;
+		}
 	}
-	regs->esp = (unsigned long) frame;
-	regs->eip = eip;		/* "return" to the first handler */
-	regs->eflags &= ~TF_MASK;
-	current->tss.trap_no = current->tss.error_code = 0;
-	return 1;
+	return 0;
 }
diff --git a/arch/i386/kernel/smp.c b/arch/i386/kernel/smp.c
new file mode 100644
index 000000000..05c80b711
--- /dev/null
+++ b/arch/i386/kernel/smp.c
@@ -0,0 +1,1204 @@
+/*
+ *	Intel MP v1.1/v1.4 specification support routines for multi-pentium 
+ *	hosts.
+ *
+ *	(c) 1995 Alan Cox, CymruNET Ltd  <alan@cymru.net>
+ *	Supported by Caldera http://www.caldera.com.
+ *	Much of the core SMP work is based on previous work by Thomas Radke, to
+ *	whom a great many thanks are extended.
+ *
+ *	Thanks to Intel for making available several different Pentium and
+ *	Pentium Pro MP machines.
+ *
+ *	This code is released under the GNU public license version 2 or
+ *	later.
+ *
+ *	Fixes
+ *		Felix Koop	:	NR_CPUS used properly
+ *		Jose Renau	:	Handle single CPU case.
+ *		Alan Cox	:	By repeated request 8) - Total BogoMIP report.
+ *		Greg Wright	:	Fix for kernel stacks panic.
+ *		Erich Boleyn	:	MP v1.4 and additional changes.
+ */
+
+#include <linux/kernel.h>
+#include <linux/string.h>
+#include <linux/timer.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/kernel_stat.h>
+#include <linux/delay.h>
+#include <linux/mc146818rtc.h>
+#include <asm/i82489.h>
+#include <linux/smp.h>
+#include <asm/pgtable.h>
+#include <asm/bitops.h>
+#include <asm/pgtable.h>
+#include <asm/smp.h>
+
+/*
+ *	Why isn't this somewhere standard ??
+ */
+ 
+extern __inline int max(int a,int b)
+{
+	if(a>b)
+		return a;
+	return b;
+}
+
+
+int smp_found_config=0;					/* Have we found an SMP box 				*/
+
+unsigned long cpu_present_map = 0;			/* Bitmask of existing CPU's 				*/
+int smp_num_cpus = 1;					/* Total count of live CPU's 				*/
+int smp_threads_ready=0;				/* Set when the idlers are all forked 			*/
+volatile int cpu_number_map[NR_CPUS];			/* which CPU maps to which logical number		*/
+volatile int cpu_logical_map[NR_CPUS];			/* which logical number maps to which CPU		*/
+volatile unsigned long cpu_callin_map[NR_CPUS] = {0,};	/* We always use 0 the rest is ready for parallel delivery */
+volatile unsigned long smp_invalidate_needed;		/* Used for the invalidate map that's also checked in the spinlock */
+struct cpuinfo_x86 cpu_data[NR_CPUS];			/* Per cpu bogomips and other parameters 		*/
+static unsigned int num_processors = 1;			/* Internal processor count				*/
+static unsigned long io_apic_addr = 0xFEC00000;		/* Address of the I/O apic (not yet used) 		*/
+unsigned char boot_cpu_id = 0;				/* Processor that is doing the boot up 			*/
+static unsigned char *kstack_base,*kstack_end;		/* Kernel stack list pointers 				*/
+static int smp_activated = 0;				/* Tripped once we need to start cross invalidating 	*/
+int apic_version[NR_CPUS];				/* APIC version number					*/
+static volatile int smp_commenced=0;			/* Tripped when we start scheduling 		    	*/
+unsigned long apic_addr=0xFEE00000;			/* Address of APIC (defaults to 0xFEE00000)		*/
+unsigned long nlong = 0;				/* dummy used for apic_reg address + 0x20		*/
+unsigned char *apic_reg=((unsigned char *)(&nlong))-0x20;/* Later set to the ioremap() of the APIC 		*/
+unsigned long apic_retval;				/* Just debugging the assembler.. 			*/
+unsigned char *kernel_stacks[NR_CPUS];			/* Kernel stack pointers for CPU's (debugging)		*/
+
+static volatile unsigned char smp_cpu_in_msg[NR_CPUS];	/* True if this processor is sending an IPI		*/
+static volatile unsigned long smp_msg_data;		/* IPI data pointer					*/
+static volatile int smp_src_cpu;			/* IPI sender processor					*/
+static volatile int smp_msg_id;				/* Message being sent					*/
+
+volatile unsigned long kernel_flag=0;			/* Kernel spinlock 					*/
+volatile unsigned char active_kernel_processor = NO_PROC_ID;	/* Processor holding kernel spinlock		*/
+volatile unsigned long kernel_counter=0;		/* Number of times the processor holds the lock		*/
+volatile unsigned long syscall_count=0;			/* Number of times the processor holds the syscall lock	*/
+
+volatile unsigned long ipi_count;			/* Number of IPI's delivered				*/
+#ifdef __SMP_PROF__
+volatile unsigned long smp_spins[NR_CPUS]={0};          /* Count interrupt spins 				*/
+volatile unsigned long smp_spins_syscall[NR_CPUS]={0};  /* Count syscall spins                   		*/
+volatile unsigned long smp_spins_syscall_cur[NR_CPUS]={0};/* Count spins for the actual syscall                 */
+volatile unsigned long smp_spins_sys_idle[NR_CPUS]={0}; /* Count spins for sys_idle 				*/
+volatile unsigned long smp_idle_count[1+NR_CPUS]={0,};	/* Count idle ticks					*/
+#endif
+#if defined (__SMP_PROF__)
+volatile unsigned long smp_idle_map=0;			/* Map for idle processors 				*/
+#endif
+
+volatile unsigned long  smp_proc_in_lock[NR_CPUS] = {0,};/* for computing process time */
+volatile unsigned long smp_process_available=0;
+
+/*#define SMP_DEBUG*/
+
+#ifdef SMP_DEBUG
+#define SMP_PRINTK(x)	printk x
+#else
+#define SMP_PRINTK(x)
+#endif
+
+
+/* 
+ *	Checksum an MP configuration block.
+ */
+ 
+static int mpf_checksum(unsigned char *mp, int len)
+{
+	int sum=0;
+	while(len--)
+		sum+=*mp++;
+	return sum&0xFF;
+}
+
+/*
+ *	Processor encoding in an MP configuration block
+ */
+ 
+static char *mpc_family(int family,int model)
+{
+	static char n[32];
+	static char *model_defs[]=
+	{
+		"80486DX","80486DX",
+		"80486SX","80486DX/2 or 80487",
+		"80486SL","Intel5X2(tm)",
+		"Unknown","Unknown",
+		"80486DX/4"
+	};
+	if(family==0x6)
+		return("Pentium(tm) Pro");
+	if(family==0x5)
+		return("Pentium(tm)");
+	if(family==0x0F && model==0x0F)
+		return("Special controller");
+	if(family==0x04 && model<9)
+		return model_defs[model];
+	sprintf(n,"Unknown CPU [%d:%d]",family, model);
+	return n;
+}
+
+/*
+ *	Read the MPC
+ */
+
+static int smp_read_mpc(struct mp_config_table *mpc)
+{
+	char str[16];
+	int count=sizeof(*mpc);
+	int apics=0;
+	unsigned char *mpt=((unsigned char *)mpc)+count;
+
+	if(memcmp(mpc->mpc_signature,MPC_SIGNATURE,4))
+	{
+		printk("Bad signature [%c%c%c%c].\n",
+			mpc->mpc_signature[0],
+			mpc->mpc_signature[1],
+			mpc->mpc_signature[2],
+			mpc->mpc_signature[3]);
+		return 1;
+	}
+	if(mpf_checksum((unsigned char *)mpc,mpc->mpc_length))
+	{
+		printk("Checksum error.\n");
+		return 1;
+	}
+	if(mpc->mpc_spec!=0x01 && mpc->mpc_spec!=0x04)
+	{
+		printk("Bad Config Table version (%d)!!\n",mpc->mpc_spec);
+		return 1;
+	}
+	memcpy(str,mpc->mpc_oem,8);
+	str[8]=0;
+	printk("OEM ID: %s ",str);
+	memcpy(str,mpc->mpc_productid,12);
+	str[12]=0;
+	printk("Product ID: %s ",str);
+	printk("APIC at: 0x%lX\n",mpc->mpc_lapic);
+
+	/* set the local APIC address */
+	apic_addr = mpc->mpc_lapic;
+	
+	/*
+	 *	Now process the configuration blocks.
+	 */
+	 
+	while(count<mpc->mpc_length)
+	{
+		switch(*mpt)
+		{
+			case MP_PROCESSOR:
+			{
+				struct mpc_config_processor *m=
+					(struct mpc_config_processor *)mpt;
+				if(m->mpc_cpuflag&CPU_ENABLED)
+				{
+					printk("Processor #%d %s APIC version %d\n",
+						m->mpc_apicid, 
+						mpc_family((m->mpc_cpufeature&
+							CPU_FAMILY_MASK)>>8,
+							(m->mpc_cpufeature&
+								CPU_MODEL_MASK)>>4),
+						m->mpc_apicver);
+#ifdef SMP_DEBUG						
+					if(m->mpc_featureflag&(1<<0))
+						printk("    Floating point unit present.\n");
+					if(m->mpc_featureflag&(1<<7))
+						printk("    Machine Exception supported.\n");
+					if(m->mpc_featureflag&(1<<8))
+						printk("    64 bit compare & exchange supported.\n");
+					if(m->mpc_featureflag&(1<<9))
+						printk("    Internal APIC present.\n");
+#endif						
+					if(m->mpc_cpuflag&CPU_BOOTPROCESSOR)
+					{
+						SMP_PRINTK(("    Bootup CPU\n"));
+						boot_cpu_id=m->mpc_apicid;
+					}
+					else	/* Boot CPU already counted */
+						num_processors++;
+						
+					if(m->mpc_apicid>NR_CPUS)
+						printk("Processor #%d unused. (Max %d processors).\n",m->mpc_apicid, NR_CPUS);
+					else						
+					{
+						cpu_present_map|=(1<<m->mpc_apicid);
+						apic_version[m->mpc_apicid]=m->mpc_apicver;
+					}
+				}
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			case MP_BUS:
+			{
+				struct mpc_config_bus *m=
+					(struct mpc_config_bus *)mpt;
+				memcpy(str,m->mpc_bustype,6);
+				str[6]=0;
+				SMP_PRINTK(("Bus #%d is %s\n",
+					m->mpc_busid,
+					str));
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break; 
+			}
+			case MP_IOAPIC:
+			{
+				struct mpc_config_ioapic *m=
+					(struct mpc_config_ioapic *)mpt;
+				if(m->mpc_flags&MPC_APIC_USABLE)
+				{
+					apics++;
+	                                printk("I/O APIC #%d Version %d at 0x%lX.\n",
+	                                	m->mpc_apicid,m->mpc_apicver,
+	                                	m->mpc_apicaddr);
+	                                io_apic_addr = m->mpc_apicaddr;
+	                        }
+                                mpt+=sizeof(*m);
+                                count+=sizeof(*m); 
+                                break;
+			}
+			case MP_INTSRC:
+			{
+				struct mpc_config_intsrc *m=
+					(struct mpc_config_intsrc *)mpt;
+				
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+			case MP_LINTSRC:
+			{
+				struct mpc_config_intlocal *m=
+					(struct mpc_config_intlocal *)mpt;
+				mpt+=sizeof(*m);
+				count+=sizeof(*m);
+				break;
+			}
+		}
+	}
+	if(apics>1)
+		printk("Warning: Multiple APIC's not supported.\n");
+	return num_processors;				
+}
+
+/*
+ *	Scan the memory blocks for an SMP configuration block.
+ */
+ 
+int smp_scan_config(unsigned long base, unsigned long length)
+{
+	unsigned long *bp=(unsigned long *)base;
+	struct intel_mp_floating *mpf;
+	
+	SMP_PRINTK(("Scan SMP from %p for %ld bytes.\n",
+		bp,length));
+	if(sizeof(*mpf)!=16)
+		printk("Error: MPF size\n");
+	
+	while(length>0)
+	{
+		if(*bp==SMP_MAGIC_IDENT)
+		{
+			mpf=(struct intel_mp_floating *)bp;
+			if(mpf->mpf_length==1 && 
+				!mpf_checksum((unsigned char *)bp,16) &&
+				(mpf->mpf_specification == 1
+				 || mpf->mpf_specification == 4) )
+			{
+				printk("Intel MultiProcessor Specification v1.%d\n", mpf->mpf_specification);
+				if(mpf->mpf_feature2&(1<<7))
+					printk("    IMCR and PIC compatibility mode.\n");
+				else
+					printk("    Virtual Wire compatibility mode.\n");
+				smp_found_config=1;
+				/*
+				 *	Now see if we need to read further.
+				 */
+				if(mpf->mpf_feature1!=0)
+				{
+					unsigned long cfg;
+
+					/*
+					 *	We need to know what the local
+					 *	APIC id of the boot CPU is!
+					 */
+
+/*
+ *
+ *	HACK HACK HACK HACK HACK HACK HACK HACK HACK HACK HACK HACK HACK
+ *
+ *	It's not just a crazy hack...  ;-)
+ */
+					/*
+					 *	Standard page mapping
+					 *	functions don't work yet.
+					 *	We know that page 0 is not
+					 *	used.  Steal it for now!
+					 */
+			 
+					cfg=pg0[0];
+					pg0[0] = (apic_addr | 7);
+					local_flush_tlb();
+
+					boot_cpu_id = GET_APIC_ID(*((volatile unsigned long *) APIC_ID));
+
+					/*
+					 *	Give it back
+					 */
+
+					pg0[0]= cfg;
+					local_flush_tlb();
+
+/*
+ *
+ *	END OF HACK   END OF HACK   END OF HACK   END OF HACK   END OF HACK
+ *
+ */					
+					/*
+					 *	2 CPUs, numbered 0 & 1.
+					 */
+					cpu_present_map=3;
+					num_processors=2;
+					printk("I/O APIC at 0xFEC00000.\n");
+					printk("Bus#0 is ");
+				}
+				switch(mpf->mpf_feature1)
+				{
+					case 1:
+					case 5:
+						printk("ISA\n");
+						break;
+					case 2:
+						printk("EISA with no IRQ8 chaining\n");
+						break;
+					case 6:
+					case 3:
+						printk("EISA\n");
+						break;
+					case 4:
+					case 7:
+						printk("MCA\n");
+						break;
+					case 0:
+						break;
+					default:
+						printk("???\nUnknown standard configuration %d\n",
+							mpf->mpf_feature1);
+						return 1;
+				}
+				if(mpf->mpf_feature1>4)
+				{
+					printk("Bus #1 is PCI\n");
+
+					/*
+					 *	Set local APIC version to
+					 *	the integrated form.
+					 *	It's initialized to zero
+					 *	otherwise, representing
+					 *	a discrete 82489DX.
+					 */
+					apic_version[0] = 0x10;
+					apic_version[1] = 0x10;
+				}
+				/*
+				 *	Read the physical hardware table.
+				 *	Anything here will override the
+				 *	defaults.
+				 */
+				if(mpf->mpf_physptr)
+					smp_read_mpc((void *)mpf->mpf_physptr);
+
+				/*
+				 *	Now that the boot CPU id is known,
+				 *	set some other information about it.
+				 */
+				nlong = boot_cpu_id<<24;	/* Dummy 'self' for bootup */
+				cpu_logical_map[0] = boot_cpu_id;
+
+				printk("Processors: %d\n", num_processors);
+				/*
+				 *	Only use the first configuration found.
+				 */
+				return 1;
+			}
+		}
+		bp+=4;
+		length-=16;
+	}
+
+	return 0;
+}
+
+/*
+ *	Trampoline 80x86 program as an array.
+ */
+
+static unsigned char trampoline_data[]={ 
+#include  "trampoline.hex"
+};
+
+/*
+ *	Currently trivial. Write the real->protected mode
+ *	bootstrap into the page concerned. The caller
+ *	has made sure it's suitably aligned.
+ */
+ 
+static void install_trampoline(unsigned char *mp)
+{
+	memcpy(mp,trampoline_data,sizeof(trampoline_data));
+}
+
+/*
+ *	We are called very early to get the low memory for the trampoline/kernel stacks
+ *	This has to be done by mm/init.c to parcel us out nice low memory. We allocate
+ *	the kernel stacks at 4K, 8K, 12K... currently (0-03FF is preserved for SMM and
+ *	other things).
+ */
+ 
+unsigned long smp_alloc_memory(unsigned long mem_base)
+{
+	int size=(num_processors-1)*PAGE_SIZE;		/* Number of stacks needed */
+	/*
+	 *	Our stacks have to be below the 1Mb line, and mem_base on entry
+	 *	is 4K aligned.
+	 */
+	 
+	if(mem_base+size>=0x9F000)
+		panic("smp_alloc_memory: Insufficient low memory for kernel stacks.\n");
+	kstack_base=(void *)mem_base;
+	mem_base+=size;
+	kstack_end=(void *)mem_base;
+	return mem_base;
+}
+	
+/*
+ *	Hand out stacks one at a time.
+ */
+ 
+static void *get_kernel_stack(void)
+{
+	void *stack=kstack_base;
+	if(kstack_base>=kstack_end)
+		return NULL;
+	kstack_base+=PAGE_SIZE;
+	return stack;
+}
+
+
+/*
+ *	The bootstrap kernel entry code has set these up. Save them for
+ *	a given CPU
+ */
+ 
+void smp_store_cpu_info(int id)
+{
+	struct cpuinfo_x86 *c=&cpu_data[id];
+	c->hard_math=hard_math;			/* Always assumed same currently */
+	c->x86=x86;
+	c->x86_model=x86_model;
+	c->x86_mask=x86_mask;
+	c->x86_capability=x86_capability;
+	c->fdiv_bug=fdiv_bug;
+	c->wp_works_ok=wp_works_ok;		/* Always assumed the same currently */
+	c->hlt_works_ok=hlt_works_ok;
+	c->have_cpuid=have_cpuid;
+	c->udelay_val=loops_per_sec;
+	strcpy(c->x86_vendor_id, x86_vendor_id);
+}
+
+/*
+ *	Architecture specific routine called by the kernel just before init is
+ *	fired off. This allows the BP to have everything in order [we hope].
+ *	At the end of this all the AP's will hit the system scheduling and off
+ *	we go. Each AP will load the system gdt's and jump through the kernel
+ *	init into idle(). At this point the scheduler will one day take over 
+ * 	and give them jobs to do. smp_callin is a standard routine
+ *	we use to track CPU's as they power up.
+ */
+
+void smp_commence(void)
+{
+	/*
+	 *	Lets the callin's below out of their loop.
+	 */
+	smp_commenced=1;
+}
+ 
+void smp_callin(void)
+{
+	extern void calibrate_delay(void);
+	int cpuid=GET_APIC_ID(apic_read(APIC_ID));
+	unsigned long l;
+	
+	/*
+	 *	Activate our APIC
+	 */
+	 
+	SMP_PRINTK(("CALLIN %d\n",smp_processor_id()));
+ 	l=apic_read(APIC_SPIV);
+ 	l|=(1<<8);		/* Enable */
+ 	apic_write(APIC_SPIV,l);
+ 	sti();
+	/*
+	 *	Get our bogomips.
+	 */	
+	calibrate_delay();
+	/*
+	 *	Save our processor parameters
+	 */
+ 	smp_store_cpu_info(cpuid);
+	/*
+	 *	Allow the master to continue.
+	 */	
+	set_bit(cpuid, (unsigned long *)&cpu_callin_map[0]);
+	/*
+	 *	Until we are ready for SMP scheduling
+	 */
+	load_ldt(0);
+/*	printk("Testing faulting...\n");
+	*(long *)0=1;		 OOPS... */
+	local_flush_tlb();
+	while(!smp_commenced);
+	if (cpu_number_map[cpuid] == -1)
+		while(1);
+	local_flush_tlb();
+	SMP_PRINTK(("Commenced..\n"));
+	
+	load_TR(cpu_number_map[cpuid]);
+/*	while(1);*/
+}
+
+/*
+ *	Cycle through the processors sending pentium IPI's to boot each.
+ */
+ 
+void smp_boot_cpus(void)
+{
+	int i;
+	int cpucount=0;
+	unsigned long cfg;
+	void *stack;
+	extern unsigned long init_user_stack[];
+	
+	/*
+	 *	Initialize the logical to physical cpu number mapping
+	 */
+
+	for (i = 0; i < NR_CPUS; i++)
+		cpu_number_map[i] = -1;
+
+	/*
+	 *	Setup boot CPU information
+	 */
+ 
+	kernel_stacks[boot_cpu_id]=(void *)init_user_stack;	/* Set up for boot processor first */
+
+	smp_store_cpu_info(boot_cpu_id);			/* Final full version of the data */
+
+	cpu_present_map |= (1 << smp_processor_id());
+	cpu_number_map[boot_cpu_id] = 0;
+	active_kernel_processor=boot_cpu_id;
+
+	/*
+	 *	If we don't conform to the Intel MPS standard, get out
+	 *	of here now!
+	 */
+
+	if (!smp_found_config)
+		return;
+
+	/*
+	 *	Map the local APIC into kernel space
+	 */
+
+	apic_reg = ioremap(apic_addr,4096);
+	
+	if(apic_reg == NULL)
+		panic("Unable to map local apic.\n");
+		
+#ifdef SMP_DEBUG		
+	{
+		int reg;
+
+		/*
+		 *	This is to verify that we're looking at
+		 *	a real local APIC.  Check these against
+		 *	your board if the CPUs aren't getting
+		 *	started for no apparent reason.
+		 */
+
+		reg = apic_read(APIC_VERSION);
+		SMP_PRINTK(("Getting VERSION: %x\n", reg));
+
+		apic_write(APIC_VERSION, 0);
+		reg = apic_read(APIC_VERSION);
+		SMP_PRINTK(("Getting VERSION: %x\n", reg));
+
+		/*
+		 *	The two version reads above should print the same
+		 *	NON-ZERO!!! numbers.  If the second one is zero,
+		 *	there is a problem with the APIC write/read
+		 *	definitions.
+		 *
+		 *	The next two are just to see if we have sane values.
+		 *	They're only really relevant if we're in Virtual Wire
+		 *	compatibility mode, but most boxes are anymore.
+		 */
+
+
+		reg = apic_read(APIC_LVT0);
+		SMP_PRINTK(("Getting LVT0: %x\n", reg));
+
+		reg = apic_read(APIC_LVT1);
+		SMP_PRINTK(("Getting LVT1: %x\n", reg));
+	}
+#endif
+	
+	/*
+	 *	Enable the local APIC
+	 */
+ 
+	cfg=apic_read(APIC_SPIV);
+	cfg|=(1<<8);		/* Enable APIC */
+	apic_write(APIC_SPIV,cfg);
+
+	udelay(10);
+			
+	/*
+	 *	Now scan the cpu present map and fire up the other CPUs.
+	 */
+ 
+	SMP_PRINTK(("CPU map: %lx\n", cpu_present_map));
+		
+	for(i=0;i<NR_CPUS;i++)
+	{
+		/*
+		 *	Don't even attempt to start the boot CPU!
+		 */
+		if (i == boot_cpu_id)
+			continue;
+		
+		if (cpu_present_map & (1 << i))
+		{
+			unsigned long send_status, accept_status;
+			int timeout, num_starts, j;
+			
+			/*
+			 *	We need a kernel stack for each processor.
+			 */
+			
+			stack=get_kernel_stack();	/* We allocated these earlier */
+			if(stack==NULL)
+				panic("No memory for processor stacks.\n");
+			kernel_stacks[i]=stack;
+			install_trampoline(stack);
+
+			printk("Booting processor %d stack %p: ",i,stack);			/* So we set what's up   */
+
+			/*				
+			 *	This grunge runs the startup process for
+			 *	the targeted processor.
+			 */
+
+			SMP_PRINTK(("Setting warm reset code and vector.\n"));
+
+			/*
+			 *	Install a writable page 0 entry.
+			 */
+			 
+			cfg=pg0[0];
+			
+			CMOS_WRITE(0xa, 0xf);
+			pg0[0]=7;
+			local_flush_tlb();
+			*((volatile unsigned short *) 0x469) = ((unsigned long)stack)>>4;
+			*((volatile unsigned short *) 0x467) = 0;
+			
+			/*
+			 *	Protect it again
+			 */
+			 
+			pg0[0]= cfg;
+			local_flush_tlb();
+
+			/*
+			 *	Be paranoid about clearing APIC errors.
+			 */
+
+			if ( apic_version[i] & 0xF0 )
+			{
+				apic_write(APIC_ESR, 0);
+				accept_status = (apic_read(APIC_ESR) & 0xEF);
+			}
+			
+			/*
+			 *	Status is now clean
+			 */
+			 
+			send_status = 	0;
+			accept_status = 0;
+
+			/*
+			 *	Starting actual IPI sequence...
+			 */
+
+			SMP_PRINTK(("Asserting INIT.\n"));
+
+			/*
+			 *	Turn INIT on
+			 */
+			 
+			cfg=apic_read(APIC_ICR2);
+			cfg&=0x00FFFFFF;
+			apic_write(APIC_ICR2, cfg|SET_APIC_DEST_FIELD(i)); 			/* Target chip     	*/
+			cfg=apic_read(APIC_ICR);
+			cfg&=~0xCDFFF;								/* Clear bits 		*/
+			cfg |= (APIC_DEST_FIELD | APIC_DEST_LEVELTRIG
+				| APIC_DEST_ASSERT | APIC_DEST_DM_INIT);
+			apic_write(APIC_ICR, cfg);						/* Send IPI */
+
+			udelay(200);
+			SMP_PRINTK(("Deasserting INIT.\n"));
+
+			cfg=apic_read(APIC_ICR2);
+			cfg&=0x00FFFFFF;
+			apic_write(APIC_ICR2, cfg|SET_APIC_DEST_FIELD(i));			/* Target chip     	*/
+			cfg=apic_read(APIC_ICR);
+			cfg&=~0xCDFFF;								/* Clear bits 		*/
+			cfg |= (APIC_DEST_FIELD | APIC_DEST_LEVELTRIG
+				| APIC_DEST_DM_INIT);
+			apic_write(APIC_ICR, cfg);						/* Send IPI */
+			
+			/*
+			 *	Should we send STARTUP IPIs ?
+			 *
+			 *	Determine this based on the APIC version.
+			 *	If we don't have an integrated APIC, don't
+			 *	send the STARTUP IPIs.
+			 */
+
+			if ( apic_version[i] & 0xF0 )
+				num_starts = 2;
+			else
+				num_starts = 0;
+
+			/*
+			 *	Run STARTUP IPI loop.
+			 */
+
+			for (j = 1; !(send_status || accept_status)
+				    && (j <= num_starts) ; j++)
+			{
+				SMP_PRINTK(("Sending STARTUP #%d.\n",j));
+
+				apic_write(APIC_ESR, 0);
+			
+				/*
+				 *	STARTUP IPI
+				 */
+
+				cfg=apic_read(APIC_ICR2);
+				cfg&=0x00FFFFFF;
+				apic_write(APIC_ICR2, cfg|SET_APIC_DEST_FIELD(i));			/* Target chip     	*/
+				cfg=apic_read(APIC_ICR);
+				cfg&=~0xCDFFF;								/* Clear bits 		*/
+				cfg |= (APIC_DEST_FIELD
+					| APIC_DEST_DM_STARTUP
+					| (((int) stack) >> 12) );					/* Boot on the stack 	*/		
+				apic_write(APIC_ICR, cfg);						/* Kick the second 	*/
+
+				timeout = 0;
+				do {
+					udelay(10);
+				} while ( (send_status = (apic_read(APIC_ICR) & 0x1000))
+					  && (timeout++ < 1000));
+				udelay(200);
+
+				accept_status = (apic_read(APIC_ESR) & 0xEF);
+			}
+
+			if (send_status)		/* APIC never delivered?? */
+				printk("APIC never delivered???\n");
+			if (accept_status)		/* Send accept error */
+				printk("APIC delivery error (%lx).\n", accept_status);
+			
+			if( !(send_status || accept_status) )
+			{
+				for(timeout=0;timeout<50000;timeout++)
+				{
+					if(cpu_callin_map[0]&(1<<i))
+						break;				/* It has booted */
+					udelay(100);				/* Wait 5s total for a response */
+				}
+				if(cpu_callin_map[0]&(1<<i))
+				{
+					cpucount++;
+					/* number CPUs logically, starting from 1 (BSP is 0) */
+					cpu_number_map[i] = cpucount;
+					cpu_logical_map[cpucount] = i;
+				}
+				else
+				{
+					if(*((volatile unsigned char *)8192)==0xA5)
+						printk("Stuck ??\n");
+					else
+						printk("Not responding.\n");
+				}
+			}
+
+			/* mark "stuck" area as not stuck */
+			*((volatile unsigned long *)8192) = 0;
+		}
+		
+		/* 
+		 *	Make sure we unmap all failed CPUs
+		 */
+		 
+		if (cpu_number_map[i] == -1)
+			cpu_present_map &= ~(1 << i);
+	}
+
+	/*
+	 *	Cleanup possible dangling ends...
+	 */
+
+	/*
+	 *	Install writable page 0 entry.
+	 */
+
+	cfg = pg0[0];
+	pg0[0] = 3;	/* writeable, present, addr 0 */
+	local_flush_tlb();
+
+	/*
+	 *	Paranoid:  Set warm reset code and vector here back
+	 *	to default values.
+	 */
+
+	CMOS_WRITE(0, 0xf);
+
+	*((volatile long *) 0x467) = 0;
+
+	/*
+	 *	Restore old page 0 entry.
+	 */
+
+	pg0[0] = cfg;
+	local_flush_tlb();
+
+	/*
+	 *	Allow the user to impress friends.
+	 */
+	
+	if(cpucount==0)
+	{
+		printk("Error: only one processor found.\n");
+		cpu_present_map=(1<<smp_processor_id());
+	}
+	else
+	{
+		unsigned long bogosum=0;
+		for(i=0;i<32;i++)
+		{
+			if(cpu_present_map&(1<<i))
+				bogosum+=cpu_data[i].udelay_val;
+		}
+		printk("Total of %d processors activated (%lu.%02lu BogoMIPS).\n", 
+			cpucount+1, 
+			(bogosum+2500)/500000,
+			((bogosum+2500)/5000)%100);
+		smp_activated=1;
+		smp_num_cpus=cpucount+1;
+	}
+}
+
+
+/*
+ *	A non wait message cannot pass data or cpu source info. This current setup
+ *	is only safe because the kernel lock owner is the only person who can send a message.
+ *
+ *	Wrapping this whole block in a spinlock is not the safe answer either. A processor may
+ *	get stuck with irq's off waiting to send a message and thus not replying to the person
+ *	spinning for a reply....
+ *
+ *	In the end flush tlb ought to be the NMI and a very very short function (to avoid the old
+ *	IDE disk problems), and other messages sent with IRQ's enabled in a civilised fashion. That
+ *	will also boost performance.
+ */
+ 
+void smp_message_pass(int target, int msg, unsigned long data, int wait)
+{
+	unsigned long cfg;
+	unsigned long target_map;
+	int p=smp_processor_id();
+	int irq=0x2d;								/* IRQ 13 */
+	int ct=0;
+	static volatile int message_cpu = NO_PROC_ID;
+
+	/*
+	 *	During boot up send no messages
+	 */
+	 
+	if(!smp_activated || !smp_commenced)
+		return;
+		
+	
+	/*
+	 *	Skip the reschedule if we are waiting to clear a
+	 *	message at this time. The reschedule cannot wait
+	 *	but is not critical.
+	 */
+	
+	if(msg==MSG_RESCHEDULE)							/* Reschedules we do via trap 0x30 */
+	{
+		irq=0x30;
+		if(smp_cpu_in_msg[p])
+			return;
+	}
+
+	/*
+	 *	Sanity check we don't re-enter this across CPU's. Only the kernel
+	 *	lock holder may send messages. For a STOP_CPU we are bringing the
+	 *	entire box to the fastest halt we can.. A reschedule carries
+	 *	no data and can occur during a flush.. guess what panic
+	 *	I got to notice this bug...
+	 */
+	 
+	if(message_cpu!=NO_PROC_ID && msg!=MSG_STOP_CPU && msg!=MSG_RESCHEDULE)
+	{
+		panic("CPU #%d: Message pass %d but pass in progress by %d of %d\n",
+			smp_processor_id(),msg,message_cpu, smp_msg_id);
+	}
+	message_cpu=smp_processor_id();
+	
+
+	/*
+	 *	We are busy
+	 */
+	 	
+	smp_cpu_in_msg[p]++;
+	
+	/*
+	 *	Reschedule is currently special
+	 */
+	 
+	if(msg!=MSG_RESCHEDULE)
+	{
+		smp_src_cpu=p;
+		smp_msg_id=msg;
+		smp_msg_data=data;
+	}
+	
+/*	printk("SMP message pass #%d to %d of %d\n",
+		p, msg, target);*/
+	
+	/*
+	 *	Wait for the APIC to become ready - this should never occur. Its
+	 *	a debugging check really.
+	 */
+	 
+	while(ct<1000)
+	{
+		cfg=apic_read(APIC_ICR);
+		if(!(cfg&(1<<12)))
+			break;
+		ct++;
+		udelay(10);
+	}
+	
+	/*
+	 *	Just pray... there is nothing more we can do
+	 */
+	 
+	if(ct==1000)
+		printk("CPU #%d: previous IPI still not cleared after 10mS", smp_processor_id());
+		
+	/*
+	 *	Program the APIC to deliver the IPI
+	 */
+	 
+	cfg=apic_read(APIC_ICR2);
+	cfg&=0x00FFFFFF;
+	apic_write(APIC_ICR2, cfg|SET_APIC_DEST_FIELD(target));			/* Target chip     		*/
+	cfg=apic_read(APIC_ICR);
+	cfg&=~0xFDFFF;								/* Clear bits 			*/
+	cfg|=APIC_DEST_FIELD|APIC_DEST_DM_FIXED|irq;				/* Send an IRQ 13		*/		
+
+	/*
+	 *	Set the target requirement
+	 */
+	 
+	if(target==MSG_ALL_BUT_SELF)
+	{
+		cfg|=APIC_DEST_ALLBUT;
+		target_map=cpu_present_map;
+		cpu_callin_map[0]=(1<<smp_src_cpu);
+	}
+	else if(target==MSG_ALL)
+	{
+		cfg|=APIC_DEST_ALLINC;
+		target_map=cpu_present_map;
+		cpu_callin_map[0]=0;
+	}
+	else
+	{
+		target_map=(1<<target);
+		cpu_callin_map[0]=0;
+	}
+		
+	/*
+	 *	Send the IPI. The write to APIC_ICR fires this off.
+	 */
+	 
+	apic_write(APIC_ICR, cfg);	
+	
+	/*
+	 *	Spin waiting for completion
+	 */
+	 
+	switch(wait)
+	{
+		case 1:
+			while(cpu_callin_map[0]!=target_map);		/* Spin on the pass		*/
+			break;
+		case 2:
+			while(smp_invalidate_needed);			/* Wait for invalidate map to clear */
+			break;
+	}
+	
+	/*
+	 *	Record our completion
+	 */
+	 
+	smp_cpu_in_msg[p]--;
+	message_cpu=NO_PROC_ID;
+}
+
+/*
+ *	This is fraught with deadlocks. Linus does a flush tlb at a whim
+ *	even with IRQ's off. We have to avoid a pair of crossing flushes
+ *	or we are doomed.  See the notes about smp_message_pass.
+ */
+ 
+void smp_flush_tlb(void)
+{
+	unsigned long flags;
+	if(smp_activated && smp_processor_id()!=active_kernel_processor)
+		panic("CPU #%d:Attempted flush tlb IPI when not AKP(=%d)\n",smp_processor_id(),active_kernel_processor);
+/*	printk("SMI-");*/
+	
+	/*
+	 *	The assignment is safe because it's volatile so the compiler cannot reorder it,
+	 *	because the i586 has strict memory ordering and because only the kernel lock holder
+	 *	may issue a tlb flush. If you break any one of those three change this to an atomic
+	 *	bus locked or.
+	 */
+	
+	smp_invalidate_needed=cpu_present_map&~(1<<smp_processor_id());
+	
+	/*
+	 *	Processors spinning on the lock will see this IRQ late. The smp_invalidate_needed map will
+	 *	ensure they don't do a spurious flush tlb or miss one.
+	 */
+	 
+	save_flags(flags);
+	cli();
+	smp_message_pass(MSG_ALL_BUT_SELF, MSG_INVALIDATE_TLB, 0L, 2);
+	
+	/*
+	 *	Flush the local TLB
+	 */
+	 
+	local_flush_tlb();
+	
+	restore_flags(flags);
+	
+	/*
+	 *	Completed.
+	 */
+	 
+/*	printk("SMID\n");*/
+}
+
+/*	
+ *	Reschedule call back
+ */
+
+void smp_reschedule_irq(int cpl, struct pt_regs *regs)
+{
+#ifdef DEBUGGING_SMP_RESCHED
+	static int ct=0;
+	if(ct==0)
+	{
+		printk("Beginning scheduling on CPU#%d\n",smp_processor_id());
+		ct=1;
+	}
+#endif	
+	if(smp_processor_id()!=active_kernel_processor)
+		panic("SMP Reschedule on CPU #%d, but #%d is active.\n",
+			smp_processor_id(), active_kernel_processor);
+
+	need_resched=1;
+
+	/*
+	 *	Clear the IPI
+	 */
+	apic_read(APIC_SPIV);		/* Dummy read */
+	apic_write(APIC_EOI, 0);	/* Docs say use 0 for future compatibility */
+}	
+
+/*
+ *	Message call back.
+ */
+ 
+void smp_message_irq(int cpl, void *dev_id, struct pt_regs *regs)
+{
+	int i=smp_processor_id();
+/*	static int n=0;
+	if(n++<NR_CPUS)
+		printk("IPI %d->%d(%d,%ld)\n",smp_src_cpu,i,smp_msg_id,smp_msg_data);*/
+	switch(smp_msg_id)
+	{
+		case 0:	/* IRQ 13 testing - boring */
+			return;
+			
+		/*
+		 *	A TLB flush is needed.
+		 */
+		 
+		case MSG_INVALIDATE_TLB:
+			if(clear_bit(i,(unsigned long *)&smp_invalidate_needed))
+				local_flush_tlb();
+			set_bit(i, (unsigned long *)&cpu_callin_map[0]);
+		/*	cpu_callin_map[0]|=1<<smp_processor_id();*/
+			break;
+			
+		/*
+		 *	Halt other CPU's for a panic or reboot
+		 */
+		case MSG_STOP_CPU:
+			while(1)
+			{
+				if(cpu_data[smp_processor_id()].hlt_works_ok)
+					__asm__("hlt");
+			}
+		default:
+			printk("CPU #%d sent invalid cross CPU message to CPU #%d: %X(%lX).\n",
+				smp_src_cpu,smp_processor_id(),smp_msg_id,smp_msg_data);
+			break;
+	}
+	/*
+	 *	Clear the IPI, so we can receive future IPI's
+	 */
+	 
+	apic_read(APIC_SPIV);		/* Dummy read */
+	apic_write(APIC_EOI, 0);	/* Docs say use 0 for future compatibility */
+}
diff --git a/arch/i386/kernel/sys_i386.c b/arch/i386/kernel/sys_i386.c
new file mode 100644
index 000000000..619b372d8
--- /dev/null
+++ b/arch/i386/kernel/sys_i386.c
@@ -0,0 +1,181 @@
+/*
+ * linux/arch/i386/kernel/sys_i386.c
+ *
+ * This file contains various random system calls that
+ * have a non-standard calling sequence on the Linux/i386
+ * platform.
+ */
+
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/mm.h>
+#include <linux/sem.h>
+#include <linux/msg.h>
+#include <linux/shm.h>
+#include <linux/stat.h>
+#include <linux/mman.h>
+#include <linux/file.h>
+
+#include <asm/uaccess.h>
+#include <asm/ipc.h>
+
+/*
+ * sys_pipe() is the normal C calling standard for creating
+ * a pipe. It's not the way unix traditionally does this, though.
+ */
+asmlinkage int sys_pipe(unsigned long * fildes)
+{
+	int fd[2];
+	int error;
+
+	error = verify_area(VERIFY_WRITE,fildes,8);
+	if (error)
+		return error;
+	error = do_pipe(fd);
+	if (error)
+		return error;
+	put_user(fd[0],0+fildes);
+	put_user(fd[1],1+fildes);
+	return 0;
+}
+
+/*
+ * Perform the select(nd, in, out, ex, tv) and mmap() system
+ * calls. Linux/i386 didn't use to be able to handle more than
+ * 4 system call parameters, so these system calls used a memory
+ * block for parameter passing..
+ */
+struct mmap_arg_struct {
+	unsigned long addr;
+	unsigned long len;
+	unsigned long prot;
+	unsigned long flags;
+	unsigned long fd;
+	unsigned long offset;
+};
+
+asmlinkage int old_mmap(struct mmap_arg_struct *arg)
+{
+	int error;
+	struct file * file = NULL;
+	struct mmap_arg_struct a;
+
+	error = verify_area(VERIFY_READ, arg, sizeof(*arg));
+	if (error)
+		return error;
+	copy_from_user(&a, arg, sizeof(a));
+	if (!(a.flags & MAP_ANONYMOUS)) {
+		if (a.fd >= NR_OPEN || !(file = current->files->fd[a.fd]))
+			return -EBADF;
+	}
+	a.flags &= ~(MAP_EXECUTABLE | MAP_DENYWRITE);
+	error = do_mmap(file, a.addr, a.len, a.prot, a.flags, a.offset);
+	return error;
+}
+
+extern asmlinkage int sys_select(int, fd_set *, fd_set *, fd_set *, struct timeval *);
+
+struct sel_arg_struct {
+	unsigned long n;
+	fd_set *inp, *outp, *exp;
+	struct timeval *tvp;
+};
+
+asmlinkage int old_select(struct sel_arg_struct *arg)
+{
+	struct sel_arg_struct a;
+
+	if (copy_from_user(&a, arg, sizeof(a)))
+		return -EFAULT;
+	return sys_select(a.n, a.inp, a.outp, a.exp, a.tvp);
+}
+
+/*
+ * sys_ipc() is the de-multiplexer for the SysV IPC calls..
+ *
+ * This is really horribly ugly.
+ */
+asmlinkage int sys_ipc (uint call, int first, int second, int third, void *ptr, long fifth)
+{
+	int version;
+
+	version = call >> 16; /* hack for backward compatibility */
+	call &= 0xffff;
+
+	if (call <= SEMCTL)
+		switch (call) {
+		case SEMOP:
+			return sys_semop (first, (struct sembuf *)ptr, second);
+		case SEMGET:
+			return sys_semget (first, second, third);
+		case SEMCTL: {
+			union semun fourth;
+			int err;
+			if (!ptr)
+				return -EINVAL;
+			if ((err = verify_area (VERIFY_READ, ptr, sizeof(long))))
+				return err;
+			get_user(fourth.__pad, (void **) ptr);
+			return sys_semctl (first, second, third, fourth);
+			}
+		default:
+			return -EINVAL;
+		}
+	if (call <= MSGCTL) 
+		switch (call) {
+		case MSGSND:
+			return sys_msgsnd (first, (struct msgbuf *) ptr, 
+					   second, third);
+		case MSGRCV:
+			switch (version) {
+			case 0: {
+				struct ipc_kludge tmp;
+				int err;
+				if (!ptr)
+					return -EINVAL;
+				if ((err = verify_area (VERIFY_READ, ptr, sizeof(tmp))))
+					return err;
+				copy_from_user(&tmp,(struct ipc_kludge *) ptr, sizeof (tmp));
+				return sys_msgrcv (first, tmp.msgp, second, tmp.msgtyp, third);
+				}
+			case 1: default:
+				return sys_msgrcv (first, (struct msgbuf *) ptr, second, fifth, third);
+			}
+		case MSGGET:
+			return sys_msgget ((key_t) first, second);
+		case MSGCTL:
+			return sys_msgctl (first, second, (struct msqid_ds *) ptr);
+		default:
+			return -EINVAL;
+		}
+	if (call <= SHMCTL) 
+		switch (call) {
+		case SHMAT:
+			switch (version) {
+			case 0: default: {
+				ulong raddr;
+				int err;
+				if ((err = verify_area(VERIFY_WRITE, (ulong*) third, sizeof(ulong))))
+					return err;
+				err = sys_shmat (first, (char *) ptr, second, &raddr);
+				if (err)
+					return err;
+				put_user (raddr, (ulong *) third);
+				return 0;
+				}
+			case 1:	/* iBCS2 emulator entry point */
+				if (get_fs() != get_ds())
+					return -EINVAL;
+				return sys_shmat (first, (char *) ptr, second, (ulong *) third);
+			}
+		case SHMDT: 
+			return sys_shmdt ((char *)ptr);
+		case SHMGET:
+			return sys_shmget (first, second, third);
+		case SHMCTL:
+			return sys_shmctl (first, second, (struct shmid_ds *) ptr);
+		default:
+			return -EINVAL;
+		}
+	return -EINVAL;
+}
diff --git a/arch/i386/kernel/time.c b/arch/i386/kernel/time.c
new file mode 100644
index 000000000..0eddd66e5
--- /dev/null
+++ b/arch/i386/kernel/time.c
@@ -0,0 +1,500 @@
+/*
+ *  linux/arch/i386/kernel/time.c
+ *
+ *  Copyright (C) 1991, 1992, 1995  Linus Torvalds
+ *
+ * This file contains the PC-specific time handling details:
+ * reading the RTC at bootup, etc..
+ * 1994-07-02    Alan Modra
+ *	fixed set_rtc_mmss, fixed time.year for >= 2000, new mktime
+ * 1995-03-26    Markus Kuhn
+ *      fixed 500 ms bug at call to set_rtc_mmss, fixed DS12887
+ *      precision CMOS clock update
+ * 1996-05-03    Ingo Molnar
+ *      fixed time warps in do_[slow|fast]_gettimeoffset()
+ */
+#include <linux/errno.h>
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/param.h>
+#include <linux/string.h>
+#include <linux/mm.h>
+#include <linux/interrupt.h>
+#include <linux/time.h>
+#include <linux/delay.h>
+
+#include <asm/uaccess.h>
+#include <asm/io.h>
+#include <asm/irq.h>
+#include <asm/delay.h>
+
+#include <linux/mc146818rtc.h>
+#include <linux/timex.h>
+#include <linux/config.h>
+
+extern int setup_x86_irq(int, struct irqaction *);
+
+#ifndef	CONFIG_APM	/* cycle counter may be unreliable */
+/* Cycle counter value at the previous timer interrupt.. */
+static struct {
+	unsigned long low;
+	unsigned long high;
+} init_timer_cc, last_timer_cc;
+
+/*
+ * This is more assembly than C, but it's also rather
+ * timing-critical and we have to use assembler to get
+ * reasonable 64-bit arithmetic
+ */
+static unsigned long do_fast_gettimeoffset(void)
+{
+	register unsigned long eax asm("ax");
+	register unsigned long edx asm("dx");
+	unsigned long tmp, quotient, low_timer, missing_time;
+
+	/* Last jiffy when do_fast_gettimeoffset() was called.. */
+	static unsigned long last_jiffies=0;
+
+	/* Cached "clocks per usec" value.. */
+	static unsigned long cached_quotient=0;
+
+	/* The "clocks per usec" value is calculated once each jiffy */
+	tmp = jiffies;
+	quotient = cached_quotient;
+	low_timer = last_timer_cc.low;
+	missing_time = 0;
+	if (last_jiffies != tmp) {
+		last_jiffies = tmp;
+		/*
+		 * test for hanging bottom handler (this means xtime is not 
+		 * updated yet)
+		 */
+		if (test_bit(TIMER_BH, &bh_active) )
+		{
+			missing_time = 1000020/HZ;
+		}
+
+		/* Get last timer tick in absolute kernel time */
+		eax = low_timer;
+		edx = last_timer_cc.high;
+		__asm__("subl "SYMBOL_NAME_STR(init_timer_cc)",%0\n\t"
+			"sbbl "SYMBOL_NAME_STR(init_timer_cc)"+4,%1"
+			:"=a" (eax), "=d" (edx)
+			:"0" (eax), "1" (edx));
+
+		/*
+		 * Divide the 64-bit time with the 32-bit jiffy counter,
+		 * getting the quotient in clocks.
+		 *
+		 * Giving quotient = "average internal clocks per usec"
+		 */
+		__asm__("divl %2"
+			:"=a" (eax), "=d" (edx)
+			:"r" (tmp),
+			 "0" (eax), "1" (edx));
+
+		edx = 1000020/HZ;
+		tmp = eax;
+		eax = 0;
+
+		__asm__("divl %2"
+			:"=a" (eax), "=d" (edx)
+			:"r" (tmp),
+			 "0" (eax), "1" (edx));
+		cached_quotient = eax;
+		quotient = eax;
+	}
+
+	/* Read the time counter */
+	__asm__(".byte 0x0f,0x31"
+		:"=a" (eax), "=d" (edx));
+
+	/* .. relative to previous jiffy (32 bits is enough) */
+	edx = 0;
+	eax -= low_timer;
+
+	/*
+	 * Time offset = (1000020/HZ * time_low) / quotient.
+	 */
+
+	__asm__("mull %2"
+		:"=a" (eax), "=d" (edx)
+		:"r" (quotient),
+		 "0" (eax), "1" (edx));
+
+	/*
+ 	 * Due to rounding errors (and jiffies inconsistencies),
+	 * we need to check the result so that we'll get a timer
+	 * that is monotonic.
+	 */
+	if (edx >= 1000020/HZ)
+		edx = 1000020/HZ-1;
+
+	eax = edx + missing_time;
+	return eax;
+}
+#endif
+
+/* This function must be called with interrupts disabled 
+ * It was inspired by Steve McCanne's microtime-i386 for BSD.  -- jrs
+ * 
+ * However, the pc-audio speaker driver changes the divisor so that
+ * it gets interrupted rather more often - it loads 64 into the
+ * counter rather than 11932! This has an adverse impact on
+ * do_gettimeoffset() -- it stops working! What is also not
+ * good is that the interval that our timer function gets called
+ * is no longer 10.0002 ms, but 9.9767 ms. To get around this
+ * would require using a different timing source. Maybe someone
+ * could use the RTC - I know that this can interrupt at frequencies
+ * ranging from 8192Hz to 2Hz. If I had the energy, I'd somehow fix
+ * it so that at startup, the timer code in sched.c would select
+ * using either the RTC or the 8253 timer. The decision would be
+ * based on whether there was any other device around that needed
+ * to trample on the 8253. I'd set up the RTC to interrupt at 1024 Hz,
+ * and then do some jiggery to have a version of do_timer that 
+ * advanced the clock by 1/1024 s. Every time that reached over 1/100
+ * of a second, then do all the old code. If the time was kept correct
+ * then do_gettimeoffset could just return 0 - there is no low order
+ * divider that can be accessed.
+ *
+ * Ideally, you would be able to use the RTC for the speaker driver,
+ * but it appears that the speaker driver really needs interrupt more
+ * often than every 120 us or so.
+ *
+ * Anyway, this needs more thought....		pjsg (1993-08-28)
+ * 
+ * If you are really that interested, you should be reading
+ * comp.protocols.time.ntp!
+ */
+
+#define TICK_SIZE tick
+
+static unsigned long do_slow_gettimeoffset(void)
+{
+	int count;
+	static int count_p = 0;
+	unsigned long offset = 0;
+	static unsigned long jiffies_p = 0;
+
+	/*
+	 * cache volatile jiffies temporarily; we have IRQs turned off. 
+	 */
+	unsigned long jiffies_t;
+
+	/* timer count may underflow right here */
+	outb_p(0x00, 0x43);	/* latch the count ASAP */
+	count = inb_p(0x40);	/* read the latched count */
+	count |= inb(0x40) << 8;
+
+ 	jiffies_t = jiffies;
+
+	/*
+	 * avoiding timer inconsistencies (they are rare, but they happen)...
+	 * there are three kinds of problems that must be avoided here:
+	 *  1. the timer counter underflows
+	 *  2. hardware problem with the timer, not giving us continuous time,
+	 *     the counter does small "jumps" upwards on some Pentium systems,
+	 *     thus causes time warps
+	 *  3. we are after the timer interrupt, but the bottom half handler
+	 *     hasn't executed yet.
+	 */
+	if( count > count_p ) {
+		if( jiffies_t == jiffies_p ) {
+			if( count > LATCH-LATCH/100 )
+				offset = TICK_SIZE;
+			else
+				/*
+				 * argh, the timer is bugging we cant do nothing 
+				 * but to give the previous clock value.
+				 */
+				count = count_p;
+		} else {
+			if( test_bit(TIMER_BH, &bh_active) ) {
+				/*
+				 * we have detected a counter underflow.
+			 	 */
+				offset = TICK_SIZE;
+				count_p = count;		
+			} else {
+				count_p = count;
+				jiffies_p = jiffies_t;
+			}
+		}
+	} else {
+		count_p = count;
+		jiffies_p = jiffies_t;
+ 	}
+
+
+	count = ((LATCH-1) - count) * TICK_SIZE;
+	count = (count + LATCH/2) / LATCH;
+
+	return offset + count;
+}
+
+/*
+ * this is only used if we have fast gettimeoffset:
+ */
+void do_x86_get_fast_time(struct timeval * tv)
+{
+	do_gettimeofday(tv);
+}
+
+static unsigned long (*do_gettimeoffset)(void) = do_slow_gettimeoffset;
+
+/*
+ * This version of gettimeofday has near microsecond resolution.
+ */
+void do_gettimeofday(struct timeval *tv)
+{
+	unsigned long flags;
+
+	save_flags(flags);
+	cli();
+	*tv = xtime;
+	tv->tv_usec += do_gettimeoffset();
+	if (tv->tv_usec >= 1000000) {
+		tv->tv_usec -= 1000000;
+		tv->tv_sec++;
+	}
+	restore_flags(flags);
+}
+
+void do_settimeofday(struct timeval *tv)
+{
+	cli();
+	/* This is revolting. We need to set the xtime.tv_usec
+	 * correctly. However, the value in this location is
+	 * is value at the last tick.
+	 * Discover what correction gettimeofday
+	 * would have done, and then undo it!
+	 */
+	tv->tv_usec -= do_gettimeoffset();
+
+	if (tv->tv_usec < 0) {
+		tv->tv_usec += 1000000;
+		tv->tv_sec--;
+	}
+
+	xtime = *tv;
+	time_state = TIME_BAD;
+	time_maxerror = MAXPHASE;
+	time_esterror = MAXPHASE;
+	sti();
+}
+
+
+/*
+ * In order to set the CMOS clock precisely, set_rtc_mmss has to be
+ * called 500 ms after the second nowtime has started, because when
+ * nowtime is written into the registers of the CMOS clock, it will
+ * jump to the next second precisely 500 ms later. Check the Motorola
+ * MC146818A or Dallas DS12887 data sheet for details.
+ */
+static int set_rtc_mmss(unsigned long nowtime)
+{
+	int retval = 0;
+	int real_seconds, real_minutes, cmos_minutes;
+	unsigned char save_control, save_freq_select;
+
+	save_control = CMOS_READ(RTC_CONTROL); /* tell the clock it's being set */
+	CMOS_WRITE((save_control|RTC_SET), RTC_CONTROL);
+
+	save_freq_select = CMOS_READ(RTC_FREQ_SELECT); /* stop and reset prescaler */
+	CMOS_WRITE((save_freq_select|RTC_DIV_RESET2), RTC_FREQ_SELECT);
+
+	cmos_minutes = CMOS_READ(RTC_MINUTES);
+	if (!(save_control & RTC_DM_BINARY) || RTC_ALWAYS_BCD)
+		BCD_TO_BIN(cmos_minutes);
+
+	/*
+	 * since we're only adjusting minutes and seconds,
+	 * don't interfere with hour overflow. This avoids
+	 * messing with unknown time zones but requires your
+	 * RTC not to be off by more than 15 minutes
+	 */
+	real_seconds = nowtime % 60;
+	real_minutes = nowtime / 60;
+	if (((abs(real_minutes - cmos_minutes) + 15)/30) & 1)
+		real_minutes += 30;		/* correct for half hour time zone */
+	real_minutes %= 60;
+
+	if (abs(real_minutes - cmos_minutes) < 30) {
+		if (!(save_control & RTC_DM_BINARY) || RTC_ALWAYS_BCD) {
+			BIN_TO_BCD(real_seconds);
+			BIN_TO_BCD(real_minutes);
+		}
+		CMOS_WRITE(real_seconds,RTC_SECONDS);
+		CMOS_WRITE(real_minutes,RTC_MINUTES);
+	} else
+		retval = -1;
+
+	/* The following flags have to be released exactly in this order,
+	 * otherwise the DS12887 (popular MC146818A clone with integrated
+	 * battery and quartz) will not reset the oscillator and will not
+	 * update precisely 500 ms later. You won't find this mentioned in
+	 * the Dallas Semiconductor data sheets, but who believes data
+	 * sheets anyway ...                           -- Markus Kuhn
+	 */
+	CMOS_WRITE(save_control, RTC_CONTROL);
+	CMOS_WRITE(save_freq_select, RTC_FREQ_SELECT);
+
+	return retval;
+}
+
+/* last time the cmos clock got updated */
+static long last_rtc_update = 0;
+
+/*
+ * timer_interrupt() needs to keep up the real-time clock,
+ * as well as call the "do_timer()" routine every clocktick
+ */
+static inline void timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
+{
+	do_timer(regs);
+
+	/*
+	 * If we have an externally synchronized Linux clock, then update
+	 * CMOS clock accordingly every ~11 minutes. Set_rtc_mmss() has to be
+	 * called as close as possible to 500 ms before the new second starts.
+	 */
+	if (time_state != TIME_BAD && xtime.tv_sec > last_rtc_update + 660 &&
+	    xtime.tv_usec > 500000 - (tick >> 1) &&
+	    xtime.tv_usec < 500000 + (tick >> 1))
+	  if (set_rtc_mmss(xtime.tv_sec) == 0)
+	    last_rtc_update = xtime.tv_sec;
+	  else
+	    last_rtc_update = xtime.tv_sec - 600; /* do it again in 60 s */
+	/* As we return to user mode fire off the other CPU schedulers.. this is 
+	   basically because we don't yet share IRQ's around. This message is
+	   rigged to be safe on the 386 - basically it's a hack, so don't look
+	   closely for now.. */
+	/*smp_message_pass(MSG_ALL_BUT_SELF, MSG_RESCHEDULE, 0L, 0); */
+	    
+}
+
+#ifndef	CONFIG_APM	/* cycle counter may be unreliable */
+/*
+ * This is the same as the above, except we _also_ save the current
+ * cycle counter value at the time of the timer interrupt, so that
+ * we later on can estimate the time of day more exactly.
+ */
+static void pentium_timer_interrupt(int irq, void *dev_id, struct pt_regs *regs)
+{
+	/* read Pentium cycle counter */
+	__asm__(".byte 0x0f,0x31"
+		:"=a" (last_timer_cc.low),
+		 "=d" (last_timer_cc.high));
+	timer_interrupt(irq, NULL, regs);
+}
+#endif
+
+/* Converts Gregorian date to seconds since 1970-01-01 00:00:00.
+ * Assumes input in normal date format, i.e. 1980-12-31 23:59:59
+ * => year=1980, mon=12, day=31, hour=23, min=59, sec=59.
+ *
+ * [For the Julian calendar (which was used in Russia before 1917,
+ * Britain & colonies before 1752, anywhere else before 1582,
+ * and is still in use by some communities) leave out the
+ * -year/100+year/400 terms, and add 10.]
+ *
+ * This algorithm was first published by Gauss (I think).
+ *
+ * WARNING: this function will overflow on 2106-02-07 06:28:16 on
+ * machines were long is 32-bit! (However, as time_t is signed, we
+ * will already get problems at other places on 2038-01-19 03:14:08)
+ */
+static inline unsigned long mktime(unsigned int year, unsigned int mon,
+	unsigned int day, unsigned int hour,
+	unsigned int min, unsigned int sec)
+{
+	if (0 >= (int) (mon -= 2)) {	/* 1..12 -> 11,12,1..10 */
+		mon += 12;	/* Puts Feb last since it has leap day */
+		year -= 1;
+	}
+	return (((
+	    (unsigned long)(year/4 - year/100 + year/400 + 367*mon/12 + day) +
+	      year*365 - 719499
+	    )*24 + hour /* now have hours */
+	   )*60 + min /* now have minutes */
+	  )*60 + sec; /* finally seconds */
+}
+
+unsigned long get_cmos_time(void)
+{
+	unsigned int year, mon, day, hour, min, sec;
+	int i;
+
+	/* The Linux interpretation of the CMOS clock register contents:
+	 * When the Update-In-Progress (UIP) flag goes from 1 to 0, the
+	 * RTC registers show the second which has precisely just started.
+	 * Let's hope other operating systems interpret the RTC the same way.
+	 */
+	/* read RTC exactly on falling edge of update flag */
+	for (i = 0 ; i < 1000000 ; i++)	/* may take up to 1 second... */
+		if (CMOS_READ(RTC_FREQ_SELECT) & RTC_UIP)
+			break;
+	for (i = 0 ; i < 1000000 ; i++)	/* must try at least 2.228 ms */
+		if (!(CMOS_READ(RTC_FREQ_SELECT) & RTC_UIP))
+			break;
+	do { /* Isn't this overkill ? UIP above should guarantee consistency */
+		sec = CMOS_READ(RTC_SECONDS);
+		min = CMOS_READ(RTC_MINUTES);
+		hour = CMOS_READ(RTC_HOURS);
+		day = CMOS_READ(RTC_DAY_OF_MONTH);
+		mon = CMOS_READ(RTC_MONTH);
+		year = CMOS_READ(RTC_YEAR);
+	} while (sec != CMOS_READ(RTC_SECONDS));
+	if (!(CMOS_READ(RTC_CONTROL) & RTC_DM_BINARY) || RTC_ALWAYS_BCD)
+	  {
+	    BCD_TO_BIN(sec);
+	    BCD_TO_BIN(min);
+	    BCD_TO_BIN(hour);
+	    BCD_TO_BIN(day);
+	    BCD_TO_BIN(mon);
+	    BCD_TO_BIN(year);
+	  }
+	if ((year += 1900) < 1970)
+		year += 100;
+	return mktime(year, mon, day, hour, min, sec);
+}
+
+static struct irqaction irq0  = { timer_interrupt, 0, 0, "timer", NULL, NULL};
+
+void time_init(void)
+{
+	xtime.tv_sec = get_cmos_time();
+	xtime.tv_usec = 0;
+
+	/* If we have the CPU hardware time counters, use them */
+#ifndef CONFIG_APM
+				/* Don't use them if a suspend/resume could
+                                   corrupt the timer value.  This problem
+                                   needs more debugging. */
+	if (x86_capability & 16) {
+		do_gettimeoffset = do_fast_gettimeoffset;
+		do_get_fast_time = do_x86_get_fast_time;
+
+		if( strcmp( x86_vendor_id, "AuthenticAMD" ) == 0 ) {
+			if( x86 == 5 ) {
+				if( x86_model == 0 ) {
+					/* turn on cycle counters during power down */
+					__asm__ __volatile__ (" movl $0x83, %%ecx \n \
+								.byte 0x0f,0x32 \n \
+								orl $1,%%eax \n \
+								.byte 0x0f,0x30 \n " 
+                                                                : : : "ax", "cx", "dx" );
+					udelay(500);
+				}
+			}
+		}
+
+		/* read Pentium cycle counter */
+		__asm__(".byte 0x0f,0x31"
+			:"=a" (init_timer_cc.low),
+			 "=d" (init_timer_cc.high));
+		irq0.handler = pentium_timer_interrupt;
+	}
+#endif
+	setup_x86_irq(0, &irq0);
+}
diff --git a/arch/i386/kernel/trampoline.S b/arch/i386/kernel/trampoline.S
new file mode 100644
index 000000000..29bcc9040
--- /dev/null
+++ b/arch/i386/kernel/trampoline.S
@@ -0,0 +1,74 @@
+!
+!	Trampoline.S	Derived from Setup.S by Linus Torvalds
+!
+!	Entry: CS:IP point to the start of our code, we are 
+!	in real mode with no stack, but the rest of the 
+!	trampoline page to make our stack and everything else
+!	is a mystery.
+!
+!	In fact we don't actually need a stack so we don't
+!	set one up.
+!
+!	We jump into the boot/compressed/head.S code. So you'd
+!	better be running a compressed kernel image or you
+!	won't get very far.
+!
+#define __ASSEMBLY__
+#include <asm/segment.h>
+
+.text
+	extrn startup32
+
+entry start
+start:
+!	nop
+!	jmp start	! 	Test
+	mov ax,cs	!	Code and data in the same place
+	mov ds,ax	!
+	mov cx,ax	!	Pass stack info to the 32bit boot
+	add cx,cx
+	add cx,cx
+	add cx,cx
+	add cx,cx	!	Segment -> Offset
+	add cx, #4096   !	End of page is wanted
+	mov     bx,#1	!	Flag an SMP trampoline
+	cli		!	We should be safe anyway
+
+	lidt	idt_48	!	load idt with 0,0
+	lgdt	gdt_48	!	load gdt with whatever is appropriate
+
+	xor	ax,ax	
+	inc	ax	!	protected mode (PE) bit
+	lmsw	ax	!	Into protected mode
+	jmp	flush_instr
+flush_instr:
+	jmpi	8192+startup32,KERNEL_CS	!	Jump to the 32bit trampoline code
+!	jmpi	0x100000,KERNEL_CS		!	Jump into the 32bit startup
+!	.byte   0x66,0x67			!	32bit
+!	.byte	0xea,0x00,0x00,0x10,0x00,0x10,0x00	!jmpi	.0x100000,KERNEL_CS
+
+gdt:
+	.word	0,0,0,0		! dummy
+
+	.word	0,0,0,0		! unused
+
+	.word	0x07FF		! 8Mb - limit=2047 (2048*4096=8Mb)
+	.word	0x0000		! base address=0
+	.word	0x9A00		! code read/exec
+	.word	0x00C0		! granularity=4096, 386
+
+	.word	0x07FF		! 8Mb - limit=2047 (2048*4096=8Mb)
+	.word	0x0000		! base address=0
+	.word	0x9200		! data read/write
+	.word	0x00C0		! granularity=4096, 386
+
+idt_48:
+	.word	0			! idt limit=0
+	.word	0,0			! idt base=0L
+
+gdt_48:
+	.word	0x800		! gdt limit=2048, 256 GDT entries
+	.word	8192+gdt,0x0	! gdt base = 8192+gdt (first SMP CPU)
+				! we load the others with the first table
+				! saves rewriting gdt_48 for each
+
diff --git a/arch/i386/kernel/trampoline32.S b/arch/i386/kernel/trampoline32.S
new file mode 100644
index 000000000..fafa972a0
--- /dev/null
+++ b/arch/i386/kernel/trampoline32.S
@@ -0,0 +1,20 @@
+!
+!	32bit side of the trampoline code
+!
+#define __ASSEMBLY__
+#include <asm/segment.h>
+!
+!
+!	Anything but a relative address here will be wrong by 8K...
+!
+	.globl startup32
+.text
+startup32:
+!	Run the kernel
+	mov	eax,#KERNEL_DS
+	mov	ds,ax
+	mov	eax,#0xA5A5A5A5
+	mov	[8192],eax
+	jmpi	0x100000,KERNEL_CS
+l1:
+	.byte	0xEA,0x00,0x00,0x10,0x00,0x10,0x00
diff --git a/arch/i386/kernel/traps.c b/arch/i386/kernel/traps.c
index 6dd7fc65b..8dfab81c5 100644
--- a/arch/i386/kernel/traps.c
+++ b/arch/i386/kernel/traps.c
@@ -10,6 +10,7 @@
  * to mainly kill the offending process (probably by giving it a signal,
  * but possibly by killing it outright if necessary).
  */
+#include <linux/config.h>
 #include <linux/head.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
@@ -21,12 +22,12 @@
 #include <linux/mm.h>
 
 #include <asm/system.h>
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 #include <asm/io.h>
 
 asmlinkage int system_call(void);
 asmlinkage void lcall7(void);
-struct desc_struct default_ldt;
+struct desc_struct default_ldt = { 0, 0 };
 
 static inline void console_verbose(void)
 {
@@ -39,9 +40,7 @@ asmlinkage void do_##name(struct pt_regs * regs, long error_code) \
 { \
 	tsk->tss.error_code = error_code; \
 	tsk->tss.trap_no = trapnr; \
-	if (signr == SIGTRAP && current->flags & PF_PTRACED) \
-		current->blocked &= ~(1 << (SIGTRAP-1)); \
-	send_sig(signr, tsk, 1); \
+	force_sig(signr, tsk); \
 	die_if_kernel(str,regs,error_code); \
 }
 
@@ -82,6 +81,7 @@ asmlinkage void page_fault(void);
 asmlinkage void coprocessor_error(void);
 asmlinkage void reserved(void);
 asmlinkage void alignment_check(void);
+asmlinkage void spurious_interrupt_bug(void);
 
 int kstack_depth_to_print = 24;
 
@@ -93,31 +93,32 @@ int kstack_depth_to_print = 24;
 #define VMALLOC_OFFSET (8*1024*1024)
 #define MODULE_RANGE (8*1024*1024)
 
-/*static*/ void die_if_kernel(char * str, struct pt_regs * regs, long err)
+/*static*/ void die_if_kernel(const char * str, struct pt_regs * regs, long err)
 {
 	int i;
 	unsigned long esp;
 	unsigned short ss;
 	unsigned long *stack, addr, module_start, module_end;
-	extern char start_kernel, etext;
+	extern char start_kernel, _etext;
 
 	esp = (unsigned long) &regs->esp;
 	ss = KERNEL_DS;
-	if ((regs->eflags & VM_MASK) || (3 & regs->cs) == 3)
+	if ((regs->eflags & VM_MASK) || (3 & regs->xcs) == 3)
 		return;
-	if (regs->cs & 3) {
+	if (regs->xcs & 3) {
 		esp = regs->esp;
-		ss = regs->ss;
+		ss = regs->xss & 0xffff;
 	}
 	console_verbose();
 	printk("%s: %04lx\n", str, err & 0xffff);
-	printk("EIP:    %04x:%08lx\nEFLAGS: %08lx\n", 0xffff & regs->cs,regs->eip,regs->eflags);
+	printk("CPU:    %d\n", smp_processor_id());
+	printk("EIP:    %04x:[<%08lx>]\nEFLAGS: %08lx\n", 0xffff & regs->xcs,regs->eip,regs->eflags);
 	printk("eax: %08lx   ebx: %08lx   ecx: %08lx   edx: %08lx\n",
 		regs->eax, regs->ebx, regs->ecx, regs->edx);
 	printk("esi: %08lx   edi: %08lx   ebp: %08lx   esp: %08lx\n",
 		regs->esi, regs->edi, regs->ebp, esp);
-	printk("ds: %04x   es: %04x   fs: %04x   gs: %04x   ss: %04x\n",
-		regs->ds, regs->es, regs->fs, regs->gs, ss);
+	printk("ds: %04x   es: %04x   ss: %04x\n",
+		regs->xds & 0xffff, regs->xes & 0xffff, ss);
 	store_TR(i);
 	if (STACK_MAGIC != *(unsigned long *)current->kernel_stack_page)
 		printk("Corrupted stack page\n");
@@ -134,7 +135,8 @@ int kstack_depth_to_print = 24;
 	printk("\nCall Trace: ");
 	stack = (unsigned long *) esp;
 	i = 1;
-	module_start = ((high_memory + VMALLOC_OFFSET) & ~(VMALLOC_OFFSET-1));
+	module_start = PAGE_OFFSET + (max_mapnr << PAGE_SHIFT);
+	module_start = ((module_start + VMALLOC_OFFSET) & ~(VMALLOC_OFFSET-1));
 	module_end = module_start + MODULE_RANGE;
 	while (((long) stack & 4095) != 0) {
 		addr = get_seg_long(ss, stack++);
@@ -147,17 +149,17 @@ int kstack_depth_to_print = 24;
 		 * out the call path that was taken.
 		 */
 		if (((addr >= (unsigned long) &start_kernel) &&
-		     (addr <= (unsigned long) &etext)) ||
+		     (addr <= (unsigned long) &_etext)) ||
 		    ((addr >= module_start) && (addr <= module_end))) {
 			if (i && ((i % 8) == 0))
 				printk("\n       ");
-			printk("%08lx ", addr);
+			printk("[<%08lx>] ", addr);
 			i++;
 		}
 	}
 	printk("\nCode: ");
 	for(i=0;i<20;i++)
-		printk("%02x ",0xff & get_seg_byte(regs->cs,(i+(char *)regs->eip)));
+		printk("%02x ",0xff & get_seg_byte(regs->xcs & 0xffff,(i+(char *)regs->eip)));
 	printk("\n");
 	do_exit(SIGSEGV);
 }
@@ -173,8 +175,8 @@ DO_ERROR( 9, SIGFPE,  "coprocessor segment overrun", coprocessor_segment_overrun
 DO_ERROR(10, SIGSEGV, "invalid TSS", invalid_TSS, current)
 DO_ERROR(11, SIGBUS,  "segment not present", segment_not_present, current)
 DO_ERROR(12, SIGBUS,  "stack segment", stack_segment, current)
-DO_ERROR(15, SIGSEGV, "reserved", reserved, current)
 DO_ERROR(17, SIGSEGV, "alignment check", alignment_check, current)
+DO_ERROR(18, SIGSEGV, "reserved", reserved, current)
 
 asmlinkage void do_general_protection(struct pt_regs * regs, long error_code)
 {
@@ -185,16 +187,20 @@ asmlinkage void do_general_protection(struct pt_regs * regs, long error_code)
 	die_if_kernel("general protection",regs,error_code);
 	current->tss.error_code = error_code;
 	current->tss.trap_no = 13;
-	send_sig(SIGSEGV, current, 1);	
+	force_sig(SIGSEGV, current);	
 }
 
 asmlinkage void do_nmi(struct pt_regs * regs, long error_code)
 {
+#ifdef CONFIG_SMP_NMI_INVAL
+	smp_flush_tlb_rcv();
+#else
 #ifndef CONFIG_IGNORE_NMI
 	printk("Uhhuh. NMI received. Dazed and confused, but trying to continue\n");
 	printk("You probably have a hardware problem with your RAM chips or a\n");
 	printk("power saving mode enabled.\n");
 #endif	
+#endif
 }
 
 asmlinkage void do_debug(struct pt_regs * regs, long error_code)
@@ -203,12 +209,10 @@ asmlinkage void do_debug(struct pt_regs * regs, long error_code)
 		handle_vm86_debug((struct vm86_regs *) regs, error_code);
 		return;
 	}
-	if (current->flags & PF_PTRACED)
-		current->blocked &= ~(1 << (SIGTRAP-1));
-	send_sig(SIGTRAP, current, 1);
+	force_sig(SIGTRAP, current);
 	current->tss.trap_no = 1;
 	current->tss.error_code = error_code;
-	if ((regs->cs & 3) == 0) {
+	if ((regs->xcs & 3) == 0) {
 		/* If this is a kernel mode trap, then reset db7 and allow us to continue */
 		__asm__("movl %0,%%db7"
 			: /* no output */
@@ -219,40 +223,35 @@ asmlinkage void do_debug(struct pt_regs * regs, long error_code)
 }
 
 /*
- * Allow the process which triggered the interrupt to recover the error
- * condition.
- *  - the status word is saved in the cs selector.
- *  - the tag word is saved in the operand selector.
- *  - the status word is then cleared and the tags all set to Empty.
- *
- * This will give sufficient information for complete recovery provided that
- * the affected process knows or can deduce the code and data segments
- * which were in force when the exception condition arose.
- *
  * Note that we play around with the 'TS' bit to hopefully get
  * the correct behaviour even in the presence of the asynchronous
  * IRQ13 behaviour
  */
 void math_error(void)
 {
-	struct i387_hard_struct * env;
+	struct task_struct * task;
 
 	clts();
-	if (!last_task_used_math) {
+#ifdef __SMP__
+	task = current;
+#else
+	task = last_task_used_math;
+	last_task_used_math = NULL;
+	if (!task) {
 		__asm__("fnclex");
 		return;
 	}
-	env = &last_task_used_math->tss.i387.hard;
-	send_sig(SIGFPE, last_task_used_math, 1);
-	last_task_used_math->tss.trap_no = 16;
-	last_task_used_math->tss.error_code = 0;
-	__asm__ __volatile__("fnsave %0":"=m" (*env));
-	last_task_used_math = NULL;
+#endif
+	/*
+	 *	Save the info for the exception handler
+	 */
+	__asm__ __volatile__("fnsave %0":"=m" (task->tss.i387.hard));
+	task->flags&=~PF_USEDFPU;
 	stts();
-	env->fcs = (env->swd & 0x0000ffff) | (env->fcs & 0xffff0000);
-	env->fos = env->twd;
-	env->swd &= 0xffff3800;
-	env->twd = 0xffffffff;
+
+	force_sig(SIGFPE, task);
+	task->tss.trap_no = 16;
+	task->tss.error_code = 0;
 }
 
 asmlinkage void do_coprocessor_error(struct pt_regs * regs, long error_code)
@@ -261,6 +260,12 @@ asmlinkage void do_coprocessor_error(struct pt_regs * regs, long error_code)
 	math_error();
 }
 
+asmlinkage void do_spurious_interrupt_bug(struct pt_regs * regs,
+					  long error_code)
+{
+	printk("Ignoring P6 Local APIC Spurious Interrupt Bug...\n");
+}
+
 /*
  *  'math_state_restore()' saves the current math information in the
  * old math state array, and gets the new ones from the current task
@@ -270,23 +275,37 @@ asmlinkage void do_coprocessor_error(struct pt_regs * regs, long error_code)
  */
 asmlinkage void math_state_restore(void)
 {
-	__asm__ __volatile__("clts");
+	__asm__ __volatile__("clts");		/* Allow maths ops (or we recurse) */
+
+/*
+ *	SMP is actually simpler than uniprocessor for once. Because
+ *	we can't pull the delayed FPU switching trick Linus does
+ *	we simply have to do the restore each context switch and
+ *	set the flag. switch_to() will always save the state in
+ *	case we swap processors. We also don't use the coprocessor
+ *	timer - IRQ 13 mode isn't used with SMP machines (thank god).
+ */
+#ifndef __SMP__
 	if (last_task_used_math == current)
 		return;
-	timer_table[COPRO_TIMER].expires = jiffies+50;
-	timer_active |= 1<<COPRO_TIMER;	
 	if (last_task_used_math)
 		__asm__("fnsave %0":"=m" (last_task_used_math->tss.i387));
 	else
 		__asm__("fnclex");
 	last_task_used_math = current;
-	if (current->used_math) {
+#endif
+
+	if(current->used_math)
 		__asm__("frstor %0": :"m" (current->tss.i387));
-	} else {
+	else
+	{
+		/*
+		 *	Our first FPU usage, clean the chip.
+		 */
 		__asm__("fninit");
-		current->used_math=1;
+		current->used_math = 1;
 	}
-	timer_active &= ~(1<<COPRO_TIMER);
+	current->flags|=PF_USEDFPU;		/* So we fnsave on switch_to() */
 }
 
 #ifndef CONFIG_MATH_EMULATION
@@ -295,7 +314,7 @@ asmlinkage void math_emulate(long arg)
 {
   printk("math-emulation not enabled and no coprocessor found.\n");
   printk("killing %s.\n",current->comm);
-  send_sig(SIGFPE,current,1);
+  force_sig(SIGFPE,current);
   schedule();
 }
 
@@ -305,8 +324,16 @@ void trap_init(void)
 {
 	int i;
 	struct desc_struct * p;
-
-	if (strncmp((char*)0x0FFFD9, "EISA", 4) == 0)
+	static int smptrap=0;
+	
+	if(smptrap)
+	{
+		__asm__("pushfl ; andl $0xffffbfff,(%esp) ; popfl");
+		load_ldt(0);
+		return;
+	}
+	smptrap++;
+	if (readl(0x0FFFD9) == 'E' + ('I'<<8) + ('S'<<16) + ('A'<<24))
 		EISA_bus = 1;
 	set_call_gate(&default_ldt,lcall7);
 	set_trap_gate(0,&divide_error);
@@ -324,7 +351,7 @@ void trap_init(void)
 	set_trap_gate(12,&stack_segment);
 	set_trap_gate(13,&general_protection);
 	set_trap_gate(14,&page_fault);
-	set_trap_gate(15,&reserved);
+	set_trap_gate(15,&spurious_interrupt_bug);
 	set_trap_gate(16,&coprocessor_error);
 	set_trap_gate(17,&alignment_check);
 	for (i=18;i<48;i++)
diff --git a/arch/i386/kernel/vm86.c b/arch/i386/kernel/vm86.c
index d55f8248f..6cceba36c 100644
--- a/arch/i386/kernel/vm86.c
+++ b/arch/i386/kernel/vm86.c
@@ -11,7 +11,7 @@
 #include <linux/ptrace.h>
 #include <linux/mm.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 #include <asm/pgtable.h>
 #include <asm/io.h>
 
@@ -59,8 +59,12 @@ asmlinkage struct pt_regs * save_v86_state(struct vm86_regs * regs)
 		do_exit(SIGSEGV);
 	}
 	set_flags(regs->eflags, VEFLAGS, VIF_MASK | current->tss.v86mask);
-	memcpy_tofs(&current->tss.vm86_info->regs,regs,sizeof(*regs));
-	put_fs_long(current->tss.screen_bitmap,&current->tss.vm86_info->screen_bitmap);
+	tmp = copy_to_user(&current->tss.vm86_info->regs,regs,sizeof(*regs));
+	tmp += put_user(current->tss.screen_bitmap,&current->tss.vm86_info->screen_bitmap);
+	if (tmp) {
+		printk("vm86: could not access userspace vm86_info\n");
+		do_exit(SIGSEGV);
+	}
 	tmp = current->tss.esp0;
 	current->tss.esp0 = current->saved_kernel_stack;
 	current->saved_kernel_stack = 0;
@@ -74,7 +78,7 @@ static void mark_screen_rdonly(struct task_struct * tsk)
 	pte_t *pte;
 	int i;
 
-	pgd = pgd_offset(tsk, 0xA0000);
+	pgd = pgd_offset(tsk->mm, 0xA0000);
 	if (pgd_none(*pgd))
 		return;
 	if (pgd_bad(*pgd)) {
@@ -93,32 +97,32 @@ static void mark_screen_rdonly(struct task_struct * tsk)
 	pte = pte_offset(pmd, 0xA0000);
 	for (i = 0; i < 32; i++) {
 		if (pte_present(*pte))
-			*pte = pte_wrprotect(*pte);
+			set_pte(pte, pte_wrprotect(*pte));
 		pte++;
 	}
-	invalidate();
+	flush_tlb();
 }
 
 asmlinkage int sys_vm86(struct vm86_struct * v86)
 {
 	struct vm86_struct info;
+	struct task_struct *tsk = current;
 	struct pt_regs * pt_regs = (struct pt_regs *) &v86;
-	int error;
 
-	if (current->saved_kernel_stack)
+	if (tsk->saved_kernel_stack)
 		return -EPERM;
-	/* v86 must be readable (now) and writable (for save_v86_state) */
-	error = verify_area(VERIFY_WRITE,v86,sizeof(*v86));
-	if (error)
-		return error;
-	memcpy_fromfs(&info,v86,sizeof(info));
+	if (copy_from_user(&info,v86,sizeof(info)))
+		return -EFAULT;
 /*
  * make sure the vm86() system call doesn't try to do anything silly
  */
 	info.regs.__null_ds = 0;
 	info.regs.__null_es = 0;
-	info.regs.__null_fs = 0;
-	info.regs.__null_gs = 0;
+
+/* we are clearing fs,gs later just before "jmp ret_from_sys_call",
+ * because starting with Linux 2.1.x they aren't no longer saved/restored
+ */
+	
 /*
  * The eflags register is also special: we cannot trust that the user
  * has set it up safely, so this makes sure interrupt etc flags are
@@ -131,16 +135,16 @@ asmlinkage int sys_vm86(struct vm86_struct * v86)
 
 	switch (info.cpu_type) {
 		case CPU_286:
-			current->tss.v86mask = 0;
+			tsk->tss.v86mask = 0;
 			break;
 		case CPU_386:
-			current->tss.v86mask = NT_MASK | IOPL_MASK;
+			tsk->tss.v86mask = NT_MASK | IOPL_MASK;
 			break;
 		case CPU_486:
-			current->tss.v86mask = AC_MASK | NT_MASK | IOPL_MASK;
+			tsk->tss.v86mask = AC_MASK | NT_MASK | IOPL_MASK;
 			break;
 		default:
-			current->tss.v86mask = ID_MASK | AC_MASK | NT_MASK | IOPL_MASK;
+			tsk->tss.v86mask = ID_MASK | AC_MASK | NT_MASK | IOPL_MASK;
 			break;
 	}
 
@@ -148,17 +152,19 @@ asmlinkage int sys_vm86(struct vm86_struct * v86)
  * Save old state, set default return value (%eax) to 0
  */
 	pt_regs->eax = 0;
-	current->saved_kernel_stack = current->tss.esp0;
-	current->tss.esp0 = (unsigned long) pt_regs;
-	current->tss.vm86_info = v86;
+	tsk->saved_kernel_stack = tsk->tss.esp0;
+	tsk->tss.esp0 = (unsigned long) pt_regs;
+	tsk->tss.vm86_info = v86;
 
-	current->tss.screen_bitmap = info.screen_bitmap;
+	tsk->tss.screen_bitmap = info.screen_bitmap;
 	if (info.flags & VM86_SCREEN_BITMAP)
-		mark_screen_rdonly(current);
-	__asm__ __volatile__("movl %0,%%esp\n\t"
+		mark_screen_rdonly(tsk);
+	__asm__ __volatile__(
+		"xorl %%eax,%%eax; mov %%ax,%%fs; mov %%ax,%%gs\n\t"
+		"movl %0,%%esp\n\t"
 		"jmp ret_from_sys_call"
 		: /* no outputs */
-		:"r" (&info.regs));
+		:"r" (&info.regs), "b" (tsk) : "ax");
 	return 0;
 }
 
@@ -170,7 +176,7 @@ static inline void return_to_32bit(struct vm86_regs * regs16, int retval)
 	regs32->eax = retval;
 	__asm__ __volatile__("movl %0,%%esp\n\t"
 		"jmp ret_from_sys_call"
-		: : "r" (regs32));
+		: : "r" (regs32), "b" (current));
 }
 
 static inline void set_IF(struct vm86_regs * regs)
@@ -217,10 +223,10 @@ static inline unsigned long get_vflags(struct vm86_regs * regs)
 
 static inline int is_revectored(int nr, struct revectored_struct * bitmap)
 {
-	__asm__ __volatile__("btl %2,%%fs:%1\n\tsbbl %0,%0"
-		:"=r" (nr)
-		:"m" (*bitmap),"r" (nr));
-	return nr;
+	unsigned long map;
+	if (get_user(map, bitmap->__map + (nr >> 5)))
+		return 1;
+	return test_bit(nr & ((1 << 5)-1), &map);
 }
 
 /*
@@ -231,16 +237,16 @@ static inline int is_revectored(int nr, struct revectored_struct * bitmap)
 #define pushb(base, ptr, val) \
 __asm__ __volatile__( \
 	"decw %w0\n\t" \
-	"movb %2,%%fs:0(%1,%0)" \
+	"movb %2,0(%1,%0)" \
 	: "=r" (ptr) \
 	: "r" (base), "q" (val), "0" (ptr))
 
 #define pushw(base, ptr, val) \
 __asm__ __volatile__( \
 	"decw %w0\n\t" \
-	"movb %h2,%%fs:0(%1,%0)\n\t" \
+	"movb %h2,0(%1,%0)\n\t" \
 	"decw %w0\n\t" \
-	"movb %b2,%%fs:0(%1,%0)" \
+	"movb %b2,0(%1,%0)" \
 	: "=r" (ptr) \
 	: "r" (base), "q" (val), "0" (ptr))
 
@@ -248,21 +254,21 @@ __asm__ __volatile__( \
 __asm__ __volatile__( \
 	"decw %w0\n\t" \
 	"rorl $16,%2\n\t" \
-	"movb %h2,%%fs:0(%1,%0)\n\t" \
+	"movb %h2,0(%1,%0)\n\t" \
 	"decw %w0\n\t" \
-	"movb %b2,%%fs:0(%1,%0)\n\t" \
+	"movb %b2,0(%1,%0)\n\t" \
 	"decw %w0\n\t" \
 	"rorl $16,%2\n\t" \
-	"movb %h2,%%fs:0(%1,%0)\n\t" \
+	"movb %h2,0(%1,%0)\n\t" \
 	"decw %w0\n\t" \
-	"movb %b2,%%fs:0(%1,%0)" \
+	"movb %b2,0(%1,%0)" \
 	: "=r" (ptr) \
 	: "r" (base), "q" (val), "0" (ptr))
 
 #define popb(base, ptr) \
 ({ unsigned long __res; \
 __asm__ __volatile__( \
-	"movb %%fs:0(%1,%0),%b2\n\t" \
+	"movb 0(%1,%0),%b2\n\t" \
 	"incw %w0" \
 	: "=r" (ptr), "=r" (base), "=q" (__res) \
 	: "0" (ptr), "1" (base), "2" (0)); \
@@ -271,9 +277,9 @@ __res; })
 #define popw(base, ptr) \
 ({ unsigned long __res; \
 __asm__ __volatile__( \
-	"movb %%fs:0(%1,%0),%b2\n\t" \
+	"movb 0(%1,%0),%b2\n\t" \
 	"incw %w0\n\t" \
-	"movb %%fs:0(%1,%0),%h2\n\t" \
+	"movb 0(%1,%0),%h2\n\t" \
 	"incw %w0" \
 	: "=r" (ptr), "=r" (base), "=q" (__res) \
 	: "0" (ptr), "1" (base), "2" (0)); \
@@ -282,14 +288,14 @@ __res; })
 #define popl(base, ptr) \
 ({ unsigned long __res; \
 __asm__ __volatile__( \
-	"movb %%fs:0(%1,%0),%b2\n\t" \
+	"movb 0(%1,%0),%b2\n\t" \
 	"incw %w0\n\t" \
-	"movb %%fs:0(%1,%0),%h2\n\t" \
+	"movb 0(%1,%0),%h2\n\t" \
 	"incw %w0\n\t" \
 	"rorl $16,%2\n\t" \
-	"movb %%fs:0(%1,%0),%b2\n\t" \
+	"movb 0(%1,%0),%b2\n\t" \
 	"incw %w0\n\t" \
-	"movb %%fs:0(%1,%0),%h2\n\t" \
+	"movb 0(%1,%0),%h2\n\t" \
 	"incw %w0\n\t" \
 	"rorl $16,%2" \
 	: "=r" (ptr), "=r" (base), "=q" (__res) \
@@ -298,22 +304,31 @@ __res; })
 
 static void do_int(struct vm86_regs *regs, int i, unsigned char * ssp, unsigned long sp)
 {
-	unsigned short seg = get_fs_word((void *) ((i<<2)+2));
-
-	if (seg == BIOSSEG || regs->cs == BIOSSEG ||
-	    is_revectored(i, &current->tss.vm86_info->int_revectored))
-		return_to_32bit(regs, VM86_INTx + (i << 8));
+	unsigned long *intr_ptr, segoffs;
+	
+	if (regs->cs == BIOSSEG)
+		goto cannot_handle;
+	if (is_revectored(i, &current->tss.vm86_info->int_revectored))
+		goto cannot_handle;
 	if (i==0x21 && is_revectored(AH(regs),&current->tss.vm86_info->int21_revectored))
-		return_to_32bit(regs, VM86_INTx + (i << 8));
+		goto cannot_handle;
+	intr_ptr = (unsigned long *) (i << 2);
+	if (get_user(segoffs, intr_ptr))
+		goto cannot_handle;
+	if ((segoffs >> 16) == BIOSSEG)
+		goto cannot_handle;
 	pushw(ssp, sp, get_vflags(regs));
 	pushw(ssp, sp, regs->cs);
 	pushw(ssp, sp, IP(regs));
-	regs->cs = seg;
+	regs->cs = segoffs >> 16;
 	SP(regs) -= 6;
-	IP(regs) = get_fs_word((void *) (i<<2));
+	IP(regs) = segoffs & 0xffff;
 	clear_TF(regs);
 	clear_IF(regs);
 	return;
+
+cannot_handle:
+	return_to_32bit(regs, VM86_INTx + (i << 8));
 }
 
 void handle_vm86_debug(struct vm86_regs * regs, long error_code)
@@ -367,6 +382,7 @@ void handle_vm86_fault(struct vm86_regs * regs, long error_code)
 			set_vflags_long(popl(ssp, sp), regs);
 			return;
 		}
+		break;
 
 	/* pushf */
 	case 0x9c:
@@ -413,8 +429,10 @@ void handle_vm86_fault(struct vm86_regs * regs, long error_code)
 		IP(regs)++;
 		set_IF(regs);
 		return;
-
-	default:
-		return_to_32bit(regs, VM86_UNKNOWN);
 	}
+
+	/*
+	 * We didn't recognize it, let the emulator take care of it..
+	 */
+	return_to_32bit(regs, VM86_UNKNOWN);
 }
diff --git a/arch/i386/lib/Makefile b/arch/i386/lib/Makefile
new file mode 100644
index 000000000..cf54dd7df
--- /dev/null
+++ b/arch/i386/lib/Makefile
@@ -0,0 +1,16 @@
+#
+# Makefile for i386-specific library files..
+#
+
+ifdef SMP
+.S.o:
+	$(CC) -D__ASSEMBLY__ $(AFLAGS) -traditional -c $< -o $*.o
+else
+.S.o:
+	$(CC) -D__ASSEMBLY__ -traditional -c $< -o $*.o
+endif
+
+L_TARGET = lib.a
+L_OBJS  = checksum.o semaphore.o
+
+include $(TOPDIR)/Rules.make
diff --git a/arch/i386/lib/checksum.c b/arch/i386/lib/checksum.c
new file mode 100644
index 000000000..d2bf48b08
--- /dev/null
+++ b/arch/i386/lib/checksum.c
@@ -0,0 +1,194 @@
+/*
+ * 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.
+ *
+ *		IP/TCP/UDP checksumming routines
+ *
+ * Authors:	Jorge Cwik, <jorge@laser.satlink.net>
+ *		Arnt Gulbrandsen, <agulbra@nvg.unit.no>
+ *		Tom May, <ftom@netcom.com>
+ *		Lots of code moved 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.
+ */
+
+#include <net/checksum.h>
+
+/*
+ * computes a partial checksum, e.g. for TCP/UDP fragments
+ */
+
+unsigned int csum_partial(const unsigned char * buff, int len, unsigned int sum) {
+	  /*
+	   * Experiments with ethernet and slip connections show that buff
+	   * is aligned on either a 2-byte or 4-byte boundary.  We get at
+	   * least a 2x speedup on 486 and Pentium if it is 4-byte aligned.
+	   * Fortunately, it is easy to convert 2-byte alignment to 4-byte
+	   * alignment for the unrolled loop.
+	   */
+	__asm__("
+	    testl $2, %%esi		# Check alignment.
+	    jz 2f			# Jump if alignment is ok.
+	    subl $2, %%ecx		# Alignment uses up two bytes.
+	    jae 1f			# Jump if we had at least two bytes.
+	    addl $2, %%ecx		# ecx was < 2.  Deal with it.
+	    jmp 4f
+1:	    movw (%%esi), %%bx
+	    addl $2, %%esi
+	    addw %%bx, %%ax
+	    adcl $0, %%eax
+2:
+	    movl %%ecx, %%edx
+	    shrl $5, %%ecx
+	    jz 2f
+	    testl %%esi, %%esi
+1:	    movl (%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 4(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 8(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 12(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 16(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 20(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 24(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    movl 28(%%esi), %%ebx
+	    adcl %%ebx, %%eax
+	    lea 32(%%esi), %%esi
+	    dec %%ecx
+	    jne 1b
+	    adcl $0, %%eax
+2:	    movl %%edx, %%ecx
+	    andl $0x1c, %%edx
+	    je 4f
+	    shrl $2, %%edx
+	    testl %%esi, %%esi
+3:	    adcl (%%esi), %%eax
+	    lea 4(%%esi), %%esi
+	    dec %%edx
+	    jne 3b
+	    adcl $0, %%eax
+4:	    andl $3, %%ecx
+	    jz 7f
+	    cmpl $2, %%ecx
+	    jb 5f
+	    movw (%%esi),%%cx
+	    leal 2(%%esi),%%esi
+	    je 6f
+	    shll $16,%%ecx
+5:	    movb (%%esi),%%cl
+6:	    addl %%ecx,%%eax
+	    adcl $0, %%eax
+7:	    "
+	: "=a"(sum)
+	: "0"(sum), "c"(len), "S"(buff)
+	: "bx", "cx", "dx", "si");
+	return(sum);
+}
+
+
+
+/*
+ * copy from ds while checksumming, otherwise like csum_partial
+ */
+
+unsigned int csum_partial_copy(const char *src, char *dst, 
+				  int len, int sum) {
+    __asm__("
+	testl $2, %%edi		# Check alignment.
+	jz 2f			# Jump if alignment is ok.
+	subl $2, %%ecx		# Alignment uses up two bytes.
+	jae 1f			# Jump if we had at least two bytes.
+	addl $2, %%ecx		# ecx was < 2.  Deal with it.
+	jmp 4f
+1:	movw (%%esi), %%bx
+	addl $2, %%esi
+	movw %%bx, (%%edi)
+	addl $2, %%edi
+	addw %%bx, %%ax
+	adcl $0, %%eax
+2:
+	movl %%ecx, %%edx
+	shrl $5, %%ecx
+	jz 2f
+	testl %%esi, %%esi
+1:	movl (%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, (%%edi)
+
+	movl 4(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 4(%%edi)
+
+	movl 8(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 8(%%edi)
+
+	movl 12(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 12(%%edi)
+
+	movl 16(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 16(%%edi)
+
+	movl 20(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 20(%%edi)
+
+	movl 24(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 24(%%edi)
+
+	movl 28(%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, 28(%%edi)
+
+	lea 32(%%esi), %%esi
+	lea 32(%%edi), %%edi
+	dec %%ecx
+	jne 1b
+	adcl $0, %%eax
+2:	movl %%edx, %%ecx
+	andl $28, %%edx
+	je 4f
+	shrl $2, %%edx
+	testl %%esi, %%esi
+3:	movl (%%esi), %%ebx
+	adcl %%ebx, %%eax
+	movl %%ebx, (%%edi)
+	lea 4(%%esi), %%esi
+	lea 4(%%edi), %%edi
+	dec %%edx
+	jne 3b
+	adcl $0, %%eax
+4:	andl $3, %%ecx
+	jz 7f
+	cmpl $2, %%ecx
+	jb 5f
+	movw (%%esi), %%cx
+	leal 2(%%esi), %%esi
+	movw %%cx, (%%edi)
+	leal 2(%%edi), %%edi
+	je 6f
+	shll $16,%%ecx
+5:	movb (%%esi), %%cl
+	movb %%cl, (%%edi)
+6:	addl %%ecx, %%eax
+	adcl $0, %%eax
+7:
+	"
+	: "=a" (sum)
+	: "0"(sum), "c"(len), "S"(src), "D" (dst)
+	: "bx", "cx", "dx", "si", "di" );
+    return(sum);
+}
diff --git a/arch/i386/lib/semaphore.S b/arch/i386/lib/semaphore.S
new file mode 100644
index 000000000..f2d11099f
--- /dev/null
+++ b/arch/i386/lib/semaphore.S
@@ -0,0 +1,35 @@
+/*
+ *  linux/arch/i386/lib/semaphore.S
+ *
+ *  Copyright (C) 1996  Linus Torvalds
+ */
+
+#include <linux/linkage.h>
+
+/*
+ * "down_failed" is called with the eventual return address
+ * in %eax, and the address of the semaphore in %ecx. We need
+ * to call "__down()", and then re-try until we succeed..
+ */
+ENTRY(__down_failed)
+	pushl %eax	/* return address */
+	pushl %edx	/* save %edx */
+1:	pushl %ecx	/* save %ecx (and argument) */
+	call SYMBOL_NAME(__down)
+	popl %ecx	/* restore %ecx (count on __down not changing it) */
+#ifdef __SMP__
+	lock
+#endif
+	decl (%ecx)
+	js 1b
+	popl %edx	/* restore %edx */
+	ret
+
+ENTRY(__up_wakeup)
+	pushl %eax	/* return address */
+	pushl %edx	/* save %edx */
+	pushl %ecx	/* save %ecx (and argument) */
+	call SYMBOL_NAME(__up)
+	popl %ecx	/* restore %ecx (count on __up not changing it) */
+	popl %edx	/* restore %edx */
+	ret
diff --git a/arch/i386/math-emu/Makefile b/arch/i386/math-emu/Makefile
index 2d391a9e6..1bd2cb40d 100644
--- a/arch/i386/math-emu/Makefile
+++ b/arch/i386/math-emu/Makefile
@@ -2,21 +2,17 @@
 #               Makefile for wm-FPU-emu
 #
 
+L_TARGET := math.a
+
 #DEBUG	= -DDEBUGGING
 DEBUG	=
 PARANOID = -DPARANOID
-CFLAGS	:= $(CFLAGS) $(PARANOID) $(DEBUG) -fno-builtin
-
-.c.o:
-	$(CC) $(CFLAGS) $(MATH_EMULATION) -c $<
+CFLAGS	:= $(CFLAGS) $(PARANOID) $(DEBUG) -fno-builtin $(MATH_EMULATION)
 
 .S.o:
-	$(CC) -D__ASSEMBLER__ $(PARANOID) -c $<
-
-.s.o:
-	$(CC) -c $<
+	$(CC) -D__ASSEMBLY__ $(PARANOID) -c $<
 
-OBJS =	fpu_entry.o div_small.o errors.o \
+L_OBJS =fpu_entry.o div_small.o errors.o \
 	fpu_arith.o fpu_aux.o fpu_etc.o fpu_trig.o \
 	load_store.o get_address.o \
 	poly_atan.o poly_l2.o poly_2xm1.o poly_sin.o poly_tan.o \
@@ -28,25 +24,7 @@ OBJS =	fpu_entry.o div_small.o errors.o \
 	div_Xsig.o polynom_Xsig.o round_Xsig.o \
 	shr_Xsig.o mul_Xsig.o
 
-math.a: $(OBJS)
-	rm -f math.a
-	$(AR) rcs math.a $(OBJS)
-	sync
-
-dep:
-	$(CPP) -M *.c > .depend
-	$(CPP) -D__ASSEMBLER__ -M *.S >> .depend
+include $(TOPDIR)/Rules.make
 
 proto:
 	cproto -e -DMAKING_PROTO *.c >fpu_proto.h
-
-modules:
-
-dummy:
-
-#
-# include a dependency file if one exists
-#
-ifeq (.depend,$(wildcard .depend))
-include .depend
-endif
diff --git a/arch/i386/math-emu/README b/arch/i386/math-emu/README
index 2c0acb423..5158b4b81 100644
--- a/arch/i386/math-emu/README
+++ b/arch/i386/math-emu/README
@@ -1,9 +1,9 @@
  +---------------------------------------------------------------------------+
  |  wm-FPU-emu   an FPU emulator for 80386 and 80486SX microprocessors.      |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
+ | Copyright (C) 1992,1993,1994,1995,1996                                    |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@suburbia.net               |
  |                                                                           |
  |    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      |
@@ -42,13 +42,11 @@ but is very close.  See "Limitations" later in this file for a list of
 some differences.
 
 Please report bugs, etc to me at:
-       billm@vaxc.cc.monash.edu.au
-  or at:
-       billm@jacobi.maths.monash.edu.au
+       billm@suburbia.net
 
 
 --Bill Metzenthen
-  August 1994
+  October 1996
 
 
 ----------------------- Internals of wm-FPU-emu -----------------------
diff --git a/arch/i386/math-emu/control_w.h b/arch/i386/math-emu/control_w.h
index ef5fced39..6bd7654b7 100644
--- a/arch/i386/math-emu/control_w.h
+++ b/arch/i386/math-emu/control_w.h
@@ -10,7 +10,7 @@
 #ifndef _CONTROLW_H_
 #define _CONTROLW_H_
 
-#ifdef __ASSEMBLER__
+#ifdef __ASSEMBLY__
 #define	_Const_(x)	$##x
 #else
 #define	_Const_(x)	x
diff --git a/arch/i386/math-emu/div_Xsig.S b/arch/i386/math-emu/div_Xsig.S
index 67d8be964..fd83732fc 100644
--- a/arch/i386/math-emu/div_Xsig.S
+++ b/arch/i386/math-emu/div_Xsig.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Division subroutine for 96 bit quantities                                 |
  |                                                                           |
- | Copyright (C) 1994                                                        |
+ | Copyright (C) 1994,1995                                                   |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -28,7 +28,7 @@
  +---------------------------------------------------------------------------*/
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 
 #define	XsigLL(x)	(x)
@@ -74,11 +74,7 @@ FPU_result_1:
 
 
 .text
-	.align 2,144
-
-.globl _div_Xsig
-
-_div_Xsig:
+ENTRY(div_Xsig)
 	pushl	%ebp
 	movl	%esp,%ebp
 #ifndef NON_REENTRANT_FPU
diff --git a/arch/i386/math-emu/div_small.S b/arch/i386/math-emu/div_small.S
index 0225a96d4..13ab2b7ae 100644
--- a/arch/i386/math-emu/div_small.S
+++ b/arch/i386/math-emu/div_small.S
@@ -4,8 +4,9 @@
  |                                                                           |
  | Divide a 64 bit integer by a 32 bit integer & return remainder.           |
  |                                                                           |
- | Copyright (C) 1992    W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1995                                                   |
+ |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -14,14 +15,10 @@
  |    unsigned long div_small(unsigned long long *x, unsigned long y)        |
  +---------------------------------------------------------------------------*/
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 .text
-	.align 2,144
-
-.globl _div_small
-
-_div_small:
+ENTRY(div_small)
 	pushl	%ebp
 	movl	%esp,%ebp
 
diff --git a/arch/i386/math-emu/errors.c b/arch/i386/math-emu/errors.c
index e34eec942..f76020d1f 100644
--- a/arch/i386/math-emu/errors.c
+++ b/arch/i386/math-emu/errors.c
@@ -3,9 +3,9 @@
  |                                                                           |
  |  The error handling functions for wm-FPU-emu                              |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
- |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1993,1994,1996                                         |
+ |                  W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
+ |                  E-mail   billm@jacobi.maths.monash.edu.au                |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -19,7 +19,7 @@
 
 #include <linux/signal.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 
 #include "fpu_system.h"
 #include "exception.h"
@@ -46,13 +46,13 @@ void Un_impl(void)
     {
       while ( 1 )
 	{
-	  byte1 = get_fs_byte((unsigned char *) address);
+	  get_user(byte1, (unsigned char *) address);
 	  if ( (byte1 & 0xf8) == 0xd8 ) break;
 	  printk("[%02x]", byte1);
 	  address++;
 	}
       printk("%02x ", byte1);
-      FPU_modrm = get_fs_byte(1 + (unsigned char *) address);
+      get_user(FPU_modrm, 1 + (unsigned char *) address);
       
       if (FPU_modrm >= 0300)
 	printk("%02x (%02x+%d)\n", FPU_modrm, FPU_modrm & 0xf8, FPU_modrm & 7);
@@ -82,10 +82,10 @@ void FPU_illegal(void)
 
 
 
-void emu_printall()
+void emu_printall(void)
 {
   int i;
-  static char *tag_desc[] = { "Valid", "Zero", "ERROR", "ERROR",
+  static const char *tag_desc[] = { "Valid", "Zero", "ERROR", "ERROR",
                               "DeNorm", "Inf", "NaN", "Empty" };
   unsigned char byte1, FPU_modrm;
   unsigned long address = FPU_ORIG_EIP;
@@ -98,7 +98,7 @@ void emu_printall()
 #define MAX_PRINTED_BYTES 20
       for ( i = 0; i < MAX_PRINTED_BYTES; i++ )
 	{
-	  byte1 = get_fs_byte((unsigned char *) address);
+	  get_user(byte1, (unsigned char *) address);
 	  if ( (byte1 & 0xf8) == 0xd8 )
 	    {
 	      printk(" %02x", byte1);
@@ -111,7 +111,7 @@ void emu_printall()
 	printk(" [more..]\n");
       else
 	{
-	  FPU_modrm = get_fs_byte(1 + (unsigned char *) address);
+	  get_user(FPU_modrm, 1 + (unsigned char *) address);
 	  
 	  if (FPU_modrm >= 0300)
 	    printk(" %02x (%02x+%d)\n", FPU_modrm, FPU_modrm & 0xf8, FPU_modrm & 7);
@@ -166,7 +166,8 @@ printk(" CW: ic=%d rc=%ld%ld pc=%ld%ld iem=%d     ef=%d%d%d%d%d%d\n",
   for ( i = 0; i < 8; i++ )
     {
       FPU_REG *r = &st(i);
-      switch (r->tag)
+      char tagi = r->tag;
+      switch (tagi)
 	{
 	case TW_Empty:
 	  continue;
@@ -190,29 +191,20 @@ printk(" CW: ic=%d rc=%ld%ld pc=%ld%ld iem=%d     ef=%d%d%d%d%d%d\n",
 		 r->exp - EXP_BIAS + 1);
 	  break;
 	default:
-	  printk("Whoops! Error in errors.c      ");
+	  printk("Whoops! Error in errors.c: tag%d is %d ", i, tagi);
+	  continue;
 	  break;
 	}
-      printk("%s\n", tag_desc[(int) (unsigned) r->tag]);
+      printk("%s\n", tag_desc[(int) (unsigned) tagi]);
     }
 
-#ifdef OBSOLETE
-  printk("[data] %c .%04lx %04lx %04lx %04lx e%+-6ld ",
-	 FPU_loaded_data.sign ? '-' : '+',
-	 (long)(FPU_loaded_data.sigh >> 16),
-	 (long)(FPU_loaded_data.sigh & 0xFFFF),
-	 (long)(FPU_loaded_data.sigl >> 16),
-	 (long)(FPU_loaded_data.sigl & 0xFFFF),
-	 FPU_loaded_data.exp - EXP_BIAS + 1);
-  printk("%s\n", tag_desc[(int) (unsigned) FPU_loaded_data.tag]);
-#endif OBSOLETE
   RE_ENTRANT_CHECK_ON;
 
 }
 
 static struct {
   int type;
-  char *name;
+  const char *name;
 } exception_names[] = {
   { EX_StackOver, "stack overflow" },
   { EX_StackUnder, "stack underflow" },
@@ -300,7 +292,7 @@ static struct {
 	      0x242  in div_Xsig.S
  */
 
-void exception(int n)
+void FPU_exception(int n)
 {
   int i, int_type;
 
@@ -365,12 +357,8 @@ void exception(int n)
       /*
        * The 80486 generates an interrupt on the next non-control FPU
        * instruction. So we need some means of flagging it.
-       * We use the ES (Error Summary) bit for this, assuming that
-       * this is the way a real FPU does it (until I can check it out),
-       * if not, then some method such as the following kludge might
-       * be needed.
+       * We use the ES (Error Summary) bit for this.
        */
-/*      regs[0].tag |= TW_FPU_Interrupt; */
     }
   RE_ENTRANT_CHECK_ON;
 
@@ -494,7 +482,7 @@ int set_precision_flag(int flags)
     }
   else
     {
-      exception(flags);
+      EXCEPTION(flags);
       return 1;
     }
 }
@@ -506,7 +494,7 @@ asmlinkage void set_precision_flag_up(void)
   if ( control_word & CW_Precision )
     partial_status |= (SW_Precision | SW_C1);   /* The masked response */
   else
-    exception(EX_Precision | SW_C1);
+    EXCEPTION(EX_Precision | SW_C1);
 
 }
 
@@ -520,7 +508,7 @@ asmlinkage void set_precision_flag_down(void)
       partial_status |= SW_Precision;
     }
   else
-    exception(EX_Precision);
+    EXCEPTION(EX_Precision);
 }
 
 
@@ -533,7 +521,7 @@ asmlinkage int denormal_operand(void)
     }
   else
     {
-      exception(EX_Denormal);
+      EXCEPTION(EX_Denormal);
       return 1;
     }
 }
@@ -568,7 +556,7 @@ asmlinkage int arith_overflow(FPU_REG *dest)
       return !(control_word & CW_Precision);
     }
 
-  return !(control_word & CW_Overflow);
+  return 0;
 
 }
 
@@ -599,7 +587,7 @@ asmlinkage int arith_underflow(FPU_REG *dest)
       return !(control_word & CW_Precision);
     }
 
-  return !(control_word & CW_Underflow);
+  return 0;
 
 }
 
diff --git a/arch/i386/math-emu/exception.h b/arch/i386/math-emu/exception.h
index 2e629a30c..b90e5e9c3 100644
--- a/arch/i386/math-emu/exception.h
+++ b/arch/i386/math-emu/exception.h
@@ -10,7 +10,7 @@
 #define _EXCEPTION_H_
 
 
-#ifdef __ASSEMBLER__
+#ifdef __ASSEMBLY__
 #define	Const_(x)	$##x
 #else
 #define	Const_(x)	x
@@ -39,15 +39,15 @@
 #define PRECISION_LOST_DOWN  Const_(EX_Precision)
 
 
-#ifndef __ASSEMBLER__
+#ifndef __ASSEMBLY__
 
 #ifdef DEBUG
 #define	EXCEPTION(x)	{ printk("exception in %s at line %d\n", \
-	__FILE__, __LINE__); exception(x); }
+	__FILE__, __LINE__); FPU_exception(x); }
 #else
-#define	EXCEPTION(x)	exception(x)
+#define	EXCEPTION(x)	FPU_exception(x)
 #endif
 
-#endif __ASSEMBLER__
+#endif __ASSEMBLY__
 
 #endif _EXCEPTION_H_
diff --git a/arch/i386/math-emu/fpu_asm.h b/arch/i386/math-emu/fpu_asm.h
index 8eb60148d..d08fbc874 100644
--- a/arch/i386/math-emu/fpu_asm.h
+++ b/arch/i386/math-emu/fpu_asm.h
@@ -1,17 +1,18 @@
 /*---------------------------------------------------------------------------+
  |  fpu_asm.h                                                                |
  |                                                                           |
- | Copyright (C) 1992    W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1995                                                   |
+ |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  +---------------------------------------------------------------------------*/
 
 #ifndef _FPU_ASM_H_
 #define _FPU_ASM_H_
 
-#include "fpu_emu.h"
+#include <linux/linkage.h>
 
-#define	EXCEPTION	_exception
+#define	EXCEPTION	SYMBOL_NAME(FPU_exception)
 
 
 #define PARAM1	8(%ebp)
diff --git a/arch/i386/math-emu/fpu_emu.h b/arch/i386/math-emu/fpu_emu.h
index 9d2c5dd13..b8385db1f 100644
--- a/arch/i386/math-emu/fpu_emu.h
+++ b/arch/i386/math-emu/fpu_emu.h
@@ -28,7 +28,7 @@
  */
 #define PECULIAR_486
 
-#ifdef __ASSEMBLER__
+#ifdef __ASSEMBLY__
 #include "fpu_asm.h"
 #define	Const(x)	$##x
 #else
@@ -55,9 +55,11 @@
 #define TW_Empty	Const(7)	/* empty */
 
 
-#ifndef __ASSEMBLER__
+#ifndef __ASSEMBLY__
+
+#include <asm/sigcontext.h>	/* for struct _fpstate */
+#include <asm/math_emu.h>
 
-#include <linux/math_emu.h>
 #include <linux/linkage.h>
 
 /*
@@ -95,9 +97,9 @@ extern char emulating;
 
 struct address {
   unsigned int offset;
-  unsigned int selector:16;
-  unsigned int opcode:11;
-  unsigned int empty:5;
+  unsigned short selector;
+  unsigned short opcode:11,
+		 empty:5;
 };
 typedef void (*FUNC)(void);
 typedef struct fpu_reg FPU_REG;
@@ -133,9 +135,9 @@ extern unsigned char const data_sizes_16[32];
 
 
 #define reg_move(x, y) { \
-		 *(short *)&((y)->sign) = *(short *)&((x)->sign); \
-		 *(long *)&((y)->exp) = *(long *)&((x)->exp); \
-		 *(long long *)&((y)->sigl) = *(long long *)&((x)->sigl); }
+		 *(short *)&((y)->sign) = *(const short *)&((x)->sign); \
+		 *(long *)&((y)->exp) = *(const long *)&((x)->exp); \
+		 *(long long *)&((y)->sigl) = *(const long long *)&((x)->sigl); }
 
 #define significand(x) ( ((unsigned long long *)&((x)->sigl))[0] )
 
@@ -166,6 +168,6 @@ asmlinkage void round_reg(FPU_REG *arg, unsigned int extent,
 #include "fpu_proto.h"
 #endif
 
-#endif __ASSEMBLER__
+#endif __ASSEMBLY__
 
 #endif _FPU_EMU_H_
diff --git a/arch/i386/math-emu/fpu_entry.c b/arch/i386/math-emu/fpu_entry.c
index b2777a722..07c32db9a 100644
--- a/arch/i386/math-emu/fpu_entry.c
+++ b/arch/i386/math-emu/fpu_entry.c
@@ -1,11 +1,11 @@
 /*---------------------------------------------------------------------------+
  |  fpu_entry.c                                                              |
  |                                                                           |
- | The entry function for wm-FPU-emu                                         |
+ | The entry functions for wm-FPU-emu                                        |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
- |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1993,1994,1996                                         |
+ |                  W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
+ |                  E-mail   billm@jacobi.maths.monash.edu.au                |
  |                                                                           |
  | See the files "README" and "COPYING" for further copyright and warranty   |
  | information.                                                              |
@@ -20,12 +20,13 @@
  +---------------------------------------------------------------------------*/
 
 /*---------------------------------------------------------------------------+
- | math_emulate() is the sole entry point for wm-FPU-emu                     |
+ | math_emulate(), restore_i387_soft() and save_i387_soft() are the only     |
+ | entry points for wm-FPU-emu.                                              |
  +---------------------------------------------------------------------------*/
 
 #include <linux/signal.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 
 #include "fpu_system.h"
 #include "fpu_emu.h"
@@ -159,8 +160,9 @@ asmlinkage void math_emulate(long arg)
 	{
 	  /* Make sure that the registers are compatible
 	     with the assumptions of the emulator. */
-	  regs[i].exp = 0;
-	  regs[i].sigh = 0x80000000;
+	  if ( !((regs[i].exp == EXP_UNDER) && (regs[i].sigh == 0)
+		 && (regs[i].sigl == 0)) )
+	    regs[i].sigh |= 0x80000000;
 	}
       finit();
       current->used_math = 1;
@@ -262,7 +264,7 @@ do_another_FPU_instruction:
 
   RE_ENTRANT_CHECK_OFF;
   FPU_code_verify_area(1);
-  FPU_modrm = get_fs_byte((unsigned char *) FPU_EIP);
+  get_user(FPU_modrm, (unsigned char *) FPU_EIP);
   RE_ENTRANT_CHECK_ON;
   FPU_EIP++;
 
@@ -283,24 +285,8 @@ do_another_FPU_instruction:
 	  /*
 	   *  We need to simulate the action of the kernel to FPU
 	   *  interrupts here.
-	   *  Currently, the "real FPU" part of the kernel (0.99.10)
-	   *  clears the exception flags, sets the registers to empty,
-	   *  and passes information back to the interrupted process
-	   *  via the cs selector and operand selector, so we do the same.
 	   */
 	do_the_FPU_interrupt:
-	  instruction_address.selector = status_word();
-      	  operand_address.selector = tag_word();
-	  partial_status = 0;
-	  top = 0;
-	  {
-	    int r;
-	    for (r = 0; r < 8; r++)
-	      {
-		regs[r].tag = TW_Empty;
-	      }
-	  }
-
 	  FPU_EIP = FPU_ORIG_EIP;	/* Point to current FPU instruction. */
 
 	  RE_ENTRANT_CHECK_OFF;
@@ -605,7 +591,7 @@ static int valid_prefix(unsigned char *Byte, unsigned char **fpu_eip,
 
   RE_ENTRANT_CHECK_OFF;
   FPU_code_verify_area(1);
-  byte = get_fs_byte(ip);
+  get_user(byte, ip);
   RE_ENTRANT_CHECK_ON;
 
   while ( 1 )
@@ -651,7 +637,7 @@ static int valid_prefix(unsigned char *Byte, unsigned char **fpu_eip,
 	  ip++;
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_code_verify_area(1);
-	  byte = get_fs_byte(ip);
+	  get_user(byte, ip);
 	  RE_ENTRANT_CHECK_ON;
 	  break;
 	case FWAIT_OPCODE:
@@ -688,3 +674,22 @@ void math_abort(struct info * info, unsigned int signal)
       printk("ERROR: wm-FPU-emu math_abort failed!\n");
 #endif PARANOID
 }
+
+
+
+void restore_i387_soft(struct _fpstate *buf)
+{
+  fpu_addr_modes addr_modes = {{ 0, 0, PREFIX_DEFAULT }, 0};
+
+  frstor(addr_modes, (char *)buf);
+}
+
+
+struct _fpstate * save_i387_soft(struct _fpstate * buf)
+{
+  fpu_addr_modes addr_modes = {{ 0, 0, PREFIX_DEFAULT }, 0};
+
+  fsave(addr_modes, (char *)buf);
+
+  return buf;
+}
diff --git a/arch/i386/math-emu/fpu_proto.h b/arch/i386/math-emu/fpu_proto.h
index b4392fe57..d0e58ed0e 100644
--- a/arch/i386/math-emu/fpu_proto.h
+++ b/arch/i386/math-emu/fpu_proto.h
@@ -7,7 +7,7 @@ extern void stack_underflow(void);
 extern void stack_underflow_i(int i);
 extern void stack_underflow_pop(int i);
 extern int set_precision_flag(int flags);
-asmlinkage void exception(int n);
+asmlinkage void FPU_exception(int n);
 asmlinkage int real_2op_NaN(FPU_REG const *a, FPU_REG const *b, FPU_REG *dest);
 asmlinkage int arith_invalid(FPU_REG *dest);
 asmlinkage int divide_by_zero(int sign, FPU_REG *dest);
diff --git a/arch/i386/math-emu/get_address.c b/arch/i386/math-emu/get_address.c
index 6f3270ae3..0749cdc70 100644
--- a/arch/i386/math-emu/get_address.c
+++ b/arch/i386/math-emu/get_address.c
@@ -21,7 +21,7 @@
 #include <linux/stddef.h>
 #include <linux/head.h>
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 
 #include "fpu_system.h"
 #include "exception.h"
@@ -56,12 +56,16 @@ static int reg_offset_vm86[] = {
 #define VM86_REG_(x) (*(unsigned short *) \
 		      (reg_offset_vm86[((unsigned)x)]+(char *) FPU_info))
 
+/* These are dummy, fs and gs are not saved on the stack. */
+#define ___FS ___ds
+#define ___GS ___ds
+
 static int reg_offset_pm[] = {
 	offsetof(struct info,___cs),
 	offsetof(struct info,___ds),
 	offsetof(struct info,___es),
-	offsetof(struct info,___fs),
-	offsetof(struct info,___gs),
+	offsetof(struct info,___FS),
+	offsetof(struct info,___GS),
 	offsetof(struct info,___ss),
 	offsetof(struct info,___ds)
       };
@@ -78,7 +82,7 @@ static int sib(int mod, unsigned long *fpu_eip)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_code_verify_area(1);
-  base = get_fs_byte((char *) (*fpu_eip));   /* The SIB byte */
+  get_user(base, (unsigned char *) (*fpu_eip));   /* The SIB byte */
   RE_ENTRANT_CHECK_ON;
   (*fpu_eip)++;
   ss = base >> 6;
@@ -105,18 +109,22 @@ static int sib(int mod, unsigned long *fpu_eip)
   if (mod == 1)
     {
       /* 8 bit signed displacement */
+      long displacement;
       RE_ENTRANT_CHECK_OFF;
       FPU_code_verify_area(1);
-      offset += (signed char) get_fs_byte((char *) (*fpu_eip));
+      get_user(displacement, (signed char *) (*fpu_eip));
+      offset += displacement;
       RE_ENTRANT_CHECK_ON;
       (*fpu_eip)++;
     }
   else if (mod == 2 || base == 5) /* The second condition also has mod==0 */
     {
       /* 32 bit displacement */
+      long displacement;
       RE_ENTRANT_CHECK_OFF;
       FPU_code_verify_area(4);
-      offset += (signed) get_fs_long((unsigned long *) (*fpu_eip));
+      get_user(displacement, (signed long *) (*fpu_eip));
+      offset += displacement;
       RE_ENTRANT_CHECK_ON;
       (*fpu_eip) += 4;
     }
@@ -149,7 +157,9 @@ static long pm_address(unsigned char FPU_modrm, unsigned char segment,
   unsigned long base_address, limit, address, seg_top;
 
   segment--;
+
 #ifdef PARANOID
+  /* segment is unsigned, so this also detects if segment was 0: */
   if ( segment > PREFIX_SS_ )
     {
       EXCEPTION(EX_INTERNAL|0x132);
@@ -157,7 +167,19 @@ static long pm_address(unsigned char FPU_modrm, unsigned char segment,
     }
 #endif PARANOID
 
-  *selector = PM_REG_(segment);
+  switch ( segment )
+    {
+      /* fs and gs aren't used by the kernel, so they still have their
+	 user-space values. */
+    case PREFIX_FS_-1:
+      __asm__("mov %%fs,%0":"=r" (*selector));
+      break;
+    case PREFIX_GS_-1:
+      __asm__("mov %%gs,%0":"=r" (*selector));
+      break;
+    default:
+      *selector = PM_REG_(segment);
+    }
 
   descriptor = LDT_DESCRIPTOR(PM_REG_(segment));
   base_address = SEG_BASE_ADDR(descriptor);
@@ -248,7 +270,7 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
 	      /* Special case: disp32 */
 	      RE_ENTRANT_CHECK_OFF;
 	      FPU_code_verify_area(4);
-	      address = get_fs_long((unsigned long *) (*fpu_eip));
+	      get_user(address, (unsigned long *) (*fpu_eip));
 	      (*fpu_eip) += 4;
 	      RE_ENTRANT_CHECK_ON;
 	      addr->offset = address;
@@ -265,7 +287,7 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
 	  /* 8 bit signed displacement */
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_code_verify_area(1);
-	  address = (signed char) get_fs_byte((char *) (*fpu_eip));
+	  get_user(address, (signed char *) (*fpu_eip));
 	  RE_ENTRANT_CHECK_ON;
 	  (*fpu_eip)++;
 	  break;
@@ -273,7 +295,7 @@ void *get_address(unsigned char FPU_modrm, unsigned long *fpu_eip,
 	  /* 32 bit displacement */
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_code_verify_area(4);
-	  address = (signed) get_fs_long((unsigned long *) (*fpu_eip));
+	  get_user(address, (long *) (*fpu_eip));
 	  (*fpu_eip) += 4;
 	  RE_ENTRANT_CHECK_ON;
 	  break;
@@ -336,7 +358,7 @@ void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
 	  /* Special case: disp16 */
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_code_verify_area(2);
-	  address = (unsigned short)get_fs_word((unsigned short *) (*fpu_eip));
+	  get_user(address, (unsigned short *) (*fpu_eip));
 	  (*fpu_eip) += 2;
 	  RE_ENTRANT_CHECK_ON;
 	  goto add_segment;
@@ -346,7 +368,7 @@ void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
       /* 8 bit signed displacement */
       RE_ENTRANT_CHECK_OFF;
       FPU_code_verify_area(1);
-      address = (signed char) get_fs_byte((signed char *) (*fpu_eip));
+      get_user(address, (signed char *) (*fpu_eip));
       RE_ENTRANT_CHECK_ON;
       (*fpu_eip)++;
       break;
@@ -354,7 +376,7 @@ void *get_address_16(unsigned char FPU_modrm, unsigned long *fpu_eip,
       /* 16 bit displacement */
       RE_ENTRANT_CHECK_OFF;
       FPU_code_verify_area(2);
-      address = (unsigned) get_fs_word((unsigned short *) (*fpu_eip));
+      get_user(address, (unsigned short *) (*fpu_eip));
       (*fpu_eip) += 2;
       RE_ENTRANT_CHECK_ON;
       break;
diff --git a/arch/i386/math-emu/load_store.c b/arch/i386/math-emu/load_store.c
index 6f0e167d6..cc288a9ab 100644
--- a/arch/i386/math-emu/load_store.c
+++ b/arch/i386/math-emu/load_store.c
@@ -18,7 +18,7 @@
  |    other processes using the emulator while swapping is in progress.      |
  +---------------------------------------------------------------------------*/
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 
 #include "fpu_system.h"
 #include "exception.h"
@@ -201,7 +201,7 @@ int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
     case 024:     /* fldcw */
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_READ, data_address, 2);
-      control_word = get_fs_word((unsigned short *) data_address);
+      get_user(control_word, (unsigned short *) data_address);
       RE_ENTRANT_CHECK_ON;
       if ( partial_status & ~control_word & CW_Exceptions )
 	partial_status |= (SW_Summary | SW_Backward);
@@ -234,7 +234,7 @@ int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
     case 034:      /* fstcw m16int */
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_WRITE,data_address,2);
-      put_fs_word(control_word, (short *) data_address);
+      put_user(control_word, (unsigned short *) data_address);
       RE_ENTRANT_CHECK_ON;
       return 1;
     case 035:      /* fstp m80real */
@@ -246,7 +246,7 @@ int load_store_instr(unsigned char type, fpu_addr_modes addr_modes,
     case 036:      /* fstsw m2byte */
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_WRITE,data_address,2);
-      put_fs_word(status_word(),(short *) data_address);
+      put_user(status_word(),(unsigned short *) data_address);
       RE_ENTRANT_CHECK_ON;
       return 1;
     case 037:      /* fistp m64int */
diff --git a/arch/i386/math-emu/mul_Xsig.S b/arch/i386/math-emu/mul_Xsig.S
index 1d88d4466..717785a53 100644
--- a/arch/i386/math-emu/mul_Xsig.S
+++ b/arch/i386/math-emu/mul_Xsig.S
@@ -3,9 +3,9 @@
  |                                                                           |
  | Multiply a 12 byte fixed point number by another fixed point number.      |
  |                                                                           |
- | Copyright (C) 1992,1994                                                   |
+ | Copyright (C) 1992,1994,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void mul32_Xsig(Xsig *x, unsigned b)                                    |
@@ -21,12 +21,10 @@
 	.file	"mul_Xsig.S"
 
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 .text
-	.align 2,144
-.globl _mul32_Xsig
-_mul32_Xsig:
+ENTRY(mul32_Xsig)
 	pushl %ebp
 	movl %esp,%ebp
 	subl $16,%esp
@@ -66,9 +64,7 @@ _mul32_Xsig:
 	ret
 
 
-	.align 2,144
-.globl _mul64_Xsig
-_mul64_Xsig:
+ENTRY(mul64_Xsig)
 	pushl %ebp
 	movl %esp,%ebp
 	subl $16,%esp
@@ -121,9 +117,7 @@ _mul64_Xsig:
 
 
 
-	.align 2,144
-.globl _mul_Xsig_Xsig
-_mul_Xsig_Xsig:
+ENTRY(mul_Xsig_Xsig)
 	pushl %ebp
 	movl %esp,%ebp
 	subl $16,%esp
diff --git a/arch/i386/math-emu/polynom_Xsig.S b/arch/i386/math-emu/polynom_Xsig.S
index 585221f96..17315c89f 100644
--- a/arch/i386/math-emu/polynom_Xsig.S
+++ b/arch/i386/math-emu/polynom_Xsig.S
@@ -3,9 +3,9 @@
  |                                                                           |
  | Fixed point arithmetic polynomial evaluation.                             |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
+ | Copyright (C) 1992,1993,1994,1995                                         |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void polynomial_Xsig(Xsig *accum, unsigned long long x,                 |
@@ -23,7 +23,7 @@
  +---------------------------------------------------------------------------*/
 	.file	"polynomial_Xsig.S"
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 
 #define	TERM_SIZE	$8
@@ -36,9 +36,7 @@
 #define OVERFLOWED      -16(%ebp)	/* addition overflow flag */
 
 .text
-	.align 2,144
-.globl _polynomial_Xsig
-_polynomial_Xsig:
+ENTRY(polynomial_Xsig)
 	pushl	%ebp
 	movl	%esp,%ebp
 	subl	$32,%esp
diff --git a/arch/i386/math-emu/reg_constant.c b/arch/i386/math-emu/reg_constant.c
index c1981ce24..1b2458eea 100644
--- a/arch/i386/math-emu/reg_constant.c
+++ b/arch/i386/math-emu/reg_constant.c
@@ -3,9 +3,9 @@
  |                                                                           |
  | All of the constant FPU_REGs                                              |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
- |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1993,1994,1996                                         |
+ |                     W. Metzenthen, 22 Parker St, Ormond, Vic 3163,        |
+ |                     Australia.  E-mail   billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -14,6 +14,7 @@
 #include "fpu_emu.h"
 #include "status_w.h"
 #include "reg_constant.h"
+#include "control_w.h"
 
 
 FPU_REG const CONST_1    = { SIGN_POS, TW_Valid, EXP_BIAS,
@@ -56,7 +57,7 @@ FPU_REG const CONST_INF  = { SIGN_POS, TW_Infinity, EXP_OVER, 0x00000000, 0x8000
 
 
 
-static void fld_const(FPU_REG const *c)
+static void fld_const(FPU_REG const *c, int adj)
 {
   FPU_REG *st_new_ptr;
 
@@ -67,50 +68,58 @@ static void fld_const(FPU_REG const *c)
     }
   push();
   reg_move(c, st_new_ptr);
+  st_new_ptr->sigl += adj;  /* For all our fldxxx constants, we don't need to
+			       borrow or carry. */
   clear_C1();
 }
 
+/* A fast way to find out whether x is one of RC_DOWN or RC_CHOP
+   (and not one of RC_RND or RC_UP).
+   */
+#define DOWN_OR_CHOP(x)  (x & RC_DOWN)
 
-static void fld1(void)
+static void fld1(int rc)
 {
-  fld_const(&CONST_1);
+  fld_const(&CONST_1, 0);
 }
 
-static void fldl2t(void)
+static void fldl2t(int rc)
 {
-  fld_const(&CONST_L2T);
+  fld_const(&CONST_L2T, (rc == RC_UP) ? 1 : 0);
 }
 
-static void fldl2e(void)
+static void fldl2e(int rc)
 {
-  fld_const(&CONST_L2E);
+  fld_const(&CONST_L2E, DOWN_OR_CHOP(rc) ? -1 : 0);
 }
 
-static void fldpi(void)
+static void fldpi(int rc)
 {
-  fld_const(&CONST_PI);
+  fld_const(&CONST_PI, DOWN_OR_CHOP(rc) ? -1 : 0);
 }
 
-static void fldlg2(void)
+static void fldlg2(int rc)
 {
-  fld_const(&CONST_LG2);
+  fld_const(&CONST_LG2, DOWN_OR_CHOP(rc) ? -1 : 0);
 }
 
-static void fldln2(void)
+static void fldln2(int rc)
 {
-  fld_const(&CONST_LN2);
+  fld_const(&CONST_LN2, DOWN_OR_CHOP(rc) ? -1 : 0);
 }
 
-static void fldz(void)
+static void fldz(int rc)
 {
-  fld_const(&CONST_Z);
+  fld_const(&CONST_Z, 0);
 }
 
-static FUNC constants_table[] = {
-  fld1, fldl2t, fldl2e, fldpi, fldlg2, fldln2, fldz, FPU_illegal
+typedef void (*FUNC_RC)(int);
+
+static FUNC_RC constants_table[] = {
+  fld1, fldl2t, fldl2e, fldpi, fldlg2, fldln2, fldz, (FUNC_RC)FPU_illegal
 };
 
 void fconst(void)
 {
-  (constants_table[FPU_rm])();
+  (constants_table[FPU_rm])(control_word & CW_RC);
 }
diff --git a/arch/i386/math-emu/reg_div.S b/arch/i386/math-emu/reg_div.S
index 2fbc5f7c4..24d44ac6c 100644
--- a/arch/i386/math-emu/reg_div.S
+++ b/arch/i386/math-emu/reg_div.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Divide one FPU_REG by another and put the result in a destination FPU_REG.|
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
+ | Copyright (C) 1992,1993,1994,1995                                         |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void reg_div(FPU_REG *a, FPU_REG *b, FPU_REG *dest,                     |
@@ -15,14 +15,11 @@
  +---------------------------------------------------------------------------*/
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 
 .text
-	.align 2
-
-.globl	_reg_div
-_reg_div:
+ENTRY(reg_div)
 	pushl	%ebp
 	movl	%esp,%ebp
 #ifndef NON_REENTRANT_FPU
@@ -47,7 +44,7 @@ _reg_div:
 	cmpl	EXP_UNDER,EXP(%esi)
 	jg	xL_arg1_not_denormal
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
@@ -55,7 +52,7 @@ xL_arg1_not_denormal:
 	cmpl	EXP_UNDER,EXP(%ebx)
 	jg	xL_arg2_not_denormal
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
@@ -75,7 +72,7 @@ xL_arg2_not_denormal:
 	addl	EXP_BIAS,%edx
 	movl	%edx,EXP(%edi)
 
-	jmp	_divide_kernel
+	jmp	SYMBOL_NAME(divide_kernel)
 
 
 /*-----------------------------------------------------------------------*/
@@ -92,14 +89,14 @@ L_arg2_NaN:
 	pushl	%edi			/* Destination */
 	pushl	%esi
 	pushl	%ebx			/* Ordering is important here */
-	call	_real_2op_NaN
+	call	SYMBOL_NAME(real_2op_NaN)
 	jmp	LDiv_exit
 
 /* Invalid operations */
 L_zero_zero:
 L_inf_inf:
 	pushl	%edi			/* Destination */
-	call	_arith_invalid		/* 0/0 or Infinity/Infinity */
+	call	SYMBOL_NAME(arith_invalid) /* 0/0 or Infinity/Infinity */
 	jmp	LDiv_exit
 
 L_no_NaN_arg:
@@ -126,7 +123,7 @@ L_inf_valid:
 	cmpl	EXP_UNDER,EXP(%ebx)
 	jg	L_copy_arg1		/* Answer is Inf */
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 #endif DENORM_OPERAND
@@ -151,7 +148,7 @@ L_arg1_not_inf:
 	movb	SIGN(%esi),%al
 	xorb	SIGN(%ebx),%al
 	pushl	%eax			/* lower 8 bits have the sign */
-	call	_divide_by_zero
+	call	SYMBOL_NAME(divide_by_zero)
 	jmp	LDiv_exit
 
 L_arg2_not_zero:
@@ -165,7 +162,7 @@ L_arg2_not_zero:
 	cmpl	EXP_UNDER,EXP(%esi)
 	jg	L_return_zero		/* Answer is zero */
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 #endif DENORM_OPERAND
@@ -185,7 +182,7 @@ L_arg2_not_inf:
 	cmpl	EXP_UNDER,EXP(%ebx)
 	jg	L_copy_arg1		/* Answer is zero */
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 #endif DENORM_OPERAND
@@ -241,11 +238,11 @@ L_unknown_tags:
 	call	EXCEPTION
 
 	/* Generate a NaN for unknown tags */
-	movl	_CONST_QNaN,%eax
+	movl	SYMBOL_NAME(CONST_QNaN),%eax
 	movl	%eax,(%edi)
-	movl	_CONST_QNaN+4,%eax
+	movl	SYMBOL_NAME(CONST_QNaN)+4,%eax
 	movl	%eax,SIGL(%edi)
-	movl	_CONST_QNaN+8,%eax
+	movl	SYMBOL_NAME(CONST_QNaN)+8,%eax
 	movl	%eax,SIGH(%edi)
 	jmp	LDiv_exit		/* %eax is nz */
 #endif PARANOID
diff --git a/arch/i386/math-emu/reg_ld_str.c b/arch/i386/math-emu/reg_ld_str.c
index efec9e010..3e258a0c6 100644
--- a/arch/i386/math-emu/reg_ld_str.c
+++ b/arch/i386/math-emu/reg_ld_str.c
@@ -3,9 +3,9 @@
  |                                                                           |
  | All of the functions which transfer data between user memory and FPU_REGs.|
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
- |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1993,1994,1996                                         |
+ |                  W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
+ |                  E-mail   billm@jacobi.maths.monash.edu.au                |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -17,7 +17,7 @@
  |    other processes using the emulator while swapping is in progress.      |
  +---------------------------------------------------------------------------*/
 
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 
 #include "fpu_system.h"
 #include "exception.h"
@@ -48,9 +48,9 @@ int reg_load_extended(long double *s, FPU_REG *loaded_data)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_READ, s, 10);
-  sigl = get_fs_long((unsigned long *) s);
-  sigh = get_fs_long(1 + (unsigned long *) s);
-  exp = get_fs_word(4 + (unsigned short *) s);
+  get_user(sigl, (unsigned long *) s);
+  get_user(sigh, 1 + (unsigned long *) s);
+  get_user(exp, 4 + (unsigned short *) s);
   RE_ENTRANT_CHECK_ON;
 
   loaded_data->tag = TW_Valid;   /* Default */
@@ -144,8 +144,8 @@ int reg_load_double(double *dfloat, FPU_REG *loaded_data)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_READ, dfloat, 8);
-  m64 = get_fs_long(1 + (unsigned long *) dfloat);
-  l64 = get_fs_long((unsigned long *) dfloat);
+  get_user(m64, 1 + (unsigned long *) dfloat);
+  get_user(l64, (unsigned long *) dfloat);
   RE_ENTRANT_CHECK_ON;
 
   if (m64 & 0x80000000)
@@ -221,7 +221,7 @@ int reg_load_single(float *single, FPU_REG *loaded_data)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_READ, single, 4);
-  m32 = get_fs_long((unsigned long *) single);
+  get_user(m32, (unsigned long *) single);
   RE_ENTRANT_CHECK_ON;
 
   if (m32 & 0x80000000)
@@ -289,8 +289,7 @@ void reg_load_int64(long long *_s, FPU_REG *loaded_data)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_READ, _s, 8);
-  ((unsigned long *)&s)[0] = get_fs_long((unsigned long *) _s);
-  ((unsigned long *)&s)[1] = get_fs_long(1 + (unsigned long *) _s);
+  copy_from_user(&s,_s,8);
   RE_ENTRANT_CHECK_ON;
 
   if (s == 0)
@@ -320,7 +319,7 @@ void reg_load_int32(long *_s, FPU_REG *loaded_data)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_READ, _s, 4);
-  s = (long)get_fs_long((unsigned long *) _s);
+  get_user(s, _s);
   RE_ENTRANT_CHECK_ON;
 
   if (s == 0)
@@ -351,7 +350,7 @@ void reg_load_int16(short *_s, FPU_REG *loaded_data)
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_READ, _s, 2);
   /* Cast as short to get the sign extended. */
-  s = (short)get_fs_word((unsigned short *) _s);
+  get_user(s, _s);
   RE_ENTRANT_CHECK_ON;
 
   if (s == 0)
@@ -389,7 +388,7 @@ void reg_load_bcd(char *s, FPU_REG *loaded_data)
     {
       l *= 10;
       RE_ENTRANT_CHECK_OFF;
-      bcd = (unsigned char)get_fs_byte((unsigned char *) s+pos);
+      get_user(bcd, (unsigned char *) s+pos);
       RE_ENTRANT_CHECK_ON;
       l += bcd >> 4;
       l *= 10;
@@ -397,9 +396,11 @@ void reg_load_bcd(char *s, FPU_REG *loaded_data)
     }
  
   RE_ENTRANT_CHECK_OFF;
-  loaded_data->sign =
-    ((unsigned char)get_fs_byte((unsigned char *) s+9)) & 0x80 ?
-      SIGN_NEG : SIGN_POS;
+  {
+    unsigned char sign;
+    get_user(sign, (unsigned char *) s+9);
+    loaded_data->sign = (sign & 0x80) ? SIGN_NEG : SIGN_POS;
+  }
   RE_ENTRANT_CHECK_ON;
 
   if (l == 0)
@@ -445,9 +446,9 @@ int reg_store_extended(long double *d, FPU_REG *st0_ptr)
       /* Put out the QNaN indefinite */
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_WRITE,d,10);
-      put_fs_long(0, (unsigned long *) d);
-      put_fs_long(0xc0000000, 1 + (unsigned long *) d);
-      put_fs_word(0xffff, 4 + (short *) d);
+      put_user(0, (unsigned long *) d);
+      put_user(0xc0000000, 1 + (unsigned long *) d);
+      put_user(0xffff, 4 + (short *) d);
       RE_ENTRANT_CHECK_ON;
       return 1;
     }
@@ -466,6 +467,7 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 
   if (st0_tag == TW_Valid)
     {
+      int precision_loss;
       int exp;
       FPU_REG tmp;
 
@@ -474,8 +476,6 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 
       if ( exp < DOUBLE_Emin )     /* It may be a denormal */
 	{
-	  int precision_loss;
-
 	  /* A denormal will always underflow. */
 #ifndef PECULIAR_486
 	  /* An 80486 is supposed to be able to generate
@@ -518,6 +518,7 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 	{
 	  if ( tmp.sigl & 0x000007ff )
 	    {
+	      precision_loss = 1;
 	      switch (control_word & CW_RC)
 		{
 		case RC_RND:
@@ -541,8 +542,6 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 	  
 	      if ( increment )
 		{
-		  set_precision_flag_up();
-
 		  if ( tmp.sigl >= 0xfffff800 )
 		    {
 		      /* the sigl part overflows */
@@ -566,9 +565,9 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 		      tmp.sigl += 0x00000800;
 		    }
 		}
-	      else
-		set_precision_flag_down();
 	    }
+	  else
+	    precision_loss = 0;
 	  
 	  l[0] = (tmp.sigl >> 11) | (tmp.sigh << 21);
 	  l[1] = ((tmp.sigh >> 11) & 0xfffff);
@@ -590,6 +589,13 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 	    }
 	  else
 	    {
+	      if ( precision_loss )
+		{
+		  if ( increment )
+		    set_precision_flag_up();
+		  else
+		    set_precision_flag_down();
+		}
 	      /* Add the exponent */
 	      l[1] |= (((exp+DOUBLE_Ebias) & 0x7ff) << 20);
 	    }
@@ -631,8 +637,8 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 	  /* Put out the QNaN indefinite */
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_verify_area(VERIFY_WRITE,(void *)dfloat,8);
-	  put_fs_long(0, (unsigned long *) dfloat);
-	  put_fs_long(0xfff80000, 1 + (unsigned long *) dfloat);
+	  put_user(0, (unsigned long *) dfloat);
+	  put_user(0xfff80000, 1 + (unsigned long *) dfloat);
 	  RE_ENTRANT_CHECK_ON;
 	  return 1;
 	}
@@ -644,8 +650,8 @@ int reg_store_double(double *dfloat, FPU_REG *st0_ptr)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_WRITE,(void *)dfloat,8);
-  put_fs_long(l[0], (unsigned long *)dfloat);
-  put_fs_long(l[1], 1 + (unsigned long *)dfloat);
+  put_user(l[0], (unsigned long *)dfloat);
+  put_user(l[1], 1 + (unsigned long *)dfloat);
   RE_ENTRANT_CHECK_ON;
 
   return 1;
@@ -661,6 +667,7 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 
   if (st0_tag == TW_Valid)
     {
+      int precision_loss;
       int exp;
       FPU_REG tmp;
 
@@ -669,8 +676,6 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 
       if ( exp < SINGLE_Emin )
 	{
-	  int precision_loss;
-
 	  /* A denormal will always underflow. */
 #ifndef PECULIAR_486
 	  /* An 80486 is supposed to be able to generate
@@ -715,6 +720,7 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 	      unsigned long sigh = tmp.sigh;
 	      unsigned long sigl = tmp.sigl;
 	      
+	      precision_loss = 1;
 	      switch (control_word & CW_RC)
 		{
 		case RC_RND:
@@ -740,8 +746,6 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 	  
 	      if (increment)
 		{
-		  set_precision_flag_up();
-
 		  if ( sigh >= 0xffffff00 )
 		    {
 		      /* The sigh part overflows */
@@ -758,10 +762,11 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 		}
 	      else
 		{
-		  set_precision_flag_down();
 		  tmp.sigh &= 0xffffff00;  /* Finish the truncation */
 		}
 	    }
+	  else
+	    precision_loss = 0;
 
 	  templ = (tmp.sigh >> 8) & 0x007fffff;
 
@@ -780,7 +785,17 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 	      templ = 0x7f800000;
 	    }
 	  else
-	    templ |= ((exp+SINGLE_Ebias) & 0xff) << 23;
+	    {
+	      if ( precision_loss )
+		{
+		  if ( increment )
+		    set_precision_flag_up();
+		  else
+		    set_precision_flag_down();
+		}
+	      /* Add the exponent */
+	      templ |= ((exp+SINGLE_Ebias) & 0xff) << 23;
+	    }
 	}
     }
   else if (st0_tag == TW_Zero)
@@ -815,7 +830,7 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 	  /* Put out the QNaN indefinite */
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_verify_area(VERIFY_WRITE,(void *)single,4);
-	  put_fs_long(0xffc00000, (unsigned long *) single);
+	  put_user(0xffc00000, (unsigned long *) single);
 	  RE_ENTRANT_CHECK_ON;
 	  return 1;
 	}
@@ -834,7 +849,7 @@ int reg_store_single(float *single, FPU_REG *st0_ptr)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_WRITE,(void *)single,4);
-  put_fs_long(templ,(unsigned long *) single);
+  put_user(templ,(unsigned long *) single);
   RE_ENTRANT_CHECK_ON;
 
   return 1;
@@ -892,8 +907,7 @@ int reg_store_int64(long long *d, FPU_REG *st0_ptr)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_WRITE,(void *)d,8);
-  put_fs_long(((long *)&tll)[0],(unsigned long *) d);
-  put_fs_long(((long *)&tll)[1],1 + (unsigned long *) d);
+  copy_to_user(d, &tll, 8);
   RE_ENTRANT_CHECK_ON;
 
   return 1;
@@ -947,7 +961,7 @@ int reg_store_int32(long *d, FPU_REG *st0_ptr)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_WRITE,d,4);
-  put_fs_long(t.sigl, (unsigned long *) d);
+  put_user(t.sigl, (unsigned long *) d);
   RE_ENTRANT_CHECK_ON;
 
   return 1;
@@ -1001,7 +1015,7 @@ int reg_store_int16(short *d, FPU_REG *st0_ptr)
 
   RE_ENTRANT_CHECK_OFF;
   FPU_verify_area(VERIFY_WRITE,d,2);
-  put_fs_word((short)t.sigl,(short *) d);
+  put_user((short)t.sigl,(short *) d);
   RE_ENTRANT_CHECK_ON;
 
   return 1;
@@ -1042,10 +1056,10 @@ int reg_store_bcd(char *d, FPU_REG *st0_ptr)
 	  RE_ENTRANT_CHECK_OFF;
 	  FPU_verify_area(VERIFY_WRITE,d,10);
 	  for ( i = 0; i < 7; i++)
-	    put_fs_byte(0, (unsigned char *) d+i); /* These bytes "undefined" */
-	  put_fs_byte(0xc0, (unsigned char *) d+7); /* This byte "undefined" */
-	  put_fs_byte(0xff, (unsigned char *) d+8);
-	  put_fs_byte(0xff, (unsigned char *) d+9);
+	    put_user(0, (unsigned char *) d+i); /* These bytes "undefined" */
+	  put_user(0xc0, (unsigned char *) d+7); /* This byte "undefined" */
+	  put_user(0xff, (unsigned char *) d+8);
+	  put_user(0xff, (unsigned char *) d+9);
 	  RE_ENTRANT_CHECK_ON;
 	  return 1;
 	}
@@ -1066,11 +1080,11 @@ int reg_store_bcd(char *d, FPU_REG *st0_ptr)
       b = div_small(&ll, 10);
       b |= (div_small(&ll, 10)) << 4;
       RE_ENTRANT_CHECK_OFF;
-      put_fs_byte(b,(unsigned char *) d+i);
+      put_user(b,(unsigned char *) d+i);
       RE_ENTRANT_CHECK_ON;
     }
   RE_ENTRANT_CHECK_OFF;
-  put_fs_byte(sign,(unsigned char *) d+9);
+  put_user(sign,(unsigned char *) d+9);
   RE_ENTRANT_CHECK_ON;
 
   return 1;
@@ -1158,13 +1172,13 @@ char *fldenv(fpu_addr_modes addr_modes, char *s)
     {
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_READ, s, 0x0e);
-      control_word = get_fs_word((unsigned short *) s);
-      partial_status = get_fs_word((unsigned short *) (s+2));
-      tag_word = get_fs_word((unsigned short *) (s+4));
-      instruction_address.offset = get_fs_word((unsigned short *) (s+6));
-      instruction_address.selector = get_fs_word((unsigned short *) (s+8));
-      operand_address.offset = get_fs_word((unsigned short *) (s+0x0a));
-      operand_address.selector = get_fs_word((unsigned short *) (s+0x0c));
+      get_user(control_word, (unsigned short *) s);
+      get_user(partial_status, (unsigned short *) (s+2));
+      get_user(tag_word, (unsigned short *) (s+4));
+      get_user(instruction_address.offset, (unsigned short *) (s+6));
+      get_user(instruction_address.selector, (unsigned short *) (s+8));
+      get_user(operand_address.offset, (unsigned short *) (s+0x0a));
+      get_user(operand_address.selector, (unsigned short *) (s+0x0c));
       RE_ENTRANT_CHECK_ON;
       s += 0x0e;
       if ( addr_modes.default_mode == VM86 )
@@ -1178,14 +1192,14 @@ char *fldenv(fpu_addr_modes addr_modes, char *s)
     {
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_READ, s, 0x1c);
-      control_word = get_fs_word((unsigned short *) s);
-      partial_status = get_fs_word((unsigned short *) (s+4));
-      tag_word = get_fs_word((unsigned short *) (s+8));
-      instruction_address.offset = get_fs_long((unsigned long *) (s+0x0c));
-      instruction_address.selector = get_fs_word((unsigned short *) (s+0x10));
-      instruction_address.opcode = get_fs_word((unsigned short *) (s+0x12));
-      operand_address.offset = get_fs_long((unsigned long *) (s+0x14));
-      operand_address.selector = get_fs_long((unsigned long *) (s+0x18));
+      get_user(control_word, (unsigned short *) s);
+      get_user(partial_status, (unsigned short *) (s+4));
+      get_user(tag_word, (unsigned short *) (s+8));
+      get_user(instruction_address.offset, (unsigned long *) (s+0x0c));
+      get_user(instruction_address.selector, (unsigned short *) (s+0x10));
+      get_user(instruction_address.opcode, (unsigned short *) (s+0x12));
+      get_user(operand_address.offset, (unsigned long *) (s+0x14));
+      get_user(operand_address.selector, (unsigned long *) (s+0x18));
       RE_ENTRANT_CHECK_ON;
       s += 0x1c;
     }
@@ -1296,25 +1310,25 @@ char *fstenv(fpu_addr_modes addr_modes, char *d)
       RE_ENTRANT_CHECK_OFF;
       FPU_verify_area(VERIFY_WRITE,d,14);
 #ifdef PECULIAR_486
-      put_fs_long(control_word & ~0xe080, (unsigned short *) d);
+      put_user(control_word & ~0xe080, (unsigned long *) d);
 #else
-      put_fs_word(control_word, (unsigned short *) d);
+      put_user(control_word, (unsigned short *) d);
 #endif PECULIAR_486
-      put_fs_word(status_word(), (unsigned short *) (d+2));
-      put_fs_word(tag_word(), (unsigned short *) (d+4));
-      put_fs_word(instruction_address.offset, (unsigned short *) (d+6));
-      put_fs_word(operand_address.offset, (unsigned short *) (d+0x0a));
+      put_user(status_word(), (unsigned short *) (d+2));
+      put_user(tag_word(), (unsigned short *) (d+4));
+      put_user(instruction_address.offset, (unsigned short *) (d+6));
+      put_user(operand_address.offset, (unsigned short *) (d+0x0a));
       if ( addr_modes.default_mode == VM86 )
 	{
-	  put_fs_word((instruction_address.offset & 0xf0000) >> 4,
+	  put_user((instruction_address.offset & 0xf0000) >> 4,
 		      (unsigned short *) (d+8));
-	  put_fs_word((operand_address.offset & 0xf0000) >> 4,
+	  put_user((operand_address.offset & 0xf0000) >> 4,
 		      (unsigned short *) (d+0x0c));
 	}
       else
 	{
-	  put_fs_word(instruction_address.selector, (unsigned short *) (d+8));
-	  put_fs_word(operand_address.selector, (unsigned short *) (d+0x0c));
+	  put_user(instruction_address.selector, (unsigned short *) (d+8));
+	  put_user(operand_address.selector, (unsigned short *) (d+0x0c));
 	}
       RE_ENTRANT_CHECK_ON;
       d += 0x0e;
@@ -1325,24 +1339,24 @@ char *fstenv(fpu_addr_modes addr_modes, char *d)
       FPU_verify_area(VERIFY_WRITE,d,28);
 #ifdef PECULIAR_486
       /* An 80486 sets all the reserved bits to 1. */
-      put_fs_long(0xffff0040 | (control_word & ~0xe080), (unsigned long *) d);
-      put_fs_long(0xffff0000 | status_word(), (unsigned long *) (d+4));
-      put_fs_long(0xffff0000 | tag_word(), (unsigned long *) (d+8));
+      put_user(0xffff0040 | (control_word & ~0xe080), (unsigned long *) d);
+      put_user(0xffff0000 | status_word(), (unsigned long *) (d+4));
+      put_user(0xffff0000 | tag_word(), (unsigned long *) (d+8));
 #else
-      put_fs_word(control_word, (unsigned short *) d);
-      put_fs_word(status_word(), (unsigned short *) (d+4));
-      put_fs_word(tag_word(), (unsigned short *) (d+8));
+      put_user(control_word, (unsigned short *) d);
+      put_user(status_word(), (unsigned short *) (d+4));
+      put_user(tag_word(), (unsigned short *) (d+8));
 #endif PECULIAR_486
-      put_fs_long(instruction_address.offset, (unsigned long *) (d+0x0c));
-      put_fs_word(instruction_address.selector, (unsigned short *) (d+0x10));
-      put_fs_word(instruction_address.opcode, (unsigned short *) (d+0x12));
-      put_fs_long(operand_address.offset, (unsigned long *) (d+0x14));
+      put_user(instruction_address.offset, (unsigned long *) (d+0x0c));
+      put_user(instruction_address.selector, (unsigned short *) (d+0x10));
+      put_user(instruction_address.opcode, (unsigned short *) (d+0x12));
+      put_user(operand_address.offset, (unsigned long *) (d+0x14));
 #ifdef PECULIAR_486
       /* An 80486 sets all the reserved bits to 1. */
-      put_fs_word(operand_address.selector, (unsigned short *) (d+0x18));
-      put_fs_word(0xffff, (unsigned short *) (d+0x1a));
+      put_user(operand_address.selector, (unsigned short *) (d+0x18));
+      put_user(0xffff, (unsigned short *) (d+0x1a));
 #else
-      put_fs_long(operand_address.selector, (unsigned long *) (d+0x18));
+      put_user(operand_address.selector, (unsigned long *) (d+0x18));
 #endif PECULIAR_486
       RE_ENTRANT_CHECK_ON;
       d += 0x1c;
@@ -1411,8 +1425,8 @@ static void write_to_extended(FPU_REG *rp, char *d)
     {
       /* just copy the reg */
       RE_ENTRANT_CHECK_OFF;
-      put_fs_long(rp->sigl, (unsigned long *) d);
-      put_fs_long(rp->sigh, (unsigned long *) (d + 4));
+      put_user(rp->sigl, (unsigned long *) d);
+      put_user(rp->sigh, (unsigned long *) (d + 4));
       RE_ENTRANT_CHECK_ON;
     }
   else
@@ -1427,12 +1441,12 @@ static void write_to_extended(FPU_REG *rp, char *d)
       round_to_int(&tmp);
       e = 0;
       RE_ENTRANT_CHECK_OFF;
-      put_fs_long(tmp.sigl, (unsigned long *) d);
-      put_fs_long(tmp.sigh, (unsigned long *) (d + 4));
+      put_user(tmp.sigl, (unsigned long *) d);
+      put_user(tmp.sigh, (unsigned long *) (d + 4));
       RE_ENTRANT_CHECK_ON;
     }
   e |= rp->sign == SIGN_POS ? 0 : 0x8000;
   RE_ENTRANT_CHECK_OFF;
-  put_fs_word(e, (unsigned short *) (d + 8));
+  put_user(e, (unsigned short *) (d + 8));
   RE_ENTRANT_CHECK_ON;
 }
diff --git a/arch/i386/math-emu/reg_norm.S b/arch/i386/math-emu/reg_norm.S
index 9b7a9d77d..781a2d466 100644
--- a/arch/i386/math-emu/reg_norm.S
+++ b/arch/i386/math-emu/reg_norm.S
@@ -1,9 +1,9 @@
 /*---------------------------------------------------------------------------+
  |  reg_norm.S                                                               |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
+ | Copyright (C) 1992,1993,1994,1995                                         |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Normalize the value in a FPU_REG.                                         |
  |                                                                           |
@@ -14,15 +14,11 @@
  |                                                                           |
  +---------------------------------------------------------------------------*/
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 
 .text
-
-	.align 2,144
-.globl _normalize
-
-_normalize:
+ENTRY(normalize)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%ebx
@@ -34,7 +30,7 @@ _normalize:
 	je	L_ok
 
 	pushl	$0x220
-	call	_exception
+	call	SYMBOL_NAME(FPU_exception)
 	addl	$4,%esp
 
 L_ok:
@@ -86,23 +82,20 @@ L_zero:
 
 L_underflow:
 	push	%ebx
-	call	_arith_underflow
+	call	SYMBOL_NAME(arith_underflow)
 	pop	%ebx
 	jmp	L_exit
 
 L_overflow:
 	push	%ebx
-	call	_arith_overflow
+	call	SYMBOL_NAME(arith_overflow)
 	pop	%ebx
 	jmp	L_exit
 
 
 
 /* Normalise without reporting underflow or overflow */
-	.align 2,144
-.globl _normalize_nuo
-
-_normalize_nuo:
+ENTRY(normalize_nuo)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%ebx
@@ -114,7 +107,7 @@ _normalize_nuo:
 	je	L_ok_nuo
 
 	pushl	$0x221
-	call	_exception
+	call	SYMBOL_NAME(FPU_exception)
 	addl	$4,%esp
 
 L_ok_nuo:
diff --git a/arch/i386/math-emu/reg_round.S b/arch/i386/math-emu/reg_round.S
index bd8a40dc4..4aac507a1 100644
--- a/arch/i386/math-emu/reg_round.S
+++ b/arch/i386/math-emu/reg_round.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Rounding/truncation/etc for FPU basic arithmetic functions.               |
  |                                                                           |
- | Copyright (C) 1993                                                        |
+ | Copyright (C) 1993,1995                                                   |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | This code has four possible entry points.                                 |
  | The following must be entered by a jmp instruction:                       |
@@ -69,7 +69,7 @@
  +---------------------------------------------------------------------------*/
 
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 #include "exception.h"
 #include "control_w.h"
 
@@ -101,14 +101,12 @@ FPU_denormal:
 
 
 .text
-	.align 2,144
 .globl fpu_reg_round
 .globl fpu_reg_round_sqrt
 .globl fpu_Arith_exit
-.globl _round_reg
 
 /* Entry point when called from C */
-_round_reg:
+ENTRY(round_reg)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%esi
@@ -435,7 +433,7 @@ fpu_Arith_exit:
  */
 xL_precision_lost_up:
 	push	%eax
-	call	_set_precision_flag_up
+	call	SYMBOL_NAME(set_precision_flag_up)
 	popl	%eax
 	jmp	xL_no_precision_loss
 
@@ -445,7 +443,7 @@ xL_precision_lost_up:
  */
 xL_precision_lost_down:
 	push	%eax
-	call	_set_precision_flag_down
+	call	SYMBOL_NAME(set_precision_flag_down)
 	popl	%eax
 	jmp	xL_no_precision_loss
 
@@ -598,7 +596,7 @@ LNormalise_shift_done:
 	/* There must be a masked underflow */
 	push	%eax
 	pushl	EX_Underflow
-	call	_exception
+	call	SYMBOL_NAME(FPU_exception)
 	popl	%eax
 	popl	%eax
 	jmp	xL_Normalised
@@ -610,12 +608,12 @@ LNormalise_shift_done:
  */
 L_underflow_to_zero:
 	push	%eax
-	call	_set_precision_flag_down
+	call	SYMBOL_NAME(set_precision_flag_down)
 	popl	%eax
 
 	push	%eax
 	pushl	EX_Underflow
-	call	_exception
+	call	SYMBOL_NAME(FPU_exception)
 	popl	%eax
 	popl	%eax
 
@@ -628,7 +626,7 @@ L_underflow_to_zero:
 /* The operations resulted in a number too large to represent. */
 L_overflow:
 	push	%edi
-	call	_arith_overflow
+	call	SYMBOL_NAME(arith_overflow)
 	pop	%edi
 	jmp	fpu_reg_round_exit
 
diff --git a/arch/i386/math-emu/reg_u_add.S b/arch/i386/math-emu/reg_u_add.S
index 4410f8fd4..1dc0d41df 100644
--- a/arch/i386/math-emu/reg_u_add.S
+++ b/arch/i386/math-emu/reg_u_add.S
@@ -5,9 +5,9 @@
  | Add two valid (TW_Valid) FPU_REG numbers, of the same sign, and put the   |
  |   result in a destination FPU_REG.                                        |
  |                                                                           |
- | Copyright (C) 1992,1993                                                   |
+ | Copyright (C) 1992,1993,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void reg_u_add(FPU_REG *arg1, FPU_REG *arg2, FPU_REG *answ,             |
@@ -25,13 +25,11 @@
  */
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 #include "control_w.h"
 
 .text
-	.align 2,144
-.globl _reg_u_add
-_reg_u_add:
+ENTRY(reg_u_add)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%esi
@@ -45,7 +43,7 @@ _reg_u_add:
 	cmpl	EXP_UNDER,EXP(%esi)
 	jg	xOp1_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
@@ -53,7 +51,7 @@ xOp1_not_denorm:
 	cmpl	EXP_UNDER,EXP(%edi)
 	jg	xOp2_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
diff --git a/arch/i386/math-emu/reg_u_div.S b/arch/i386/math-emu/reg_u_div.S
index 328e9116e..5bba98dd8 100644
--- a/arch/i386/math-emu/reg_u_div.S
+++ b/arch/i386/math-emu/reg_u_div.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Core division routines                                                    |
  |                                                                           |
- | Copyright (C) 1992,1993                                                   |
+ | Copyright (C) 1992,1993,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -21,7 +21,7 @@
  +---------------------------------------------------------------------------*/
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 #include "control_w.h"
 
 
@@ -69,13 +69,7 @@ FPU_ovfl_flag:
 
 
 .text
-	.align 2,144
-
-.globl _reg_u_div
-
-.globl _divide_kernel
-
-_reg_u_div:
+ENTRY(reg_u_div)
 	pushl	%ebp
 	movl	%esp,%ebp
 #ifndef NON_REENTRANT_FPU
@@ -95,7 +89,7 @@ _reg_u_div:
 	cmpl	EXP_UNDER,%eax
 	jg	xOp1_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
@@ -104,14 +98,14 @@ xOp1_not_denorm:
 	cmpl	EXP_UNDER,%eax
 	jg	xOp2_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
 xOp2_not_denorm:
 #endif DENORM_OPERAND
 
-_divide_kernel:
+ENTRY(divide_kernel)
 #ifdef PARANOID
 /*	testl	$0x80000000, SIGH(%esi)	// Dividend */
 /*	je	L_bugged */
diff --git a/arch/i386/math-emu/reg_u_mul.S b/arch/i386/math-emu/reg_u_mul.S
index 8250666bd..682fbec15 100644
--- a/arch/i386/math-emu/reg_u_mul.S
+++ b/arch/i386/math-emu/reg_u_mul.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Core multiplication routine                                               |
  |                                                                           |
- | Copyright (C) 1992,1993                                                   |
+ | Copyright (C) 1992,1993,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
@@ -22,7 +22,7 @@
  +---------------------------------------------------------------------------*/
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 #include "control_w.h"
 
 
@@ -44,10 +44,7 @@ FPU_accum_1:
 
 
 .text
-	.align 2,144
-
-.globl _reg_u_mul
-_reg_u_mul:
+ENTRY(reg_u_mul)
 	pushl	%ebp
 	movl	%esp,%ebp
 #ifndef NON_REENTRANT_FPU
@@ -73,7 +70,7 @@ _reg_u_mul:
 	cmpl	EXP_UNDER,%eax
 	jg	xOp1_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
@@ -82,7 +79,7 @@ xOp1_not_denorm:
 	cmpl	EXP_UNDER,%eax
 	jg	xOp2_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
diff --git a/arch/i386/math-emu/reg_u_sub.S b/arch/i386/math-emu/reg_u_sub.S
index fbec17dfb..891670260 100644
--- a/arch/i386/math-emu/reg_u_sub.S
+++ b/arch/i386/math-emu/reg_u_sub.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Core floating point subtraction routine.                                  |
  |                                                                           |
- | Copyright (C) 1992,1993                                                   |
+ | Copyright (C) 1992,1993,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void reg_u_sub(FPU_REG *arg1, FPU_REG *arg2, FPU_REG *answ,             |
@@ -26,13 +26,11 @@
  */
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 #include "control_w.h"
 
 .text
-	.align 2,144
-.globl _reg_u_sub
-_reg_u_sub:
+ENTRY(reg_u_sub)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%esi
@@ -46,7 +44,7 @@ _reg_u_sub:
 	cmpl	EXP_UNDER,EXP(%esi)
 	jg	xOp1_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
@@ -54,7 +52,7 @@ xOp1_not_denorm:
 	cmpl	EXP_UNDER,EXP(%edi)
 	jg	xOp2_not_denorm
 
-	call	_denormal_operand
+	call	SYMBOL_NAME(denormal_operand)
 	orl	%eax,%eax
 	jnz	fpu_Arith_exit
 
diff --git a/arch/i386/math-emu/round_Xsig.S b/arch/i386/math-emu/round_Xsig.S
index 163755878..bbe0e8771 100644
--- a/arch/i386/math-emu/round_Xsig.S
+++ b/arch/i386/math-emu/round_Xsig.S
@@ -1,9 +1,9 @@
 /*---------------------------------------------------------------------------+
  |  round_Xsig.S                                                             |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
+ | Copyright (C) 1992,1993,1994,1995                                         |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Normalize and round a 12 byte quantity.                                   |
  | Call from C as:                                                           |
@@ -18,15 +18,11 @@
  +---------------------------------------------------------------------------*/
 	.file	"round_Xsig.S"
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 
 .text
-
-	.align 2,144
-.globl _round_Xsig
-
-_round_Xsig:
+ENTRY(round_Xsig)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%ebx		/* Reserve some space */
@@ -86,10 +82,7 @@ L_exit:
 
 
 
-	.align 2,144
-.globl _norm_Xsig
-
-_norm_Xsig:
+ENTRY(norm_Xsig)
 	pushl	%ebp
 	movl	%esp,%ebp
 	pushl	%ebx		/* Reserve some space */
diff --git a/arch/i386/math-emu/shr_Xsig.S b/arch/i386/math-emu/shr_Xsig.S
index d6724a204..31cdd118e 100644
--- a/arch/i386/math-emu/shr_Xsig.S
+++ b/arch/i386/math-emu/shr_Xsig.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | 12 byte right shift function                                              |
  |                                                                           |
- | Copyright (C) 1992,1994                                                   |
+ | Copyright (C) 1992,1994,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void shr_Xsig(Xsig *arg, unsigned nr)                                   |
@@ -18,13 +18,10 @@
  |                                                                           |
  +---------------------------------------------------------------------------*/
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 .text
-	.align 2,144
-
-	.globl	_shr_Xsig
-_shr_Xsig:
+ENTRY(shr_Xsig)
 	push	%ebp
 	movl	%esp,%ebp
 	pushl	%esi
diff --git a/arch/i386/math-emu/status_w.h b/arch/i386/math-emu/status_w.h
index 96607d0e1..57d1ae6bb 100644
--- a/arch/i386/math-emu/status_w.h
+++ b/arch/i386/math-emu/status_w.h
@@ -12,7 +12,7 @@
 
 #include "fpu_emu.h"    /* for definition of PECULIAR_486 */
 
-#ifdef __ASSEMBLER__
+#ifdef __ASSEMBLY__
 #define	Const__(x)	$##x
 #else
 #define	Const__(x)	x
@@ -36,7 +36,7 @@
 
 #define SW_Exc_Mask     Const__(0x27f)  /* Status word exception bit mask */
 
-#ifndef __ASSEMBLER__
+#ifndef __ASSEMBLY__
 
 #define COMP_A_gt_B	1
 #define COMP_A_eq_B	2
@@ -60,6 +60,6 @@
 #  define clear_C1()
 #endif PECULIAR_486
 
-#endif __ASSEMBLER__
+#endif __ASSEMBLY__
 
 #endif _STATUS_H_
diff --git a/arch/i386/math-emu/version.h b/arch/i386/math-emu/version.h
index 4c75a4792..d966ab193 100644
--- a/arch/i386/math-emu/version.h
+++ b/arch/i386/math-emu/version.h
@@ -2,11 +2,11 @@
  |  version.h                                                                |
  |                                                                           |
  |                                                                           |
- | Copyright (C) 1992,1993,1994                                              |
- |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1993,1994,1996                                         |
+ |                  W. Metzenthen, 22 Parker St, Ormond, Vic 3163, Australia |
+ |                  E-mail   billm@jacobi.maths.monash.edu.au                |
  |                                                                           |
  |                                                                           |
  +---------------------------------------------------------------------------*/
 
-#define FPU_VERSION "wm-FPU-emu version 1.20"
+#define FPU_VERSION "wm-FPU-emu version 1.22"
diff --git a/arch/i386/math-emu/wm_shrx.S b/arch/i386/math-emu/wm_shrx.S
index bef0e1963..1ea7ff7da 100644
--- a/arch/i386/math-emu/wm_shrx.S
+++ b/arch/i386/math-emu/wm_shrx.S
@@ -4,8 +4,9 @@
  |                                                                           |
  | 64 bit right shift functions                                              |
  |                                                                           |
- | Copyright (C) 1992    W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ | Copyright (C) 1992,1995                                                   |
+ |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   unsigned shrx(void *arg1, unsigned arg2)                                |
@@ -14,11 +15,9 @@
  |                                                                           |
  +---------------------------------------------------------------------------*/
 
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 .text
-	.align 2,144
-
 /*---------------------------------------------------------------------------+
  |   unsigned shrx(void *arg1, unsigned arg2)                                |
  |                                                                           |
@@ -33,9 +32,7 @@
  |   Results returned in the 64 bit arg and eax.                             |
  +---------------------------------------------------------------------------*/
 
-	.globl	_shrx
-
-_shrx:
+ENTRY(shrx)
 	push	%ebp
 	movl	%esp,%ebp
 	pushl	%esi
@@ -113,8 +110,7 @@ L_more_than_95:
  |   part which has been shifted out of the arg.                             |
  |   Results returned in the 64 bit arg and eax.                             |
  +---------------------------------------------------------------------------*/
-	.globl	_shrxs
-_shrxs:
+ENTRY(shrxs)
 	push	%ebp
 	movl	%esp,%ebp
 	pushl	%esi
diff --git a/arch/i386/math-emu/wm_sqrt.S b/arch/i386/math-emu/wm_sqrt.S
index 4e028cb80..848796188 100644
--- a/arch/i386/math-emu/wm_sqrt.S
+++ b/arch/i386/math-emu/wm_sqrt.S
@@ -4,9 +4,9 @@
  |                                                                           |
  | Fixed point arithmetic square root evaluation.                            |
  |                                                                           |
- | Copyright (C) 1992,1993                                                   |
+ | Copyright (C) 1992,1993,1995                                              |
  |                       W. Metzenthen, 22 Parker St, Ormond, Vic 3163,      |
- |                       Australia.  E-mail   billm@vaxc.cc.monash.edu.au    |
+ |                       Australia.  E-mail billm@jacobi.maths.monash.edu.au |
  |                                                                           |
  | Call from C as:                                                           |
  |   void wm_sqrt(FPU_REG *n, unsigned int control_word)                     |
@@ -26,7 +26,7 @@
  +---------------------------------------------------------------------------*/
 
 #include "exception.h"
-#include "fpu_asm.h"
+#include "fpu_emu.h"
 
 
 #ifndef NON_REENTRANT_FPU
@@ -74,10 +74,7 @@ FPU_fsqrt_arg_0:
 
 
 .text
-	.align 2,144
-
-.globl _wm_sqrt
-_wm_sqrt:
+ENTRY(wm_sqrt)
 	pushl	%ebp
 	movl	%esp,%ebp
 #ifndef NON_REENTRANT_FPU
diff --git a/arch/i386/mm/Makefile b/arch/i386/mm/Makefile
index af75a20a4..cee7d4e6d 100644
--- a/arch/i386/mm/Makefile
+++ b/arch/i386/mm/Makefile
@@ -7,26 +7,7 @@
 #
 # Note 2! The CFLAGS definition is now in the main makefile...
 
-.c.o:
-	$(CC) $(CFLAGS) -c $<
-.s.o:
-	$(AS) -o $*.o $<
-.c.s:
-	$(CC) $(CFLAGS) -S $<
+O_TARGET := mm.o
+O_OBJS	 := init.o fault.o ioremap.o extable.o
 
-OBJS	= init.o fault.o
-
-mm.o: $(OBJS)
-	$(LD) -r -o mm.o $(OBJS)
-
-modules:
-
-dep:
-	$(CPP) -M *.c > .depend
-
-#
-# include a dependency file if one exists
-#
-ifeq (.depend,$(wildcard .depend))
-include .depend
-endif
+include $(TOPDIR)/Rules.make
diff --git a/arch/i386/mm/extable.c b/arch/i386/mm/extable.c
new file mode 100644
index 000000000..c43a8d43d
--- /dev/null
+++ b/arch/i386/mm/extable.c
@@ -0,0 +1,63 @@
+/*
+ * linux/arch/i386/mm/extable.c
+ */
+
+#include <linux/config.h>
+#include <linux/module.h>
+#include <asm/uaccess.h>
+
+extern const struct exception_table_entry __start___ex_table[];
+extern const struct exception_table_entry __stop___ex_table[];
+
+static inline unsigned long
+search_one_table(const struct exception_table_entry *first,
+		 const struct exception_table_entry *last,
+		 unsigned long value)
+{
+	/* Some versions of the linker are buggy and do not align the
+	   __start pointer along with the section, thus we may be low.  */
+	if ((long)first & 3)
+		(long)first = ((long)first | 3) + 1;
+
+        while (first <= last) {
+		const struct exception_table_entry *mid;
+		long diff;
+
+		mid = (last - first) / 2 + first;
+		diff = mid->insn - value;
+                if (diff == 0)
+                        return mid->fixup;
+                else if (diff < 0)
+                        first = mid+1;
+                else
+                        last = mid-1;
+        }
+        return 0;
+}
+
+unsigned long
+search_exception_table(unsigned long addr)
+{
+	unsigned long ret;
+#ifdef CONFIG_MODULES
+	struct module *mp;
+#endif
+
+	/* Search the kernel's table first.  */
+	ret = search_one_table(__start___ex_table,
+			       __stop___ex_table-1, addr);
+	if (ret)
+		return ret;
+
+#ifdef CONFIG_MODULES
+	for (mp = module_list; mp != NULL; mp = mp->next) {
+		if (mp->exceptinfo.start != NULL) {
+			ret = search_one_table(mp->exceptinfo.start,
+				mp->exceptinfo.stop-1, addr);
+			if (ret)
+				return ret;
+		}
+	}
+#endif
+	return 0;
+}
diff --git a/arch/i386/mm/fault.c b/arch/i386/mm/fault.c
index 01c259f0f..50dcd0735 100644
--- a/arch/i386/mm/fault.c
+++ b/arch/i386/mm/fault.c
@@ -4,7 +4,6 @@
  *  Copyright (C) 1995  Linus Torvalds
  */
 
-#include <linux/config.h>
 #include <linux/signal.h>
 #include <linux/sched.h>
 #include <linux/head.h>
@@ -17,10 +16,61 @@
 #include <linux/mm.h>
 
 #include <asm/system.h>
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 #include <asm/pgtable.h>
 
-extern void die_if_kernel(char *,struct pt_regs *,long);
+extern void die_if_kernel(const char *,struct pt_regs *,long);
+
+/*
+ * Ugly, ugly, but the goto's result in better assembly..
+ */
+int __verify_write(const void * addr, unsigned long size)
+{
+	struct vm_area_struct * vma;
+	unsigned long start = (unsigned long) addr;
+
+	if (!size)
+		return 1;
+
+	vma = find_vma(current->mm, start);
+	if (!vma)
+		goto bad_area;
+	if (vma->vm_start > start)
+		goto check_stack;
+
+good_area:
+	if (!(vma->vm_flags & VM_WRITE))
+		goto bad_area;
+	size--;
+	size += start & ~PAGE_MASK;
+	size >>= PAGE_SHIFT;
+	start &= PAGE_MASK;
+
+	for (;;) {
+		do_wp_page(current, vma, start, 1);
+		if (!size)
+			break;
+		size--;
+		start += PAGE_SIZE;
+		if (start < vma->vm_end)
+			continue;
+		vma = vma->vm_next;
+		if (!vma || vma->vm_start != start)
+			goto bad_area;
+		if (!(vma->vm_flags & VM_WRITE))
+			goto bad_area;;
+	}
+	return 1;
+
+check_stack:
+	if (!(vma->vm_flags & VM_GROWSDOWN))
+		goto bad_area;
+	if (expand_stack(vma, start) == 0)
+		goto good_area;
+
+bad_area:
+	return 0;
+}
 
 /*
  * This routine handles page faults.  It determines the address,
@@ -34,58 +84,76 @@ extern void die_if_kernel(char *,struct pt_regs *,long);
  */
 asmlinkage void do_page_fault(struct pt_regs *regs, unsigned long error_code)
 {
+	void (*handler)(struct task_struct *,
+			struct vm_area_struct *,
+			unsigned long,
+			int);
+	struct task_struct *tsk = current;
+	struct mm_struct *mm = tsk->mm;
 	struct vm_area_struct * vma;
 	unsigned long address;
 	unsigned long page;
+	unsigned long fixup;
+	int write;
 
 	/* get the address */
 	__asm__("movl %%cr2,%0":"=r" (address));
-	vma = find_vma(current, address);
+	down(&mm->mmap_sem);
+	vma = find_vma(mm, address);
 	if (!vma)
 		goto bad_area;
 	if (vma->vm_start <= address)
 		goto good_area;
 	if (!(vma->vm_flags & VM_GROWSDOWN))
 		goto bad_area;
-	if (vma->vm_end - address > current->rlim[RLIMIT_STACK].rlim_cur)
+	if (error_code & 4) {
+		/*
+		 * accessing the stack below %esp is always a bug.
+		 * The "+ 32" is there due to some instructions (like
+		 * pusha) doing pre-decrement on the stack and that
+		 * doesn't show up until later..
+		 */
+		if (address + 32 < regs->esp)
+			goto bad_area;
+	}
+	if (expand_stack(vma, address))
 		goto bad_area;
-	vma->vm_offset -= vma->vm_start - (address & PAGE_MASK);
-	vma->vm_start = (address & PAGE_MASK);
 /*
  * Ok, we have a good vm_area for this memory access, so
  * we can handle it..
  */
 good_area:
-	/*
-	 * was it a write?
-	 */
-	if (error_code & 2) {
-		if (!(vma->vm_flags & VM_WRITE))
-			goto bad_area;
-	} else {
-		/* read with protection fault? */
-		if (error_code & 1)
-			goto bad_area;
-		if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
+	write = 0;
+	handler = do_no_page;
+	switch (error_code & 3) {
+		default:	/* 3: write, present */
+			handler = do_wp_page;
+#ifdef TEST_VERIFY_AREA
+			if (regs->cs == KERNEL_CS)
+				printk("WP fault at %08lx\n", regs->eip);
+#endif
+			/* fall through */
+		case 2:		/* write, not present */
+			if (!(vma->vm_flags & VM_WRITE))
+				goto bad_area;
+			write++;
+			break;
+		case 1:		/* read, present */
 			goto bad_area;
+		case 0:		/* read, not present */
+			if (!(vma->vm_flags & (VM_READ | VM_EXEC)))
+				goto bad_area;
 	}
+	handler(tsk, vma, address, write);
+	up(&mm->mmap_sem);
 	/*
 	 * Did it hit the DOS screen memory VA from vm86 mode?
 	 */
 	if (regs->eflags & VM_MASK) {
 		unsigned long bit = (address - 0xA0000) >> PAGE_SHIFT;
 		if (bit < 32)
-			current->tss.screen_bitmap |= 1 << bit;
-	}
-	if (error_code & 1) {
-#ifdef CONFIG_TEST_VERIFY_AREA
-		if (regs->cs == KERNEL_CS)
-			printk("WP fault at %08x\n", regs->eip);
-#endif
-		do_wp_page(vma, address, error_code & 2);
-		return;
+			tsk->tss.screen_bitmap |= 1 << bit;
 	}
-	do_no_page(vma, address, error_code & 2);
 	return;
 
 /*
@@ -93,11 +161,20 @@ good_area:
  * Fix it, but check if it's kernel or user first..
  */
 bad_area:
+	up(&mm->mmap_sem);
+
+	/* Are we prepared to handle this fault?  */
+	if ((fixup = search_exception_table(regs->eip)) != 0) {
+		printk("Exception at %lx (%lx)\n", regs->eip, fixup);
+		regs->eip = fixup;
+		return;
+	}
+
 	if (error_code & 4) {
-		current->tss.cr2 = address;
-		current->tss.error_code = error_code;
-		current->tss.trap_no = 14;
-		send_sig(SIGSEGV, current, 1);
+		tsk->tss.cr2 = address;
+		tsk->tss.error_code = error_code;
+		tsk->tss.trap_no = 14;
+		force_sig(SIGSEGV, tsk);
 		return;
 	}
 /*
@@ -106,14 +183,14 @@ bad_area:
  *
  * First we check if it was the bootup rw-test, though..
  */
-	if (wp_works_ok < 0 && address == TASK_SIZE && (error_code & 1)) {
+	if (wp_works_ok < 0 && !address && (error_code & 1)) {
 		wp_works_ok = 1;
 		pg0[0] = pte_val(mk_pte(0, PAGE_SHARED));
-		invalidate();
+		flush_tlb();
 		printk("This processor honours the WP bit even when in supervisor mode. Good.\n");
 		return;
 	}
-	if ((unsigned long) (address-TASK_SIZE) < PAGE_SIZE) {
+	if (address < PAGE_SIZE) {
 		printk(KERN_ALERT "Unable to handle kernel NULL pointer dereference");
 		pg0[0] = pte_val(mk_pte(0, PAGE_SHARED));
 	} else
@@ -121,13 +198,13 @@ bad_area:
 	printk(" at virtual address %08lx\n",address);
 	__asm__("movl %%cr3,%0" : "=r" (page));
 	printk(KERN_ALERT "current->tss.cr3 = %08lx, %%cr3 = %08lx\n",
-		current->tss.cr3, page);
-	page = ((unsigned long *) page)[address >> 22];
+		tsk->tss.cr3, page);
+	page = ((unsigned long *) __va(page))[address >> 22];
 	printk(KERN_ALERT "*pde = %08lx\n", page);
 	if (page & 1) {
 		page &= PAGE_MASK;
 		address &= 0x003ff000;
-		page = ((unsigned long *) page)[address >> PAGE_SHIFT];
+		page = ((unsigned long *) __va(page))[address >> PAGE_SHIFT];
 		printk(KERN_ALERT "*pte = %08lx\n", page);
 	}
 	die_if_kernel("Oops", regs, error_code);
diff --git a/arch/i386/mm/init.c b/arch/i386/mm/init.c
index 296595644..cf258d6de 100644
--- a/arch/i386/mm/init.c
+++ b/arch/i386/mm/init.c
@@ -15,13 +15,17 @@
 #include <linux/ptrace.h>
 #include <linux/mman.h>
 #include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/smp.h>
+#ifdef CONFIG_BLK_DEV_INITRD
+#include <linux/blk.h>
+#endif
 
 #include <asm/system.h>
-#include <asm/segment.h>
+#include <asm/uaccess.h>
 #include <asm/pgtable.h>
+#include <asm/dma.h>
 
-extern void scsi_mem_init(unsigned long);
-extern void sound_mem_init(void);
 extern void die_if_kernel(char *,struct pt_regs *,long);
 extern void show_net_buffers(void);
 
@@ -29,7 +33,7 @@ extern void show_net_buffers(void);
  * BAD_PAGE is the page that is used for page faults when linux
  * is out-of-memory. Older versions of linux just did a
  * do_exit(), but using this instead means there is less risk
- * for a process dying in kernel mode, possibly leaving a inode
+ * for a process dying in kernel mode, possibly leaving an inode
  * unused etc..
  *
  * BAD_PAGETABLE is the accompanying page-table: it is initialized
@@ -70,15 +74,15 @@ void show_mem(void)
 	printk("Mem-info:\n");
 	show_free_areas();
 	printk("Free swap:       %6dkB\n",nr_swap_pages<<(PAGE_SHIFT-10));
-	i = high_memory >> PAGE_SHIFT;
+	i = max_mapnr;
 	while (i-- > 0) {
 		total++;
-		if (mem_map[i] & MAP_PAGE_RESERVED)
+		if (PageReserved(mem_map+i))
 			reserved++;
-		else if (!mem_map[i])
+		else if (!mem_map[i].count)
 			free++;
 		else
-			shared += mem_map[i]-1;
+			shared += mem_map[i].count-1;
 	}
 	printk("%d pages of RAM\n",total);
 	printk("%d free pages\n",free);
@@ -111,34 +115,90 @@ unsigned long paging_init(unsigned long start_mem, unsigned long end_mem)
  * and SMM (for laptops with [34]86/SL chips) may need it.  It is read
  * and write protected to detect null pointer references in the
  * kernel.
+ * It may also hold the MP configuration table when we are booting SMP.
  */
 #if 0
 	memset((void *) 0, 0, PAGE_SIZE);
 #endif
+#ifdef __SMP__
+	if (!smp_scan_config(0x0,0x400))	/* Scan the bottom 1K for a signature */
+	{
+		/*
+		 *	FIXME: Linux assumes you have 640K of base ram.. this continues
+		 *	the error...
+		 */
+		if (!smp_scan_config(639*0x400,0x400))	/* Scan the top 1K of base RAM */
+			smp_scan_config(0xF0000,0x10000);	/* Scan the 64K of bios */
+	}
+	/*
+	 *	If it is an SMP machine we should know now, unless the configuration
+	 *	is in an EISA/MCA bus machine with an extended bios data area. I don't
+	 *	have such a machine so someone else can fill in the check of the EBDA
+	 *	here.
+	 */
+/*	smp_alloc_memory(8192); */
+#endif
+#ifdef TEST_VERIFY_AREA
+	wp_works_ok = 0;
+#endif
 	start_mem = PAGE_ALIGN(start_mem);
-	address = 0;
+	address = PAGE_OFFSET;
 	pg_dir = swapper_pg_dir;
+	/* unmap the original low memory mappings */
+	pgd_val(pg_dir[0]) = 0;
 	while (address < end_mem) {
+		/*
+		 * The following code enabled 4MB page tables for the
+		 * Intel Pentium cpu, unfortunately the SMP kernel can't
+		 * handle the 4MB page table optimizations yet
+		 */
+#ifndef __SMP__
+		/*
+		 * This will create page tables that
+		 * span up to the next 4MB virtual
+		 * memory boundary, but that's ok,
+		 * we won't use that memory anyway.
+		 */
+		if (x86_capability & 8) {
+#ifdef GAS_KNOWS_CR4
+			__asm__("movl %%cr4,%%eax\n\t"
+				"orl $16,%%eax\n\t"
+				"movl %%eax,%%cr4"
+				: : :"ax");
+#else
+			__asm__(".byte 0x0f,0x20,0xe0\n\t"
+				"orl $16,%%eax\n\t"
+				".byte 0x0f,0x22,0xe0"
+				: : :"ax");
+#endif
+			wp_works_ok = 1;
+			pgd_val(pg_dir[768]) = _PAGE_TABLE + _PAGE_4M + __pa(address);
+			pg_dir++;
+			address += 4*1024*1024;
+			continue;
+		}
+#endif
 		/* map the memory at virtual addr 0xC0000000 */
+		/* pg_table is physical at this point */
 		pg_table = (pte_t *) (PAGE_MASK & pgd_val(pg_dir[768]));
 		if (!pg_table) {
-			pg_table = (pte_t *) start_mem;
+			pg_table = (pte_t *) __pa(start_mem);
 			start_mem += PAGE_SIZE;
 		}
 
-		/* also map it temporarily at 0x0000000 for init */
-		pgd_val(pg_dir[0])   = _PAGE_TABLE | (unsigned long) pg_table;
 		pgd_val(pg_dir[768]) = _PAGE_TABLE | (unsigned long) pg_table;
 		pg_dir++;
+		/* now change pg_table to kernel virtual addresses */
+		pg_table = (pte_t *) __va(pg_table);
 		for (tmp = 0 ; tmp < PTRS_PER_PTE ; tmp++,pg_table++) {
-			if (address < end_mem)
-				*pg_table = mk_pte(address, PAGE_SHARED);
-			else
-				pte_clear(pg_table);
+			pte_t pte = mk_pte(address, PAGE_KERNEL);
+			if (address >= end_mem)
+				pte_val(pte) = 0;
+			set_pte(pg_table, pte);
 			address += PAGE_SIZE;
 		}
 	}
-	invalidate();
+	local_flush_tlb();
 	return free_area_init(start_mem, end_mem);
 }
 
@@ -149,16 +209,25 @@ void mem_init(unsigned long start_mem, unsigned long end_mem)
 	int reservedpages = 0;
 	int datapages = 0;
 	unsigned long tmp;
-	extern int etext;
+	extern int _etext;
 
 	end_mem &= PAGE_MASK;
-	high_memory = end_mem;
+	high_memory = (void *) end_mem;
+	max_mapnr = MAP_NR(end_mem);
 
 	/* clear the zero-page */
 	memset(empty_zero_page, 0, PAGE_SIZE);
 
 	/* mark usable pages in the mem_map[] */
-	start_low_mem = PAGE_ALIGN(start_low_mem);
+	start_low_mem = PAGE_ALIGN(start_low_mem)+PAGE_OFFSET;
+
+#ifdef __SMP__
+	/*
+	 * But first pinch a few for the stack/trampoline stuff
+	 */
+	start_low_mem += PAGE_SIZE;				/* 32bit startup code */
+	start_low_mem = smp_alloc_memory(start_low_mem); 	/* AP processor stacks */
+#endif
 	start_mem = PAGE_ALIGN(start_mem);
 
 	/*
@@ -166,53 +235,56 @@ void mem_init(unsigned long start_mem, unsigned long end_mem)
 	 * They seem to have done something stupid with the floppy
 	 * controller as well..
 	 */
-	while (start_low_mem < 0x9f000) {
-		mem_map[MAP_NR(start_low_mem)] = 0;
+	while (start_low_mem < 0x9f000+PAGE_OFFSET) {
+		clear_bit(PG_reserved, &mem_map[MAP_NR(start_low_mem)].flags);
 		start_low_mem += PAGE_SIZE;
 	}
 
-	while (start_mem < high_memory) {
-		mem_map[MAP_NR(start_mem)] = 0;
+	while (start_mem < end_mem) {
+		clear_bit(PG_reserved, &mem_map[MAP_NR(start_mem)].flags);
 		start_mem += PAGE_SIZE;
 	}
-#ifdef CONFIG_SCSI
-	scsi_mem_init(high_memory);
-#endif
-#ifdef CONFIG_SOUND
-	sound_mem_init();
-#endif
-	for (tmp = 0 ; tmp < high_memory ; tmp += PAGE_SIZE) {
-		if (mem_map[MAP_NR(tmp)]) {
-			if (tmp >= 0xA0000 && tmp < 0x100000)
+	for (tmp = PAGE_OFFSET ; tmp < end_mem ; tmp += PAGE_SIZE) {
+		if (tmp >= MAX_DMA_ADDRESS)
+			clear_bit(PG_DMA, &mem_map[MAP_NR(tmp)].flags);
+		if (PageReserved(mem_map+MAP_NR(tmp))) {
+			if (tmp >= 0xA0000+PAGE_OFFSET && tmp < 0x100000+PAGE_OFFSET)
 				reservedpages++;
-			else if (tmp < (unsigned long) &etext)
+			else if (tmp < (unsigned long) &_etext)
 				codepages++;
 			else
 				datapages++;
 			continue;
 		}
-		mem_map[MAP_NR(tmp)] = 1;
-		free_page(tmp);
+		mem_map[MAP_NR(tmp)].count = 1;
+#ifdef CONFIG_BLK_DEV_INITRD
+		if (!initrd_start || (tmp < initrd_start || tmp >=
+		    initrd_end))
+#endif
+			free_page(tmp);
 	}
-	tmp = nr_free_pages << PAGE_SHIFT;
 	printk("Memory: %luk/%luk available (%dk kernel code, %dk reserved, %dk data)\n",
-		tmp >> 10,
-		high_memory >> 10,
+		(unsigned long) nr_free_pages << (PAGE_SHIFT-10),
+		max_mapnr << (PAGE_SHIFT-10),
 		codepages << (PAGE_SHIFT-10),
 		reservedpages << (PAGE_SHIFT-10),
 		datapages << (PAGE_SHIFT-10));
 /* test if the WP bit is honoured in supervisor mode */
-	wp_works_ok = -1;
-	pg0[0] = pte_val(mk_pte(0, PAGE_READONLY));
-	invalidate();
-	__asm__ __volatile__("movb 0,%%al ; movb %%al,0": : :"ax", "memory");
-	pg0[0] = 0;
-	invalidate();
-	if (wp_works_ok < 0)
-		wp_works_ok = 0;
-#ifdef CONFIG_TEST_VERIFY_AREA
-	wp_works_ok = 0;
-#endif
+	if (wp_works_ok < 0) {
+		unsigned char tmp_reg;
+		pg0[0] = pte_val(mk_pte(PAGE_OFFSET, PAGE_READONLY));
+		local_flush_tlb();
+		__asm__ __volatile__(
+			"movb %0,%1 ; movb %1,%0"
+			:"=m" (*(char *) __va(0)),
+			 "=q" (tmp_reg)
+			:/* no inputs */
+			:"memory");
+		pg0[0] = pte_val(mk_pte(PAGE_OFFSET, PAGE_KERNEL));
+		local_flush_tlb();
+		if (wp_works_ok < 0)
+			wp_works_ok = 0;
+	}
 	return;
 }
 
@@ -220,18 +292,18 @@ void si_meminfo(struct sysinfo *val)
 {
 	int i;
 
-	i = high_memory >> PAGE_SHIFT;
+	i = max_mapnr;
 	val->totalram = 0;
 	val->sharedram = 0;
 	val->freeram = nr_free_pages << PAGE_SHIFT;
 	val->bufferram = buffermem;
 	while (i-- > 0)  {
-		if (mem_map[i] & MAP_PAGE_RESERVED)
+		if (PageReserved(mem_map+i))
 			continue;
 		val->totalram++;
-		if (!mem_map[i])
+		if (!mem_map[i].count)
 			continue;
-		val->sharedram += mem_map[i]-1;
+		val->sharedram += mem_map[i].count-1;
 	}
 	val->totalram <<= PAGE_SHIFT;
 	val->sharedram <<= PAGE_SHIFT;
diff --git a/arch/i386/mm/ioremap.c b/arch/i386/mm/ioremap.c
new file mode 100644
index 000000000..42b386c1c
--- /dev/null
+++ b/arch/i386/mm/ioremap.c
@@ -0,0 +1,109 @@
+/*
+ * arch/i386/mm/ioremap.c
+ *
+ * Re-map IO memory to kernel address space so that we can access it.
+ * This is needed for high PCI addresses that aren't mapped in the
+ * 640k-1MB IO memory area on PC's
+ *
+ * (C) Copyright 1995 1996 Linus Torvalds
+ */
+
+#include <linux/vmalloc.h>
+
+#include <asm/io.h>
+
+static inline void remap_area_pte(pte_t * pte, unsigned long address, unsigned long size,
+	unsigned long phys_addr)
+{
+	unsigned long end;
+
+	address &= ~PMD_MASK;
+	end = address + size;
+	if (end > PMD_SIZE)
+		end = PMD_SIZE;
+	do {
+		if (!pte_none(*pte))
+			printk("remap_area_pte: page already exists\n");
+		set_pte(pte, mk_pte_phys(phys_addr, PAGE_KERNEL));
+		address += PAGE_SIZE;
+		phys_addr += PAGE_SIZE;
+		pte++;
+	} while (address < end);
+}
+
+static inline int remap_area_pmd(pmd_t * pmd, unsigned long address, unsigned long size,
+	unsigned long phys_addr)
+{
+	unsigned long end;
+
+	address &= ~PGDIR_MASK;
+	end = address + size;
+	if (end > PGDIR_SIZE)
+		end = PGDIR_SIZE;
+	phys_addr -= address;
+	do {
+		pte_t * pte = pte_alloc_kernel(pmd, address);
+		if (!pte)
+			return -ENOMEM;
+		remap_area_pte(pte, address, end - address, address + phys_addr);
+		address = (address + PMD_SIZE) & PMD_MASK;
+		pmd++;
+	} while (address < end);
+	return 0;
+}
+
+static int remap_area_pages(unsigned long address, unsigned long phys_addr, unsigned long size)
+{
+	pgd_t * dir;
+	unsigned long end = address + size;
+
+	phys_addr -= address;
+	dir = pgd_offset(&init_mm, address);
+	flush_cache_all();
+	while (address < end) {
+		pmd_t *pmd = pmd_alloc_kernel(dir, address);
+		if (!pmd)
+			return -ENOMEM;
+		if (remap_area_pmd(pmd, address, end - address, phys_addr + address))
+			return -ENOMEM;
+		set_pgdir(address, *dir);
+		address = (address + PGDIR_SIZE) & PGDIR_MASK;
+		dir++;
+	}
+	flush_tlb_all();
+	return 0;
+}
+
+/*
+ * Remap an arbitrary physical address space into the kernel virtual
+ * address space. Needed when the kernel wants to access high addresses
+ * directly.
+ */
+void * ioremap(unsigned long phys_addr, unsigned long size)
+{
+	void * addr;
+	struct vm_struct * area;
+
+	if (phys_addr < virt_to_phys(high_memory))
+		return phys_to_virt(phys_addr);
+	if (phys_addr & ~PAGE_MASK)
+		return NULL;
+	size = PAGE_ALIGN(size);
+	if (!size || size > phys_addr + size)
+		return NULL;
+	area = get_vm_area(size);
+	if (!area)
+		return NULL;
+	addr = area->addr;
+	if (remap_area_pages(VMALLOC_VMADDR(addr), phys_addr, size)) {
+		vfree(addr);
+		return NULL;
+	}
+	return addr;
+}
+
+void iounmap(void *addr)
+{
+	if (addr > high_memory)
+		return vfree(addr);
+}