Merge HEAD from master.kernel.org:/pub/scm/linux/kernel/git/paulus/ppc64-2.6

55 files changed, 1920 insertions, 1619 deletions

diff --git a/arch/ppc64/Kconfig b/arch/ppc64/Kconfig
index 2ce8783..13b262f 100644
--- a/arch/ppc64/Kconfig
+++ b/arch/ppc64/Kconfig
@@ -302,12 +302,6 @@ config GENERIC_HARDIRQS
 	bool
 	default y
 
-config MSCHUNKS
-	bool
-	depends on PPC_ISERIES
-	default y
-
-
 config PPC_RTAS
 	bool
 	depends on PPC_PSERIES || PPC_BPA
@@ -350,13 +344,46 @@ config SECCOMP
 
 	  If unsure, say Y. Only embedded should say N here.
 
+source "fs/Kconfig.binfmt"
+
+config HOTPLUG_CPU
+	bool "Support for hot-pluggable CPUs"
+	depends on SMP && EXPERIMENTAL && (PPC_PSERIES || PPC_PMAC)
+	select HOTPLUG
+	---help---
+	  Say Y here to be able to turn CPUs off and on.
+
+	  Say N if you are unsure.
+
+config PROC_DEVICETREE
+	bool "Support for Open Firmware device tree in /proc"
+	depends on !PPC_ISERIES
+	help
+	  This option adds a device-tree directory under /proc which contains
+	  an image of the device tree that the kernel copies from Open
+	  Firmware. If unsure, say Y here.
+
+config CMDLINE_BOOL
+	bool "Default bootloader kernel arguments"
+	depends on !PPC_ISERIES
+
+config CMDLINE
+	string "Initial kernel command string"
+	depends on CMDLINE_BOOL
+	default "console=ttyS0,9600 console=tty0 root=/dev/sda2"
+	help
+	  On some platforms, there is currently no way for the boot loader to
+	  pass arguments to the kernel. For these platforms, you can supply
+	  some command-line options at build time by entering them here.  In
+	  most cases you will need to specify the root device here.
+
 endmenu
 
 config ISA_DMA_API
 	bool
 	default y
 
-menu "General setup"
+menu "Bus Options"
 
 config ISA
 	bool
@@ -389,45 +416,12 @@ config PCI_DOMAINS
 	bool
 	default PCI
 
-source "fs/Kconfig.binfmt"
-
 source "drivers/pci/Kconfig"
 
-config HOTPLUG_CPU
-	bool "Support for hot-pluggable CPUs"
-	depends on SMP && EXPERIMENTAL && (PPC_PSERIES || PPC_PMAC)
-	select HOTPLUG
-	---help---
-	  Say Y here to be able to turn CPUs off and on.
-
-	  Say N if you are unsure.
-
 source "drivers/pcmcia/Kconfig"
 
 source "drivers/pci/hotplug/Kconfig"
 
-config PROC_DEVICETREE
-	bool "Support for Open Firmware device tree in /proc"
-	depends on !PPC_ISERIES
-	help
-	  This option adds a device-tree directory under /proc which contains
-	  an image of the device tree that the kernel copies from Open
-	  Firmware. If unsure, say Y here.
-
-config CMDLINE_BOOL
-	bool "Default bootloader kernel arguments"
-	depends on !PPC_ISERIES
-
-config CMDLINE
-	string "Initial kernel command string"
-	depends on CMDLINE_BOOL
-	default "console=ttyS0,9600 console=tty0 root=/dev/sda2"
-	help
-	  On some platforms, there is currently no way for the boot loader to
-	  pass arguments to the kernel. For these platforms, you can supply
-	  some command-line options at build time by entering them here.  In
-	  most cases you will need to specify the root device here.
-
 endmenu
 
 source "net/Kconfig"
diff --git a/arch/ppc64/boot/Makefile b/arch/ppc64/boot/Makefile
index 683b2d4..2c5f5e7 100644
--- a/arch/ppc64/boot/Makefile
+++ b/arch/ppc64/boot/Makefile
@@ -22,8 +22,8 @@
 
 
 HOSTCC		:= gcc
-BOOTCFLAGS	:= $(HOSTCFLAGS) $(LINUXINCLUDE) -fno-builtin 
-BOOTAFLAGS	:= -D__ASSEMBLY__ $(BOOTCFLAGS) -traditional
+BOOTCFLAGS	:= $(HOSTCFLAGS) -fno-builtin -nostdinc -isystem $(shell $(CROSS32CC) -print-file-name=include)
+BOOTAFLAGS	:= -D__ASSEMBLY__ $(BOOTCFLAGS) -traditional -nostdinc
 BOOTLFLAGS	:= -Ttext 0x00400000 -e _start -T $(srctree)/$(src)/zImage.lds
 OBJCOPYFLAGS    := contents,alloc,load,readonly,data
 
diff --git a/arch/ppc64/boot/addnote.c b/arch/ppc64/boot/addnote.c
index 719663a..8041a98 100644
--- a/arch/ppc64/boot/addnote.c
+++ b/arch/ppc64/boot/addnote.c
@@ -157,7 +157,7 @@ main(int ac, char **av)
 	PUT_32BE(ns, strlen(arch) + 1);
 	PUT_32BE(ns + 4, N_DESCR * 4);
 	PUT_32BE(ns + 8, 0x1275);
-	strcpy(&buf[ns + 12], arch);
+	strcpy((char *) &buf[ns + 12], arch);
 	ns += 12 + strlen(arch) + 1;
 	for (i = 0; i < N_DESCR; ++i, ns += 4)
 		PUT_32BE(ns, descr[i]);
@@ -172,7 +172,7 @@ main(int ac, char **av)
 	PUT_32BE(ns, strlen(rpaname) + 1);
 	PUT_32BE(ns + 4, sizeof(rpanote));
 	PUT_32BE(ns + 8, 0x12759999);
-	strcpy(&buf[ns + 12], rpaname);
+	strcpy((char *) &buf[ns + 12], rpaname);
 	ns += 12 + ROUNDUP(strlen(rpaname) + 1);
 	for (i = 0; i < N_RPA_DESCR; ++i, ns += 4)
 		PUT_32BE(ns, rpanote[i]);
diff --git a/arch/ppc64/boot/crt0.S b/arch/ppc64/boot/crt0.S
index 04d3e74..3861e7f 100644
--- a/arch/ppc64/boot/crt0.S
+++ b/arch/ppc64/boot/crt0.S
@@ -9,7 +9,7 @@
  * NOTE: this code runs in 32 bit mode and is packaged as ELF32.
  */
 
-#include <asm/ppc_asm.h>
+#include "ppc_asm.h"
 
 	.text
 	.globl	_start
diff --git a/arch/ppc64/boot/div64.S b/arch/ppc64/boot/div64.S
index 38f7e46..722f360 100644
--- a/arch/ppc64/boot/div64.S
+++ b/arch/ppc64/boot/div64.S
@@ -13,7 +13,7 @@
  * as published by the Free Software Foundation; either version
  * 2 of the License, or (at your option) any later version.
  */
-#include <asm/ppc_asm.h>
+#include "ppc_asm.h"
 
 	.globl __div64_32
 __div64_32:
diff --git a/arch/ppc64/boot/elf.h b/arch/ppc64/boot/elf.h
new file mode 100644
index 0000000..d4828fc
--- /dev/null
+++ b/arch/ppc64/boot/elf.h
@@ -0,0 +1,149 @@
+#ifndef _PPC_BOOT_ELF_H_
+#define _PPC_BOOT_ELF_H_
+
+/* 32-bit ELF base types. */
+typedef unsigned int Elf32_Addr;
+typedef unsigned short Elf32_Half;
+typedef unsigned int Elf32_Off;
+typedef signed int Elf32_Sword;
+typedef unsigned int Elf32_Word;
+
+/* 64-bit ELF base types. */
+typedef unsigned long long Elf64_Addr;
+typedef unsigned short Elf64_Half;
+typedef signed short Elf64_SHalf;
+typedef unsigned long long Elf64_Off;
+typedef signed int Elf64_Sword;
+typedef unsigned int Elf64_Word;
+typedef unsigned long long Elf64_Xword;
+typedef signed long long Elf64_Sxword;
+
+/* These constants are for the segment types stored in the image headers */
+#define PT_NULL    0
+#define PT_LOAD    1
+#define PT_DYNAMIC 2
+#define PT_INTERP  3
+#define PT_NOTE    4
+#define PT_SHLIB   5
+#define PT_PHDR    6
+#define PT_TLS     7		/* Thread local storage segment */
+#define PT_LOOS    0x60000000	/* OS-specific */
+#define PT_HIOS    0x6fffffff	/* OS-specific */
+#define PT_LOPROC  0x70000000
+#define PT_HIPROC  0x7fffffff
+#define PT_GNU_EH_FRAME		0x6474e550
+
+#define PT_GNU_STACK	(PT_LOOS + 0x474e551)
+
+/* These constants define the different elf file types */
+#define ET_NONE   0
+#define ET_REL    1
+#define ET_EXEC   2
+#define ET_DYN    3
+#define ET_CORE   4
+#define ET_LOPROC 0xff00
+#define ET_HIPROC 0xffff
+
+/* These constants define the various ELF target machines */
+#define EM_NONE  0
+#define EM_PPC	       20	/* PowerPC */
+#define EM_PPC64       21	/* PowerPC64 */
+
+#define EI_NIDENT	16
+
+typedef struct elf32_hdr {
+	unsigned char e_ident[EI_NIDENT];
+	Elf32_Half e_type;
+	Elf32_Half e_machine;
+	Elf32_Word e_version;
+	Elf32_Addr e_entry;	/* Entry point */
+	Elf32_Off e_phoff;
+	Elf32_Off e_shoff;
+	Elf32_Word e_flags;
+	Elf32_Half e_ehsize;
+	Elf32_Half e_phentsize;
+	Elf32_Half e_phnum;
+	Elf32_Half e_shentsize;
+	Elf32_Half e_shnum;
+	Elf32_Half e_shstrndx;
+} Elf32_Ehdr;
+
+typedef struct elf64_hdr {
+	unsigned char e_ident[16];	/* ELF "magic number" */
+	Elf64_Half e_type;
+	Elf64_Half e_machine;
+	Elf64_Word e_version;
+	Elf64_Addr e_entry;	/* Entry point virtual address */
+	Elf64_Off e_phoff;	/* Program header table file offset */
+	Elf64_Off e_shoff;	/* Section header table file offset */
+	Elf64_Word e_flags;
+	Elf64_Half e_ehsize;
+	Elf64_Half e_phentsize;
+	Elf64_Half e_phnum;
+	Elf64_Half e_shentsize;
+	Elf64_Half e_shnum;
+	Elf64_Half e_shstrndx;
+} Elf64_Ehdr;
+
+/* These constants define the permissions on sections in the program
+   header, p_flags. */
+#define PF_R		0x4
+#define PF_W		0x2
+#define PF_X		0x1
+
+typedef struct elf32_phdr {
+	Elf32_Word p_type;
+	Elf32_Off p_offset;
+	Elf32_Addr p_vaddr;
+	Elf32_Addr p_paddr;
+	Elf32_Word p_filesz;
+	Elf32_Word p_memsz;
+	Elf32_Word p_flags;
+	Elf32_Word p_align;
+} Elf32_Phdr;
+
+typedef struct elf64_phdr {
+	Elf64_Word p_type;
+	Elf64_Word p_flags;
+	Elf64_Off p_offset;	/* Segment file offset */
+	Elf64_Addr p_vaddr;	/* Segment virtual address */
+	Elf64_Addr p_paddr;	/* Segment physical address */
+	Elf64_Xword p_filesz;	/* Segment size in file */
+	Elf64_Xword p_memsz;	/* Segment size in memory */
+	Elf64_Xword p_align;	/* Segment alignment, file & memory */
+} Elf64_Phdr;
+
+#define	EI_MAG0		0	/* e_ident[] indexes */
+#define	EI_MAG1		1
+#define	EI_MAG2		2
+#define	EI_MAG3		3
+#define	EI_CLASS	4
+#define	EI_DATA		5
+#define	EI_VERSION	6
+#define	EI_OSABI	7
+#define	EI_PAD		8
+
+#define	ELFMAG0		0x7f	/* EI_MAG */
+#define	ELFMAG1		'E'
+#define	ELFMAG2		'L'
+#define	ELFMAG3		'F'
+#define	ELFMAG		"\177ELF"
+#define	SELFMAG		4
+
+#define	ELFCLASSNONE	0	/* EI_CLASS */
+#define	ELFCLASS32	1
+#define	ELFCLASS64	2
+#define	ELFCLASSNUM	3
+
+#define ELFDATANONE	0	/* e_ident[EI_DATA] */
+#define ELFDATA2LSB	1
+#define ELFDATA2MSB	2
+
+#define EV_NONE		0	/* e_version, EI_VERSION */
+#define EV_CURRENT	1
+#define EV_NUM		2
+
+#define ELFOSABI_NONE	0
+#define ELFOSABI_LINUX	3
+
+#endif				/* _PPC_BOOT_ELF_H_ */
diff --git a/arch/ppc64/boot/main.c b/arch/ppc64/boot/main.c
index 199d980..99e68cf 100644
--- a/arch/ppc64/boot/main.c
+++ b/arch/ppc64/boot/main.c
@@ -8,36 +8,28 @@
  * as published by the Free Software Foundation; either version
  * 2 of the License, or (at your option) any later version.
  */
-#include "ppc32-types.h"
+#include <stdarg.h>
+#include <stddef.h>
+#include "elf.h"
+#include "page.h"
+#include "string.h"
+#include "stdio.h"
+#include "prom.h"
 #include "zlib.h"
-#include <linux/elf.h>
-#include <linux/string.h>
-#include <asm/processor.h>
-#include <asm/page.h>
-
-extern void *finddevice(const char *);
-extern int getprop(void *, const char *, void *, int);
-extern void printf(const char *fmt, ...);
-extern int sprintf(char *buf, const char *fmt, ...);
-void gunzip(void *, int, unsigned char *, int *);
-void *claim(unsigned int, unsigned int, unsigned int);
-void flush_cache(void *, unsigned long);
-void pause(void);
-extern void exit(void);
-
-unsigned long strlen(const char *s);
-void *memmove(void *dest, const void *src, unsigned long n);
-void *memcpy(void *dest, const void *src, unsigned long n);
+
+static void gunzip(void *, int, unsigned char *, int *);
+extern void flush_cache(void *, unsigned long);
+
 
 /* Value picked to match that used by yaboot */
 #define PROG_START	0x01400000
 #define RAM_END		(256<<20) // Fixme: use OF */
 
-char *avail_ram;
-char *begin_avail, *end_avail;
-char *avail_high;
-unsigned int heap_use;
-unsigned int heap_max;
+static char *avail_ram;
+static char *begin_avail, *end_avail;
+static char *avail_high;
+static unsigned int heap_use;
+static unsigned int heap_max;
 
 extern char _start[];
 extern char _vmlinux_start[];
@@ -52,9 +44,9 @@ struct addr_range {
 	unsigned long size;
 	unsigned long memsize;
 };
-struct addr_range vmlinux = {0, 0, 0};
-struct addr_range vmlinuz = {0, 0, 0};
-struct addr_range initrd  = {0, 0, 0};
+static struct addr_range vmlinux = {0, 0, 0};
+static struct addr_range vmlinuz = {0, 0, 0};
+static struct addr_range initrd  = {0, 0, 0};
 
 static char scratch[128<<10];	/* 128kB of scratch space for gunzip */
 
@@ -64,13 +56,6 @@ typedef void (*kernel_entry_t)( unsigned long,
 				void *);
 
 
-int (*prom)(void *);
-
-void *chosen_handle;
-void *stdin;
-void *stdout;
-void *stderr;
-
 #undef DEBUG
 
 static unsigned long claim_base = PROG_START;
@@ -277,7 +262,7 @@ void zfree(void *x, void *addr, unsigned nb)
 
 #define DEFLATED	8
 
-void gunzip(void *dst, int dstlen, unsigned char *src, int *lenp)
+static void gunzip(void *dst, int dstlen, unsigned char *src, int *lenp)
 {
 	z_stream s;
 	int r, i, flags;
diff --git a/arch/ppc64/boot/page.h b/arch/ppc64/boot/page.h
new file mode 100644
index 0000000..14eca30f
--- /dev/null
+++ b/arch/ppc64/boot/page.h
@@ -0,0 +1,34 @@
+#ifndef _PPC_BOOT_PAGE_H
+#define _PPC_BOOT_PAGE_H
+/*
+ * Copyright (C) 2001 PPC64 Team, IBM Corp
+ *
+ * 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.
+ */
+
+#ifdef __ASSEMBLY__
+#define ASM_CONST(x) x
+#else
+#define __ASM_CONST(x) x##UL
+#define ASM_CONST(x) __ASM_CONST(x)
+#endif
+
+/* PAGE_SHIFT determines the page size */
+#define PAGE_SHIFT	12
+#define PAGE_SIZE	(ASM_CONST(1) << PAGE_SHIFT)
+#define PAGE_MASK	(~(PAGE_SIZE-1))
+
+/* align addr on a size boundary - adjust address up/down if needed */
+#define _ALIGN_UP(addr,size)	(((addr)+((size)-1))&(~((size)-1)))
+#define _ALIGN_DOWN(addr,size)	((addr)&(~((size)-1)))
+
+/* align addr on a size boundary - adjust address up if needed */
+#define _ALIGN(addr,size)     _ALIGN_UP(addr,size)
+
+/* to align the pointer to the (next) page boundary */
+#define PAGE_ALIGN(addr)	_ALIGN(addr, PAGE_SIZE)
+
+#endif				/* _PPC_BOOT_PAGE_H */
diff --git a/arch/ppc64/boot/ppc32-types.h b/arch/ppc64/boot/ppc32-types.h
deleted file mode 100644
index f7b8884..0000000
--- a/arch/ppc64/boot/ppc32-types.h
+++ /dev/null
@@ -1,36 +0,0 @@
-#ifndef _PPC64_TYPES_H
-#define _PPC64_TYPES_H
-
-typedef __signed__ char __s8;
-typedef unsigned char __u8;
-
-typedef __signed__ short __s16;
-typedef unsigned short __u16;
-
-typedef __signed__ int __s32;
-typedef unsigned int __u32;
-
-typedef __signed__ long long __s64;
-typedef unsigned long long __u64;
-
-typedef signed char s8;
-typedef unsigned char u8;
-
-typedef signed short s16;
-typedef unsigned short u16;
-
-typedef signed int s32;
-typedef unsigned int u32;
-
-typedef signed long long s64;
-typedef unsigned long long u64;
-
-typedef struct {
-	__u32 u[4];
-} __attribute((aligned(16))) __vector128;
-
-#define BITS_PER_LONG 32
-
-typedef __vector128 vector128;
-
-#endif /* _PPC64_TYPES_H */
diff --git a/arch/ppc64/boot/ppc_asm.h b/arch/ppc64/boot/ppc_asm.h
new file mode 100644
index 0000000..1c2c281
--- /dev/null
+++ b/arch/ppc64/boot/ppc_asm.h
@@ -0,0 +1,62 @@
+#ifndef _PPC64_PPC_ASM_H
+#define _PPC64_PPC_ASM_H
+/*
+ *
+ * Definitions used by various bits of low-level assembly code on PowerPC.
+ *
+ * Copyright (C) 1995-1999 Gary Thomas, Paul Mackerras, Cort Dougan.
+ *
+ *  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.
+ */
+
+/* Condition Register Bit Fields */
+
+#define	cr0	0
+#define	cr1	1
+#define	cr2	2
+#define	cr3	3
+#define	cr4	4
+#define	cr5	5
+#define	cr6	6
+#define	cr7	7
+
+
+/* General Purpose Registers (GPRs) */
+
+#define	r0	0
+#define	r1	1
+#define	r2	2
+#define	r3	3
+#define	r4	4
+#define	r5	5
+#define	r6	6
+#define	r7	7
+#define	r8	8
+#define	r9	9
+#define	r10	10
+#define	r11	11
+#define	r12	12
+#define	r13	13
+#define	r14	14
+#define	r15	15
+#define	r16	16
+#define	r17	17
+#define	r18	18
+#define	r19	19
+#define	r20	20
+#define	r21	21
+#define	r22	22
+#define	r23	23
+#define	r24	24
+#define	r25	25
+#define	r26	26
+#define	r27	27
+#define	r28	28
+#define	r29	29
+#define	r30	30
+#define	r31	31
+
+#endif /* _PPC64_PPC_ASM_H */
diff --git a/arch/ppc64/boot/prom.c b/arch/ppc64/boot/prom.c
index 5e48b80..4bea2f4 100644
--- a/arch/ppc64/boot/prom.c
+++ b/arch/ppc64/boot/prom.c
@@ -7,43 +7,19 @@
  * 2 of the License, or (at your option) any later version.
  */
 #include <stdarg.h>
-#include <linux/types.h>
-#include <linux/string.h>
-#include <linux/ctype.h>
-
-extern __u32 __div64_32(unsigned long long *dividend, __u32 divisor);
-
-/* The unnecessary pointer compare is there
- * to check for type safety (n must be 64bit)
- */
-# define do_div(n,base) ({				\
-	__u32 __base = (base);			\
-	__u32 __rem;					\
-	(void)(((typeof((n)) *)0) == ((unsigned long long *)0));	\
-	if (((n) >> 32) == 0) {			\
-		__rem = (__u32)(n) % __base;		\
-		(n) = (__u32)(n) / __base;		\
-	} else 						\
-		__rem = __div64_32(&(n), __base);	\
-	__rem;						\
- })
+#include <stddef.h>
+#include "string.h"
+#include "stdio.h"
+#include "prom.h"
 
 int (*prom)(void *);
 
 void *chosen_handle;
+
 void *stdin;
 void *stdout;
 void *stderr;
 
-void exit(void);
-void *finddevice(const char *name);
-int getprop(void *phandle, const char *name, void *buf, int buflen);
-void chrpboot(int a1, int a2, void *prom);	/* in main.c */
-
-int printf(char *fmt, ...);
-
-/* there is no convenient header to get this from...  -- paulus */
-extern unsigned long strlen(const char *);
 
 int
 write(void *handle, void *ptr, int nb)
@@ -210,107 +186,6 @@ fputs(char *str, void *f)
 	return write(f, str, n) == n? 0: -1;
 }
 
-int
-readchar(void)
-{
-	char ch;
-
-	for (;;) {
-		switch (read(stdin, &ch, 1)) {
-		case 1:
-			return ch;
-		case -1:
-			printf("read(stdin) returned -1\r\n");
-			return -1;
-		}
-	}
-}
-
-static char line[256];
-static char *lineptr;
-static int lineleft;
-
-int
-getchar(void)
-{
-	int c;
-
-	if (lineleft == 0) {
-		lineptr = line;
-		for (;;) {
-			c = readchar();
-			if (c == -1 || c == 4)
-				break;
-			if (c == '\r' || c == '\n') {
-				*lineptr++ = '\n';
-				putchar('\n');
-				break;
-			}
-			switch (c) {
-			case 0177:
-			case '\b':
-				if (lineptr > line) {
-					putchar('\b');
-					putchar(' ');
-					putchar('\b');
-					--lineptr;
-				}
-				break;
-			case 'U' & 0x1F:
-				while (lineptr > line) {
-					putchar('\b');
-					putchar(' ');
-					putchar('\b');
-					--lineptr;
-				}
-				break;
-			default:
-				if (lineptr >= &line[sizeof(line) - 1])
-					putchar('\a');
-				else {
-					putchar(c);
-					*lineptr++ = c;
-				}
-			}
-		}
-		lineleft = lineptr - line;
-		lineptr = line;
-	}
-	if (lineleft == 0)
-		return -1;
-	--lineleft;
-	return *lineptr++;
-}
-
-
-
-/* String functions lifted from lib/vsprintf.c and lib/ctype.c */
-unsigned char _ctype[] = {
-_C,_C,_C,_C,_C,_C,_C,_C,			/* 0-7 */
-_C,_C|_S,_C|_S,_C|_S,_C|_S,_C|_S,_C,_C,		/* 8-15 */
-_C,_C,_C,_C,_C,_C,_C,_C,			/* 16-23 */
-_C,_C,_C,_C,_C,_C,_C,_C,			/* 24-31 */
-_S|_SP,_P,_P,_P,_P,_P,_P,_P,			/* 32-39 */
-_P,_P,_P,_P,_P,_P,_P,_P,			/* 40-47 */
-_D,_D,_D,_D,_D,_D,_D,_D,			/* 48-55 */
-_D,_D,_P,_P,_P,_P,_P,_P,			/* 56-63 */
-_P,_U|_X,_U|_X,_U|_X,_U|_X,_U|_X,_U|_X,_U,	/* 64-71 */
-_U,_U,_U,_U,_U,_U,_U,_U,			/* 72-79 */
-_U,_U,_U,_U,_U,_U,_U,_U,			/* 80-87 */
-_U,_U,_U,_P,_P,_P,_P,_P,			/* 88-95 */
-_P,_L|_X,_L|_X,_L|_X,_L|_X,_L|_X,_L|_X,_L,	/* 96-103 */
-_L,_L,_L,_L,_L,_L,_L,_L,			/* 104-111 */
-_L,_L,_L,_L,_L,_L,_L,_L,			/* 112-119 */
-_L,_L,_L,_P,_P,_P,_P,_C,			/* 120-127 */
-0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,		/* 128-143 */
-0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,		/* 144-159 */
-_S|_SP,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,   /* 160-175 */
-_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,_P,       /* 176-191 */
-_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,_U,       /* 192-207 */
-_U,_U,_U,_U,_U,_U,_U,_P,_U,_U,_U,_U,_U,_U,_U,_L,       /* 208-223 */
-_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,_L,       /* 224-239 */
-_L,_L,_L,_L,_L,_L,_L,_P,_L,_L,_L,_L,_L,_L,_L,_L};      /* 240-255 */
-
 size_t strnlen(const char * s, size_t count)
 {
 	const char *sc;
@@ -320,44 +195,30 @@ size_t strnlen(const char * s, size_t count)
 	return sc - s;
 }
 
-unsigned long simple_strtoul(const char *cp,char **endp,unsigned int base)
-{
-	unsigned long result = 0,value;
+extern unsigned int __div64_32(unsigned long long *dividend,
+			       unsigned int divisor);
 
-	if (!base) {
-		base = 10;
-		if (*cp == '0') {
-			base = 8;
-			cp++;
-			if ((*cp == 'x') && isxdigit(cp[1])) {
-				cp++;
-				base = 16;
-			}
-		}
-	}
-	while (isxdigit(*cp) &&
-	       (value = isdigit(*cp) ? *cp-'0' : toupper(*cp)-'A'+10) < base) {
-		result = result*base + value;
-		cp++;
-	}
-	if (endp)
-		*endp = (char *)cp;
-	return result;
-}
-
-long simple_strtol(const char *cp,char **endp,unsigned int base)
-{
-	if(*cp=='-')
-		return -simple_strtoul(cp+1,endp,base);
-	return simple_strtoul(cp,endp,base);
-}
+/* The unnecessary pointer compare is there
+ * to check for type safety (n must be 64bit)
+ */
+# define do_div(n,base) ({						\
+	unsigned int __base = (base);					\
+	unsigned int __rem;						\
+	(void)(((typeof((n)) *)0) == ((unsigned long long *)0));	\
+	if (((n) >> 32) == 0) {						\
+		__rem = (unsigned int)(n) % __base;			\
+		(n) = (unsigned int)(n) / __base;			\
+	} else								\
+		__rem = __div64_32(&(n), __base);			\
+	__rem;								\
+ })
 
 static int skip_atoi(const char **s)
 {
-	int i=0;
+	int i, c;
 
-	while (isdigit(**s))
-		i = i*10 + *((*s)++) - '0';
+	for (i = 0; '0' <= (c = **s) && c <= '9'; ++*s)
+		i = i*10 + c - '0';
 	return i;
 }
 
@@ -436,9 +297,6 @@ static char * number(char * str, unsigned long long num, int base, int size, int
 	return str;
 }
 
-/* Forward decl. needed for IP address printing stuff... */
-int sprintf(char * buf, const char *fmt, ...);
-
 int vsprintf(char *buf, const char *fmt, va_list args)
 {
 	int len;
@@ -477,7 +335,7 @@ int vsprintf(char *buf, const char *fmt, va_list args)
 		
 		/* get field width */
 		field_width = -1;
-		if (isdigit(*fmt))
+		if ('0' <= *fmt && *fmt <= '9')
 			field_width = skip_atoi(&fmt);
 		else if (*fmt == '*') {
 			++fmt;
@@ -493,7 +351,7 @@ int vsprintf(char *buf, const char *fmt, va_list args)
 		precision = -1;
 		if (*fmt == '.') {
 			++fmt;	
-			if (isdigit(*fmt))
+			if ('0' <= *fmt && *fmt <= '9')
 				precision = skip_atoi(&fmt);
 			else if (*fmt == '*') {
 				++fmt;
@@ -628,7 +486,7 @@ int sprintf(char * buf, const char *fmt, ...)
 static char sprint_buf[1024];
 
 int
-printf(char *fmt, ...)
+printf(const char *fmt, ...)
 {
 	va_list args;
 	int n;
diff --git a/arch/ppc64/boot/prom.h b/arch/ppc64/boot/prom.h
new file mode 100644
index 0000000..96ab5ae
--- /dev/null
+++ b/arch/ppc64/boot/prom.h
@@ -0,0 +1,18 @@
+#ifndef _PPC_BOOT_PROM_H_
+#define _PPC_BOOT_PROM_H_
+
+extern int (*prom) (void *);
+extern void *chosen_handle;
+
+extern void *stdin;
+extern void *stdout;
+extern void *stderr;
+
+extern int write(void *handle, void *ptr, int nb);
+extern int read(void *handle, void *ptr, int nb);
+extern void exit(void);
+extern void pause(void);
+extern void *finddevice(const char *);
+extern void *claim(unsigned long virt, unsigned long size, unsigned long align);
+extern int getprop(void *phandle, const char *name, void *buf, int buflen);
+#endif				/* _PPC_BOOT_PROM_H_ */
diff --git a/arch/ppc64/boot/stdio.h b/arch/ppc64/boot/stdio.h
new file mode 100644
index 0000000..24bd3a8
--- /dev/null
+++ b/arch/ppc64/boot/stdio.h
@@ -0,0 +1,16 @@
+#ifndef _PPC_BOOT_STDIO_H_
+#define _PPC_BOOT_STDIO_H_
+
+extern int printf(const char *fmt, ...);
+
+extern int sprintf(char *buf, const char *fmt, ...);
+
+extern int vsprintf(char *buf, const char *fmt, va_list args);
+
+extern int putc(int c, void *f);
+extern int putchar(int c);
+extern int getchar(void);
+
+extern int fputs(char *str, void *f);
+
+#endif				/* _PPC_BOOT_STDIO_H_ */
diff --git a/arch/ppc64/boot/string.S b/arch/ppc64/boot/string.S
index ba5f2d2..7ade87a 100644
--- a/arch/ppc64/boot/string.S
+++ b/arch/ppc64/boot/string.S
@@ -9,7 +9,7 @@
  * NOTE: this code runs in 32 bit mode and is packaged as ELF32.
  */
 
-#include <asm/ppc_asm.h>
+#include "ppc_asm.h"
 
 	.text
 	.globl	strcpy
diff --git a/arch/ppc64/boot/string.h b/arch/ppc64/boot/string.h
new file mode 100644
index 0000000..9289258
--- /dev/null
+++ b/arch/ppc64/boot/string.h
@@ -0,0 +1,16 @@
+#ifndef _PPC_BOOT_STRING_H_
+#define _PPC_BOOT_STRING_H_
+
+extern char *strcpy(char *dest, const char *src);
+extern char *strncpy(char *dest, const char *src, size_t n);
+extern char *strcat(char *dest, const char *src);
+extern int strcmp(const char *s1, const char *s2);
+extern size_t strlen(const char *s);
+extern size_t strnlen(const char *s, size_t count);
+
+extern void *memset(void *s, int c, size_t n);
+extern void *memmove(void *dest, const void *src, unsigned long n);
+extern void *memcpy(void *dest, const void *src, unsigned long n);
+extern int memcmp(const void *s1, const void *s2, size_t n);
+
+#endif	/* _PPC_BOOT_STRING_H_ */
diff --git a/arch/ppc64/boot/zlib.c b/arch/ppc64/boot/zlib.c
index 78837e8..0d910cd 100644
--- a/arch/ppc64/boot/zlib.c
+++ b/arch/ppc64/boot/zlib.c
@@ -107,7 +107,7 @@ extern void *memcpy(void *, const void *, unsigned long);
 
 /* Diagnostic functions */
 #ifdef DEBUG_ZLIB
-#  include <stdio.h>
+#  include "stdio.h"
 #  ifndef verbose
 #    define verbose 0
 #  endif
diff --git a/arch/ppc64/configs/iSeries_defconfig b/arch/ppc64/configs/iSeries_defconfig
index 394ba18..219c667 100644
--- a/arch/ppc64/configs/iSeries_defconfig
+++ b/arch/ppc64/configs/iSeries_defconfig
@@ -99,7 +99,6 @@ CONFIG_HZ_100=y
 # CONFIG_HZ_1000 is not set
 CONFIG_HZ=100
 CONFIG_GENERIC_HARDIRQS=y
-CONFIG_MSCHUNKS=y
 CONFIG_LPARCFG=y
 CONFIG_SECCOMP=y
 CONFIG_ISA_DMA_API=y
diff --git a/arch/ppc64/kernel/LparData.c b/arch/ppc64/kernel/LparData.c
index 1c11031..0a9c23c 100644
--- a/arch/ppc64/kernel/LparData.c
+++ b/arch/ppc64/kernel/LparData.c
@@ -51,6 +51,17 @@ struct HvReleaseData hvReleaseData = {
 		0xf4, 0x4b, 0xf6, 0xf4 },
 };
 
+/*
+ * The NACA.  The first dword of the naca is required by the iSeries
+ * hypervisor to point to itVpdAreas.  The hypervisor finds the NACA
+ * through the pointer in hvReleaseData.
+ */
+struct naca_struct naca = {
+	.xItVpdAreas = &itVpdAreas,
+	.xRamDisk = 0,
+	.xRamDiskSize = 0,
+};
+
 extern void system_reset_iSeries(void);
 extern void machine_check_iSeries(void);
 extern void data_access_iSeries(void);
@@ -214,29 +225,3 @@ struct ItVpdAreas itVpdAreas = {
 		0,0
 	}
 };
-
-struct msChunks msChunks;
-EXPORT_SYMBOL(msChunks);
-
-/* Depending on whether this is called from iSeries or pSeries setup
- * code, the location of the msChunks struct may or may not have
- * to be reloc'd, so we force the caller to do that for us by passing
- * in a pointer to the structure.
- */
-unsigned long
-msChunks_alloc(unsigned long mem, unsigned long num_chunks, unsigned long chunk_size)
-{
-	unsigned long offset = reloc_offset();
-	struct msChunks *_msChunks = PTRRELOC(&msChunks);
-
-	_msChunks->num_chunks  = num_chunks;
-	_msChunks->chunk_size  = chunk_size;
-	_msChunks->chunk_shift = __ilog2(chunk_size);
-	_msChunks->chunk_mask  = (1UL<<_msChunks->chunk_shift)-1;
-
-	mem = _ALIGN(mem, sizeof(msChunks_entry));
-	_msChunks->abs = (msChunks_entry *)(mem + offset);
-	mem += num_chunks * sizeof(msChunks_entry);
-
-	return mem;
-}
diff --git a/arch/ppc64/kernel/Makefile b/arch/ppc64/kernel/Makefile
index 2ecccb6..f4b3bfc 100644
--- a/arch/ppc64/kernel/Makefile
+++ b/arch/ppc64/kernel/Makefile
@@ -11,7 +11,7 @@ obj-y               :=	setup.o entry.o traps.o irq.o idle.o dma.o \
 			udbg.o binfmt_elf32.o sys_ppc32.o ioctl32.o \
 			ptrace32.o signal32.o rtc.o init_task.o \
 			lmb.o cputable.o cpu_setup_power4.o idle_power4.o \
-			iommu.o sysfs.o vdso.o pmc.o
+			iommu.o sysfs.o vdso.o pmc.o firmware.o
 obj-y += vdso32/ vdso64/
 
 obj-$(CONFIG_PPC_OF) +=	of_device.o
@@ -50,7 +50,10 @@ obj-$(CONFIG_LPARCFG)		+= lparcfg.o
 obj-$(CONFIG_HVC_CONSOLE)	+= hvconsole.o
 obj-$(CONFIG_BOOTX_TEXT)	+= btext.o
 obj-$(CONFIG_HVCS)		+= hvcserver.o
-obj-$(CONFIG_IBMVIO)		+= vio.o
+
+vio-obj-$(CONFIG_PPC_PSERIES)	+= pSeries_vio.o
+vio-obj-$(CONFIG_PPC_ISERIES)	+= iSeries_vio.o
+obj-$(CONFIG_IBMVIO)		+= vio.o $(vio-obj-y)
 obj-$(CONFIG_XICS)		+= xics.o
 obj-$(CONFIG_MPIC)		+= mpic.o
 
diff --git a/arch/ppc64/kernel/asm-offsets.c b/arch/ppc64/kernel/asm-offsets.c
index abb9e5b..17e35d0 100644
--- a/arch/ppc64/kernel/asm-offsets.c
+++ b/arch/ppc64/kernel/asm-offsets.c
@@ -94,7 +94,8 @@ int main(void)
 	DEFINE(PACASLBCACHEPTR, offsetof(struct paca_struct, slb_cache_ptr));
 	DEFINE(PACACONTEXTID, offsetof(struct paca_struct, context.id));
 #ifdef CONFIG_HUGETLB_PAGE
-	DEFINE(PACAHTLBSEGS, offsetof(struct paca_struct, context.htlb_segs));
+	DEFINE(PACALOWHTLBAREAS, offsetof(struct paca_struct, context.low_htlb_areas));
+	DEFINE(PACAHIGHHTLBAREAS, offsetof(struct paca_struct, context.high_htlb_areas));
 #endif /* CONFIG_HUGETLB_PAGE */
 	DEFINE(PACADEFAULTDECR, offsetof(struct paca_struct, default_decr));
         DEFINE(PACA_EXGEN, offsetof(struct paca_struct, exgen));
diff --git a/arch/ppc64/kernel/cputable.c b/arch/ppc64/kernel/cputable.c
index 77cec42..4847f2a 100644
--- a/arch/ppc64/kernel/cputable.c
+++ b/arch/ppc64/kernel/cputable.c
@@ -5,7 +5,7 @@
  *
  *  Modifications for ppc64:
  *      Copyright (C) 2003 Dave Engebretsen <engebret@us.ibm.com>
- * 
+ *
  *  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
@@ -60,7 +60,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power3,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Power3+ */
 		.pvr_mask		= 0xffff0000,
@@ -73,7 +72,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power3,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Northstar */
 		.pvr_mask		= 0xffff0000,
@@ -86,7 +84,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power3,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Pulsar */
 		.pvr_mask		= 0xffff0000,
@@ -99,7 +96,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power3,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* I-star */
 		.pvr_mask		= 0xffff0000,
@@ -112,7 +108,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power3,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* S-star */
 		.pvr_mask		= 0xffff0000,
@@ -125,7 +120,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power3,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Power4 */
 		.pvr_mask		= 0xffff0000,
@@ -138,7 +132,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power4,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Power4+ */
 		.pvr_mask		= 0xffff0000,
@@ -151,7 +144,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power4,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* PPC970 */
 		.pvr_mask		= 0xffff0000,
@@ -166,7 +158,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_ppc970,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* PPC970FX */
 		.pvr_mask		= 0xffff0000,
@@ -181,7 +172,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_ppc970,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* PPC970MP */
 		.pvr_mask		= 0xffff0000,
@@ -196,7 +186,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_ppc970,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Power5 */
 		.pvr_mask		= 0xffff0000,
@@ -211,7 +200,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power4,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* Power5 */
 		.pvr_mask		= 0xffff0000,
@@ -226,7 +214,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power4,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* BE DD1.x */
 		.pvr_mask		= 0xffff0000,
@@ -241,7 +228,6 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_be,
-		.firmware_features	= COMMON_PPC64_FW,
 	},
 	{	/* default match */
 		.pvr_mask		= 0x00000000,
@@ -254,29 +240,5 @@ struct cpu_spec	cpu_specs[] = {
 		.icache_bsize		= 128,
 		.dcache_bsize		= 128,
 		.cpu_setup		= __setup_cpu_power4,
-		.firmware_features	= COMMON_PPC64_FW,
 	}
 };
-
-firmware_feature_t firmware_features_table[FIRMWARE_MAX_FEATURES] = {
-	{FW_FEATURE_PFT,		"hcall-pft"},
-	{FW_FEATURE_TCE,		"hcall-tce"},
-	{FW_FEATURE_SPRG0,		"hcall-sprg0"},
-	{FW_FEATURE_DABR,		"hcall-dabr"},
-	{FW_FEATURE_COPY,		"hcall-copy"},
-	{FW_FEATURE_ASR,		"hcall-asr"},
-	{FW_FEATURE_DEBUG,		"hcall-debug"},
-	{FW_FEATURE_PERF,		"hcall-perf"},
-	{FW_FEATURE_DUMP,		"hcall-dump"},
-	{FW_FEATURE_INTERRUPT,		"hcall-interrupt"},
-	{FW_FEATURE_MIGRATE,		"hcall-migrate"},
-	{FW_FEATURE_PERFMON,		"hcall-perfmon"},
-	{FW_FEATURE_CRQ,		"hcall-crq"},
-	{FW_FEATURE_VIO,		"hcall-vio"},
-	{FW_FEATURE_RDMA,		"hcall-rdma"},
-	{FW_FEATURE_LLAN,		"hcall-lLAN"},
-	{FW_FEATURE_BULK,		"hcall-bulk"},
-	{FW_FEATURE_XDABR,		"hcall-xdabr"},
-	{FW_FEATURE_MULTITCE,		"hcall-multi-tce"},
-	{FW_FEATURE_SPLPAR,		"hcall-splpar"},
-};
diff --git a/arch/ppc64/kernel/firmware.c b/arch/ppc64/kernel/firmware.c
new file mode 100644
index 0000000..d8432c0
--- /dev/null
+++ b/arch/ppc64/kernel/firmware.c
@@ -0,0 +1,47 @@
+/*
+ *  arch/ppc64/kernel/firmware.c
+ *
+ *  Extracted from cputable.c
+ *
+ *  Copyright (C) 2001 Ben. Herrenschmidt (benh@kernel.crashing.org)
+ *
+ *  Modifications for ppc64:
+ *      Copyright (C) 2003 Dave Engebretsen <engebret@us.ibm.com>
+ *  Copyright (C) 2005 Stephen Rothwell, IBM Corporation
+ *
+ *  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 <linux/config.h>
+
+#include <asm/firmware.h>
+
+unsigned long ppc64_firmware_features;
+
+#ifdef CONFIG_PPC_PSERIES
+firmware_feature_t firmware_features_table[FIRMWARE_MAX_FEATURES] = {
+	{FW_FEATURE_PFT,		"hcall-pft"},
+	{FW_FEATURE_TCE,		"hcall-tce"},
+	{FW_FEATURE_SPRG0,		"hcall-sprg0"},
+	{FW_FEATURE_DABR,		"hcall-dabr"},
+	{FW_FEATURE_COPY,		"hcall-copy"},
+	{FW_FEATURE_ASR,		"hcall-asr"},
+	{FW_FEATURE_DEBUG,		"hcall-debug"},
+	{FW_FEATURE_PERF,		"hcall-perf"},
+	{FW_FEATURE_DUMP,		"hcall-dump"},
+	{FW_FEATURE_INTERRUPT,		"hcall-interrupt"},
+	{FW_FEATURE_MIGRATE,		"hcall-migrate"},
+	{FW_FEATURE_PERFMON,		"hcall-perfmon"},
+	{FW_FEATURE_CRQ,		"hcall-crq"},
+	{FW_FEATURE_VIO,		"hcall-vio"},
+	{FW_FEATURE_RDMA,		"hcall-rdma"},
+	{FW_FEATURE_LLAN,		"hcall-lLAN"},
+	{FW_FEATURE_BULK,		"hcall-bulk"},
+	{FW_FEATURE_XDABR,		"hcall-xdabr"},
+	{FW_FEATURE_MULTITCE,		"hcall-multi-tce"},
+	{FW_FEATURE_SPLPAR,		"hcall-splpar"},
+};
+#endif
diff --git a/arch/ppc64/kernel/head.S b/arch/ppc64/kernel/head.S
index accaa05..cccec49 100644
--- a/arch/ppc64/kernel/head.S
+++ b/arch/ppc64/kernel/head.S
@@ -23,14 +23,11 @@
  *  2 of the License, or (at your option) any later version.
  */
 
-#define SECONDARY_PROCESSORS
-
 #include <linux/config.h>
 #include <linux/threads.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
-#include <asm/naca.h>
 #include <asm/systemcfg.h>
 #include <asm/ppc_asm.h>
 #include <asm/offsets.h>
@@ -45,18 +42,13 @@
 #endif
 
 /*
- * hcall interface to pSeries LPAR
- */
-#define H_SET_ASR	0x30
-
-/*
  * We layout physical memory as follows:
  * 0x0000 - 0x00ff : Secondary processor spin code
  * 0x0100 - 0x2fff : pSeries Interrupt prologs
- * 0x3000 - 0x3fff : Interrupt support
- * 0x4000 - 0x4fff : NACA
- * 0x6000	   : iSeries and common interrupt prologs
- * 0x9000 - 0x9fff : Initial segment table
+ * 0x3000 - 0x5fff : interrupt support, iSeries and common interrupt prologs
+ * 0x6000 - 0x6fff : Initial (CPU0) segment table
+ * 0x7000 - 0x7fff : FWNMI data area
+ * 0x8000 -        : Early init and support code
  */
 
 /*
@@ -94,6 +86,7 @@ END_FTR_SECTION(0, 1)
 
 	/* Catch branch to 0 in real mode */
 	trap
+
 #ifdef CONFIG_PPC_ISERIES
 	/*
 	 * At offset 0x20, there is a pointer to iSeries LPAR data.
@@ -103,12 +96,12 @@ END_FTR_SECTION(0, 1)
 	.llong hvReleaseData-KERNELBASE
 
 	/*
-	 * At offset 0x28 and 0x30 are offsets to the msChunks
+	 * At offset 0x28 and 0x30 are offsets to the mschunks_map
 	 * array (used by the iSeries LPAR debugger to do translation
 	 * between physical addresses and absolute addresses) and
 	 * to the pidhash table (also used by the debugger)
 	 */
-	.llong msChunks-KERNELBASE
+	.llong mschunks_map-KERNELBASE
 	.llong 0	/* pidhash-KERNELBASE SFRXXX */
 
 	/* Offset 0x38 - Pointer to start of embedded System.map */
@@ -120,7 +113,7 @@ embedded_sysmap_start:
 embedded_sysmap_end:
 	.llong	0
 
-#else /* CONFIG_PPC_ISERIES */
+#endif /* CONFIG_PPC_ISERIES */
 
 	/* Secondary processors spin on this value until it goes to 1. */
 	.globl  __secondary_hold_spinloop
@@ -155,7 +148,7 @@ _GLOBAL(__secondary_hold)
 	std	r24,__secondary_hold_acknowledge@l(0)
 	sync
 
-	/* All secondary cpu's wait here until told to start. */
+	/* All secondary cpus wait here until told to start. */
 100:	ld	r4,__secondary_hold_spinloop@l(0)
 	cmpdi	0,r4,1
 	bne	100b
@@ -170,7 +163,6 @@ _GLOBAL(__secondary_hold)
 	BUG_OPCODE
 #endif
 #endif
-#endif
 
 /* This value is used to mark exception frames on the stack. */
 	.section ".toc","aw"
@@ -502,33 +494,37 @@ system_call_pSeries:
 	STD_EXCEPTION_PSERIES(0x1300, instruction_breakpoint)
 	STD_EXCEPTION_PSERIES(0x1700, altivec_assist)
 
+	. = 0x3000
+
+/*** pSeries interrupt support ***/
+
 	/* moved from 0xf00 */
-	STD_EXCEPTION_PSERIES(0x3000, performance_monitor)
+	STD_EXCEPTION_PSERIES(., performance_monitor)
 
-	. = 0x3100
+	.align	7
 _GLOBAL(do_stab_bolted_pSeries)
 	mtcrf	0x80,r12
 	mfspr	r12,SPRG2
 	EXCEPTION_PROLOG_PSERIES(PACA_EXSLB, .do_stab_bolted)
 
-	
-	/* Space for the naca.  Architected to be located at real address
-	 * NACA_PHYS_ADDR.  Various tools rely on this location being fixed.
-	 * The first dword of the naca is required by iSeries LPAR to
-	 * point to itVpdAreas.  On pSeries native, this value is not used.
-	 */
-	. = NACA_PHYS_ADDR
-	.globl __end_interrupts
-__end_interrupts:
-#ifdef CONFIG_PPC_ISERIES
-	.globl naca
-naca:
-	.llong	itVpdAreas
-	.llong	0		/* xRamDisk */
-	.llong	0		/* xRamDiskSize */
+/*
+ * Vectors for the FWNMI option.  Share common code.
+ */
+      .globl system_reset_fwnmi
+system_reset_fwnmi:
+      HMT_MEDIUM
+      mtspr   SPRG1,r13               /* save r13 */
+      RUNLATCH_ON(r13)
+      EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, system_reset_common)
 
-	. = 0x6100
+      .globl machine_check_fwnmi
+machine_check_fwnmi:
+      HMT_MEDIUM
+      mtspr   SPRG1,r13               /* save r13 */
+      RUNLATCH_ON(r13)
+      EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common)
 
+#ifdef CONFIG_PPC_ISERIES
 /***  ISeries-LPAR interrupt handlers ***/
 
 	STD_EXCEPTION_ISERIES(0x200, machine_check, PACA_EXMC)
@@ -626,9 +622,7 @@ system_reset_iSeries:
 
 	cmpwi	0,r23,0
 	beq	iSeries_secondary_smp_loop	/* Loop until told to go */
-#ifdef SECONDARY_PROCESSORS
 	bne	.__secondary_start		/* Loop until told to go */
-#endif
 iSeries_secondary_smp_loop:
 	/* Let the Hypervisor know we are alive */
 	/* 8002 is a call to HvCallCfg::getLps, a harmless Hypervisor function */
@@ -671,51 +665,8 @@ hardware_interrupt_iSeries_masked:
 	ld	r13,PACA_EXGEN+EX_R13(r13)
 	rfid
 	b	.	/* prevent speculative execution */
-#endif
-
-/*
- * Data area reserved for FWNMI option.
- */
-	.= 0x7000
-	.globl fwnmi_data_area
-fwnmi_data_area:
-
-#ifdef CONFIG_PPC_ISERIES
-	. = LPARMAP_PHYS
-#include "lparmap.s"
 #endif /* CONFIG_PPC_ISERIES */
 
-/*
- * Vectors for the FWNMI option.  Share common code.
- */
-	. = 0x8000
-	.globl system_reset_fwnmi
-system_reset_fwnmi:
-	HMT_MEDIUM
-	mtspr	SPRG1,r13		/* save r13 */
-	RUNLATCH_ON(r13)
-	EXCEPTION_PROLOG_PSERIES(PACA_EXGEN, system_reset_common)
-	.globl machine_check_fwnmi
-machine_check_fwnmi:
-	HMT_MEDIUM
-	mtspr	SPRG1,r13		/* save r13 */
-	RUNLATCH_ON(r13)
-	EXCEPTION_PROLOG_PSERIES(PACA_EXMC, machine_check_common)
-
-	/*
-	 * Space for the initial segment table
-	 * For LPAR, the hypervisor must fill in at least one entry
-	 * before we get control (with relocate on)
-	 */
-	. = STAB0_PHYS_ADDR
-	.globl __start_stab
-__start_stab:
-
-	. = (STAB0_PHYS_ADDR + PAGE_SIZE)
-	.globl __end_stab
-__end_stab:
-
-
 /*** Common interrupt handlers ***/
 
 	STD_EXCEPTION_COMMON(0x100, system_reset, .system_reset_exception)
@@ -752,8 +703,8 @@ machine_check_common:
  * R9 contains the saved CR, r13 points to the paca,
  * r10 contains the (bad) kernel stack pointer,
  * r11 and r12 contain the saved SRR0 and SRR1.
- * We switch to using the paca guard page as an emergency stack,
- * save the registers there, and call kernel_bad_stack(), which panics.
+ * We switch to using an emergency stack, save the registers there,
+ * and call kernel_bad_stack(), which panics.
  */
 bad_stack:
 	ld	r1,PACAEMERGSP(r13)
@@ -906,6 +857,62 @@ fp_unavailable_common:
 	bl	.kernel_fp_unavailable_exception
 	BUG_OPCODE
 
+/*
+ * load_up_fpu(unused, unused, tsk)
+ * Disable FP for the task which had the FPU previously,
+ * and save its floating-point registers in its thread_struct.
+ * Enables the FPU for use in the kernel on return.
+ * On SMP we know the fpu is free, since we give it up every
+ * switch (ie, no lazy save of the FP registers).
+ * On entry: r13 == 'current' && last_task_used_math != 'current'
+ */
+_STATIC(load_up_fpu)
+	mfmsr	r5			/* grab the current MSR */
+	ori	r5,r5,MSR_FP
+	mtmsrd	r5			/* enable use of fpu now */
+	isync
+/*
+ * For SMP, we don't do lazy FPU switching because it just gets too
+ * horrendously complex, especially when a task switches from one CPU
+ * to another.  Instead we call giveup_fpu in switch_to.
+ *
+ */
+#ifndef CONFIG_SMP
+	ld	r3,last_task_used_math@got(r2)
+	ld	r4,0(r3)
+	cmpdi	0,r4,0
+	beq	1f
+	/* Save FP state to last_task_used_math's THREAD struct */
+	addi	r4,r4,THREAD
+	SAVE_32FPRS(0, r4)
+	mffs	fr0
+	stfd	fr0,THREAD_FPSCR(r4)
+	/* Disable FP for last_task_used_math */
+	ld	r5,PT_REGS(r4)
+	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+	li	r6,MSR_FP|MSR_FE0|MSR_FE1
+	andc	r4,r4,r6
+	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+1:
+#endif /* CONFIG_SMP */
+	/* enable use of FP after return */
+	ld	r4,PACACURRENT(r13)
+	addi	r5,r4,THREAD		/* Get THREAD */
+	ld	r4,THREAD_FPEXC_MODE(r5)
+	ori	r12,r12,MSR_FP
+	or	r12,r12,r4
+	std	r12,_MSR(r1)
+	lfd	fr0,THREAD_FPSCR(r5)
+	mtfsf	0xff,fr0
+	REST_32FPRS(0, r5)
+#ifndef CONFIG_SMP
+	/* Update last_task_used_math to 'current' */
+	subi	r4,r5,THREAD		/* Back to 'current' */
+	std	r4,0(r3)
+#endif /* CONFIG_SMP */
+	/* restore registers and return */
+	b	fast_exception_return
+
 	.align	7
 	.globl altivec_unavailable_common
 altivec_unavailable_common:
@@ -921,6 +928,80 @@ END_FTR_SECTION_IFSET(CPU_FTR_ALTIVEC)
 	bl	.altivec_unavailable_exception
 	b	.ret_from_except
 
+#ifdef CONFIG_ALTIVEC
+/*
+ * load_up_altivec(unused, unused, tsk)
+ * Disable VMX for the task which had it previously,
+ * and save its vector registers in its thread_struct.
+ * Enables the VMX for use in the kernel on return.
+ * On SMP we know the VMX is free, since we give it up every
+ * switch (ie, no lazy save of the vector registers).
+ * On entry: r13 == 'current' && last_task_used_altivec != 'current'
+ */
+_STATIC(load_up_altivec)
+	mfmsr	r5			/* grab the current MSR */
+	oris	r5,r5,MSR_VEC@h
+	mtmsrd	r5			/* enable use of VMX now */
+	isync
+
+/*
+ * For SMP, we don't do lazy VMX switching because it just gets too
+ * horrendously complex, especially when a task switches from one CPU
+ * to another.  Instead we call giveup_altvec in switch_to.
+ * VRSAVE isn't dealt with here, that is done in the normal context
+ * switch code. Note that we could rely on vrsave value to eventually
+ * avoid saving all of the VREGs here...
+ */
+#ifndef CONFIG_SMP
+	ld	r3,last_task_used_altivec@got(r2)
+	ld	r4,0(r3)
+	cmpdi	0,r4,0
+	beq	1f
+	/* Save VMX state to last_task_used_altivec's THREAD struct */
+	addi	r4,r4,THREAD
+	SAVE_32VRS(0,r5,r4)
+	mfvscr	vr0
+	li	r10,THREAD_VSCR
+	stvx	vr0,r10,r4
+	/* Disable VMX for last_task_used_altivec */
+	ld	r5,PT_REGS(r4)
+	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+	lis	r6,MSR_VEC@h
+	andc	r4,r4,r6
+	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+1:
+#endif /* CONFIG_SMP */
+	/* Hack: if we get an altivec unavailable trap with VRSAVE
+	 * set to all zeros, we assume this is a broken application
+	 * that fails to set it properly, and thus we switch it to
+	 * all 1's
+	 */
+	mfspr	r4,SPRN_VRSAVE
+	cmpdi	0,r4,0
+	bne+	1f
+	li	r4,-1
+	mtspr	SPRN_VRSAVE,r4
+1:
+	/* enable use of VMX after return */
+	ld	r4,PACACURRENT(r13)
+	addi	r5,r4,THREAD		/* Get THREAD */
+	oris	r12,r12,MSR_VEC@h
+	std	r12,_MSR(r1)
+	li	r4,1
+	li	r10,THREAD_VSCR
+	stw	r4,THREAD_USED_VR(r5)
+	lvx	vr0,r10,r5
+	mtvscr	vr0
+	REST_32VRS(0,r4,r5)
+#ifndef CONFIG_SMP
+	/* Update last_task_used_math to 'current' */
+	subi	r4,r5,THREAD		/* Back to 'current' */
+	std	r4,0(r3)
+#endif /* CONFIG_SMP */
+	/* restore registers and return */
+	b	fast_exception_return
+#endif /* CONFIG_ALTIVEC */
+
 /*
  * Hash table stuff
  */
@@ -1167,6 +1248,28 @@ unrecov_slb:
 	bl	.unrecoverable_exception
 	b	1b
 
+/*
+ * Space for CPU0's segment table.
+ *
+ * On iSeries, the hypervisor must fill in at least one entry before
+ * we get control (with relocate on).  The address is give to the hv
+ * as a page number (see xLparMap in LparData.c), so this must be at a
+ * fixed address (the linker can't compute (u64)&initial_stab >>
+ * PAGE_SHIFT).
+ */
+	. = STAB0_PHYS_ADDR	/* 0x6000 */
+	.globl initial_stab
+initial_stab:
+	.space	4096
+
+/*
+ * Data area reserved for FWNMI option.
+ * This address (0x7000) is fixed by the RPA.
+ */
+	.= 0x7000
+	.globl fwnmi_data_area
+fwnmi_data_area:
+	.space	PAGE_SIZE
 
 /*
  * On pSeries, secondary processors spin in the following code.
@@ -1200,7 +1303,7 @@ _GLOBAL(pSeries_secondary_smp_init)
 	b	.kexec_wait		/* next kernel might do better	 */
 
 2:	mtspr	SPRG3,r13		/* Save vaddr of paca in SPRG3	 */
-	/* From now on, r24 is expected to be logica cpuid */
+	/* From now on, r24 is expected to be logical cpuid */
 	mr	r24,r5
 3:	HMT_LOW
 	lbz	r23,PACAPROCSTART(r13)	/* Test if this processor should */
@@ -1213,10 +1316,8 @@ _GLOBAL(pSeries_secondary_smp_init)
 
 	cmpwi	0,r23,0
 #ifdef CONFIG_SMP
-#ifdef SECONDARY_PROCESSORS
 	bne	.__secondary_start
 #endif
-#endif
 	b 	3b			/* Loop until told to go	 */
 
 #ifdef CONFIG_PPC_ISERIES
@@ -1430,228 +1531,6 @@ _GLOBAL(copy_and_flush)
 .align 8
 copy_to_here:
 
-/*
- * load_up_fpu(unused, unused, tsk)
- * Disable FP for the task which had the FPU previously,
- * and save its floating-point registers in its thread_struct.
- * Enables the FPU for use in the kernel on return.
- * On SMP we know the fpu is free, since we give it up every
- * switch (ie, no lazy save of the FP registers).
- * On entry: r13 == 'current' && last_task_used_math != 'current'
- */
-_STATIC(load_up_fpu)
-	mfmsr	r5			/* grab the current MSR */
-	ori	r5,r5,MSR_FP
-	mtmsrd	r5			/* enable use of fpu now */
-	isync
-/*
- * For SMP, we don't do lazy FPU switching because it just gets too
- * horrendously complex, especially when a task switches from one CPU
- * to another.  Instead we call giveup_fpu in switch_to.
- *
- */
-#ifndef CONFIG_SMP
-	ld	r3,last_task_used_math@got(r2)
-	ld	r4,0(r3)
-	cmpdi	0,r4,0
-	beq	1f
-	/* Save FP state to last_task_used_math's THREAD struct */
-	addi	r4,r4,THREAD
-	SAVE_32FPRS(0, r4)
-	mffs	fr0
-	stfd	fr0,THREAD_FPSCR(r4)
-	/* Disable FP for last_task_used_math */
-	ld	r5,PT_REGS(r4)
-	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-	li	r6,MSR_FP|MSR_FE0|MSR_FE1
-	andc	r4,r4,r6
-	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-1:
-#endif /* CONFIG_SMP */
-	/* enable use of FP after return */
-	ld	r4,PACACURRENT(r13)
-	addi	r5,r4,THREAD		/* Get THREAD */
-	ld	r4,THREAD_FPEXC_MODE(r5)
-	ori	r12,r12,MSR_FP
-	or	r12,r12,r4
-	std	r12,_MSR(r1)
-	lfd	fr0,THREAD_FPSCR(r5)
-	mtfsf	0xff,fr0
-	REST_32FPRS(0, r5)
-#ifndef CONFIG_SMP
-	/* Update last_task_used_math to 'current' */
-	subi	r4,r5,THREAD		/* Back to 'current' */
-	std	r4,0(r3)
-#endif /* CONFIG_SMP */
-	/* restore registers and return */
-	b	fast_exception_return
-
-/*
- * disable_kernel_fp()
- * Disable the FPU.
- */
-_GLOBAL(disable_kernel_fp)
-	mfmsr	r3
-	rldicl	r0,r3,(63-MSR_FP_LG),1
-	rldicl	r3,r0,(MSR_FP_LG+1),0
-	mtmsrd	r3			/* disable use of fpu now */
-	isync
-	blr
-
-/*
- * giveup_fpu(tsk)
- * Disable FP for the task given as the argument,
- * and save the floating-point registers in its thread_struct.
- * Enables the FPU for use in the kernel on return.
- */
-_GLOBAL(giveup_fpu)
-	mfmsr	r5
-	ori	r5,r5,MSR_FP
-	mtmsrd	r5			/* enable use of fpu now */
-	isync
-	cmpdi	0,r3,0
-	beqlr-				/* if no previous owner, done */
-	addi	r3,r3,THREAD		/* want THREAD of task */
-	ld	r5,PT_REGS(r3)
-	cmpdi	0,r5,0
-	SAVE_32FPRS(0, r3)
-	mffs	fr0
-	stfd	fr0,THREAD_FPSCR(r3)
-	beq	1f
-	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-	li	r3,MSR_FP|MSR_FE0|MSR_FE1
-	andc	r4,r4,r3		/* disable FP for previous task */
-	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-1:
-#ifndef CONFIG_SMP
-	li	r5,0
-	ld	r4,last_task_used_math@got(r2)
-	std	r5,0(r4)
-#endif /* CONFIG_SMP */
-	blr
-
-
-#ifdef CONFIG_ALTIVEC
-		
-/*
- * load_up_altivec(unused, unused, tsk)
- * Disable VMX for the task which had it previously,
- * and save its vector registers in its thread_struct.
- * Enables the VMX for use in the kernel on return.
- * On SMP we know the VMX is free, since we give it up every
- * switch (ie, no lazy save of the vector registers).
- * On entry: r13 == 'current' && last_task_used_altivec != 'current'
- */
-_STATIC(load_up_altivec)
-	mfmsr	r5			/* grab the current MSR */
-	oris	r5,r5,MSR_VEC@h
-	mtmsrd	r5			/* enable use of VMX now */
-	isync
-	
-/*
- * For SMP, we don't do lazy VMX switching because it just gets too
- * horrendously complex, especially when a task switches from one CPU
- * to another.  Instead we call giveup_altvec in switch_to.
- * VRSAVE isn't dealt with here, that is done in the normal context
- * switch code. Note that we could rely on vrsave value to eventually
- * avoid saving all of the VREGs here...
- */
-#ifndef CONFIG_SMP
-	ld	r3,last_task_used_altivec@got(r2)
-	ld	r4,0(r3)
-	cmpdi	0,r4,0
-	beq	1f
-	/* Save VMX state to last_task_used_altivec's THREAD struct */
-	addi	r4,r4,THREAD
-	SAVE_32VRS(0,r5,r4)
-	mfvscr	vr0
-	li	r10,THREAD_VSCR
-	stvx	vr0,r10,r4
-	/* Disable VMX for last_task_used_altivec */
-	ld	r5,PT_REGS(r4)
-	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-	lis	r6,MSR_VEC@h
-	andc	r4,r4,r6
-	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-1:
-#endif /* CONFIG_SMP */
-	/* Hack: if we get an altivec unavailable trap with VRSAVE
-	 * set to all zeros, we assume this is a broken application
-	 * that fails to set it properly, and thus we switch it to
-	 * all 1's
-	 */
-	mfspr	r4,SPRN_VRSAVE
-	cmpdi	0,r4,0
-	bne+	1f
-	li	r4,-1
-	mtspr	SPRN_VRSAVE,r4
-1:
-	/* enable use of VMX after return */
-	ld	r4,PACACURRENT(r13)
-	addi	r5,r4,THREAD		/* Get THREAD */
-	oris	r12,r12,MSR_VEC@h
-	std	r12,_MSR(r1)
-	li	r4,1
-	li	r10,THREAD_VSCR
-	stw	r4,THREAD_USED_VR(r5)
-	lvx	vr0,r10,r5
-	mtvscr	vr0
-	REST_32VRS(0,r4,r5)
-#ifndef CONFIG_SMP
-	/* Update last_task_used_math to 'current' */
-	subi	r4,r5,THREAD		/* Back to 'current' */
-	std	r4,0(r3)
-#endif /* CONFIG_SMP */
-	/* restore registers and return */
-	b	fast_exception_return
-
-/*
- * disable_kernel_altivec()
- * Disable the VMX.
- */
-_GLOBAL(disable_kernel_altivec)
-	mfmsr	r3
-	rldicl	r0,r3,(63-MSR_VEC_LG),1
-	rldicl	r3,r0,(MSR_VEC_LG+1),0
-	mtmsrd	r3			/* disable use of VMX now */
-	isync
-	blr
-
-/*
- * giveup_altivec(tsk)
- * Disable VMX for the task given as the argument,
- * and save the vector registers in its thread_struct.
- * Enables the VMX for use in the kernel on return.
- */
-_GLOBAL(giveup_altivec)
-	mfmsr	r5
-	oris	r5,r5,MSR_VEC@h
-	mtmsrd	r5			/* enable use of VMX now */
-	isync
-	cmpdi	0,r3,0
-	beqlr-				/* if no previous owner, done */
-	addi	r3,r3,THREAD		/* want THREAD of task */
-	ld	r5,PT_REGS(r3)
-	cmpdi	0,r5,0
-	SAVE_32VRS(0,r4,r3)
-	mfvscr	vr0
-	li	r4,THREAD_VSCR
-	stvx	vr0,r4,r3
-	beq	1f
-	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-	lis	r3,MSR_VEC@h
-	andc	r4,r4,r3		/* disable FP for previous task */
-	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
-1:
-#ifndef CONFIG_SMP
-	li	r5,0
-	ld	r4,last_task_used_altivec@got(r2)
-	std	r5,0(r4)
-#endif /* CONFIG_SMP */
-	blr
-
-#endif /* CONFIG_ALTIVEC */
-
 #ifdef CONFIG_SMP
 #ifdef CONFIG_PPC_PMAC
 /*
@@ -2002,9 +1881,6 @@ _STATIC(start_here_common)
 
 	bl .start_kernel
 
-_GLOBAL(__setup_cpu_power3)
-	blr
-
 _GLOBAL(hmt_init)
 #ifdef CONFIG_HMT
 	LOADADDR(r5, hmt_thread_data)
@@ -2095,20 +1971,19 @@ _GLOBAL(smp_release_cpus)
 
 /*
  * We put a few things here that have to be page-aligned.
- * This stuff goes at the beginning of the data segment,
- * which is page-aligned.
+ * This stuff goes at the beginning of the bss, which is page-aligned.
  */
-	.data
+	.section ".bss"
+
 	.align	12
-	.globl	sdata
-sdata:
+
 	.globl	empty_zero_page
 empty_zero_page:
-	.space	4096
+	.space	PAGE_SIZE
 
 	.globl	swapper_pg_dir
 swapper_pg_dir:
-	.space	4096
+	.space	PAGE_SIZE
 
 /*
  * This space gets a copy of optional info passed to us by the bootstrap
diff --git a/arch/ppc64/kernel/iSeries_htab.c b/arch/ppc64/kernel/iSeries_htab.c
index b0250ae..2192055 100644
--- a/arch/ppc64/kernel/iSeries_htab.c
+++ b/arch/ppc64/kernel/iSeries_htab.c
@@ -41,6 +41,7 @@ static long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va,
 				unsigned long prpn, unsigned long vflags,
 				unsigned long rflags)
 {
+	unsigned long arpn;
 	long slot;
 	hpte_t lhpte;
 	int secondary = 0;
@@ -70,8 +71,10 @@ static long iSeries_hpte_insert(unsigned long hpte_group, unsigned long va,
 		slot &= 0x7fffffffffffffff;
 	}
 
+	arpn = phys_to_abs(prpn << PAGE_SHIFT) >> PAGE_SHIFT;
+
 	lhpte.v = (va >> 23) << HPTE_V_AVPN_SHIFT | vflags | HPTE_V_VALID;
-	lhpte.r = (physRpn_to_absRpn(prpn) << HPTE_R_RPN_SHIFT) | rflags;
+	lhpte.r = (arpn << HPTE_R_RPN_SHIFT) | rflags;
 
 	/* Now fill in the actual HPTE */
 	HvCallHpt_addValidate(slot, secondary, &lhpte);
diff --git a/arch/ppc64/kernel/iSeries_setup.c b/arch/ppc64/kernel/iSeries_setup.c
index a649edb..3ffefbb 100644
--- a/arch/ppc64/kernel/iSeries_setup.c
+++ b/arch/ppc64/kernel/iSeries_setup.c
@@ -39,6 +39,7 @@
 #include <asm/cputable.h>
 #include <asm/sections.h>
 #include <asm/iommu.h>
+#include <asm/firmware.h>
 
 #include <asm/time.h>
 #include "iSeries_setup.h"
@@ -314,6 +315,8 @@ static void __init iSeries_init_early(void)
 
 	DBG(" -> iSeries_init_early()\n");
 
+	ppc64_firmware_features = FW_FEATURE_ISERIES;
+
 	ppcdbg_initialize();
 
 #if defined(CONFIG_BLK_DEV_INITRD)
@@ -412,6 +415,22 @@ static void __init iSeries_init_early(void)
 	DBG(" <- iSeries_init_early()\n");
 }
 
+struct mschunks_map mschunks_map = {
+	/* XXX We don't use these, but Piranha might need them. */
+	.chunk_size  = MSCHUNKS_CHUNK_SIZE,
+	.chunk_shift = MSCHUNKS_CHUNK_SHIFT,
+	.chunk_mask  = MSCHUNKS_OFFSET_MASK,
+};
+EXPORT_SYMBOL(mschunks_map);
+
+void mschunks_alloc(unsigned long num_chunks)
+{
+	klimit = _ALIGN(klimit, sizeof(u32));
+	mschunks_map.mapping = (u32 *)klimit;
+	klimit += num_chunks * sizeof(u32);
+	mschunks_map.num_chunks = num_chunks;
+}
+
 /*
  * The iSeries may have very large memories ( > 128 GB ) and a partition
  * may get memory in "chunks" that may be anywhere in the 2**52 real
@@ -449,7 +468,7 @@ static void __init build_iSeries_Memory_Map(void)
 
 	/* Chunk size on iSeries is 256K bytes */
 	totalChunks = (u32)HvLpConfig_getMsChunks();
-	klimit = msChunks_alloc(klimit, totalChunks, 1UL << 18);
+	mschunks_alloc(totalChunks);
 
 	/*
 	 * Get absolute address of our load area
@@ -486,7 +505,7 @@ static void __init build_iSeries_Memory_Map(void)
 	printk("Load area size %dK\n", loadAreaSize * 256);
 
 	for (nextPhysChunk = 0; nextPhysChunk < loadAreaSize; ++nextPhysChunk)
-		msChunks.abs[nextPhysChunk] =
+		mschunks_map.mapping[nextPhysChunk] =
 			loadAreaFirstChunk + nextPhysChunk;
 
 	/*
@@ -495,7 +514,7 @@ static void __init build_iSeries_Memory_Map(void)
 	 */
 	hptFirstChunk = (u32)addr_to_chunk(HvCallHpt_getHptAddress());
 	hptSizePages = (u32)HvCallHpt_getHptPages();
-	hptSizeChunks = hptSizePages >> (msChunks.chunk_shift - PAGE_SHIFT);
+	hptSizeChunks = hptSizePages >> (MSCHUNKS_CHUNK_SHIFT - PAGE_SHIFT);
 	hptLastChunk = hptFirstChunk + hptSizeChunks - 1;
 
 	printk("HPT absolute addr = %016lx, size = %dK\n",
@@ -552,7 +571,8 @@ static void __init build_iSeries_Memory_Map(void)
 				     (absChunk > hptLastChunk)) &&
 				    ((absChunk < loadAreaFirstChunk) ||
 				     (absChunk > loadAreaLastChunk))) {
-					msChunks.abs[nextPhysChunk] = absChunk;
+					mschunks_map.mapping[nextPhysChunk] =
+						absChunk;
 					++nextPhysChunk;
 				}
 			}
@@ -944,6 +964,8 @@ void __init iSeries_early_setup(void)
 	ppc_md.calibrate_decr = iSeries_calibrate_decr;
 	ppc_md.progress = iSeries_progress;
 
+	/* XXX Implement enable_pmcs for iSeries */
+
 	if (get_paca()->lppaca.shared_proc) {
 		ppc_md.idle_loop = iseries_shared_idle;
 		printk(KERN_INFO "Using shared processor idle loop\n");
diff --git a/arch/ppc64/kernel/iSeries_vio.c b/arch/ppc64/kernel/iSeries_vio.c
new file mode 100644
index 0000000..b4268cc
--- /dev/null
+++ b/arch/ppc64/kernel/iSeries_vio.c
@@ -0,0 +1,144 @@
+/*
+ * IBM PowerPC iSeries Virtual I/O Infrastructure Support.
+ *
+ *    Copyright (c) 2005 Stephen Rothwell, IBM Corp.
+ *
+ *      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 <linux/types.h>
+#include <linux/device.h>
+#include <linux/init.h>
+
+#include <asm/vio.h>
+#include <asm/iommu.h>
+#include <asm/abs_addr.h>
+#include <asm/page.h>
+#include <asm/iSeries/vio.h>
+#include <asm/iSeries/HvTypes.h>
+#include <asm/iSeries/HvLpConfig.h>
+#include <asm/iSeries/HvCallXm.h>
+
+struct device *iSeries_vio_dev = &vio_bus_device.dev;
+EXPORT_SYMBOL(iSeries_vio_dev);
+
+static struct iommu_table veth_iommu_table;
+static struct iommu_table vio_iommu_table;
+
+static void __init iommu_vio_init(void)
+{
+	struct iommu_table *t;
+	struct iommu_table_cb cb;
+	unsigned long cbp;
+	unsigned long itc_entries;
+
+	cb.itc_busno = 255;    /* Bus 255 is the virtual bus */
+	cb.itc_virtbus = 0xff; /* Ask for virtual bus */
+
+	cbp = virt_to_abs(&cb);
+	HvCallXm_getTceTableParms(cbp);
+
+	itc_entries = cb.itc_size * PAGE_SIZE / sizeof(union tce_entry);
+	veth_iommu_table.it_size        = itc_entries / 2;
+	veth_iommu_table.it_busno       = cb.itc_busno;
+	veth_iommu_table.it_offset      = cb.itc_offset;
+	veth_iommu_table.it_index       = cb.itc_index;
+	veth_iommu_table.it_type        = TCE_VB;
+	veth_iommu_table.it_blocksize	= 1;
+
+	t = iommu_init_table(&veth_iommu_table);
+
+	if (!t)
+		printk("Virtual Bus VETH TCE table failed.\n");
+
+	vio_iommu_table.it_size         = itc_entries - veth_iommu_table.it_size;
+	vio_iommu_table.it_busno        = cb.itc_busno;
+	vio_iommu_table.it_offset       = cb.itc_offset +
+					  veth_iommu_table.it_size;
+	vio_iommu_table.it_index        = cb.itc_index;
+	vio_iommu_table.it_type         = TCE_VB;
+	vio_iommu_table.it_blocksize	= 1;
+
+	t = iommu_init_table(&vio_iommu_table);
+
+	if (!t)
+		printk("Virtual Bus VIO TCE table failed.\n");
+}
+
+/**
+ * vio_register_device: - Register a new vio device.
+ * @voidev:	The device to register.
+ */
+static struct vio_dev *__init vio_register_device_iseries(char *type,
+		uint32_t unit_num)
+{
+	struct vio_dev *viodev;
+
+	/* allocate a vio_dev for this node */
+	viodev = kmalloc(sizeof(struct vio_dev), GFP_KERNEL);
+	if (!viodev)
+		return NULL;
+	memset(viodev, 0, sizeof(struct vio_dev));
+
+	snprintf(viodev->dev.bus_id, BUS_ID_SIZE, "%s%d", type, unit_num);
+
+	return vio_register_device_common(viodev, viodev->dev.bus_id, type,
+			unit_num, &vio_iommu_table);
+}
+
+void __init probe_bus_iseries(void)
+{
+	HvLpIndexMap vlan_map;
+	struct vio_dev *viodev;
+	int i;
+
+	/* there is only one of each of these */
+	vio_register_device_iseries("viocons", 0);
+	vio_register_device_iseries("vscsi", 0);
+
+	vlan_map = HvLpConfig_getVirtualLanIndexMap();
+	for (i = 0; i < HVMAXARCHITECTEDVIRTUALLANS; i++) {
+		if ((vlan_map & (0x8000 >> i)) == 0)
+			continue;
+		viodev = vio_register_device_iseries("vlan", i);
+		/* veth is special and has it own iommu_table */
+		viodev->iommu_table = &veth_iommu_table;
+	}
+	for (i = 0; i < HVMAXARCHITECTEDVIRTUALDISKS; i++)
+		vio_register_device_iseries("viodasd", i);
+	for (i = 0; i < HVMAXARCHITECTEDVIRTUALCDROMS; i++)
+		vio_register_device_iseries("viocd", i);
+	for (i = 0; i < HVMAXARCHITECTEDVIRTUALTAPES; i++)
+		vio_register_device_iseries("viotape", i);
+}
+
+/**
+ * vio_match_device_iseries: - Tell if a iSeries VIO device matches a
+ *	vio_device_id
+ */
+static int vio_match_device_iseries(const struct vio_device_id *id,
+		const struct vio_dev *dev)
+{
+	return strncmp(dev->type, id->type, strlen(id->type)) == 0;
+}
+
+/**
+ * vio_bus_init_iseries: - Initialize the iSeries virtual IO bus
+ */
+static int __init vio_bus_init_iseries(void)
+{
+	int err;
+
+	err = vio_bus_init(vio_match_device_iseries, NULL, NULL);
+	if (err == 0) {
+		iommu_vio_init();
+		vio_bus_device.iommu_table = &vio_iommu_table;
+		iSeries_vio_dev = &vio_bus_device.dev;
+		probe_bus_iseries();
+	}
+	return err;
+}
+
+__initcall(vio_bus_init_iseries);
diff --git a/arch/ppc64/kernel/lmb.c b/arch/ppc64/kernel/lmb.c
index d6c6bd0..5adaca2 100644
--- a/arch/ppc64/kernel/lmb.c
+++ b/arch/ppc64/kernel/lmb.c
@@ -28,33 +28,28 @@ void lmb_dump_all(void)
 {
 #ifdef DEBUG
 	unsigned long i;
-	struct lmb *_lmb  = &lmb;
 
 	udbg_printf("lmb_dump_all:\n");
 	udbg_printf("    memory.cnt		  = 0x%lx\n",
-		    _lmb->memory.cnt);
+		    lmb.memory.cnt);
 	udbg_printf("    memory.size		  = 0x%lx\n",
-		    _lmb->memory.size);
-	for (i=0; i < _lmb->memory.cnt ;i++) {
+		    lmb.memory.size);
+	for (i=0; i < lmb.memory.cnt ;i++) {
 		udbg_printf("    memory.region[0x%x].base       = 0x%lx\n",
-			    i, _lmb->memory.region[i].base);
-		udbg_printf("		      .physbase = 0x%lx\n",
-			    _lmb->memory.region[i].physbase);
+			    i, lmb.memory.region[i].base);
 		udbg_printf("		      .size     = 0x%lx\n",
-			    _lmb->memory.region[i].size);
+			    lmb.memory.region[i].size);
 	}
 
 	udbg_printf("\n    reserved.cnt	  = 0x%lx\n",
-		    _lmb->reserved.cnt);
+		    lmb.reserved.cnt);
 	udbg_printf("    reserved.size	  = 0x%lx\n",
-		    _lmb->reserved.size);
-	for (i=0; i < _lmb->reserved.cnt ;i++) {
+		    lmb.reserved.size);
+	for (i=0; i < lmb.reserved.cnt ;i++) {
 		udbg_printf("    reserved.region[0x%x].base       = 0x%lx\n",
-			    i, _lmb->reserved.region[i].base);
-		udbg_printf("		      .physbase = 0x%lx\n",
-			    _lmb->reserved.region[i].physbase);
+			    i, lmb.reserved.region[i].base);
 		udbg_printf("		      .size     = 0x%lx\n",
-			    _lmb->reserved.region[i].size);
+			    lmb.reserved.region[i].size);
 	}
 #endif /* DEBUG */
 }
@@ -98,7 +93,6 @@ lmb_coalesce_regions(struct lmb_region *rgn, unsigned long r1, unsigned long r2)
 	rgn->region[r1].size += rgn->region[r2].size;
 	for (i=r2; i < rgn->cnt-1; i++) {
 		rgn->region[i].base = rgn->region[i+1].base;
-		rgn->region[i].physbase = rgn->region[i+1].physbase;
 		rgn->region[i].size = rgn->region[i+1].size;
 	}
 	rgn->cnt--;
@@ -108,49 +102,29 @@ lmb_coalesce_regions(struct lmb_region *rgn, unsigned long r1, unsigned long r2)
 void __init
 lmb_init(void)
 {
-	struct lmb *_lmb = &lmb;
-
 	/* Create a dummy zero size LMB which will get coalesced away later.
 	 * This simplifies the lmb_add() code below...
 	 */
-	_lmb->memory.region[0].base = 0;
-	_lmb->memory.region[0].size = 0;
-	_lmb->memory.cnt = 1;
+	lmb.memory.region[0].base = 0;
+	lmb.memory.region[0].size = 0;
+	lmb.memory.cnt = 1;
 
 	/* Ditto. */
-	_lmb->reserved.region[0].base = 0;
-	_lmb->reserved.region[0].size = 0;
-	_lmb->reserved.cnt = 1;
+	lmb.reserved.region[0].base = 0;
+	lmb.reserved.region[0].size = 0;
+	lmb.reserved.cnt = 1;
 }
 
 /* This routine called with relocation disabled. */
 void __init
 lmb_analyze(void)
 {
-	unsigned long i;
-	unsigned long mem_size = 0;
-	unsigned long size_mask = 0;
-	struct lmb *_lmb = &lmb;
-#ifdef CONFIG_MSCHUNKS
-	unsigned long physbase = 0;
-#endif
-
-	for (i=0; i < _lmb->memory.cnt; i++) {
-		unsigned long lmb_size;
-
-		lmb_size = _lmb->memory.region[i].size;
-
-#ifdef CONFIG_MSCHUNKS
-		_lmb->memory.region[i].physbase = physbase;
-		physbase += lmb_size;
-#else
-		_lmb->memory.region[i].physbase = _lmb->memory.region[i].base;
-#endif
-		mem_size += lmb_size;
-		size_mask |= lmb_size;
-	}
+	int i;
+
+	lmb.memory.size = 0;
 
-	_lmb->memory.size = mem_size;
+	for (i = 0; i < lmb.memory.cnt; i++)
+		lmb.memory.size += lmb.memory.region[i].size;
 }
 
 /* This routine called with relocation disabled. */
@@ -168,7 +142,6 @@ lmb_add_region(struct lmb_region *rgn, unsigned long base, unsigned long size)
 		adjacent = lmb_addrs_adjacent(base,size,rgnbase,rgnsize);
 		if ( adjacent > 0 ) {
 			rgn->region[i].base -= size;
-			rgn->region[i].physbase -= size;
 			rgn->region[i].size += size;
 			coalesced++;
 			break;
@@ -195,11 +168,9 @@ lmb_add_region(struct lmb_region *rgn, unsigned long base, unsigned long size)
 	for (i=rgn->cnt-1; i >= 0; i--) {
 		if (base < rgn->region[i].base) {
 			rgn->region[i+1].base = rgn->region[i].base;
-			rgn->region[i+1].physbase = rgn->region[i].physbase;
 			rgn->region[i+1].size = rgn->region[i].size;
 		}  else {
 			rgn->region[i+1].base = base;
-			rgn->region[i+1].physbase = lmb_abs_to_phys(base);
 			rgn->region[i+1].size = size;
 			break;
 		}
@@ -213,12 +184,11 @@ lmb_add_region(struct lmb_region *rgn, unsigned long base, unsigned long size)
 long __init
 lmb_add(unsigned long base, unsigned long size)
 {
-	struct lmb *_lmb = &lmb;
-	struct lmb_region *_rgn = &(_lmb->memory);
+	struct lmb_region *_rgn = &(lmb.memory);
 
 	/* On pSeries LPAR systems, the first LMB is our RMO region. */
 	if ( base == 0 )
-		_lmb->rmo_size = size;
+		lmb.rmo_size = size;
 
 	return lmb_add_region(_rgn, base, size);
 
@@ -227,8 +197,7 @@ lmb_add(unsigned long base, unsigned long size)
 long __init
 lmb_reserve(unsigned long base, unsigned long size)
 {
-	struct lmb *_lmb = &lmb;
-	struct lmb_region *_rgn = &(_lmb->reserved);
+	struct lmb_region *_rgn = &(lmb.reserved);
 
 	return lmb_add_region(_rgn, base, size);
 }
@@ -260,13 +229,10 @@ lmb_alloc_base(unsigned long size, unsigned long align, unsigned long max_addr)
 {
 	long i, j;
 	unsigned long base = 0;
-	struct lmb *_lmb = &lmb;
-	struct lmb_region *_mem = &(_lmb->memory);
-	struct lmb_region *_rsv = &(_lmb->reserved);
 
-	for (i=_mem->cnt-1; i >= 0; i--) {
-		unsigned long lmbbase = _mem->region[i].base;
-		unsigned long lmbsize = _mem->region[i].size;
+	for (i=lmb.memory.cnt-1; i >= 0; i--) {
+		unsigned long lmbbase = lmb.memory.region[i].base;
+		unsigned long lmbsize = lmb.memory.region[i].size;
 
 		if ( max_addr == LMB_ALLOC_ANYWHERE )
 			base = _ALIGN_DOWN(lmbbase+lmbsize-size, align);
@@ -276,8 +242,8 @@ lmb_alloc_base(unsigned long size, unsigned long align, unsigned long max_addr)
 			continue;
 
 		while ( (lmbbase <= base) &&
-			((j = lmb_overlaps_region(_rsv,base,size)) >= 0) ) {
-			base = _ALIGN_DOWN(_rsv->region[j].base-size, align);
+			((j = lmb_overlaps_region(&lmb.reserved,base,size)) >= 0) ) {
+			base = _ALIGN_DOWN(lmb.reserved.region[j].base-size, align);
 		}
 
 		if ( (base != 0) && (lmbbase <= base) )
@@ -287,62 +253,24 @@ lmb_alloc_base(unsigned long size, unsigned long align, unsigned long max_addr)
 	if ( i < 0 )
 		return 0;
 
-	lmb_add_region(_rsv, base, size);
+	lmb_add_region(&lmb.reserved, base, size);
 
 	return base;
 }
 
+/* You must call lmb_analyze() before this. */
 unsigned long __init
 lmb_phys_mem_size(void)
 {
-	struct lmb *_lmb = &lmb;
-#ifdef CONFIG_MSCHUNKS
-	return _lmb->memory.size;
-#else
-	struct lmb_region *_mem = &(_lmb->memory);
-	unsigned long total = 0;
-	int i;
-
-	/* add all physical memory to the bootmem map */
-	for (i=0; i < _mem->cnt; i++)
-		total += _mem->region[i].size;
-	return total;
-#endif /* CONFIG_MSCHUNKS */
+	return lmb.memory.size;
 }
 
 unsigned long __init
 lmb_end_of_DRAM(void)
 {
-	struct lmb *_lmb = &lmb;
-	struct lmb_region *_mem = &(_lmb->memory);
-	int idx = _mem->cnt - 1;
-
-#ifdef CONFIG_MSCHUNKS
-	return (_mem->region[idx].physbase + _mem->region[idx].size);
-#else
-	return (_mem->region[idx].base + _mem->region[idx].size);
-#endif /* CONFIG_MSCHUNKS */
-
-	return 0;
-}
-
-unsigned long __init
-lmb_abs_to_phys(unsigned long aa)
-{
-	unsigned long i, pa = aa;
-	struct lmb *_lmb = &lmb;
-	struct lmb_region *_mem = &(_lmb->memory);
-
-	for (i=0; i < _mem->cnt; i++) {
-		unsigned long lmbbase = _mem->region[i].base;
-		unsigned long lmbsize = _mem->region[i].size;
-		if ( lmb_addrs_overlap(aa,1,lmbbase,lmbsize) ) {
-			pa = _mem->region[i].physbase + (aa - lmbbase);
-			break;
-		}
-	}
+	int idx = lmb.memory.cnt - 1;
 
-	return pa;
+	return (lmb.memory.region[idx].base + lmb.memory.region[idx].size);
 }
 
 /*
@@ -353,20 +281,19 @@ void __init lmb_enforce_memory_limit(void)
 {
 	extern unsigned long memory_limit;
 	unsigned long i, limit;
-	struct lmb_region *mem = &(lmb.memory);
 
 	if (! memory_limit)
 		return;
 
 	limit = memory_limit;
-	for (i = 0; i < mem->cnt; i++) {
-		if (limit > mem->region[i].size) {
-			limit -= mem->region[i].size;
+	for (i = 0; i < lmb.memory.cnt; i++) {
+		if (limit > lmb.memory.region[i].size) {
+			limit -= lmb.memory.region[i].size;
 			continue;
 		}
 
-		mem->region[i].size = limit;
-		mem->cnt = i + 1;
+		lmb.memory.region[i].size = limit;
+		lmb.memory.cnt = i + 1;
 		break;
 	}
 }
diff --git a/arch/ppc64/kernel/lparcfg.c b/arch/ppc64/kernel/lparcfg.c
index 02e9662..9d034ff 100644
--- a/arch/ppc64/kernel/lparcfg.c
+++ b/arch/ppc64/kernel/lparcfg.c
@@ -29,7 +29,7 @@
 #include <asm/iSeries/HvLpConfig.h>
 #include <asm/lppaca.h>
 #include <asm/hvcall.h>
-#include <asm/cputable.h>
+#include <asm/firmware.h>
 #include <asm/rtas.h>
 #include <asm/system.h>
 #include <asm/time.h>
@@ -377,7 +377,7 @@ static int lparcfg_data(struct seq_file *m, void *v)
 
 	partition_active_processors = lparcfg_count_active_processors();
 
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR) {
+	if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
 		unsigned long h_entitled, h_unallocated;
 		unsigned long h_aggregation, h_resource;
 		unsigned long pool_idle_time, pool_procs;
@@ -571,7 +571,7 @@ int __init lparcfg_init(void)
 	mode_t mode = S_IRUSR;
 
 	/* Allow writing if we have FW_FEATURE_SPLPAR */
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR) {
+	if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
 		lparcfg_fops.write = lparcfg_write;
 		mode |= S_IWUSR;
 	}
diff --git a/arch/ppc64/kernel/misc.S b/arch/ppc64/kernel/misc.S
index a05b50b..474df0a 100644
--- a/arch/ppc64/kernel/misc.S
+++ b/arch/ppc64/kernel/misc.S
@@ -680,6 +680,104 @@ _GLOBAL(kernel_thread)
 	ld	r30,-16(r1)
 	blr
 
+/*
+ * disable_kernel_fp()
+ * Disable the FPU.
+ */
+_GLOBAL(disable_kernel_fp)
+	mfmsr	r3
+	rldicl	r0,r3,(63-MSR_FP_LG),1
+	rldicl	r3,r0,(MSR_FP_LG+1),0
+	mtmsrd	r3			/* disable use of fpu now */
+	isync
+	blr
+
+/*
+ * giveup_fpu(tsk)
+ * Disable FP for the task given as the argument,
+ * and save the floating-point registers in its thread_struct.
+ * Enables the FPU for use in the kernel on return.
+ */
+_GLOBAL(giveup_fpu)
+	mfmsr	r5
+	ori	r5,r5,MSR_FP
+	mtmsrd	r5			/* enable use of fpu now */
+	isync
+	cmpdi	0,r3,0
+	beqlr-				/* if no previous owner, done */
+	addi	r3,r3,THREAD		/* want THREAD of task */
+	ld	r5,PT_REGS(r3)
+	cmpdi	0,r5,0
+	SAVE_32FPRS(0, r3)
+	mffs	fr0
+	stfd	fr0,THREAD_FPSCR(r3)
+	beq	1f
+	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+	li	r3,MSR_FP|MSR_FE0|MSR_FE1
+	andc	r4,r4,r3		/* disable FP for previous task */
+	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+1:
+#ifndef CONFIG_SMP
+	li	r5,0
+	ld	r4,last_task_used_math@got(r2)
+	std	r5,0(r4)
+#endif /* CONFIG_SMP */
+	blr
+
+#ifdef CONFIG_ALTIVEC
+
+#if 0 /* this has no callers for now */
+/*
+ * disable_kernel_altivec()
+ * Disable the VMX.
+ */
+_GLOBAL(disable_kernel_altivec)
+	mfmsr	r3
+	rldicl	r0,r3,(63-MSR_VEC_LG),1
+	rldicl	r3,r0,(MSR_VEC_LG+1),0
+	mtmsrd	r3			/* disable use of VMX now */
+	isync
+	blr
+#endif /* 0 */
+
+/*
+ * giveup_altivec(tsk)
+ * Disable VMX for the task given as the argument,
+ * and save the vector registers in its thread_struct.
+ * Enables the VMX for use in the kernel on return.
+ */
+_GLOBAL(giveup_altivec)
+	mfmsr	r5
+	oris	r5,r5,MSR_VEC@h
+	mtmsrd	r5			/* enable use of VMX now */
+	isync
+	cmpdi	0,r3,0
+	beqlr-				/* if no previous owner, done */
+	addi	r3,r3,THREAD		/* want THREAD of task */
+	ld	r5,PT_REGS(r3)
+	cmpdi	0,r5,0
+	SAVE_32VRS(0,r4,r3)
+	mfvscr	vr0
+	li	r4,THREAD_VSCR
+	stvx	vr0,r4,r3
+	beq	1f
+	ld	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+	lis	r3,MSR_VEC@h
+	andc	r4,r4,r3		/* disable FP for previous task */
+	std	r4,_MSR-STACK_FRAME_OVERHEAD(r5)
+1:
+#ifndef CONFIG_SMP
+	li	r5,0
+	ld	r4,last_task_used_altivec@got(r2)
+	std	r5,0(r4)
+#endif /* CONFIG_SMP */
+	blr
+
+#endif /* CONFIG_ALTIVEC */
+
+_GLOBAL(__setup_cpu_power3)
+	blr
+
 /* kexec_wait(phys_cpu)
  *
  * wait for the flag to change, indicating this kernel is going away but
diff --git a/arch/ppc64/kernel/pSeries_iommu.c b/arch/ppc64/kernel/pSeries_iommu.c
index 6913052..9d5e1e7 100644
--- a/arch/ppc64/kernel/pSeries_iommu.c
+++ b/arch/ppc64/kernel/pSeries_iommu.c
@@ -45,6 +45,7 @@
 #include <asm/plpar_wrappers.h>
 #include <asm/pSeries_reconfig.h>
 #include <asm/systemcfg.h>
+#include <asm/firmware.h>
 #include "pci.h"
 
 #define DBG(fmt...)
@@ -546,7 +547,7 @@ void iommu_init_early_pSeries(void)
 	}
 
 	if (systemcfg->platform & PLATFORM_LPAR) {
-		if (cur_cpu_spec->firmware_features & FW_FEATURE_MULTITCE) {
+		if (firmware_has_feature(FW_FEATURE_MULTITCE)) {
 			ppc_md.tce_build = tce_buildmulti_pSeriesLP;
 			ppc_md.tce_free	 = tce_freemulti_pSeriesLP;
 		} else {
diff --git a/arch/ppc64/kernel/pSeries_lpar.c b/arch/ppc64/kernel/pSeries_lpar.c
index 74dd144..0a3ddc9 100644
--- a/arch/ppc64/kernel/pSeries_lpar.c
+++ b/arch/ppc64/kernel/pSeries_lpar.c
@@ -52,7 +52,6 @@ EXPORT_SYMBOL(plpar_hcall_4out);
 EXPORT_SYMBOL(plpar_hcall_norets);
 EXPORT_SYMBOL(plpar_hcall_8arg_2ret);
 
-extern void fw_feature_init(void);
 extern void pSeries_find_serial_port(void);
 
 
@@ -279,7 +278,6 @@ long pSeries_lpar_hpte_insert(unsigned long hpte_group,
 			      unsigned long va, unsigned long prpn,
 			      unsigned long vflags, unsigned long rflags)
 {
-	unsigned long arpn = physRpn_to_absRpn(prpn);
 	unsigned long lpar_rc;
 	unsigned long flags;
 	unsigned long slot;
@@ -290,7 +288,7 @@ long pSeries_lpar_hpte_insert(unsigned long hpte_group,
 	if (vflags & HPTE_V_LARGE)
 		hpte_v &= ~(1UL << HPTE_V_AVPN_SHIFT);
 
-	hpte_r = (arpn << HPTE_R_RPN_SHIFT) | rflags;
+	hpte_r = (prpn << HPTE_R_RPN_SHIFT) | rflags;
 
 	/* Now fill in the actual HPTE */
 	/* Set CEC cookie to 0         */
diff --git a/arch/ppc64/kernel/pSeries_setup.c b/arch/ppc64/kernel/pSeries_setup.c
index 5bec956..f0f0630 100644
--- a/arch/ppc64/kernel/pSeries_setup.c
+++ b/arch/ppc64/kernel/pSeries_setup.c
@@ -60,7 +60,8 @@
 #include <asm/nvram.h>
 #include <asm/plpar_wrappers.h>
 #include <asm/xics.h>
-#include <asm/cputable.h>
+#include <asm/firmware.h>
+#include <asm/pmc.h>
 
 #include "i8259.h"
 #include "mpic.h"
@@ -187,6 +188,21 @@ static void __init pSeries_setup_mpic(void)
 				  " MPIC     ");
 }
 
+static void pseries_lpar_enable_pmcs(void)
+{
+	unsigned long set, reset;
+
+	power4_enable_pmcs();
+
+	set = 1UL << 63;
+	reset = 0;
+	plpar_hcall_norets(H_PERFMON, set, reset);
+
+	/* instruct hypervisor to maintain PMCs */
+	if (firmware_has_feature(FW_FEATURE_SPLPAR))
+		get_paca()->lppaca.pmcregs_in_use = 1;
+}
+
 static void __init pSeries_setup_arch(void)
 {
 	/* Fixup ppc_md depending on the type of interrupt controller */
@@ -231,11 +247,9 @@ static void __init pSeries_setup_arch(void)
 
 	pSeries_nvram_init();
 
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR)
-		vpa_init(boot_cpuid);
-
 	/* Choose an idle loop */
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR) {
+	if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
+		vpa_init(boot_cpuid);
 		if (get_paca()->lppaca.shared_proc) {
 			printk(KERN_INFO "Using shared processor idle loop\n");
 			ppc_md.idle_loop = pseries_shared_idle;
@@ -247,6 +261,11 @@ static void __init pSeries_setup_arch(void)
 		printk(KERN_INFO "Using default idle loop\n");
 		ppc_md.idle_loop = default_idle;
 	}
+
+	if (systemcfg->platform & PLATFORM_LPAR)
+		ppc_md.enable_pmcs = pseries_lpar_enable_pmcs;
+	else
+		ppc_md.enable_pmcs = power4_enable_pmcs;
 }
 
 static int __init pSeries_init_panel(void)
@@ -260,11 +279,11 @@ static int __init pSeries_init_panel(void)
 arch_initcall(pSeries_init_panel);
 
 
-/* Build up the firmware_features bitmask field
+/* Build up the ppc64_firmware_features bitmask field
  * using contents of device-tree/ibm,hypertas-functions.
  * Ultimately this functionality may be moved into prom.c prom_init().
  */
-void __init fw_feature_init(void)
+static void __init fw_feature_init(void)
 {
 	struct device_node * dn;
 	char * hypertas;
@@ -272,7 +291,7 @@ void __init fw_feature_init(void)
 
 	DBG(" -> fw_feature_init()\n");
 
-	cur_cpu_spec->firmware_features = 0;
+	ppc64_firmware_features = 0;
 	dn = of_find_node_by_path("/rtas");
 	if (dn == NULL) {
 		printk(KERN_ERR "WARNING ! Cannot find RTAS in device-tree !\n");
@@ -288,7 +307,7 @@ void __init fw_feature_init(void)
 				if ((firmware_features_table[i].name) &&
 				    (strcmp(firmware_features_table[i].name,hypertas))==0) {
 					/* we have a match */
-					cur_cpu_spec->firmware_features |= 
+					ppc64_firmware_features |= 
 						(firmware_features_table[i].val);
 					break;
 				} 
@@ -302,7 +321,7 @@ void __init fw_feature_init(void)
 	of_node_put(dn);
  no_rtas:
 	printk(KERN_INFO "firmware_features = 0x%lx\n", 
-	       cur_cpu_spec->firmware_features);
+	       ppc64_firmware_features);
 
 	DBG(" <- fw_feature_init()\n");
 }
diff --git a/arch/ppc64/kernel/pSeries_smp.c b/arch/ppc64/kernel/pSeries_smp.c
index 62c55a1..79c7f32 100644
--- a/arch/ppc64/kernel/pSeries_smp.c
+++ b/arch/ppc64/kernel/pSeries_smp.c
@@ -41,6 +41,7 @@
 #include <asm/machdep.h>
 #include <asm/xics.h>
 #include <asm/cputable.h>
+#include <asm/firmware.h>
 #include <asm/system.h>
 #include <asm/rtas.h>
 #include <asm/plpar_wrappers.h>
@@ -326,7 +327,7 @@ static void __devinit smp_xics_setup_cpu(int cpu)
 	if (cpu != boot_cpuid)
 		xics_setup_cpu();
 
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR)
+	if (firmware_has_feature(FW_FEATURE_SPLPAR))
 		vpa_init(cpu);
 
 	cpu_clear(cpu, of_spin_map);
diff --git a/arch/ppc64/kernel/pSeries_vio.c b/arch/ppc64/kernel/pSeries_vio.c
new file mode 100644
index 0000000..338f9e1
--- /dev/null
+++ b/arch/ppc64/kernel/pSeries_vio.c
@@ -0,0 +1,266 @@
+/*
+ * IBM PowerPC pSeries Virtual I/O Infrastructure Support.
+ *
+ *    Copyright (c) 2003-2005 IBM Corp.
+ *     Dave Engebretsen engebret@us.ibm.com
+ *     Santiago Leon santil@us.ibm.com
+ *     Hollis Blanchard <hollisb@us.ibm.com>
+ *     Stephen Rothwell
+ *
+ *      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 <linux/init.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/kobject.h>
+#include <asm/iommu.h>
+#include <asm/dma.h>
+#include <asm/vio.h>
+#include <asm/hvcall.h>
+
+extern struct subsystem devices_subsys; /* needed for vio_find_name() */
+
+static void probe_bus_pseries(void)
+{
+	struct device_node *node_vroot, *of_node;
+
+	node_vroot = find_devices("vdevice");
+	if ((node_vroot == NULL) || (node_vroot->child == NULL))
+		/* this machine doesn't do virtual IO, and that's ok */
+		return;
+
+	/*
+	 * Create struct vio_devices for each virtual device in the device tree.
+	 * Drivers will associate with them later.
+	 */
+	for (of_node = node_vroot->child; of_node != NULL;
+			of_node = of_node->sibling) {
+		printk(KERN_DEBUG "%s: processing %p\n", __FUNCTION__, of_node);
+		vio_register_device_node(of_node);
+	}
+}
+
+/**
+ * vio_match_device_pseries: - Tell if a pSeries VIO device matches a
+ *	vio_device_id
+ */
+static int vio_match_device_pseries(const struct vio_device_id *id,
+		const struct vio_dev *dev)
+{
+	return (strncmp(dev->type, id->type, strlen(id->type)) == 0) &&
+			device_is_compatible(dev->dev.platform_data, id->compat);
+}
+
+static void vio_release_device_pseries(struct device *dev)
+{
+	/* XXX free TCE table */
+	of_node_put(dev->platform_data);
+}
+
+static ssize_t viodev_show_devspec(struct device *dev,
+		struct device_attribute *attr, char *buf)
+{
+	struct device_node *of_node = dev->platform_data;
+
+	return sprintf(buf, "%s\n", of_node->full_name);
+}
+DEVICE_ATTR(devspec, S_IRUSR | S_IRGRP | S_IROTH, viodev_show_devspec, NULL);
+
+static void vio_unregister_device_pseries(struct vio_dev *viodev)
+{
+	device_remove_file(&viodev->dev, &dev_attr_devspec);
+}
+
+/**
+ * vio_bus_init_pseries: - Initialize the pSeries virtual IO bus
+ */
+static int __init vio_bus_init_pseries(void)
+{
+	int err;
+
+	err = vio_bus_init(vio_match_device_pseries,
+			vio_unregister_device_pseries,
+			vio_release_device_pseries);
+	if (err == 0)
+		probe_bus_pseries();
+	return err;
+}
+
+__initcall(vio_bus_init_pseries);
+
+/**
+ * vio_build_iommu_table: - gets the dma information from OF and
+ *	builds the TCE tree.
+ * @dev: the virtual device.
+ *
+ * Returns a pointer to the built tce tree, or NULL if it can't
+ * find property.
+*/
+static struct iommu_table *vio_build_iommu_table(struct vio_dev *dev)
+{
+	unsigned int *dma_window;
+	struct iommu_table *newTceTable;
+	unsigned long offset;
+	int dma_window_property_size;
+
+	dma_window = (unsigned int *) get_property(dev->dev.platform_data, "ibm,my-dma-window", &dma_window_property_size);
+	if(!dma_window) {
+		return NULL;
+	}
+
+	newTceTable = (struct iommu_table *) kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
+
+	/*  There should be some code to extract the phys-encoded offset
+		using prom_n_addr_cells(). However, according to a comment
+		on earlier versions, it's always zero, so we don't bother */
+	offset = dma_window[1] >>  PAGE_SHIFT;
+
+	/* TCE table size - measured in tce entries */
+	newTceTable->it_size		= dma_window[4] >> PAGE_SHIFT;
+	/* offset for VIO should always be 0 */
+	newTceTable->it_offset		= offset;
+	newTceTable->it_busno		= 0;
+	newTceTable->it_index		= (unsigned long)dma_window[0];
+	newTceTable->it_type		= TCE_VB;
+
+	return iommu_init_table(newTceTable);
+}
+
+/**
+ * vio_register_device_node: - Register a new vio device.
+ * @of_node:	The OF node for this device.
+ *
+ * Creates and initializes a vio_dev structure from the data in
+ * of_node (dev.platform_data) and adds it to the list of virtual devices.
+ * Returns a pointer to the created vio_dev or NULL if node has
+ * NULL device_type or compatible fields.
+ */
+struct vio_dev * __devinit vio_register_device_node(struct device_node *of_node)
+{
+	struct vio_dev *viodev;
+	unsigned int *unit_address;
+	unsigned int *irq_p;
+
+	/* we need the 'device_type' property, in order to match with drivers */
+	if ((NULL == of_node->type)) {
+		printk(KERN_WARNING
+			"%s: node %s missing 'device_type'\n", __FUNCTION__,
+			of_node->name ? of_node->name : "<unknown>");
+		return NULL;
+	}
+
+	unit_address = (unsigned int *)get_property(of_node, "reg", NULL);
+	if (!unit_address) {
+		printk(KERN_WARNING "%s: node %s missing 'reg'\n", __FUNCTION__,
+			of_node->name ? of_node->name : "<unknown>");
+		return NULL;
+	}
+
+	/* allocate a vio_dev for this node */
+	viodev = kmalloc(sizeof(struct vio_dev), GFP_KERNEL);
+	if (!viodev) {
+		return NULL;
+	}
+	memset(viodev, 0, sizeof(struct vio_dev));
+
+	viodev->dev.platform_data = of_node_get(of_node);
+
+	viodev->irq = NO_IRQ;
+	irq_p = (unsigned int *)get_property(of_node, "interrupts", NULL);
+	if (irq_p) {
+		int virq = virt_irq_create_mapping(*irq_p);
+		if (virq == NO_IRQ) {
+			printk(KERN_ERR "Unable to allocate interrupt "
+			       "number for %s\n", of_node->full_name);
+		} else
+			viodev->irq = irq_offset_up(virq);
+	}
+
+	snprintf(viodev->dev.bus_id, BUS_ID_SIZE, "%x", *unit_address);
+
+	/* register with generic device framework */
+	if (vio_register_device_common(viodev, of_node->name, of_node->type,
+				*unit_address, vio_build_iommu_table(viodev))
+			== NULL) {
+		/* XXX free TCE table */
+		kfree(viodev);
+		return NULL;
+	}
+	device_create_file(&viodev->dev, &dev_attr_devspec);
+
+	return viodev;
+}
+EXPORT_SYMBOL(vio_register_device_node);
+
+/**
+ * vio_get_attribute: - get attribute for virtual device
+ * @vdev:	The vio device to get property.
+ * @which:	The property/attribute to be extracted.
+ * @length:	Pointer to length of returned data size (unused if NULL).
+ *
+ * Calls prom.c's get_property() to return the value of the
+ * attribute specified by the preprocessor constant @which
+*/
+const void * vio_get_attribute(struct vio_dev *vdev, void* which, int* length)
+{
+	return get_property(vdev->dev.platform_data, (char*)which, length);
+}
+EXPORT_SYMBOL(vio_get_attribute);
+
+/* vio_find_name() - internal because only vio.c knows how we formatted the
+ * kobject name
+ * XXX once vio_bus_type.devices is actually used as a kset in
+ * drivers/base/bus.c, this function should be removed in favor of
+ * "device_find(kobj_name, &vio_bus_type)"
+ */
+static struct vio_dev *vio_find_name(const char *kobj_name)
+{
+	struct kobject *found;
+
+	found = kset_find_obj(&devices_subsys.kset, kobj_name);
+	if (!found)
+		return NULL;
+
+	return to_vio_dev(container_of(found, struct device, kobj));
+}
+
+/**
+ * vio_find_node - find an already-registered vio_dev
+ * @vnode: device_node of the virtual device we're looking for
+ */
+struct vio_dev *vio_find_node(struct device_node *vnode)
+{
+	uint32_t *unit_address;
+	char kobj_name[BUS_ID_SIZE];
+
+	/* construct the kobject name from the device node */
+	unit_address = (uint32_t *)get_property(vnode, "reg", NULL);
+	if (!unit_address)
+		return NULL;
+	snprintf(kobj_name, BUS_ID_SIZE, "%x", *unit_address);
+
+	return vio_find_name(kobj_name);
+}
+EXPORT_SYMBOL(vio_find_node);
+
+int vio_enable_interrupts(struct vio_dev *dev)
+{
+	int rc = h_vio_signal(dev->unit_address, VIO_IRQ_ENABLE);
+	if (rc != H_Success)
+		printk(KERN_ERR "vio: Error 0x%x enabling interrupts\n", rc);
+	return rc;
+}
+EXPORT_SYMBOL(vio_enable_interrupts);
+
+int vio_disable_interrupts(struct vio_dev *dev)
+{
+	int rc = h_vio_signal(dev->unit_address, VIO_IRQ_DISABLE);
+	if (rc != H_Success)
+		printk(KERN_ERR "vio: Error 0x%x disabling interrupts\n", rc);
+	return rc;
+}
+EXPORT_SYMBOL(vio_disable_interrupts);
diff --git a/arch/ppc64/kernel/pacaData.c b/arch/ppc64/kernel/pacaData.c
index 6316188..6182a2c 100644
--- a/arch/ppc64/kernel/pacaData.c
+++ b/arch/ppc64/kernel/pacaData.c
@@ -78,7 +78,7 @@ extern unsigned long __toc_start;
 
 #define BOOTCPU_PACA_INIT(number)					    \
 {									    \
-	PACA_INIT_COMMON(number, 1, 0, STAB0_VIRT_ADDR)			    \
+	PACA_INIT_COMMON(number, 1, 0, (u64)&initial_stab)		    \
 	PACA_INIT_ISERIES(number)					    \
 }
 
@@ -90,7 +90,7 @@ extern unsigned long __toc_start;
 
 #define BOOTCPU_PACA_INIT(number)					    \
 {									    \
-	PACA_INIT_COMMON(number, 1, STAB0_PHYS_ADDR, STAB0_VIRT_ADDR)	    \
+	PACA_INIT_COMMON(number, 1, STAB0_PHYS_ADDR, (u64)&initial_stab)    \
 }
 #endif
 
diff --git a/arch/ppc64/kernel/pmac_setup.c b/arch/ppc64/kernel/pmac_setup.c
index e40877f..8ff86a7 100644
--- a/arch/ppc64/kernel/pmac_setup.c
+++ b/arch/ppc64/kernel/pmac_setup.c
@@ -71,6 +71,7 @@
 #include <asm/of_device.h>
 #include <asm/lmb.h>
 #include <asm/smu.h>
+#include <asm/pmc.h>
 
 #include "pmac.h"
 #include "mpic.h"
@@ -511,4 +512,5 @@ struct machdep_calls __initdata pmac_md = {
 	.progress		= pmac_progress,
 	.check_legacy_ioport	= pmac_check_legacy_ioport,
 	.idle_loop		= native_idle,
+	.enable_pmcs		= power4_enable_pmcs,
 };
diff --git a/arch/ppc64/kernel/pmc.c b/arch/ppc64/kernel/pmc.c
index 67be773..cdfec74 100644
--- a/arch/ppc64/kernel/pmc.c
+++ b/arch/ppc64/kernel/pmc.c
@@ -65,3 +65,24 @@ void release_pmc_hardware(void)
 	spin_unlock(&pmc_owner_lock);
 }
 EXPORT_SYMBOL_GPL(release_pmc_hardware);
+
+void power4_enable_pmcs(void)
+{
+	unsigned long hid0;
+
+	hid0 = mfspr(HID0);
+	hid0 |= 1UL << (63 - 20);
+
+	/* POWER4 requires the following sequence */
+	asm volatile(
+		"sync\n"
+		"mtspr     %1, %0\n"
+		"mfspr     %0, %1\n"
+		"mfspr     %0, %1\n"
+		"mfspr     %0, %1\n"
+		"mfspr     %0, %1\n"
+		"mfspr     %0, %1\n"
+		"mfspr     %0, %1\n"
+		"isync" : "=&r" (hid0) : "i" (HID0), "0" (hid0):
+		"memory");
+}
diff --git a/arch/ppc64/kernel/process.c b/arch/ppc64/kernel/process.c
index f7cae05..7a7e027 100644
--- a/arch/ppc64/kernel/process.c
+++ b/arch/ppc64/kernel/process.c
@@ -50,6 +50,7 @@
 #include <asm/machdep.h>
 #include <asm/iSeries/HvCallHpt.h>
 #include <asm/cputable.h>
+#include <asm/firmware.h>
 #include <asm/sections.h>
 #include <asm/tlbflush.h>
 #include <asm/time.h>
@@ -202,11 +203,10 @@ struct task_struct *__switch_to(struct task_struct *prev,
 	new_thread = &new->thread;
 	old_thread = &current->thread;
 
-/* Collect purr utilization data per process and per processor wise */
-/* purr is nothing but processor time base                          */
-
-#if defined(CONFIG_PPC_PSERIES)
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR) {
+	/* Collect purr utilization data per process and per processor
+	 * wise purr is nothing but processor time base
+	 */
+	if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
 		struct cpu_usage *cu = &__get_cpu_var(cpu_usage_array);
 		long unsigned start_tb, current_tb;
 		start_tb = old_thread->start_tb;
@@ -214,8 +214,6 @@ struct task_struct *__switch_to(struct task_struct *prev,
 		old_thread->accum_tb += (current_tb - start_tb);
 		new_thread->start_tb = current_tb;
 	}
-#endif
-
 
 	local_irq_save(flags);
 	last = _switch(old_thread, new_thread);
diff --git a/arch/ppc64/kernel/prom.c b/arch/ppc64/kernel/prom.c
index 5aca01d..b218488 100644
--- a/arch/ppc64/kernel/prom.c
+++ b/arch/ppc64/kernel/prom.c
@@ -625,8 +625,8 @@ void __init finish_device_tree(void)
 
 static inline char *find_flat_dt_string(u32 offset)
 {
-	return ((char *)initial_boot_params) + initial_boot_params->off_dt_strings
-		+ offset;
+	return ((char *)initial_boot_params) +
+		initial_boot_params->off_dt_strings + offset;
 }
 
 /**
@@ -635,26 +635,33 @@ static inline char *find_flat_dt_string(u32 offset)
  * unflatten the tree
  */
 static int __init scan_flat_dt(int (*it)(unsigned long node,
-					 const char *full_path, void *data),
+					 const char *uname, int depth,
+					 void *data),
 			       void *data)
 {
 	unsigned long p = ((unsigned long)initial_boot_params) +
 		initial_boot_params->off_dt_struct;
 	int rc = 0;
+	int depth = -1;
 
 	do {
 		u32 tag = *((u32 *)p);
 		char *pathp;
 		
 		p += 4;
-		if (tag == OF_DT_END_NODE)
+		if (tag == OF_DT_END_NODE) {
+			depth --;
+			continue;
+		}
+		if (tag == OF_DT_NOP)
 			continue;
 		if (tag == OF_DT_END)
 			break;
 		if (tag == OF_DT_PROP) {
 			u32 sz = *((u32 *)p);
 			p += 8;
-			p = _ALIGN(p, sz >= 8 ? 8 : 4);
+			if (initial_boot_params->version < 0x10)
+				p = _ALIGN(p, sz >= 8 ? 8 : 4);
 			p += sz;
 			p = _ALIGN(p, 4);
 			continue;
@@ -664,9 +671,18 @@ static int __init scan_flat_dt(int (*it)(unsigned long node,
 			       " device tree !\n", tag);
 			return -EINVAL;
 		}
+		depth++;
 		pathp = (char *)p;
 		p = _ALIGN(p + strlen(pathp) + 1, 4);
-		rc = it(p, pathp, data);
+		if ((*pathp) == '/') {
+			char *lp, *np;
+			for (lp = NULL, np = pathp; *np; np++)
+				if ((*np) == '/')
+					lp = np+1;
+			if (lp != NULL)
+				pathp = lp;
+		}
+		rc = it(p, pathp, depth, data);
 		if (rc != 0)
 			break;		
 	} while(1);
@@ -689,17 +705,21 @@ static void* __init get_flat_dt_prop(unsigned long node, const char *name,
 		const char *nstr;
 
 		p += 4;
+		if (tag == OF_DT_NOP)
+			continue;
 		if (tag != OF_DT_PROP)
 			return NULL;
 
 		sz = *((u32 *)p);
 		noff = *((u32 *)(p + 4));
 		p += 8;
-		p = _ALIGN(p, sz >= 8 ? 8 : 4);
+		if (initial_boot_params->version < 0x10)
+			p = _ALIGN(p, sz >= 8 ? 8 : 4);
 
 		nstr = find_flat_dt_string(noff);
 		if (nstr == NULL) {
-			printk(KERN_WARNING "Can't find property index name !\n");
+			printk(KERN_WARNING "Can't find property index"
+			       " name !\n");
 			return NULL;
 		}
 		if (strcmp(name, nstr) == 0) {
@@ -713,7 +733,7 @@ static void* __init get_flat_dt_prop(unsigned long node, const char *name,
 }
 
 static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size,
-					       unsigned long align)
+				       unsigned long align)
 {
 	void *res;
 
@@ -727,13 +747,16 @@ static void *__init unflatten_dt_alloc(unsigned long *mem, unsigned long size,
 static unsigned long __init unflatten_dt_node(unsigned long mem,
 					      unsigned long *p,
 					      struct device_node *dad,
-					      struct device_node ***allnextpp)
+					      struct device_node ***allnextpp,
+					      unsigned long fpsize)
 {
 	struct device_node *np;
 	struct property *pp, **prev_pp = NULL;
 	char *pathp;
 	u32 tag;
-	unsigned int l;
+	unsigned int l, allocl;
+	int has_name = 0;
+	int new_format = 0;
 
 	tag = *((u32 *)(*p));
 	if (tag != OF_DT_BEGIN_NODE) {
@@ -742,21 +765,62 @@ static unsigned long __init unflatten_dt_node(unsigned long mem,
 	}
 	*p += 4;
 	pathp = (char *)*p;
-	l = strlen(pathp) + 1;
+	l = allocl = strlen(pathp) + 1;
 	*p = _ALIGN(*p + l, 4);
 
-	np = unflatten_dt_alloc(&mem, sizeof(struct device_node) + l,
+	/* version 0x10 has a more compact unit name here instead of the full
+	 * path. we accumulate the full path size using "fpsize", we'll rebuild
+	 * it later. We detect this because the first character of the name is
+	 * not '/'.
+	 */
+	if ((*pathp) != '/') {
+		new_format = 1;
+		if (fpsize == 0) {
+			/* root node: special case. fpsize accounts for path
+			 * plus terminating zero. root node only has '/', so
+			 * fpsize should be 2, but we want to avoid the first
+			 * level nodes to have two '/' so we use fpsize 1 here
+			 */
+			fpsize = 1;
+			allocl = 2;
+		} else {
+			/* account for '/' and path size minus terminal 0
+			 * already in 'l'
+			 */
+			fpsize += l;
+			allocl = fpsize;
+		}
+	}
+
+
+	np = unflatten_dt_alloc(&mem, sizeof(struct device_node) + allocl,
 				__alignof__(struct device_node));
 	if (allnextpp) {
 		memset(np, 0, sizeof(*np));
 		np->full_name = ((char*)np) + sizeof(struct device_node);
-		memcpy(np->full_name, pathp, l);
+		if (new_format) {
+			char *p = np->full_name;
+			/* rebuild full path for new format */
+			if (dad && dad->parent) {
+				strcpy(p, dad->full_name);
+#ifdef DEBUG
+				if ((strlen(p) + l + 1) != allocl) {
+					DBG("%s: p: %d, l: %d, a: %d\n",
+					    pathp, strlen(p), l, allocl);
+				}
+#endif
+				p += strlen(p);
+			}
+			*(p++) = '/';
+			memcpy(p, pathp, l);
+		} else
+			memcpy(np->full_name, pathp, l);
 		prev_pp = &np->properties;
 		**allnextpp = np;
 		*allnextpp = &np->allnext;
 		if (dad != NULL) {
 			np->parent = dad;
-			/* we temporarily use the `next' field as `last_child'. */
+			/* we temporarily use the next field as `last_child'*/
 			if (dad->next == 0)
 				dad->child = np;
 			else
@@ -770,18 +834,26 @@ static unsigned long __init unflatten_dt_node(unsigned long mem,
 		char *pname;
 
 		tag = *((u32 *)(*p));
+		if (tag == OF_DT_NOP) {
+			*p += 4;
+			continue;
+		}
 		if (tag != OF_DT_PROP)
 			break;
 		*p += 4;
 		sz = *((u32 *)(*p));
 		noff = *((u32 *)((*p) + 4));
-		*p = _ALIGN((*p) + 8, sz >= 8 ? 8 : 4);
+		*p += 8;
+		if (initial_boot_params->version < 0x10)
+			*p = _ALIGN(*p, sz >= 8 ? 8 : 4);
 
 		pname = find_flat_dt_string(noff);
 		if (pname == NULL) {
 			printk("Can't find property name in list !\n");
 			break;
 		}
+		if (strcmp(pname, "name") == 0)
+			has_name = 1;
 		l = strlen(pname) + 1;
 		pp = unflatten_dt_alloc(&mem, sizeof(struct property),
 					__alignof__(struct property));
@@ -801,6 +873,36 @@ static unsigned long __init unflatten_dt_node(unsigned long mem,
 		}
 		*p = _ALIGN((*p) + sz, 4);
 	}
+	/* with version 0x10 we may not have the name property, recreate
+	 * it here from the unit name if absent
+	 */
+	if (!has_name) {
+		char *p = pathp, *ps = pathp, *pa = NULL;
+		int sz;
+
+		while (*p) {
+			if ((*p) == '@')
+				pa = p;
+			if ((*p) == '/')
+				ps = p + 1;
+			p++;
+		}
+		if (pa < ps)
+			pa = p;
+		sz = (pa - ps) + 1;
+		pp = unflatten_dt_alloc(&mem, sizeof(struct property) + sz,
+					__alignof__(struct property));
+		if (allnextpp) {
+			pp->name = "name";
+			pp->length = sz;
+			pp->value = (unsigned char *)(pp + 1);
+			*prev_pp = pp;
+			prev_pp = &pp->next;
+			memcpy(pp->value, ps, sz - 1);
+			((char *)pp->value)[sz - 1] = 0;
+			DBG("fixed up name for %s -> %s\n", pathp, pp->value);
+		}
+	}
 	if (allnextpp) {
 		*prev_pp = NULL;
 		np->name = get_property(np, "name", NULL);
@@ -812,11 +914,11 @@ static unsigned long __init unflatten_dt_node(unsigned long mem,
 			np->type = "<NULL>";
 	}
 	while (tag == OF_DT_BEGIN_NODE) {
-		mem = unflatten_dt_node(mem, p, np, allnextpp);
+		mem = unflatten_dt_node(mem, p, np, allnextpp, fpsize);
 		tag = *((u32 *)(*p));
 	}
 	if (tag != OF_DT_END_NODE) {
-		printk("Weird tag at start of node: %x\n", tag);
+		printk("Weird tag at end of node: %x\n", tag);
 		return mem;
 	}
 	*p += 4;
@@ -842,21 +944,32 @@ void __init unflatten_device_tree(void)
 	/* First pass, scan for size */
 	start = ((unsigned long)initial_boot_params) +
 		initial_boot_params->off_dt_struct;
-	size = unflatten_dt_node(0, &start, NULL, NULL);
+	size = unflatten_dt_node(0, &start, NULL, NULL, 0);
+	size = (size | 3) + 1;
 
 	DBG("  size is %lx, allocating...\n", size);
 
 	/* Allocate memory for the expanded device tree */
-	mem = (unsigned long)abs_to_virt(lmb_alloc(size,
-						   __alignof__(struct device_node)));
+	mem = lmb_alloc(size + 4, __alignof__(struct device_node));
+	if (!mem) {
+		DBG("Couldn't allocate memory with lmb_alloc()!\n");
+		panic("Couldn't allocate memory with lmb_alloc()!\n");
+	}
+	mem = (unsigned long)abs_to_virt(mem);
+
+	((u32 *)mem)[size / 4] = 0xdeadbeef;
+
 	DBG("  unflattening...\n", mem);
 
 	/* Second pass, do actual unflattening */
 	start = ((unsigned long)initial_boot_params) +
 		initial_boot_params->off_dt_struct;
-	unflatten_dt_node(mem, &start, NULL, &allnextp);
+	unflatten_dt_node(mem, &start, NULL, &allnextp, 0);
 	if (*((u32 *)start) != OF_DT_END)
-		printk(KERN_WARNING "Weird tag at end of tree: %x\n", *((u32 *)start));
+		printk(KERN_WARNING "Weird tag at end of tree: %08x\n", *((u32 *)start));
+	if (((u32 *)mem)[size / 4] != 0xdeadbeef)
+		printk(KERN_WARNING "End of tree marker overwritten: %08x\n",
+		       ((u32 *)mem)[size / 4] );
 	*allnextp = NULL;
 
 	/* Get pointer to OF "/chosen" node for use everywhere */
@@ -880,7 +993,7 @@ void __init unflatten_device_tree(void)
 
 
 static int __init early_init_dt_scan_cpus(unsigned long node,
-					  const char *full_path, void *data)
+					  const char *uname, int depth, void *data)
 {
 	char *type = get_flat_dt_prop(node, "device_type", NULL);
 	u32 *prop;
@@ -947,13 +1060,15 @@ static int __init early_init_dt_scan_cpus(unsigned long node,
 }
 
 static int __init early_init_dt_scan_chosen(unsigned long node,
-					    const char *full_path, void *data)
+					    const char *uname, int depth, void *data)
 {
 	u32 *prop;
 	u64 *prop64;
 	extern unsigned long memory_limit, tce_alloc_start, tce_alloc_end;
 
-	if (strcmp(full_path, "/chosen") != 0)
+	DBG("search \"chosen\", depth: %d, uname: %s\n", depth, uname);
+
+	if (depth != 1 || strcmp(uname, "chosen") != 0)
 		return 0;
 
 	/* get platform type */
@@ -1003,18 +1118,20 @@ static int __init early_init_dt_scan_chosen(unsigned long node,
 }
 
 static int __init early_init_dt_scan_root(unsigned long node,
-					  const char *full_path, void *data)
+					  const char *uname, int depth, void *data)
 {
 	u32 *prop;
 
-	if (strcmp(full_path, "/") != 0)
+	if (depth != 0)
 		return 0;
 
 	prop = (u32 *)get_flat_dt_prop(node, "#size-cells", NULL);
 	dt_root_size_cells = (prop == NULL) ? 1 : *prop;
-		
+	DBG("dt_root_size_cells = %x\n", dt_root_size_cells);
+
 	prop = (u32 *)get_flat_dt_prop(node, "#address-cells", NULL);
 	dt_root_addr_cells = (prop == NULL) ? 2 : *prop;
+	DBG("dt_root_addr_cells = %x\n", dt_root_addr_cells);
 	
 	/* break now */
 	return 1;
@@ -1042,7 +1159,7 @@ static unsigned long __init dt_mem_next_cell(int s, cell_t **cellp)
 
 
 static int __init early_init_dt_scan_memory(unsigned long node,
-					    const char *full_path, void *data)
+					    const char *uname, int depth, void *data)
 {
 	char *type = get_flat_dt_prop(node, "device_type", NULL);
 	cell_t *reg, *endp;
@@ -1058,7 +1175,9 @@ static int __init early_init_dt_scan_memory(unsigned long node,
 
 	endp = reg + (l / sizeof(cell_t));
 
-	DBG("memory scan node %s ...\n", full_path);
+	DBG("memory scan node %s ..., reg size %ld, data: %x %x %x %x, ...\n",
+	    uname, l, reg[0], reg[1], reg[2], reg[3]);
+
 	while ((endp - reg) >= (dt_root_addr_cells + dt_root_size_cells)) {
 		unsigned long base, size;
 
@@ -1469,10 +1588,11 @@ struct device_node *of_find_node_by_path(const char *path)
 	struct device_node *np = allnodes;
 
 	read_lock(&devtree_lock);
-	for (; np != 0; np = np->allnext)
+	for (; np != 0; np = np->allnext) {
 		if (np->full_name != 0 && strcasecmp(np->full_name, path) == 0
 		    && of_node_get(np))
 			break;
+	}
 	read_unlock(&devtree_lock);
 	return np;
 }
diff --git a/arch/ppc64/kernel/prom_init.c b/arch/ppc64/kernel/prom_init.c
index dbbe6c79..adcf972 100644
--- a/arch/ppc64/kernel/prom_init.c
+++ b/arch/ppc64/kernel/prom_init.c
@@ -1534,7 +1534,8 @@ static unsigned long __init dt_find_string(char *str)
  */
 #define MAX_PROPERTY_NAME 64
 
-static void __init scan_dt_build_strings(phandle node, unsigned long *mem_start,
+static void __init scan_dt_build_strings(phandle node,
+					 unsigned long *mem_start,
 					 unsigned long *mem_end)
 {
 	unsigned long offset = reloc_offset();
@@ -1547,16 +1548,21 @@ static void __init scan_dt_build_strings(phandle node, unsigned long *mem_start,
 	/* get and store all property names */
 	prev_name = RELOC("");
 	for (;;) {
-		int rc;
-
 		/* 64 is max len of name including nul. */
 		namep = make_room(mem_start, mem_end, MAX_PROPERTY_NAME, 1);
-		rc = call_prom("nextprop", 3, 1, node, prev_name, namep);
-		if (rc != 1) {
+		if (call_prom("nextprop", 3, 1, node, prev_name, namep) != 1) {
 			/* No more nodes: unwind alloc */
 			*mem_start = (unsigned long)namep;
 			break;
 		}
+
+ 		/* skip "name" */
+ 		if (strcmp(namep, RELOC("name")) == 0) {
+ 			*mem_start = (unsigned long)namep;
+ 			prev_name = RELOC("name");
+ 			continue;
+ 		}
+		/* get/create string entry */
 		soff = dt_find_string(namep);
 		if (soff != 0) {
 			*mem_start = (unsigned long)namep;
@@ -1571,7 +1577,7 @@ static void __init scan_dt_build_strings(phandle node, unsigned long *mem_start,
 
 	/* do all our children */
 	child = call_prom("child", 1, 1, node);
-	while (child != (phandle)0) {
+	while (child != 0) {
 		scan_dt_build_strings(child, mem_start, mem_end);
 		child = call_prom("peer", 1, 1, child);
 	}
@@ -1580,16 +1586,13 @@ static void __init scan_dt_build_strings(phandle node, unsigned long *mem_start,
 static void __init scan_dt_build_struct(phandle node, unsigned long *mem_start,
 					unsigned long *mem_end)
 {
-	int l, align;
 	phandle child;
-	char *namep, *prev_name, *sstart, *p, *ep;
+	char *namep, *prev_name, *sstart, *p, *ep, *lp, *path;
 	unsigned long soff;
 	unsigned char *valp;
 	unsigned long offset = reloc_offset();
-	char pname[MAX_PROPERTY_NAME];
-	char *path;
-
-	path = RELOC(prom_scratch);
+	static char pname[MAX_PROPERTY_NAME];
+	int l;
 
 	dt_push_token(OF_DT_BEGIN_NODE, mem_start, mem_end);
 
@@ -1599,23 +1602,33 @@ static void __init scan_dt_build_struct(phandle node, unsigned long *mem_start,
 		      namep, *mem_end - *mem_start);
 	if (l >= 0) {
 		/* Didn't fit?  Get more room. */
-		if (l+1 > *mem_end - *mem_start) {
+		if ((l+1) > (*mem_end - *mem_start)) {
 			namep = make_room(mem_start, mem_end, l+1, 1);
 			call_prom("package-to-path", 3, 1, node, namep, l);
 		}
 		namep[l] = '\0';
+
 		/* Fixup an Apple bug where they have bogus \0 chars in the
 		 * middle of the path in some properties
 		 */
 		for (p = namep, ep = namep + l; p < ep; p++)
 			if (*p == '\0') {
 				memmove(p, p+1, ep - p);
-				ep--; l--;
+				ep--; l--; p--;
 			}
-		*mem_start = _ALIGN(((unsigned long) namep) + strlen(namep) + 1, 4);
+
+		/* now try to extract the unit name in that mess */
+		for (p = namep, lp = NULL; *p; p++)
+			if (*p == '/')
+				lp = p + 1;
+		if (lp != NULL)
+			memmove(namep, lp, strlen(lp) + 1);
+		*mem_start = _ALIGN(((unsigned long) namep) +
+				    strlen(namep) + 1, 4);
 	}
 
 	/* get it again for debugging */
+	path = RELOC(prom_scratch);
 	memset(path, 0, PROM_SCRATCH_SIZE);
 	call_prom("package-to-path", 3, 1, node, path, PROM_SCRATCH_SIZE-1);
 
@@ -1623,23 +1636,27 @@ static void __init scan_dt_build_struct(phandle node, unsigned long *mem_start,
 	prev_name = RELOC("");
 	sstart = (char *)RELOC(dt_string_start);
 	for (;;) {
-		int rc;
-
-		rc = call_prom("nextprop", 3, 1, node, prev_name, pname);
-		if (rc != 1)
+		if (call_prom("nextprop", 3, 1, node, prev_name,
+			      RELOC(pname)) != 1)
 			break;
 
+ 		/* skip "name" */
+ 		if (strcmp(RELOC(pname), RELOC("name")) == 0) {
+ 			prev_name = RELOC("name");
+ 			continue;
+ 		}
+
 		/* find string offset */
-		soff = dt_find_string(pname);
+		soff = dt_find_string(RELOC(pname));
 		if (soff == 0) {
-			prom_printf("WARNING: Can't find string index for <%s>, node %s\n",
-				    pname, path);
+			prom_printf("WARNING: Can't find string index for"
+				    " <%s>, node %s\n", RELOC(pname), path);
 			break;
 		}
 		prev_name = sstart + soff;
 
 		/* get length */
-		l = call_prom("getproplen", 2, 1, node, pname);
+		l = call_prom("getproplen", 2, 1, node, RELOC(pname));
 
 		/* sanity checks */
 		if (l == PROM_ERROR)
@@ -1648,7 +1665,7 @@ static void __init scan_dt_build_struct(phandle node, unsigned long *mem_start,
 			prom_printf("WARNING: ignoring large property ");
 			/* It seems OF doesn't null-terminate the path :-( */
 			prom_printf("[%s] ", path);
-			prom_printf("%s length 0x%x\n", pname, l);
+			prom_printf("%s length 0x%x\n", RELOC(pname), l);
 			continue;
 		}
 
@@ -1658,17 +1675,16 @@ static void __init scan_dt_build_struct(phandle node, unsigned long *mem_start,
 		dt_push_token(soff, mem_start, mem_end);
 
 		/* push property content */
-		align = (l >= 8) ? 8 : 4;
-		valp = make_room(mem_start, mem_end, l, align);
-		call_prom("getprop", 4, 1, node, pname, valp, l);
+		valp = make_room(mem_start, mem_end, l, 4);
+		call_prom("getprop", 4, 1, node, RELOC(pname), valp, l);
 		*mem_start = _ALIGN(*mem_start, 4);
 	}
 
 	/* Add a "linux,phandle" property. */
 	soff = dt_find_string(RELOC("linux,phandle"));
 	if (soff == 0)
-		prom_printf("WARNING: Can't find string index for <linux-phandle>"
-			    " node %s\n", path);
+		prom_printf("WARNING: Can't find string index for"
+			    " <linux-phandle> node %s\n", path);
 	else {
 		dt_push_token(OF_DT_PROP, mem_start, mem_end);
 		dt_push_token(4, mem_start, mem_end);
@@ -1679,7 +1695,7 @@ static void __init scan_dt_build_struct(phandle node, unsigned long *mem_start,
 
 	/* do all our children */
 	child = call_prom("child", 1, 1, node);
-	while (child != (phandle)0) {
+	while (child != 0) {
 		scan_dt_build_struct(child, mem_start, mem_end);
 		child = call_prom("peer", 1, 1, child);
 	}
@@ -1718,7 +1734,8 @@ static void __init flatten_device_tree(void)
 
 	/* Build header and make room for mem rsv map */ 
 	mem_start = _ALIGN(mem_start, 4);
-	hdr = make_room(&mem_start, &mem_end, sizeof(struct boot_param_header), 4);
+	hdr = make_room(&mem_start, &mem_end,
+			sizeof(struct boot_param_header), 4);
 	RELOC(dt_header_start) = (unsigned long)hdr;
 	rsvmap = make_room(&mem_start, &mem_end, sizeof(mem_reserve_map), 8);
 
@@ -1731,11 +1748,11 @@ static void __init flatten_device_tree(void)
 	namep = make_room(&mem_start, &mem_end, 16, 1);
 	strcpy(namep, RELOC("linux,phandle"));
 	mem_start = (unsigned long)namep + strlen(namep) + 1;
-	RELOC(dt_string_end) = mem_start;
 
 	/* Build string array */
 	prom_printf("Building dt strings...\n"); 
 	scan_dt_build_strings(root, &mem_start, &mem_end);
+	RELOC(dt_string_end) = mem_start;
 
 	/* Build structure */
 	mem_start = PAGE_ALIGN(mem_start);
@@ -1750,9 +1767,11 @@ static void __init flatten_device_tree(void)
 	hdr->totalsize = RELOC(dt_struct_end) - RELOC(dt_header_start);
 	hdr->off_dt_struct = RELOC(dt_struct_start) - RELOC(dt_header_start);
 	hdr->off_dt_strings = RELOC(dt_string_start) - RELOC(dt_header_start);
+	hdr->dt_strings_size = RELOC(dt_string_end) - RELOC(dt_string_start);
 	hdr->off_mem_rsvmap = ((unsigned long)rsvmap) - RELOC(dt_header_start);
 	hdr->version = OF_DT_VERSION;
-	hdr->last_comp_version = 1;
+	/* Version 16 is not backward compatible */
+	hdr->last_comp_version = 0x10;
 
 	/* Reserve the whole thing and copy the reserve map in, we
 	 * also bump mem_reserve_cnt to cause further reservations to
@@ -1808,6 +1827,9 @@ static void __init fixup_device_tree(void)
 	/* does it need fixup ? */
 	if (prom_getproplen(i2c, "interrupts") > 0)
 		return;
+
+	prom_printf("fixing up bogus interrupts for u3 i2c...\n");
+
 	/* interrupt on this revision of u3 is number 0 and level */
 	interrupts[0] = 0;
 	interrupts[1] = 1;
diff --git a/arch/ppc64/kernel/rtas_pci.c b/arch/ppc64/kernel/rtas_pci.c
index 1048817..1dccada 100644
--- a/arch/ppc64/kernel/rtas_pci.c
+++ b/arch/ppc64/kernel/rtas_pci.c
@@ -58,6 +58,21 @@ static int config_access_valid(struct device_node *dn, int where)
 	return 0;
 }
 
+static int of_device_available(struct device_node * dn)
+{
+        char * status;
+
+        status = get_property(dn, "status", NULL);
+
+        if (!status)
+                return 1;
+
+        if (!strcmp(status, "okay"))
+                return 1;
+
+        return 0;
+}
+
 static int rtas_read_config(struct device_node *dn, int where, int size, u32 *val)
 {
 	int returnval = -1;
@@ -103,7 +118,7 @@ static int rtas_pci_read_config(struct pci_bus *bus,
 
 	/* Search only direct children of the bus */
 	for (dn = busdn->child; dn; dn = dn->sibling)
-		if (dn->devfn == devfn)
+		if (dn->devfn == devfn && of_device_available(dn))
 			return rtas_read_config(dn, where, size, val);
 	return PCIBIOS_DEVICE_NOT_FOUND;
 }
@@ -146,7 +161,7 @@ static int rtas_pci_write_config(struct pci_bus *bus,
 
 	/* Search only direct children of the bus */
 	for (dn = busdn->child; dn; dn = dn->sibling)
-		if (dn->devfn == devfn)
+		if (dn->devfn == devfn && of_device_available(dn))
 			return rtas_write_config(dn, where, size, val);
 	return PCIBIOS_DEVICE_NOT_FOUND;
 }
diff --git a/arch/ppc64/kernel/setup.c b/arch/ppc64/kernel/setup.c
index e9c24d2..ee3b20d 100644
--- a/arch/ppc64/kernel/setup.c
+++ b/arch/ppc64/kernel/setup.c
@@ -536,15 +536,19 @@ static void __init check_for_initrd(void)
 
 	DBG(" -> check_for_initrd()\n");
 
-	prop = (u64 *)get_property(of_chosen, "linux,initrd-start", NULL);
-	if (prop != NULL) {
-		initrd_start = (unsigned long)__va(*prop);
-		prop = (u64 *)get_property(of_chosen, "linux,initrd-end", NULL);
+	if (of_chosen) {
+		prop = (u64 *)get_property(of_chosen,
+				"linux,initrd-start", NULL);
 		if (prop != NULL) {
-			initrd_end = (unsigned long)__va(*prop);
-			initrd_below_start_ok = 1;
-		} else
-			initrd_start = 0;
+			initrd_start = (unsigned long)__va(*prop);
+			prop = (u64 *)get_property(of_chosen,
+					"linux,initrd-end", NULL);
+			if (prop != NULL) {
+				initrd_end = (unsigned long)__va(*prop);
+				initrd_below_start_ok = 1;
+			} else
+				initrd_start = 0;
+		}
 	}
 
 	/* If we were passed an initrd, set the ROOT_DEV properly if the values
@@ -627,7 +631,7 @@ void __init setup_system(void)
 	 * Initialize xmon
 	 */
 #ifdef CONFIG_XMON_DEFAULT
-	xmon_init();
+	xmon_init(1);
 #endif
 	/*
 	 * Register early console
@@ -1343,11 +1347,13 @@ static int __init early_xmon(char *p)
 	/* ensure xmon is enabled */
 	if (p) {
 		if (strncmp(p, "on", 2) == 0)
-			xmon_init();
+			xmon_init(1);
+		if (strncmp(p, "off", 3) == 0)
+			xmon_init(0);
 		if (strncmp(p, "early", 5) != 0)
 			return 0;
 	}
-	xmon_init();
+	xmon_init(1);
 	debugger(NULL);
 
 	return 0;
diff --git a/arch/ppc64/kernel/sysfs.c b/arch/ppc64/kernel/sysfs.c
index 02b8ac4..f311ee7 100644
--- a/arch/ppc64/kernel/sysfs.c
+++ b/arch/ppc64/kernel/sysfs.c
@@ -13,6 +13,7 @@
 #include <asm/current.h>
 #include <asm/processor.h>
 #include <asm/cputable.h>
+#include <asm/firmware.h>
 #include <asm/hvcall.h>
 #include <asm/prom.h>
 #include <asm/systemcfg.h>
@@ -100,6 +101,8 @@ static int __init setup_smt_snooze_delay(char *str)
 }
 __setup("smt-snooze-delay=", setup_smt_snooze_delay);
 
+#endif /* CONFIG_PPC_MULTIPLATFORM */
+
 /*
  * Enabling PMCs will slow partition context switch times so we only do
  * it the first time we write to the PMCs.
@@ -109,65 +112,15 @@ static DEFINE_PER_CPU(char, pmcs_enabled);
 
 void ppc64_enable_pmcs(void)
 {
-	unsigned long hid0;
-#ifdef CONFIG_PPC_PSERIES
-	unsigned long set, reset;
-#endif /* CONFIG_PPC_PSERIES */
-
 	/* Only need to enable them once */
 	if (__get_cpu_var(pmcs_enabled))
 		return;
 
 	__get_cpu_var(pmcs_enabled) = 1;
 
-	switch (systemcfg->platform) {
-	case PLATFORM_PSERIES:
-	case PLATFORM_POWERMAC:
-		hid0 = mfspr(HID0);
-		hid0 |= 1UL << (63 - 20);
-
-		/* POWER4 requires the following sequence */
-		asm volatile(
-			     "sync\n"
-			     "mtspr	%1, %0\n"
-			     "mfspr	%0, %1\n"
-			     "mfspr	%0, %1\n"
-			     "mfspr	%0, %1\n"
-			     "mfspr	%0, %1\n"
-			     "mfspr	%0, %1\n"
-			     "mfspr	%0, %1\n"
-			     "isync" : "=&r" (hid0) : "i" (HID0), "0" (hid0):
-			     "memory");
-		break;
-
-#ifdef CONFIG_PPC_PSERIES
-	case PLATFORM_PSERIES_LPAR:
-		set = 1UL << 63;
-		reset = 0;
-		plpar_hcall_norets(H_PERFMON, set, reset);
-		break;
-#endif /* CONFIG_PPC_PSERIES */
-
-	default:
-		break;
-	}
-
-#ifdef CONFIG_PPC_PSERIES
-	/* instruct hypervisor to maintain PMCs */
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR)
-		get_paca()->lppaca.pmcregs_in_use = 1;
-#endif /* CONFIG_PPC_PSERIES */
+	if (ppc_md.enable_pmcs)
+		ppc_md.enable_pmcs();
 }
-
-#else
-
-/* PMC stuff */
-void ppc64_enable_pmcs(void)
-{
-	/* XXX Implement for iseries */
-}
-#endif /* CONFIG_PPC_MULTIPLATFORM */
-
 EXPORT_SYMBOL(ppc64_enable_pmcs);
 
 /* XXX convert to rusty's on_one_cpu */
diff --git a/arch/ppc64/kernel/time.c b/arch/ppc64/kernel/time.c
index 909462e..1696e1b 100644
--- a/arch/ppc64/kernel/time.c
+++ b/arch/ppc64/kernel/time.c
@@ -67,6 +67,7 @@
 #include <asm/prom.h>
 #include <asm/sections.h>
 #include <asm/systemcfg.h>
+#include <asm/firmware.h>
 
 u64 jiffies_64 __cacheline_aligned_in_smp = INITIAL_JIFFIES;
 
@@ -370,13 +371,11 @@ int timer_interrupt(struct pt_regs * regs)
 		process_hvlpevents(regs);
 #endif
 
-/* collect purr register values often, for accurate calculations */
-#if defined(CONFIG_PPC_PSERIES)
-	if (cur_cpu_spec->firmware_features & FW_FEATURE_SPLPAR) {
+	/* collect purr register values often, for accurate calculations */
+	if (firmware_has_feature(FW_FEATURE_SPLPAR)) {
 		struct cpu_usage *cu = &__get_cpu_var(cpu_usage_array);
 		cu->current_tb = mfspr(SPRN_PURR);
 	}
-#endif
 
 	irq_exit();
 
diff --git a/arch/ppc64/kernel/vio.c b/arch/ppc64/kernel/vio.c
index 0c0ba71..3b790ba 100644
--- a/arch/ppc64/kernel/vio.c
+++ b/arch/ppc64/kernel/vio.c
@@ -1,10 +1,11 @@
 /*
  * IBM PowerPC Virtual I/O Infrastructure Support.
  *
- *    Copyright (c) 2003 IBM Corp.
+ *    Copyright (c) 2003-2005 IBM Corp.
  *     Dave Engebretsen engebret@us.ibm.com
  *     Santiago Leon santil@us.ibm.com
  *     Hollis Blanchard <hollisb@us.ibm.com>
+ *     Stephen Rothwell
  *
  *      This program is free software; you can redistribute it and/or
  *      modify it under the terms of the GNU General Public License
@@ -14,57 +15,27 @@
 
 #include <linux/init.h>
 #include <linux/console.h>
-#include <linux/version.h>
 #include <linux/module.h>
-#include <linux/kobject.h>
 #include <linux/mm.h>
 #include <linux/dma-mapping.h>
-#include <asm/rtas.h>
 #include <asm/iommu.h>
 #include <asm/dma.h>
-#include <asm/ppcdebug.h>
 #include <asm/vio.h>
-#include <asm/hvcall.h>
-#include <asm/iSeries/vio.h>
-#include <asm/iSeries/HvTypes.h>
-#include <asm/iSeries/HvCallXm.h>
-#include <asm/iSeries/HvLpConfig.h>
-
-#define DBGENTER() pr_debug("%s entered\n", __FUNCTION__)
-
-extern struct subsystem devices_subsys; /* needed for vio_find_name() */
 
 static const struct vio_device_id *vio_match_device(
 		const struct vio_device_id *, const struct vio_dev *);
 
-#ifdef CONFIG_PPC_PSERIES
-static struct iommu_table *vio_build_iommu_table(struct vio_dev *);
-static int vio_num_address_cells;
-#endif
-#ifdef CONFIG_PPC_ISERIES
-static struct iommu_table veth_iommu_table;
-static struct iommu_table vio_iommu_table;
-#endif
-static struct vio_dev vio_bus_device  = { /* fake "parent" device */
+struct vio_dev vio_bus_device  = { /* fake "parent" device */
 	.name = vio_bus_device.dev.bus_id,
 	.type = "",
-#ifdef CONFIG_PPC_ISERIES
-	.iommu_table = &vio_iommu_table,
-#endif
 	.dev.bus_id = "vio",
 	.dev.bus = &vio_bus_type,
 };
 
-#ifdef CONFIG_PPC_ISERIES
-static struct vio_dev *__init vio_register_device_iseries(char *type,
-		uint32_t unit_num);
-
-struct device *iSeries_vio_dev = &vio_bus_device.dev;
-EXPORT_SYMBOL(iSeries_vio_dev);
-
-#define device_is_compatible(a, b)	1
-
-#endif
+static int (*is_match)(const struct vio_device_id *id,
+		const struct vio_dev *dev);
+static void (*unregister_device_callback)(struct vio_dev *dev);
+static void (*release_device_callback)(struct device *dev);
 
 /* convert from struct device to struct vio_dev and pass to driver.
  * dev->driver has already been set by generic code because vio_bus_match
@@ -76,8 +47,6 @@ static int vio_bus_probe(struct device *dev)
 	const struct vio_device_id *id;
 	int error = -ENODEV;
 
-	DBGENTER();
-
 	if (!viodrv->probe)
 		return error;
 
@@ -95,8 +64,6 @@ static int vio_bus_remove(struct device *dev)
 	struct vio_dev *viodev = to_vio_dev(dev);
 	struct vio_driver *viodrv = to_vio_driver(dev->driver);
 
-	DBGENTER();
-
 	if (viodrv->remove) {
 		return viodrv->remove(viodev);
 	}
@@ -146,178 +113,65 @@ EXPORT_SYMBOL(vio_unregister_driver);
 static const struct vio_device_id * vio_match_device(const struct vio_device_id *ids,
 	const struct vio_dev *dev)
 {
-	DBGENTER();
-
 	while (ids->type) {
-		if ((strncmp(dev->type, ids->type, strlen(ids->type)) == 0) &&
-			device_is_compatible(dev->dev.platform_data, ids->compat))
+		if (is_match(ids, dev))
 			return ids;
 		ids++;
 	}
 	return NULL;
 }
 
-#ifdef CONFIG_PPC_ISERIES
-void __init iommu_vio_init(void)
-{
-	struct iommu_table *t;
-	struct iommu_table_cb cb;
-	unsigned long cbp;
-	unsigned long itc_entries;
-
-	cb.itc_busno = 255;    /* Bus 255 is the virtual bus */
-	cb.itc_virtbus = 0xff; /* Ask for virtual bus */
-
-	cbp = virt_to_abs(&cb);
-	HvCallXm_getTceTableParms(cbp);
-
-	itc_entries = cb.itc_size * PAGE_SIZE / sizeof(union tce_entry);
-	veth_iommu_table.it_size        = itc_entries / 2;
-	veth_iommu_table.it_busno       = cb.itc_busno;
-	veth_iommu_table.it_offset      = cb.itc_offset;
-	veth_iommu_table.it_index       = cb.itc_index;
-	veth_iommu_table.it_type        = TCE_VB;
-	veth_iommu_table.it_blocksize	= 1;
-
-	t = iommu_init_table(&veth_iommu_table);
-
-	if (!t)
-		printk("Virtual Bus VETH TCE table failed.\n");
-
-	vio_iommu_table.it_size         = itc_entries - veth_iommu_table.it_size;
-	vio_iommu_table.it_busno        = cb.itc_busno;
-	vio_iommu_table.it_offset       = cb.itc_offset +
-					  veth_iommu_table.it_size;
-	vio_iommu_table.it_index        = cb.itc_index;
-	vio_iommu_table.it_type         = TCE_VB;
-	vio_iommu_table.it_blocksize	= 1;
-
-	t = iommu_init_table(&vio_iommu_table);
-
-	if (!t)
-		printk("Virtual Bus VIO TCE table failed.\n");
-}
-#endif
-
-#ifdef CONFIG_PPC_PSERIES
-static void probe_bus_pseries(void)
-{
-	struct device_node *node_vroot, *of_node;
-
-	node_vroot = find_devices("vdevice");
-	if ((node_vroot == NULL) || (node_vroot->child == NULL))
-		/* this machine doesn't do virtual IO, and that's ok */
-		return;
-
-	vio_num_address_cells = prom_n_addr_cells(node_vroot->child);
-
-	/*
-	 * Create struct vio_devices for each virtual device in the device tree.
-	 * Drivers will associate with them later.
-	 */
-	for (of_node = node_vroot->child; of_node != NULL;
-			of_node = of_node->sibling) {
-		printk(KERN_DEBUG "%s: processing %p\n", __FUNCTION__, of_node);
-		vio_register_device_node(of_node);
-	}
-}
-#endif
-
-#ifdef CONFIG_PPC_ISERIES
-static void probe_bus_iseries(void)
-{
-	HvLpIndexMap vlan_map = HvLpConfig_getVirtualLanIndexMap();
-	struct vio_dev *viodev;
-	int i;
-
-	/* there is only one of each of these */
-	vio_register_device_iseries("viocons", 0);
-	vio_register_device_iseries("vscsi", 0);
-
-	vlan_map = HvLpConfig_getVirtualLanIndexMap();
-	for (i = 0; i < HVMAXARCHITECTEDVIRTUALLANS; i++) {
-		if ((vlan_map & (0x8000 >> i)) == 0)
-			continue;
-		viodev = vio_register_device_iseries("vlan", i);
-		/* veth is special and has it own iommu_table */
-		viodev->iommu_table = &veth_iommu_table;
-	}
-	for (i = 0; i < HVMAXARCHITECTEDVIRTUALDISKS; i++)
-		vio_register_device_iseries("viodasd", i);
-	for (i = 0; i < HVMAXARCHITECTEDVIRTUALCDROMS; i++)
-		vio_register_device_iseries("viocd", i);
-	for (i = 0; i < HVMAXARCHITECTEDVIRTUALTAPES; i++)
-		vio_register_device_iseries("viotape", i);
-}
-#endif
-
 /**
  * vio_bus_init: - Initialize the virtual IO bus
  */
-static int __init vio_bus_init(void)
+int __init vio_bus_init(int (*match_func)(const struct vio_device_id *id,
+			const struct vio_dev *dev),
+		void (*unregister_dev)(struct vio_dev *),
+		void (*release_dev)(struct device *))
 {
 	int err;
 
+	is_match = match_func;
+	unregister_device_callback = unregister_dev;
+	release_device_callback = release_dev;
+
 	err = bus_register(&vio_bus_type);
 	if (err) {
 		printk(KERN_ERR "failed to register VIO bus\n");
 		return err;
 	}
 
-	/* the fake parent of all vio devices, just to give us a nice directory */
+	/* the fake parent of all vio devices, just to give us
+	 * a nice directory
+	 */
 	err = device_register(&vio_bus_device.dev);
 	if (err) {
-		printk(KERN_WARNING "%s: device_register returned %i\n", __FUNCTION__,
-			err);
+		printk(KERN_WARNING "%s: device_register returned %i\n",
+				__FUNCTION__, err);
 		return err;
 	}
 
-#ifdef CONFIG_PPC_PSERIES
-	probe_bus_pseries();
-#endif
-#ifdef CONFIG_PPC_ISERIES
-	probe_bus_iseries();
-#endif
-
 	return 0;
 }
 
-__initcall(vio_bus_init);
-
 /* vio_dev refcount hit 0 */
 static void __devinit vio_dev_release(struct device *dev)
 {
-	DBGENTER();
-
-#ifdef CONFIG_PPC_PSERIES
-	/* XXX free TCE table */
-	of_node_put(dev->platform_data);
-#endif
+	if (release_device_callback)
+		release_device_callback(dev);
 	kfree(to_vio_dev(dev));
 }
 
-#ifdef CONFIG_PPC_PSERIES
-static ssize_t viodev_show_devspec(struct device *dev, struct device_attribute *attr, char *buf)
-{
-	struct device_node *of_node = dev->platform_data;
-
-	return sprintf(buf, "%s\n", of_node->full_name);
-}
-DEVICE_ATTR(devspec, S_IRUSR | S_IRGRP | S_IROTH, viodev_show_devspec, NULL);
-#endif
-
 static ssize_t viodev_show_name(struct device *dev, struct device_attribute *attr, char *buf)
 {
 	return sprintf(buf, "%s\n", to_vio_dev(dev)->name);
 }
 DEVICE_ATTR(name, S_IRUSR | S_IRGRP | S_IROTH, viodev_show_name, NULL);
 
-static struct vio_dev * __devinit vio_register_device_common(
+struct vio_dev * __devinit vio_register_device_common(
 		struct vio_dev *viodev, char *name, char *type,
 		uint32_t unit_address, struct iommu_table *iommu_table)
 {
-	DBGENTER();
-
 	viodev->name = name;
 	viodev->type = type;
 	viodev->unit_address = unit_address;
@@ -338,222 +192,15 @@ static struct vio_dev * __devinit vio_register_device_common(
 	return viodev;
 }
 
-#ifdef CONFIG_PPC_PSERIES
-/**
- * vio_register_device_node: - Register a new vio device.
- * @of_node:	The OF node for this device.
- *
- * Creates and initializes a vio_dev structure from the data in
- * of_node (dev.platform_data) and adds it to the list of virtual devices.
- * Returns a pointer to the created vio_dev or NULL if node has
- * NULL device_type or compatible fields.
- */
-struct vio_dev * __devinit vio_register_device_node(struct device_node *of_node)
-{
-	struct vio_dev *viodev;
-	unsigned int *unit_address;
-	unsigned int *irq_p;
-
-	DBGENTER();
-
-	/* we need the 'device_type' property, in order to match with drivers */
-	if ((NULL == of_node->type)) {
-		printk(KERN_WARNING
-			"%s: node %s missing 'device_type'\n", __FUNCTION__,
-			of_node->name ? of_node->name : "<unknown>");
-		return NULL;
-	}
-
-	unit_address = (unsigned int *)get_property(of_node, "reg", NULL);
-	if (!unit_address) {
-		printk(KERN_WARNING "%s: node %s missing 'reg'\n", __FUNCTION__,
-			of_node->name ? of_node->name : "<unknown>");
-		return NULL;
-	}
-
-	/* allocate a vio_dev for this node */
-	viodev = kmalloc(sizeof(struct vio_dev), GFP_KERNEL);
-	if (!viodev) {
-		return NULL;
-	}
-	memset(viodev, 0, sizeof(struct vio_dev));
-
-	viodev->dev.platform_data = of_node_get(of_node);
-
-	viodev->irq = NO_IRQ;
-	irq_p = (unsigned int *)get_property(of_node, "interrupts", NULL);
-	if (irq_p) {
-		int virq = virt_irq_create_mapping(*irq_p);
-		if (virq == NO_IRQ) {
-			printk(KERN_ERR "Unable to allocate interrupt "
-			       "number for %s\n", of_node->full_name);
-		} else
-			viodev->irq = irq_offset_up(virq);
-	}
-
-	snprintf(viodev->dev.bus_id, BUS_ID_SIZE, "%x", *unit_address);
-
-	/* register with generic device framework */
-	if (vio_register_device_common(viodev, of_node->name, of_node->type,
-				*unit_address, vio_build_iommu_table(viodev))
-			== NULL) {
-		/* XXX free TCE table */
-		kfree(viodev);
-		return NULL;
-	}
-	device_create_file(&viodev->dev, &dev_attr_devspec);
-
-	return viodev;
-}
-EXPORT_SYMBOL(vio_register_device_node);
-#endif
-
-#ifdef CONFIG_PPC_ISERIES
-/**
- * vio_register_device: - Register a new vio device.
- * @voidev:	The device to register.
- */
-static struct vio_dev *__init vio_register_device_iseries(char *type,
-		uint32_t unit_num)
-{
-	struct vio_dev *viodev;
-
-	DBGENTER();
-
-	/* allocate a vio_dev for this node */
-	viodev = kmalloc(sizeof(struct vio_dev), GFP_KERNEL);
-	if (!viodev)
-		return NULL;
-	memset(viodev, 0, sizeof(struct vio_dev));
-
-	snprintf(viodev->dev.bus_id, BUS_ID_SIZE, "%s%d", type, unit_num);
-
-	return vio_register_device_common(viodev, viodev->dev.bus_id, type,
-			unit_num, &vio_iommu_table);
-}
-#endif
-
 void __devinit vio_unregister_device(struct vio_dev *viodev)
 {
-	DBGENTER();
-#ifdef CONFIG_PPC_PSERIES
-	device_remove_file(&viodev->dev, &dev_attr_devspec);
-#endif
+	if (unregister_device_callback)
+		unregister_device_callback(viodev);
 	device_remove_file(&viodev->dev, &dev_attr_name);
 	device_unregister(&viodev->dev);
 }
 EXPORT_SYMBOL(vio_unregister_device);
 
-#ifdef CONFIG_PPC_PSERIES
-/**
- * vio_get_attribute: - get attribute for virtual device
- * @vdev:	The vio device to get property.
- * @which:	The property/attribute to be extracted.
- * @length:	Pointer to length of returned data size (unused if NULL).
- *
- * Calls prom.c's get_property() to return the value of the
- * attribute specified by the preprocessor constant @which
-*/
-const void * vio_get_attribute(struct vio_dev *vdev, void* which, int* length)
-{
-	return get_property(vdev->dev.platform_data, (char*)which, length);
-}
-EXPORT_SYMBOL(vio_get_attribute);
-
-/* vio_find_name() - internal because only vio.c knows how we formatted the
- * kobject name
- * XXX once vio_bus_type.devices is actually used as a kset in
- * drivers/base/bus.c, this function should be removed in favor of
- * "device_find(kobj_name, &vio_bus_type)"
- */
-static struct vio_dev *vio_find_name(const char *kobj_name)
-{
-	struct kobject *found;
-
-	found = kset_find_obj(&devices_subsys.kset, kobj_name);
-	if (!found)
-		return NULL;
-
-	return to_vio_dev(container_of(found, struct device, kobj));
-}
-
-/**
- * vio_find_node - find an already-registered vio_dev
- * @vnode: device_node of the virtual device we're looking for
- */
-struct vio_dev *vio_find_node(struct device_node *vnode)
-{
-	uint32_t *unit_address;
-	char kobj_name[BUS_ID_SIZE];
-
-	/* construct the kobject name from the device node */
-	unit_address = (uint32_t *)get_property(vnode, "reg", NULL);
-	if (!unit_address)
-		return NULL;
-	snprintf(kobj_name, BUS_ID_SIZE, "%x", *unit_address);
-
-	return vio_find_name(kobj_name);
-}
-EXPORT_SYMBOL(vio_find_node);
-
-/**
- * vio_build_iommu_table: - gets the dma information from OF and builds the TCE tree.
- * @dev: the virtual device.
- *
- * Returns a pointer to the built tce tree, or NULL if it can't
- * find property.
-*/
-static struct iommu_table * vio_build_iommu_table(struct vio_dev *dev)
-{
-	unsigned int *dma_window;
-	struct iommu_table *newTceTable;
-	unsigned long offset;
-	int dma_window_property_size;
-
-	dma_window = (unsigned int *) get_property(dev->dev.platform_data, "ibm,my-dma-window", &dma_window_property_size);
-	if(!dma_window) {
-		return NULL;
-	}
-
-	newTceTable = (struct iommu_table *) kmalloc(sizeof(struct iommu_table), GFP_KERNEL);
-
-	/*  There should be some code to extract the phys-encoded offset
-		using prom_n_addr_cells(). However, according to a comment
-		on earlier versions, it's always zero, so we don't bother */
-	offset = dma_window[1] >>  PAGE_SHIFT;
-
-	/* TCE table size - measured in tce entries */
-	newTceTable->it_size		= dma_window[4] >> PAGE_SHIFT;
-	/* offset for VIO should always be 0 */
-	newTceTable->it_offset		= offset;
-	newTceTable->it_busno		= 0;
-	newTceTable->it_index		= (unsigned long)dma_window[0];
-	newTceTable->it_type		= TCE_VB;
-
-	return iommu_init_table(newTceTable);
-}
-
-int vio_enable_interrupts(struct vio_dev *dev)
-{
-	int rc = h_vio_signal(dev->unit_address, VIO_IRQ_ENABLE);
-	if (rc != H_Success) {
-		printk(KERN_ERR "vio: Error 0x%x enabling interrupts\n", rc);
-	}
-	return rc;
-}
-EXPORT_SYMBOL(vio_enable_interrupts);
-
-int vio_disable_interrupts(struct vio_dev *dev)
-{
-	int rc = h_vio_signal(dev->unit_address, VIO_IRQ_DISABLE);
-	if (rc != H_Success) {
-		printk(KERN_ERR "vio: Error 0x%x disabling interrupts\n", rc);
-	}
-	return rc;
-}
-EXPORT_SYMBOL(vio_disable_interrupts);
-#endif
-
 static dma_addr_t vio_map_single(struct device *dev, void *vaddr,
 			  size_t size, enum dma_data_direction direction)
 {
@@ -617,8 +264,6 @@ static int vio_bus_match(struct device *dev, struct device_driver *drv)
 	const struct vio_device_id *ids = vio_drv->id_table;
 	const struct vio_device_id *found_id;
 
-	DBGENTER();
-
 	if (!ids)
 		return 0;
 
diff --git a/arch/ppc64/mm/hash_native.c b/arch/ppc64/mm/hash_native.c
index a6abd3a..7626bb5 100644
--- a/arch/ppc64/mm/hash_native.c
+++ b/arch/ppc64/mm/hash_native.c
@@ -51,7 +51,6 @@ long native_hpte_insert(unsigned long hpte_group, unsigned long va,
 			unsigned long prpn, unsigned long vflags,
 			unsigned long rflags)
 {
-	unsigned long arpn = physRpn_to_absRpn(prpn);
 	hpte_t *hptep = htab_address + hpte_group;
 	unsigned long hpte_v, hpte_r;
 	int i;
@@ -74,7 +73,7 @@ long native_hpte_insert(unsigned long hpte_group, unsigned long va,
 	hpte_v = (va >> 23) << HPTE_V_AVPN_SHIFT | vflags | HPTE_V_VALID;
 	if (vflags & HPTE_V_LARGE)
 		va &= ~(1UL << HPTE_V_AVPN_SHIFT);
-	hpte_r = (arpn << HPTE_R_RPN_SHIFT) | rflags;
+	hpte_r = (prpn << HPTE_R_RPN_SHIFT) | rflags;
 
 	hptep->r = hpte_r;
 	/* Guarantee the second dword is visible before the valid bit */
diff --git a/arch/ppc64/mm/hash_utils.c b/arch/ppc64/mm/hash_utils.c
index 623b5d1..09475c8 100644
--- a/arch/ppc64/mm/hash_utils.c
+++ b/arch/ppc64/mm/hash_utils.c
@@ -210,7 +210,7 @@ void __init htab_initialize(void)
 
 	/* create bolted the linear mapping in the hash table */
 	for (i=0; i < lmb.memory.cnt; i++) {
-		base = lmb.memory.region[i].physbase + KERNELBASE;
+		base = lmb.memory.region[i].base + KERNELBASE;
 		size = lmb.memory.region[i].size;
 
 		DBG("creating mapping for region: %lx : %lx\n", base, size);
@@ -302,7 +302,7 @@ int hash_page(unsigned long ea, unsigned long access, unsigned long trap)
 	int local = 0;
 	cpumask_t tmp;
 
-	if ((ea & ~REGION_MASK) > EADDR_MASK)
+	if ((ea & ~REGION_MASK) >= PGTABLE_RANGE)
 		return 1;
 
  	switch (REGION_ID(ea)) {
diff --git a/arch/ppc64/mm/hugetlbpage.c b/arch/ppc64/mm/hugetlbpage.c
index f952460..e7833c8 100644
--- a/arch/ppc64/mm/hugetlbpage.c
+++ b/arch/ppc64/mm/hugetlbpage.c
@@ -27,124 +27,94 @@
 
 #include <linux/sysctl.h>
 
-#define	HUGEPGDIR_SHIFT		(HPAGE_SHIFT + PAGE_SHIFT - 3)
-#define HUGEPGDIR_SIZE		(1UL << HUGEPGDIR_SHIFT)
-#define HUGEPGDIR_MASK		(~(HUGEPGDIR_SIZE-1))
+#define NUM_LOW_AREAS	(0x100000000UL >> SID_SHIFT)
+#define NUM_HIGH_AREAS	(PGTABLE_RANGE >> HTLB_AREA_SHIFT)
 
-#define HUGEPTE_INDEX_SIZE	9
-#define HUGEPGD_INDEX_SIZE	10
-
-#define PTRS_PER_HUGEPTE	(1 << HUGEPTE_INDEX_SIZE)
-#define PTRS_PER_HUGEPGD	(1 << HUGEPGD_INDEX_SIZE)
-
-static inline int hugepgd_index(unsigned long addr)
-{
-	return (addr & ~REGION_MASK) >> HUGEPGDIR_SHIFT;
-}
-
-static pud_t *hugepgd_offset(struct mm_struct *mm, unsigned long addr)
+/* Modelled after find_linux_pte() */
+pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
 {
-	int index;
+	pgd_t *pg;
+	pud_t *pu;
+	pmd_t *pm;
+	pte_t *pt;
 
-	if (! mm->context.huge_pgdir)
-		return NULL;
+	BUG_ON(! in_hugepage_area(mm->context, addr));
 
+	addr &= HPAGE_MASK;
+
+	pg = pgd_offset(mm, addr);
+	if (!pgd_none(*pg)) {
+		pu = pud_offset(pg, addr);
+		if (!pud_none(*pu)) {
+			pm = pmd_offset(pu, addr);
+			pt = (pte_t *)pm;
+			BUG_ON(!pmd_none(*pm)
+			       && !(pte_present(*pt) && pte_huge(*pt)));
+			return pt;
+		}
+	}
 
-	index = hugepgd_index(addr);
-	BUG_ON(index >= PTRS_PER_HUGEPGD);
-	return (pud_t *)(mm->context.huge_pgdir + index);
+	return NULL;
 }
 
-static inline pte_t *hugepte_offset(pud_t *dir, unsigned long addr)
+pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr)
 {
-	int index;
-
-	if (pud_none(*dir))
-		return NULL;
+	pgd_t *pg;
+	pud_t *pu;
+	pmd_t *pm;
+	pte_t *pt;
 
-	index = (addr >> HPAGE_SHIFT) % PTRS_PER_HUGEPTE;
-	return (pte_t *)pud_page(*dir) + index;
-}
-
-static pud_t *hugepgd_alloc(struct mm_struct *mm, unsigned long addr)
-{
 	BUG_ON(! in_hugepage_area(mm->context, addr));
 
-	if (! mm->context.huge_pgdir) {
-		pgd_t *new;
-		spin_unlock(&mm->page_table_lock);
-		/* Don't use pgd_alloc(), because we want __GFP_REPEAT */
-		new = kmem_cache_alloc(zero_cache, GFP_KERNEL | __GFP_REPEAT);
-		BUG_ON(memcmp(new, empty_zero_page, PAGE_SIZE));
-		spin_lock(&mm->page_table_lock);
+	addr &= HPAGE_MASK;
 
-		/*
-		 * Because we dropped the lock, we should re-check the
-		 * entry, as somebody else could have populated it..
-		 */
-		if (mm->context.huge_pgdir)
-			pgd_free(new);
-		else
-			mm->context.huge_pgdir = new;
-	}
-	return hugepgd_offset(mm, addr);
-}
+	pg = pgd_offset(mm, addr);
+	pu = pud_alloc(mm, pg, addr);
 
-static pte_t *hugepte_alloc(struct mm_struct *mm, pud_t *dir, unsigned long addr)
-{
-	if (! pud_present(*dir)) {
-		pte_t *new;
-
-		spin_unlock(&mm->page_table_lock);
-		new = kmem_cache_alloc(zero_cache, GFP_KERNEL | __GFP_REPEAT);
-		BUG_ON(memcmp(new, empty_zero_page, PAGE_SIZE));
-		spin_lock(&mm->page_table_lock);
-		/*
-		 * Because we dropped the lock, we should re-check the
-		 * entry, as somebody else could have populated it..
-		 */
-		if (pud_present(*dir)) {
-			if (new)
-				kmem_cache_free(zero_cache, new);
-		} else {
-			struct page *ptepage;
-
-			if (! new)
-				return NULL;
-			ptepage = virt_to_page(new);
-			ptepage->mapping = (void *) mm;
-			ptepage->index = addr & HUGEPGDIR_MASK;
-			pud_populate(mm, dir, new);
+	if (pu) {
+		pm = pmd_alloc(mm, pu, addr);
+		if (pm) {
+			pt = (pte_t *)pm;
+			BUG_ON(!pmd_none(*pm)
+			       && !(pte_present(*pt) && pte_huge(*pt)));
+			return pt;
 		}
 	}
 
-	return hugepte_offset(dir, addr);
+	return NULL;
 }
 
-pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr)
-{
-	pud_t *pud;
+#define HUGEPTE_BATCH_SIZE	(HPAGE_SIZE / PMD_SIZE)
 
-	BUG_ON(! in_hugepage_area(mm->context, addr));
+void set_huge_pte_at(struct mm_struct *mm, unsigned long addr,
+		     pte_t *ptep, pte_t pte)
+{
+	int i;
 
-	pud = hugepgd_offset(mm, addr);
-	if (! pud)
-		return NULL;
+	if (pte_present(*ptep)) {
+		pte_clear(mm, addr, ptep);
+		flush_tlb_pending();
+	}
 
-	return hugepte_offset(pud, addr);
+	for (i = 0; i < HUGEPTE_BATCH_SIZE; i++) {
+		*ptep = __pte(pte_val(pte) & ~_PAGE_HPTEFLAGS);
+		ptep++;
+	}
 }
 
-pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr)
+pte_t huge_ptep_get_and_clear(struct mm_struct *mm, unsigned long addr,
+			      pte_t *ptep)
 {
-	pud_t *pud;
+	unsigned long old = pte_update(ptep, ~0UL);
+	int i;
 
-	BUG_ON(! in_hugepage_area(mm->context, addr));
+	if (old & _PAGE_HASHPTE)
+		hpte_update(mm, addr, old, 0);
 
-	pud = hugepgd_alloc(mm, addr);
-	if (! pud)
-		return NULL;
+	for (i = 1; i < HUGEPTE_BATCH_SIZE; i++)
+		ptep[i] = __pte(0);
 
-	return hugepte_alloc(mm, pud, addr);
+	return __pte(old);
 }
 
 /*
@@ -162,15 +132,17 @@ int is_aligned_hugepage_range(unsigned long addr, unsigned long len)
 	return 0;
 }
 
-static void flush_segments(void *parm)
+static void flush_low_segments(void *parm)
 {
-	u16 segs = (unsigned long) parm;
+	u16 areas = (unsigned long) parm;
 	unsigned long i;
 
 	asm volatile("isync" : : : "memory");
 
-	for (i = 0; i < 16; i++) {
-		if (! (segs & (1U << i)))
+	BUILD_BUG_ON((sizeof(areas)*8) != NUM_LOW_AREAS);
+
+	for (i = 0; i < NUM_LOW_AREAS; i++) {
+		if (! (areas & (1U << i)))
 			continue;
 		asm volatile("slbie %0" : : "r" (i << SID_SHIFT));
 	}
@@ -178,13 +150,33 @@ static void flush_segments(void *parm)
 	asm volatile("isync" : : : "memory");
 }
 
-static int prepare_low_seg_for_htlb(struct mm_struct *mm, unsigned long seg)
+static void flush_high_segments(void *parm)
 {
-	unsigned long start = seg << SID_SHIFT;
-	unsigned long end = (seg+1) << SID_SHIFT;
+	u16 areas = (unsigned long) parm;
+	unsigned long i, j;
+
+	asm volatile("isync" : : : "memory");
+
+	BUILD_BUG_ON((sizeof(areas)*8) != NUM_HIGH_AREAS);
+
+	for (i = 0; i < NUM_HIGH_AREAS; i++) {
+		if (! (areas & (1U << i)))
+			continue;
+		for (j = 0; j < (1UL << (HTLB_AREA_SHIFT-SID_SHIFT)); j++)
+			asm volatile("slbie %0"
+				     :: "r" ((i << HTLB_AREA_SHIFT) + (j << SID_SHIFT)));
+	}
+
+	asm volatile("isync" : : : "memory");
+}
+
+static int prepare_low_area_for_htlb(struct mm_struct *mm, unsigned long area)
+{
+	unsigned long start = area << SID_SHIFT;
+	unsigned long end = (area+1) << SID_SHIFT;
 	struct vm_area_struct *vma;
 
-	BUG_ON(seg >= 16);
+	BUG_ON(area >= NUM_LOW_AREAS);
 
 	/* Check no VMAs are in the region */
 	vma = find_vma(mm, start);
@@ -194,20 +186,39 @@ static int prepare_low_seg_for_htlb(struct mm_struct *mm, unsigned long seg)
 	return 0;
 }
 
-static int open_low_hpage_segs(struct mm_struct *mm, u16 newsegs)
+static int prepare_high_area_for_htlb(struct mm_struct *mm, unsigned long area)
+{
+	unsigned long start = area << HTLB_AREA_SHIFT;
+	unsigned long end = (area+1) << HTLB_AREA_SHIFT;
+	struct vm_area_struct *vma;
+
+	BUG_ON(area >= NUM_HIGH_AREAS);
+
+	/* Check no VMAs are in the region */
+	vma = find_vma(mm, start);
+	if (vma && (vma->vm_start < end))
+		return -EBUSY;
+
+	return 0;
+}
+
+static int open_low_hpage_areas(struct mm_struct *mm, u16 newareas)
 {
 	unsigned long i;
 
-	newsegs &= ~(mm->context.htlb_segs);
-	if (! newsegs)
+	BUILD_BUG_ON((sizeof(newareas)*8) != NUM_LOW_AREAS);
+	BUILD_BUG_ON((sizeof(mm->context.low_htlb_areas)*8) != NUM_LOW_AREAS);
+
+	newareas &= ~(mm->context.low_htlb_areas);
+	if (! newareas)
 		return 0; /* The segments we want are already open */
 
-	for (i = 0; i < 16; i++)
-		if ((1 << i) & newsegs)
-			if (prepare_low_seg_for_htlb(mm, i) != 0)
+	for (i = 0; i < NUM_LOW_AREAS; i++)
+		if ((1 << i) & newareas)
+			if (prepare_low_area_for_htlb(mm, i) != 0)
 				return -EBUSY;
 
-	mm->context.htlb_segs |= newsegs;
+	mm->context.low_htlb_areas |= newareas;
 
 	/* update the paca copy of the context struct */
 	get_paca()->context = mm->context;
@@ -215,29 +226,63 @@ static int open_low_hpage_segs(struct mm_struct *mm, u16 newsegs)
 	/* the context change must make it to memory before the flush,
 	 * so that further SLB misses do the right thing. */
 	mb();
-	on_each_cpu(flush_segments, (void *)(unsigned long)newsegs, 0, 1);
+	on_each_cpu(flush_low_segments, (void *)(unsigned long)newareas, 0, 1);
+
+	return 0;
+}
+
+static int open_high_hpage_areas(struct mm_struct *mm, u16 newareas)
+{
+	unsigned long i;
+
+	BUILD_BUG_ON((sizeof(newareas)*8) != NUM_HIGH_AREAS);
+	BUILD_BUG_ON((sizeof(mm->context.high_htlb_areas)*8)
+		     != NUM_HIGH_AREAS);
+
+	newareas &= ~(mm->context.high_htlb_areas);
+	if (! newareas)
+		return 0; /* The areas we want are already open */
+
+	for (i = 0; i < NUM_HIGH_AREAS; i++)
+		if ((1 << i) & newareas)
+			if (prepare_high_area_for_htlb(mm, i) != 0)
+				return -EBUSY;
+
+	mm->context.high_htlb_areas |= newareas;
+
+	/* update the paca copy of the context struct */
+	get_paca()->context = mm->context;
+
+	/* the context change must make it to memory before the flush,
+	 * so that further SLB misses do the right thing. */
+	mb();
+	on_each_cpu(flush_high_segments, (void *)(unsigned long)newareas, 0, 1);
 
 	return 0;
 }
 
 int prepare_hugepage_range(unsigned long addr, unsigned long len)
 {
-	if (within_hugepage_high_range(addr, len))
-		return 0;
-	else if ((addr < 0x100000000UL) && ((addr+len) < 0x100000000UL)) {
-		int err;
-		/* Yes, we need both tests, in case addr+len overflows
-		 * 64-bit arithmetic */
-		err = open_low_hpage_segs(current->mm,
+	int err;
+
+	if ( (addr+len) < addr )
+		return -EINVAL;
+
+	if ((addr + len) < 0x100000000UL)
+		err = open_low_hpage_areas(current->mm,
 					  LOW_ESID_MASK(addr, len));
-		if (err)
-			printk(KERN_DEBUG "prepare_hugepage_range(%lx, %lx)"
-			       " failed (segs: 0x%04hx)\n", addr, len,
-			       LOW_ESID_MASK(addr, len));
+	else
+		err = open_high_hpage_areas(current->mm,
+					    HTLB_AREA_MASK(addr, len));
+	if (err) {
+		printk(KERN_DEBUG "prepare_hugepage_range(%lx, %lx)"
+		       " failed (lowmask: 0x%04hx, highmask: 0x%04hx)\n",
+		       addr, len,
+		       LOW_ESID_MASK(addr, len), HTLB_AREA_MASK(addr, len));
 		return err;
 	}
 
-	return -EINVAL;
+	return 0;
 }
 
 struct page *
@@ -309,8 +354,8 @@ full_search:
 			vma = find_vma(mm, addr);
 			continue;
 		}
-		if (touches_hugepage_high_range(addr, len)) {
-			addr = TASK_HPAGE_END;
+		if (touches_hugepage_high_range(mm, addr, len)) {
+			addr = ALIGN(addr+1, 1UL<<HTLB_AREA_SHIFT);
 			vma = find_vma(mm, addr);
 			continue;
 		}
@@ -389,8 +434,9 @@ hugepage_recheck:
 		if (touches_hugepage_low_range(mm, addr, len)) {
 			addr = (addr & ((~0) << SID_SHIFT)) - len;
 			goto hugepage_recheck;
-		} else if (touches_hugepage_high_range(addr, len)) {
-			addr = TASK_HPAGE_BASE - len;
+		} else if (touches_hugepage_high_range(mm, addr, len)) {
+			addr = (addr & ((~0UL) << HTLB_AREA_SHIFT)) - len;
+			goto hugepage_recheck;
 		}
 
 		/*
@@ -481,23 +527,28 @@ static unsigned long htlb_get_low_area(unsigned long len, u16 segmask)
 	return -ENOMEM;
 }
 
-static unsigned long htlb_get_high_area(unsigned long len)
+static unsigned long htlb_get_high_area(unsigned long len, u16 areamask)
 {
-	unsigned long addr = TASK_HPAGE_BASE;
+	unsigned long addr = 0x100000000UL;
 	struct vm_area_struct *vma;
 
 	vma = find_vma(current->mm, addr);
-	for (vma = find_vma(current->mm, addr);
-	     addr + len <= TASK_HPAGE_END;
-	     vma = vma->vm_next) {
+	while (addr + len <= TASK_SIZE_USER64) {
 		BUG_ON(vma && (addr >= vma->vm_end)); /* invariant */
-		BUG_ON(! within_hugepage_high_range(addr, len));
+
+		if (! __within_hugepage_high_range(addr, len, areamask)) {
+			addr = ALIGN(addr+1, 1UL<<HTLB_AREA_SHIFT);
+			vma = find_vma(current->mm, addr);
+			continue;
+		}
 
 		if (!vma || (addr + len) <= vma->vm_start)
 			return addr;
 		addr = ALIGN(vma->vm_end, HPAGE_SIZE);
-		/* Because we're in a hugepage region, this alignment
-		 * should not skip us over any VMAs */
+		/* Depending on segmask this might not be a confirmed
+		 * hugepage region, so the ALIGN could have skipped
+		 * some VMAs */
+		vma = find_vma(current->mm, addr);
 	}
 
 	return -ENOMEM;
@@ -507,6 +558,9 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 					unsigned long len, unsigned long pgoff,
 					unsigned long flags)
 {
+	int lastshift;
+	u16 areamask, curareas;
+
 	if (len & ~HPAGE_MASK)
 		return -EINVAL;
 
@@ -514,67 +568,49 @@ unsigned long hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 		return -EINVAL;
 
 	if (test_thread_flag(TIF_32BIT)) {
-		int lastshift = 0;
-		u16 segmask, cursegs = current->mm->context.htlb_segs;
+		curareas = current->mm->context.low_htlb_areas;
 
 		/* First see if we can do the mapping in the existing
-		 * low hpage segments */
-		addr = htlb_get_low_area(len, cursegs);
+		 * low areas */
+		addr = htlb_get_low_area(len, curareas);
 		if (addr != -ENOMEM)
 			return addr;
 
-		for (segmask = LOW_ESID_MASK(0x100000000UL-len, len);
-		     ! lastshift; segmask >>=1) {
-			if (segmask & 1)
+		lastshift = 0;
+		for (areamask = LOW_ESID_MASK(0x100000000UL-len, len);
+		     ! lastshift; areamask >>=1) {
+			if (areamask & 1)
 				lastshift = 1;
 
-			addr = htlb_get_low_area(len, cursegs | segmask);
+			addr = htlb_get_low_area(len, curareas | areamask);
 			if ((addr != -ENOMEM)
-			    && open_low_hpage_segs(current->mm, segmask) == 0)
+			    && open_low_hpage_areas(current->mm, areamask) == 0)
 				return addr;
 		}
-		printk(KERN_DEBUG "hugetlb_get_unmapped_area() unable to open"
-		       " enough segments\n");
-		return -ENOMEM;
 	} else {
-		return htlb_get_high_area(len);
-	}
-}
-
-void hugetlb_mm_free_pgd(struct mm_struct *mm)
-{
-	int i;
-	pgd_t *pgdir;
-
-	spin_lock(&mm->page_table_lock);
-
-	pgdir = mm->context.huge_pgdir;
-	if (! pgdir)
-		goto out;
-
-	mm->context.huge_pgdir = NULL;
+		curareas = current->mm->context.high_htlb_areas;
 
-	/* cleanup any hugepte pages leftover */
-	for (i = 0; i < PTRS_PER_HUGEPGD; i++) {
-		pud_t *pud = (pud_t *)(pgdir + i);
-
-		if (! pud_none(*pud)) {
-			pte_t *pte = (pte_t *)pud_page(*pud);
-			struct page *ptepage = virt_to_page(pte);
+		/* First see if we can do the mapping in the existing
+		 * high areas */
+		addr = htlb_get_high_area(len, curareas);
+		if (addr != -ENOMEM)
+			return addr;
 
-			ptepage->mapping = NULL;
+		lastshift = 0;
+		for (areamask = HTLB_AREA_MASK(TASK_SIZE_USER64-len, len);
+		     ! lastshift; areamask >>=1) {
+			if (areamask & 1)
+				lastshift = 1;
 
-			BUG_ON(memcmp(pte, empty_zero_page, PAGE_SIZE));
-			kmem_cache_free(zero_cache, pte);
+			addr = htlb_get_high_area(len, curareas | areamask);
+			if ((addr != -ENOMEM)
+			    && open_high_hpage_areas(current->mm, areamask) == 0)
+				return addr;
 		}
-		pud_clear(pud);
 	}
-
-	BUG_ON(memcmp(pgdir, empty_zero_page, PAGE_SIZE));
-	kmem_cache_free(zero_cache, pgdir);
-
- out:
-	spin_unlock(&mm->page_table_lock);
+	printk(KERN_DEBUG "hugetlb_get_unmapped_area() unable to open"
+	       " enough areas\n");
+	return -ENOMEM;
 }
 
 int hash_huge_page(struct mm_struct *mm, unsigned long access,
diff --git a/arch/ppc64/mm/imalloc.c b/arch/ppc64/mm/imalloc.c
index b6e75b8..c65b87b 100644
--- a/arch/ppc64/mm/imalloc.c
+++ b/arch/ppc64/mm/imalloc.c
@@ -31,7 +31,7 @@ static int get_free_im_addr(unsigned long size, unsigned long *im_addr)
 			break;
 		if ((unsigned long)tmp->addr >= ioremap_bot)
 			addr = tmp->size + (unsigned long) tmp->addr;
-		if (addr > IMALLOC_END-size) 
+		if (addr >= IMALLOC_END-size)
 			return 1;
 	}
 	*im_addr = addr;
diff --git a/arch/ppc64/mm/init.c b/arch/ppc64/mm/init.c
index e58a24d..c02dc98 100644
--- a/arch/ppc64/mm/init.c
+++ b/arch/ppc64/mm/init.c
@@ -42,7 +42,6 @@
 
 #include <asm/pgalloc.h>
 #include <asm/page.h>
-#include <asm/abs_addr.h>
 #include <asm/prom.h>
 #include <asm/lmb.h>
 #include <asm/rtas.h>
@@ -66,6 +65,14 @@
 #include <asm/vdso.h>
 #include <asm/imalloc.h>
 
+#if PGTABLE_RANGE > USER_VSID_RANGE
+#warning Limited user VSID range means pagetable space is wasted
+#endif
+
+#if (TASK_SIZE_USER64 < PGTABLE_RANGE) && (TASK_SIZE_USER64 < USER_VSID_RANGE)
+#warning TASK_SIZE is smaller than it needs to be.
+#endif
+
 int mem_init_done;
 unsigned long ioremap_bot = IMALLOC_BASE;
 static unsigned long phbs_io_bot = PHBS_IO_BASE;
@@ -159,7 +166,6 @@ static int map_io_page(unsigned long ea, unsigned long pa, int flags)
 		ptep = pte_alloc_kernel(&init_mm, pmdp, ea);
 		if (!ptep)
 			return -ENOMEM;
-		pa = abs_to_phys(pa);
 		set_pte_at(&init_mm, ea, ptep, pfn_pte(pa >> PAGE_SHIFT,
 							  __pgprot(flags)));
 		spin_unlock(&init_mm.page_table_lock);
@@ -226,7 +232,7 @@ void __iomem * __ioremap(unsigned long addr, unsigned long size,
 	 * Before that, we map using addresses going
 	 * up from ioremap_bot.  imalloc will use
 	 * the addresses from ioremap_bot through
-	 * IMALLOC_END (0xE000001fffffffff)
+	 * IMALLOC_END
 	 * 
 	 */
 	pa = addr & PAGE_MASK;
@@ -417,12 +423,6 @@ int init_new_context(struct task_struct *tsk, struct mm_struct *mm)
 	int index;
 	int err;
 
-#ifdef CONFIG_HUGETLB_PAGE
-	/* We leave htlb_segs as it was, but for a fork, we need to
-	 * clear the huge_pgdir. */
-	mm->context.huge_pgdir = NULL;
-#endif
-
 again:
 	if (!idr_pre_get(&mmu_context_idr, GFP_KERNEL))
 		return -ENOMEM;
@@ -453,8 +453,6 @@ void destroy_context(struct mm_struct *mm)
 	spin_unlock(&mmu_context_lock);
 
 	mm->context.id = NO_CONTEXT;
-
-	hugetlb_mm_free_pgd(mm);
 }
 
 /*
@@ -484,9 +482,9 @@ void __init mm_init_ppc64(void)
 	for (i = 1; i < lmb.memory.cnt; i++) {
 		unsigned long base, prevbase, prevsize;
 
-		prevbase = lmb.memory.region[i-1].physbase;
+		prevbase = lmb.memory.region[i-1].base;
 		prevsize = lmb.memory.region[i-1].size;
-		base = lmb.memory.region[i].physbase;
+		base = lmb.memory.region[i].base;
 		if (base > (prevbase + prevsize)) {
 			io_hole_start = prevbase + prevsize;
 			io_hole_size = base  - (prevbase + prevsize);
@@ -513,11 +511,8 @@ int page_is_ram(unsigned long pfn)
 	for (i=0; i < lmb.memory.cnt; i++) {
 		unsigned long base;
 
-#ifdef CONFIG_MSCHUNKS
-		base = lmb.memory.region[i].physbase;
-#else
 		base = lmb.memory.region[i].base;
-#endif
+
 		if ((paddr >= base) &&
 			(paddr < (base + lmb.memory.region[i].size))) {
 			return 1;
@@ -547,7 +542,7 @@ void __init do_init_bootmem(void)
 	 */
 	bootmap_pages = bootmem_bootmap_pages(total_pages);
 
-	start = abs_to_phys(lmb_alloc(bootmap_pages<<PAGE_SHIFT, PAGE_SIZE));
+	start = lmb_alloc(bootmap_pages<<PAGE_SHIFT, PAGE_SIZE);
 	BUG_ON(!start);
 
 	boot_mapsize = init_bootmem(start >> PAGE_SHIFT, total_pages);
@@ -558,25 +553,25 @@ void __init do_init_bootmem(void)
 	 * present.
 	 */
 	for (i=0; i < lmb.memory.cnt; i++) {
-		unsigned long physbase, size;
+		unsigned long base, size;
 		unsigned long start_pfn, end_pfn;
 
-		physbase = lmb.memory.region[i].physbase;
+		base = lmb.memory.region[i].base;
 		size = lmb.memory.region[i].size;
 
-		start_pfn = physbase >> PAGE_SHIFT;
+		start_pfn = base >> PAGE_SHIFT;
 		end_pfn = start_pfn + (size >> PAGE_SHIFT);
 		memory_present(0, start_pfn, end_pfn);
 
-		free_bootmem(physbase, size);
+		free_bootmem(base, size);
 	}
 
 	/* reserve the sections we're already using */
 	for (i=0; i < lmb.reserved.cnt; i++) {
-		unsigned long physbase = lmb.reserved.region[i].physbase;
+		unsigned long base = lmb.reserved.region[i].base;
 		unsigned long size = lmb.reserved.region[i].size;
 
-		reserve_bootmem(physbase, size);
+		reserve_bootmem(base, size);
 	}
 }
 
@@ -615,10 +610,10 @@ static int __init setup_kcore(void)
 	int i;
 
 	for (i=0; i < lmb.memory.cnt; i++) {
-		unsigned long physbase, size;
+		unsigned long base, size;
 		struct kcore_list *kcore_mem;
 
-		physbase = lmb.memory.region[i].physbase;
+		base = lmb.memory.region[i].base;
 		size = lmb.memory.region[i].size;
 
 		/* GFP_ATOMIC to avoid might_sleep warnings during boot */
@@ -626,7 +621,7 @@ static int __init setup_kcore(void)
 		if (!kcore_mem)
 			panic("mem_init: kmalloc failed\n");
 
-		kclist_add(kcore_mem, __va(physbase), size);
+		kclist_add(kcore_mem, __va(base), size);
 	}
 
 	kclist_add(&kcore_vmem, (void *)VMALLOC_START, VMALLOC_END-VMALLOC_START);
@@ -686,9 +681,6 @@ void __init mem_init(void)
 
 	mem_init_done = 1;
 
-#ifdef CONFIG_PPC_ISERIES
-	iommu_vio_init();
-#endif
 	/* Initialize the vDSO */
 	vdso_init();
 }
@@ -833,23 +825,43 @@ void __iomem * reserve_phb_iospace(unsigned long size)
 	return virt_addr;
 }
 
-kmem_cache_t *zero_cache;
-
-static void zero_ctor(void *pte, kmem_cache_t *cache, unsigned long flags)
+static void zero_ctor(void *addr, kmem_cache_t *cache, unsigned long flags)
 {
-	memset(pte, 0, PAGE_SIZE);
+	memset(addr, 0, kmem_cache_size(cache));
 }
 
+static const int pgtable_cache_size[2] = {
+	PTE_TABLE_SIZE, PMD_TABLE_SIZE
+};
+static const char *pgtable_cache_name[ARRAY_SIZE(pgtable_cache_size)] = {
+	"pgd_pte_cache", "pud_pmd_cache",
+};
+
+kmem_cache_t *pgtable_cache[ARRAY_SIZE(pgtable_cache_size)];
+
 void pgtable_cache_init(void)
 {
-	zero_cache = kmem_cache_create("zero",
-				PAGE_SIZE,
-				0,
-				SLAB_HWCACHE_ALIGN | SLAB_MUST_HWCACHE_ALIGN,
-				zero_ctor,
-				NULL);
-	if (!zero_cache)
-		panic("pgtable_cache_init(): could not create zero_cache!\n");
+	int i;
+
+	BUILD_BUG_ON(PTE_TABLE_SIZE != pgtable_cache_size[PTE_CACHE_NUM]);
+	BUILD_BUG_ON(PMD_TABLE_SIZE != pgtable_cache_size[PMD_CACHE_NUM]);
+	BUILD_BUG_ON(PUD_TABLE_SIZE != pgtable_cache_size[PUD_CACHE_NUM]);
+	BUILD_BUG_ON(PGD_TABLE_SIZE != pgtable_cache_size[PGD_CACHE_NUM]);
+
+	for (i = 0; i < ARRAY_SIZE(pgtable_cache_size); i++) {
+		int size = pgtable_cache_size[i];
+		const char *name = pgtable_cache_name[i];
+
+		pgtable_cache[i] = kmem_cache_create(name,
+						     size, size,
+						     SLAB_HWCACHE_ALIGN
+						     | SLAB_MUST_HWCACHE_ALIGN,
+						     zero_ctor,
+						     NULL);
+		if (! pgtable_cache[i])
+			panic("pgtable_cache_init(): could not create %s!\n",
+			      name);
+	}
 }
 
 pgprot_t phys_mem_access_prot(struct file *file, unsigned long addr,
diff --git a/arch/ppc64/mm/numa.c b/arch/ppc64/mm/numa.c
index 0b191f2..c3116f0 100644
--- a/arch/ppc64/mm/numa.c
+++ b/arch/ppc64/mm/numa.c
@@ -671,7 +671,7 @@ new_range:
 		 * Mark reserved regions on this node
 		 */
 		for (i = 0; i < lmb.reserved.cnt; i++) {
-			unsigned long physbase = lmb.reserved.region[i].physbase;
+			unsigned long physbase = lmb.reserved.region[i].base;
 			unsigned long size = lmb.reserved.region[i].size;
 
 			if (pa_to_nid(physbase) != nid &&
diff --git a/arch/ppc64/mm/slb_low.S b/arch/ppc64/mm/slb_low.S
index 8379d67..bab2558 100644
--- a/arch/ppc64/mm/slb_low.S
+++ b/arch/ppc64/mm/slb_low.S
@@ -89,28 +89,29 @@ END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE)
 	b	9f
 
 0:	/* user address: proto-VSID = context<<15 | ESID */
-	li	r11,SLB_VSID_USER
-
-	srdi.	r9,r3,13
+	srdi.	r9,r3,USER_ESID_BITS
 	bne-	8f			/* invalid ea bits set */
 
 #ifdef CONFIG_HUGETLB_PAGE
 BEGIN_FTR_SECTION
-	/* check against the hugepage ranges */
-	cmpldi	r3,(TASK_HPAGE_END>>SID_SHIFT)
-	bge	6f			/* >= TASK_HPAGE_END */
-	cmpldi	r3,(TASK_HPAGE_BASE>>SID_SHIFT)
-	bge	5f			/* TASK_HPAGE_BASE..TASK_HPAGE_END */
+	lhz	r9,PACAHIGHHTLBAREAS(r13)
+	srdi	r11,r3,(HTLB_AREA_SHIFT-SID_SHIFT)
+	srd	r9,r9,r11
+	andi.	r9,r9,1
+	bne	5f
+
+	li	r11,SLB_VSID_USER
+
 	cmpldi	r3,16
-	bge	6f			/* 4GB..TASK_HPAGE_BASE */
+	bge	6f
 
-	lhz	r9,PACAHTLBSEGS(r13)
+	lhz	r9,PACALOWHTLBAREAS(r13)
 	srd	r9,r9,r3
 	andi.	r9,r9,1
+
 	beq	6f
 
-5:	/* this is a hugepage user address */
-	li	r11,(SLB_VSID_USER|SLB_VSID_L)
+5:	li	r11,SLB_VSID_USER|SLB_VSID_L
 END_FTR_SECTION_IFSET(CPU_FTR_16M_PAGE)
 #endif /* CONFIG_HUGETLB_PAGE */
 
diff --git a/arch/ppc64/mm/tlb.c b/arch/ppc64/mm/tlb.c
index 26f0172..d8a6593 100644
--- a/arch/ppc64/mm/tlb.c
+++ b/arch/ppc64/mm/tlb.c
@@ -41,7 +41,58 @@ DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 DEFINE_PER_CPU(struct pte_freelist_batch *, pte_freelist_cur);
 unsigned long pte_freelist_forced_free;
 
-void __pte_free_tlb(struct mmu_gather *tlb, struct page *ptepage)
+struct pte_freelist_batch
+{
+	struct rcu_head	rcu;
+	unsigned int	index;
+	pgtable_free_t	tables[0];
+};
+
+DEFINE_PER_CPU(struct pte_freelist_batch *, pte_freelist_cur);
+unsigned long pte_freelist_forced_free;
+
+#define PTE_FREELIST_SIZE \
+	((PAGE_SIZE - sizeof(struct pte_freelist_batch)) \
+	  / sizeof(pgtable_free_t))
+
+#ifdef CONFIG_SMP
+static void pte_free_smp_sync(void *arg)
+{
+	/* Do nothing, just ensure we sync with all CPUs */
+}
+#endif
+
+/* This is only called when we are critically out of memory
+ * (and fail to get a page in pte_free_tlb).
+ */
+static void pgtable_free_now(pgtable_free_t pgf)
+{
+	pte_freelist_forced_free++;
+
+	smp_call_function(pte_free_smp_sync, NULL, 0, 1);
+
+	pgtable_free(pgf);
+}
+
+static void pte_free_rcu_callback(struct rcu_head *head)
+{
+	struct pte_freelist_batch *batch =
+		container_of(head, struct pte_freelist_batch, rcu);
+	unsigned int i;
+
+	for (i = 0; i < batch->index; i++)
+		pgtable_free(batch->tables[i]);
+
+	free_page((unsigned long)batch);
+}
+
+static void pte_free_submit(struct pte_freelist_batch *batch)
+{
+	INIT_RCU_HEAD(&batch->rcu);
+	call_rcu(&batch->rcu, pte_free_rcu_callback);
+}
+
+void pgtable_free_tlb(struct mmu_gather *tlb, pgtable_free_t pgf)
 {
 	/* This is safe as we are holding page_table_lock */
         cpumask_t local_cpumask = cpumask_of_cpu(smp_processor_id());
@@ -49,19 +100,19 @@ void __pte_free_tlb(struct mmu_gather *tlb, struct page *ptepage)
 
 	if (atomic_read(&tlb->mm->mm_users) < 2 ||
 	    cpus_equal(tlb->mm->cpu_vm_mask, local_cpumask)) {
-		pte_free(ptepage);
+		pgtable_free(pgf);
 		return;
 	}
 
 	if (*batchp == NULL) {
 		*batchp = (struct pte_freelist_batch *)__get_free_page(GFP_ATOMIC);
 		if (*batchp == NULL) {
-			pte_free_now(ptepage);
+			pgtable_free_now(pgf);
 			return;
 		}
 		(*batchp)->index = 0;
 	}
-	(*batchp)->pages[(*batchp)->index++] = ptepage;
+	(*batchp)->tables[(*batchp)->index++] = pgf;
 	if ((*batchp)->index == PTE_FREELIST_SIZE) {
 		pte_free_submit(*batchp);
 		*batchp = NULL;
@@ -132,42 +183,6 @@ void __flush_tlb_pending(struct ppc64_tlb_batch *batch)
 	put_cpu();
 }
 
-#ifdef CONFIG_SMP
-static void pte_free_smp_sync(void *arg)
-{
-	/* Do nothing, just ensure we sync with all CPUs */
-}
-#endif
-
-/* This is only called when we are critically out of memory
- * (and fail to get a page in pte_free_tlb).
- */
-void pte_free_now(struct page *ptepage)
-{
-	pte_freelist_forced_free++;
-
-	smp_call_function(pte_free_smp_sync, NULL, 0, 1);
-
-	pte_free(ptepage);
-}
-
-static void pte_free_rcu_callback(struct rcu_head *head)
-{
-	struct pte_freelist_batch *batch =
-		container_of(head, struct pte_freelist_batch, rcu);
-	unsigned int i;
-
-	for (i = 0; i < batch->index; i++)
-		pte_free(batch->pages[i]);
-	free_page((unsigned long)batch);
-}
-
-void pte_free_submit(struct pte_freelist_batch *batch)
-{
-	INIT_RCU_HEAD(&batch->rcu);
-	call_rcu(&batch->rcu, pte_free_rcu_callback);
-}
-
 void pte_free_finish(void)
 {
 	/* This is safe as we are holding page_table_lock */
diff --git a/arch/ppc64/xmon/start.c b/arch/ppc64/xmon/start.c
index a9265bc..f86b584 100644
--- a/arch/ppc64/xmon/start.c
+++ b/arch/ppc64/xmon/start.c
@@ -27,7 +27,7 @@ static void sysrq_handle_xmon(int key, struct pt_regs *pt_regs,
 			      struct tty_struct *tty) 
 {
 	/* ensure xmon is enabled */
-	xmon_init();
+	xmon_init(1);
 	debugger(pt_regs);
 }
 
diff --git a/arch/ppc64/xmon/xmon.c b/arch/ppc64/xmon/xmon.c
index 0553943..45908b1 100644
--- a/arch/ppc64/xmon/xmon.c
+++ b/arch/ppc64/xmon/xmon.c
@@ -2496,15 +2496,25 @@ static void dump_stab(void)
 	}
 }
 
-void xmon_init(void)
-{
-	__debugger = xmon;
-	__debugger_ipi = xmon_ipi;
-	__debugger_bpt = xmon_bpt;
-	__debugger_sstep = xmon_sstep;
-	__debugger_iabr_match = xmon_iabr_match;
-	__debugger_dabr_match = xmon_dabr_match;
-	__debugger_fault_handler = xmon_fault_handler;
+void xmon_init(int enable)
+{
+	if (enable) {
+		__debugger = xmon;
+		__debugger_ipi = xmon_ipi;
+		__debugger_bpt = xmon_bpt;
+		__debugger_sstep = xmon_sstep;
+		__debugger_iabr_match = xmon_iabr_match;
+		__debugger_dabr_match = xmon_dabr_match;
+		__debugger_fault_handler = xmon_fault_handler;
+	} else {
+		__debugger = NULL;
+		__debugger_ipi = NULL;
+		__debugger_bpt = NULL;
+		__debugger_sstep = NULL;
+		__debugger_iabr_match = NULL;
+		__debugger_dabr_match = NULL;
+		__debugger_fault_handler = NULL;
+	}
 }
 
 void dump_segments(void)