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

28 files changed, 1177 insertions, 1044 deletions

diff --git a/arch/ppc64/kernel/LparData.c b/arch/ppc64/kernel/LparData.c
index 1c11031c838e..0a9c23ca2f0c 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 2ecccb6b4f8c..f4b3bfcc109d 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 abb9e5b5da03..17e35d0fed09 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 77cec42f9525..4847f2ac8c9f 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 000000000000..d8432c0fb27d
--- /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 accaa052d31f..cccec4902646 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 b0250ae4a72a..2192055a90a0 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 a649edbb23b6..3ffefbbc6623 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 000000000000..b4268cc4ba48
--- /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 d6c6bd03d2a4..5adaca2ddc9d 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 02e96627fa66..9d034ff062b1 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 a05b50b738e9..474df0a862bf 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 69130522a87e..9d5e1e7fc389 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 74dd144dcce8..0a3ddc9227c5 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 5bec956e44a0..f0f0630cf07c 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 62c55a123560..79c7f3223665 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 000000000000..338f9e1bdc09
--- /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 6316188737b6..6182a2cd90a5 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 e40877fa67cd..8ff86a766cdf 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 67be773f9c00..cdfec7438d01 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 f7cae05e40fb..7a7e027653ad 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 5aca01ddd81f..b21848826791 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 dbbe6c79d8da..adcf972711fc 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 1048817befb8..1dccadaddd1d 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 e9c24d2dbd91..ee3b20de2e7a 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 02b8ac4e0168..f311ee7c0070 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 909462e1adea..1696e1b05bb9 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 0c0ba71ac0e8..3b790bafcaad 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;