Merge branch 'dma-debug' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/linux-2.6-iommu into x86/urgent

128 files changed, 3744 insertions, 460 deletions

diff --git a/arch/alpha/include/asm/spinlock.h b/arch/alpha/include/asm/spinlock.h
index aeeb125f6851..e38fb95cb335 100644
--- a/arch/alpha/include/asm/spinlock.h
+++ b/arch/alpha/include/asm/spinlock.h
@@ -166,6 +166,9 @@ static inline void __raw_write_unlock(raw_rwlock_t * lock)
         lock->lock = 0;
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
 #define _raw_write_relax(lock)  cpu_relax()
diff --git a/arch/alpha/kernel/process.c b/arch/alpha/kernel/process.c
index 8d0097f10208..3a2fb7a02db4 100644
--- a/arch/alpha/kernel/process.c
+++ b/arch/alpha/kernel/process.c
@@ -272,7 +272,7 @@ alpha_vfork(struct pt_regs *regs)
  */
 
 int
-copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+copy_thread(unsigned long clone_flags, unsigned long usp,
             unsigned long unused,
             struct task_struct * p, struct pt_regs * regs)
 {
diff --git a/arch/arm/include/asm/spinlock.h b/arch/arm/include/asm/spinlock.h
index 2b41ebbfa7ff..c13681ac1ede 100644
--- a/arch/arm/include/asm/spinlock.h
+++ b/arch/arm/include/asm/spinlock.h
@@ -217,6 +217,9 @@ static inline int __raw_read_trylock(raw_rwlock_t *rw)
 /* read_can_lock - would read_trylock() succeed? */
 #define __raw_read_can_lock(x)          ((x)->lock < 0x80000000)
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
 #define _raw_write_relax(lock)  cpu_relax()
diff --git a/arch/arm/kernel/process.c b/arch/arm/kernel/process.c
index 2de14e2afdc5..c3265a2e7cd4 100644
--- a/arch/arm/kernel/process.c
+++ b/arch/arm/kernel/process.c
@@ -301,7 +301,7 @@ void release_thread(struct task_struct *dead_task)
 asmlinkage void ret_from_fork(void) __asm__("ret_from_fork");
 
 int
-copy_thread(int nr, unsigned long clone_flags, unsigned long stack_start,
+copy_thread(unsigned long clone_flags, unsigned long stack_start,
             unsigned long stk_sz, struct task_struct *p, struct pt_regs *regs)
 {
         struct thread_info *thread = task_thread_info(p);
diff --git a/arch/avr32/kernel/process.c b/arch/avr32/kernel/process.c
index 43ae555ecb33..1bbe1da54869 100644
--- a/arch/avr32/kernel/process.c
+++ b/arch/avr32/kernel/process.c
@@ -332,7 +332,7 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 
 asmlinkage void ret_from_fork(void);
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/blackfin/kernel/process.c b/arch/blackfin/kernel/process.c
index 33e2e8993f7f..f49427293ca1 100644
--- a/arch/blackfin/kernel/process.c
+++ b/arch/blackfin/kernel/process.c
@@ -193,7 +193,7 @@ asmlinkage int bfin_clone(struct pt_regs *regs)
 }
 
 int
-copy_thread(int nr, unsigned long clone_flags,
+copy_thread(unsigned long clone_flags,
             unsigned long usp, unsigned long topstk,
             struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/cris/arch-v10/kernel/process.c b/arch/cris/arch-v10/kernel/process.c
index bd9b3ff63f6c..c4c69cf721e5 100644
--- a/arch/cris/arch-v10/kernel/process.c
+++ b/arch/cris/arch-v10/kernel/process.c
@@ -115,7 +115,7 @@ int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
  */
 asmlinkage void ret_from_fork(void);
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/cris/arch-v32/kernel/process.c b/arch/cris/arch-v32/kernel/process.c
index ced5b725d9bd..120e7f796fea 100644
--- a/arch/cris/arch-v32/kernel/process.c
+++ b/arch/cris/arch-v32/kernel/process.c
@@ -131,7 +131,7 @@ kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
 extern asmlinkage void ret_from_fork(void);
 
 int
-copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+copy_thread(unsigned long clone_flags, unsigned long usp,
         unsigned long unused,
         struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/cris/include/arch-v32/arch/spinlock.h b/arch/cris/include/arch-v32/arch/spinlock.h
index 0d5709b983a1..129756b96661 100644
--- a/arch/cris/include/arch-v32/arch/spinlock.h
+++ b/arch/cris/include/arch-v32/arch/spinlock.h
@@ -121,6 +121,8 @@ static  inline int __raw_write_trylock(raw_rwlock_t *rw)
         return 1;
 }
 
+#define _raw_read_lock_flags(lock, flags) _raw_read_lock(lock)
+#define _raw_write_lock_flags(lock, flags) _raw_write_lock(lock)
 
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
diff --git a/arch/cris/kernel/process.c b/arch/cris/kernel/process.c
index 60816e876455..4df0b320d524 100644
--- a/arch/cris/kernel/process.c
+++ b/arch/cris/kernel/process.c
@@ -19,7 +19,6 @@
 #include <asm/system.h>
 #include <linux/module.h>
 #include <linux/spinlock.h>
-#include <linux/fs_struct.h>
 #include <linux/init_task.h>
 #include <linux/sched.h>
 #include <linux/fs.h>
diff --git a/arch/frv/kernel/process.c b/arch/frv/kernel/process.c
index 9583a338e9d6..0de50df74970 100644
--- a/arch/frv/kernel/process.c
+++ b/arch/frv/kernel/process.c
@@ -204,7 +204,7 @@ void prepare_to_copy(struct task_struct *tsk)
 /*
  * set up the kernel stack and exception frames for a new process
  */
-int copy_thread(int nr, unsigned long clone_flags,
+int copy_thread(unsigned long clone_flags,
                 unsigned long usp, unsigned long topstk,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/h8300/kernel/process.c b/arch/h8300/kernel/process.c
index a8ef654a5a0b..e2f33d0f9969 100644
--- a/arch/h8300/kernel/process.c
+++ b/arch/h8300/kernel/process.c
@@ -191,7 +191,7 @@ asmlinkage int h8300_clone(struct pt_regs *regs)
 
 }
 
-int copy_thread(int nr, unsigned long clone_flags,
+int copy_thread(unsigned long clone_flags,
                 unsigned long usp, unsigned long topstk,
                  struct task_struct * p, struct pt_regs * regs)
 {
diff --git a/arch/ia64/configs/generic_defconfig b/arch/ia64/configs/generic_defconfig
index a109db30ce55..75645495c2dd 100644
--- a/arch/ia64/configs/generic_defconfig
+++ b/arch/ia64/configs/generic_defconfig
@@ -193,7 +193,6 @@ CONFIG_BOUNCE=y
 CONFIG_NR_QUICK=1
 CONFIG_VIRT_TO_BUS=y
 CONFIG_UNEVICTABLE_LRU=y
-CONFIG_MMU_NOTIFIER=y
 CONFIG_ARCH_SELECT_MEMORY_MODEL=y
 CONFIG_ARCH_DISCONTIGMEM_ENABLE=y
 CONFIG_ARCH_FLATMEM_ENABLE=y
@@ -416,8 +415,6 @@ CONFIG_SGI_IOC4=y
 # CONFIG_ENCLOSURE_SERVICES is not set
 CONFIG_SGI_XP=m
 # CONFIG_HP_ILO is not set
-CONFIG_SGI_GRU=m
-# CONFIG_SGI_GRU_DEBUG is not set
 # CONFIG_C2PORT is not set
 CONFIG_HAVE_IDE=y
 CONFIG_IDE=y
diff --git a/arch/ia64/include/asm/intrinsics.h b/arch/ia64/include/asm/intrinsics.h
index c47830e26cb7..111ed5222892 100644
--- a/arch/ia64/include/asm/intrinsics.h
+++ b/arch/ia64/include/asm/intrinsics.h
@@ -202,7 +202,11 @@ extern long ia64_cmpxchg_called_with_bad_pointer (void);
 
 #ifndef __ASSEMBLY__
 #if defined(CONFIG_PARAVIRT) && defined(__KERNEL__)
-#define IA64_INTRINSIC_API(name)        pv_cpu_ops.name
+#ifdef ASM_SUPPORTED
+# define IA64_INTRINSIC_API(name)       paravirt_ ## name
+#else
+# define IA64_INTRINSIC_API(name)       pv_cpu_ops.name
+#endif
 #define IA64_INTRINSIC_MACRO(name)      paravirt_ ## name
 #else
 #define IA64_INTRINSIC_API(name)        ia64_native_ ## name
diff --git a/arch/ia64/include/asm/mmu_context.h b/arch/ia64/include/asm/mmu_context.h
index 040bc87db930..7f2a456603cb 100644
--- a/arch/ia64/include/asm/mmu_context.h
+++ b/arch/ia64/include/asm/mmu_context.h
@@ -87,7 +87,7 @@ get_mmu_context (struct mm_struct *mm)
         /* re-check, now that we've got the lock: */
         context = mm->context;
         if (context == 0) {
-                cpus_clear(mm->cpu_vm_mask);
+                cpumask_clear(mm_cpumask(mm));
                 if (ia64_ctx.next >= ia64_ctx.limit) {
                         ia64_ctx.next = find_next_zero_bit(ia64_ctx.bitmap,
                                         ia64_ctx.max_ctx, ia64_ctx.next);
@@ -166,8 +166,8 @@ activate_context (struct mm_struct *mm)
 
         do {
                 context = get_mmu_context(mm);
-                if (!cpu_isset(smp_processor_id(), mm->cpu_vm_mask))
-                        cpu_set(smp_processor_id(), mm->cpu_vm_mask);
+                if (!cpumask_test_cpu(smp_processor_id(), mm_cpumask(mm)))
+                        cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm));
                 reload_context(context);
                 /*
                  * in the unlikely event of a TLB-flush by another thread,
diff --git a/arch/ia64/include/asm/module.h b/arch/ia64/include/asm/module.h
index d2da61e4c49b..908eaef42a08 100644
--- a/arch/ia64/include/asm/module.h
+++ b/arch/ia64/include/asm/module.h
@@ -16,6 +16,12 @@ struct mod_arch_specific {
         struct elf64_shdr *got;         /* global offset table */
         struct elf64_shdr *opd;         /* official procedure descriptors */
         struct elf64_shdr *unwind;      /* unwind-table section */
+#ifdef CONFIG_PARAVIRT
+        struct elf64_shdr *paravirt_bundles;
+                                        /* paravirt_alt_bundle_patch table */
+        struct elf64_shdr *paravirt_insts;
+                                        /* paravirt_alt_inst_patch table */
+#endif
         unsigned long gp;               /* global-pointer for module */
 
         void *core_unw_table;           /* core unwind-table cookie returned by unwinder */
diff --git a/arch/ia64/include/asm/native/inst.h b/arch/ia64/include/asm/native/inst.h
index 0a1026cca4fa..d2d46efb3e6e 100644
--- a/arch/ia64/include/asm/native/inst.h
+++ b/arch/ia64/include/asm/native/inst.h
@@ -30,6 +30,9 @@
 #define __paravirt_work_processed_syscall_target \
                                                 ia64_work_processed_syscall
 
+#define paravirt_fsyscall_table                 ia64_native_fsyscall_table
+#define paravirt_fsys_bubble_down               ia64_native_fsys_bubble_down
+
 #ifdef CONFIG_PARAVIRT_GUEST_ASM_CLOBBER_CHECK
 # define PARAVIRT_POISON        0xdeadbeefbaadf00d
 # define CLOBBER(clob)                          \
@@ -74,6 +77,11 @@
 (pred)  mov reg = psr                   \
         CLOBBER(clob)
 
+#define MOV_FROM_ITC(pred, pred_clob, reg, clob)        \
+(pred)  mov reg = ar.itc                                \
+        CLOBBER(clob)                                   \
+        CLOBBER_PRED(pred_clob)
+
 #define MOV_TO_IFA(reg, clob)   \
         mov cr.ifa = reg        \
         CLOBBER(clob)
@@ -158,6 +166,11 @@
 #define RSM_PSR_DT              \
         rsm psr.dt
 
+#define RSM_PSR_BE_I(clob0, clob1)      \
+        rsm psr.be | psr.i              \
+        CLOBBER(clob0)                  \
+        CLOBBER(clob1)
+
 #define SSM_PSR_DT_AND_SRLZ_I   \
         ssm psr.dt              \
         ;;                      \
diff --git a/arch/ia64/include/asm/native/patchlist.h b/arch/ia64/include/asm/native/patchlist.h
new file mode 100644
index 000000000000..be16ca9311bf
--- /dev/null
+++ b/arch/ia64/include/asm/native/patchlist.h
@@ -0,0 +1,38 @@
+/******************************************************************************
+ * arch/ia64/include/asm/native/inst.h
+ *
+ * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#define __paravirt_start_gate_fsyscall_patchlist                \
+        __ia64_native_start_gate_fsyscall_patchlist
+#define __paravirt_end_gate_fsyscall_patchlist                  \
+        __ia64_native_end_gate_fsyscall_patchlist
+#define __paravirt_start_gate_brl_fsys_bubble_down_patchlist    \
+        __ia64_native_start_gate_brl_fsys_bubble_down_patchlist
+#define __paravirt_end_gate_brl_fsys_bubble_down_patchlist      \
+        __ia64_native_end_gate_brl_fsys_bubble_down_patchlist
+#define __paravirt_start_gate_vtop_patchlist                    \
+        __ia64_native_start_gate_vtop_patchlist
+#define __paravirt_end_gate_vtop_patchlist                      \
+        __ia64_native_end_gate_vtop_patchlist
+#define __paravirt_start_gate_mckinley_e9_patchlist             \
+        __ia64_native_start_gate_mckinley_e9_patchlist
+#define __paravirt_end_gate_mckinley_e9_patchlist               \
+        __ia64_native_end_gate_mckinley_e9_patchlist
diff --git a/arch/ia64/include/asm/native/pvchk_inst.h b/arch/ia64/include/asm/native/pvchk_inst.h
index b8e6eb1090d7..8d72962ec838 100644
--- a/arch/ia64/include/asm/native/pvchk_inst.h
+++ b/arch/ia64/include/asm/native/pvchk_inst.h
@@ -180,6 +180,11 @@
         IS_PRED_IN(pred)                        \
         IS_RREG_OUT(reg)                        \
         IS_RREG_CLOB(clob)
+#define MOV_FROM_ITC(pred, pred_clob, reg, clob)        \
+        IS_PRED_IN(pred)                                \
+        IS_PRED_CLOB(pred_clob)                         \
+        IS_RREG_OUT(reg)                                \
+        IS_RREG_CLOB(clob)
 #define MOV_TO_IFA(reg, clob)                   \
         IS_RREG_IN(reg)                         \
         IS_RREG_CLOB(clob)
@@ -246,6 +251,9 @@
         IS_RREG_CLOB(clob2)
 #define RSM_PSR_DT                              \
         nop 0
+#define RSM_PSR_BE_I(clob0, clob1)              \
+        IS_RREG_CLOB(clob0)                     \
+        IS_RREG_CLOB(clob1)
 #define SSM_PSR_DT_AND_SRLZ_I                   \
         nop 0
 #define BSW_0(clob0, clob1, clob2)              \
diff --git a/arch/ia64/include/asm/paravirt.h b/arch/ia64/include/asm/paravirt.h
index 2bf3636473fe..2eb0a981a09a 100644
--- a/arch/ia64/include/asm/paravirt.h
+++ b/arch/ia64/include/asm/paravirt.h
@@ -22,6 +22,56 @@
 #ifndef __ASM_PARAVIRT_H
 #define __ASM_PARAVIRT_H
 
+#ifndef __ASSEMBLY__
+/******************************************************************************
+ * fsys related addresses
+ */
+struct pv_fsys_data {
+        unsigned long *fsyscall_table;
+        void *fsys_bubble_down;
+};
+
+extern struct pv_fsys_data pv_fsys_data;
+
+unsigned long *paravirt_get_fsyscall_table(void);
+char *paravirt_get_fsys_bubble_down(void);
+
+/******************************************************************************
+ * patchlist addresses for gate page
+ */
+enum pv_gate_patchlist {
+        PV_GATE_START_FSYSCALL,
+        PV_GATE_END_FSYSCALL,
+
+        PV_GATE_START_BRL_FSYS_BUBBLE_DOWN,
+        PV_GATE_END_BRL_FSYS_BUBBLE_DOWN,
+
+        PV_GATE_START_VTOP,
+        PV_GATE_END_VTOP,
+
+        PV_GATE_START_MCKINLEY_E9,
+        PV_GATE_END_MCKINLEY_E9,
+};
+
+struct pv_patchdata {
+        unsigned long start_fsyscall_patchlist;
+        unsigned long end_fsyscall_patchlist;
+        unsigned long start_brl_fsys_bubble_down_patchlist;
+        unsigned long end_brl_fsys_bubble_down_patchlist;
+        unsigned long start_vtop_patchlist;
+        unsigned long end_vtop_patchlist;
+        unsigned long start_mckinley_e9_patchlist;
+        unsigned long end_mckinley_e9_patchlist;
+
+        void *gate_section;
+};
+
+extern struct pv_patchdata pv_patchdata;
+
+unsigned long paravirt_get_gate_patchlist(enum pv_gate_patchlist type);
+void *paravirt_get_gate_section(void);
+#endif
+
 #ifdef CONFIG_PARAVIRT_GUEST
 
 #define PARAVIRT_HYPERVISOR_TYPE_DEFAULT        0
@@ -68,6 +118,14 @@ struct pv_init_ops {
         int (*arch_setup_nomca)(void);
 
         void (*post_smp_prepare_boot_cpu)(void);
+
+#ifdef ASM_SUPPORTED
+        unsigned long (*patch_bundle)(void *sbundle, void *ebundle,
+                                      unsigned long type);
+        unsigned long (*patch_inst)(unsigned long stag, unsigned long etag,
+                                    unsigned long type);
+#endif
+        void (*patch_branch)(unsigned long tag, unsigned long type);
 };
 
 extern struct pv_init_ops pv_init_ops;
@@ -210,6 +268,8 @@ struct pv_time_ops {
         int (*do_steal_accounting)(unsigned long *new_itm);
 
         void (*clocksource_resume)(void);
+
+        unsigned long long (*sched_clock)(void);
 };
 
 extern struct pv_time_ops pv_time_ops;
@@ -227,6 +287,11 @@ paravirt_do_steal_accounting(unsigned long *new_itm)
         return pv_time_ops.do_steal_accounting(new_itm);
 }
 
+static inline unsigned long long paravirt_sched_clock(void)
+{
+        return pv_time_ops.sched_clock();
+}
+
 #endif /* !__ASSEMBLY__ */
 
 #else
diff --git a/arch/ia64/include/asm/paravirt_patch.h b/arch/ia64/include/asm/paravirt_patch.h
new file mode 100644
index 000000000000..128ff5db6e67
--- /dev/null
+++ b/arch/ia64/include/asm/paravirt_patch.h
@@ -0,0 +1,143 @@
+/******************************************************************************
+ * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#ifndef __ASM_PARAVIRT_PATCH_H
+#define __ASM_PARAVIRT_PATCH_H
+
+#ifdef __ASSEMBLY__
+
+        .section .paravirt_branches, "a"
+        .previous
+#define PARAVIRT_PATCH_SITE_BR(type)            \
+        {                                       \
+        [1:] ;                                  \
+        br.cond.sptk.many 2f ;                  \
+        nop.b 0 ;                               \
+        nop.b 0;; ;                             \
+        } ;                                     \
+        2:                                      \
+        .xdata8 ".paravirt_branches", 1b, type
+
+#else
+
+#include <linux/stringify.h>
+#include <asm/intrinsics.h>
+
+/* for binary patch */
+struct paravirt_patch_site_bundle {
+        void            *sbundle;
+        void            *ebundle;
+        unsigned long   type;
+};
+
+/* label means the beginning of new bundle */
+#define paravirt_alt_bundle(instr, privop)                              \
+        "\t998:\n"                                                      \
+        "\t" instr "\n"                                                 \
+        "\t999:\n"                                                      \
+        "\t.pushsection .paravirt_bundles, \"a\"\n"                     \
+        "\t.popsection\n"                                               \
+        "\t.xdata8 \".paravirt_bundles\", 998b, 999b, "                 \
+        __stringify(privop) "\n"
+
+
+struct paravirt_patch_bundle_elem {
+        const void      *sbundle;
+        const void      *ebundle;
+        unsigned long   type;
+};
+
+
+struct paravirt_patch_site_inst {
+        unsigned long   stag;
+        unsigned long   etag;
+        unsigned long   type;
+};
+
+#define paravirt_alt_inst(instr, privop)                                \
+        "\t[998:]\n"                                                    \
+        "\t" instr "\n"                                                 \
+        "\t[999:]\n"                                                    \
+        "\t.pushsection .paravirt_insts, \"a\"\n"                       \
+        "\t.popsection\n"                                               \
+        "\t.xdata8 \".paravirt_insts\", 998b, 999b, "                   \
+        __stringify(privop) "\n"
+
+struct paravirt_patch_site_branch {
+        unsigned long   tag;
+        unsigned long   type;
+};
+
+struct paravirt_patch_branch_target {
+        const void      *entry;
+        unsigned long   type;
+};
+
+void
+__paravirt_patch_apply_branch(
+        unsigned long tag, unsigned long type,
+        const struct paravirt_patch_branch_target *entries,
+        unsigned int nr_entries);
+
+void
+paravirt_patch_reloc_br(unsigned long tag, const void *target);
+
+void
+paravirt_patch_reloc_brl(unsigned long tag, const void *target);
+
+
+#if defined(ASM_SUPPORTED) && defined(CONFIG_PARAVIRT)
+unsigned long
+ia64_native_patch_bundle(void *sbundle, void *ebundle, unsigned long type);
+
+unsigned long
+__paravirt_patch_apply_bundle(void *sbundle, void *ebundle, unsigned long type,
+                              const struct paravirt_patch_bundle_elem *elems,
+                              unsigned long nelems,
+                              const struct paravirt_patch_bundle_elem **found);
+
+void
+paravirt_patch_apply_bundle(const struct paravirt_patch_site_bundle *start,
+                            const struct paravirt_patch_site_bundle *end);
+
+void
+paravirt_patch_apply_inst(const struct paravirt_patch_site_inst *start,
+                          const struct paravirt_patch_site_inst *end);
+
+void paravirt_patch_apply(void);
+#else
+#define paravirt_patch_apply_bundle(start, end) do { } while (0)
+#define paravirt_patch_apply_inst(start, end)   do { } while (0)
+#define paravirt_patch_apply()                  do { } while (0)
+#endif
+
+#endif /* !__ASSEMBLEY__ */
+
+#endif /* __ASM_PARAVIRT_PATCH_H */
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "linux"
+ * c-basic-offset: 8
+ * tab-width: 8
+ * indent-tabs-mode: t
+ * End:
+ */
diff --git a/arch/ia64/include/asm/paravirt_privop.h b/arch/ia64/include/asm/paravirt_privop.h
index 33c8e55f5775..3d2951130b5f 100644
--- a/arch/ia64/include/asm/paravirt_privop.h
+++ b/arch/ia64/include/asm/paravirt_privop.h
@@ -33,7 +33,7 @@
  */
 
 struct pv_cpu_ops {
-        void (*fc)(unsigned long addr);
+        void (*fc)(void *addr);
         unsigned long (*thash)(unsigned long addr);
         unsigned long (*get_cpuid)(int index);
         unsigned long (*get_pmd)(int index);
@@ -60,12 +60,18 @@ extern unsigned long ia64_native_getreg_func(int regnum);
 /* Instructions paravirtualized for performance */
 /************************************************/
 
+#ifndef ASM_SUPPORTED
+#define paravirt_ssm_i()        pv_cpu_ops.ssm_i()
+#define paravirt_rsm_i()        pv_cpu_ops.rsm_i()
+#define __paravirt_getreg()     pv_cpu_ops.getreg()
+#endif
+
 /* mask for ia64_native_ssm/rsm() must be constant.("i" constraing).
  * static inline function doesn't satisfy it. */
 #define paravirt_ssm(mask)                      \
         do {                                    \
                 if ((mask) == IA64_PSR_I)       \
-                        pv_cpu_ops.ssm_i();     \
+                        paravirt_ssm_i();       \
                 else                            \
                         ia64_native_ssm(mask);  \
         } while (0)
@@ -73,7 +79,7 @@ extern unsigned long ia64_native_getreg_func(int regnum);
 #define paravirt_rsm(mask)                      \
         do {                                    \
                 if ((mask) == IA64_PSR_I)       \
-                        pv_cpu_ops.rsm_i();     \
+                        paravirt_rsm_i();       \
                 else                            \
                         ia64_native_rsm(mask);  \
         } while (0)
@@ -86,7 +92,7 @@ extern unsigned long ia64_native_getreg_func(int regnum);
                 if ((reg) == _IA64_REG_IP)                      \
                         res = ia64_native_getreg(_IA64_REG_IP); \
                 else                                            \
-                        res = pv_cpu_ops.getreg(reg);           \
+                        res = __paravirt_getreg(reg);           \
                 res;                                            \
         })
 
@@ -112,6 +118,12 @@ void paravirt_cpu_asm_init(const struct pv_cpu_asm_switch *cpu_asm_switch);
 
 #endif /* CONFIG_PARAVIRT */
 
+#if defined(CONFIG_PARAVIRT) && defined(ASM_SUPPORTED)
+#define paravirt_dv_serialize_data()    ia64_dv_serialize_data()
+#else
+#define paravirt_dv_serialize_data()    /* nothing */
+#endif
+
 /* these routines utilize privilege-sensitive or performance-sensitive
  * privileged instructions so the code must be replaced with
  * paravirtualized versions */
@@ -121,4 +133,349 @@ void paravirt_cpu_asm_init(const struct pv_cpu_asm_switch *cpu_asm_switch);
         IA64_PARAVIRT_ASM_FUNC(work_processed_syscall)
 #define ia64_leave_kernel               IA64_PARAVIRT_ASM_FUNC(leave_kernel)
 
+
+#if defined(CONFIG_PARAVIRT)
+/******************************************************************************
+ * binary patching infrastructure
+ */
+#define PARAVIRT_PATCH_TYPE_FC                          1
+#define PARAVIRT_PATCH_TYPE_THASH                       2
+#define PARAVIRT_PATCH_TYPE_GET_CPUID                   3
+#define PARAVIRT_PATCH_TYPE_GET_PMD                     4
+#define PARAVIRT_PATCH_TYPE_PTCGA                       5
+#define PARAVIRT_PATCH_TYPE_GET_RR                      6
+#define PARAVIRT_PATCH_TYPE_SET_RR                      7
+#define PARAVIRT_PATCH_TYPE_SET_RR0_TO_RR4              8
+#define PARAVIRT_PATCH_TYPE_SSM_I                       9
+#define PARAVIRT_PATCH_TYPE_RSM_I                       10
+#define PARAVIRT_PATCH_TYPE_GET_PSR_I                   11
+#define PARAVIRT_PATCH_TYPE_INTRIN_LOCAL_IRQ_RESTORE    12
+
+/* PARAVIRT_PATY_TYPE_[GS]ETREG + _IA64_REG_xxx */
+#define PARAVIRT_PATCH_TYPE_GETREG                      0x10000000
+#define PARAVIRT_PATCH_TYPE_SETREG                      0x20000000
+
+/*
+ * struct task_struct* (*ia64_switch_to)(void* next_task);
+ * void *ia64_leave_syscall;
+ * void *ia64_work_processed_syscall
+ * void *ia64_leave_kernel;
+ */
+
+#define PARAVIRT_PATCH_TYPE_BR_START                    0x30000000
+#define PARAVIRT_PATCH_TYPE_BR_SWITCH_TO                \
+        (PARAVIRT_PATCH_TYPE_BR_START + 0)
+#define PARAVIRT_PATCH_TYPE_BR_LEAVE_SYSCALL            \
+        (PARAVIRT_PATCH_TYPE_BR_START + 1)
+#define PARAVIRT_PATCH_TYPE_BR_WORK_PROCESSED_SYSCALL   \
+        (PARAVIRT_PATCH_TYPE_BR_START + 2)
+#define PARAVIRT_PATCH_TYPE_BR_LEAVE_KERNEL             \
+        (PARAVIRT_PATCH_TYPE_BR_START + 3)
+
+#ifdef ASM_SUPPORTED
+#include <asm/paravirt_patch.h>
+
+/*
+ * pv_cpu_ops calling stub.
+ * normal function call convension can't be written by gcc
+ * inline assembly.
+ *
+ * from the caller's point of view,
+ * the following registers will be clobbered.
+ * r2, r3
+ * r8-r15
+ * r16, r17
+ * b6, b7
+ * p6-p15
+ * ar.ccv
+ *
+ * from the callee's point of view ,
+ * the following registers can be used.
+ * r2, r3: scratch
+ * r8: scratch, input argument0 and return value
+ * r0-r15: scratch, input argument1-5
+ * b6: return pointer
+ * b7: scratch
+ * p6-p15: scratch
+ * ar.ccv: scratch
+ *
+ * other registers must not be changed. especially
+ * b0: rp: preserved. gcc ignores b0 in clobbered register.
+ * r16: saved gp
+ */
+/* 5 bundles */
+#define __PARAVIRT_BR                                                   \
+        ";;\n"                                                          \
+        "{ .mlx\n"                                                      \
+        "nop 0\n"                                                       \
+        "movl r2 = %[op_addr]\n"/* get function pointer address */      \
+        ";;\n"                                                          \
+        "}\n"                                                           \
+        "1:\n"                                                          \
+        "{ .mii\n"                                                      \
+        "ld8 r2 = [r2]\n"       /* load function descriptor address */  \
+        "mov r17 = ip\n"        /* get ip to calc return address */     \
+        "mov r16 = gp\n"        /* save gp */                           \
+        ";;\n"                                                          \
+        "}\n"                                                           \
+        "{ .mii\n"                                                      \
+        "ld8 r3 = [r2], 8\n"    /* load entry address */                \
+        "adds r17 =  1f - 1b, r17\n"    /* calculate return address */  \
+        ";;\n"                                                          \
+        "mov b7 = r3\n"         /* set entry address */                 \
+        "}\n"                                                           \
+        "{ .mib\n"                                                      \
+        "ld8 gp = [r2]\n"       /* load gp value */                     \
+        "mov b6 = r17\n"        /* set return address */                \
+        "br.cond.sptk.few b7\n" /* intrinsics are very short isns */    \
+        "}\n"                                                           \
+        "1:\n"                                                          \
+        "{ .mii\n"                                                      \
+        "mov gp = r16\n"        /* restore gp value */                  \
+        "nop 0\n"                                                       \
+        "nop 0\n"                                                       \
+        ";;\n"                                                          \
+        "}\n"
+
+#define PARAVIRT_OP(op)                         \
+        [op_addr] "i"(&pv_cpu_ops.op)
+
+#define PARAVIRT_TYPE(type)                     \
+        PARAVIRT_PATCH_TYPE_ ## type
+
+#define PARAVIRT_REG_CLOBBERS0                                  \
+        "r2", "r3", /*"r8",*/ "r9", "r10", "r11", "r14",        \
+                "r15", "r16", "r17"
+
+#define PARAVIRT_REG_CLOBBERS1                                  \
+        "r2","r3", /*"r8",*/ "r9", "r10", "r11", "r14",         \
+                "r15", "r16", "r17"
+
+#define PARAVIRT_REG_CLOBBERS2                                  \
+        "r2", "r3", /*"r8", "r9",*/ "r10", "r11", "r14",        \
+                "r15", "r16", "r17"
+
+#define PARAVIRT_REG_CLOBBERS5                                  \
+        "r2", "r3", /*"r8", "r9", "r10", "r11", "r14",*/        \
+                "r15", "r16", "r17"
+
+#define PARAVIRT_BR_CLOBBERS                    \
+        "b6", "b7"
+
+#define PARAVIRT_PR_CLOBBERS                                            \
+        "p6", "p7", "p8", "p9", "p10", "p11", "p12", "p13", "p14", "p15"
+
+#define PARAVIRT_AR_CLOBBERS                    \
+        "ar.ccv"
+
+#define PARAVIRT_CLOBBERS0                      \
+                PARAVIRT_REG_CLOBBERS0,         \
+                PARAVIRT_BR_CLOBBERS,           \
+                PARAVIRT_PR_CLOBBERS,           \
+                PARAVIRT_AR_CLOBBERS,           \
+                "memory"
+
+#define PARAVIRT_CLOBBERS1                      \
+                PARAVIRT_REG_CLOBBERS1,         \
+                PARAVIRT_BR_CLOBBERS,           \
+                PARAVIRT_PR_CLOBBERS,           \
+                PARAVIRT_AR_CLOBBERS,           \
+                "memory"
+
+#define PARAVIRT_CLOBBERS2                      \
+                PARAVIRT_REG_CLOBBERS2,         \
+                PARAVIRT_BR_CLOBBERS,           \
+                PARAVIRT_PR_CLOBBERS,           \
+                PARAVIRT_AR_CLOBBERS,           \
+                "memory"
+
+#define PARAVIRT_CLOBBERS5                      \
+                PARAVIRT_REG_CLOBBERS5,         \
+                PARAVIRT_BR_CLOBBERS,           \
+                PARAVIRT_PR_CLOBBERS,           \
+                PARAVIRT_AR_CLOBBERS,           \
+                "memory"
+
+#define PARAVIRT_BR0(op, type)                                  \
+        register unsigned long ia64_clobber asm ("r8");         \
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                          PARAVIRT_TYPE(type))  \
+                      : "=r"(ia64_clobber)                      \
+                      : PARAVIRT_OP(op)                         \
+                      : PARAVIRT_CLOBBERS0)
+
+#define PARAVIRT_BR0_RET(op, type)                              \
+        register unsigned long ia64_intri_res asm ("r8");       \
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                          PARAVIRT_TYPE(type))  \
+                      : "=r"(ia64_intri_res)                    \
+                      : PARAVIRT_OP(op)                         \
+                      : PARAVIRT_CLOBBERS0)
+
+#define PARAVIRT_BR1(op, type, arg1)                            \
+        register unsigned long __##arg1 asm ("r8") = arg1;      \
+        register unsigned long ia64_clobber asm ("r8");         \
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                          PARAVIRT_TYPE(type))  \
+                      : "=r"(ia64_clobber)                      \
+                      : PARAVIRT_OP(op), "0"(__##arg1)          \
+                      : PARAVIRT_CLOBBERS1)
+
+#define PARAVIRT_BR1_RET(op, type, arg1)                        \
+        register unsigned long ia64_intri_res asm ("r8");       \
+        register unsigned long __##arg1 asm ("r8") = arg1;      \
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                          PARAVIRT_TYPE(type))  \
+                      : "=r"(ia64_intri_res)                    \
+                      : PARAVIRT_OP(op), "0"(__##arg1)          \
+                      : PARAVIRT_CLOBBERS1)
+
+#define PARAVIRT_BR1_VOID(op, type, arg1)                       \
+        register void *__##arg1 asm ("r8") = arg1;              \
+        register unsigned long ia64_clobber asm ("r8");         \
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                          PARAVIRT_TYPE(type))  \
+                      : "=r"(ia64_clobber)                      \
+                      : PARAVIRT_OP(op), "0"(__##arg1)          \
+                      : PARAVIRT_CLOBBERS1)
+
+#define PARAVIRT_BR2(op, type, arg1, arg2)                              \
+        register unsigned long __##arg1 asm ("r8") = arg1;              \
+        register unsigned long __##arg2 asm ("r9") = arg2;              \
+        register unsigned long ia64_clobber1 asm ("r8");                \
+        register unsigned long ia64_clobber2 asm ("r9");                \
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,                \
+                                          PARAVIRT_TYPE(type))          \
+                      : "=r"(ia64_clobber1), "=r"(ia64_clobber2)        \
+                      : PARAVIRT_OP(op), "0"(__##arg1), "1"(__##arg2)   \
+                      : PARAVIRT_CLOBBERS2)
+
+
+#define PARAVIRT_DEFINE_CPU_OP0(op, type)               \
+        static inline void                              \
+        paravirt_ ## op (void)                          \
+        {                                               \
+                PARAVIRT_BR0(op, type);                 \
+        }
+
+#define PARAVIRT_DEFINE_CPU_OP0_RET(op, type)           \
+        static inline unsigned long                     \
+        paravirt_ ## op (void)                          \
+        {                                               \
+                PARAVIRT_BR0_RET(op, type);             \
+                return ia64_intri_res;                  \
+        }
+
+#define PARAVIRT_DEFINE_CPU_OP1_VOID(op, type)          \
+        static inline void                              \
+        paravirt_ ## op (void *arg1)                    \
+        {                                               \
+                PARAVIRT_BR1_VOID(op, type, arg1);      \
+        }
+
+#define PARAVIRT_DEFINE_CPU_OP1(op, type)               \
+        static inline void                              \
+        paravirt_ ## op (unsigned long arg1)            \
+        {                                               \
+                PARAVIRT_BR1(op, type, arg1);           \
+        }
+
+#define PARAVIRT_DEFINE_CPU_OP1_RET(op, type)           \
+        static inline unsigned long                     \
+        paravirt_ ## op (unsigned long arg1)            \
+        {                                               \
+                PARAVIRT_BR1_RET(op, type, arg1);       \
+                return ia64_intri_res;                  \
+        }
+
+#define PARAVIRT_DEFINE_CPU_OP2(op, type)               \
+        static inline void                              \
+        paravirt_ ## op (unsigned long arg1,            \
+                         unsigned long arg2)            \
+        {                                               \
+                PARAVIRT_BR2(op, type, arg1, arg2);     \
+        }
+
+
+PARAVIRT_DEFINE_CPU_OP1_VOID(fc, FC);
+PARAVIRT_DEFINE_CPU_OP1_RET(thash, THASH)
+PARAVIRT_DEFINE_CPU_OP1_RET(get_cpuid, GET_CPUID)
+PARAVIRT_DEFINE_CPU_OP1_RET(get_pmd, GET_PMD)
+PARAVIRT_DEFINE_CPU_OP2(ptcga, PTCGA)
+PARAVIRT_DEFINE_CPU_OP1_RET(get_rr, GET_RR)
+PARAVIRT_DEFINE_CPU_OP2(set_rr, SET_RR)
+PARAVIRT_DEFINE_CPU_OP0(ssm_i, SSM_I)
+PARAVIRT_DEFINE_CPU_OP0(rsm_i, RSM_I)
+PARAVIRT_DEFINE_CPU_OP0_RET(get_psr_i, GET_PSR_I)
+PARAVIRT_DEFINE_CPU_OP1(intrin_local_irq_restore, INTRIN_LOCAL_IRQ_RESTORE)
+
+static inline void
+paravirt_set_rr0_to_rr4(unsigned long val0, unsigned long val1,
+                        unsigned long val2, unsigned long val3,
+                        unsigned long val4)
+{
+        register unsigned long __val0 asm ("r8") = val0;
+        register unsigned long __val1 asm ("r9") = val1;
+        register unsigned long __val2 asm ("r10") = val2;
+        register unsigned long __val3 asm ("r11") = val3;
+        register unsigned long __val4 asm ("r14") = val4;
+
+        register unsigned long ia64_clobber0 asm ("r8");
+        register unsigned long ia64_clobber1 asm ("r9");
+        register unsigned long ia64_clobber2 asm ("r10");
+        register unsigned long ia64_clobber3 asm ("r11");
+        register unsigned long ia64_clobber4 asm ("r14");
+
+        asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,
+                                          PARAVIRT_TYPE(SET_RR0_TO_RR4))
+                      : "=r"(ia64_clobber0),
+                        "=r"(ia64_clobber1),
+                        "=r"(ia64_clobber2),
+                        "=r"(ia64_clobber3),
+                        "=r"(ia64_clobber4)
+                      : PARAVIRT_OP(set_rr0_to_rr4),
+                        "0"(__val0), "1"(__val1), "2"(__val2),
+                        "3"(__val3), "4"(__val4)
+                      : PARAVIRT_CLOBBERS5);
+}
+
+/* unsigned long paravirt_getreg(int reg) */
+#define __paravirt_getreg(reg)                                          \
+        ({                                                              \
+                register unsigned long ia64_intri_res asm ("r8");       \
+                register unsigned long __reg asm ("r8") = (reg);        \
+                                                                        \
+                BUILD_BUG_ON(!__builtin_constant_p(reg));               \
+                asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                                  PARAVIRT_TYPE(GETREG) \
+                                                  + (reg))              \
+                              : "=r"(ia64_intri_res)                    \
+                              : PARAVIRT_OP(getreg), "0"(__reg)         \
+                              : PARAVIRT_CLOBBERS1);                    \
+                                                                        \
+                ia64_intri_res;                                         \
+        })
+
+/* void paravirt_setreg(int reg, unsigned long val) */
+#define paravirt_setreg(reg, val)                                       \
+        do {                                                            \
+                register unsigned long __val asm ("r8") = val;          \
+                register unsigned long __reg asm ("r9") = reg;          \
+                register unsigned long ia64_clobber1 asm ("r8");        \
+                register unsigned long ia64_clobber2 asm ("r9");        \
+                                                                        \
+                BUILD_BUG_ON(!__builtin_constant_p(reg));               \
+                asm volatile (paravirt_alt_bundle(__PARAVIRT_BR,        \
+                                                  PARAVIRT_TYPE(SETREG) \
+                                                  + (reg))              \
+                              : "=r"(ia64_clobber1),                    \
+                                "=r"(ia64_clobber2)                     \
+                              : PARAVIRT_OP(setreg),                    \
+                                "1"(__reg), "0"(__val)                  \
+                              : PARAVIRT_CLOBBERS2);                    \
+        } while (0)
+
+#endif /* ASM_SUPPORTED */
+#endif /* CONFIG_PARAVIRT && ASM_SUPPOTED */
+
 #endif /* _ASM_IA64_PARAVIRT_PRIVOP_H */
diff --git a/arch/ia64/include/asm/smp.h b/arch/ia64/include/asm/smp.h
index 21c402365d0e..598408336251 100644
--- a/arch/ia64/include/asm/smp.h
+++ b/arch/ia64/include/asm/smp.h
@@ -126,7 +126,8 @@ extern void identify_siblings (struct cpuinfo_ia64 *);
 extern int is_multithreading_enabled(void);
 
 extern void arch_send_call_function_single_ipi(int cpu);
-extern void arch_send_call_function_ipi(cpumask_t mask);
+extern void arch_send_call_function_ipi_mask(const struct cpumask *mask);
+#define arch_send_call_function_ipi_mask arch_send_call_function_ipi_mask
 
 #else /* CONFIG_SMP */
 
diff --git a/arch/ia64/include/asm/spinlock.h b/arch/ia64/include/asm/spinlock.h
index 0229fb95fb38..13ab71576bc7 100644
--- a/arch/ia64/include/asm/spinlock.h
+++ b/arch/ia64/include/asm/spinlock.h
@@ -120,6 +120,38 @@ do {											\
 #define __raw_read_can_lock(rw)         (*(volatile int *)(rw) >= 0)
 #define __raw_write_can_lock(rw)        (*(volatile int *)(rw) == 0)
 
+#ifdef ASM_SUPPORTED
+
+static __always_inline void
+__raw_read_lock_flags(raw_rwlock_t *lock, unsigned long flags)
+{
+        __asm__ __volatile__ (
+                "tbit.nz p6, p0 = %1,%2\n"
+                "br.few 3f\n"
+                "1:\n"
+                "fetchadd4.rel r2 = [%0], -1;;\n"
+                "(p6) ssm psr.i\n"
+                "2:\n"
+                "hint @pause\n"
+                "ld4 r2 = [%0];;\n"
+                "cmp4.lt p7,p0 = r2, r0\n"
+                "(p7) br.cond.spnt.few 2b\n"
+                "(p6) rsm psr.i\n"
+                ";;\n"
+                "3:\n"
+                "fetchadd4.acq r2 = [%0], 1;;\n"
+                "cmp4.lt p7,p0 = r2, r0\n"
+                "(p7) br.cond.spnt.few 1b\n"
+                : : "r"(lock), "r"(flags), "i"(IA64_PSR_I_BIT)
+                : "p6", "p7", "r2", "memory");
+}
+
+#define __raw_read_lock(lock) __raw_read_lock_flags(lock, 0)
+
+#else /* !ASM_SUPPORTED */
+
+#define __raw_read_lock_flags(rw, flags) __raw_read_lock(rw)
+
 #define __raw_read_lock(rw)                                                             \
 do {                                                                                    \
         raw_rwlock_t *__read_lock_ptr = (rw);                                           \
@@ -131,6 +163,8 @@ do {											\
         }                                                                               \
 } while (0)
 
+#endif /* !ASM_SUPPORTED */
+
 #define __raw_read_unlock(rw)                                   \
 do {                                                            \
         raw_rwlock_t *__read_lock_ptr = (rw);                   \
@@ -138,20 +172,33 @@ do {								\
 } while (0)
 
 #ifdef ASM_SUPPORTED
-#define __raw_write_lock(rw)                                                    \
-do {                                                                            \
-        __asm__ __volatile__ (                                                  \
-                "mov ar.ccv = r0\n"                                             \
-                "dep r29 = -1, r0, 31, 1;;\n"                                   \
-                "1:\n"                                                          \
-                "ld4 r2 = [%0];;\n"                                             \
-                "cmp4.eq p0,p7 = r0,r2\n"                                       \
-                "(p7) br.cond.spnt.few 1b \n"                                   \
-                "cmpxchg4.acq r2 = [%0], r29, ar.ccv;;\n"                       \
-                "cmp4.eq p0,p7 = r0, r2\n"                                      \
-                "(p7) br.cond.spnt.few 1b;;\n"                                  \
-                :: "r"(rw) : "ar.ccv", "p7", "r2", "r29", "memory");            \
-} while(0)
+
+static __always_inline void
+__raw_write_lock_flags(raw_rwlock_t *lock, unsigned long flags)
+{
+        __asm__ __volatile__ (
+                "tbit.nz p6, p0 = %1, %2\n"
+                "mov ar.ccv = r0\n"
+                "dep r29 = -1, r0, 31, 1\n"
+                "br.few 3f;;\n"
+                "1:\n"
+                "(p6) ssm psr.i\n"
+                "2:\n"
+                "hint @pause\n"
+                "ld4 r2 = [%0];;\n"
+                "cmp4.eq p0,p7 = r0, r2\n"
+                "(p7) br.cond.spnt.few 2b\n"
+                "(p6) rsm psr.i\n"
+                ";;\n"
+                "3:\n"
+                "cmpxchg4.acq r2 = [%0], r29, ar.ccv;;\n"
+                "cmp4.eq p0,p7 = r0, r2\n"
+                "(p7) br.cond.spnt.few 1b;;\n"
+                : : "r"(lock), "r"(flags), "i"(IA64_PSR_I_BIT)
+                : "ar.ccv", "p6", "p7", "r2", "r29", "memory");
+}
+
+#define __raw_write_lock(rw) __raw_write_lock_flags(rw, 0)
 
 #define __raw_write_trylock(rw)                                                 \
 ({                                                                              \
@@ -174,6 +221,8 @@ static inline void __raw_write_unlock(raw_rwlock_t *x)
 
 #else /* !ASM_SUPPORTED */
 
+#define __raw_write_lock_flags(l, flags) __raw_write_lock(l)
+
 #define __raw_write_lock(l)                                                             \
 ({                                                                                      \
         __u64 ia64_val, ia64_set_val = ia64_dep_mi(-1, 0, 31, 1);                       \
diff --git a/arch/ia64/include/asm/timex.h b/arch/ia64/include/asm/timex.h
index 4e03cfe74a0c..86c7db861180 100644
--- a/arch/ia64/include/asm/timex.h
+++ b/arch/ia64/include/asm/timex.h
@@ -40,5 +40,6 @@ get_cycles (void)
 }
 
 extern void ia64_cpu_local_tick (void);
+extern unsigned long long ia64_native_sched_clock (void);
 
 #endif /* _ASM_IA64_TIMEX_H */
diff --git a/arch/ia64/include/asm/topology.h b/arch/ia64/include/asm/topology.h
index f260dcf21515..7b4c8c70b2d1 100644
--- a/arch/ia64/include/asm/topology.h
+++ b/arch/ia64/include/asm/topology.h
@@ -112,11 +112,6 @@ void build_cpu_to_node_map(void);
 
 extern void arch_fix_phys_package_id(int num, u32 slot);
 
-#define pcibus_to_cpumask(bus)  (pcibus_to_node(bus) == -1 ? \
-                                        CPU_MASK_ALL : \
-                                        node_to_cpumask(pcibus_to_node(bus)) \
-                                )
-
 #define cpumask_of_pcibus(bus)  (pcibus_to_node(bus) == -1 ?            \
                                  cpu_all_mask :                         \
                                  cpumask_of_node(pcibus_to_node(bus)))
diff --git a/arch/ia64/include/asm/uv/uv_hub.h b/arch/ia64/include/asm/uv/uv_hub.h
index f607018af4a1..53e9dfacd073 100644
--- a/arch/ia64/include/asm/uv/uv_hub.h
+++ b/arch/ia64/include/asm/uv/uv_hub.h
@@ -305,5 +305,11 @@ static inline int uv_num_possible_blades(void)
         return 1;
 }
 
+static inline void uv_hub_send_ipi(int pnode, int apicid, int vector)
+{
+        /* not currently needed on ia64 */
+}
+
+
 #endif /* __ASM_IA64_UV_HUB__ */
 
diff --git a/arch/ia64/include/asm/uv/uv_mmrs.h b/arch/ia64/include/asm/uv/uv_mmrs.h
index c149ef085437..fe0b8f05e1a8 100644
--- a/arch/ia64/include/asm/uv/uv_mmrs.h
+++ b/arch/ia64/include/asm/uv/uv_mmrs.h
@@ -8,8 +8,8 @@
  * Copyright (C) 2007-2008 Silicon Graphics, Inc. All rights reserved.
  */
 
-#ifndef __ASM_IA64_UV_MMRS__
-#define __ASM_IA64_UV_MMRS__
+#ifndef _ASM_IA64_UV_UV_MMRS_H
+#define _ASM_IA64_UV_UV_MMRS_H
 
 #define UV_MMR_ENABLE           (1UL << 63)
 
@@ -243,6 +243,158 @@ union uvh_event_occurred0_u {
 #define UVH_EVENT_OCCURRED0_ALIAS_32 0x005f0
 
 /* ========================================================================= */
+/*                         UVH_GR0_TLB_INT0_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR0_TLB_INT0_CONFIG 0x61b00UL
+
+#define UVH_GR0_TLB_INT0_CONFIG_VECTOR_SHFT 0
+#define UVH_GR0_TLB_INT0_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR0_TLB_INT0_CONFIG_DM_SHFT 8
+#define UVH_GR0_TLB_INT0_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR0_TLB_INT0_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR0_TLB_INT0_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR0_TLB_INT0_CONFIG_STATUS_SHFT 12
+#define UVH_GR0_TLB_INT0_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR0_TLB_INT0_CONFIG_P_SHFT 13
+#define UVH_GR0_TLB_INT0_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR0_TLB_INT0_CONFIG_T_SHFT 15
+#define UVH_GR0_TLB_INT0_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR0_TLB_INT0_CONFIG_M_SHFT 16
+#define UVH_GR0_TLB_INT0_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR0_TLB_INT0_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR0_TLB_INT0_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr0_tlb_int0_config_u {
+    unsigned long       v;
+    struct uvh_gr0_tlb_int0_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
+/*                         UVH_GR0_TLB_INT1_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR0_TLB_INT1_CONFIG 0x61b40UL
+
+#define UVH_GR0_TLB_INT1_CONFIG_VECTOR_SHFT 0
+#define UVH_GR0_TLB_INT1_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR0_TLB_INT1_CONFIG_DM_SHFT 8
+#define UVH_GR0_TLB_INT1_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR0_TLB_INT1_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR0_TLB_INT1_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR0_TLB_INT1_CONFIG_STATUS_SHFT 12
+#define UVH_GR0_TLB_INT1_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR0_TLB_INT1_CONFIG_P_SHFT 13
+#define UVH_GR0_TLB_INT1_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR0_TLB_INT1_CONFIG_T_SHFT 15
+#define UVH_GR0_TLB_INT1_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR0_TLB_INT1_CONFIG_M_SHFT 16
+#define UVH_GR0_TLB_INT1_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR0_TLB_INT1_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR0_TLB_INT1_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr0_tlb_int1_config_u {
+    unsigned long       v;
+    struct uvh_gr0_tlb_int1_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
+/*                         UVH_GR1_TLB_INT0_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR1_TLB_INT0_CONFIG 0x61f00UL
+
+#define UVH_GR1_TLB_INT0_CONFIG_VECTOR_SHFT 0
+#define UVH_GR1_TLB_INT0_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR1_TLB_INT0_CONFIG_DM_SHFT 8
+#define UVH_GR1_TLB_INT0_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR1_TLB_INT0_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR1_TLB_INT0_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR1_TLB_INT0_CONFIG_STATUS_SHFT 12
+#define UVH_GR1_TLB_INT0_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR1_TLB_INT0_CONFIG_P_SHFT 13
+#define UVH_GR1_TLB_INT0_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR1_TLB_INT0_CONFIG_T_SHFT 15
+#define UVH_GR1_TLB_INT0_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR1_TLB_INT0_CONFIG_M_SHFT 16
+#define UVH_GR1_TLB_INT0_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR1_TLB_INT0_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR1_TLB_INT0_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr1_tlb_int0_config_u {
+    unsigned long       v;
+    struct uvh_gr1_tlb_int0_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
+/*                         UVH_GR1_TLB_INT1_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR1_TLB_INT1_CONFIG 0x61f40UL
+
+#define UVH_GR1_TLB_INT1_CONFIG_VECTOR_SHFT 0
+#define UVH_GR1_TLB_INT1_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR1_TLB_INT1_CONFIG_DM_SHFT 8
+#define UVH_GR1_TLB_INT1_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR1_TLB_INT1_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR1_TLB_INT1_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR1_TLB_INT1_CONFIG_STATUS_SHFT 12
+#define UVH_GR1_TLB_INT1_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR1_TLB_INT1_CONFIG_P_SHFT 13
+#define UVH_GR1_TLB_INT1_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR1_TLB_INT1_CONFIG_T_SHFT 15
+#define UVH_GR1_TLB_INT1_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR1_TLB_INT1_CONFIG_M_SHFT 16
+#define UVH_GR1_TLB_INT1_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR1_TLB_INT1_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR1_TLB_INT1_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr1_tlb_int1_config_u {
+    unsigned long       v;
+    struct uvh_gr1_tlb_int1_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
 /*                               UVH_INT_CMPB                                */
 /* ========================================================================= */
 #define UVH_INT_CMPB 0x22080UL
@@ -670,4 +822,4 @@ union uvh_si_alias2_overlay_config_u {
 };
 
 
-#endif /* __ASM_IA64_UV_MMRS__ */
+#endif /* _ASM_IA64_UV_UV_MMRS_H */
diff --git a/arch/ia64/include/asm/xen/hypervisor.h b/arch/ia64/include/asm/xen/hypervisor.h
index 7a804e80fc67..e425227a418e 100644
--- a/arch/ia64/include/asm/xen/hypervisor.h
+++ b/arch/ia64/include/asm/xen/hypervisor.h
@@ -33,9 +33,6 @@
 #ifndef _ASM_IA64_XEN_HYPERVISOR_H
 #define _ASM_IA64_XEN_HYPERVISOR_H
 
-#ifdef CONFIG_XEN
-
-#include <linux/init.h>
 #include <xen/interface/xen.h>
 #include <xen/interface/version.h>      /* to compile feature.c */
 #include <xen/features.h>               /* to comiple xen-netfront.c */
@@ -43,22 +40,32 @@
 
 /* xen_domain_type is set before executing any C code by early_xen_setup */
 enum xen_domain_type {
-        XEN_NATIVE,
-        XEN_PV_DOMAIN,
-        XEN_HVM_DOMAIN,
+        XEN_NATIVE,     /* running on bare hardware */
+        XEN_PV_DOMAIN,  /* running in a PV domain */
+        XEN_HVM_DOMAIN, /* running in a Xen hvm domain*/
 };
 
+#ifdef CONFIG_XEN
 extern enum xen_domain_type xen_domain_type;
+#else
+#define xen_domain_type         XEN_NATIVE
+#endif
 
 #define xen_domain()            (xen_domain_type != XEN_NATIVE)
-#define xen_pv_domain()         (xen_domain_type == XEN_PV_DOMAIN)
-#define xen_initial_domain()    (xen_pv_domain() && \
+#define xen_pv_domain()         (xen_domain() &&                        \
+                                 xen_domain_type == XEN_PV_DOMAIN)
+#define xen_hvm_domain()        (xen_domain() &&                        \
+                                 xen_domain_type == XEN_HVM_DOMAIN)
+
+#ifdef CONFIG_XEN_DOM0
+#define xen_initial_domain()    (xen_pv_domain() &&                     \
                                  (xen_start_info->flags & SIF_INITDOMAIN))
-#define xen_hvm_domain()        (xen_domain_type == XEN_HVM_DOMAIN)
+#else
+#define xen_initial_domain()    (0)
+#endif
 
-/* deprecated. remove this */
-#define is_running_on_xen()     (xen_domain_type == XEN_PV_DOMAIN)
 
+#ifdef CONFIG_XEN
 extern struct shared_info *HYPERVISOR_shared_info;
 extern struct start_info *xen_start_info;
 
@@ -74,16 +81,6 @@ void force_evtchn_callback(void);
 
 /* For setup_arch() in arch/ia64/kernel/setup.c */
 void xen_ia64_enable_opt_feature(void);
-
-#else /* CONFIG_XEN */
-
-#define xen_domain()            (0)
-#define xen_pv_domain()         (0)
-#define xen_initial_domain()    (0)
-#define xen_hvm_domain()        (0)
-#define is_running_on_xen()     (0)     /* deprecated. remove this */
 #endif
 
-#define is_initial_xendomain()  (0)     /* deprecated. remove this */
-
 #endif /* _ASM_IA64_XEN_HYPERVISOR_H */
diff --git a/arch/ia64/include/asm/xen/inst.h b/arch/ia64/include/asm/xen/inst.h
index 19c2ae1d878a..c53a47611208 100644
--- a/arch/ia64/include/asm/xen/inst.h
+++ b/arch/ia64/include/asm/xen/inst.h
@@ -33,6 +33,9 @@
 #define __paravirt_work_processed_syscall_target \
                                                 xen_work_processed_syscall
 
+#define paravirt_fsyscall_table                 xen_fsyscall_table
+#define paravirt_fsys_bubble_down               xen_fsys_bubble_down
+
 #define MOV_FROM_IFA(reg)       \
         movl reg = XSI_IFA;     \
         ;;                      \
@@ -110,6 +113,27 @@
 .endm
 #define MOV_FROM_PSR(pred, reg, clob)   __MOV_FROM_PSR pred, reg, clob
 
+/* assuming ar.itc is read with interrupt disabled. */
+#define MOV_FROM_ITC(pred, pred_clob, reg, clob)                \
+(pred)  movl clob = XSI_ITC_OFFSET;                             \
+        ;;                                                      \
+(pred)  ld8 clob = [clob];                                      \
+(pred)  mov reg = ar.itc;                                       \
+        ;;                                                      \
+(pred)  add reg = reg, clob;                                    \
+        ;;                                                      \
+(pred)  movl clob = XSI_ITC_LAST;                               \
+        ;;                                                      \
+(pred)  ld8 clob = [clob];                                      \
+        ;;                                                      \
+(pred)  cmp.geu.unc pred_clob, p0 = clob, reg;                  \
+        ;;                                                      \
+(pred_clob)     add reg = 1, clob;                              \
+        ;;                                                      \
+(pred)  movl clob = XSI_ITC_LAST;                               \
+        ;;                                                      \
+(pred)  st8 [clob] = reg
+
 
 #define MOV_TO_IFA(reg, clob)   \
         movl clob = XSI_IFA;    \
@@ -362,6 +386,10 @@
 #define RSM_PSR_DT              \
         XEN_HYPER_RSM_PSR_DT
 
+#define RSM_PSR_BE_I(clob0, clob1)      \
+        RSM_PSR_I(p0, clob0, clob1);    \
+        rum psr.be
+
 #define SSM_PSR_DT_AND_SRLZ_I   \
         XEN_HYPER_SSM_PSR_DT
 
diff --git a/arch/ia64/include/asm/xen/interface.h b/arch/ia64/include/asm/xen/interface.h
index f00fab40854d..e951e740bdf2 100644
--- a/arch/ia64/include/asm/xen/interface.h
+++ b/arch/ia64/include/asm/xen/interface.h
@@ -209,6 +209,15 @@ struct mapped_regs {
                         unsigned long krs[8];   /* kernel registers */
                         unsigned long tmp[16];  /* temp registers
                                                    (e.g. for hyperprivops) */
+
+                        /* itc paravirtualization
+                         * vAR.ITC = mAR.ITC + itc_offset
+                         * itc_last is one which was lastly passed to
+                         * the guest OS in order to prevent it from
+                         * going backwords.
+                         */
+                        unsigned long itc_offset;
+                        unsigned long itc_last;
                 };
         };
 };
diff --git a/arch/ia64/include/asm/xen/minstate.h b/arch/ia64/include/asm/xen/minstate.h
index 4d92d9bbda7b..c57fa910f2c9 100644
--- a/arch/ia64/include/asm/xen/minstate.h
+++ b/arch/ia64/include/asm/xen/minstate.h
@@ -1,3 +1,12 @@
+
+#ifdef CONFIG_VIRT_CPU_ACCOUNTING
+/* read ar.itc in advance, and use it before leaving bank 0 */
+#define XEN_ACCOUNT_GET_STAMP           \
+        MOV_FROM_ITC(pUStk, p6, r20, r2);
+#else
+#define XEN_ACCOUNT_GET_STAMP
+#endif
+
 /*
  * DO_SAVE_MIN switches to the kernel stacks (if necessary) and saves
  * the minimum state necessary that allows us to turn psr.ic back
@@ -123,7 +132,7 @@
         ;;                                                                                      \
 .mem.offset 0,0; st8.spill [r16]=r2,16;                                                         \
 .mem.offset 8,0; st8.spill [r17]=r3,16;                                                         \
-        ACCOUNT_GET_STAMP                                                                       \
+        XEN_ACCOUNT_GET_STAMP                                                                   \
         adds r2=IA64_PT_REGS_R16_OFFSET,r1;                                                     \
         ;;                                                                                      \
         EXTRA;                                                                                  \
diff --git a/arch/ia64/include/asm/xen/patchlist.h b/arch/ia64/include/asm/xen/patchlist.h
new file mode 100644
index 000000000000..eae944e88846
--- /dev/null
+++ b/arch/ia64/include/asm/xen/patchlist.h
@@ -0,0 +1,38 @@
+/******************************************************************************
+ * arch/ia64/include/asm/xen/patchlist.h
+ *
+ * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#define __paravirt_start_gate_fsyscall_patchlist                \
+        __xen_start_gate_fsyscall_patchlist
+#define __paravirt_end_gate_fsyscall_patchlist                  \
+        __xen_end_gate_fsyscall_patchlist
+#define __paravirt_start_gate_brl_fsys_bubble_down_patchlist    \
+        __xen_start_gate_brl_fsys_bubble_down_patchlist
+#define __paravirt_end_gate_brl_fsys_bubble_down_patchlist      \
+        __xen_end_gate_brl_fsys_bubble_down_patchlist
+#define __paravirt_start_gate_vtop_patchlist                    \
+        __xen_start_gate_vtop_patchlist
+#define __paravirt_end_gate_vtop_patchlist                      \
+        __xen_end_gate_vtop_patchlist
+#define __paravirt_start_gate_mckinley_e9_patchlist             \
+        __xen_start_gate_mckinley_e9_patchlist
+#define __paravirt_end_gate_mckinley_e9_patchlist               \
+        __xen_end_gate_mckinley_e9_patchlist
diff --git a/arch/ia64/include/asm/xen/privop.h b/arch/ia64/include/asm/xen/privop.h
index 71ec7546e100..fb4ec5e0b066 100644
--- a/arch/ia64/include/asm/xen/privop.h
+++ b/arch/ia64/include/asm/xen/privop.h
@@ -55,6 +55,8 @@
 #define XSI_BANK1_R16                   (XSI_BASE + XSI_BANK1_R16_OFS)
 #define XSI_BANKNUM                     (XSI_BASE + XSI_BANKNUM_OFS)
 #define XSI_IHA                         (XSI_BASE + XSI_IHA_OFS)
+#define XSI_ITC_OFFSET                  (XSI_BASE + XSI_ITC_OFFSET_OFS)
+#define XSI_ITC_LAST                    (XSI_BASE + XSI_ITC_LAST_OFS)
 #endif
 
 #ifndef __ASSEMBLY__
@@ -67,7 +69,7 @@
  *  may have different semantics depending on whether they are executed
  *  at PL0 vs PL!=0.  When paravirtualized, these instructions mustn't
  *  be allowed to execute directly, lest incorrect semantics result. */
-extern void xen_fc(unsigned long addr);
+extern void xen_fc(void *addr);
 extern unsigned long xen_thash(unsigned long addr);
 
 /* Note that "ttag" and "cover" are also privilege-sensitive; "ttag"
@@ -80,8 +82,10 @@ extern unsigned long xen_thash(unsigned long addr);
 extern unsigned long xen_get_cpuid(int index);
 extern unsigned long xen_get_pmd(int index);
 
+#ifndef ASM_SUPPORTED
 extern unsigned long xen_get_eflag(void);       /* see xen_ia64_getreg */
 extern void xen_set_eflag(unsigned long);       /* see xen_ia64_setreg */
+#endif
 
 /************************************************/
 /* Instructions paravirtualized for performance */
@@ -106,6 +110,7 @@ extern void xen_set_eflag(unsigned long);	/* see xen_ia64_setreg */
 #define xen_get_virtual_pend()          \
         (*(((uint8_t *)XEN_MAPPEDREGS->interrupt_mask_addr) - 1))
 
+#ifndef ASM_SUPPORTED
 /* Although all privileged operations can be left to trap and will
  * be properly handled by Xen, some are frequent enough that we use
  * hyperprivops for performance. */
@@ -123,6 +128,7 @@ extern void xen_set_rr0_to_rr4(unsigned long val0, unsigned long val1,
                                unsigned long val4);
 extern void xen_set_kr(unsigned long index, unsigned long val);
 extern void xen_ptcga(unsigned long addr, unsigned long size);
+#endif /* !ASM_SUPPORTED */
 
 #endif /* !__ASSEMBLY__ */
 
diff --git a/arch/ia64/kernel/Makefile b/arch/ia64/kernel/Makefile
index f2778f2c4fd9..5628e9a990a6 100644
--- a/arch/ia64/kernel/Makefile
+++ b/arch/ia64/kernel/Makefile
@@ -5,7 +5,7 @@
 extra-y := head.o init_task.o vmlinux.lds
 
 obj-y := acpi.o entry.o efi.o efi_stub.o gate-data.o fsys.o ia64_ksyms.o irq.o irq_ia64.o       \
-         irq_lsapic.o ivt.o machvec.o pal.o patch.o process.o perfmon.o ptrace.o sal.o          \
+         irq_lsapic.o ivt.o machvec.o pal.o paravirt_patchlist.o patch.o process.o perfmon.o ptrace.o sal.o             \
          salinfo.o setup.o signal.o sys_ia64.o time.o traps.o unaligned.o \
          unwind.o mca.o mca_asm.o topology.o dma-mapping.o
 
@@ -36,7 +36,8 @@ obj-$(CONFIG_PCI_MSI)		+= msi_ia64.o
 mca_recovery-y                  += mca_drv.o mca_drv_asm.o
 obj-$(CONFIG_IA64_MC_ERR_INJECT)+= err_inject.o
 
-obj-$(CONFIG_PARAVIRT)          += paravirt.o paravirtentry.o
+obj-$(CONFIG_PARAVIRT)          += paravirt.o paravirtentry.o \
+                                   paravirt_patch.o
 
 obj-$(CONFIG_IA64_ESI)          += esi.o
 ifneq ($(CONFIG_IA64_ESI),)
@@ -45,35 +46,13 @@ endif
 obj-$(CONFIG_DMAR)              += pci-dma.o
 obj-$(CONFIG_SWIOTLB)           += pci-swiotlb.o
 
-# The gate DSO image is built using a special linker script.
-targets += gate.so gate-syms.o
-
-extra-y += gate.so gate-syms.o gate.lds gate.o
-
 # fp_emulate() expects f2-f5,f16-f31 to contain the user-level state.
 CFLAGS_traps.o  += -mfixed-range=f2-f5,f16-f31
 
-CPPFLAGS_gate.lds := -P -C -U$(ARCH)
-
-quiet_cmd_gate = GATE $@
-      cmd_gate = $(CC) -nostdlib $(GATECFLAGS_$(@F)) -Wl,-T,$(filter-out FORCE,$^) -o $@
-
-GATECFLAGS_gate.so = -shared -s -Wl,-soname=linux-gate.so.1 \
-                     $(call ld-option, -Wl$(comma)--hash-style=sysv)
-$(obj)/gate.so: $(obj)/gate.lds $(obj)/gate.o FORCE
-        $(call if_changed,gate)
-
-$(obj)/built-in.o: $(obj)/gate-syms.o
-$(obj)/built-in.o: ld_flags += -R $(obj)/gate-syms.o
-
-GATECFLAGS_gate-syms.o = -r
-$(obj)/gate-syms.o: $(obj)/gate.lds $(obj)/gate.o FORCE
-        $(call if_changed,gate)
-
-# gate-data.o contains the gate DSO image as data in section .data.gate.
-# We must build gate.so before we can assemble it.
-# Note: kbuild does not track this dependency due to usage of .incbin
-$(obj)/gate-data.o: $(obj)/gate.so
+# The gate DSO image is built using a special linker script.
+include $(srctree)/arch/ia64/kernel/Makefile.gate
+# tell compiled for native
+CPPFLAGS_gate.lds += -D__IA64_GATE_PARAVIRTUALIZED_NATIVE
 
 # Calculate NR_IRQ = max(IA64_NATIVE_NR_IRQS, XEN_NR_IRQS, ...) based on config
 define sed-y
@@ -109,9 +88,9 @@ include/asm-ia64/nr-irqs.h: arch/$(SRCARCH)/kernel/nr-irqs.s
 clean-files += $(objtree)/include/asm-ia64/nr-irqs.h
 
 #
-# native ivt.S and entry.S
+# native ivt.S, entry.S and fsys.S
 #
-ASM_PARAVIRT_OBJS = ivt.o entry.o
+ASM_PARAVIRT_OBJS = ivt.o entry.o fsys.o
 define paravirtualized_native
 AFLAGS_$(1) += -D__IA64_ASM_PARAVIRTUALIZED_NATIVE
 AFLAGS_pvchk-sed-$(1) += -D__IA64_ASM_PARAVIRTUALIZED_PVCHECK
diff --git a/arch/ia64/kernel/Makefile.gate b/arch/ia64/kernel/Makefile.gate
new file mode 100644
index 000000000000..1d87f84069b3
--- /dev/null
+++ b/arch/ia64/kernel/Makefile.gate
@@ -0,0 +1,27 @@
+# The gate DSO image is built using a special linker script.
+
+targets += gate.so gate-syms.o
+
+extra-y += gate.so gate-syms.o gate.lds gate.o
+
+CPPFLAGS_gate.lds := -P -C -U$(ARCH)
+
+quiet_cmd_gate = GATE $@
+      cmd_gate = $(CC) -nostdlib $(GATECFLAGS_$(@F)) -Wl,-T,$(filter-out FORCE,$^) -o $@
+
+GATECFLAGS_gate.so = -shared -s -Wl,-soname=linux-gate.so.1 \
+                     $(call ld-option, -Wl$(comma)--hash-style=sysv)
+$(obj)/gate.so: $(obj)/gate.lds $(obj)/gate.o FORCE
+        $(call if_changed,gate)
+
+$(obj)/built-in.o: $(obj)/gate-syms.o
+$(obj)/built-in.o: ld_flags += -R $(obj)/gate-syms.o
+
+GATECFLAGS_gate-syms.o = -r
+$(obj)/gate-syms.o: $(obj)/gate.lds $(obj)/gate.o FORCE
+        $(call if_changed,gate)
+
+# gate-data.o contains the gate DSO image as data in section .data.gate.
+# We must build gate.so before we can assemble it.
+# Note: kbuild does not track this dependency due to usage of .incbin
+$(obj)/gate-data.o: $(obj)/gate.so
diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c
index bdef2ce38c8b..5510317db37b 100644
--- a/arch/ia64/kernel/acpi.c
+++ b/arch/ia64/kernel/acpi.c
@@ -890,7 +890,7 @@ __init void prefill_possible_map(void)
                 possible, max((possible - available_cpus), 0));
 
         for (i = 0; i < possible; i++)
-                cpu_set(i, cpu_possible_map);
+                set_cpu_possible(i, true);
 }
 
 int acpi_map_lsapic(acpi_handle handle, int *pcpu)
@@ -928,9 +928,9 @@ int acpi_map_lsapic(acpi_handle handle, int *pcpu)
         buffer.length = ACPI_ALLOCATE_BUFFER;
         buffer.pointer = NULL;
 
-        cpus_complement(tmp_map, cpu_present_map);
-        cpu = first_cpu(tmp_map);
-        if (cpu >= NR_CPUS)
+        cpumask_complement(&tmp_map, cpu_present_mask);
+        cpu = cpumask_first(&tmp_map);
+        if (cpu >= nr_cpu_ids)
                 return -EINVAL;
 
         acpi_map_cpu2node(handle, cpu, physid);
diff --git a/arch/ia64/kernel/asm-offsets.c b/arch/ia64/kernel/asm-offsets.c
index 742dbb1d5a4f..af5650169043 100644
--- a/arch/ia64/kernel/asm-offsets.c
+++ b/arch/ia64/kernel/asm-offsets.c
@@ -316,5 +316,7 @@ void foo(void)
         DEFINE_MAPPED_REG_OFS(XSI_BANK1_R16_OFS, bank1_regs[0]);
         DEFINE_MAPPED_REG_OFS(XSI_B0NATS_OFS, vbnat);
         DEFINE_MAPPED_REG_OFS(XSI_B1NATS_OFS, vnat);
+        DEFINE_MAPPED_REG_OFS(XSI_ITC_OFFSET_OFS, itc_offset);
+        DEFINE_MAPPED_REG_OFS(XSI_ITC_LAST_OFS, itc_last);
 #endif /* CONFIG_XEN */
 }
diff --git a/arch/ia64/kernel/efi.c b/arch/ia64/kernel/efi.c
index efaff15d8cf1..7ef80e8161ce 100644
--- a/arch/ia64/kernel/efi.c
+++ b/arch/ia64/kernel/efi.c
@@ -456,6 +456,7 @@ efi_map_pal_code (void)
                  GRANULEROUNDDOWN((unsigned long) pal_vaddr),
                  pte_val(pfn_pte(__pa(pal_vaddr) >> PAGE_SHIFT, PAGE_KERNEL)),
                  IA64_GRANULE_SHIFT);
+        paravirt_dv_serialize_data();
         ia64_set_psr(psr);              /* restore psr */
 }
 
diff --git a/arch/ia64/kernel/entry.S b/arch/ia64/kernel/entry.S
index e5341e2c1175..ccfdeee9d89f 100644
--- a/arch/ia64/kernel/entry.S
+++ b/arch/ia64/kernel/entry.S
@@ -735,7 +735,7 @@ GLOBAL_ENTRY(__paravirt_leave_syscall)
 __paravirt_work_processed_syscall:
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
         adds r2=PT(LOADRS)+16,r12
-(pUStk) mov.m r22=ar.itc                        // fetch time at leave
+        MOV_FROM_ITC(pUStk, p9, r22, r19)       // fetch time at leave
         adds r18=TI_FLAGS+IA64_TASK_SIZE,r13
         ;;
 (p6)    ld4 r31=[r18]                           // load current_thread_info()->flags
@@ -984,7 +984,7 @@ GLOBAL_ENTRY(__paravirt_leave_kernel)
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
         .pred.rel.mutex pUStk,pKStk
         MOV_FROM_PSR(pKStk, r22, r29)   // M2 read PSR now that interrupts are disabled
-(pUStk) mov.m r22=ar.itc        // M  fetch time at leave
+        MOV_FROM_ITC(pUStk, p9, r22, r29)       // M  fetch time at leave
         nop.i 0
         ;;
 #else
diff --git a/arch/ia64/kernel/fsys.S b/arch/ia64/kernel/fsys.S
index c1625c7e1779..3567d54f8cee 100644
--- a/arch/ia64/kernel/fsys.S
+++ b/arch/ia64/kernel/fsys.S
@@ -25,6 +25,7 @@
 #include <asm/unistd.h>
 
 #include "entry.h"
+#include "paravirt_inst.h"
 
 /*
  * See Documentation/ia64/fsys.txt for details on fsyscalls.
@@ -279,7 +280,7 @@ ENTRY(fsys_gettimeofday)
 (p9)    cmp.eq p13,p0 = 0,r30   // if mmio_ptr, clear p13 jitter control
         ;;
         .pred.rel.mutex p8,p9
-(p8)    mov r2 = ar.itc         // CPU_TIMER. 36 clocks latency!!!
+        MOV_FROM_ITC(p8, p6, r2, r10)   // CPU_TIMER. 36 clocks latency!!!
 (p9)    ld8 r2 = [r30]          // MMIO_TIMER. Could also have latency issues..
 (p13)   ld8 r25 = [r19]         // get itc_lastcycle value
         ld8 r9 = [r22],IA64_TIMESPEC_TV_NSEC_OFFSET     // tv_sec
@@ -418,7 +419,7 @@ EX(.fail_efault, ld8 r14=[r33])			// r14 <- *set
         mov r17=(1 << (SIGKILL - 1)) | (1 << (SIGSTOP - 1))
         ;;
 
-        rsm psr.i                               // mask interrupt delivery
+        RSM_PSR_I(p0, r18, r19)                 // mask interrupt delivery
         mov ar.ccv=0
         andcm r14=r14,r17                       // filter out SIGKILL & SIGSTOP
 
@@ -491,7 +492,7 @@ EX(.fail_efault, ld8 r14=[r33])			// r14 <- *set
 #ifdef CONFIG_SMP
         st4.rel [r31]=r0                        // release the lock
 #endif
-        ssm psr.i
+        SSM_PSR_I(p0, p9, r31)
         ;;
 
         srlz.d                                  // ensure psr.i is set again
@@ -513,7 +514,7 @@ EX(.fail_efault, (p15) st8 [r34]=r3)
 #ifdef CONFIG_SMP
         st4.rel [r31]=r0                        // release the lock
 #endif
-        ssm psr.i
+        SSM_PSR_I(p0, p9, r17)
         ;;
         srlz.d
         br.sptk.many fsys_fallback_syscall      // with signal pending, do the heavy-weight syscall
@@ -521,7 +522,7 @@ EX(.fail_efault, (p15) st8 [r34]=r3)
 #ifdef CONFIG_SMP
 .lock_contention:
         /* Rather than spinning here, fall back on doing a heavy-weight syscall.  */
-        ssm psr.i
+        SSM_PSR_I(p0, p9, r17)
         ;;
         srlz.d
         br.sptk.many fsys_fallback_syscall
@@ -592,17 +593,17 @@ ENTRY(fsys_fallback_syscall)
         adds r17=-1024,r15
         movl r14=sys_call_table
         ;;
-        rsm psr.i
+        RSM_PSR_I(p0, r26, r27)
         shladd r18=r17,3,r14
         ;;
         ld8 r18=[r18]                           // load normal (heavy-weight) syscall entry-point
-        mov r29=psr                             // read psr (12 cyc load latency)
+        MOV_FROM_PSR(p0, r29, r26)              // read psr (12 cyc load latency)
         mov r27=ar.rsc
         mov r21=ar.fpsr
         mov r26=ar.pfs
 END(fsys_fallback_syscall)
         /* FALL THROUGH */
-GLOBAL_ENTRY(fsys_bubble_down)
+GLOBAL_ENTRY(paravirt_fsys_bubble_down)
         .prologue
         .altrp b6
         .body
@@ -640,7 +641,7 @@ GLOBAL_ENTRY(fsys_bubble_down)
          *
          * PSR.BE : already is turned off in __kernel_syscall_via_epc()
          * PSR.AC : don't care (kernel normally turns PSR.AC on)
-         * PSR.I  : already turned off by the time fsys_bubble_down gets
+         * PSR.I  : already turned off by the time paravirt_fsys_bubble_down gets
          *          invoked
          * PSR.DFL: always 0 (kernel never turns it on)
          * PSR.DFH: don't care --- kernel never touches f32-f127 on its own
@@ -650,7 +651,7 @@ GLOBAL_ENTRY(fsys_bubble_down)
          * PSR.DB : don't care --- kernel never enables kernel-level
          *          breakpoints
          * PSR.TB : must be 0 already; if it wasn't zero on entry to
-         *          __kernel_syscall_via_epc, the branch to fsys_bubble_down
+         *          __kernel_syscall_via_epc, the branch to paravirt_fsys_bubble_down
          *          will trigger a taken branch; the taken-trap-handler then
          *          converts the syscall into a break-based system-call.
          */
@@ -683,7 +684,7 @@ GLOBAL_ENTRY(fsys_bubble_down)
         ;;
         mov ar.rsc=0                            // M2   set enforced lazy mode, pl 0, LE, loadrs=0
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
-        mov.m r30=ar.itc                        // M    get cycle for accounting
+        MOV_FROM_ITC(p0, p6, r30, r23)          // M    get cycle for accounting
 #else
         nop.m 0
 #endif
@@ -734,21 +735,21 @@ GLOBAL_ENTRY(fsys_bubble_down)
         mov rp=r14                              // I0   set the real return addr
         and r3=_TIF_SYSCALL_TRACEAUDIT,r3       // A
         ;;
-        ssm psr.i                               // M2   we're on kernel stacks now, reenable irqs
+        SSM_PSR_I(p0, p6, r22)                  // M2   we're on kernel stacks now, reenable irqs
         cmp.eq p8,p0=r3,r0                      // A
 (p10)   br.cond.spnt.many ia64_ret_from_syscall // B    return if bad call-frame or r15 is a NaT
 
         nop.m 0
 (p8)    br.call.sptk.many b6=b6                 // B    (ignore return address)
         br.cond.spnt ia64_trace_syscall         // B
-END(fsys_bubble_down)
+END(paravirt_fsys_bubble_down)
 
         .rodata
         .align 8
-        .globl fsyscall_table
+        .globl paravirt_fsyscall_table
 
-        data8 fsys_bubble_down
-fsyscall_table:
+        data8 paravirt_fsys_bubble_down
+paravirt_fsyscall_table:
         data8 fsys_ni_syscall
         data8 0                         // exit                 // 1025
         data8 0                         // read
@@ -1033,4 +1034,4 @@ fsyscall_table:
 
         // fill in zeros for the remaining entries
         .zero:
-        .space fsyscall_table + 8*NR_syscalls - .zero, 0
+        .space paravirt_fsyscall_table + 8*NR_syscalls - .zero, 0
diff --git a/arch/ia64/kernel/gate.S b/arch/ia64/kernel/gate.S
index 74b1ccce4e84..cf5e0a105e16 100644
--- a/arch/ia64/kernel/gate.S
+++ b/arch/ia64/kernel/gate.S
@@ -13,6 +13,7 @@
 #include <asm/sigcontext.h>
 #include <asm/system.h>
 #include <asm/unistd.h>
+#include "paravirt_inst.h"
 
 /*
  * We can't easily refer to symbols inside the kernel.  To avoid full runtime relocation,
@@ -48,87 +49,6 @@ GLOBAL_ENTRY(__kernel_syscall_via_break)
 }
 END(__kernel_syscall_via_break)
 
-/*
- * On entry:
- *      r11 = saved ar.pfs
- *      r15 = system call #
- *      b0  = saved return address
- *      b6  = return address
- * On exit:
- *      r11 = saved ar.pfs
- *      r15 = system call #
- *      b0  = saved return address
- *      all other "scratch" registers:  undefined
- *      all "preserved" registers:      same as on entry
- */
-
-GLOBAL_ENTRY(__kernel_syscall_via_epc)
-        .prologue
-        .altrp b6
-        .body
-{
-        /*
-         * Note: the kernel cannot assume that the first two instructions in this
-         * bundle get executed.  The remaining code must be safe even if
-         * they do not get executed.
-         */
-        adds r17=-1024,r15                      // A
-        mov r10=0                               // A    default to successful syscall execution
-        epc                                     // B    causes split-issue
-}
-        ;;
-        rsm psr.be | psr.i                      // M2 (5 cyc to srlz.d)
-        LOAD_FSYSCALL_TABLE(r14)                // X
-        ;;
-        mov r16=IA64_KR(CURRENT)                // M2 (12 cyc)
-        shladd r18=r17,3,r14                    // A
-        mov r19=NR_syscalls-1                   // A
-        ;;
-        lfetch [r18]                            // M0|1
-        mov r29=psr                             // M2 (12 cyc)
-        // If r17 is a NaT, p6 will be zero
-        cmp.geu p6,p7=r19,r17                   // A    (sysnr > 0 && sysnr < 1024+NR_syscalls)?
-        ;;
-        mov r21=ar.fpsr                         // M2 (12 cyc)
-        tnat.nz p10,p9=r15                      // I0
-        mov.i r26=ar.pfs                        // I0 (would stall anyhow due to srlz.d...)
-        ;;
-        srlz.d                                  // M0 (forces split-issue) ensure PSR.BE==0
-(p6)    ld8 r18=[r18]                           // M0|1
-        nop.i 0
-        ;;
-        nop.m 0
-(p6)    tbit.z.unc p8,p0=r18,0                  // I0 (dual-issues with "mov b7=r18"!)
-        nop.i 0
-        ;;
-(p8)    ssm psr.i
-(p6)    mov b7=r18                              // I0
-(p8)    br.dptk.many b7                         // B
-
-        mov r27=ar.rsc                          // M2 (12 cyc)
-/*
- * brl.cond doesn't work as intended because the linker would convert this branch
- * into a branch to a PLT.  Perhaps there will be a way to avoid this with some
- * future version of the linker.  In the meantime, we just use an indirect branch
- * instead.
- */
-#ifdef CONFIG_ITANIUM
-(p6)    add r14=-8,r14                          // r14 <- addr of fsys_bubble_down entry
-        ;;
-(p6)    ld8 r14=[r14]                           // r14 <- fsys_bubble_down
-        ;;
-(p6)    mov b7=r14
-(p6)    br.sptk.many b7
-#else
-        BRL_COND_FSYS_BUBBLE_DOWN(p6)
-#endif
-        ssm psr.i
-        mov r10=-1
-(p10)   mov r8=EINVAL
-(p9)    mov r8=ENOSYS
-        FSYS_RETURN
-END(__kernel_syscall_via_epc)
-
 #       define ARG0_OFF         (16 + IA64_SIGFRAME_ARG0_OFFSET)
 #       define ARG1_OFF         (16 + IA64_SIGFRAME_ARG1_OFFSET)
 #       define ARG2_OFF         (16 + IA64_SIGFRAME_ARG2_OFFSET)
@@ -374,3 +294,92 @@ restore_rbs:
         // invala not necessary as that will happen when returning to user-mode
         br.cond.sptk back_from_restore_rbs
 END(__kernel_sigtramp)
+
+/*
+ * On entry:
+ *      r11 = saved ar.pfs
+ *      r15 = system call #
+ *      b0  = saved return address
+ *      b6  = return address
+ * On exit:
+ *      r11 = saved ar.pfs
+ *      r15 = system call #
+ *      b0  = saved return address
+ *      all other "scratch" registers:  undefined
+ *      all "preserved" registers:      same as on entry
+ */
+
+GLOBAL_ENTRY(__kernel_syscall_via_epc)
+        .prologue
+        .altrp b6
+        .body
+{
+        /*
+         * Note: the kernel cannot assume that the first two instructions in this
+         * bundle get executed.  The remaining code must be safe even if
+         * they do not get executed.
+         */
+        adds r17=-1024,r15                      // A
+        mov r10=0                               // A    default to successful syscall execution
+        epc                                     // B    causes split-issue
+}
+        ;;
+        RSM_PSR_BE_I(r20, r22)                  // M2 (5 cyc to srlz.d)
+        LOAD_FSYSCALL_TABLE(r14)                // X
+        ;;
+        mov r16=IA64_KR(CURRENT)                // M2 (12 cyc)
+        shladd r18=r17,3,r14                    // A
+        mov r19=NR_syscalls-1                   // A
+        ;;
+        lfetch [r18]                            // M0|1
+        MOV_FROM_PSR(p0, r29, r8)               // M2 (12 cyc)
+        // If r17 is a NaT, p6 will be zero
+        cmp.geu p6,p7=r19,r17                   // A    (sysnr > 0 && sysnr < 1024+NR_syscalls)?
+        ;;
+        mov r21=ar.fpsr                         // M2 (12 cyc)
+        tnat.nz p10,p9=r15                      // I0
+        mov.i r26=ar.pfs                        // I0 (would stall anyhow due to srlz.d...)
+        ;;
+        srlz.d                                  // M0 (forces split-issue) ensure PSR.BE==0
+(p6)    ld8 r18=[r18]                           // M0|1
+        nop.i 0
+        ;;
+        nop.m 0
+(p6)    tbit.z.unc p8,p0=r18,0                  // I0 (dual-issues with "mov b7=r18"!)
+        nop.i 0
+        ;;
+        SSM_PSR_I(p8, p14, r25)
+(p6)    mov b7=r18                              // I0
+(p8)    br.dptk.many b7                         // B
+
+        mov r27=ar.rsc                          // M2 (12 cyc)
+/*
+ * brl.cond doesn't work as intended because the linker would convert this branch
+ * into a branch to a PLT.  Perhaps there will be a way to avoid this with some
+ * future version of the linker.  In the meantime, we just use an indirect branch
+ * instead.
+ */
+#ifdef CONFIG_ITANIUM
+(p6)    add r14=-8,r14                          // r14 <- addr of fsys_bubble_down entry
+        ;;
+(p6)    ld8 r14=[r14]                           // r14 <- fsys_bubble_down
+        ;;
+(p6)    mov b7=r14
+(p6)    br.sptk.many b7
+#else
+        BRL_COND_FSYS_BUBBLE_DOWN(p6)
+#endif
+        SSM_PSR_I(p0, p14, r10)
+        mov r10=-1
+(p10)   mov r8=EINVAL
+(p9)    mov r8=ENOSYS
+        FSYS_RETURN
+
+#ifdef CONFIG_PARAVIRT
+        /*
+         * padd to make the size of this symbol constant
+         * independent of paravirtualization.
+         */
+        .align PAGE_SIZE / 8
+#endif
+END(__kernel_syscall_via_epc)
diff --git a/arch/ia64/kernel/gate.lds.S b/arch/ia64/kernel/gate.lds.S
index 3cb1abc00e24..88c64ed47c36 100644
--- a/arch/ia64/kernel/gate.lds.S
+++ b/arch/ia64/kernel/gate.lds.S
@@ -7,6 +7,7 @@
 
 
 #include <asm/system.h>
+#include "paravirt_patchlist.h"
 
 SECTIONS
 {
@@ -33,21 +34,21 @@ SECTIONS
         . = GATE_ADDR + 0x600;
 
         .data.patch             : {
-                __start_gate_mckinley_e9_patchlist = .;
+                __paravirt_start_gate_mckinley_e9_patchlist = .;
                 *(.data.patch.mckinley_e9)
-                __end_gate_mckinley_e9_patchlist = .;
+                __paravirt_end_gate_mckinley_e9_patchlist = .;
 
-                __start_gate_vtop_patchlist = .;
+                __paravirt_start_gate_vtop_patchlist = .;
                 *(.data.patch.vtop)
-                __end_gate_vtop_patchlist = .;
+                __paravirt_end_gate_vtop_patchlist = .;
 
-                __start_gate_fsyscall_patchlist = .;
+                __paravirt_start_gate_fsyscall_patchlist = .;
                 *(.data.patch.fsyscall_table)
-                __end_gate_fsyscall_patchlist = .;
+                __paravirt_end_gate_fsyscall_patchlist = .;
 
-                __start_gate_brl_fsys_bubble_down_patchlist = .;
+                __paravirt_start_gate_brl_fsys_bubble_down_patchlist = .;
                 *(.data.patch.brl_fsys_bubble_down)
-                __end_gate_brl_fsys_bubble_down_patchlist = .;
+                __paravirt_end_gate_brl_fsys_bubble_down_patchlist = .;
         }                                               :readable
 
         .IA_64.unwind_info      : { *(.IA_64.unwind_info*) }
diff --git a/arch/ia64/kernel/head.S b/arch/ia64/kernel/head.S
index 59301c472800..23f846de62d5 100644
--- a/arch/ia64/kernel/head.S
+++ b/arch/ia64/kernel/head.S
@@ -1050,7 +1050,7 @@ END(ia64_delay_loop)
  * except that the multiplication and the shift are done with 128-bit
  * intermediate precision so that we can produce a full 64-bit result.
  */
-GLOBAL_ENTRY(sched_clock)
+GLOBAL_ENTRY(ia64_native_sched_clock)
         addl r8=THIS_CPU(cpu_info) + IA64_CPUINFO_NSEC_PER_CYC_OFFSET,r0
         mov.m r9=ar.itc         // fetch cycle-counter                          (35 cyc)
         ;;
@@ -1066,7 +1066,13 @@ GLOBAL_ENTRY(sched_clock)
         ;;
         shrp r8=r9,r8,IA64_NSEC_PER_CYC_SHIFT
         br.ret.sptk.many rp
-END(sched_clock)
+END(ia64_native_sched_clock)
+#ifndef CONFIG_PARAVIRT
+        //unsigned long long
+        //sched_clock(void) __attribute__((alias("ia64_native_sched_clock")));
+        .global sched_clock
+sched_clock = ia64_native_sched_clock
+#endif
 
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
 GLOBAL_ENTRY(cycle_to_cputime)
diff --git a/arch/ia64/kernel/ivt.S b/arch/ia64/kernel/ivt.S
index f675d8e33853..ec9a5fdfa1b9 100644
--- a/arch/ia64/kernel/ivt.S
+++ b/arch/ia64/kernel/ivt.S
@@ -804,7 +804,7 @@ ENTRY(break_fault)
 ///////////////////////////////////////////////////////////////////////
         st1 [r16]=r0                            // M2|3 clear current->thread.on_ustack flag
 #ifdef CONFIG_VIRT_CPU_ACCOUNTING
-        mov.m r30=ar.itc                        // M    get cycle for accounting
+        MOV_FROM_ITC(p0, p14, r30, r18)         // M    get cycle for accounting
 #else
         mov b6=r30                              // I0   setup syscall handler branch reg early
 #endif
diff --git a/arch/ia64/kernel/mca.c b/arch/ia64/kernel/mca.c
index bab1de2d2f6a..8f33a8840422 100644
--- a/arch/ia64/kernel/mca.c
+++ b/arch/ia64/kernel/mca.c
@@ -1456,9 +1456,9 @@ ia64_mca_cmc_int_caller(int cmc_irq, void *arg)
 
         ia64_mca_cmc_int_handler(cmc_irq, arg);
 
-        for (++cpuid ; cpuid < NR_CPUS && !cpu_online(cpuid) ; cpuid++);
+        cpuid = cpumask_next(cpuid+1, cpu_online_mask);
 
-        if (cpuid < NR_CPUS) {
+        if (cpuid < nr_cpu_ids) {
                 platform_send_ipi(cpuid, IA64_CMCP_VECTOR, IA64_IPI_DM_INT, 0);
         } else {
                 /* If no log record, switch out of polling mode */
@@ -1525,7 +1525,7 @@ ia64_mca_cpe_int_caller(int cpe_irq, void *arg)
 
         ia64_mca_cpe_int_handler(cpe_irq, arg);
 
-        for (++cpuid ; cpuid < NR_CPUS && !cpu_online(cpuid) ; cpuid++);
+        cpuid = cpumask_next(cpuid+1, cpu_online_mask);
 
         if (cpuid < NR_CPUS) {
                 platform_send_ipi(cpuid, IA64_CPEP_VECTOR, IA64_IPI_DM_INT, 0);
diff --git a/arch/ia64/kernel/module.c b/arch/ia64/kernel/module.c
index aaa7d901521f..da3b0cf495a3 100644
--- a/arch/ia64/kernel/module.c
+++ b/arch/ia64/kernel/module.c
@@ -446,6 +446,14 @@ module_frob_arch_sections (Elf_Ehdr *ehdr, Elf_Shdr *sechdrs, char *secstrings,
                         mod->arch.opd = s;
                 else if (strcmp(".IA_64.unwind", secstrings + s->sh_name) == 0)
                         mod->arch.unwind = s;
+#ifdef CONFIG_PARAVIRT
+                else if (strcmp(".paravirt_bundles",
+                                secstrings + s->sh_name) == 0)
+                        mod->arch.paravirt_bundles = s;
+                else if (strcmp(".paravirt_insts",
+                                secstrings + s->sh_name) == 0)
+                        mod->arch.paravirt_insts = s;
+#endif
 
         if (!mod->arch.core_plt || !mod->arch.init_plt || !mod->arch.got || !mod->arch.opd) {
                 printk(KERN_ERR "%s: sections missing\n", mod->name);
@@ -525,8 +533,7 @@ get_ltoff (struct module *mod, uint64_t value, int *okp)
                         goto found;
 
         /* Not enough GOT entries? */
-        if (e >= (struct got_entry *) (mod->arch.got->sh_addr + mod->arch.got->sh_size))
-                BUG();
+        BUG_ON(e >= (struct got_entry *) (mod->arch.got->sh_addr + mod->arch.got->sh_size));
 
         e->val = value;
         ++mod->arch.next_got_entry;
@@ -921,6 +928,30 @@ module_finalize (const Elf_Ehdr *hdr, const Elf_Shdr *sechdrs, struct module *mo
         DEBUGP("%s: init: entry=%p\n", __func__, mod->init);
         if (mod->arch.unwind)
                 register_unwind_table(mod);
+#ifdef CONFIG_PARAVIRT
+        if (mod->arch.paravirt_bundles) {
+                struct paravirt_patch_site_bundle *start =
+                        (struct paravirt_patch_site_bundle *)
+                        mod->arch.paravirt_bundles->sh_addr;
+                struct paravirt_patch_site_bundle *end =
+                        (struct paravirt_patch_site_bundle *)
+                        (mod->arch.paravirt_bundles->sh_addr +
+                         mod->arch.paravirt_bundles->sh_size);
+
+                paravirt_patch_apply_bundle(start, end);
+        }
+        if (mod->arch.paravirt_insts) {
+                struct paravirt_patch_site_inst *start =
+                        (struct paravirt_patch_site_inst *)
+                        mod->arch.paravirt_insts->sh_addr;
+                struct paravirt_patch_site_inst *end =
+                        (struct paravirt_patch_site_inst *)
+                        (mod->arch.paravirt_insts->sh_addr +
+                         mod->arch.paravirt_insts->sh_size);
+
+                paravirt_patch_apply_inst(start, end);
+        }
+#endif
         return 0;
 }
 
diff --git a/arch/ia64/kernel/paravirt.c b/arch/ia64/kernel/paravirt.c
index 9f14c16f6369..a21d7bb9c69c 100644
--- a/arch/ia64/kernel/paravirt.c
+++ b/arch/ia64/kernel/paravirt.c
@@ -46,13 +46,23 @@ struct pv_info pv_info = {
  * initialization hooks.
  */
 
-struct pv_init_ops pv_init_ops;
+static void __init
+ia64_native_patch_branch(unsigned long tag, unsigned long type);
+
+struct pv_init_ops pv_init_ops =
+{
+#ifdef ASM_SUPPORTED
+        .patch_bundle = ia64_native_patch_bundle,
+#endif
+        .patch_branch = ia64_native_patch_branch,
+};
 
 /***************************************************************************
  * pv_cpu_ops
  * intrinsics hooks.
  */
 
+#ifndef ASM_SUPPORTED
 /* ia64_native_xxx are macros so that we have to make them real functions */
 
 #define DEFINE_VOID_FUNC1(name)                                 \
@@ -60,7 +70,14 @@ struct pv_init_ops pv_init_ops;
         ia64_native_ ## name ## _func(unsigned long arg)        \
         {                                                       \
                 ia64_native_ ## name(arg);                      \
-        }                                                       \
+        }
+
+#define DEFINE_VOID_FUNC1_VOID(name)                            \
+        static void                                             \
+        ia64_native_ ## name ## _func(void *arg)                \
+        {                                                       \
+                ia64_native_ ## name(arg);                      \
+        }
 
 #define DEFINE_VOID_FUNC2(name)                                 \
         static void                                             \
@@ -68,7 +85,7 @@ struct pv_init_ops pv_init_ops;
                                       unsigned long arg1)       \
         {                                                       \
                 ia64_native_ ## name(arg0, arg1);               \
-        }                                                       \
+        }
 
 #define DEFINE_FUNC0(name)                      \
         static unsigned long                    \
@@ -84,7 +101,7 @@ struct pv_init_ops pv_init_ops;
                 return ia64_native_ ## name(arg);       \
         }                                               \
 
-DEFINE_VOID_FUNC1(fc);
+DEFINE_VOID_FUNC1_VOID(fc);
 DEFINE_VOID_FUNC1(intrin_local_irq_restore);
 
 DEFINE_VOID_FUNC2(ptcga);
@@ -274,6 +291,266 @@ ia64_native_setreg_func(int regnum, unsigned long val)
                 break;
         }
 }
+#else
+
+#define __DEFINE_FUNC(name, code)                                       \
+        extern const char ia64_native_ ## name ## _direct_start[];      \
+        extern const char ia64_native_ ## name ## _direct_end[];        \
+        asm (".align 32\n"                                              \
+             ".proc ia64_native_" #name "_func\n"                       \
+             "ia64_native_" #name "_func:\n"                            \
+             "ia64_native_" #name "_direct_start:\n"                    \
+             code                                                       \
+             "ia64_native_" #name "_direct_end:\n"                      \
+             "br.cond.sptk.many b6\n"                                   \
+             ".endp ia64_native_" #name "_func\n")
+
+#define DEFINE_VOID_FUNC0(name, code)                           \
+        extern void                                             \
+        ia64_native_ ## name ## _func(void);                    \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_VOID_FUNC1(name, code)                           \
+        extern void                                             \
+        ia64_native_ ## name ## _func(unsigned long arg);       \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_VOID_FUNC1_VOID(name, code)                      \
+        extern void                                             \
+        ia64_native_ ## name ## _func(void *arg);               \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_VOID_FUNC2(name, code)                           \
+        extern void                                             \
+        ia64_native_ ## name ## _func(unsigned long arg0,       \
+                                      unsigned long arg1);      \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_FUNC0(name, code)                \
+        extern unsigned long                    \
+        ia64_native_ ## name ## _func(void);    \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_FUNC1(name, type, code)                  \
+        extern unsigned long                            \
+        ia64_native_ ## name ## _func(type arg);        \
+        __DEFINE_FUNC(name, code)
+
+DEFINE_VOID_FUNC1_VOID(fc,
+                       "fc r8\n");
+DEFINE_VOID_FUNC1(intrin_local_irq_restore,
+                  ";;\n"
+                  "     cmp.ne p6, p7 = r8, r0\n"
+                  ";;\n"
+                  "(p6) ssm psr.i\n"
+                  "(p7) rsm psr.i\n"
+                  ";;\n"
+                  "(p6) srlz.d\n");
+
+DEFINE_VOID_FUNC2(ptcga,
+                  "ptc.ga r8, r9\n");
+DEFINE_VOID_FUNC2(set_rr,
+                  "mov rr[r8] = r9\n");
+
+/* ia64_native_getreg(_IA64_REG_PSR) & IA64_PSR_I */
+DEFINE_FUNC0(get_psr_i,
+             "mov r2 = " __stringify(1 << IA64_PSR_I_BIT) "\n"
+             "mov r8 = psr\n"
+             ";;\n"
+             "and r8 = r2, r8\n");
+
+DEFINE_FUNC1(thash, unsigned long,
+             "thash r8 = r8\n");
+DEFINE_FUNC1(get_cpuid, int,
+             "mov r8 = cpuid[r8]\n");
+DEFINE_FUNC1(get_pmd, int,
+             "mov r8 = pmd[r8]\n");
+DEFINE_FUNC1(get_rr, unsigned long,
+             "mov r8 = rr[r8]\n");
+
+DEFINE_VOID_FUNC0(ssm_i,
+                  "ssm psr.i\n");
+DEFINE_VOID_FUNC0(rsm_i,
+                  "rsm psr.i\n");
+
+extern void
+ia64_native_set_rr0_to_rr4_func(unsigned long val0, unsigned long val1,
+                                unsigned long val2, unsigned long val3,
+                                unsigned long val4);
+__DEFINE_FUNC(set_rr0_to_rr4,
+              "mov rr[r0] = r8\n"
+              "movl r2 = 0x2000000000000000\n"
+              ";;\n"
+              "mov rr[r2] = r9\n"
+              "shl r3 = r2, 1\n"        /* movl r3 = 0x4000000000000000 */
+              ";;\n"
+              "add r2 = r2, r3\n"       /* movl r2 = 0x6000000000000000 */
+              "mov rr[r3] = r10\n"
+              ";;\n"
+              "mov rr[r2] = r11\n"
+              "shl r3 = r3, 1\n"        /* movl r3 = 0x8000000000000000 */
+              ";;\n"
+              "mov rr[r3] = r14\n");
+
+extern unsigned long ia64_native_getreg_func(int regnum);
+asm(".global ia64_native_getreg_func\n");
+#define __DEFINE_GET_REG(id, reg)                       \
+        "mov r2 = " __stringify(_IA64_REG_ ## id) "\n"  \
+        ";;\n"                                          \
+        "cmp.eq p6, p0 = r2, r8\n"                      \
+        ";;\n"                                          \
+        "(p6) mov r8 = " #reg "\n"                      \
+        "(p6) br.cond.sptk.many b6\n"                   \
+        ";;\n"
+#define __DEFINE_GET_AR(id, reg)        __DEFINE_GET_REG(AR_ ## id, ar.reg)
+#define __DEFINE_GET_CR(id, reg)        __DEFINE_GET_REG(CR_ ## id, cr.reg)
+
+__DEFINE_FUNC(getreg,
+              __DEFINE_GET_REG(GP, gp)
+              /*__DEFINE_GET_REG(IP, ip)*/ /* returned ip value shouldn't be constant */
+              __DEFINE_GET_REG(PSR, psr)
+              __DEFINE_GET_REG(TP, tp)
+              __DEFINE_GET_REG(SP, sp)
+
+              __DEFINE_GET_REG(AR_KR0, ar0)
+              __DEFINE_GET_REG(AR_KR1, ar1)
+              __DEFINE_GET_REG(AR_KR2, ar2)
+              __DEFINE_GET_REG(AR_KR3, ar3)
+              __DEFINE_GET_REG(AR_KR4, ar4)
+              __DEFINE_GET_REG(AR_KR5, ar5)
+              __DEFINE_GET_REG(AR_KR6, ar6)
+              __DEFINE_GET_REG(AR_KR7, ar7)
+              __DEFINE_GET_AR(RSC, rsc)
+              __DEFINE_GET_AR(BSP, bsp)
+              __DEFINE_GET_AR(BSPSTORE, bspstore)
+              __DEFINE_GET_AR(RNAT, rnat)
+              __DEFINE_GET_AR(FCR, fcr)
+              __DEFINE_GET_AR(EFLAG, eflag)
+              __DEFINE_GET_AR(CSD, csd)
+              __DEFINE_GET_AR(SSD, ssd)
+              __DEFINE_GET_REG(AR_CFLAG, ar27)
+              __DEFINE_GET_AR(FSR, fsr)
+              __DEFINE_GET_AR(FIR, fir)
+              __DEFINE_GET_AR(FDR, fdr)
+              __DEFINE_GET_AR(CCV, ccv)
+              __DEFINE_GET_AR(UNAT, unat)
+              __DEFINE_GET_AR(FPSR, fpsr)
+              __DEFINE_GET_AR(ITC, itc)
+              __DEFINE_GET_AR(PFS, pfs)
+              __DEFINE_GET_AR(LC, lc)
+              __DEFINE_GET_AR(EC, ec)
+
+              __DEFINE_GET_CR(DCR, dcr)
+              __DEFINE_GET_CR(ITM, itm)
+              __DEFINE_GET_CR(IVA, iva)
+              __DEFINE_GET_CR(PTA, pta)
+              __DEFINE_GET_CR(IPSR, ipsr)
+              __DEFINE_GET_CR(ISR, isr)
+              __DEFINE_GET_CR(IIP, iip)
+              __DEFINE_GET_CR(IFA, ifa)
+              __DEFINE_GET_CR(ITIR, itir)
+              __DEFINE_GET_CR(IIPA, iipa)
+              __DEFINE_GET_CR(IFS, ifs)
+              __DEFINE_GET_CR(IIM, iim)
+              __DEFINE_GET_CR(IHA, iha)
+              __DEFINE_GET_CR(LID, lid)
+              __DEFINE_GET_CR(IVR, ivr)
+              __DEFINE_GET_CR(TPR, tpr)
+              __DEFINE_GET_CR(EOI, eoi)
+              __DEFINE_GET_CR(IRR0, irr0)
+              __DEFINE_GET_CR(IRR1, irr1)
+              __DEFINE_GET_CR(IRR2, irr2)
+              __DEFINE_GET_CR(IRR3, irr3)
+              __DEFINE_GET_CR(ITV, itv)
+              __DEFINE_GET_CR(PMV, pmv)
+              __DEFINE_GET_CR(CMCV, cmcv)
+              __DEFINE_GET_CR(LRR0, lrr0)
+              __DEFINE_GET_CR(LRR1, lrr1)
+
+              "mov r8 = -1\n"   /* unsupported case */
+        );
+
+extern void ia64_native_setreg_func(int regnum, unsigned long val);
+asm(".global ia64_native_setreg_func\n");
+#define __DEFINE_SET_REG(id, reg)                       \
+        "mov r2 = " __stringify(_IA64_REG_ ## id) "\n"  \
+        ";;\n"                                          \
+        "cmp.eq p6, p0 = r2, r9\n"                      \
+        ";;\n"                                          \
+        "(p6) mov " #reg " = r8\n"                      \
+        "(p6) br.cond.sptk.many b6\n"                   \
+        ";;\n"
+#define __DEFINE_SET_AR(id, reg)        __DEFINE_SET_REG(AR_ ## id, ar.reg)
+#define __DEFINE_SET_CR(id, reg)        __DEFINE_SET_REG(CR_ ## id, cr.reg)
+__DEFINE_FUNC(setreg,
+              "mov r2 = " __stringify(_IA64_REG_PSR_L) "\n"
+              ";;\n"
+              "cmp.eq p6, p0 = r2, r9\n"
+              ";;\n"
+              "(p6) mov psr.l = r8\n"
+#ifdef HAVE_SERIALIZE_DIRECTIVE
+              ".serialize.data\n"
+#endif
+              "(p6) br.cond.sptk.many b6\n"
+              __DEFINE_SET_REG(GP, gp)
+              __DEFINE_SET_REG(SP, sp)
+
+              __DEFINE_SET_REG(AR_KR0, ar0)
+              __DEFINE_SET_REG(AR_KR1, ar1)
+              __DEFINE_SET_REG(AR_KR2, ar2)
+              __DEFINE_SET_REG(AR_KR3, ar3)
+              __DEFINE_SET_REG(AR_KR4, ar4)
+              __DEFINE_SET_REG(AR_KR5, ar5)
+              __DEFINE_SET_REG(AR_KR6, ar6)
+              __DEFINE_SET_REG(AR_KR7, ar7)
+              __DEFINE_SET_AR(RSC, rsc)
+              __DEFINE_SET_AR(BSP, bsp)
+              __DEFINE_SET_AR(BSPSTORE, bspstore)
+              __DEFINE_SET_AR(RNAT, rnat)
+              __DEFINE_SET_AR(FCR, fcr)
+              __DEFINE_SET_AR(EFLAG, eflag)
+              __DEFINE_SET_AR(CSD, csd)
+              __DEFINE_SET_AR(SSD, ssd)
+              __DEFINE_SET_REG(AR_CFLAG, ar27)
+              __DEFINE_SET_AR(FSR, fsr)
+              __DEFINE_SET_AR(FIR, fir)
+              __DEFINE_SET_AR(FDR, fdr)
+              __DEFINE_SET_AR(CCV, ccv)
+              __DEFINE_SET_AR(UNAT, unat)
+              __DEFINE_SET_AR(FPSR, fpsr)
+              __DEFINE_SET_AR(ITC, itc)
+              __DEFINE_SET_AR(PFS, pfs)
+              __DEFINE_SET_AR(LC, lc)
+              __DEFINE_SET_AR(EC, ec)
+
+              __DEFINE_SET_CR(DCR, dcr)
+              __DEFINE_SET_CR(ITM, itm)
+              __DEFINE_SET_CR(IVA, iva)
+              __DEFINE_SET_CR(PTA, pta)
+              __DEFINE_SET_CR(IPSR, ipsr)
+              __DEFINE_SET_CR(ISR, isr)
+              __DEFINE_SET_CR(IIP, iip)
+              __DEFINE_SET_CR(IFA, ifa)
+              __DEFINE_SET_CR(ITIR, itir)
+              __DEFINE_SET_CR(IIPA, iipa)
+              __DEFINE_SET_CR(IFS, ifs)
+              __DEFINE_SET_CR(IIM, iim)
+              __DEFINE_SET_CR(IHA, iha)
+              __DEFINE_SET_CR(LID, lid)
+              __DEFINE_SET_CR(IVR, ivr)
+              __DEFINE_SET_CR(TPR, tpr)
+              __DEFINE_SET_CR(EOI, eoi)
+              __DEFINE_SET_CR(IRR0, irr0)
+              __DEFINE_SET_CR(IRR1, irr1)
+              __DEFINE_SET_CR(IRR2, irr2)
+              __DEFINE_SET_CR(IRR3, irr3)
+              __DEFINE_SET_CR(ITV, itv)
+              __DEFINE_SET_CR(PMV, pmv)
+              __DEFINE_SET_CR(CMCV, cmcv)
+              __DEFINE_SET_CR(LRR0, lrr0)
+              __DEFINE_SET_CR(LRR1, lrr1)
+        );
+#endif
 
 struct pv_cpu_ops pv_cpu_ops = {
         .fc             = ia64_native_fc_func,
@@ -366,4 +643,258 @@ ia64_native_do_steal_accounting(unsigned long *new_itm)
 
 struct pv_time_ops pv_time_ops = {
         .do_steal_accounting = ia64_native_do_steal_accounting,
+        .sched_clock = ia64_native_sched_clock,
+};
+
+/***************************************************************************
+ * binary pacthing
+ * pv_init_ops.patch_bundle
+ */
+
+#ifdef ASM_SUPPORTED
+#define IA64_NATIVE_PATCH_DEFINE_GET_REG(name, reg)     \
+        __DEFINE_FUNC(get_ ## name,                     \
+                      ";;\n"                            \
+                      "mov r8 = " #reg "\n"             \
+                      ";;\n")
+
+#define IA64_NATIVE_PATCH_DEFINE_SET_REG(name, reg)     \
+        __DEFINE_FUNC(set_ ## name,                     \
+                      ";;\n"                            \
+                      "mov " #reg " = r8\n"             \
+                      ";;\n")
+
+#define IA64_NATIVE_PATCH_DEFINE_REG(name, reg)         \
+        IA64_NATIVE_PATCH_DEFINE_GET_REG(name, reg);    \
+        IA64_NATIVE_PATCH_DEFINE_SET_REG(name, reg)     \
+
+#define IA64_NATIVE_PATCH_DEFINE_AR(name, reg)                  \
+        IA64_NATIVE_PATCH_DEFINE_REG(ar_ ## name, ar.reg)
+
+#define IA64_NATIVE_PATCH_DEFINE_CR(name, reg)                  \
+        IA64_NATIVE_PATCH_DEFINE_REG(cr_ ## name, cr.reg)
+
+
+IA64_NATIVE_PATCH_DEFINE_GET_REG(psr, psr);
+IA64_NATIVE_PATCH_DEFINE_GET_REG(tp, tp);
+
+/* IA64_NATIVE_PATCH_DEFINE_SET_REG(psr_l, psr.l); */
+__DEFINE_FUNC(set_psr_l,
+              ";;\n"
+              "mov psr.l = r8\n"
+#ifdef HAVE_SERIALIZE_DIRECTIVE
+              ".serialize.data\n"
+#endif
+              ";;\n");
+
+IA64_NATIVE_PATCH_DEFINE_REG(gp, gp);
+IA64_NATIVE_PATCH_DEFINE_REG(sp, sp);
+
+IA64_NATIVE_PATCH_DEFINE_REG(kr0, ar0);
+IA64_NATIVE_PATCH_DEFINE_REG(kr1, ar1);
+IA64_NATIVE_PATCH_DEFINE_REG(kr2, ar2);
+IA64_NATIVE_PATCH_DEFINE_REG(kr3, ar3);
+IA64_NATIVE_PATCH_DEFINE_REG(kr4, ar4);
+IA64_NATIVE_PATCH_DEFINE_REG(kr5, ar5);
+IA64_NATIVE_PATCH_DEFINE_REG(kr6, ar6);
+IA64_NATIVE_PATCH_DEFINE_REG(kr7, ar7);
+
+IA64_NATIVE_PATCH_DEFINE_AR(rsc, rsc);
+IA64_NATIVE_PATCH_DEFINE_AR(bsp, bsp);
+IA64_NATIVE_PATCH_DEFINE_AR(bspstore, bspstore);
+IA64_NATIVE_PATCH_DEFINE_AR(rnat, rnat);
+IA64_NATIVE_PATCH_DEFINE_AR(fcr, fcr);
+IA64_NATIVE_PATCH_DEFINE_AR(eflag, eflag);
+IA64_NATIVE_PATCH_DEFINE_AR(csd, csd);
+IA64_NATIVE_PATCH_DEFINE_AR(ssd, ssd);
+IA64_NATIVE_PATCH_DEFINE_REG(ar27, ar27);
+IA64_NATIVE_PATCH_DEFINE_AR(fsr, fsr);
+IA64_NATIVE_PATCH_DEFINE_AR(fir, fir);
+IA64_NATIVE_PATCH_DEFINE_AR(fdr, fdr);
+IA64_NATIVE_PATCH_DEFINE_AR(ccv, ccv);
+IA64_NATIVE_PATCH_DEFINE_AR(unat, unat);
+IA64_NATIVE_PATCH_DEFINE_AR(fpsr, fpsr);
+IA64_NATIVE_PATCH_DEFINE_AR(itc, itc);
+IA64_NATIVE_PATCH_DEFINE_AR(pfs, pfs);
+IA64_NATIVE_PATCH_DEFINE_AR(lc, lc);
+IA64_NATIVE_PATCH_DEFINE_AR(ec, ec);
+
+IA64_NATIVE_PATCH_DEFINE_CR(dcr, dcr);
+IA64_NATIVE_PATCH_DEFINE_CR(itm, itm);
+IA64_NATIVE_PATCH_DEFINE_CR(iva, iva);
+IA64_NATIVE_PATCH_DEFINE_CR(pta, pta);
+IA64_NATIVE_PATCH_DEFINE_CR(ipsr, ipsr);
+IA64_NATIVE_PATCH_DEFINE_CR(isr, isr);
+IA64_NATIVE_PATCH_DEFINE_CR(iip, iip);
+IA64_NATIVE_PATCH_DEFINE_CR(ifa, ifa);
+IA64_NATIVE_PATCH_DEFINE_CR(itir, itir);
+IA64_NATIVE_PATCH_DEFINE_CR(iipa, iipa);
+IA64_NATIVE_PATCH_DEFINE_CR(ifs, ifs);
+IA64_NATIVE_PATCH_DEFINE_CR(iim, iim);
+IA64_NATIVE_PATCH_DEFINE_CR(iha, iha);
+IA64_NATIVE_PATCH_DEFINE_CR(lid, lid);
+IA64_NATIVE_PATCH_DEFINE_CR(ivr, ivr);
+IA64_NATIVE_PATCH_DEFINE_CR(tpr, tpr);
+IA64_NATIVE_PATCH_DEFINE_CR(eoi, eoi);
+IA64_NATIVE_PATCH_DEFINE_CR(irr0, irr0);
+IA64_NATIVE_PATCH_DEFINE_CR(irr1, irr1);
+IA64_NATIVE_PATCH_DEFINE_CR(irr2, irr2);
+IA64_NATIVE_PATCH_DEFINE_CR(irr3, irr3);
+IA64_NATIVE_PATCH_DEFINE_CR(itv, itv);
+IA64_NATIVE_PATCH_DEFINE_CR(pmv, pmv);
+IA64_NATIVE_PATCH_DEFINE_CR(cmcv, cmcv);
+IA64_NATIVE_PATCH_DEFINE_CR(lrr0, lrr0);
+IA64_NATIVE_PATCH_DEFINE_CR(lrr1, lrr1);
+
+static const struct paravirt_patch_bundle_elem ia64_native_patch_bundle_elems[]
+__initdata_or_module =
+{
+#define IA64_NATIVE_PATCH_BUNDLE_ELEM(name, type)               \
+        {                                                       \
+                (void*)ia64_native_ ## name ## _direct_start,   \
+                (void*)ia64_native_ ## name ## _direct_end,     \
+                PARAVIRT_PATCH_TYPE_ ## type,                   \
+        }
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(fc, FC),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(thash, THASH),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(get_cpuid, GET_CPUID),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(get_pmd, GET_PMD),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(ptcga, PTCGA),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(get_rr, GET_RR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(set_rr, SET_RR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(set_rr0_to_rr4, SET_RR0_TO_RR4),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(ssm_i, SSM_I),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(rsm_i, RSM_I),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(get_psr_i, GET_PSR_I),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM(intrin_local_irq_restore,
+                                      INTRIN_LOCAL_IRQ_RESTORE),
+
+#define IA64_NATIVE_PATCH_BUNDLE_ELEM_GETREG(name, reg)                 \
+        {                                                               \
+                (void*)ia64_native_get_ ## name ## _direct_start,       \
+                (void*)ia64_native_get_ ## name ## _direct_end,         \
+                PARAVIRT_PATCH_TYPE_GETREG + _IA64_REG_ ## reg,         \
+        }
+
+#define IA64_NATIVE_PATCH_BUNDLE_ELEM_SETREG(name, reg)                 \
+        {                                                               \
+                (void*)ia64_native_set_ ## name ## _direct_start,       \
+                (void*)ia64_native_set_ ## name ## _direct_end,         \
+                PARAVIRT_PATCH_TYPE_SETREG + _IA64_REG_ ## reg,         \
+        }
+
+#define IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(name, reg)            \
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_GETREG(name, reg),        \
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_SETREG(name, reg)         \
+
+#define IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(name, reg)             \
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(ar_ ## name, AR_ ## reg)
+
+#define IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(name, reg)             \
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(cr_ ## name, CR_ ## reg)
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_GETREG(psr, PSR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_GETREG(tp, TP),
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_SETREG(psr_l, PSR_L),
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(gp, GP),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(sp, SP),
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr0, AR_KR0),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr1, AR_KR1),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr2, AR_KR2),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr3, AR_KR3),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr4, AR_KR4),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr5, AR_KR5),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr6, AR_KR6),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(kr7, AR_KR7),
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(rsc, RSC),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(bsp, BSP),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(bspstore, BSPSTORE),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(rnat, RNAT),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(fcr, FCR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(eflag, EFLAG),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(csd, CSD),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(ssd, SSD),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_REG(ar27, AR_CFLAG),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(fsr, FSR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(fir, FIR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(fdr, FDR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(ccv, CCV),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(unat, UNAT),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(fpsr, FPSR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(itc, ITC),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(pfs, PFS),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(lc, LC),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_AR(ec, EC),
+
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(dcr, DCR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(itm, ITM),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(iva, IVA),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(pta, PTA),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(ipsr, IPSR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(isr, ISR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(iip, IIP),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(ifa, IFA),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(itir, ITIR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(iipa, IIPA),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(ifs, IFS),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(iim, IIM),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(iha, IHA),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(lid, LID),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(ivr, IVR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(tpr, TPR),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(eoi, EOI),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(irr0, IRR0),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(irr1, IRR1),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(irr2, IRR2),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(irr3, IRR3),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(itv, ITV),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(pmv, PMV),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(cmcv, CMCV),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(lrr0, LRR0),
+        IA64_NATIVE_PATCH_BUNDLE_ELEM_CR(lrr1, LRR1),
 };
+
+unsigned long __init_or_module
+ia64_native_patch_bundle(void *sbundle, void *ebundle, unsigned long type)
+{
+        const unsigned long nelems = sizeof(ia64_native_patch_bundle_elems) /
+                sizeof(ia64_native_patch_bundle_elems[0]);
+
+        return __paravirt_patch_apply_bundle(sbundle, ebundle, type,
+                                              ia64_native_patch_bundle_elems,
+                                              nelems, NULL);
+}
+#endif /* ASM_SUPPOTED */
+
+extern const char ia64_native_switch_to[];
+extern const char ia64_native_leave_syscall[];
+extern const char ia64_native_work_processed_syscall[];
+extern const char ia64_native_leave_kernel[];
+
+const struct paravirt_patch_branch_target ia64_native_branch_target[]
+__initconst = {
+#define PARAVIRT_BR_TARGET(name, type)                  \
+        {                                               \
+                ia64_native_ ## name,                   \
+                PARAVIRT_PATCH_TYPE_BR_ ## type,        \
+        }
+        PARAVIRT_BR_TARGET(switch_to, SWITCH_TO),
+        PARAVIRT_BR_TARGET(leave_syscall, LEAVE_SYSCALL),
+        PARAVIRT_BR_TARGET(work_processed_syscall, WORK_PROCESSED_SYSCALL),
+        PARAVIRT_BR_TARGET(leave_kernel, LEAVE_KERNEL),
+};
+
+static void __init
+ia64_native_patch_branch(unsigned long tag, unsigned long type)
+{
+        const unsigned long nelem =
+                sizeof(ia64_native_branch_target) /
+                sizeof(ia64_native_branch_target[0]);
+        __paravirt_patch_apply_branch(tag, type,
+                                      ia64_native_branch_target, nelem);
+}
diff --git a/arch/ia64/kernel/paravirt_patch.c b/arch/ia64/kernel/paravirt_patch.c
new file mode 100644
index 000000000000..bfdfef1b1ffd
--- /dev/null
+++ b/arch/ia64/kernel/paravirt_patch.c
@@ -0,0 +1,514 @@
+/******************************************************************************
+ * linux/arch/ia64/xen/paravirt_patch.c
+ *
+ * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#include <linux/init.h>
+#include <asm/intrinsics.h>
+#include <asm/kprobes.h>
+#include <asm/paravirt.h>
+#include <asm/paravirt_patch.h>
+
+typedef union ia64_inst {
+        struct {
+                unsigned long long qp : 6;
+                unsigned long long : 31;
+                unsigned long long opcode : 4;
+                unsigned long long reserved : 23;
+        } generic;
+        unsigned long long l;
+} ia64_inst_t;
+
+/*
+ * flush_icache_range() can't be used here.
+ * we are here before cpu_init() which initializes
+ * ia64_i_cache_stride_shift. flush_icache_range() uses it.
+ */
+void __init_or_module
+paravirt_flush_i_cache_range(const void *instr, unsigned long size)
+{
+        extern void paravirt_fc_i(const void *addr);
+        unsigned long i;
+
+        for (i = 0; i < size; i += sizeof(bundle_t))
+                paravirt_fc_i(instr + i);
+}
+
+bundle_t* __init_or_module
+paravirt_get_bundle(unsigned long tag)
+{
+        return (bundle_t *)(tag & ~3UL);
+}
+
+unsigned long __init_or_module
+paravirt_get_slot(unsigned long tag)
+{
+        return tag & 3UL;
+}
+
+unsigned long __init_or_module
+paravirt_get_num_inst(unsigned long stag, unsigned long etag)
+{
+        bundle_t *sbundle = paravirt_get_bundle(stag);
+        unsigned long sslot = paravirt_get_slot(stag);
+        bundle_t *ebundle = paravirt_get_bundle(etag);
+        unsigned long eslot = paravirt_get_slot(etag);
+
+        return (ebundle - sbundle) * 3 + eslot - sslot + 1;
+}
+
+unsigned long __init_or_module
+paravirt_get_next_tag(unsigned long tag)
+{
+        unsigned long slot = paravirt_get_slot(tag);
+
+        switch (slot) {
+        case 0:
+        case 1:
+                return tag + 1;
+        case 2: {
+                bundle_t *bundle = paravirt_get_bundle(tag);
+                return (unsigned long)(bundle + 1);
+        }
+        default:
+                BUG();
+        }
+        /* NOTREACHED */
+}
+
+ia64_inst_t __init_or_module
+paravirt_read_slot0(const bundle_t *bundle)
+{
+        ia64_inst_t inst;
+        inst.l = bundle->quad0.slot0;
+        return inst;
+}
+
+ia64_inst_t __init_or_module
+paravirt_read_slot1(const bundle_t *bundle)
+{
+        ia64_inst_t inst;
+        inst.l = bundle->quad0.slot1_p0 |
+                ((unsigned long long)bundle->quad1.slot1_p1 << 18UL);
+        return inst;
+}
+
+ia64_inst_t __init_or_module
+paravirt_read_slot2(const bundle_t *bundle)
+{
+        ia64_inst_t inst;
+        inst.l = bundle->quad1.slot2;
+        return inst;
+}
+
+ia64_inst_t __init_or_module
+paravirt_read_inst(unsigned long tag)
+{
+        bundle_t *bundle = paravirt_get_bundle(tag);
+        unsigned long slot = paravirt_get_slot(tag);
+
+        switch (slot) {
+        case 0:
+                return paravirt_read_slot0(bundle);
+        case 1:
+                return paravirt_read_slot1(bundle);
+        case 2:
+                return paravirt_read_slot2(bundle);
+        default:
+                BUG();
+        }
+        /* NOTREACHED */
+}
+
+void __init_or_module
+paravirt_write_slot0(bundle_t *bundle, ia64_inst_t inst)
+{
+        bundle->quad0.slot0 = inst.l;
+}
+
+void __init_or_module
+paravirt_write_slot1(bundle_t *bundle, ia64_inst_t inst)
+{
+        bundle->quad0.slot1_p0 = inst.l;
+        bundle->quad1.slot1_p1 = inst.l >> 18UL;
+}
+
+void __init_or_module
+paravirt_write_slot2(bundle_t *bundle, ia64_inst_t inst)
+{
+        bundle->quad1.slot2 = inst.l;
+}
+
+void __init_or_module
+paravirt_write_inst(unsigned long tag, ia64_inst_t inst)
+{
+        bundle_t *bundle = paravirt_get_bundle(tag);
+        unsigned long slot = paravirt_get_slot(tag);
+
+        switch (slot) {
+        case 0:
+                paravirt_write_slot0(bundle, inst);
+                break;
+        case 1:
+                paravirt_write_slot1(bundle, inst);
+                break;
+        case 2:
+                paravirt_write_slot2(bundle, inst);
+                break;
+        default:
+                BUG();
+                break;
+        }
+        paravirt_flush_i_cache_range(bundle, sizeof(*bundle));
+}
+
+/* for debug */
+void
+paravirt_print_bundle(const bundle_t *bundle)
+{
+        const unsigned long *quad = (const unsigned long *)bundle;
+        ia64_inst_t slot0 = paravirt_read_slot0(bundle);
+        ia64_inst_t slot1 = paravirt_read_slot1(bundle);
+        ia64_inst_t slot2 = paravirt_read_slot2(bundle);
+
+        printk(KERN_DEBUG
+               "bundle 0x%p 0x%016lx 0x%016lx\n", bundle, quad[0], quad[1]);
+        printk(KERN_DEBUG
+               "bundle template 0x%x\n",
+               bundle->quad0.template);
+        printk(KERN_DEBUG
+               "slot0 0x%lx slot1_p0 0x%lx slot1_p1 0x%lx slot2 0x%lx\n",
+               (unsigned long)bundle->quad0.slot0,
+               (unsigned long)bundle->quad0.slot1_p0,
+               (unsigned long)bundle->quad1.slot1_p1,
+               (unsigned long)bundle->quad1.slot2);
+        printk(KERN_DEBUG
+               "slot0 0x%016llx slot1 0x%016llx slot2 0x%016llx\n",
+               slot0.l, slot1.l, slot2.l);
+}
+
+static int noreplace_paravirt __init_or_module = 0;
+
+static int __init setup_noreplace_paravirt(char *str)
+{
+        noreplace_paravirt = 1;
+        return 1;
+}
+__setup("noreplace-paravirt", setup_noreplace_paravirt);
+
+#ifdef ASM_SUPPORTED
+static void __init_or_module
+fill_nop_bundle(void *sbundle, void *ebundle)
+{
+        extern const char paravirt_nop_bundle[];
+        extern const unsigned long paravirt_nop_bundle_size;
+
+        void *bundle = sbundle;
+
+        BUG_ON((((unsigned long)sbundle) % sizeof(bundle_t)) != 0);
+        BUG_ON((((unsigned long)ebundle) % sizeof(bundle_t)) != 0);
+
+        while (bundle < ebundle) {
+                memcpy(bundle, paravirt_nop_bundle, paravirt_nop_bundle_size);
+
+                bundle += paravirt_nop_bundle_size;
+        }
+}
+
+/* helper function */
+unsigned long __init_or_module
+__paravirt_patch_apply_bundle(void *sbundle, void *ebundle, unsigned long type,
+                              const struct paravirt_patch_bundle_elem *elems,
+                              unsigned long nelems,
+                              const struct paravirt_patch_bundle_elem **found)
+{
+        unsigned long used = 0;
+        unsigned long i;
+
+        BUG_ON((((unsigned long)sbundle) % sizeof(bundle_t)) != 0);
+        BUG_ON((((unsigned long)ebundle) % sizeof(bundle_t)) != 0);
+
+        found = NULL;
+        for (i = 0; i < nelems; i++) {
+                const struct paravirt_patch_bundle_elem *p = &elems[i];
+                if (p->type == type) {
+                        unsigned long need = p->ebundle - p->sbundle;
+                        unsigned long room = ebundle - sbundle;
+
+                        if (found != NULL)
+                                *found = p;
+
+                        if (room < need) {
+                                /* no room to replace. skip it */
+                                printk(KERN_DEBUG
+                                       "the space is too small to put "
+                                       "bundles. type %ld need %ld room %ld\n",
+                                       type, need, room);
+                                break;
+                        }
+
+                        used = need;
+                        memcpy(sbundle, p->sbundle, used);
+                        break;
+                }
+        }
+
+        return used;
+}
+
+void __init_or_module
+paravirt_patch_apply_bundle(const struct paravirt_patch_site_bundle *start,
+                            const struct paravirt_patch_site_bundle *end)
+{
+        const struct paravirt_patch_site_bundle *p;
+
+        if (noreplace_paravirt)
+                return;
+        if (pv_init_ops.patch_bundle == NULL)
+                return;
+
+        for (p = start; p < end; p++) {
+                unsigned long used;
+
+                used = (*pv_init_ops.patch_bundle)(p->sbundle, p->ebundle,
+                                                   p->type);
+                if (used == 0)
+                        continue;
+
+                fill_nop_bundle(p->sbundle + used, p->ebundle);
+                paravirt_flush_i_cache_range(p->sbundle,
+                                             p->ebundle - p->sbundle);
+        }
+        ia64_sync_i();
+        ia64_srlz_i();
+}
+
+/*
+ * nop.i, nop.m, nop.f instruction are same format.
+ * but nop.b has differennt format.
+ * This doesn't support nop.b for now.
+ */
+static void __init_or_module
+fill_nop_inst(unsigned long stag, unsigned long etag)
+{
+        extern const bundle_t paravirt_nop_mfi_inst_bundle[];
+        unsigned long tag;
+        const ia64_inst_t nop_inst =
+                paravirt_read_slot0(paravirt_nop_mfi_inst_bundle);
+
+        for (tag = stag; tag < etag; tag = paravirt_get_next_tag(tag))
+                paravirt_write_inst(tag, nop_inst);
+}
+
+void __init_or_module
+paravirt_patch_apply_inst(const struct paravirt_patch_site_inst *start,
+                          const struct paravirt_patch_site_inst *end)
+{
+        const struct paravirt_patch_site_inst *p;
+
+        if (noreplace_paravirt)
+                return;
+        if (pv_init_ops.patch_inst == NULL)
+                return;
+
+        for (p = start; p < end; p++) {
+                unsigned long tag;
+                bundle_t *sbundle;
+                bundle_t *ebundle;
+
+                tag = (*pv_init_ops.patch_inst)(p->stag, p->etag, p->type);
+                if (tag == p->stag)
+                        continue;
+
+                fill_nop_inst(tag, p->etag);
+                sbundle = paravirt_get_bundle(p->stag);
+                ebundle = paravirt_get_bundle(p->etag) + 1;
+                paravirt_flush_i_cache_range(sbundle, (ebundle - sbundle) *
+                                             sizeof(bundle_t));
+        }
+        ia64_sync_i();
+        ia64_srlz_i();
+}
+#endif /* ASM_SUPPOTED */
+
+/* brl.cond.sptk.many <target64> X3 */
+typedef union inst_x3_op {
+        ia64_inst_t inst;
+        struct {
+                unsigned long qp: 6;
+                unsigned long btyp: 3;
+                unsigned long unused: 3;
+                unsigned long p: 1;
+                unsigned long imm20b: 20;
+                unsigned long wh: 2;
+                unsigned long d: 1;
+                unsigned long i: 1;
+                unsigned long opcode: 4;
+        };
+        unsigned long l;
+} inst_x3_op_t;
+
+typedef union inst_x3_imm {
+        ia64_inst_t inst;
+        struct {
+                unsigned long unused: 2;
+                unsigned long imm39: 39;
+        };
+        unsigned long l;
+} inst_x3_imm_t;
+
+void __init_or_module
+paravirt_patch_reloc_brl(unsigned long tag, const void *target)
+{
+        unsigned long tag_op = paravirt_get_next_tag(tag);
+        unsigned long tag_imm = tag;
+        bundle_t *bundle = paravirt_get_bundle(tag);
+
+        ia64_inst_t inst_op = paravirt_read_inst(tag_op);
+        ia64_inst_t inst_imm = paravirt_read_inst(tag_imm);
+
+        inst_x3_op_t inst_x3_op = { .l = inst_op.l };
+        inst_x3_imm_t inst_x3_imm = { .l = inst_imm.l };
+
+        unsigned long imm60 =
+                ((unsigned long)target - (unsigned long)bundle) >> 4;
+
+        BUG_ON(paravirt_get_slot(tag) != 1); /* MLX */
+        BUG_ON(((unsigned long)target & (sizeof(bundle_t) - 1)) != 0);
+
+        /* imm60[59] 1bit */
+        inst_x3_op.i = (imm60 >> 59) & 1;
+        /* imm60[19:0] 20bit */
+        inst_x3_op.imm20b = imm60 & ((1UL << 20) - 1);
+        /* imm60[58:20] 39bit */
+        inst_x3_imm.imm39 = (imm60 >> 20) & ((1UL << 39) - 1);
+
+        inst_op.l = inst_x3_op.l;
+        inst_imm.l = inst_x3_imm.l;
+
+        paravirt_write_inst(tag_op, inst_op);
+        paravirt_write_inst(tag_imm, inst_imm);
+}
+
+/* br.cond.sptk.many <target25> B1 */
+typedef union inst_b1 {
+        ia64_inst_t inst;
+        struct {
+                unsigned long qp: 6;
+                unsigned long btype: 3;
+                unsigned long unused: 3;
+                unsigned long p: 1;
+                unsigned long imm20b: 20;
+                unsigned long wh: 2;
+                unsigned long d: 1;
+                unsigned long s: 1;
+                unsigned long opcode: 4;
+        };
+        unsigned long l;
+} inst_b1_t;
+
+void __init
+paravirt_patch_reloc_br(unsigned long tag, const void *target)
+{
+        bundle_t *bundle = paravirt_get_bundle(tag);
+        ia64_inst_t inst = paravirt_read_inst(tag);
+        unsigned long target25 = (unsigned long)target - (unsigned long)bundle;
+        inst_b1_t inst_b1;
+
+        BUG_ON(((unsigned long)target & (sizeof(bundle_t) - 1)) != 0);
+
+        inst_b1.l = inst.l;
+        if (target25 & (1UL << 63))
+                inst_b1.s = 1;
+        else
+                inst_b1.s = 0;
+
+        inst_b1.imm20b = target25 >> 4;
+        inst.l = inst_b1.l;
+
+        paravirt_write_inst(tag, inst);
+}
+
+void __init
+__paravirt_patch_apply_branch(
+        unsigned long tag, unsigned long type,
+        const struct paravirt_patch_branch_target *entries,
+        unsigned int nr_entries)
+{
+        unsigned int i;
+        for (i = 0; i < nr_entries; i++) {
+                if (entries[i].type == type) {
+                        paravirt_patch_reloc_br(tag, entries[i].entry);
+                        break;
+                }
+        }
+}
+
+static void __init
+paravirt_patch_apply_branch(const struct paravirt_patch_site_branch *start,
+                            const struct paravirt_patch_site_branch *end)
+{
+        const struct paravirt_patch_site_branch *p;
+
+        if (noreplace_paravirt)
+                return;
+        if (pv_init_ops.patch_branch == NULL)
+                return;
+
+        for (p = start; p < end; p++)
+                (*pv_init_ops.patch_branch)(p->tag, p->type);
+
+        ia64_sync_i();
+        ia64_srlz_i();
+}
+
+void __init
+paravirt_patch_apply(void)
+{
+        extern const char __start_paravirt_bundles[];
+        extern const char __stop_paravirt_bundles[];
+        extern const char __start_paravirt_insts[];
+        extern const char __stop_paravirt_insts[];
+        extern const char __start_paravirt_branches[];
+        extern const char __stop_paravirt_branches[];
+
+        paravirt_patch_apply_bundle((const struct paravirt_patch_site_bundle *)
+                                    __start_paravirt_bundles,
+                                    (const struct paravirt_patch_site_bundle *)
+                                    __stop_paravirt_bundles);
+        paravirt_patch_apply_inst((const struct paravirt_patch_site_inst *)
+                                  __start_paravirt_insts,
+                                  (const struct paravirt_patch_site_inst *)
+                                  __stop_paravirt_insts);
+        paravirt_patch_apply_branch((const struct paravirt_patch_site_branch *)
+                                    __start_paravirt_branches,
+                                    (const struct paravirt_patch_site_branch *)
+                                    __stop_paravirt_branches);
+}
+
+/*
+ * Local variables:
+ * mode: C
+ * c-set-style: "linux"
+ * c-basic-offset: 8
+ * tab-width: 8
+ * indent-tabs-mode: t
+ * End:
+ */
diff --git a/arch/ia64/kernel/paravirt_patchlist.c b/arch/ia64/kernel/paravirt_patchlist.c
new file mode 100644
index 000000000000..b28082a95d45
--- /dev/null
+++ b/arch/ia64/kernel/paravirt_patchlist.c
@@ -0,0 +1,79 @@
+/******************************************************************************
+ * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#include <linux/bug.h>
+#include <asm/paravirt.h>
+
+#define DECLARE(name)                                           \
+        extern unsigned long                                    \
+                __ia64_native_start_gate_##name##_patchlist[];  \
+        extern unsigned long                                    \
+                __ia64_native_end_gate_##name##_patchlist[]
+
+DECLARE(fsyscall);
+DECLARE(brl_fsys_bubble_down);
+DECLARE(vtop);
+DECLARE(mckinley_e9);
+
+extern unsigned long __start_gate_section[];
+
+#define ASSIGN(name)                                                        \
+        .start_##name##_patchlist =                                         \
+                (unsigned long)__ia64_native_start_gate_##name##_patchlist, \
+        .end_##name##_patchlist =                                           \
+                (unsigned long)__ia64_native_end_gate_##name##_patchlist
+
+struct pv_patchdata pv_patchdata __initdata = {
+        ASSIGN(fsyscall),
+        ASSIGN(brl_fsys_bubble_down),
+        ASSIGN(vtop),
+        ASSIGN(mckinley_e9),
+
+        .gate_section = (void*)__start_gate_section,
+};
+
+
+unsigned long __init
+paravirt_get_gate_patchlist(enum pv_gate_patchlist type)
+{
+
+#define CASE(NAME, name)                                        \
+        case PV_GATE_START_##NAME:                              \
+                return pv_patchdata.start_##name##_patchlist;   \
+        case PV_GATE_END_##NAME:                                \
+                return pv_patchdata.end_##name##_patchlist;     \
+
+        switch (type) {
+                CASE(FSYSCALL, fsyscall);
+                CASE(BRL_FSYS_BUBBLE_DOWN, brl_fsys_bubble_down);
+                CASE(VTOP, vtop);
+                CASE(MCKINLEY_E9, mckinley_e9);
+        default:
+                BUG();
+                break;
+        }
+        return 0;
+}
+
+void * __init
+paravirt_get_gate_section(void)
+{
+        return pv_patchdata.gate_section;
+}
diff --git a/arch/ia64/kernel/paravirt_patchlist.h b/arch/ia64/kernel/paravirt_patchlist.h
new file mode 100644
index 000000000000..0684aa6c6507
--- /dev/null
+++ b/arch/ia64/kernel/paravirt_patchlist.h
@@ -0,0 +1,28 @@
+/******************************************************************************
+ * linux/arch/ia64/xen/paravirt_patchlist.h
+ *
+ * Copyright (c) 2008 Isaku Yamahata <yamahata at valinux co jp>
+ *                    VA Linux Systems Japan K.K.
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
+ *
+ */
+
+#if defined(__IA64_GATE_PARAVIRTUALIZED_XEN)
+#include <asm/xen/patchlist.h>
+#else
+#include <asm/native/patchlist.h>
+#endif
+
diff --git a/arch/ia64/kernel/paravirtentry.S b/arch/ia64/kernel/paravirtentry.S
index 2f42fcb9776a..6158560d7f17 100644
--- a/arch/ia64/kernel/paravirtentry.S
+++ b/arch/ia64/kernel/paravirtentry.S
@@ -20,8 +20,11 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/asmmacro.h>
 #include <asm/asm-offsets.h>
+#include <asm/paravirt_privop.h>
+#include <asm/paravirt_patch.h>
 #include "entry.h"
 
 #define DATA8(sym, init_value)                  \
@@ -32,29 +35,87 @@
         data8 init_value ;                      \
         .popsection
 
-#define BRANCH(targ, reg, breg)         \
-        movl reg=targ ;                 \
-        ;;                              \
-        ld8 reg=[reg] ;                 \
-        ;;                              \
-        mov breg=reg ;                  \
+#define BRANCH(targ, reg, breg, type)                                   \
+        PARAVIRT_PATCH_SITE_BR(PARAVIRT_PATCH_TYPE_BR_ ## type) ;       \
+        ;;                                                              \
+        movl reg=targ ;                                                 \
+        ;;                                                              \
+        ld8 reg=[reg] ;                                                 \
+        ;;                                                              \
+        mov breg=reg ;                                                  \
         br.cond.sptk.many breg
 
-#define BRANCH_PROC(sym, reg, breg)                             \
-        DATA8(paravirt_ ## sym ## _targ, ia64_native_ ## sym) ; \
-        GLOBAL_ENTRY(paravirt_ ## sym) ;                        \
-                BRANCH(paravirt_ ## sym ## _targ, reg, breg) ;  \
+#define BRANCH_PROC(sym, reg, breg, type)                               \
+        DATA8(paravirt_ ## sym ## _targ, ia64_native_ ## sym) ;         \
+        GLOBAL_ENTRY(paravirt_ ## sym) ;                                \
+                BRANCH(paravirt_ ## sym ## _targ, reg, breg, type) ;    \
         END(paravirt_ ## sym)
 
-#define BRANCH_PROC_UNWINFO(sym, reg, breg)                     \
-        DATA8(paravirt_ ## sym ## _targ, ia64_native_ ## sym) ; \
-        GLOBAL_ENTRY(paravirt_ ## sym) ;                        \
-                PT_REGS_UNWIND_INFO(0) ;                        \
-                BRANCH(paravirt_ ## sym ## _targ, reg, breg) ;  \
+#define BRANCH_PROC_UNWINFO(sym, reg, breg, type)                       \
+        DATA8(paravirt_ ## sym ## _targ, ia64_native_ ## sym) ;         \
+        GLOBAL_ENTRY(paravirt_ ## sym) ;                                \
+                PT_REGS_UNWIND_INFO(0) ;                                \
+                BRANCH(paravirt_ ## sym ## _targ, reg, breg, type) ;    \
         END(paravirt_ ## sym)
 
 
-BRANCH_PROC(switch_to, r22, b7)
-BRANCH_PROC_UNWINFO(leave_syscall, r22, b7)
-BRANCH_PROC(work_processed_syscall, r2, b7)
-BRANCH_PROC_UNWINFO(leave_kernel, r22, b7)
+BRANCH_PROC(switch_to, r22, b7, SWITCH_TO)
+BRANCH_PROC_UNWINFO(leave_syscall, r22, b7, LEAVE_SYSCALL)
+BRANCH_PROC(work_processed_syscall, r2, b7, WORK_PROCESSED_SYSCALL)
+BRANCH_PROC_UNWINFO(leave_kernel, r22, b7, LEAVE_KERNEL)
+
+
+#ifdef CONFIG_MODULES
+#define __INIT_OR_MODULE        .text
+#define __INITDATA_OR_MODULE    .data
+#else
+#define __INIT_OR_MODULE        __INIT
+#define __INITDATA_OR_MODULE    __INITDATA
+#endif /* CONFIG_MODULES */
+
+        __INIT_OR_MODULE
+        GLOBAL_ENTRY(paravirt_fc_i)
+        fc.i r32
+        br.ret.sptk.many rp
+        END(paravirt_fc_i)
+        __FINIT
+
+        __INIT_OR_MODULE
+        .align 32
+        GLOBAL_ENTRY(paravirt_nop_b_inst_bundle)
+        {
+                nop.b 0
+                nop.b 0
+                nop.b 0
+        }
+        END(paravirt_nop_b_inst_bundle)
+        __FINIT
+
+        /* NOTE: nop.[mfi] has same format */
+        __INIT_OR_MODULE
+        GLOBAL_ENTRY(paravirt_nop_mfi_inst_bundle)
+        {
+                nop.m 0
+                nop.f 0
+                nop.i 0
+        }
+        END(paravirt_nop_mfi_inst_bundle)
+        __FINIT
+
+        __INIT_OR_MODULE
+        GLOBAL_ENTRY(paravirt_nop_bundle)
+paravirt_nop_bundle_start:
+        {
+                nop 0
+                nop 0
+                nop 0
+        }
+paravirt_nop_bundle_end:
+        END(paravirt_nop_bundle)
+        __FINIT
+
+        __INITDATA_OR_MODULE
+        .align 8
+        .global paravirt_nop_bundle_size
+paravirt_nop_bundle_size:
+        data8   paravirt_nop_bundle_end - paravirt_nop_bundle_start
diff --git a/arch/ia64/kernel/patch.c b/arch/ia64/kernel/patch.c
index b83b2c516008..68a1311db806 100644
--- a/arch/ia64/kernel/patch.c
+++ b/arch/ia64/kernel/patch.c
@@ -7,6 +7,7 @@
 #include <linux/init.h>
 #include <linux/string.h>
 
+#include <asm/paravirt.h>
 #include <asm/patch.h>
 #include <asm/processor.h>
 #include <asm/sections.h>
@@ -169,16 +170,35 @@ ia64_patch_mckinley_e9 (unsigned long start, unsigned long end)
         ia64_srlz_i();
 }
 
+extern unsigned long ia64_native_fsyscall_table[NR_syscalls];
+extern char ia64_native_fsys_bubble_down[];
+struct pv_fsys_data pv_fsys_data __initdata = {
+        .fsyscall_table = (unsigned long *)ia64_native_fsyscall_table,
+        .fsys_bubble_down = (void *)ia64_native_fsys_bubble_down,
+};
+
+unsigned long * __init
+paravirt_get_fsyscall_table(void)
+{
+        return pv_fsys_data.fsyscall_table;
+}
+
+char * __init
+paravirt_get_fsys_bubble_down(void)
+{
+        return pv_fsys_data.fsys_bubble_down;
+}
+
 static void __init
 patch_fsyscall_table (unsigned long start, unsigned long end)
 {
-        extern unsigned long fsyscall_table[NR_syscalls];
+        u64 fsyscall_table = (u64)paravirt_get_fsyscall_table();
         s32 *offp = (s32 *) start;
         u64 ip;
 
         while (offp < (s32 *) end) {
                 ip = (u64) ia64_imva((char *) offp + *offp);
-                ia64_patch_imm64(ip, (u64) fsyscall_table);
+                ia64_patch_imm64(ip, fsyscall_table);
                 ia64_fc((void *) ip);
                 ++offp;
         }
@@ -189,7 +209,7 @@ patch_fsyscall_table (unsigned long start, unsigned long end)
 static void __init
 patch_brl_fsys_bubble_down (unsigned long start, unsigned long end)
 {
-        extern char fsys_bubble_down[];
+        u64 fsys_bubble_down = (u64)paravirt_get_fsys_bubble_down();
         s32 *offp = (s32 *) start;
         u64 ip;
 
@@ -207,13 +227,13 @@ patch_brl_fsys_bubble_down (unsigned long start, unsigned long end)
 void __init
 ia64_patch_gate (void)
 {
-#       define START(name)      ((unsigned long) __start_gate_##name##_patchlist)
-#       define END(name)        ((unsigned long)__end_gate_##name##_patchlist)
+#       define START(name)      paravirt_get_gate_patchlist(PV_GATE_START_##name)
+#       define END(name)        paravirt_get_gate_patchlist(PV_GATE_END_##name)
 
-        patch_fsyscall_table(START(fsyscall), END(fsyscall));
-        patch_brl_fsys_bubble_down(START(brl_fsys_bubble_down), END(brl_fsys_bubble_down));
-        ia64_patch_vtop(START(vtop), END(vtop));
-        ia64_patch_mckinley_e9(START(mckinley_e9), END(mckinley_e9));
+        patch_fsyscall_table(START(FSYSCALL), END(FSYSCALL));
+        patch_brl_fsys_bubble_down(START(BRL_FSYS_BUBBLE_DOWN), END(BRL_FSYS_BUBBLE_DOWN));
+        ia64_patch_vtop(START(VTOP), END(VTOP));
+        ia64_patch_mckinley_e9(START(MCKINLEY_E9), END(MCKINLEY_E9));
 }
 
 void ia64_patch_phys_stack_reg(unsigned long val)
@@ -229,7 +249,7 @@ void ia64_patch_phys_stack_reg(unsigned long val)
         while (offp < end) {
                 ip = (u64) offp + *offp;
                 ia64_patch(ip, mask, imm);
-                ia64_fc(ip);
+                ia64_fc((void *)ip);
                 ++offp;
         }
         ia64_sync_i();
diff --git a/arch/ia64/kernel/perfmon.c b/arch/ia64/kernel/perfmon.c
index 5c0f408cfd71..8a06dc480594 100644
--- a/arch/ia64/kernel/perfmon.c
+++ b/arch/ia64/kernel/perfmon.c
@@ -5603,7 +5603,7 @@ pfm_interrupt_handler(int irq, void *arg)
  * /proc/perfmon interface, for debug only
  */
 
-#define PFM_PROC_SHOW_HEADER    ((void *)NR_CPUS+1)
+#define PFM_PROC_SHOW_HEADER    ((void *)nr_cpu_ids+1)
 
 static void *
 pfm_proc_start(struct seq_file *m, loff_t *pos)
@@ -5612,7 +5612,7 @@ pfm_proc_start(struct seq_file *m, loff_t *pos)
                 return PFM_PROC_SHOW_HEADER;
         }
 
-        while (*pos <= NR_CPUS) {
+        while (*pos <= nr_cpu_ids) {
                 if (cpu_online(*pos - 1)) {
                         return (void *)*pos;
                 }
diff --git a/arch/ia64/kernel/process.c b/arch/ia64/kernel/process.c
index c57162705147..5d7c0e5b9e76 100644
--- a/arch/ia64/kernel/process.c
+++ b/arch/ia64/kernel/process.c
@@ -413,7 +413,7 @@ ia64_load_extra (struct task_struct *task)
  * so there is nothing to worry about.
  */
 int
-copy_thread (int nr, unsigned long clone_flags,
+copy_thread(unsigned long clone_flags,
              unsigned long user_stack_base, unsigned long user_stack_size,
              struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/ia64/kernel/salinfo.c b/arch/ia64/kernel/salinfo.c
index ecb9eb78d687..7053c55b7649 100644
--- a/arch/ia64/kernel/salinfo.c
+++ b/arch/ia64/kernel/salinfo.c
@@ -317,7 +317,7 @@ retry:
         }
 
         n = data->cpu_check;
-        for (i = 0; i < NR_CPUS; i++) {
+        for (i = 0; i < nr_cpu_ids; i++) {
                 if (cpu_isset(n, data->cpu_event)) {
                         if (!cpu_online(n)) {
                                 cpu_clear(n, data->cpu_event);
@@ -326,7 +326,7 @@ retry:
                         cpu = n;
                         break;
                 }
-                if (++n == NR_CPUS)
+                if (++n == nr_cpu_ids)
                         n = 0;
         }
 
@@ -337,7 +337,7 @@ retry:
 
         /* for next read, start checking at next CPU */
         data->cpu_check = cpu;
-        if (++data->cpu_check == NR_CPUS)
+        if (++data->cpu_check == nr_cpu_ids)
                 data->cpu_check = 0;
 
         snprintf(cmd, sizeof(cmd), "read %d\n", cpu);
diff --git a/arch/ia64/kernel/setup.c b/arch/ia64/kernel/setup.c
index 865af27c7737..714066aeda7f 100644
--- a/arch/ia64/kernel/setup.c
+++ b/arch/ia64/kernel/setup.c
@@ -52,6 +52,7 @@
 #include <asm/meminit.h>
 #include <asm/page.h>
 #include <asm/paravirt.h>
+#include <asm/paravirt_patch.h>
 #include <asm/patch.h>
 #include <asm/pgtable.h>
 #include <asm/processor.h>
@@ -537,6 +538,7 @@ setup_arch (char **cmdline_p)
         paravirt_arch_setup_early();
 
         ia64_patch_vtop((u64) __start___vtop_patchlist, (u64) __end___vtop_patchlist);
+        paravirt_patch_apply();
 
         *cmdline_p = __va(ia64_boot_param->command_line);
         strlcpy(boot_command_line, *cmdline_p, COMMAND_LINE_SIZE);
@@ -730,10 +732,10 @@ static void *
 c_start (struct seq_file *m, loff_t *pos)
 {
 #ifdef CONFIG_SMP
-        while (*pos < NR_CPUS && !cpu_isset(*pos, cpu_online_map))
+        while (*pos < nr_cpu_ids && !cpu_online(*pos))
                 ++*pos;
 #endif
-        return *pos < NR_CPUS ? cpu_data(*pos) : NULL;
+        return *pos < nr_cpu_ids ? cpu_data(*pos) : NULL;
 }
 
 static void *
@@ -1016,8 +1018,7 @@ cpu_init (void)
                                         | IA64_DCR_DA | IA64_DCR_DD | IA64_DCR_LC));
         atomic_inc(&init_mm.mm_count);
         current->active_mm = &init_mm;
-        if (current->mm)
-                BUG();
+        BUG_ON(current->mm);
 
         ia64_mmu_init(ia64_imva(cpu_data));
         ia64_mca_cpu_init(ia64_imva(cpu_data));
diff --git a/arch/ia64/kernel/smp.c b/arch/ia64/kernel/smp.c
index da8f020d82c1..2ea4199d9c57 100644
--- a/arch/ia64/kernel/smp.c
+++ b/arch/ia64/kernel/smp.c
@@ -166,11 +166,11 @@ send_IPI_allbutself (int op)
  * Called with preemption disabled.
  */
 static inline void
-send_IPI_mask(cpumask_t mask, int op)
+send_IPI_mask(const struct cpumask *mask, int op)
 {
         unsigned int cpu;
 
-        for_each_cpu_mask(cpu, mask) {
+        for_each_cpu(cpu, mask) {
                         send_IPI_single(cpu, op);
         }
 }
@@ -316,7 +316,7 @@ void arch_send_call_function_single_ipi(int cpu)
         send_IPI_single(cpu, IPI_CALL_FUNC_SINGLE);
 }
 
-void arch_send_call_function_ipi(cpumask_t mask)
+void arch_send_call_function_ipi_mask(const struct cpumask *mask)
 {
         send_IPI_mask(mask, IPI_CALL_FUNC);
 }
diff --git a/arch/ia64/kernel/smpboot.c b/arch/ia64/kernel/smpboot.c
index 52290547c85b..7700e23034bb 100644
--- a/arch/ia64/kernel/smpboot.c
+++ b/arch/ia64/kernel/smpboot.c
@@ -581,14 +581,14 @@ smp_build_cpu_map (void)
 
         ia64_cpu_to_sapicid[0] = boot_cpu_id;
         cpus_clear(cpu_present_map);
-        cpu_set(0, cpu_present_map);
-        cpu_set(0, cpu_possible_map);
+        set_cpu_present(0, true);
+        set_cpu_possible(0, true);
         for (cpu = 1, i = 0; i < smp_boot_data.cpu_count; i++) {
                 sapicid = smp_boot_data.cpu_phys_id[i];
                 if (sapicid == boot_cpu_id)
                         continue;
-                cpu_set(cpu, cpu_present_map);
-                cpu_set(cpu, cpu_possible_map);
+                set_cpu_present(cpu, true);
+                set_cpu_possible(cpu, true);
                 ia64_cpu_to_sapicid[cpu] = sapicid;
                 cpu++;
         }
@@ -626,12 +626,9 @@ smp_prepare_cpus (unsigned int max_cpus)
          */
         if (!max_cpus) {
                 printk(KERN_INFO "SMP mode deactivated.\n");
-                cpus_clear(cpu_online_map);
-                cpus_clear(cpu_present_map);
-                cpus_clear(cpu_possible_map);
-                cpu_set(0, cpu_online_map);
-                cpu_set(0, cpu_present_map);
-                cpu_set(0, cpu_possible_map);
+                init_cpu_online(cpumask_of(0));
+                init_cpu_present(cpumask_of(0));
+                init_cpu_possible(cpumask_of(0));
                 return;
         }
 }
diff --git a/arch/ia64/kernel/time.c b/arch/ia64/kernel/time.c
index d6747bae52d8..641c8b61c4f1 100644
--- a/arch/ia64/kernel/time.c
+++ b/arch/ia64/kernel/time.c
@@ -51,6 +51,15 @@ EXPORT_SYMBOL(last_cli_ip);
 #endif
 
 #ifdef CONFIG_PARAVIRT
+/* We need to define a real function for sched_clock, to override the
+   weak default version */
+unsigned long long sched_clock(void)
+{
+        return paravirt_sched_clock();
+}
+#endif
+
+#ifdef CONFIG_PARAVIRT
 static void
 paravirt_clocksource_resume(void)
 {
diff --git a/arch/ia64/kernel/vmlinux.lds.S b/arch/ia64/kernel/vmlinux.lds.S
index 3765efc5f963..4a95e86b9ac2 100644
--- a/arch/ia64/kernel/vmlinux.lds.S
+++ b/arch/ia64/kernel/vmlinux.lds.S
@@ -169,6 +169,30 @@ SECTIONS
           __end___mckinley_e9_bundles = .;
         }
 
+#if defined(CONFIG_PARAVIRT)
+  . = ALIGN(16);
+  .paravirt_bundles : AT(ADDR(.paravirt_bundles) - LOAD_OFFSET)
+        {
+          __start_paravirt_bundles = .;
+          *(.paravirt_bundles)
+          __stop_paravirt_bundles = .;
+        }
+  . = ALIGN(16);
+  .paravirt_insts : AT(ADDR(.paravirt_insts) - LOAD_OFFSET)
+        {
+          __start_paravirt_insts = .;
+          *(.paravirt_insts)
+          __stop_paravirt_insts = .;
+        }
+  . = ALIGN(16);
+  .paravirt_branches : AT(ADDR(.paravirt_branches) - LOAD_OFFSET)
+        {
+          __start_paravirt_branches = .;
+          *(.paravirt_branches)
+          __stop_paravirt_branches = .;
+        }
+#endif
+
 #if defined(CONFIG_IA64_GENERIC)
   /* Machine Vector */
   . = ALIGN(16);
@@ -201,6 +225,12 @@ SECTIONS
           __start_gate_section = .;
           *(.data.gate)
           __stop_gate_section = .;
+#ifdef CONFIG_XEN
+          . = ALIGN(PAGE_SIZE);
+          __xen_start_gate_section = .;
+          *(.data.gate.xen)
+          __xen_stop_gate_section = .;
+#endif
         }
   . = ALIGN(PAGE_SIZE);         /* make sure the gate page doesn't expose
                                  * kernel data
diff --git a/arch/ia64/kvm/kvm-ia64.c b/arch/ia64/kvm/kvm-ia64.c
index 076b00d1dbff..28af6a731bb8 100644
--- a/arch/ia64/kvm/kvm-ia64.c
+++ b/arch/ia64/kvm/kvm-ia64.c
@@ -70,7 +70,7 @@ static void kvm_flush_icache(unsigned long start, unsigned long len)
         int l;
 
         for (l = 0; l < (len + 32); l += 32)
-                ia64_fc(start + l);
+                ia64_fc((void *)(start + l));
 
         ia64_sync_i();
         ia64_srlz_i();
diff --git a/arch/ia64/kvm/vcpu.c b/arch/ia64/kvm/vcpu.c
index d4d280505878..a18ee17b9192 100644
--- a/arch/ia64/kvm/vcpu.c
+++ b/arch/ia64/kvm/vcpu.c
@@ -386,7 +386,7 @@ void set_rse_reg(struct kvm_pt_regs *regs, unsigned long r1,
                 else
                         *rnat_addr = (*rnat_addr) & (~nat_mask);
 
-                ia64_setreg(_IA64_REG_AR_BSPSTORE, bspstore);
+                ia64_setreg(_IA64_REG_AR_BSPSTORE, (unsigned long)bspstore);
                 ia64_setreg(_IA64_REG_AR_RNAT, rnat);
         }
         local_irq_restore(psr);
diff --git a/arch/ia64/kvm/vtlb.c b/arch/ia64/kvm/vtlb.c
index 38232b37668b..2c2501f13159 100644
--- a/arch/ia64/kvm/vtlb.c
+++ b/arch/ia64/kvm/vtlb.c
@@ -210,6 +210,7 @@ void thash_vhpt_insert(struct kvm_vcpu *v, u64 pte, u64 itir, u64 va, int type)
                 phy_pte  &= ~PAGE_FLAGS_RV_MASK;
                 psr = ia64_clear_ic();
                 ia64_itc(type, va, phy_pte, itir_ps(itir));
+                paravirt_dv_serialize_data();
                 ia64_set_psr(psr);
         }
 
@@ -456,6 +457,7 @@ void  thash_purge_and_insert(struct kvm_vcpu *v, u64 pte, u64 itir,
                 phy_pte  &= ~PAGE_FLAGS_RV_MASK;
                 psr = ia64_clear_ic();
                 ia64_itc(type, ifa, phy_pte, ps);
+                paravirt_dv_serialize_data();
                 ia64_set_psr(psr);
         }
         if (!(pte&VTLB_PTE_IO))
diff --git a/arch/ia64/mm/init.c b/arch/ia64/mm/init.c
index 56e12903973c..c0f3bee69042 100644
--- a/arch/ia64/mm/init.c
+++ b/arch/ia64/mm/init.c
@@ -35,6 +35,7 @@
 #include <asm/uaccess.h>
 #include <asm/unistd.h>
 #include <asm/mca.h>
+#include <asm/paravirt.h>
 
 DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 
@@ -259,6 +260,7 @@ put_kernel_page (struct page *page, unsigned long address, pgprot_t pgprot)
 static void __init
 setup_gate (void)
 {
+        void *gate_section;
         struct page *page;
 
         /*
@@ -266,10 +268,11 @@ setup_gate (void)
          * headers etc. and once execute-only page to enable
          * privilege-promotion via "epc":
          */
-        page = virt_to_page(ia64_imva(__start_gate_section));
+        gate_section = paravirt_get_gate_section();
+        page = virt_to_page(ia64_imva(gate_section));
         put_kernel_page(page, GATE_ADDR, PAGE_READONLY);
 #ifdef HAVE_BUGGY_SEGREL
-        page = virt_to_page(ia64_imva(__start_gate_section + PAGE_SIZE));
+        page = virt_to_page(ia64_imva(gate_section + PAGE_SIZE));
         put_kernel_page(page, GATE_ADDR + PAGE_SIZE, PAGE_GATE);
 #else
         put_kernel_page(page, GATE_ADDR + PERCPU_PAGE_SIZE, PAGE_GATE);
@@ -633,8 +636,7 @@ mem_init (void)
 #endif
 
 #ifdef CONFIG_FLATMEM
-        if (!mem_map)
-                BUG();
+        BUG_ON(!mem_map);
         max_mapnr = max_low_pfn;
 #endif
 
@@ -667,8 +669,8 @@ mem_init (void)
          * code can tell them apart.
          */
         for (i = 0; i < NR_syscalls; ++i) {
-                extern unsigned long fsyscall_table[NR_syscalls];
                 extern unsigned long sys_call_table[NR_syscalls];
+                unsigned long *fsyscall_table = paravirt_get_fsyscall_table();
 
                 if (!fsyscall_table[i] || nolwsys)
                         fsyscall_table[i] = sys_call_table[i] | 1;
diff --git a/arch/ia64/mm/tlb.c b/arch/ia64/mm/tlb.c
index bd9818a36b47..b9f3d7bbb338 100644
--- a/arch/ia64/mm/tlb.c
+++ b/arch/ia64/mm/tlb.c
@@ -309,7 +309,7 @@ flush_tlb_range (struct vm_area_struct *vma, unsigned long start,
 
         preempt_disable();
 #ifdef CONFIG_SMP
-        if (mm != current->active_mm || cpus_weight(mm->cpu_vm_mask) != 1) {
+        if (mm != current->active_mm || cpumask_weight(mm_cpumask(mm)) != 1) {
                 platform_global_tlb_purge(mm, start, end, nbits);
                 preempt_enable();
                 return;
diff --git a/arch/ia64/scripts/pvcheck.sed b/arch/ia64/scripts/pvcheck.sed
index ba66ac2e4c60..e59809a3fc01 100644
--- a/arch/ia64/scripts/pvcheck.sed
+++ b/arch/ia64/scripts/pvcheck.sed
@@ -17,6 +17,7 @@ s/mov.*=.*cr\.iip/.warning \"cr.iip should not used directly\"/g
 s/mov.*=.*cr\.ivr/.warning \"cr.ivr should not used directly\"/g
 s/mov.*=[^\.]*psr/.warning \"psr should not used directly\"/g   # avoid ar.fpsr
 s/mov.*=.*ar\.eflags/.warning \"ar.eflags should not used directly\"/g
+s/mov.*=.*ar\.itc.*/.warning \"ar.itc should not used directly\"/g
 s/mov.*cr\.ifa.*=.*/.warning \"cr.ifa should not used directly\"/g
 s/mov.*cr\.itir.*=.*/.warning \"cr.itir should not used directly\"/g
 s/mov.*cr\.iha.*=.*/.warning \"cr.iha should not used directly\"/g
diff --git a/arch/ia64/sn/kernel/io_common.c b/arch/ia64/sn/kernel/io_common.c
index 0d4ffa4da1da..57f280dd9def 100644
--- a/arch/ia64/sn/kernel/io_common.c
+++ b/arch/ia64/sn/kernel/io_common.c
@@ -135,8 +135,7 @@ static s64 sn_device_fixup_war(u64 nasid, u64 widget, int device,
         }
 
         war_list = kzalloc(DEV_PER_WIDGET * sizeof(*war_list), GFP_KERNEL);
-        if (!war_list)
-                BUG();
+        BUG_ON(!war_list);
 
         SAL_CALL_NOLOCK(isrv, SN_SAL_IOIF_GET_WIDGET_DMAFLUSH_LIST,
                         nasid, widget, __pa(war_list), 0, 0, 0 ,0);
@@ -180,23 +179,20 @@ sn_common_hubdev_init(struct hubdev_info *hubdev)
                 sizeof(struct sn_flush_device_kernel *);
         hubdev->hdi_flush_nasid_list.widget_p =
                 kzalloc(size, GFP_KERNEL);
-        if (!hubdev->hdi_flush_nasid_list.widget_p)
-                BUG();
+        BUG_ON(!hubdev->hdi_flush_nasid_list.widget_p);
 
         for (widget = 0; widget <= HUB_WIDGET_ID_MAX; widget++) {
                 size = DEV_PER_WIDGET *
                         sizeof(struct sn_flush_device_kernel);
                 sn_flush_device_kernel = kzalloc(size, GFP_KERNEL);
-                if (!sn_flush_device_kernel)
-                        BUG();
+                BUG_ON(!sn_flush_device_kernel);
 
                 dev_entry = sn_flush_device_kernel;
                 for (device = 0; device < DEV_PER_WIDGET;
                      device++, dev_entry++) {
                         size = sizeof(struct sn_flush_device_common);
                         dev_entry->common = kzalloc(size, GFP_KERNEL);
-                        if (!dev_entry->common)
-                                BUG();
+                        BUG_ON(!dev_entry->common);
                         if (sn_prom_feature_available(PRF_DEVICE_FLUSH_LIST))
                                 status = sal_get_device_dmaflush_list(
                                              hubdev->hdi_nasid, widget, device,
@@ -326,8 +322,7 @@ sn_common_bus_fixup(struct pci_bus *bus,
          */
         controller->platform_data = kzalloc(sizeof(struct sn_platform_data),
                                             GFP_KERNEL);
-        if (controller->platform_data == NULL)
-                BUG();
+        BUG_ON(controller->platform_data == NULL);
         sn_platform_data =
                         (struct sn_platform_data *) controller->platform_data;
         sn_platform_data->provider_soft = provider_soft;
diff --git a/arch/ia64/sn/kernel/io_init.c b/arch/ia64/sn/kernel/io_init.c
index e2eb2da60f96..ee774c366a06 100644
--- a/arch/ia64/sn/kernel/io_init.c
+++ b/arch/ia64/sn/kernel/io_init.c
@@ -128,8 +128,7 @@ sn_legacy_pci_window_fixup(struct pci_controller *controller,
 {
                 controller->window = kcalloc(2, sizeof(struct pci_window),
                                              GFP_KERNEL);
-                if (controller->window == NULL)
-                        BUG();
+                BUG_ON(controller->window == NULL);
                 controller->window[0].offset = legacy_io;
                 controller->window[0].resource.name = "legacy_io";
                 controller->window[0].resource.flags = IORESOURCE_IO;
@@ -168,8 +167,7 @@ sn_pci_window_fixup(struct pci_dev *dev, unsigned int count,
         idx = controller->windows;
         new_count = controller->windows + count;
         new_window = kcalloc(new_count, sizeof(struct pci_window), GFP_KERNEL);
-        if (new_window == NULL)
-                BUG();
+        BUG_ON(new_window == NULL);
         if (controller->window) {
                 memcpy(new_window, controller->window,
                        sizeof(struct pci_window) * controller->windows);
@@ -222,8 +220,7 @@ sn_io_slot_fixup(struct pci_dev *dev)
                 (u64) __pa(pcidev_info),
                 (u64) __pa(sn_irq_info));
 
-        if (status)
-                BUG(); /* Cannot get platform pci device information */
+        BUG_ON(status); /* Cannot get platform pci device information */
 
 
         /* Copy over PIO Mapped Addresses */
@@ -307,8 +304,7 @@ sn_pci_controller_fixup(int segment, int busnum, struct pci_bus *bus)
         prom_bussoft_ptr = __va(prom_bussoft_ptr);
 
         controller = kzalloc(sizeof(*controller), GFP_KERNEL);
-        if (!controller)
-                BUG();
+        BUG_ON(!controller);
         controller->segment = segment;
 
         /*
diff --git a/arch/ia64/sn/kernel/setup.c b/arch/ia64/sn/kernel/setup.c
index 02c5b8a9fb60..e456f062f241 100644
--- a/arch/ia64/sn/kernel/setup.c
+++ b/arch/ia64/sn/kernel/setup.c
@@ -732,8 +732,7 @@ void __init build_cnode_tables(void)
                 kl_config_hdr_t *klgraph_header;
                 nasid = cnodeid_to_nasid(node);
                 klgraph_header = ia64_sn_get_klconfig_addr(nasid);
-                if (klgraph_header == NULL)
-                        BUG();
+                BUG_ON(klgraph_header == NULL);
                 brd = NODE_OFFSET_TO_LBOARD(nasid, klgraph_header->ch_board_info);
                 while (brd) {
                         if (board_needs_cnode(brd->brd_type) && physical_node_map[brd->brd_nasid] < 0) {
@@ -750,7 +749,7 @@ nasid_slice_to_cpuid(int nasid, int slice)
 {
         long cpu;
 
-        for (cpu = 0; cpu < NR_CPUS; cpu++)
+        for (cpu = 0; cpu < nr_cpu_ids; cpu++)
                 if (cpuid_to_nasid(cpu) == nasid &&
                                         cpuid_to_slice(cpu) == slice)
                         return cpu;
diff --git a/arch/ia64/sn/kernel/sn2/sn2_smp.c b/arch/ia64/sn/kernel/sn2/sn2_smp.c
index e585f9a2afb9..1176506b2bae 100644
--- a/arch/ia64/sn/kernel/sn2/sn2_smp.c
+++ b/arch/ia64/sn/kernel/sn2/sn2_smp.c
@@ -133,7 +133,7 @@ sn2_ipi_flush_all_tlb(struct mm_struct *mm)
         unsigned long itc;
 
         itc = ia64_get_itc();
-        smp_flush_tlb_cpumask(mm->cpu_vm_mask);
+        smp_flush_tlb_cpumask(*mm_cpumask(mm));
         itc = ia64_get_itc() - itc;
         __get_cpu_var(ptcstats).shub_ipi_flushes_itc_clocks += itc;
         __get_cpu_var(ptcstats).shub_ipi_flushes++;
@@ -182,7 +182,7 @@ sn2_global_tlb_purge(struct mm_struct *mm, unsigned long start,
         nodes_clear(nodes_flushed);
         i = 0;
 
-        for_each_cpu_mask(cpu, mm->cpu_vm_mask) {
+        for_each_cpu(cpu, mm_cpumask(mm)) {
                 cnode = cpu_to_node(cpu);
                 node_set(cnode, nodes_flushed);
                 lcpu = cpu;
@@ -461,7 +461,7 @@ bool sn_cpu_disable_allowed(int cpu)
 
 static void *sn2_ptc_seq_start(struct seq_file *file, loff_t * offset)
 {
-        if (*offset < NR_CPUS)
+        if (*offset < nr_cpu_ids)
                 return offset;
         return NULL;
 }
@@ -469,7 +469,7 @@ static void *sn2_ptc_seq_start(struct seq_file *file, loff_t * offset)
 static void *sn2_ptc_seq_next(struct seq_file *file, void *data, loff_t * offset)
 {
         (*offset)++;
-        if (*offset < NR_CPUS)
+        if (*offset < nr_cpu_ids)
                 return offset;
         return NULL;
 }
@@ -491,7 +491,7 @@ static int sn2_ptc_seq_show(struct seq_file *file, void *data)
                 seq_printf(file, "# ptctest %d, flushopt %d\n", sn2_ptctest, sn2_flush_opt);
         }
 
-        if (cpu < NR_CPUS && cpu_online(cpu)) {
+        if (cpu < nr_cpu_ids && cpu_online(cpu)) {
                 stat = &per_cpu(ptcstats, cpu);
                 seq_printf(file, "cpu %d %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld %ld\n", cpu, stat->ptc_l,
                                 stat->change_rid, stat->shub_ptc_flushes, stat->nodes_flushed,
@@ -554,7 +554,7 @@ static int __init sn2_ptc_init(void)
 
         proc_sn2_ptc = proc_create(PTC_BASENAME, 0444,
                                    NULL, &proc_sn2_ptc_operations);
-        if (!&proc_sn2_ptc_operations) {
+        if (!proc_sn2_ptc) {
                 printk(KERN_ERR "unable to create %s proc entry", PTC_BASENAME);
                 return -EINVAL;
         }
diff --git a/arch/ia64/sn/kernel/sn2/sn_hwperf.c b/arch/ia64/sn/kernel/sn2/sn_hwperf.c
index be339477f906..9e6491cf72bd 100644
--- a/arch/ia64/sn/kernel/sn2/sn_hwperf.c
+++ b/arch/ia64/sn/kernel/sn2/sn_hwperf.c
@@ -275,8 +275,7 @@ static int sn_hwperf_get_nearest_node_objdata(struct sn_hwperf_object_info *objb
 
         /* get it's interconnect topology */
         sz = op->ports * sizeof(struct sn_hwperf_port_info);
-        if (sz > sizeof(ptdata))
-                BUG();
+        BUG_ON(sz > sizeof(ptdata));
         e = ia64_sn_hwperf_op(sn_hwperf_master_nasid,
                               SN_HWPERF_ENUM_PORTS, nodeobj->id, sz,
                               (u64)&ptdata, 0, 0, NULL);
@@ -310,8 +309,7 @@ static int sn_hwperf_get_nearest_node_objdata(struct sn_hwperf_object_info *objb
         if (router && (!found_cpu || !found_mem)) {
                 /* search for a node connected to the same router */
                 sz = router->ports * sizeof(struct sn_hwperf_port_info);
-                if (sz > sizeof(ptdata))
-                        BUG();
+                BUG_ON(sz > sizeof(ptdata));
                 e = ia64_sn_hwperf_op(sn_hwperf_master_nasid,
                                       SN_HWPERF_ENUM_PORTS, router->id, sz,
                                       (u64)&ptdata, 0, 0, NULL);
@@ -612,7 +610,7 @@ static int sn_hwperf_op_cpu(struct sn_hwperf_op_info *op_info)
         op_info->a->arg &= SN_HWPERF_ARG_OBJID_MASK;
 
         if (cpu != SN_HWPERF_ARG_ANY_CPU) {
-                if (cpu >= NR_CPUS || !cpu_online(cpu)) {
+                if (cpu >= nr_cpu_ids || !cpu_online(cpu)) {
                         r = -EINVAL;
                         goto out;
                 }
diff --git a/arch/ia64/sn/pci/pcibr/pcibr_dma.c b/arch/ia64/sn/pci/pcibr/pcibr_dma.c
index 060df4aa9916..c659ad5613a0 100644
--- a/arch/ia64/sn/pci/pcibr/pcibr_dma.c
+++ b/arch/ia64/sn/pci/pcibr/pcibr_dma.c
@@ -256,9 +256,7 @@ void sn_dma_flush(u64 addr)
 
         hubinfo = (NODEPDA(nasid_to_cnodeid(nasid)))->pdinfo;
 
-        if (!hubinfo) {
-                BUG();
-        }
+        BUG_ON(!hubinfo);
 
         flush_nasid_list = &hubinfo->hdi_flush_nasid_list;
         if (flush_nasid_list->widget_p == NULL)
diff --git a/arch/ia64/xen/Makefile b/arch/ia64/xen/Makefile
index 0ad0224693d9..e6f4a0a74228 100644
--- a/arch/ia64/xen/Makefile
+++ b/arch/ia64/xen/Makefile
@@ -3,14 +3,29 @@
 #
 
 obj-y := hypercall.o xenivt.o xensetup.o xen_pv_ops.o irq_xen.o \
-         hypervisor.o xencomm.o xcom_hcall.o grant-table.o time.o suspend.o
+         hypervisor.o xencomm.o xcom_hcall.o grant-table.o time.o suspend.o \
+         gate-data.o
 
 obj-$(CONFIG_IA64_GENERIC) += machvec.o
 
+# The gate DSO image is built using a special linker script.
+include $(srctree)/arch/ia64/kernel/Makefile.gate
+
+# tell compiled for xen
+CPPFLAGS_gate.lds += -D__IA64_GATE_PARAVIRTUALIZED_XEN
+AFLAGS_gate.o += -D__IA64_ASM_PARAVIRTUALIZED_XEN -D__IA64_GATE_PARAVIRTUALIZED_XEN
+
+# use same file of native.
+$(obj)/gate.o: $(src)/../kernel/gate.S FORCE
+        $(call if_changed_dep,as_o_S)
+$(obj)/gate.lds: $(src)/../kernel/gate.lds.S FORCE
+        $(call if_changed_dep,cpp_lds_S)
+
+
 AFLAGS_xenivt.o += -D__IA64_ASM_PARAVIRTUALIZED_XEN
 
 # xen multi compile
-ASM_PARAVIRT_MULTI_COMPILE_SRCS = ivt.S entry.S
+ASM_PARAVIRT_MULTI_COMPILE_SRCS = ivt.S entry.S fsys.S
 ASM_PARAVIRT_OBJS = $(addprefix xen-,$(ASM_PARAVIRT_MULTI_COMPILE_SRCS:.S=.o))
 obj-y += $(ASM_PARAVIRT_OBJS)
 define paravirtualized_xen
diff --git a/arch/ia64/xen/gate-data.S b/arch/ia64/xen/gate-data.S
new file mode 100644
index 000000000000..7d4830afc91d
--- /dev/null
+++ b/arch/ia64/xen/gate-data.S
@@ -0,0 +1,3 @@
+        .section .data.gate.xen, "aw"
+
+        .incbin "arch/ia64/xen/gate.so"
diff --git a/arch/ia64/xen/hypercall.S b/arch/ia64/xen/hypercall.S
index 45e02bb64a92..e32dae444dd6 100644
--- a/arch/ia64/xen/hypercall.S
+++ b/arch/ia64/xen/hypercall.S
@@ -9,6 +9,7 @@
 #include <asm/intrinsics.h>
 #include <asm/xen/privop.h>
 
+#ifdef __INTEL_COMPILER
 /*
  * Hypercalls without parameter.
  */
@@ -72,6 +73,7 @@ GLOBAL_ENTRY(xen_set_rr0_to_rr4)
         br.ret.sptk.many rp
         ;;
 END(xen_set_rr0_to_rr4)
+#endif
 
 GLOBAL_ENTRY(xen_send_ipi)
         mov r14=r32
diff --git a/arch/ia64/xen/time.c b/arch/ia64/xen/time.c
index 68d6204c3f16..fb8332690179 100644
--- a/arch/ia64/xen/time.c
+++ b/arch/ia64/xen/time.c
@@ -175,10 +175,58 @@ static void xen_itc_jitter_data_reset(void)
         } while (unlikely(ret != lcycle));
 }
 
+/* based on xen_sched_clock() in arch/x86/xen/time.c. */
+/*
+ * This relies on HAVE_UNSTABLE_SCHED_CLOCK. If it can't be defined,
+ * something similar logic should be implemented here.
+ */
+/*
+ * Xen sched_clock implementation.  Returns the number of unstolen
+ * nanoseconds, which is nanoseconds the VCPU spent in RUNNING+BLOCKED
+ * states.
+ */
+static unsigned long long xen_sched_clock(void)
+{
+        struct vcpu_runstate_info runstate;
+
+        unsigned long long now;
+        unsigned long long offset;
+        unsigned long long ret;
+
+        /*
+         * Ideally sched_clock should be called on a per-cpu basis
+         * anyway, so preempt should already be disabled, but that's
+         * not current practice at the moment.
+         */
+        preempt_disable();
+
+        /*
+         * both ia64_native_sched_clock() and xen's runstate are
+         * based on mAR.ITC. So difference of them makes sense.
+         */
+        now = ia64_native_sched_clock();
+
+        get_runstate_snapshot(&runstate);
+
+        WARN_ON(runstate.state != RUNSTATE_running);
+
+        offset = 0;
+        if (now > runstate.state_entry_time)
+                offset = now - runstate.state_entry_time;
+        ret = runstate.time[RUNSTATE_blocked] +
+                runstate.time[RUNSTATE_running] +
+                offset;
+
+        preempt_enable();
+
+        return ret;
+}
+
 struct pv_time_ops xen_time_ops __initdata = {
         .init_missing_ticks_accounting  = xen_init_missing_ticks_accounting,
         .do_steal_accounting            = xen_do_steal_accounting,
         .clocksource_resume             = xen_itc_jitter_data_reset,
+        .sched_clock                    = xen_sched_clock,
 };
 
 /* Called after suspend, to resume time.  */
diff --git a/arch/ia64/xen/xen_pv_ops.c b/arch/ia64/xen/xen_pv_ops.c
index 936cff3c96e0..5e2270a999fa 100644
--- a/arch/ia64/xen/xen_pv_ops.c
+++ b/arch/ia64/xen/xen_pv_ops.c
@@ -24,6 +24,7 @@
 #include <linux/irq.h>
 #include <linux/kernel.h>
 #include <linux/pm.h>
+#include <linux/unistd.h>
 
 #include <asm/xen/hypervisor.h>
 #include <asm/xen/xencomm.h>
@@ -153,6 +154,13 @@ xen_post_smp_prepare_boot_cpu(void)
         xen_setup_vcpu_info_placement();
 }
 
+#ifdef ASM_SUPPORTED
+static unsigned long __init_or_module
+xen_patch_bundle(void *sbundle, void *ebundle, unsigned long type);
+#endif
+static void __init
+xen_patch_branch(unsigned long tag, unsigned long type);
+
 static const struct pv_init_ops xen_init_ops __initconst = {
         .banner = xen_banner,
 
@@ -163,6 +171,53 @@ static const struct pv_init_ops xen_init_ops __initconst = {
         .arch_setup_nomca = xen_arch_setup_nomca,
 
         .post_smp_prepare_boot_cpu = xen_post_smp_prepare_boot_cpu,
+#ifdef ASM_SUPPORTED
+        .patch_bundle = xen_patch_bundle,
+#endif
+        .patch_branch = xen_patch_branch,
+};
+
+/***************************************************************************
+ * pv_fsys_data
+ * addresses for fsys
+ */
+
+extern unsigned long xen_fsyscall_table[NR_syscalls];
+extern char xen_fsys_bubble_down[];
+struct pv_fsys_data xen_fsys_data __initdata = {
+        .fsyscall_table = (unsigned long *)xen_fsyscall_table,
+        .fsys_bubble_down = (void *)xen_fsys_bubble_down,
+};
+
+/***************************************************************************
+ * pv_patchdata
+ * patchdata addresses
+ */
+
+#define DECLARE(name)                                                   \
+        extern unsigned long __xen_start_gate_##name##_patchlist[];     \
+        extern unsigned long __xen_end_gate_##name##_patchlist[]
+
+DECLARE(fsyscall);
+DECLARE(brl_fsys_bubble_down);
+DECLARE(vtop);
+DECLARE(mckinley_e9);
+
+extern unsigned long __xen_start_gate_section[];
+
+#define ASSIGN(name)                                                    \
+        .start_##name##_patchlist =                                     \
+                (unsigned long)__xen_start_gate_##name##_patchlist,     \
+        .end_##name##_patchlist =                                       \
+                (unsigned long)__xen_end_gate_##name##_patchlist
+
+static struct pv_patchdata xen_patchdata __initdata = {
+        ASSIGN(fsyscall),
+        ASSIGN(brl_fsys_bubble_down),
+        ASSIGN(vtop),
+        ASSIGN(mckinley_e9),
+
+        .gate_section = (void*)__xen_start_gate_section,
 };
 
 /***************************************************************************
@@ -170,6 +225,76 @@ static const struct pv_init_ops xen_init_ops __initconst = {
  * intrinsics hooks.
  */
 
+#ifndef ASM_SUPPORTED
+static void
+xen_set_itm_with_offset(unsigned long val)
+{
+        /* ia64_cpu_local_tick() calls this with interrupt enabled. */
+        /* WARN_ON(!irqs_disabled()); */
+        xen_set_itm(val - XEN_MAPPEDREGS->itc_offset);
+}
+
+static unsigned long
+xen_get_itm_with_offset(void)
+{
+        /* unused at this moment */
+        printk(KERN_DEBUG "%s is called.\n", __func__);
+
+        WARN_ON(!irqs_disabled());
+        return ia64_native_getreg(_IA64_REG_CR_ITM) +
+                XEN_MAPPEDREGS->itc_offset;
+}
+
+/* ia64_set_itc() is only called by
+ * cpu_init() with ia64_set_itc(0) and ia64_sync_itc().
+ * So XEN_MAPPEDRESG->itc_offset cal be considered as almost constant.
+ */
+static void
+xen_set_itc(unsigned long val)
+{
+        unsigned long mitc;
+
+        WARN_ON(!irqs_disabled());
+        mitc = ia64_native_getreg(_IA64_REG_AR_ITC);
+        XEN_MAPPEDREGS->itc_offset = val - mitc;
+        XEN_MAPPEDREGS->itc_last = val;
+}
+
+static unsigned long
+xen_get_itc(void)
+{
+        unsigned long res;
+        unsigned long itc_offset;
+        unsigned long itc_last;
+        unsigned long ret_itc_last;
+
+        itc_offset = XEN_MAPPEDREGS->itc_offset;
+        do {
+                itc_last = XEN_MAPPEDREGS->itc_last;
+                res = ia64_native_getreg(_IA64_REG_AR_ITC);
+                res += itc_offset;
+                if (itc_last >= res)
+                        res = itc_last + 1;
+                ret_itc_last = cmpxchg(&XEN_MAPPEDREGS->itc_last,
+                                       itc_last, res);
+        } while (unlikely(ret_itc_last != itc_last));
+        return res;
+
+#if 0
+        /* ia64_itc_udelay() calls ia64_get_itc() with interrupt enabled.
+           Should it be paravirtualized instead? */
+        WARN_ON(!irqs_disabled());
+        itc_offset = XEN_MAPPEDREGS->itc_offset;
+        itc_last = XEN_MAPPEDREGS->itc_last;
+        res = ia64_native_getreg(_IA64_REG_AR_ITC);
+        res += itc_offset;
+        if (itc_last >= res)
+                res = itc_last + 1;
+        XEN_MAPPEDREGS->itc_last = res;
+        return res;
+#endif
+}
+
 static void xen_setreg(int regnum, unsigned long val)
 {
         switch (regnum) {
@@ -181,11 +306,14 @@ static void xen_setreg(int regnum, unsigned long val)
                 xen_set_eflag(val);
                 break;
 #endif
+        case _IA64_REG_AR_ITC:
+                xen_set_itc(val);
+                break;
         case _IA64_REG_CR_TPR:
                 xen_set_tpr(val);
                 break;
         case _IA64_REG_CR_ITM:
-                xen_set_itm(val);
+                xen_set_itm_with_offset(val);
                 break;
         case _IA64_REG_CR_EOI:
                 xen_eoi(val);
@@ -209,6 +337,12 @@ static unsigned long xen_getreg(int regnum)
                 res = xen_get_eflag();
                 break;
 #endif
+        case _IA64_REG_AR_ITC:
+                res = xen_get_itc();
+                break;
+        case _IA64_REG_CR_ITM:
+                res = xen_get_itm_with_offset();
+                break;
         case _IA64_REG_CR_IVR:
                 res = xen_get_ivr();
                 break;
@@ -259,8 +393,417 @@ xen_intrin_local_irq_restore(unsigned long mask)
         else
                 xen_rsm_i();
 }
+#else
+#define __DEFINE_FUNC(name, code)                                       \
+        extern const char xen_ ## name ## _direct_start[];              \
+        extern const char xen_ ## name ## _direct_end[];                \
+        asm (".align 32\n"                                              \
+             ".proc xen_" #name "\n"                                    \
+             "xen_" #name ":\n"                                         \
+             "xen_" #name "_direct_start:\n"                            \
+             code                                                       \
+             "xen_" #name "_direct_end:\n"                              \
+             "br.cond.sptk.many b6\n"                                   \
+             ".endp xen_" #name "\n")
+
+#define DEFINE_VOID_FUNC0(name, code)           \
+        extern void                             \
+        xen_ ## name (void);                    \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_VOID_FUNC1(name, code)           \
+        extern void                             \
+        xen_ ## name (unsigned long arg);       \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_VOID_FUNC1_VOID(name, code)      \
+        extern void                             \
+        xen_ ## name (void *arg);               \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_VOID_FUNC2(name, code)           \
+        extern void                             \
+        xen_ ## name (unsigned long arg0,       \
+                      unsigned long arg1);      \
+        __DEFINE_FUNC(name, code)
 
-static const struct pv_cpu_ops xen_cpu_ops __initdata = {
+#define DEFINE_FUNC0(name, code)                \
+        extern unsigned long                    \
+        xen_ ## name (void);                    \
+        __DEFINE_FUNC(name, code)
+
+#define DEFINE_FUNC1(name, type, code)          \
+        extern unsigned long                    \
+        xen_ ## name (type arg);                \
+        __DEFINE_FUNC(name, code)
+
+#define XEN_PSR_I_ADDR_ADDR     (XSI_BASE + XSI_PSR_I_ADDR_OFS)
+
+/*
+ * static void xen_set_itm_with_offset(unsigned long val)
+ *        xen_set_itm(val - XEN_MAPPEDREGS->itc_offset);
+ */
+/* 2 bundles */
+DEFINE_VOID_FUNC1(set_itm_with_offset,
+                  "mov r2 = " __stringify(XSI_BASE) " + "
+                  __stringify(XSI_ITC_OFFSET_OFS) "\n"
+                  ";;\n"
+                  "ld8 r3 = [r2]\n"
+                  ";;\n"
+                  "sub r8 = r8, r3\n"
+                  "break " __stringify(HYPERPRIVOP_SET_ITM) "\n");
+
+/*
+ * static unsigned long xen_get_itm_with_offset(void)
+ *    return ia64_native_getreg(_IA64_REG_CR_ITM) + XEN_MAPPEDREGS->itc_offset;
+ */
+/* 2 bundles */
+DEFINE_FUNC0(get_itm_with_offset,
+             "mov r2 = " __stringify(XSI_BASE) " + "
+             __stringify(XSI_ITC_OFFSET_OFS) "\n"
+             ";;\n"
+             "ld8 r3 = [r2]\n"
+             "mov r8 = cr.itm\n"
+             ";;\n"
+             "add r8 = r8, r2\n");
+
+/*
+ * static void xen_set_itc(unsigned long val)
+ *      unsigned long mitc;
+ *
+ *      WARN_ON(!irqs_disabled());
+ *      mitc = ia64_native_getreg(_IA64_REG_AR_ITC);
+ *      XEN_MAPPEDREGS->itc_offset = val - mitc;
+ *      XEN_MAPPEDREGS->itc_last = val;
+ */
+/* 2 bundles */
+DEFINE_VOID_FUNC1(set_itc,
+                  "mov r2 = " __stringify(XSI_BASE) " + "
+                  __stringify(XSI_ITC_LAST_OFS) "\n"
+                  "mov r3 = ar.itc\n"
+                  ";;\n"
+                  "sub r3 = r8, r3\n"
+                  "st8 [r2] = r8, "
+                  __stringify(XSI_ITC_LAST_OFS) " - "
+                  __stringify(XSI_ITC_OFFSET_OFS) "\n"
+                  ";;\n"
+                  "st8 [r2] = r3\n");
+
+/*
+ * static unsigned long xen_get_itc(void)
+ *      unsigned long res;
+ *      unsigned long itc_offset;
+ *      unsigned long itc_last;
+ *      unsigned long ret_itc_last;
+ *
+ *      itc_offset = XEN_MAPPEDREGS->itc_offset;
+ *      do {
+ *              itc_last = XEN_MAPPEDREGS->itc_last;
+ *              res = ia64_native_getreg(_IA64_REG_AR_ITC);
+ *              res += itc_offset;
+ *              if (itc_last >= res)
+ *                      res = itc_last + 1;
+ *              ret_itc_last = cmpxchg(&XEN_MAPPEDREGS->itc_last,
+ *                                     itc_last, res);
+ *      } while (unlikely(ret_itc_last != itc_last));
+ *      return res;
+ */
+/* 5 bundles */
+DEFINE_FUNC0(get_itc,
+             "mov r2 = " __stringify(XSI_BASE) " + "
+             __stringify(XSI_ITC_OFFSET_OFS) "\n"
+             ";;\n"
+             "ld8 r9 = [r2], " __stringify(XSI_ITC_LAST_OFS) " - "
+             __stringify(XSI_ITC_OFFSET_OFS) "\n"
+                                        /* r9 = itc_offset */
+                                        /* r2 = XSI_ITC_OFFSET */
+             "888:\n"
+             "mov r8 = ar.itc\n"        /* res = ar.itc */
+             ";;\n"
+             "ld8 r3 = [r2]\n"          /* r3 = itc_last */
+             "add r8 = r8, r9\n"        /* res = ar.itc + itc_offset */
+             ";;\n"
+             "cmp.gtu p6, p0 = r3, r8\n"
+             ";;\n"
+             "(p6) add r8 = 1, r3\n"    /* if (itc_last > res) itc_last + 1 */
+             ";;\n"
+             "mov ar.ccv = r8\n"
+             ";;\n"
+             "cmpxchg8.acq r10 = [r2], r8, ar.ccv\n"
+             ";;\n"
+             "cmp.ne p6, p0 = r10, r3\n"
+             "(p6) hint @pause\n"
+             "(p6) br.cond.spnt 888b\n");
+
+DEFINE_VOID_FUNC1_VOID(fc,
+                       "break " __stringify(HYPERPRIVOP_FC) "\n");
+
+/*
+ * psr_i_addr_addr = XEN_PSR_I_ADDR_ADDR
+ * masked_addr = *psr_i_addr_addr
+ * pending_intr_addr = masked_addr - 1
+ * if (val & IA64_PSR_I) {
+ *   masked = *masked_addr
+ *   *masked_addr = 0:xen_set_virtual_psr_i(1)
+ *   compiler barrier
+ *   if (masked) {
+ *      uint8_t pending = *pending_intr_addr;
+ *      if (pending)
+ *              XEN_HYPER_SSM_I
+ *   }
+ * } else {
+ *   *masked_addr = 1:xen_set_virtual_psr_i(0)
+ * }
+ */
+/* 6 bundles */
+DEFINE_VOID_FUNC1(intrin_local_irq_restore,
+                  /* r8 = input value: 0 or IA64_PSR_I
+                   * p6 =  (flags & IA64_PSR_I)
+                   *    = if clause
+                   * p7 = !(flags & IA64_PSR_I)
+                   *    = else clause
+                   */
+                  "cmp.ne p6, p7 = r8, r0\n"
+                  "mov r9 = " __stringify(XEN_PSR_I_ADDR_ADDR) "\n"
+                  ";;\n"
+                  /* r9 = XEN_PSR_I_ADDR */
+                  "ld8 r9 = [r9]\n"
+                  ";;\n"
+
+                  /* r10 = masked previous value */
+                  "(p6) ld1.acq r10 = [r9]\n"
+                  ";;\n"
+
+                  /* p8 = !masked interrupt masked previously? */
+                  "(p6) cmp.ne.unc p8, p0 = r10, r0\n"
+
+                  /* p7 = else clause */
+                  "(p7) mov r11 = 1\n"
+                  ";;\n"
+                  /* masked = 1 */
+                  "(p7) st1.rel [r9] = r11\n"
+
+                  /* p6 = if clause */
+                  /* masked = 0
+                   * r9 = masked_addr - 1
+                   *    = pending_intr_addr
+                   */
+                  "(p8) st1.rel [r9] = r0, -1\n"
+                  ";;\n"
+                  /* r8 = pending_intr */
+                  "(p8) ld1.acq r11 = [r9]\n"
+                  ";;\n"
+                  /* p9 = interrupt pending? */
+                  "(p8) cmp.ne.unc p9, p10 = r11, r0\n"
+                  ";;\n"
+                  "(p10) mf\n"
+                  /* issue hypercall to trigger interrupt */
+                  "(p9) break " __stringify(HYPERPRIVOP_SSM_I) "\n");
+
+DEFINE_VOID_FUNC2(ptcga,
+                  "break " __stringify(HYPERPRIVOP_PTC_GA) "\n");
+DEFINE_VOID_FUNC2(set_rr,
+                  "break " __stringify(HYPERPRIVOP_SET_RR) "\n");
+
+/*
+ * tmp = XEN_MAPPEDREGS->interrupt_mask_addr = XEN_PSR_I_ADDR_ADDR;
+ * tmp = *tmp
+ * tmp = *tmp;
+ * psr_i = tmp? 0: IA64_PSR_I;
+ */
+/* 4 bundles */
+DEFINE_FUNC0(get_psr_i,
+             "mov r9 = " __stringify(XEN_PSR_I_ADDR_ADDR) "\n"
+             ";;\n"
+             "ld8 r9 = [r9]\n"                  /* r9 = XEN_PSR_I_ADDR */
+             "mov r8 = 0\n"                     /* psr_i = 0 */
+             ";;\n"
+             "ld1.acq r9 = [r9]\n"              /* r9 = XEN_PSR_I */
+             ";;\n"
+             "cmp.eq.unc p6, p0 = r9, r0\n"     /* p6 = (XEN_PSR_I != 0) */
+             ";;\n"
+             "(p6) mov r8 = " __stringify(1 << IA64_PSR_I_BIT) "\n");
+
+DEFINE_FUNC1(thash, unsigned long,
+             "break " __stringify(HYPERPRIVOP_THASH) "\n");
+DEFINE_FUNC1(get_cpuid, int,
+             "break " __stringify(HYPERPRIVOP_GET_CPUID) "\n");
+DEFINE_FUNC1(get_pmd, int,
+             "break " __stringify(HYPERPRIVOP_GET_PMD) "\n");
+DEFINE_FUNC1(get_rr, unsigned long,
+             "break " __stringify(HYPERPRIVOP_GET_RR) "\n");
+
+/*
+ * void xen_privop_ssm_i(void)
+ *
+ * int masked = !xen_get_virtual_psr_i();
+ *      // masked = *(*XEN_MAPPEDREGS->interrupt_mask_addr)
+ * xen_set_virtual_psr_i(1)
+ *      // *(*XEN_MAPPEDREGS->interrupt_mask_addr) = 0
+ * // compiler barrier
+ * if (masked) {
+ *      uint8_t* pend_int_addr =
+ *              (uint8_t*)(*XEN_MAPPEDREGS->interrupt_mask_addr) - 1;
+ *      uint8_t pending = *pend_int_addr;
+ *      if (pending)
+ *              XEN_HYPER_SSM_I
+ * }
+ */
+/* 4 bundles */
+DEFINE_VOID_FUNC0(ssm_i,
+                  "mov r8 = " __stringify(XEN_PSR_I_ADDR_ADDR) "\n"
+                  ";;\n"
+                  "ld8 r8 = [r8]\n"             /* r8 = XEN_PSR_I_ADDR */
+                  ";;\n"
+                  "ld1.acq r9 = [r8]\n"         /* r9 = XEN_PSR_I */
+                  ";;\n"
+                  "st1.rel [r8] = r0, -1\n"     /* psr_i = 0. enable interrupt
+                                                 * r8 = XEN_PSR_I_ADDR - 1
+                                                 *    = pend_int_addr
+                                                 */
+                  "cmp.eq.unc p0, p6 = r9, r0\n"/* p6 = !XEN_PSR_I
+                                                 * previously interrupt
+                                                 * masked?
+                                                 */
+                  ";;\n"
+                  "(p6) ld1.acq r8 = [r8]\n"    /* r8 = xen_pend_int */
+                  ";;\n"
+                  "(p6) cmp.eq.unc p6, p7 = r8, r0\n"   /*interrupt pending?*/
+                  ";;\n"
+                  /* issue hypercall to get interrupt */
+                  "(p7) break " __stringify(HYPERPRIVOP_SSM_I) "\n"
+                  ";;\n");
+
+/*
+ * psr_i_addr_addr = XEN_MAPPEDREGS->interrupt_mask_addr
+ *                 = XEN_PSR_I_ADDR_ADDR;
+ * psr_i_addr = *psr_i_addr_addr;
+ * *psr_i_addr = 1;
+ */
+/* 2 bundles */
+DEFINE_VOID_FUNC0(rsm_i,
+                  "mov r8 = " __stringify(XEN_PSR_I_ADDR_ADDR) "\n"
+                                                /* r8 = XEN_PSR_I_ADDR */
+                  "mov r9 = 1\n"
+                  ";;\n"
+                  "ld8 r8 = [r8]\n"             /* r8 = XEN_PSR_I */
+                  ";;\n"
+                  "st1.rel [r8] = r9\n");       /* XEN_PSR_I = 1 */
+
+extern void
+xen_set_rr0_to_rr4(unsigned long val0, unsigned long val1,
+                   unsigned long val2, unsigned long val3,
+                   unsigned long val4);
+__DEFINE_FUNC(set_rr0_to_rr4,
+              "break " __stringify(HYPERPRIVOP_SET_RR0_TO_RR4) "\n");
+
+
+extern unsigned long xen_getreg(int regnum);
+#define __DEFINE_GET_REG(id, privop)                                    \
+        "mov r2 = " __stringify(_IA64_REG_ ## id) "\n"                  \
+        ";;\n"                                                          \
+        "cmp.eq p6, p0 = r2, r8\n"                                      \
+        ";;\n"                                                          \
+        "(p6) break " __stringify(HYPERPRIVOP_GET_ ## privop) "\n"      \
+        "(p6) br.cond.sptk.many b6\n"                                   \
+        ";;\n"
+
+__DEFINE_FUNC(getreg,
+              __DEFINE_GET_REG(PSR, PSR)
+#ifdef CONFIG_IA32_SUPPORT
+              __DEFINE_GET_REG(AR_EFLAG, EFLAG)
+#endif
+
+              /* get_itc */
+              "mov r2 = " __stringify(_IA64_REG_AR_ITC) "\n"
+              ";;\n"
+              "cmp.eq p6, p0 = r2, r8\n"
+              ";;\n"
+              "(p6) br.cond.spnt xen_get_itc\n"
+              ";;\n"
+
+              /* get itm */
+              "mov r2 = " __stringify(_IA64_REG_CR_ITM) "\n"
+              ";;\n"
+              "cmp.eq p6, p0 = r2, r8\n"
+              ";;\n"
+              "(p6) br.cond.spnt xen_get_itm_with_offset\n"
+              ";;\n"
+
+              __DEFINE_GET_REG(CR_IVR, IVR)
+              __DEFINE_GET_REG(CR_TPR, TPR)
+
+              /* fall back */
+              "movl r2 = ia64_native_getreg_func\n"
+              ";;\n"
+              "mov b7 = r2\n"
+              ";;\n"
+              "br.cond.sptk.many b7\n");
+
+extern void xen_setreg(int regnum, unsigned long val);
+#define __DEFINE_SET_REG(id, privop)                                    \
+        "mov r2 = " __stringify(_IA64_REG_ ## id) "\n"                  \
+        ";;\n"                                                          \
+        "cmp.eq p6, p0 = r2, r9\n"                                      \
+        ";;\n"                                                          \
+        "(p6) break " __stringify(HYPERPRIVOP_ ## privop) "\n"          \
+        "(p6) br.cond.sptk.many b6\n"                                   \
+        ";;\n"
+
+__DEFINE_FUNC(setreg,
+              /* kr0 .. kr 7*/
+              /*
+               * if (_IA64_REG_AR_KR0 <= regnum &&
+               *     regnum <= _IA64_REG_AR_KR7) {
+               *     register __index asm ("r8") = regnum - _IA64_REG_AR_KR0
+               *     register __val asm ("r9") = val
+               *    "break HYPERPRIVOP_SET_KR"
+               * }
+               */
+              "mov r17 = r9\n"
+              "mov r2 = " __stringify(_IA64_REG_AR_KR0) "\n"
+              ";;\n"
+              "cmp.ge p6, p0 = r9, r2\n"
+              "sub r17 = r17, r2\n"
+              ";;\n"
+              "(p6) cmp.ge.unc p7, p0 = "
+              __stringify(_IA64_REG_AR_KR7) " - " __stringify(_IA64_REG_AR_KR0)
+              ", r17\n"
+              ";;\n"
+              "(p7) mov r9 = r8\n"
+              ";;\n"
+              "(p7) mov r8 = r17\n"
+              "(p7) break " __stringify(HYPERPRIVOP_SET_KR) "\n"
+
+              /* set itm */
+              "mov r2 = " __stringify(_IA64_REG_CR_ITM) "\n"
+              ";;\n"
+              "cmp.eq p6, p0 = r2, r8\n"
+              ";;\n"
+              "(p6) br.cond.spnt xen_set_itm_with_offset\n"
+
+              /* set itc */
+              "mov r2 = " __stringify(_IA64_REG_AR_ITC) "\n"
+              ";;\n"
+              "cmp.eq p6, p0 = r2, r8\n"
+              ";;\n"
+              "(p6) br.cond.spnt xen_set_itc\n"
+
+#ifdef CONFIG_IA32_SUPPORT
+              __DEFINE_SET_REG(AR_EFLAG, SET_EFLAG)
+#endif
+              __DEFINE_SET_REG(CR_TPR, SET_TPR)
+              __DEFINE_SET_REG(CR_EOI, EOI)
+
+              /* fall back */
+              "movl r2 = ia64_native_setreg_func\n"
+              ";;\n"
+              "mov b7 = r2\n"
+              ";;\n"
+              "br.cond.sptk.many b7\n");
+#endif
+
+static const struct pv_cpu_ops xen_cpu_ops __initconst = {
         .fc             = xen_fc,
         .thash          = xen_thash,
         .get_cpuid      = xen_get_cpuid,
@@ -337,7 +880,7 @@ xen_iosapic_write(char __iomem *iosapic, unsigned int reg, u32 val)
         HYPERVISOR_physdev_op(PHYSDEVOP_apic_write, &apic_op);
 }
 
-static const struct pv_iosapic_ops xen_iosapic_ops __initconst = {
+static struct pv_iosapic_ops xen_iosapic_ops __initdata = {
         .pcat_compat_init = xen_pcat_compat_init,
         .__get_irq_chip = xen_iosapic_get_irq_chip,
 
@@ -355,6 +898,8 @@ xen_setup_pv_ops(void)
         xen_info_init();
         pv_info = xen_info;
         pv_init_ops = xen_init_ops;
+        pv_fsys_data = xen_fsys_data;
+        pv_patchdata = xen_patchdata;
         pv_cpu_ops = xen_cpu_ops;
         pv_iosapic_ops = xen_iosapic_ops;
         pv_irq_ops = xen_irq_ops;
@@ -362,3 +907,252 @@ xen_setup_pv_ops(void)
 
         paravirt_cpu_asm_init(&xen_cpu_asm_switch);
 }
+
+#ifdef ASM_SUPPORTED
+/***************************************************************************
+ * binary pacthing
+ * pv_init_ops.patch_bundle
+ */
+
+#define DEFINE_FUNC_GETREG(name, privop)                                \
+        DEFINE_FUNC0(get_ ## name,                                      \
+                     "break "__stringify(HYPERPRIVOP_GET_ ## privop) "\n")
+
+DEFINE_FUNC_GETREG(psr, PSR);
+DEFINE_FUNC_GETREG(eflag, EFLAG);
+DEFINE_FUNC_GETREG(ivr, IVR);
+DEFINE_FUNC_GETREG(tpr, TPR);
+
+#define DEFINE_FUNC_SET_KR(n)                                           \
+        DEFINE_VOID_FUNC0(set_kr ## n,                                  \
+                          ";;\n"                                        \
+                          "mov r9 = r8\n"                               \
+                          "mov r8 = " #n "\n"                           \
+                          "break " __stringify(HYPERPRIVOP_SET_KR) "\n")
+
+DEFINE_FUNC_SET_KR(0);
+DEFINE_FUNC_SET_KR(1);
+DEFINE_FUNC_SET_KR(2);
+DEFINE_FUNC_SET_KR(3);
+DEFINE_FUNC_SET_KR(4);
+DEFINE_FUNC_SET_KR(5);
+DEFINE_FUNC_SET_KR(6);
+DEFINE_FUNC_SET_KR(7);
+
+#define __DEFINE_FUNC_SETREG(name, privop)                              \
+        DEFINE_VOID_FUNC0(name,                                         \
+                          "break "__stringify(HYPERPRIVOP_ ## privop) "\n")
+
+#define DEFINE_FUNC_SETREG(name, privop)                        \
+        __DEFINE_FUNC_SETREG(set_ ## name, SET_ ## privop)
+
+DEFINE_FUNC_SETREG(eflag, EFLAG);
+DEFINE_FUNC_SETREG(tpr, TPR);
+__DEFINE_FUNC_SETREG(eoi, EOI);
+
+extern const char xen_check_events[];
+extern const char __xen_intrin_local_irq_restore_direct_start[];
+extern const char __xen_intrin_local_irq_restore_direct_end[];
+extern const unsigned long __xen_intrin_local_irq_restore_direct_reloc;
+
+asm (
+        ".align 32\n"
+        ".proc xen_check_events\n"
+        "xen_check_events:\n"
+        /* masked = 0
+         * r9 = masked_addr - 1
+         *    = pending_intr_addr
+         */
+        "st1.rel [r9] = r0, -1\n"
+        ";;\n"
+        /* r8 = pending_intr */
+        "ld1.acq r11 = [r9]\n"
+        ";;\n"
+        /* p9 = interrupt pending? */
+        "cmp.ne p9, p10 = r11, r0\n"
+        ";;\n"
+        "(p10) mf\n"
+        /* issue hypercall to trigger interrupt */
+        "(p9) break " __stringify(HYPERPRIVOP_SSM_I) "\n"
+        "br.cond.sptk.many b6\n"
+        ".endp xen_check_events\n"
+        "\n"
+        ".align 32\n"
+        ".proc __xen_intrin_local_irq_restore_direct\n"
+        "__xen_intrin_local_irq_restore_direct:\n"
+        "__xen_intrin_local_irq_restore_direct_start:\n"
+        "1:\n"
+        "{\n"
+        "cmp.ne p6, p7 = r8, r0\n"
+        "mov r17 = ip\n" /* get ip to calc return address */
+        "mov r9 = "__stringify(XEN_PSR_I_ADDR_ADDR) "\n"
+        ";;\n"
+        "}\n"
+        "{\n"
+        /* r9 = XEN_PSR_I_ADDR */
+        "ld8 r9 = [r9]\n"
+        ";;\n"
+        /* r10 = masked previous value */
+        "(p6) ld1.acq r10 = [r9]\n"
+        "adds r17 =  1f - 1b, r17\n" /* calculate return address */
+        ";;\n"
+        "}\n"
+        "{\n"
+        /* p8 = !masked interrupt masked previously? */
+        "(p6) cmp.ne.unc p8, p0 = r10, r0\n"
+        "\n"
+        /* p7 = else clause */
+        "(p7) mov r11 = 1\n"
+        ";;\n"
+        "(p8) mov b6 = r17\n" /* set return address */
+        "}\n"
+        "{\n"
+        /* masked = 1 */
+        "(p7) st1.rel [r9] = r11\n"
+        "\n"
+        "[99:]\n"
+        "(p8) brl.cond.dptk.few xen_check_events\n"
+        "}\n"
+        /* pv calling stub is 5 bundles. fill nop to adjust return address */
+        "{\n"
+        "nop 0\n"
+        "nop 0\n"
+        "nop 0\n"
+        "}\n"
+        "1:\n"
+        "__xen_intrin_local_irq_restore_direct_end:\n"
+        ".endp __xen_intrin_local_irq_restore_direct\n"
+        "\n"
+        ".align 8\n"
+        "__xen_intrin_local_irq_restore_direct_reloc:\n"
+        "data8 99b\n"
+);
+
+static struct paravirt_patch_bundle_elem xen_patch_bundle_elems[]
+__initdata_or_module =
+{
+#define XEN_PATCH_BUNDLE_ELEM(name, type)               \
+        {                                               \
+                (void*)xen_ ## name ## _direct_start,   \
+                (void*)xen_ ## name ## _direct_end,     \
+                PARAVIRT_PATCH_TYPE_ ## type,           \
+        }
+
+        XEN_PATCH_BUNDLE_ELEM(fc, FC),
+        XEN_PATCH_BUNDLE_ELEM(thash, THASH),
+        XEN_PATCH_BUNDLE_ELEM(get_cpuid, GET_CPUID),
+        XEN_PATCH_BUNDLE_ELEM(get_pmd, GET_PMD),
+        XEN_PATCH_BUNDLE_ELEM(ptcga, PTCGA),
+        XEN_PATCH_BUNDLE_ELEM(get_rr, GET_RR),
+        XEN_PATCH_BUNDLE_ELEM(set_rr, SET_RR),
+        XEN_PATCH_BUNDLE_ELEM(set_rr0_to_rr4, SET_RR0_TO_RR4),
+        XEN_PATCH_BUNDLE_ELEM(ssm_i, SSM_I),
+        XEN_PATCH_BUNDLE_ELEM(rsm_i, RSM_I),
+        XEN_PATCH_BUNDLE_ELEM(get_psr_i, GET_PSR_I),
+        {
+                (void*)__xen_intrin_local_irq_restore_direct_start,
+                (void*)__xen_intrin_local_irq_restore_direct_end,
+                PARAVIRT_PATCH_TYPE_INTRIN_LOCAL_IRQ_RESTORE,
+        },
+
+#define XEN_PATCH_BUNDLE_ELEM_GETREG(name, reg)                 \
+        {                                                       \
+                xen_get_ ## name ## _direct_start,              \
+                xen_get_ ## name ## _direct_end,                \
+                PARAVIRT_PATCH_TYPE_GETREG + _IA64_REG_ ## reg, \
+        }
+
+        XEN_PATCH_BUNDLE_ELEM_GETREG(psr, PSR),
+        XEN_PATCH_BUNDLE_ELEM_GETREG(eflag, AR_EFLAG),
+
+        XEN_PATCH_BUNDLE_ELEM_GETREG(ivr, CR_IVR),
+        XEN_PATCH_BUNDLE_ELEM_GETREG(tpr, CR_TPR),
+
+        XEN_PATCH_BUNDLE_ELEM_GETREG(itc, AR_ITC),
+        XEN_PATCH_BUNDLE_ELEM_GETREG(itm_with_offset, CR_ITM),
+
+
+#define __XEN_PATCH_BUNDLE_ELEM_SETREG(name, reg)               \
+        {                                                       \
+                xen_ ## name ## _direct_start,                  \
+                xen_ ## name ## _direct_end,                    \
+                PARAVIRT_PATCH_TYPE_SETREG + _IA64_REG_ ## reg, \
+        }
+
+#define XEN_PATCH_BUNDLE_ELEM_SETREG(name, reg)                 \
+        __XEN_PATCH_BUNDLE_ELEM_SETREG(set_ ## name, reg)
+
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr0, AR_KR0),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr1, AR_KR1),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr2, AR_KR2),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr3, AR_KR3),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr4, AR_KR4),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr5, AR_KR5),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr6, AR_KR6),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(kr7, AR_KR7),
+
+        XEN_PATCH_BUNDLE_ELEM_SETREG(eflag, AR_EFLAG),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(tpr, CR_TPR),
+        __XEN_PATCH_BUNDLE_ELEM_SETREG(eoi, CR_EOI),
+
+        XEN_PATCH_BUNDLE_ELEM_SETREG(itc, AR_ITC),
+        XEN_PATCH_BUNDLE_ELEM_SETREG(itm_with_offset, CR_ITM),
+};
+
+static unsigned long __init_or_module
+xen_patch_bundle(void *sbundle, void *ebundle, unsigned long type)
+{
+        const unsigned long nelems = sizeof(xen_patch_bundle_elems) /
+                sizeof(xen_patch_bundle_elems[0]);
+        unsigned long used;
+        const struct paravirt_patch_bundle_elem *found;
+
+        used = __paravirt_patch_apply_bundle(sbundle, ebundle, type,
+                                             xen_patch_bundle_elems, nelems,
+                                             &found);
+
+        if (found == NULL)
+                /* fallback */
+                return ia64_native_patch_bundle(sbundle, ebundle, type);
+        if (used == 0)
+                return used;
+
+        /* relocation */
+        switch (type) {
+        case PARAVIRT_PATCH_TYPE_INTRIN_LOCAL_IRQ_RESTORE: {
+                unsigned long reloc =
+                        __xen_intrin_local_irq_restore_direct_reloc;
+                unsigned long reloc_offset = reloc - (unsigned long)
+                        __xen_intrin_local_irq_restore_direct_start;
+                unsigned long tag = (unsigned long)sbundle + reloc_offset;
+                paravirt_patch_reloc_brl(tag, xen_check_events);
+                break;
+        }
+        default:
+                /* nothing */
+                break;
+        }
+        return used;
+}
+#endif /* ASM_SUPPOTED */
+
+const struct paravirt_patch_branch_target xen_branch_target[]
+__initconst = {
+#define PARAVIRT_BR_TARGET(name, type)                  \
+        {                                               \
+                &xen_ ## name,                          \
+                PARAVIRT_PATCH_TYPE_BR_ ## type,        \
+        }
+        PARAVIRT_BR_TARGET(switch_to, SWITCH_TO),
+        PARAVIRT_BR_TARGET(leave_syscall, LEAVE_SYSCALL),
+        PARAVIRT_BR_TARGET(work_processed_syscall, WORK_PROCESSED_SYSCALL),
+        PARAVIRT_BR_TARGET(leave_kernel, LEAVE_KERNEL),
+};
+
+static void __init
+xen_patch_branch(unsigned long tag, unsigned long type)
+{
+        const unsigned long nelem =
+                sizeof(xen_branch_target) / sizeof(xen_branch_target[0]);
+        __paravirt_patch_apply_branch(tag, type, xen_branch_target, nelem);
+}
diff --git a/arch/m32r/kernel/process.c b/arch/m32r/kernel/process.c
index 7103d91e1a2f..3e876f0baebc 100644
--- a/arch/m32r/kernel/process.c
+++ b/arch/m32r/kernel/process.c
@@ -225,7 +225,7 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
         return 0; /* Task didn't use the fpu at all. */
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long spu,
+int copy_thread(unsigned long clone_flags, unsigned long spu,
         unsigned long unused, struct task_struct *tsk, struct pt_regs *regs)
 {
         struct pt_regs *childregs = task_pt_regs(tsk);
diff --git a/arch/m68k/kernel/process.c b/arch/m68k/kernel/process.c
index 632ce016014d..ec37fb56c127 100644
--- a/arch/m68k/kernel/process.c
+++ b/arch/m68k/kernel/process.c
@@ -233,7 +233,7 @@ asmlinkage int m68k_clone(struct pt_regs *regs)
                        parent_tidptr, child_tidptr);
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                  unsigned long unused,
                  struct task_struct * p, struct pt_regs * regs)
 {
diff --git a/arch/m68knommu/kernel/process.c b/arch/m68knommu/kernel/process.c
index 3f2d7745f31e..1e96c6eb6312 100644
--- a/arch/m68knommu/kernel/process.c
+++ b/arch/m68knommu/kernel/process.c
@@ -199,7 +199,7 @@ asmlinkage int m68k_clone(struct pt_regs *regs)
         return do_fork(clone_flags, newsp, regs, 0, NULL, NULL);
 }
 
-int copy_thread(int nr, unsigned long clone_flags,
+int copy_thread(unsigned long clone_flags,
                 unsigned long usp, unsigned long topstk,
                 struct task_struct * p, struct pt_regs * regs)
 {
diff --git a/arch/mips/include/asm/spinlock.h b/arch/mips/include/asm/spinlock.h
index 10e82441b496..5b60a09a0f08 100644
--- a/arch/mips/include/asm/spinlock.h
+++ b/arch/mips/include/asm/spinlock.h
@@ -480,6 +480,8 @@ static inline int __raw_write_trylock(raw_rwlock_t *rw)
         return ret;
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
 
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
diff --git a/arch/mips/include/asm/unistd.h b/arch/mips/include/asm/unistd.h
index a73e1531e151..40005010827c 100644
--- a/arch/mips/include/asm/unistd.h
+++ b/arch/mips/include/asm/unistd.h
@@ -350,16 +350,18 @@
 #define __NR_dup3                       (__NR_Linux + 327)
 #define __NR_pipe2                      (__NR_Linux + 328)
 #define __NR_inotify_init1              (__NR_Linux + 329)
+#define __NR_preadv                     (__NR_Linux + 330)
+#define __NR_pwritev                    (__NR_Linux + 331)
 
 /*
  * Offset of the last Linux o32 flavoured syscall
  */
-#define __NR_Linux_syscalls             329
+#define __NR_Linux_syscalls             331
 
 #endif /* _MIPS_SIM == _MIPS_SIM_ABI32 */
 
 #define __NR_O32_Linux                  4000
-#define __NR_O32_Linux_syscalls         329
+#define __NR_O32_Linux_syscalls         331
 
 #if _MIPS_SIM == _MIPS_SIM_ABI64
 
@@ -656,16 +658,18 @@
 #define __NR_dup3                       (__NR_Linux + 286)
 #define __NR_pipe2                      (__NR_Linux + 287)
 #define __NR_inotify_init1              (__NR_Linux + 288)
+#define __NR_preadv                     (__NR_Linux + 289)
+#define __NR_pwritev                    (__NR_Linux + 290)
 
 /*
  * Offset of the last Linux 64-bit flavoured syscall
  */
-#define __NR_Linux_syscalls             288
+#define __NR_Linux_syscalls             290
 
 #endif /* _MIPS_SIM == _MIPS_SIM_ABI64 */
 
 #define __NR_64_Linux                   5000
-#define __NR_64_Linux_syscalls          288
+#define __NR_64_Linux_syscalls          290
 
 #if _MIPS_SIM == _MIPS_SIM_NABI32
 
@@ -966,16 +970,18 @@
 #define __NR_dup3                       (__NR_Linux + 290)
 #define __NR_pipe2                      (__NR_Linux + 291)
 #define __NR_inotify_init1              (__NR_Linux + 292)
+#define __NR_preadv                     (__NR_Linux + 293)
+#define __NR_pwritev                    (__NR_Linux + 294)
 
 /*
  * Offset of the last N32 flavoured syscall
  */
-#define __NR_Linux_syscalls             292
+#define __NR_Linux_syscalls             294
 
 #endif /* _MIPS_SIM == _MIPS_SIM_NABI32 */
 
 #define __NR_N32_Linux                  6000
-#define __NR_N32_Linux_syscalls         292
+#define __NR_N32_Linux_syscalls         294
 
 #ifdef __KERNEL__
 
diff --git a/arch/mips/kernel/process.c b/arch/mips/kernel/process.c
index ca2e4026ad20..1eaaa450e20c 100644
--- a/arch/mips/kernel/process.c
+++ b/arch/mips/kernel/process.c
@@ -99,7 +99,7 @@ void flush_thread(void)
 {
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
         unsigned long unused, struct task_struct *p, struct pt_regs *regs)
 {
         struct thread_info *ti = task_thread_info(p);
diff --git a/arch/mips/kernel/scall32-o32.S b/arch/mips/kernel/scall32-o32.S
index 9ab70c3b5be6..0b31b9bda048 100644
--- a/arch/mips/kernel/scall32-o32.S
+++ b/arch/mips/kernel/scall32-o32.S
@@ -650,6 +650,8 @@ einval:	li	v0, -ENOSYS
         sys     sys_dup3                3
         sys     sys_pipe2               2
         sys     sys_inotify_init1       1
+        sys     sys_preadv              6       /* 4330 */
+        sys     sys_pwritev             6
         .endm
 
         /* We pre-compute the number of _instruction_ bytes needed to
diff --git a/arch/mips/kernel/scall64-64.S b/arch/mips/kernel/scall64-64.S
index 9b4698667154..c647fd6e722f 100644
--- a/arch/mips/kernel/scall64-64.S
+++ b/arch/mips/kernel/scall64-64.S
@@ -487,4 +487,6 @@ sys_call_table:
         PTR     sys_dup3
         PTR     sys_pipe2
         PTR     sys_inotify_init1
+        PTR     sys_preadv
+        PTR     sys_pwritev                     /* 5390 */
         .size   sys_call_table,.-sys_call_table
diff --git a/arch/mips/kernel/scall64-n32.S b/arch/mips/kernel/scall64-n32.S
index f61d6b0e5731..c2c16ef9218f 100644
--- a/arch/mips/kernel/scall64-n32.S
+++ b/arch/mips/kernel/scall64-n32.S
@@ -413,4 +413,6 @@ EXPORT(sysn32_call_table)
         PTR     sys_dup3                        /* 5290 */
         PTR     sys_pipe2
         PTR     sys_inotify_init1
+        PTR     sys_preadv
+        PTR     sys_pwritev
         .size   sysn32_call_table,.-sysn32_call_table
diff --git a/arch/mips/kernel/scall64-o32.S b/arch/mips/kernel/scall64-o32.S
index 60997f1f69d4..002fac27021e 100644
--- a/arch/mips/kernel/scall64-o32.S
+++ b/arch/mips/kernel/scall64-o32.S
@@ -533,4 +533,6 @@ sys_call_table:
         PTR     sys_dup3
         PTR     sys_pipe2
         PTR     sys_inotify_init1
+        PTR     compat_sys_preadv               /* 4330 */
+        PTR     compat_sys_pwritev
         .size   sys_call_table,.-sys_call_table
diff --git a/arch/mn10300/kernel/process.c b/arch/mn10300/kernel/process.c
index b28c9a60445b..234cf344cdce 100644
--- a/arch/mn10300/kernel/process.c
+++ b/arch/mn10300/kernel/process.c
@@ -193,7 +193,7 @@ void prepare_to_copy(struct task_struct *tsk)
  * set up the kernel stack for a new thread and copy arch-specific thread
  * control information
  */
-int copy_thread(int nr, unsigned long clone_flags,
+int copy_thread(unsigned long clone_flags,
                 unsigned long c_usp, unsigned long ustk_size,
                 struct task_struct *p, struct pt_regs *kregs)
 {
diff --git a/arch/parisc/include/asm/spinlock.h b/arch/parisc/include/asm/spinlock.h
index f3d2090a18dc..fae03e136fa8 100644
--- a/arch/parisc/include/asm/spinlock.h
+++ b/arch/parisc/include/asm/spinlock.h
@@ -187,6 +187,9 @@ static __inline__ int __raw_write_can_lock(raw_rwlock_t *rw)
         return !rw->counter;
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
 #define _raw_write_relax(lock)  cpu_relax()
diff --git a/arch/parisc/kernel/process.c b/arch/parisc/kernel/process.c
index b80e02a4d81d..8aa591ed9127 100644
--- a/arch/parisc/kernel/process.c
+++ b/arch/parisc/kernel/process.c
@@ -263,7 +263,7 @@ sys_vfork(struct pt_regs *regs)
 }
 
 int
-copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+copy_thread(unsigned long clone_flags, unsigned long usp,
             unsigned long unused,       /* in ia64 this is "user_stack_size" */
             struct task_struct * p, struct pt_regs * pregs)
 {
diff --git a/arch/powerpc/Kconfig.debug b/arch/powerpc/Kconfig.debug
index 6aa0b5e087cd..a1098e23221f 100644
--- a/arch/powerpc/Kconfig.debug
+++ b/arch/powerpc/Kconfig.debug
@@ -27,16 +27,6 @@ config DEBUG_STACK_USAGE
 
           This option will slow down process creation somewhat.
 
-config DEBUG_PAGEALLOC
-        bool "Debug page memory allocations"
-        depends on DEBUG_KERNEL && !HIBERNATION
-        depends on ARCH_SUPPORTS_DEBUG_PAGEALLOC
-        help
-          Unmap pages from the kernel linear mapping after free_pages().
-          This results in a large slowdown, but helps to find certain types
-          of memory corruptions.
-
-
 config HCALL_STATS
         bool "Hypervisor call instrumentation"
         depends on PPC_PSERIES && DEBUG_FS
diff --git a/arch/powerpc/include/asm/spinlock.h b/arch/powerpc/include/asm/spinlock.h
index 36864364e601..c3b193121f81 100644
--- a/arch/powerpc/include/asm/spinlock.h
+++ b/arch/powerpc/include/asm/spinlock.h
@@ -287,6 +287,9 @@ static inline void __raw_write_unlock(raw_rwlock_t *rw)
         rw->lock = 0;
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_spin_relax(lock)   __spin_yield(lock)
 #define _raw_read_relax(lock)   __rw_yield(lock)
 #define _raw_write_relax(lock)  __rw_yield(lock)
diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
index eac064948780..7b44a33f03c2 100644
--- a/arch/powerpc/kernel/process.c
+++ b/arch/powerpc/kernel/process.c
@@ -598,7 +598,7 @@ void prepare_to_copy(struct task_struct *tsk)
 /*
  * Copy a thread..
  */
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                 unsigned long unused, struct task_struct *p,
                 struct pt_regs *regs)
 {
diff --git a/arch/powerpc/kernel/vio.c b/arch/powerpc/kernel/vio.c
index d3694498f3af..819e59f6f7c7 100644
--- a/arch/powerpc/kernel/vio.c
+++ b/arch/powerpc/kernel/vio.c
@@ -482,7 +482,7 @@ static void vio_cmo_balance(struct work_struct *work)
         cmo->excess.size = cmo->entitled - cmo->reserve.size;
         cmo->excess.free = cmo->excess.size - need;
 
-        cancel_delayed_work(container_of(work, struct delayed_work, work));
+        cancel_delayed_work(to_delayed_work(work));
         spin_unlock_irqrestore(&vio_cmo.lock, flags);
 }
 
diff --git a/arch/powerpc/platforms/cell/spufs/inode.c b/arch/powerpc/platforms/cell/spufs/inode.c
index 64f068540d0d..706eb5c7e2ee 100644
--- a/arch/powerpc/platforms/cell/spufs/inode.c
+++ b/arch/powerpc/platforms/cell/spufs/inode.c
@@ -635,7 +635,7 @@ long spufs_create(struct nameidata *nd, unsigned int flags, mode_t mode,
         if (dentry->d_inode)
                 goto out_dput;
 
-        mode &= ~current->fs->umask;
+        mode &= ~current_umask();
 
         if (flags & SPU_CREATE_GANG)
                 ret = spufs_create_gang(nd->path.dentry->d_inode,
diff --git a/arch/s390/Kconfig.debug b/arch/s390/Kconfig.debug
index 7e297a3cde34..2283933a9a93 100644
--- a/arch/s390/Kconfig.debug
+++ b/arch/s390/Kconfig.debug
@@ -6,13 +6,4 @@ config TRACE_IRQFLAGS_SUPPORT
 
 source "lib/Kconfig.debug"
 
-config DEBUG_PAGEALLOC
-        bool "Debug page memory allocations"
-        depends on DEBUG_KERNEL
-        depends on ARCH_SUPPORTS_DEBUG_PAGEALLOC
-        help
-          Unmap pages from the kernel linear mapping after free_pages().
-          This results in a slowdown, but helps to find certain types of
-          memory corruptions.
-
 endmenu
diff --git a/arch/s390/include/asm/spinlock.h b/arch/s390/include/asm/spinlock.h
index df84ae96915f..f3861b09ebb0 100644
--- a/arch/s390/include/asm/spinlock.h
+++ b/arch/s390/include/asm/spinlock.h
@@ -172,6 +172,9 @@ static inline int __raw_write_trylock(raw_rwlock_t *rw)
         return _raw_write_trylock_retry(rw);
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_read_relax(lock)   cpu_relax()
 #define _raw_write_relax(lock)  cpu_relax()
 
diff --git a/arch/s390/kernel/process.c b/arch/s390/kernel/process.c
index b48e961a38f6..a3acd8e60aff 100644
--- a/arch/s390/kernel/process.c
+++ b/arch/s390/kernel/process.c
@@ -160,7 +160,7 @@ void release_thread(struct task_struct *dead_task)
 {
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long new_stackp,
+int copy_thread(unsigned long clone_flags, unsigned long new_stackp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/sh/include/asm/spinlock.h b/arch/sh/include/asm/spinlock.h
index e793181d64da..60283565f89b 100644
--- a/arch/sh/include/asm/spinlock.h
+++ b/arch/sh/include/asm/spinlock.h
@@ -216,6 +216,9 @@ static inline int __raw_write_trylock(raw_rwlock_t *rw)
         return (oldval > (RW_LOCK_BIAS - 1));
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
 #define _raw_write_relax(lock)  cpu_relax()
diff --git a/arch/sh/kernel/process_32.c b/arch/sh/kernel/process_32.c
index ddafbbbab2ab..694bc15f84fd 100644
--- a/arch/sh/kernel/process_32.c
+++ b/arch/sh/kernel/process_32.c
@@ -170,7 +170,7 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 
 asmlinkage void ret_from_fork(void);
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/sh/kernel/process_64.c b/arch/sh/kernel/process_64.c
index c90c7e5e5fee..96be839040f8 100644
--- a/arch/sh/kernel/process_64.c
+++ b/arch/sh/kernel/process_64.c
@@ -425,7 +425,7 @@ int dump_fpu(struct pt_regs *regs, elf_fpregset_t *fpu)
 
 asmlinkage void ret_from_fork(void);
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/sparc/Kconfig.debug b/arch/sparc/Kconfig.debug
index d001b42041a5..90d5fe223a74 100644
--- a/arch/sparc/Kconfig.debug
+++ b/arch/sparc/Kconfig.debug
@@ -22,15 +22,6 @@ config DEBUG_DCFLUSH
 config STACK_DEBUG
         bool "Stack Overflow Detection Support"
 
-config DEBUG_PAGEALLOC
-        bool "Debug page memory allocations"
-        depends on DEBUG_KERNEL && !HIBERNATION
-        depends on ARCH_SUPPORTS_DEBUG_PAGEALLOC
-        help
-          Unmap pages from the kernel linear mapping after free_pages().
-          This results in a large slowdown, but helps to find certain types
-          of memory corruptions.
-
 config MCOUNT
         bool
         depends on SPARC64
diff --git a/arch/sparc/include/asm/spinlock_32.h b/arch/sparc/include/asm/spinlock_32.h
index bf2d532593e3..46f91ab66a50 100644
--- a/arch/sparc/include/asm/spinlock_32.h
+++ b/arch/sparc/include/asm/spinlock_32.h
@@ -177,6 +177,8 @@ static inline int __read_trylock(raw_rwlock_t *rw)
 #define __raw_write_unlock(rw)  do { (rw)->lock = 0; } while(0)
 
 #define __raw_spin_lock_flags(lock, flags) __raw_spin_lock(lock)
+#define __raw_read_lock_flags(rw, flags)   __raw_read_lock(rw)
+#define __raw_write_lock_flags(rw, flags)  __raw_write_lock(rw)
 
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
diff --git a/arch/sparc/include/asm/spinlock_64.h b/arch/sparc/include/asm/spinlock_64.h
index c4d274d330e9..f6b2b92ad8d2 100644
--- a/arch/sparc/include/asm/spinlock_64.h
+++ b/arch/sparc/include/asm/spinlock_64.h
@@ -211,9 +211,11 @@ static int inline __write_trylock(raw_rwlock_t *lock)
 }
 
 #define __raw_read_lock(p)      __read_lock(p)
+#define __raw_read_lock_flags(p, f) __read_lock(p)
 #define __raw_read_trylock(p)   __read_trylock(p)
 #define __raw_read_unlock(p)    __read_unlock(p)
 #define __raw_write_lock(p)     __write_lock(p)
+#define __raw_write_lock_flags(p, f) __write_lock(p)
 #define __raw_write_unlock(p)   __write_unlock(p)
 #define __raw_write_trylock(p)  __write_trylock(p)
 
diff --git a/arch/sparc/kernel/process_32.c b/arch/sparc/kernel/process_32.c
index f4bee35a1b46..2830b415e214 100644
--- a/arch/sparc/kernel/process_32.c
+++ b/arch/sparc/kernel/process_32.c
@@ -455,7 +455,7 @@ asmlinkage int sparc_do_fork(unsigned long clone_flags,
  */
 extern void ret_from_fork(void);
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
+int copy_thread(unsigned long clone_flags, unsigned long sp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/sparc/kernel/process_64.c b/arch/sparc/kernel/process_64.c
index a73954b87f0a..4041f94e7724 100644
--- a/arch/sparc/kernel/process_64.c
+++ b/arch/sparc/kernel/process_64.c
@@ -561,7 +561,7 @@ asmlinkage long sparc_do_fork(unsigned long clone_flags,
  * Parent -->  %o0 == childs  pid, %o1 == 0
  * Child  -->  %o0 == parents pid, %o1 == 1
  */
-int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
+int copy_thread(unsigned long clone_flags, unsigned long sp,
                 unsigned long unused,
                 struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/um/drivers/net_kern.c b/arch/um/drivers/net_kern.c
index 434224e2229f..434ba121e3c5 100644
--- a/arch/um/drivers/net_kern.c
+++ b/arch/um/drivers/net_kern.c
@@ -757,7 +757,7 @@ static int uml_inetaddr_event(struct notifier_block *this, unsigned long event,
         void (*proc)(unsigned char *, unsigned char *, void *);
         unsigned char addr_buf[4], netmask_buf[4];
 
-        if (dev->open != uml_net_open)
+        if (dev->netdev_ops->ndo_open != uml_net_open)
                 return NOTIFY_DONE;
 
         lp = netdev_priv(dev);
diff --git a/arch/um/kernel/process.c b/arch/um/kernel/process.c
index a1c6d07cac3e..4a28a1568d85 100644
--- a/arch/um/kernel/process.c
+++ b/arch/um/kernel/process.c
@@ -179,7 +179,7 @@ void fork_handler(void)
         userspace(&current->thread.regs.regs);
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
+int copy_thread(unsigned long clone_flags, unsigned long sp,
                 unsigned long stack_top, struct task_struct * p,
                 struct pt_regs *regs)
 {
diff --git a/arch/um/kernel/syscall.c b/arch/um/kernel/syscall.c
index c4df705b8359..a4625c7b2bf9 100644
--- a/arch/um/kernel/syscall.c
+++ b/arch/um/kernel/syscall.c
@@ -127,7 +127,8 @@ int kernel_execve(const char *filename, char *const argv[], char *const envp[])
 
         fs = get_fs();
         set_fs(KERNEL_DS);
-        ret = um_execve(filename, argv, envp);
+        ret = um_execve((char *)filename, (char __user *__user *)argv,
+                        (char __user *__user *) envp);
         set_fs(fs);
 
         return ret;
diff --git a/arch/um/sys-i386/sys_call_table.S b/arch/um/sys-i386/sys_call_table.S
index 00e5f5203eea..c6260dd6ebb9 100644
--- a/arch/um/sys-i386/sys_call_table.S
+++ b/arch/um/sys-i386/sys_call_table.S
@@ -9,6 +9,17 @@
 
 #define old_mmap old_mmap_i386
 
+#define ptregs_fork sys_fork
+#define ptregs_execve sys_execve
+#define ptregs_iopl sys_iopl
+#define ptregs_vm86old sys_vm86old
+#define ptregs_sigreturn sys_sigreturn
+#define ptregs_clone sys_clone
+#define ptregs_vm86 sys_vm86
+#define ptregs_rt_sigreturn sys_rt_sigreturn
+#define ptregs_sigaltstack sys_sigaltstack
+#define ptregs_vfork sys_vfork
+
 .section .rodata,"a"
 
 #include "../../x86/kernel/syscall_table_32.S"
diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug
index a345cb5447a8..d8359e73317f 100644
--- a/arch/x86/Kconfig.debug
+++ b/arch/x86/Kconfig.debug
@@ -72,15 +72,6 @@ config DEBUG_STACK_USAGE
 
           This option will slow down process creation somewhat.
 
-config DEBUG_PAGEALLOC
-        bool "Debug page memory allocations"
-        depends on DEBUG_KERNEL
-        depends on ARCH_SUPPORTS_DEBUG_PAGEALLOC
-        ---help---
-          Unmap pages from the kernel linear mapping after free_pages().
-          This results in a large slowdown, but helps to find certain types
-          of memory corruptions.
-
 config DEBUG_PER_CPU_MAPS
         bool "Debug access to per_cpu maps"
         depends on DEBUG_KERNEL
diff --git a/arch/x86/boot/memory.c b/arch/x86/boot/memory.c
index 8c3c25f35578..5054c2ddd1a0 100644
--- a/arch/x86/boot/memory.c
+++ b/arch/x86/boot/memory.c
@@ -2,6 +2,7 @@
  *
  *   Copyright (C) 1991, 1992 Linus Torvalds
  *   Copyright 2007 rPath, Inc. - All Rights Reserved
+ *   Copyright 2009 Intel Corporation; author H. Peter Anvin
  *
  *   This file is part of the Linux kernel, and is made available under
  *   the terms of the GNU General Public License version 2.
@@ -16,24 +17,38 @@
 
 #define SMAP    0x534d4150      /* ASCII "SMAP" */
 
+struct e820_ext_entry {
+        struct e820entry std;
+        u32 ext_flags;
+} __attribute__((packed));
+
 static int detect_memory_e820(void)
 {
         int count = 0;
         u32 next = 0;
-        u32 size, id;
+        u32 size, id, edi;
         u8 err;
         struct e820entry *desc = boot_params.e820_map;
+        static struct e820_ext_entry buf; /* static so it is zeroed */
+
+        /*
+         * Set this here so that if the BIOS doesn't change this field
+         * but still doesn't change %ecx, we're still okay...
+         */
+        buf.ext_flags = 1;
 
         do {
-                size = sizeof(struct e820entry);
+                size = sizeof buf;
 
-                /* Important: %edx is clobbered by some BIOSes,
-                   so it must be either used for the error output
-                   or explicitly marked clobbered. */
-                asm("int $0x15; setc %0"
+                /* Important: %edx and %esi are clobbered by some BIOSes,
+                   so they must be either used for the error output
+                   or explicitly marked clobbered.  Given that, assume there
+                   is something out there clobbering %ebp and %edi, too. */
+                asm("pushl %%ebp; int $0x15; popl %%ebp; setc %0"
                     : "=d" (err), "+b" (next), "=a" (id), "+c" (size),
-                      "=m" (*desc)
-                    : "D" (desc), "d" (SMAP), "a" (0xe820));
+                      "=D" (edi), "+m" (buf)
+                    : "D" (&buf), "d" (SMAP), "a" (0xe820)
+                    : "esi");
 
                 /* BIOSes which terminate the chain with CF = 1 as opposed
                    to %ebx = 0 don't always report the SMAP signature on
@@ -51,8 +66,14 @@ static int detect_memory_e820(void)
                         break;
                 }
 
+                /* ACPI 3.0 added the extended flags support.  If bit 0
+                   in the extended flags is zero, we're supposed to simply
+                   ignore the entry -- a backwards incompatible change! */
+                if (size > 20 && !(buf.ext_flags & 1))
+                        continue;
+
+                *desc++ = buf.std;
                 count++;
-                desc++;
         } while (next && count < ARRAY_SIZE(boot_params.e820_map));
 
         return boot_params.e820_entries = count;
diff --git a/arch/x86/ia32/ia32entry.S b/arch/x86/ia32/ia32entry.S
index db0c803170ab..a505202086e8 100644
--- a/arch/x86/ia32/ia32entry.S
+++ b/arch/x86/ia32/ia32entry.S
@@ -828,4 +828,6 @@ ia32_sys_call_table:
         .quad sys_dup3                  /* 330 */
         .quad sys_pipe2
         .quad sys_inotify_init1
+        .quad compat_sys_preadv
+        .quad compat_sys_pwritev
 ia32_syscall_end:
diff --git a/arch/x86/include/asm/pci.h b/arch/x86/include/asm/pci.h
index a0301bfeb954..e545ea01abcf 100644
--- a/arch/x86/include/asm/pci.h
+++ b/arch/x86/include/asm/pci.h
@@ -89,12 +89,40 @@ extern void pci_iommu_alloc(void);
 /* MSI arch hook */
 #define arch_setup_msi_irqs arch_setup_msi_irqs
 
-#endif  /* __KERNEL__ */
+#define PCI_DMA_BUS_IS_PHYS (dma_ops->is_phys)
+
+#if defined(CONFIG_X86_64) || defined(CONFIG_DMA_API_DEBUG)
+
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)       \
+                dma_addr_t ADDR_NAME;
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)         \
+                __u32 LEN_NAME;
+#define pci_unmap_addr(PTR, ADDR_NAME)                  \
+                ((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)         \
+                (((PTR)->ADDR_NAME) = (VAL))
+#define pci_unmap_len(PTR, LEN_NAME)                    \
+                ((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL)           \
+                (((PTR)->LEN_NAME) = (VAL))
 
-#ifdef CONFIG_X86_32
-# include "pci_32.h"
 #else
-# include "pci_64.h"
+
+#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)       dma_addr_t ADDR_NAME[0];
+#define DECLARE_PCI_UNMAP_LEN(LEN_NAME) unsigned LEN_NAME[0];
+#define pci_unmap_addr(PTR, ADDR_NAME)  sizeof((PTR)->ADDR_NAME)
+#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL) \
+                do { break; } while (pci_unmap_addr(PTR, ADDR_NAME))
+#define pci_unmap_len(PTR, LEN_NAME)            sizeof((PTR)->LEN_NAME)
+#define pci_unmap_len_set(PTR, LEN_NAME, VAL) \
+                do { break; } while (pci_unmap_len(PTR, LEN_NAME))
+
+#endif
+
+#endif  /* __KERNEL__ */
+
+#ifdef CONFIG_X86_64
+#include "pci_64.h"
 #endif
 
 /* implement the pci_ DMA API in terms of the generic device dma_ one */
diff --git a/arch/x86/include/asm/pci_32.h b/arch/x86/include/asm/pci_32.h
deleted file mode 100644
index 6f1213a6ef4f..000000000000
--- a/arch/x86/include/asm/pci_32.h
+++ /dev/null
@@ -1,34 +0,0 @@
-#ifndef _ASM_X86_PCI_32_H
-#define _ASM_X86_PCI_32_H
-
-
-#ifdef __KERNEL__
-
-
-/* Dynamic DMA mapping stuff.
- * i386 has everything mapped statically.
- */
-
-struct pci_dev;
-
-/* The PCI address space does equal the physical memory
- * address space.  The networking and block device layers use
- * this boolean for bounce buffer decisions.
- */
-#define PCI_DMA_BUS_IS_PHYS     (1)
-
-/* pci_unmap_{page,single} is a nop so... */
-#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)       dma_addr_t ADDR_NAME[0];
-#define DECLARE_PCI_UNMAP_LEN(LEN_NAME) unsigned LEN_NAME[0];
-#define pci_unmap_addr(PTR, ADDR_NAME)  sizeof((PTR)->ADDR_NAME)
-#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL) \
-        do { break; } while (pci_unmap_addr(PTR, ADDR_NAME))
-#define pci_unmap_len(PTR, LEN_NAME)            sizeof((PTR)->LEN_NAME)
-#define pci_unmap_len_set(PTR, LEN_NAME, VAL) \
-        do { break; } while (pci_unmap_len(PTR, LEN_NAME))
-
-
-#endif /* __KERNEL__ */
-
-
-#endif /* _ASM_X86_PCI_32_H */
diff --git a/arch/x86/include/asm/pci_64.h b/arch/x86/include/asm/pci_64.h
index 4da207982777..ae5e40f67daf 100644
--- a/arch/x86/include/asm/pci_64.h
+++ b/arch/x86/include/asm/pci_64.h
@@ -24,28 +24,6 @@ extern int (*pci_config_write)(int seg, int bus, int dev, int fn,
 
 extern void dma32_reserve_bootmem(void);
 
-/* The PCI address space does equal the physical memory
- * address space.  The networking and block device layers use
- * this boolean for bounce buffer decisions
- *
- * On AMD64 it mostly equals, but we set it to zero if a hardware
- * IOMMU (gart) of sotware IOMMU (swiotlb) is available.
- */
-#define PCI_DMA_BUS_IS_PHYS (dma_ops->is_phys)
-
-#define DECLARE_PCI_UNMAP_ADDR(ADDR_NAME)       \
-        dma_addr_t ADDR_NAME;
-#define DECLARE_PCI_UNMAP_LEN(LEN_NAME)         \
-        __u32 LEN_NAME;
-#define pci_unmap_addr(PTR, ADDR_NAME)                  \
-        ((PTR)->ADDR_NAME)
-#define pci_unmap_addr_set(PTR, ADDR_NAME, VAL)         \
-        (((PTR)->ADDR_NAME) = (VAL))
-#define pci_unmap_len(PTR, LEN_NAME)                    \
-        ((PTR)->LEN_NAME)
-#define pci_unmap_len_set(PTR, LEN_NAME, VAL)           \
-        (((PTR)->LEN_NAME) = (VAL))
-
 #endif /* __KERNEL__ */
 
 #endif /* _ASM_X86_PCI_64_H */
diff --git a/arch/x86/include/asm/spinlock.h b/arch/x86/include/asm/spinlock.h
index 3a5696656680..e5e6caffec87 100644
--- a/arch/x86/include/asm/spinlock.h
+++ b/arch/x86/include/asm/spinlock.h
@@ -295,6 +295,9 @@ static inline void __raw_write_unlock(raw_rwlock_t *rw)
                      : "+m" (rw->lock) : "i" (RW_LOCK_BIAS) : "memory");
 }
 
+#define __raw_read_lock_flags(lock, flags) __raw_read_lock(lock)
+#define __raw_write_lock_flags(lock, flags) __raw_write_lock(lock)
+
 #define _raw_spin_relax(lock)   cpu_relax()
 #define _raw_read_relax(lock)   cpu_relax()
 #define _raw_write_relax(lock)  cpu_relax()
diff --git a/arch/x86/include/asm/unistd_32.h b/arch/x86/include/asm/unistd_32.h
index f2bba78430a4..6e72d74cf8dc 100644
--- a/arch/x86/include/asm/unistd_32.h
+++ b/arch/x86/include/asm/unistd_32.h
@@ -338,6 +338,8 @@
 #define __NR_dup3               330
 #define __NR_pipe2              331
 #define __NR_inotify_init1      332
+#define __NR_preadv             333
+#define __NR_pwritev            334
 
 #ifdef __KERNEL__
 
diff --git a/arch/x86/include/asm/unistd_64.h b/arch/x86/include/asm/unistd_64.h
index d2e415e6666f..f81829462325 100644
--- a/arch/x86/include/asm/unistd_64.h
+++ b/arch/x86/include/asm/unistd_64.h
@@ -653,6 +653,10 @@ __SYSCALL(__NR_dup3, sys_dup3)
 __SYSCALL(__NR_pipe2, sys_pipe2)
 #define __NR_inotify_init1                      294
 __SYSCALL(__NR_inotify_init1, sys_inotify_init1)
+#define __NR_preadv                             295
+__SYSCALL(__NR_preadv, sys_preadv)
+#define __NR_pwritev                            296
+__SYSCALL(__NR_pwritev, sys_pwritev)
 
 
 #ifndef __NO_STUBS
diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h
index 9f4dfba33b28..d3a98ea1062e 100644
--- a/arch/x86/include/asm/uv/uv_hub.h
+++ b/arch/x86/include/asm/uv/uv_hub.h
@@ -11,11 +11,13 @@
 #ifndef _ASM_X86_UV_UV_HUB_H
 #define _ASM_X86_UV_UV_HUB_H
 
+#ifdef CONFIG_X86_64
 #include <linux/numa.h>
 #include <linux/percpu.h>
 #include <linux/timer.h>
 #include <asm/types.h>
 #include <asm/percpu.h>
+#include <asm/uv/uv_mmrs.h>
 
 
 /*
@@ -397,6 +399,7 @@ static inline void uv_set_scir_bits(unsigned char value)
                 uv_write_local_mmr8(uv_hub_info->scir.offset, value);
         }
 }
+
 static inline void uv_set_cpu_scir_bits(int cpu, unsigned char value)
 {
         if (uv_cpu_hub_info(cpu)->scir.state != value) {
@@ -405,4 +408,15 @@ static inline void uv_set_cpu_scir_bits(int cpu, unsigned char value)
         }
 }
 
+static inline void uv_hub_send_ipi(int pnode, int apicid, int vector)
+{
+        unsigned long val;
+
+        val = (1UL << UVH_IPI_INT_SEND_SHFT) |
+                        ((apicid & 0x3f) << UVH_IPI_INT_APIC_ID_SHFT) |
+                        (vector << UVH_IPI_INT_VECTOR_SHFT);
+        uv_write_global_mmr64(pnode, UVH_IPI_INT, val);
+}
+
+#endif /* CONFIG_X86_64 */
 #endif /* _ASM_X86_UV_UV_HUB_H */
diff --git a/arch/x86/include/asm/uv/uv_mmrs.h b/arch/x86/include/asm/uv/uv_mmrs.h
index dd627793a234..db68ac8a5ac2 100644
--- a/arch/x86/include/asm/uv/uv_mmrs.h
+++ b/arch/x86/include/asm/uv/uv_mmrs.h
@@ -1,3 +1,4 @@
+
 /*
  * This file is subject to the terms and conditions of the GNU General Public
  * License.  See the file "COPYING" in the main directory of this archive
@@ -243,6 +244,158 @@ union uvh_event_occurred0_u {
 #define UVH_EVENT_OCCURRED0_ALIAS_32 0x005f0
 
 /* ========================================================================= */
+/*                         UVH_GR0_TLB_INT0_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR0_TLB_INT0_CONFIG 0x61b00UL
+
+#define UVH_GR0_TLB_INT0_CONFIG_VECTOR_SHFT 0
+#define UVH_GR0_TLB_INT0_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR0_TLB_INT0_CONFIG_DM_SHFT 8
+#define UVH_GR0_TLB_INT0_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR0_TLB_INT0_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR0_TLB_INT0_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR0_TLB_INT0_CONFIG_STATUS_SHFT 12
+#define UVH_GR0_TLB_INT0_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR0_TLB_INT0_CONFIG_P_SHFT 13
+#define UVH_GR0_TLB_INT0_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR0_TLB_INT0_CONFIG_T_SHFT 15
+#define UVH_GR0_TLB_INT0_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR0_TLB_INT0_CONFIG_M_SHFT 16
+#define UVH_GR0_TLB_INT0_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR0_TLB_INT0_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR0_TLB_INT0_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr0_tlb_int0_config_u {
+    unsigned long       v;
+    struct uvh_gr0_tlb_int0_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
+/*                         UVH_GR0_TLB_INT1_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR0_TLB_INT1_CONFIG 0x61b40UL
+
+#define UVH_GR0_TLB_INT1_CONFIG_VECTOR_SHFT 0
+#define UVH_GR0_TLB_INT1_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR0_TLB_INT1_CONFIG_DM_SHFT 8
+#define UVH_GR0_TLB_INT1_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR0_TLB_INT1_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR0_TLB_INT1_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR0_TLB_INT1_CONFIG_STATUS_SHFT 12
+#define UVH_GR0_TLB_INT1_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR0_TLB_INT1_CONFIG_P_SHFT 13
+#define UVH_GR0_TLB_INT1_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR0_TLB_INT1_CONFIG_T_SHFT 15
+#define UVH_GR0_TLB_INT1_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR0_TLB_INT1_CONFIG_M_SHFT 16
+#define UVH_GR0_TLB_INT1_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR0_TLB_INT1_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR0_TLB_INT1_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr0_tlb_int1_config_u {
+    unsigned long       v;
+    struct uvh_gr0_tlb_int1_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
+/*                         UVH_GR1_TLB_INT0_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR1_TLB_INT0_CONFIG 0x61f00UL
+
+#define UVH_GR1_TLB_INT0_CONFIG_VECTOR_SHFT 0
+#define UVH_GR1_TLB_INT0_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR1_TLB_INT0_CONFIG_DM_SHFT 8
+#define UVH_GR1_TLB_INT0_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR1_TLB_INT0_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR1_TLB_INT0_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR1_TLB_INT0_CONFIG_STATUS_SHFT 12
+#define UVH_GR1_TLB_INT0_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR1_TLB_INT0_CONFIG_P_SHFT 13
+#define UVH_GR1_TLB_INT0_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR1_TLB_INT0_CONFIG_T_SHFT 15
+#define UVH_GR1_TLB_INT0_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR1_TLB_INT0_CONFIG_M_SHFT 16
+#define UVH_GR1_TLB_INT0_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR1_TLB_INT0_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR1_TLB_INT0_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr1_tlb_int0_config_u {
+    unsigned long       v;
+    struct uvh_gr1_tlb_int0_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
+/*                         UVH_GR1_TLB_INT1_CONFIG                           */
+/* ========================================================================= */
+#define UVH_GR1_TLB_INT1_CONFIG 0x61f40UL
+
+#define UVH_GR1_TLB_INT1_CONFIG_VECTOR_SHFT 0
+#define UVH_GR1_TLB_INT1_CONFIG_VECTOR_MASK 0x00000000000000ffUL
+#define UVH_GR1_TLB_INT1_CONFIG_DM_SHFT 8
+#define UVH_GR1_TLB_INT1_CONFIG_DM_MASK 0x0000000000000700UL
+#define UVH_GR1_TLB_INT1_CONFIG_DESTMODE_SHFT 11
+#define UVH_GR1_TLB_INT1_CONFIG_DESTMODE_MASK 0x0000000000000800UL
+#define UVH_GR1_TLB_INT1_CONFIG_STATUS_SHFT 12
+#define UVH_GR1_TLB_INT1_CONFIG_STATUS_MASK 0x0000000000001000UL
+#define UVH_GR1_TLB_INT1_CONFIG_P_SHFT 13
+#define UVH_GR1_TLB_INT1_CONFIG_P_MASK 0x0000000000002000UL
+#define UVH_GR1_TLB_INT1_CONFIG_T_SHFT 15
+#define UVH_GR1_TLB_INT1_CONFIG_T_MASK 0x0000000000008000UL
+#define UVH_GR1_TLB_INT1_CONFIG_M_SHFT 16
+#define UVH_GR1_TLB_INT1_CONFIG_M_MASK 0x0000000000010000UL
+#define UVH_GR1_TLB_INT1_CONFIG_APIC_ID_SHFT 32
+#define UVH_GR1_TLB_INT1_CONFIG_APIC_ID_MASK 0xffffffff00000000UL
+
+union uvh_gr1_tlb_int1_config_u {
+    unsigned long       v;
+    struct uvh_gr1_tlb_int1_config_s {
+        unsigned long   vector_  :  8;  /* RW */
+        unsigned long   dm       :  3;  /* RW */
+        unsigned long   destmode :  1;  /* RW */
+        unsigned long   status   :  1;  /* RO */
+        unsigned long   p        :  1;  /* RO */
+        unsigned long   rsvd_14  :  1;  /*    */
+        unsigned long   t        :  1;  /* RO */
+        unsigned long   m        :  1;  /* RW */
+        unsigned long   rsvd_17_31: 15;  /*    */
+        unsigned long   apic_id  : 32;  /* RW */
+    } s;
+};
+
+/* ========================================================================= */
 /*                               UVH_INT_CMPB                                */
 /* ========================================================================= */
 #define UVH_INT_CMPB 0x22080UL
diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c
index 1bd6da1f8fad..1248318436e8 100644
--- a/arch/x86/kernel/apic/x2apic_uv_x.c
+++ b/arch/x86/kernel/apic/x2apic_uv_x.c
@@ -118,17 +118,12 @@ static int uv_wakeup_secondary(int phys_apicid, unsigned long start_rip)
 
 static void uv_send_IPI_one(int cpu, int vector)
 {
-        unsigned long val, apicid;
+        unsigned long apicid;
         int pnode;
 
         apicid = per_cpu(x86_cpu_to_apicid, cpu);
         pnode = uv_apicid_to_pnode(apicid);
-
-        val = (1UL << UVH_IPI_INT_SEND_SHFT) |
-              (apicid << UVH_IPI_INT_APIC_ID_SHFT) |
-              (vector << UVH_IPI_INT_VECTOR_SHFT);
-
-        uv_write_global_mmr64(pnode, UVH_IPI_INT, val);
+        uv_hub_send_ipi(pnode, apicid, vector);
 }
 
 static void uv_send_IPI_mask(const struct cpumask *mask, int vector)
diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c
index 14014d766cad..76f8f84043a2 100644
--- a/arch/x86/kernel/process_32.c
+++ b/arch/x86/kernel/process_32.c
@@ -245,7 +245,7 @@ void prepare_to_copy(struct task_struct *tsk)
         unlazy_fpu(tsk);
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
+int copy_thread(unsigned long clone_flags, unsigned long sp,
         unsigned long unused,
         struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c
index abb7e6a7f0c6..b751a41392b1 100644
--- a/arch/x86/kernel/process_64.c
+++ b/arch/x86/kernel/process_64.c
@@ -278,7 +278,7 @@ void prepare_to_copy(struct task_struct *tsk)
         unlazy_fpu(tsk);
 }
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long sp,
+int copy_thread(unsigned long clone_flags, unsigned long sp,
                 unsigned long unused,
         struct task_struct *p, struct pt_regs *regs)
 {
diff --git a/arch/x86/kernel/ptrace.c b/arch/x86/kernel/ptrace.c
index 19378715f415..b7cc21bc6ae0 100644
--- a/arch/x86/kernel/ptrace.c
+++ b/arch/x86/kernel/ptrace.c
@@ -1455,6 +1455,6 @@ asmregparm void syscall_trace_leave(struct pt_regs *regs)
          * system call instruction.
          */
         if (test_thread_flag(TIF_SINGLESTEP) &&
-            tracehook_consider_fatal_signal(current, SIGTRAP, SIG_DFL))
+            tracehook_consider_fatal_signal(current, SIGTRAP))
                 send_sigtrap(current, regs, 0, TRAP_BRKPT);
 }
diff --git a/arch/x86/kernel/syscall_table_32.S b/arch/x86/kernel/syscall_table_32.S
index 3bdb64829b82..ff5c8736b491 100644
--- a/arch/x86/kernel/syscall_table_32.S
+++ b/arch/x86/kernel/syscall_table_32.S
@@ -332,3 +332,5 @@ ENTRY(sys_call_table)
         .long sys_dup3                  /* 330 */
         .long sys_pipe2
         .long sys_inotify_init1
+        .long sys_preadv
+        .long sys_pwritev
diff --git a/arch/xtensa/kernel/process.c b/arch/xtensa/kernel/process.c
index 9185597eb6a0..031f36685710 100644
--- a/arch/xtensa/kernel/process.c
+++ b/arch/xtensa/kernel/process.c
@@ -172,7 +172,7 @@ void prepare_to_copy(struct task_struct *tsk)
  *       childregs.
  */
 
-int copy_thread(int nr, unsigned long clone_flags, unsigned long usp,
+int copy_thread(unsigned long clone_flags, unsigned long usp,
                 unsigned long unused,
                 struct task_struct * p, struct pt_regs * regs)
 {