Merge branch 'upstream-fixes' into upstream

29 files changed, 192 insertions, 116 deletions

diff --git a/include/asm-alpha/Kbuild b/include/asm-alpha/Kbuild
index e57fd57538b8..2b06b3bad5ff 100644
--- a/include/asm-alpha/Kbuild
+++ b/include/asm-alpha/Kbuild
@@ -1,5 +1,5 @@
 include include/asm-generic/Kbuild.asm
 
-unifdef-y += console.h fpu.h sysinfo.h
+unifdef-y += console.h fpu.h sysinfo.h compiler.h
 
 header-y += gentrap.h regdef.h pal.h reg.h
diff --git a/include/asm-alpha/compiler.h b/include/asm-alpha/compiler.h
index 00c6f57ad9a7..d2768cc3d7a4 100644
--- a/include/asm-alpha/compiler.h
+++ b/include/asm-alpha/compiler.h
@@ -90,6 +90,7 @@
   __asm__("stw %1,%0" : "=m"(mem) : "r"(val))
 #endif
 
+#ifdef __KERNEL__
 /* Some idiots over in <linux/compiler.h> thought inline should imply
    always_inline.  This breaks stuff.  We'll include this file whenever
    we run into such problems.  */
@@ -101,4 +102,6 @@
 #undef __always_inline
 #define __always_inline         inline __attribute__((always_inline))
 
+#endif /* __KERNEL__ */
+
 #endif /* __ALPHA_COMPILER_H */
diff --git a/include/asm-alpha/page.h b/include/asm-alpha/page.h
index 8c7cd50d4eae..d2bed3cb33ff 100644
--- a/include/asm-alpha/page.h
+++ b/include/asm-alpha/page.h
@@ -1,6 +1,8 @@
 #ifndef _ALPHA_PAGE_H
 #define _ALPHA_PAGE_H
 
+#ifdef __KERNEL__
+
 #include <asm/pal.h>
 
 /* PAGE_SHIFT determines the page size */
@@ -8,8 +10,6 @@
 #define PAGE_SIZE       (1UL << PAGE_SHIFT)
 #define PAGE_MASK       (~(PAGE_SIZE-1))
 
-#ifdef __KERNEL__
-
 #ifndef __ASSEMBLY__
 
 #define STRICT_MM_TYPECHECKS
@@ -92,9 +92,9 @@ typedef unsigned long pgprot_t;
 
 #define VM_DATA_DEFAULT_FLAGS           (VM_READ | VM_WRITE | VM_EXEC | \
                                          VM_MAYREAD | VM_MAYWRITE | VM_MAYEXEC)
-#endif /* __KERNEL__ */
 
 #include <asm-generic/memory_model.h>
 #include <asm-generic/page.h>
 
+#endif /* __KERNEL__ */
 #endif /* _ALPHA_PAGE_H */
diff --git a/include/asm-generic/Kbuild.asm b/include/asm-generic/Kbuild.asm
index 6b16dda18115..c00de6028fa8 100644
--- a/include/asm-generic/Kbuild.asm
+++ b/include/asm-generic/Kbuild.asm
@@ -2,7 +2,7 @@ unifdef-y += a.out.h auxvec.h byteorder.h errno.h fcntl.h ioctl.h	\
         ioctls.h ipcbuf.h mman.h msgbuf.h param.h poll.h                \
         posix_types.h ptrace.h resource.h sembuf.h shmbuf.h shmparam.h  \
         sigcontext.h siginfo.h signal.h socket.h sockios.h stat.h       \
-        statfs.h termbits.h termios.h timex.h types.h unistd.h user.h
+        statfs.h termbits.h termios.h types.h unistd.h user.h
 
 # These probably shouldn't be exported
 unifdef-y += elf.h page.h
diff --git a/include/asm-i386/Kbuild b/include/asm-i386/Kbuild
index 335b2fa4e066..2308190321da 100644
--- a/include/asm-i386/Kbuild
+++ b/include/asm-i386/Kbuild
@@ -1,5 +1,5 @@
 include include/asm-generic/Kbuild.asm
 
-header-y += boot.h debugreg.h ldt.h setup.h ucontext.h
+header-y += boot.h debugreg.h ldt.h ucontext.h
 
-unifdef-y += mtrr.h vm86.h
+unifdef-y += mtrr.h setup.h vm86.h
diff --git a/include/asm-i386/elf.h b/include/asm-i386/elf.h
index 1eac92cb5b16..db4344d9f73f 100644
--- a/include/asm-i386/elf.h
+++ b/include/asm-i386/elf.h
@@ -7,10 +7,7 @@
 
 #include <asm/ptrace.h>
 #include <asm/user.h>
-#include <asm/processor.h>
-#include <asm/system.h>         /* for savesegment */
 #include <asm/auxvec.h>
-#include <asm/desc.h>
 
 #include <linux/utsname.h>
 
@@ -48,6 +45,12 @@ typedef struct user_fxsr_struct elf_fpxregset_t;
 #define ELF_DATA        ELFDATA2LSB
 #define ELF_ARCH        EM_386
 
+#ifdef __KERNEL__
+
+#include <asm/processor.h>
+#include <asm/system.h>         /* for savesegment */
+#include <asm/desc.h>
+
 /* SVR4/i386 ABI (pages 3-31, 3-32) says that when the program starts %edx
    contains a pointer to a function which might be registered using `atexit'.
    This provides a mean for the dynamic linker to call DT_FINI functions for
@@ -111,7 +114,6 @@ typedef struct user_fxsr_struct elf_fpxregset_t;
 
 #define ELF_PLATFORM  (system_utsname.machine)
 
-#ifdef __KERNEL__
 #define SET_PERSONALITY(ex, ibcs2) do { } while (0)
 
 /*
diff --git a/include/asm-i386/setup.h b/include/asm-i386/setup.h
index f737e423029e..2734909eff84 100644
--- a/include/asm-i386/setup.h
+++ b/include/asm-i386/setup.h
@@ -6,6 +6,7 @@
 #ifndef _i386_SETUP_H
 #define _i386_SETUP_H
 
+#ifdef __KERNEL__
 #include <linux/pfn.h>
 
 /*
@@ -13,6 +14,7 @@
  */
 #define MAXMEM_PFN      PFN_DOWN(MAXMEM)
 #define MAX_NONPAE_PFN  (1 << 20)
+#endif
 
 #define PARAM_SIZE 4096
 #define COMMAND_LINE_SIZE 256
diff --git a/include/asm-i386/signal.h b/include/asm-i386/signal.h
index 3824a502351f..c3e8adec5918 100644
--- a/include/asm-i386/signal.h
+++ b/include/asm-i386/signal.h
@@ -2,7 +2,6 @@
 #define _ASMi386_SIGNAL_H
 
 #include <linux/types.h>
-#include <linux/linkage.h>
 #include <linux/time.h>
 #include <linux/compiler.h>
 
@@ -10,6 +9,9 @@
 struct siginfo;
 
 #ifdef __KERNEL__
+
+#include <linux/linkage.h>
+
 /* Most things should be clean enough to redefine this at will, if care
    is taken to make libc match.  */
 
diff --git a/include/asm-i386/unistd.h b/include/asm-i386/unistd.h
index d983b74e4d9f..fc1c8ddae149 100644
--- a/include/asm-i386/unistd.h
+++ b/include/asm-i386/unistd.h
@@ -324,6 +324,8 @@
 #define __NR_vmsplice           316
 #define __NR_move_pages         317
 
+#ifdef __KERNEL__
+
 #define NR_syscalls 318
 
 /*
@@ -423,8 +425,6 @@ __asm__ volatile ("push %%ebp ; push %%ebx ; movl 4(%2),%%ebp ; " \
 __syscall_return(type,__res); \
 }
 
-#ifdef __KERNEL__
-
 #define __ARCH_WANT_IPC_PARSE_VERSION
 #define __ARCH_WANT_OLD_READDIR
 #define __ARCH_WANT_OLD_STAT
diff --git a/include/asm-ia64/Kbuild b/include/asm-ia64/Kbuild
index 85d6f8005eb4..f1cb00f39c22 100644
--- a/include/asm-ia64/Kbuild
+++ b/include/asm-ia64/Kbuild
@@ -4,4 +4,4 @@ header-y += break.h fpu.h fpswa.h gcc_intrin.h ia64regs.h		\
          intel_intrin.h intrinsics.h perfmon_default_smpl.h     \
          ptrace_offsets.h rse.h setup.h ucontext.h
 
-unifdef-y += perfmon.h
+unifdef-y += perfmon.h ustack.h
diff --git a/include/asm-ia64/page.h b/include/asm-ia64/page.h
index f5a949ec6e1e..947cb72b520e 100644
--- a/include/asm-ia64/page.h
+++ b/include/asm-ia64/page.h
@@ -7,6 +7,7 @@
  *      David Mosberger-Tang <davidm@hpl.hp.com>
  */
 
+# ifdef __KERNEL__
 
 #include <asm/intrinsics.h>
 #include <asm/types.h>
@@ -64,7 +65,6 @@
 # define __pa(x)                ((x) - PAGE_OFFSET)
 # define __va(x)                ((x) + PAGE_OFFSET)
 #else /* !__ASSEMBLY */
-# ifdef __KERNEL__
 #  define STRICT_MM_TYPECHECKS
 
 extern void clear_page (void *page);
@@ -174,7 +174,6 @@ get_order (unsigned long size)
         return order;
 }
 
-# endif /* __KERNEL__ */
 #endif /* !__ASSEMBLY__ */
 
 #ifdef STRICT_MM_TYPECHECKS
@@ -228,4 +227,5 @@ get_order (unsigned long size)
                                          (((current->personality & READ_IMPLIES_EXEC) != 0)     \
                                           ? VM_EXEC : 0))
 
+# endif /* __KERNEL__ */
 #endif /* _ASM_IA64_PAGE_H */
diff --git a/include/asm-ia64/ptrace.h b/include/asm-ia64/ptrace.h
index 415abb23b210..1414316efd40 100644
--- a/include/asm-ia64/ptrace.h
+++ b/include/asm-ia64/ptrace.h
@@ -56,6 +56,8 @@
 
 
 #include <asm/fpu.h>
+
+#ifdef __KERNEL__
 #ifndef ASM_OFFSETS_C
 #include <asm/asm-offsets.h>
 #endif
@@ -79,10 +81,9 @@
 
 #define KERNEL_STACK_SIZE               IA64_STK_OFFSET
 
-#ifndef __ASSEMBLY__
+#endif /* __KERNEL__ */
 
-#include <asm/current.h>
-#include <asm/page.h>
+#ifndef __ASSEMBLY__
 
 /*
  * This struct defines the way the registers are saved on system
@@ -229,6 +230,9 @@ struct switch_stack {
 
 #ifdef __KERNEL__
 
+#include <asm/current.h>
+#include <asm/page.h>
+
 #define __ARCH_SYS_PTRACE       1
 
 /*
diff --git a/include/asm-ia64/ustack.h b/include/asm-ia64/ustack.h
index da55c91246e3..a349467913ea 100644
--- a/include/asm-ia64/ustack.h
+++ b/include/asm-ia64/ustack.h
@@ -5,12 +5,15 @@
  * Constants for the user stack size
  */
 
+#ifdef __KERNEL__
 #include <asm/page.h>
 
 /* The absolute hard limit for stack size is 1/2 of the mappable space in the region */
 #define MAX_USER_STACK_SIZE     (RGN_MAP_LIMIT/2)
-/* Make a default stack size of 2GB */
-#define DEFAULT_USER_STACK_SIZE (1UL << 31)
 #define STACK_TOP               (0x6000000000000000UL + RGN_MAP_LIMIT)
+#endif
+
+/* Make a default stack size of 2GiB */
+#define DEFAULT_USER_STACK_SIZE (1UL << 31)
 
 #endif /* _ASM_IA64_USTACK_H */
diff --git a/include/asm-mips/page.h b/include/asm-mips/page.h
index 6ed1151a05a3..219d359861f3 100644
--- a/include/asm-mips/page.h
+++ b/include/asm-mips/page.h
@@ -14,8 +14,6 @@
 
 #include <spaces.h>
 
-#endif
-
 /*
  * PAGE_SHIFT determines the page size
  */
@@ -34,8 +32,6 @@
 #define PAGE_SIZE       (1UL << PAGE_SHIFT)
 #define PAGE_MASK       (~((1 << PAGE_SHIFT) - 1))
 
-
-#ifdef __KERNEL__
 #ifndef __ASSEMBLY__
 
 extern void clear_page(void * page);
@@ -168,8 +164,6 @@ typedef struct { unsigned long pgprot; } pgprot_t;
 #define UNCAC_ADDR(addr)        ((addr) - PAGE_OFFSET + UNCAC_BASE)
 #define CAC_ADDR(addr)          ((addr) - UNCAC_BASE + PAGE_OFFSET)
 
-#endif /* defined (__KERNEL__) */
-
 #ifdef CONFIG_LIMITED_DMA
 #define WANT_PAGE_VIRTUAL
 #endif
@@ -177,4 +171,6 @@ typedef struct { unsigned long pgprot; } pgprot_t;
 #include <asm-generic/memory_model.h>
 #include <asm-generic/page.h>
 
+#endif /* defined (__KERNEL__) */
+
 #endif /* _ASM_PAGE_H */
diff --git a/include/asm-powerpc/eeh.h b/include/asm-powerpc/eeh.h
index 4df3e80118f4..6a784396660b 100644
--- a/include/asm-powerpc/eeh.h
+++ b/include/asm-powerpc/eeh.h
@@ -205,6 +205,7 @@ static inline void eeh_memset_io(volatile void __iomem *addr, int c,
         lc |= lc << 8;
         lc |= lc << 16;
 
+        __asm__ __volatile__ ("sync" : : : "memory");
         while(n && !EEH_CHECK_ALIGN(p, 4)) {
                 *((volatile u8 *)p) = c;
                 p++;
@@ -229,6 +230,7 @@ static inline void eeh_memcpy_fromio(void *dest, const volatile void __iomem *sr
         void *destsave = dest;
         unsigned long nsave = n;
 
+        __asm__ __volatile__ ("sync" : : : "memory");
         while(n && (!EEH_CHECK_ALIGN(vsrc, 4) || !EEH_CHECK_ALIGN(dest, 4))) {
                 *((u8 *)dest) = *((volatile u8 *)vsrc);
                 __asm__ __volatile__ ("eieio" : : : "memory");
@@ -266,6 +268,7 @@ static inline void eeh_memcpy_toio(volatile void __iomem *dest, const void *src,
 {
         void *vdest = (void __force *) dest;
 
+        __asm__ __volatile__ ("sync" : : : "memory");
         while(n && (!EEH_CHECK_ALIGN(vdest, 4) || !EEH_CHECK_ALIGN(src, 4))) {
                 *((volatile u8 *)vdest) = *((u8 *)src);
                 src++;
diff --git a/include/asm-powerpc/futex.h b/include/asm-powerpc/futex.h
index f1b3c00bc1ce..936422e54891 100644
--- a/include/asm-powerpc/futex.h
+++ b/include/asm-powerpc/futex.h
@@ -84,7 +84,33 @@ static inline int futex_atomic_op_inuser (int encoded_op, int __user *uaddr)
 static inline int
 futex_atomic_cmpxchg_inatomic(int __user *uaddr, int oldval, int newval)
 {
-        return -ENOSYS;
+        int prev;
+
+        if (!access_ok(VERIFY_WRITE, uaddr, sizeof(int)))
+                return -EFAULT;
+
+        __asm__ __volatile__ (
+        LWSYNC_ON_SMP
+"1:     lwarx   %0,0,%2         # futex_atomic_cmpxchg_inatomic\n\
+        cmpw    0,%0,%3\n\
+        bne-    3f\n"
+        PPC405_ERR77(0,%2)
+"2:     stwcx.  %4,0,%2\n\
+        bne-    1b\n"
+        ISYNC_ON_SMP
+"3:     .section .fixup,\"ax\"\n\
+4:      li      %0,%5\n\
+        b       3b\n\
+        .previous\n\
+        .section __ex_table,\"a\"\n\
+        .align 3\n\
+        " PPC_LONG "1b,4b,2b,4b\n\
+        .previous" \
+        : "=&r" (prev), "+m" (*uaddr)
+        : "r" (uaddr), "r" (oldval), "r" (newval), "i" (-EFAULT)
+        : "cc", "memory");
+
+        return prev;
 }
 
 #endif /* __KERNEL__ */
diff --git a/include/asm-powerpc/io.h b/include/asm-powerpc/io.h
index 36c4c34bf565..212428db0d8b 100644
--- a/include/asm-powerpc/io.h
+++ b/include/asm-powerpc/io.h
@@ -19,6 +19,7 @@ extern int check_legacy_ioport(unsigned long base_port);
 #include <linux/compiler.h>
 #include <asm/page.h>
 #include <asm/byteorder.h>
+#include <asm/paca.h>
 #ifdef CONFIG_PPC_ISERIES 
 #include <asm/iseries/iseries_io.h>
 #endif  
@@ -162,7 +163,11 @@ extern void _outsw_ns(volatile u16 __iomem *port, const void *buf, int ns);
 extern void _insl_ns(volatile u32 __iomem *port, void *buf, int nl);
 extern void _outsl_ns(volatile u32 __iomem *port, const void *buf, int nl);
 
-#define mmiowb()
+static inline void mmiowb(void)
+{
+        __asm__ __volatile__ ("sync" : : : "memory");
+        get_paca()->io_sync = 0;
+}
 
 /*
  * output pause versions need a delay at least for the
@@ -278,22 +283,23 @@ static inline int in_8(const volatile unsigned char __iomem *addr)
 {
         int ret;
 
-        __asm__ __volatile__("lbz%U1%X1 %0,%1; twi 0,%0,0; isync"
+        __asm__ __volatile__("sync; lbz%U1%X1 %0,%1; twi 0,%0,0; isync"
                              : "=r" (ret) : "m" (*addr));
         return ret;
 }
 
 static inline void out_8(volatile unsigned char __iomem *addr, int val)
 {
-        __asm__ __volatile__("stb%U0%X0 %1,%0; sync"
+        __asm__ __volatile__("sync; stb%U0%X0 %1,%0"
                              : "=m" (*addr) : "r" (val));
+        get_paca()->io_sync = 1;
 }
 
 static inline int in_le16(const volatile unsigned short __iomem *addr)
 {
         int ret;
 
-        __asm__ __volatile__("lhbrx %0,0,%1; twi 0,%0,0; isync"
+        __asm__ __volatile__("sync; lhbrx %0,0,%1; twi 0,%0,0; isync"
                              : "=r" (ret) : "r" (addr), "m" (*addr));
         return ret;
 }
@@ -302,28 +308,30 @@ static inline int in_be16(const volatile unsigned short __iomem *addr)
 {
         int ret;
 
-        __asm__ __volatile__("lhz%U1%X1 %0,%1; twi 0,%0,0; isync"
+        __asm__ __volatile__("sync; lhz%U1%X1 %0,%1; twi 0,%0,0; isync"
                              : "=r" (ret) : "m" (*addr));
         return ret;
 }
 
 static inline void out_le16(volatile unsigned short __iomem *addr, int val)
 {
-        __asm__ __volatile__("sthbrx %1,0,%2; sync"
+        __asm__ __volatile__("sync; sthbrx %1,0,%2"
                              : "=m" (*addr) : "r" (val), "r" (addr));
+        get_paca()->io_sync = 1;
 }
 
 static inline void out_be16(volatile unsigned short __iomem *addr, int val)
 {
-        __asm__ __volatile__("sth%U0%X0 %1,%0; sync"
+        __asm__ __volatile__("sync; sth%U0%X0 %1,%0"
                              : "=m" (*addr) : "r" (val));
+        get_paca()->io_sync = 1;
 }
 
 static inline unsigned in_le32(const volatile unsigned __iomem *addr)
 {
         unsigned ret;
 
-        __asm__ __volatile__("lwbrx %0,0,%1; twi 0,%0,0; isync"
+        __asm__ __volatile__("sync; lwbrx %0,0,%1; twi 0,%0,0; isync"
                              : "=r" (ret) : "r" (addr), "m" (*addr));
         return ret;
 }
@@ -332,21 +340,23 @@ static inline unsigned in_be32(const volatile unsigned __iomem *addr)
 {
         unsigned ret;
 
-        __asm__ __volatile__("lwz%U1%X1 %0,%1; twi 0,%0,0; isync"
+        __asm__ __volatile__("sync; lwz%U1%X1 %0,%1; twi 0,%0,0; isync"
                              : "=r" (ret) : "m" (*addr));
         return ret;
 }
 
 static inline void out_le32(volatile unsigned __iomem *addr, int val)
 {
-        __asm__ __volatile__("stwbrx %1,0,%2; sync" : "=m" (*addr)
+        __asm__ __volatile__("sync; stwbrx %1,0,%2" : "=m" (*addr)
                              : "r" (val), "r" (addr));
+        get_paca()->io_sync = 1;
 }
 
 static inline void out_be32(volatile unsigned __iomem *addr, int val)
 {
-        __asm__ __volatile__("stw%U0%X0 %1,%0; sync"
+        __asm__ __volatile__("sync; stw%U0%X0 %1,%0"
                              : "=m" (*addr) : "r" (val));
+        get_paca()->io_sync = 1;
 }
 
 static inline unsigned long in_le64(const volatile unsigned long __iomem *addr)
@@ -354,6 +364,7 @@ static inline unsigned long in_le64(const volatile unsigned long __iomem *addr)
         unsigned long tmp, ret;
 
         __asm__ __volatile__(
+                             "sync\n"
                              "ld %1,0(%2)\n"
                              "twi 0,%1,0\n"
                              "isync\n"
@@ -372,7 +383,7 @@ static inline unsigned long in_be64(const volatile unsigned long __iomem *addr)
 {
         unsigned long ret;
 
-        __asm__ __volatile__("ld%U1%X1 %0,%1; twi 0,%0,0; isync"
+        __asm__ __volatile__("sync; ld%U1%X1 %0,%1; twi 0,%0,0; isync"
                              : "=r" (ret) : "m" (*addr));
         return ret;
 }
@@ -389,14 +400,16 @@ static inline void out_le64(volatile unsigned long __iomem *addr, unsigned long
                              "rldicl %1,%1,32,0\n"
                              "rlwimi %0,%1,8,8,31\n"
                              "rlwimi %0,%1,24,16,23\n"
-                             "std %0,0(%3)\n"
-                             "sync"
+                             "sync\n"
+                             "std %0,0(%3)"
                              : "=&r" (tmp) , "=&r" (val) : "1" (val) , "b" (addr) , "m" (*addr));
+        get_paca()->io_sync = 1;
 }
 
 static inline void out_be64(volatile unsigned long __iomem *addr, unsigned long val)
 {
-        __asm__ __volatile__("std%U0%X0 %1,%0; sync" : "=m" (*addr) : "r" (val));
+        __asm__ __volatile__("sync; std%U0%X0 %1,%0" : "=m" (*addr) : "r" (val));
+        get_paca()->io_sync = 1;
 }
 
 #ifndef CONFIG_PPC_ISERIES 
diff --git a/include/asm-powerpc/kdump.h b/include/asm-powerpc/kdump.h
index dc1574c945f8..10e8eb1e6f4f 100644
--- a/include/asm-powerpc/kdump.h
+++ b/include/asm-powerpc/kdump.h
@@ -7,7 +7,7 @@
 /* How many bytes to reserve at zero for kdump. The reserve limit should
  * be greater or equal to the trampoline's end address.
  * Reserve to the end of the FWNMI area, see head_64.S */
-#define KDUMP_RESERVE_LIMIT     0x8000
+#define KDUMP_RESERVE_LIMIT     0x10000 /* 64K */
 
 #ifdef CONFIG_CRASH_DUMP
 
diff --git a/include/asm-powerpc/paca.h b/include/asm-powerpc/paca.h
index 2d4585f06209..3d5d590bc4b0 100644
--- a/include/asm-powerpc/paca.h
+++ b/include/asm-powerpc/paca.h
@@ -93,6 +93,7 @@ struct paca_struct {
         u64 saved_r1;                   /* r1 save for RTAS calls */
         u64 saved_msr;                  /* MSR saved here by enter_rtas */
         u8 proc_enabled;                /* irq soft-enable flag */
+        u8 io_sync;                     /* writel() needs spin_unlock sync */
 
         /* Stuff for accurate time accounting */
         u64 user_time;                  /* accumulated usermode TB ticks */
diff --git a/include/asm-powerpc/spinlock.h b/include/asm-powerpc/spinlock.h
index 895cb6d3a42a..c31e4382a775 100644
--- a/include/asm-powerpc/spinlock.h
+++ b/include/asm-powerpc/spinlock.h
@@ -36,6 +36,19 @@
 #define LOCK_TOKEN      1
 #endif
 
+#if defined(CONFIG_PPC64) && defined(CONFIG_SMP)
+#define CLEAR_IO_SYNC   (get_paca()->io_sync = 0)
+#define SYNC_IO         do {                                            \
+                                if (unlikely(get_paca()->io_sync)) {    \
+                                        mb();                           \
+                                        get_paca()->io_sync = 0;        \
+                                }                                       \
+                        } while (0)
+#else
+#define CLEAR_IO_SYNC
+#define SYNC_IO
+#endif
+
 /*
  * This returns the old value in the lock, so we succeeded
  * in getting the lock if the return value is 0.
@@ -61,6 +74,7 @@ static __inline__ unsigned long __spin_trylock(raw_spinlock_t *lock)
 
 static int __inline__ __raw_spin_trylock(raw_spinlock_t *lock)
 {
+        CLEAR_IO_SYNC;
         return __spin_trylock(lock) == 0;
 }
 
@@ -91,6 +105,7 @@ extern void __rw_yield(raw_rwlock_t *lock);
 
 static void __inline__ __raw_spin_lock(raw_spinlock_t *lock)
 {
+        CLEAR_IO_SYNC;
         while (1) {
                 if (likely(__spin_trylock(lock) == 0))
                         break;
@@ -107,6 +122,7 @@ static void __inline__ __raw_spin_lock_flags(raw_spinlock_t *lock, unsigned long
 {
         unsigned long flags_dis;
 
+        CLEAR_IO_SYNC;
         while (1) {
                 if (likely(__spin_trylock(lock) == 0))
                         break;
@@ -124,6 +140,7 @@ static void __inline__ __raw_spin_lock_flags(raw_spinlock_t *lock, unsigned long
 
 static __inline__ void __raw_spin_unlock(raw_spinlock_t *lock)
 {
+        SYNC_IO;
         __asm__ __volatile__("# __raw_spin_unlock\n\t"
                                 LWSYNC_ON_SMP: : :"memory");
         lock->slock = 0;
diff --git a/include/asm-ppc/io.h b/include/asm-ppc/io.h
index 89c6f1bc3aab..680555be22ec 100644
--- a/include/asm-ppc/io.h
+++ b/include/asm-ppc/io.h
@@ -63,7 +63,7 @@ extern inline int in_8(const volatile unsigned char __iomem *addr)
         int ret;
 
         __asm__ __volatile__(
-                "lbz%U1%X1 %0,%1;\n"
+                "sync; lbz%U1%X1 %0,%1;\n"
                 "twi 0,%0,0;\n"
                 "isync" : "=r" (ret) : "m" (*addr));
         return ret;
@@ -78,7 +78,7 @@ extern inline int in_le16(const volatile unsigned short __iomem *addr)
 {
         int ret;
 
-        __asm__ __volatile__("lhbrx %0,0,%1;\n"
+        __asm__ __volatile__("sync; lhbrx %0,0,%1;\n"
                              "twi 0,%0,0;\n"
                              "isync" : "=r" (ret) :
                               "r" (addr), "m" (*addr));
@@ -89,7 +89,7 @@ extern inline int in_be16(const volatile unsigned short __iomem *addr)
 {
         int ret;
 
-        __asm__ __volatile__("lhz%U1%X1 %0,%1;\n"
+        __asm__ __volatile__("sync; lhz%U1%X1 %0,%1;\n"
                              "twi 0,%0,0;\n"
                              "isync" : "=r" (ret) : "m" (*addr));
         return ret;
@@ -97,20 +97,20 @@ extern inline int in_be16(const volatile unsigned short __iomem *addr)
 
 extern inline void out_le16(volatile unsigned short __iomem *addr, int val)
 {
-        __asm__ __volatile__("sthbrx %1,0,%2; eieio" : "=m" (*addr) :
+        __asm__ __volatile__("sync; sthbrx %1,0,%2" : "=m" (*addr) :
                               "r" (val), "r" (addr));
 }
 
 extern inline void out_be16(volatile unsigned short __iomem *addr, int val)
 {
-        __asm__ __volatile__("sth%U0%X0 %1,%0; eieio" : "=m" (*addr) : "r" (val));
+        __asm__ __volatile__("sync; sth%U0%X0 %1,%0" : "=m" (*addr) : "r" (val));
 }
 
 extern inline unsigned in_le32(const volatile unsigned __iomem *addr)
 {
         unsigned ret;
 
-        __asm__ __volatile__("lwbrx %0,0,%1;\n"
+        __asm__ __volatile__("sync; lwbrx %0,0,%1;\n"
                              "twi 0,%0,0;\n"
                              "isync" : "=r" (ret) :
                              "r" (addr), "m" (*addr));
@@ -121,7 +121,7 @@ extern inline unsigned in_be32(const volatile unsigned __iomem *addr)
 {
         unsigned ret;
 
-        __asm__ __volatile__("lwz%U1%X1 %0,%1;\n"
+        __asm__ __volatile__("sync; lwz%U1%X1 %0,%1;\n"
                              "twi 0,%0,0;\n"
                              "isync" : "=r" (ret) : "m" (*addr));
         return ret;
@@ -129,13 +129,13 @@ extern inline unsigned in_be32(const volatile unsigned __iomem *addr)
 
 extern inline void out_le32(volatile unsigned __iomem *addr, int val)
 {
-        __asm__ __volatile__("stwbrx %1,0,%2; eieio" : "=m" (*addr) :
+        __asm__ __volatile__("sync; stwbrx %1,0,%2" : "=m" (*addr) :
                              "r" (val), "r" (addr));
 }
 
 extern inline void out_be32(volatile unsigned __iomem *addr, int val)
 {
-        __asm__ __volatile__("stw%U0%X0 %1,%0; eieio" : "=m" (*addr) : "r" (val));
+        __asm__ __volatile__("sync; stw%U0%X0 %1,%0" : "=m" (*addr) : "r" (val));
 }
 #if defined (CONFIG_8260_PCI9)
 #define readb(addr) in_8((volatile u8 *)(addr))
@@ -259,6 +259,7 @@ extern __inline__ unsigned int name(unsigned int port)	\
 {                                                       \
         unsigned int x;                                 \
         __asm__ __volatile__(                           \
+                "sync\n"                                \
                 "0:"    op "    %0,0,%1\n"              \
                 "1:     twi     0,%0,0\n"               \
                 "2:     isync\n"                        \
@@ -284,6 +285,7 @@ extern __inline__ unsigned int name(unsigned int port)	\
 extern __inline__ void name(unsigned int val, unsigned int port) \
 {                                                       \
         __asm__ __volatile__(                           \
+                "sync\n"                                \
                 "0:" op " %0,0,%1\n"                    \
                 "1:     sync\n"                         \
                 "2:\n"                                  \
diff --git a/include/asm-s390/debug.h b/include/asm-s390/debug.h
index 7f1ef99fd1e1..c00dd2b3dc50 100644
--- a/include/asm-s390/debug.h
+++ b/include/asm-s390/debug.h
@@ -10,7 +10,6 @@
 #define DEBUG_H
 
 #include <linux/fs.h>
-#include <linux/string.h>
 
 /* Note:
  * struct __debug_entry must be defined outside of #ifdef __KERNEL__ 
@@ -35,6 +34,7 @@ struct __debug_entry{
 #define __DEBUG_FEATURE_VERSION      2  /* version of debug feature */
 
 #ifdef __KERNEL__
+#include <linux/string.h>
 #include <linux/spinlock.h>
 #include <linux/kernel.h>
 #include <linux/time.h>
diff --git a/include/asm-s390/elf.h b/include/asm-s390/elf.h
index 710646e64f7d..c0d629d61d3e 100644
--- a/include/asm-s390/elf.h
+++ b/include/asm-s390/elf.h
@@ -93,19 +93,6 @@
 #define R_390_NUM       61
 
 /*
- * ELF register definitions..
- */
-
-#include <linux/sched.h>        /* for task_struct */
-#include <asm/ptrace.h>
-#include <asm/user.h>
-#include <asm/system.h>         /* for save_access_regs */
-
-
-typedef s390_fp_regs elf_fpregset_t;
-typedef s390_regs elf_gregset_t;
-
-/*
  * These are used to set parameters in the core dumps.
  */
 #ifndef __s390x__
@@ -117,6 +104,20 @@ typedef s390_regs elf_gregset_t;
 #define ELF_ARCH        EM_S390
 
 /*
+ * ELF register definitions..
+ */
+
+#include <asm/ptrace.h>
+#include <asm/user.h>
+
+typedef s390_fp_regs elf_fpregset_t;
+typedef s390_regs elf_gregset_t;
+
+#ifdef __KERNEL__
+#include <linux/sched.h>        /* for task_struct */
+#include <asm/system.h>         /* for save_access_regs */
+
+/*
  * This is used to ensure we don't load something for the wrong architecture.
  */
 #define elf_check_arch(x) \
@@ -198,7 +199,6 @@ static inline int dump_task_fpu(struct task_struct *tsk, elf_fpregset_t *fpregs)
 
 #define ELF_PLATFORM (NULL)
 
-#ifdef __KERNEL__
 #ifndef __s390x__
 #define SET_PERSONALITY(ex, ibcs2) set_personality((ibcs2)?PER_SVR4:PER_LINUX)
 #else /* __s390x__ */
diff --git a/include/asm-x86_64/elf.h b/include/asm-x86_64/elf.h
index b4f8f4a41a6e..a406fcb1e924 100644
--- a/include/asm-x86_64/elf.h
+++ b/include/asm-x86_64/elf.h
@@ -7,8 +7,6 @@
 
 #include <asm/ptrace.h>
 #include <asm/user.h>
-#include <asm/processor.h>
-#include <asm/compat.h>
 
 /* x86-64 relocation types */
 #define R_X86_64_NONE           0       /* No reloc */
@@ -39,18 +37,23 @@ typedef elf_greg_t elf_gregset_t[ELF_NGREG];
 typedef struct user_i387_struct elf_fpregset_t;
 
 /*
- * This is used to ensure we don't load something for the wrong architecture.
- */
-#define elf_check_arch(x) \
-        ((x)->e_machine == EM_X86_64)
-
-/*
  * These are used to set parameters in the core dumps.
  */
 #define ELF_CLASS       ELFCLASS64
 #define ELF_DATA        ELFDATA2LSB
 #define ELF_ARCH        EM_X86_64
 
+#ifdef __KERNEL__
+#include <asm/processor.h>
+#include <asm/compat.h>
+
+/*
+ * This is used to ensure we don't load something for the wrong architecture.
+ */
+#define elf_check_arch(x) \
+        ((x)->e_machine == EM_X86_64)
+
+
 /* SVR4/i386 ABI (pages 3-31, 3-32) says that when the program starts %edx
    contains a pointer to a function which might be registered using `atexit'.
    This provides a mean for the dynamic linker to call DT_FINI functions for
@@ -141,7 +144,6 @@ typedef struct user_i387_struct elf_fpregset_t;
 /* I'm not sure if we can use '-' here */
 #define ELF_PLATFORM  ("x86_64")
 
-#ifdef __KERNEL__
 extern void set_personality_64bit(void);
 #define SET_PERSONALITY(ex, ibcs2) set_personality_64bit()
 /*
diff --git a/include/asm-x86_64/signal.h b/include/asm-x86_64/signal.h
index cef7a7d51b7e..3ede2a61973a 100644
--- a/include/asm-x86_64/signal.h
+++ b/include/asm-x86_64/signal.h
@@ -3,13 +3,13 @@
 
 #ifndef __ASSEMBLY__
 #include <linux/types.h>
-#include <linux/linkage.h>
 #include <linux/time.h>
 
 /* Avoid too many header ordering problems.  */
 struct siginfo;
 
 #ifdef __KERNEL__
+#include <linux/linkage.h>
 /* Most things should be clean enough to redefine this at will, if care
    is taken to make libc match.  */
 
diff --git a/include/asm-x86_64/unistd.h b/include/asm-x86_64/unistd.h
index 2d89d309a2a8..80fd48e84bbb 100644
--- a/include/asm-x86_64/unistd.h
+++ b/include/asm-x86_64/unistd.h
@@ -620,6 +620,8 @@ __SYSCALL(__NR_vmsplice, sys_vmsplice)
 #define __NR_move_pages         279
 __SYSCALL(__NR_move_pages, sys_move_pages)
 
+#ifdef __KERNEL__
+
 #define __NR_syscall_max __NR_move_pages
 
 #ifndef __NO_STUBS
@@ -744,8 +746,6 @@ __syscall_return(type,__res); \
 
 #else /* __KERNEL_SYSCALLS__ */
 
-#ifdef __KERNEL__
-
 #include <linux/syscalls.h>
 #include <asm/ptrace.h>
 
@@ -821,8 +821,6 @@ asmlinkage long sys_fork(struct pt_regs regs);
 asmlinkage long sys_vfork(struct pt_regs regs);
 asmlinkage long sys_pipe(int *fildes);
 
-#endif /* __KERNEL_SYSCALLS__ */
-
 #ifndef __ASSEMBLY__
 
 #include <linux/linkage.h>
@@ -838,9 +836,9 @@ asmlinkage long sys_rt_sigaction(int sig,
                                 struct sigaction __user *oact,
                                 size_t sigsetsize);
 
-#endif
+#endif  /* __ASSEMBLY__ */
 
-#endif
+#endif /* __KERNEL_SYSCALLS__ */
 
 /*
  * "Conditional" syscalls
@@ -850,6 +848,8 @@ asmlinkage long sys_rt_sigaction(int sig,
  */
 #define cond_syscall(x) asm(".weak\t" #x "\n\t.set\t" #x ",sys_ni_syscall")
 
-#endif
+#endif /* __NO_STUBS */
 
-#endif
+#endif /* __KERNEL__ */
+
+#endif /* _ASM_X86_64_UNISTD_H_ */
diff --git a/include/asm-x86_64/vsyscall.h b/include/asm-x86_64/vsyscall.h
index a85e16f56d73..146b24402a5f 100644
--- a/include/asm-x86_64/vsyscall.h
+++ b/include/asm-x86_64/vsyscall.h
@@ -1,8 +1,6 @@
 #ifndef _ASM_X86_64_VSYSCALL_H_
 #define _ASM_X86_64_VSYSCALL_H_
 
-#include <linux/seqlock.h>
-
 enum vsyscall_num {
         __NR_vgettimeofday,
         __NR_vtime,
@@ -14,6 +12,7 @@ enum vsyscall_num {
 #define VSYSCALL_ADDR(vsyscall_nr) (VSYSCALL_START+VSYSCALL_SIZE*(vsyscall_nr))
 
 #ifdef __KERNEL__
+#include <linux/seqlock.h>
 
 #define __section_vxtime __attribute__ ((unused, __section__ (".vxtime"), aligned(16)))
 #define __section_wall_jiffies __attribute__ ((unused, __section__ (".wall_jiffies"), aligned(16)))
diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h
index 530b1e6173b1..6c2066caeaab 100644
--- a/include/linux/nfs_fs.h
+++ b/include/linux/nfs_fs.h
@@ -9,27 +9,6 @@
 #ifndef _LINUX_NFS_FS_H
 #define _LINUX_NFS_FS_H
 
-#include <linux/in.h>
-#include <linux/mm.h>
-#include <linux/pagemap.h>
-#include <linux/rwsem.h>
-#include <linux/wait.h>
-
-#include <linux/sunrpc/debug.h>
-#include <linux/sunrpc/auth.h>
-#include <linux/sunrpc/clnt.h>
-
-#include <linux/nfs.h>
-#include <linux/nfs2.h>
-#include <linux/nfs3.h>
-#include <linux/nfs4.h>
-#include <linux/nfs_xdr.h>
-
-#include <linux/nfs_fs_sb.h>
-
-#include <linux/rwsem.h>
-#include <linux/mempool.h>
-
 /*
  * Enable debugging support for nfs client.
  * Requires RPC_DEBUG.
@@ -48,11 +27,6 @@
 #define NFS_SUPER_MAGIC                 0x6969
 
 /*
- * These are the default flags for swap requests
- */
-#define NFS_RPC_SWAPFLAGS               (RPC_TASK_SWAPPER|RPC_TASK_ROOTCREDS)
-
-/*
  * When flushing a cluster of dirty pages, there can be different
  * strategies:
  */
@@ -65,6 +39,32 @@
 
 #ifdef __KERNEL__
 
+#include <linux/in.h>
+#include <linux/mm.h>
+#include <linux/pagemap.h>
+#include <linux/rwsem.h>
+#include <linux/wait.h>
+
+#include <linux/sunrpc/debug.h>
+#include <linux/sunrpc/auth.h>
+#include <linux/sunrpc/clnt.h>
+
+#include <linux/nfs.h>
+#include <linux/nfs2.h>
+#include <linux/nfs3.h>
+#include <linux/nfs4.h>
+#include <linux/nfs_xdr.h>
+
+#include <linux/nfs_fs_sb.h>
+
+#include <linux/rwsem.h>
+#include <linux/mempool.h>
+
+/*
+ * These are the default flags for swap requests
+ */
+#define NFS_RPC_SWAPFLAGS               (RPC_TASK_SWAPPER|RPC_TASK_ROOTCREDS)
+
 /*
  * NFSv3/v4 Access mode cache entry
  */
diff --git a/include/linux/timex.h b/include/linux/timex.h
index 19bb6538b49e..d543d3871e38 100644
--- a/include/linux/timex.h
+++ b/include/linux/timex.h
@@ -57,7 +57,6 @@
 #include <linux/time.h>
 
 #include <asm/param.h>
-#include <asm/timex.h>
 
 /*
  * SHIFT_KG and SHIFT_KF establish the damping of the PLL and are chosen
@@ -191,6 +190,8 @@ struct timex {
 #define TIME_BAD        TIME_ERROR /* bw compat */
 
 #ifdef __KERNEL__
+#include <asm/timex.h>
+
 /*
  * kernel variables
  * Note: maximum error = NTP synch distance = dispersion + delay / 2;