Merge branch 'next' into for-linus

82 files changed, 437 insertions, 1166 deletions

diff --git a/arch/sparc/Kconfig b/arch/sparc/Kconfig
index 1a6f20d4e7e6..f92602e86607 100644
--- a/arch/sparc/Kconfig
+++ b/arch/sparc/Kconfig
@@ -560,12 +560,7 @@ config SUN_OPENPROMFS
           Only choose N if you know in advance that you will not need to modify
           OpenPROM settings on the running system.
 
-# Makefile helpers
-config SPARC32_PCI
-        bool
-        default y
-        depends on SPARC32 && PCI
-
+# Makefile helper
 config SPARC64_PCI
         bool
         default y
diff --git a/arch/sparc/include/asm/compat.h b/arch/sparc/include/asm/compat.h
index 6f57325bb883..b8be20d42a0a 100644
--- a/arch/sparc/include/asm/compat.h
+++ b/arch/sparc/include/asm/compat.h
@@ -134,7 +134,8 @@ struct compat_statfs {
         compat_fsid_t   f_fsid;
         int             f_namelen;      /* SunOS ignores this field. */
         int             f_frsize;
-        int             f_spare[5];
+        int             f_flags;
+        int             f_spare[4];
 };
 
 #define COMPAT_RLIM_INFINITY 0x7fffffff
diff --git a/arch/sparc/include/asm/pgtable_32.h b/arch/sparc/include/asm/pgtable_32.h
index 5b31a8e89823..a790cc657476 100644
--- a/arch/sparc/include/asm/pgtable_32.h
+++ b/arch/sparc/include/asm/pgtable_32.h
@@ -431,10 +431,6 @@ extern unsigned long *sparc_valid_addr_bitmap;
 #define kern_addr_valid(addr) \
         (test_bit(__pa((unsigned long)(addr))>>20, sparc_valid_addr_bitmap))
 
-extern int io_remap_pfn_range(struct vm_area_struct *vma,
-                              unsigned long from, unsigned long pfn,
-                              unsigned long size, pgprot_t prot);
-
 /*
  * For sparc32&64, the pfn in io_remap_pfn_range() carries <iospace> in
  * its high 4 bits.  These macros/functions put it there or get it from there.
@@ -443,6 +439,22 @@ extern int io_remap_pfn_range(struct vm_area_struct *vma,
 #define GET_IOSPACE(pfn)                (pfn >> (BITS_PER_LONG - 4))
 #define GET_PFN(pfn)                    (pfn & 0x0fffffffUL)
 
+extern int remap_pfn_range(struct vm_area_struct *, unsigned long, unsigned long,
+                           unsigned long, pgprot_t);
+
+static inline int io_remap_pfn_range(struct vm_area_struct *vma,
+                                     unsigned long from, unsigned long pfn,
+                                     unsigned long size, pgprot_t prot)
+{
+        unsigned long long offset, space, phys_base;
+
+        offset = ((unsigned long long) GET_PFN(pfn)) << PAGE_SHIFT;
+        space = GET_IOSPACE(pfn);
+        phys_base = offset | (space << 32ULL);
+
+        return remap_pfn_range(vma, from, phys_base >> PAGE_SHIFT, size, prot);
+}
+
 #define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
 #define ptep_set_access_flags(__vma, __address, __ptep, __entry, __dirty) \
 ({                                                                        \
diff --git a/arch/sparc/include/asm/pgtable_64.h b/arch/sparc/include/asm/pgtable_64.h
index adf89329af59..38ebb2c60137 100644
--- a/arch/sparc/include/asm/pgtable_64.h
+++ b/arch/sparc/include/asm/pgtable_64.h
@@ -757,10 +757,6 @@ static inline bool kern_addr_valid(unsigned long addr)
 
 extern int page_in_phys_avail(unsigned long paddr);
 
-extern int io_remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
-                               unsigned long pfn,
-                               unsigned long size, pgprot_t prot);
-
 /*
  * For sparc32&64, the pfn in io_remap_pfn_range() carries <iospace> in
  * its high 4 bits.  These macros/functions put it there or get it from there.
@@ -769,6 +765,22 @@ extern int io_remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
 #define GET_IOSPACE(pfn)                (pfn >> (BITS_PER_LONG - 4))
 #define GET_PFN(pfn)                    (pfn & 0x0fffffffffffffffUL)
 
+extern int remap_pfn_range(struct vm_area_struct *, unsigned long, unsigned long,
+                           unsigned long, pgprot_t);
+
+static inline int io_remap_pfn_range(struct vm_area_struct *vma,
+                                     unsigned long from, unsigned long pfn,
+                                     unsigned long size, pgprot_t prot)
+{
+        unsigned long offset = GET_PFN(pfn) << PAGE_SHIFT;
+        int space = GET_IOSPACE(pfn);
+        unsigned long phys_base;
+
+        phys_base = offset | (((unsigned long) space) << 32UL);
+
+        return remap_pfn_range(vma, from, phys_base >> PAGE_SHIFT, size, prot);
+}
+
 #include <asm-generic/pgtable.h>
 
 /* We provide our own get_unmapped_area to cope with VA holes and
diff --git a/arch/sparc/include/asm/pgtsrmmu.h b/arch/sparc/include/asm/pgtsrmmu.h
index 1407c07bdade..f6ae2b2b6870 100644
--- a/arch/sparc/include/asm/pgtsrmmu.h
+++ b/arch/sparc/include/asm/pgtsrmmu.h
@@ -280,7 +280,7 @@ static inline unsigned long srmmu_hwprobe(unsigned long vaddr)
         return retval;
 }
 #else
-#define srmmu_hwprobe(addr) (srmmu_swprobe(addr, 0) & SRMMU_PTE_PMASK)
+#define srmmu_hwprobe(addr) srmmu_swprobe(addr, 0)
 #endif
 
 static inline int
diff --git a/arch/sparc/include/asm/spitfire.h b/arch/sparc/include/asm/spitfire.h
index 55a17c6efeb8..d06a26601753 100644
--- a/arch/sparc/include/asm/spitfire.h
+++ b/arch/sparc/include/asm/spitfire.h
@@ -43,6 +43,8 @@
 #define SUN4V_CHIP_NIAGARA1     0x01
 #define SUN4V_CHIP_NIAGARA2     0x02
 #define SUN4V_CHIP_NIAGARA3     0x03
+#define SUN4V_CHIP_NIAGARA4     0x04
+#define SUN4V_CHIP_NIAGARA5     0x05
 #define SUN4V_CHIP_UNKNOWN      0xff
 
 #ifndef __ASSEMBLY__
diff --git a/arch/sparc/include/asm/termios.h b/arch/sparc/include/asm/termios.h
index e8ba95399643..e2f46705a210 100644
--- a/arch/sparc/include/asm/termios.h
+++ b/arch/sparc/include/asm/termios.h
@@ -40,7 +40,6 @@ struct winsize {
 };
 
 #ifdef __KERNEL__
-#include <linux/module.h>
 
 /*
  * c_cc characters in the termio structure.  Oh, how I love being
diff --git a/arch/sparc/include/asm/unistd.h b/arch/sparc/include/asm/unistd.h
index 6260d5deeabc..c7cb0af0eb59 100644
--- a/arch/sparc/include/asm/unistd.h
+++ b/arch/sparc/include/asm/unistd.h
@@ -406,8 +406,10 @@
 #define __NR_syncfs             335
 #define __NR_sendmmsg           336
 #define __NR_setns              337
+#define __NR_process_vm_readv   338
+#define __NR_process_vm_writev  339
 
-#define NR_syscalls             338
+#define NR_syscalls             340
 
 #ifdef __32bit_syscall_numbers__
 /* Sparc 32-bit only has the "setresuid32", "getresuid32" variants,
diff --git a/arch/sparc/include/asm/xor_64.h b/arch/sparc/include/asm/xor_64.h
index 9ed6ff679ab7..ee8edc68423e 100644
--- a/arch/sparc/include/asm/xor_64.h
+++ b/arch/sparc/include/asm/xor_64.h
@@ -66,6 +66,8 @@ static struct xor_block_template xor_block_niagara = {
         ((tlb_type == hypervisor && \
           (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 || \
            sun4v_chip_type == SUN4V_CHIP_NIAGARA2 || \
-           sun4v_chip_type == SUN4V_CHIP_NIAGARA3)) ? \
+           sun4v_chip_type == SUN4V_CHIP_NIAGARA3 || \
+           sun4v_chip_type == SUN4V_CHIP_NIAGARA4 || \
+           sun4v_chip_type == SUN4V_CHIP_NIAGARA5)) ? \
          &xor_block_niagara : \
          &xor_block_VIS)
diff --git a/arch/sparc/kernel/apc.c b/arch/sparc/kernel/apc.c
index caef9deb5866..812e10bbb0b3 100644
--- a/arch/sparc/kernel/apc.c
+++ b/arch/sparc/kernel/apc.c
@@ -13,6 +13,7 @@
 #include <linux/pm.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/module.h>
 
 #include <asm/io.h>
 #include <asm/oplib.h>
diff --git a/arch/sparc/kernel/auxio_32.c b/arch/sparc/kernel/auxio_32.c
index acf5151f3c1d..f7ea8f032719 100644
--- a/arch/sparc/kernel/auxio_32.c
+++ b/arch/sparc/kernel/auxio_32.c
@@ -8,6 +8,7 @@
 #include <linux/spinlock.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/export.h>
 #include <asm/oplib.h>
 #include <asm/io.h>
 #include <asm/auxio.h>
diff --git a/arch/sparc/kernel/btext.c b/arch/sparc/kernel/btext.c
index 89aa4eb20cf5..57073e56ba9e 100644
--- a/arch/sparc/kernel/btext.c
+++ b/arch/sparc/kernel/btext.c
@@ -6,7 +6,6 @@
 #include <linux/kernel.h>
 #include <linux/string.h>
 #include <linux/init.h>
-#include <linux/module.h>
 #include <linux/console.h>
 
 #include <asm/btext.h>
diff --git a/arch/sparc/kernel/central.c b/arch/sparc/kernel/central.c
index 7eef3f741963..38d48a59879c 100644
--- a/arch/sparc/kernel/central.c
+++ b/arch/sparc/kernel/central.c
@@ -6,6 +6,7 @@
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/slab.h>
+#include <linux/export.h>
 #include <linux/string.h>
 #include <linux/init.h>
 #include <linux/of_device.h>
diff --git a/arch/sparc/kernel/cpu.c b/arch/sparc/kernel/cpu.c
index 9810fd881058..2d1819641769 100644
--- a/arch/sparc/kernel/cpu.c
+++ b/arch/sparc/kernel/cpu.c
@@ -6,7 +6,7 @@
 
 #include <linux/seq_file.h>
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/init.h>
 #include <linux/smp.h>
 #include <linux/threads.h>
@@ -481,6 +481,18 @@ static void __init sun4v_cpu_probe(void)
                 sparc_pmu_type = "niagara3";
                 break;
 
+        case SUN4V_CHIP_NIAGARA4:
+                sparc_cpu_type = "UltraSparc T4 (Niagara4)";
+                sparc_fpu_type = "UltraSparc T4 integrated FPU";
+                sparc_pmu_type = "niagara4";
+                break;
+
+        case SUN4V_CHIP_NIAGARA5:
+                sparc_cpu_type = "UltraSparc T5 (Niagara5)";
+                sparc_fpu_type = "UltraSparc T5 integrated FPU";
+                sparc_pmu_type = "niagara5";
+                break;
+
         default:
                 printk(KERN_WARNING "CPU: Unknown sun4v cpu type [%s]\n",
                        prom_cpu_compatible);
diff --git a/arch/sparc/kernel/cpumap.c b/arch/sparc/kernel/cpumap.c
index 4197e8d62d4c..e4de74c2c9b0 100644
--- a/arch/sparc/kernel/cpumap.c
+++ b/arch/sparc/kernel/cpumap.c
@@ -3,7 +3,7 @@
  * Copyright (C) 2009 Hong H. Pham <hong.pham@windriver.com>
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/kernel.h>
 #include <linux/init.h>
@@ -325,6 +325,8 @@ static int iterate_cpu(struct cpuinfo_tree *t, unsigned int root_index)
         case SUN4V_CHIP_NIAGARA1:
         case SUN4V_CHIP_NIAGARA2:
         case SUN4V_CHIP_NIAGARA3:
+        case SUN4V_CHIP_NIAGARA4:
+        case SUN4V_CHIP_NIAGARA5:
                 rover_inc_table = niagara_iterate_method;
                 break;
         default:
diff --git a/arch/sparc/kernel/dma.c b/arch/sparc/kernel/dma.c
index e1ba8ee21b9a..b667aa6f28f6 100644
--- a/arch/sparc/kernel/dma.c
+++ b/arch/sparc/kernel/dma.c
@@ -1,5 +1,4 @@
 #include <linux/kernel.h>
-#include <linux/module.h>
 #include <linux/dma-mapping.h>
 #include <linux/dma-debug.h>
 
diff --git a/arch/sparc/kernel/ebus.c b/arch/sparc/kernel/ebus.c
index 77dbf6d45faf..e306fb08ee5e 100644
--- a/arch/sparc/kernel/ebus.c
+++ b/arch/sparc/kernel/ebus.c
@@ -4,7 +4,7 @@
  * Copyright (C) 1999  David S. Miller (davem@redhat.com)
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/init.h>
diff --git a/arch/sparc/kernel/entry.h b/arch/sparc/kernel/entry.h
index e27f8ea8656e..0c218e4c0881 100644
--- a/arch/sparc/kernel/entry.h
+++ b/arch/sparc/kernel/entry.h
@@ -42,6 +42,9 @@ extern void fpsave(unsigned long *fpregs, unsigned long *fsr,
 extern void fpload(unsigned long *fpregs, unsigned long *fsr);
 
 #else /* CONFIG_SPARC32 */
+
+#include <asm/trap_block.h>
+
 struct popc_3insn_patch_entry {
         unsigned int    addr;
         unsigned int    insns[3];
@@ -57,6 +60,10 @@ extern struct popc_6insn_patch_entry __popc_6insn_patch,
         __popc_6insn_patch_end;
 
 extern void __init per_cpu_patch(void);
+extern void sun4v_patch_1insn_range(struct sun4v_1insn_patch_entry *,
+                                    struct sun4v_1insn_patch_entry *);
+extern void sun4v_patch_2insn_range(struct sun4v_2insn_patch_entry *,
+                                    struct sun4v_2insn_patch_entry *);
 extern void __init sun4v_patch(void);
 extern void __init boot_cpu_id_too_large(int cpu);
 extern unsigned int dcache_parity_tl1_occurred;
diff --git a/arch/sparc/kernel/head_64.S b/arch/sparc/kernel/head_64.S
index 0eac1b2fc53d..0d810c2f1d00 100644
--- a/arch/sparc/kernel/head_64.S
+++ b/arch/sparc/kernel/head_64.S
@@ -133,7 +133,7 @@ prom_sun4v_name:
 prom_niagara_prefix:
         .asciz  "SUNW,UltraSPARC-T"
 prom_sparc_prefix:
-        .asciz  "SPARC-T"
+        .asciz  "SPARC-"
         .align  4
 prom_root_compatible:
         .skip   64
@@ -396,7 +396,7 @@ sun4v_chip_type:
         or      %g1, %lo(prom_cpu_compatible), %g1
         sethi   %hi(prom_sparc_prefix), %g7
         or      %g7, %lo(prom_sparc_prefix), %g7
-        mov     7, %g3
+        mov     6, %g3
 90:     ldub    [%g7], %g2
         ldub    [%g1], %g4
         cmp     %g2, %g4
@@ -408,10 +408,23 @@ sun4v_chip_type:
 
         sethi   %hi(prom_cpu_compatible), %g1
         or      %g1, %lo(prom_cpu_compatible), %g1
-        ldub    [%g1 + 7], %g2
+        ldub    [%g1 + 6], %g2
+        cmp     %g2, 'T'
+        be,pt   %xcc, 70f
+         cmp    %g2, 'M'
+        bne,pn  %xcc, 4f
+         nop
+
+70:     ldub    [%g1 + 7], %g2
         cmp     %g2, '3'
         be,pt   %xcc, 5f
          mov    SUN4V_CHIP_NIAGARA3, %g4
+        cmp     %g2, '4'
+        be,pt   %xcc, 5f
+         mov    SUN4V_CHIP_NIAGARA4, %g4
+        cmp     %g2, '5'
+        be,pt   %xcc, 5f
+         mov    SUN4V_CHIP_NIAGARA5, %g4
         ba,pt   %xcc, 4f
          nop
 
@@ -545,6 +558,12 @@ niagara_tlb_fixup:
         cmp     %g1, SUN4V_CHIP_NIAGARA3
         be,pt   %xcc, niagara2_patch
          nop
+        cmp     %g1, SUN4V_CHIP_NIAGARA4
+        be,pt   %xcc, niagara2_patch
+         nop
+        cmp     %g1, SUN4V_CHIP_NIAGARA5
+        be,pt   %xcc, niagara2_patch
+         nop
 
         call    generic_patch_copyops
          nop
diff --git a/arch/sparc/kernel/hvapi.c b/arch/sparc/kernel/hvapi.c
index c2d055d8ba9e..8593672838fd 100644
--- a/arch/sparc/kernel/hvapi.c
+++ b/arch/sparc/kernel/hvapi.c
@@ -3,7 +3,7 @@
  * Copyright (C) 2007 David S. Miller <davem@davemloft.net>
  */
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/init.h>
 
 #include <asm/hypervisor.h>
diff --git a/arch/sparc/kernel/idprom.c b/arch/sparc/kernel/idprom.c
index 52a15fe2db19..9167db40720e 100644
--- a/arch/sparc/kernel/idprom.c
+++ b/arch/sparc/kernel/idprom.c
@@ -8,7 +8,7 @@
 #include <linux/kernel.h>
 #include <linux/types.h>
 #include <linux/init.h>
-#include <linux/module.h>
+#include <linux/export.h>
 
 #include <asm/oplib.h>
 #include <asm/idprom.h>
diff --git a/arch/sparc/kernel/iommu.c b/arch/sparc/kernel/iommu.c
index 6f01e8c83197..4643d68713fa 100644
--- a/arch/sparc/kernel/iommu.c
+++ b/arch/sparc/kernel/iommu.c
@@ -5,7 +5,7 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/delay.h>
 #include <linux/device.h>
diff --git a/arch/sparc/kernel/irq.h b/arch/sparc/kernel/irq.h
index 100b9c204e78..42851122bbd9 100644
--- a/arch/sparc/kernel/irq.h
+++ b/arch/sparc/kernel/irq.h
@@ -88,7 +88,7 @@ BTFIXUPDEF_CALL(void, set_irq_udt, int)
 #define set_irq_udt(cpu) BTFIXUP_CALL(set_irq_udt)(cpu)
 
 /* All SUN4D IPIs are sent on this IRQ, may be shared with hard IRQs */
-#define SUN4D_IPI_IRQ 14
+#define SUN4D_IPI_IRQ 13
 
 extern void sun4d_ipi_interrupt(void);
 
diff --git a/arch/sparc/kernel/irq_32.c b/arch/sparc/kernel/irq_32.c
index 9b89d842913c..b2668afd1c34 100644
--- a/arch/sparc/kernel/irq_32.c
+++ b/arch/sparc/kernel/irq_32.c
@@ -13,6 +13,7 @@
 
 #include <linux/kernel_stat.h>
 #include <linux/seq_file.h>
+#include <linux/export.h>
 
 #include <asm/cacheflush.h>
 #include <asm/cpudata.h>
diff --git a/arch/sparc/kernel/irq_64.c b/arch/sparc/kernel/irq_64.c
index 0dd8422a469c..d45b710ea7e4 100644
--- a/arch/sparc/kernel/irq_64.c
+++ b/arch/sparc/kernel/irq_64.c
@@ -5,7 +5,6 @@
  * Copyright (C) 1998  Jakub Jelinek    (jj@ultra.linux.cz)
  */
 
-#include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/linkage.h>
 #include <linux/ptrace.h>
diff --git a/arch/sparc/kernel/jump_label.c b/arch/sparc/kernel/jump_label.c
index ea2dafc93d78..971fd435a281 100644
--- a/arch/sparc/kernel/jump_label.c
+++ b/arch/sparc/kernel/jump_label.c
@@ -36,12 +36,4 @@ void arch_jump_label_transform(struct jump_entry *entry,
         put_online_cpus();
 }
 
-void arch_jump_label_text_poke_early(jump_label_t addr)
-{
-        u32 *insn_p = (u32 *) (unsigned long) addr;
-
-        *insn_p = 0x01000000;
-        flushi(insn_p);
-}
-
 #endif
diff --git a/arch/sparc/kernel/ldc.c b/arch/sparc/kernel/ldc.c
index 732b0bce6001..435e406fdec3 100644
--- a/arch/sparc/kernel/ldc.c
+++ b/arch/sparc/kernel/ldc.c
@@ -4,7 +4,7 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/spinlock.h>
 #include <linux/delay.h>
diff --git a/arch/sparc/kernel/leon_kernel.c b/arch/sparc/kernel/leon_kernel.c
index d17255a2bbac..a19c8a063683 100644
--- a/arch/sparc/kernel/leon_kernel.c
+++ b/arch/sparc/kernel/leon_kernel.c
@@ -4,7 +4,6 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
 #include <linux/errno.h>
 #include <linux/mutex.h>
 #include <linux/of.h>
diff --git a/arch/sparc/kernel/leon_pci.c b/arch/sparc/kernel/leon_pci.c
index a8a9a275037d..f1cf6ef011a7 100644
--- a/arch/sparc/kernel/leon_pci.c
+++ b/arch/sparc/kernel/leon_pci.c
@@ -9,6 +9,7 @@
 #include <linux/of_device.h>
 #include <linux/kernel.h>
 #include <linux/pci.h>
+#include <linux/export.h>
 #include <asm/leon.h>
 #include <asm/leon_pci.h>
 
diff --git a/arch/sparc/kernel/leon_pci_grpci2.c b/arch/sparc/kernel/leon_pci_grpci2.c
index fad1bd07cb56..b1bc38853a3d 100644
--- a/arch/sparc/kernel/leon_pci_grpci2.c
+++ b/arch/sparc/kernel/leon_pci_grpci2.c
@@ -9,7 +9,7 @@
 #include <linux/kernel.h>
 #include <linux/pci.h>
 #include <linux/delay.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <asm/io.h>
 #include <asm/leon.h>
 #include <asm/vaddrs.h>
diff --git a/arch/sparc/kernel/mdesc.c b/arch/sparc/kernel/mdesc.c
index acaebb63c4fd..6dc796280589 100644
--- a/arch/sparc/kernel/mdesc.c
+++ b/arch/sparc/kernel/mdesc.c
@@ -11,11 +11,13 @@
 #include <linux/mm.h>
 #include <linux/miscdevice.h>
 #include <linux/bootmem.h>
+#include <linux/export.h>
 
 #include <asm/cpudata.h>
 #include <asm/hypervisor.h>
 #include <asm/mdesc.h>
 #include <asm/prom.h>
+#include <asm/uaccess.h>
 #include <asm/oplib.h>
 #include <asm/smp.h>
 
diff --git a/arch/sparc/kernel/module.c b/arch/sparc/kernel/module.c
index da0c6c70ccb2..e5519870c3d9 100644
--- a/arch/sparc/kernel/module.c
+++ b/arch/sparc/kernel/module.c
@@ -17,6 +17,8 @@
 #include <asm/processor.h>
 #include <asm/spitfire.h>
 
+#include "entry.h"
+
 #ifdef CONFIG_SPARC64
 
 #include <linux/jump_label.h>
@@ -203,6 +205,29 @@ int apply_relocate_add(Elf_Shdr *sechdrs,
 }
 
 #ifdef CONFIG_SPARC64
+static void do_patch_sections(const Elf_Ehdr *hdr,
+                              const Elf_Shdr *sechdrs)
+{
+        const Elf_Shdr *s, *sun4v_1insn = NULL, *sun4v_2insn = NULL;
+        char *secstrings = (void *)hdr + sechdrs[hdr->e_shstrndx].sh_offset;
+
+        for (s = sechdrs; s < sechdrs + hdr->e_shnum; s++) {
+                if (!strcmp(".sun4v_1insn_patch", secstrings + s->sh_name))
+                        sun4v_1insn = s;
+                if (!strcmp(".sun4v_2insn_patch", secstrings + s->sh_name))
+                        sun4v_2insn = s;
+        }
+
+        if (sun4v_1insn && tlb_type == hypervisor) {
+                void *p = (void *) sun4v_1insn->sh_addr;
+                sun4v_patch_1insn_range(p, p + sun4v_1insn->sh_size);
+        }
+        if (sun4v_2insn && tlb_type == hypervisor) {
+                void *p = (void *) sun4v_2insn->sh_addr;
+                sun4v_patch_2insn_range(p, p + sun4v_2insn->sh_size);
+        }
+}
+
 int module_finalize(const Elf_Ehdr *hdr,
                     const Elf_Shdr *sechdrs,
                     struct module *me)
@@ -210,6 +235,8 @@ int module_finalize(const Elf_Ehdr *hdr,
         /* make jump label nops */
         jump_label_apply_nops(me);
 
+        do_patch_sections(hdr, sechdrs);
+
         /* Cheetah's I-cache is fully coherent.  */
         if (tlb_type == spitfire) {
                 unsigned long va;
diff --git a/arch/sparc/kernel/nmi.c b/arch/sparc/kernel/nmi.c
index 300f810142f5..c76fe0b5bd94 100644
--- a/arch/sparc/kernel/nmi.c
+++ b/arch/sparc/kernel/nmi.c
@@ -10,7 +10,7 @@
 #include <linux/init.h>
 #include <linux/percpu.h>
 #include <linux/nmi.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/kprobes.h>
 #include <linux/kernel_stat.h>
 #include <linux/reboot.h>
diff --git a/arch/sparc/kernel/of_device_32.c b/arch/sparc/kernel/of_device_32.c
index a312af40ea84..4ee8ce0d5d8d 100644
--- a/arch/sparc/kernel/of_device_32.c
+++ b/arch/sparc/kernel/of_device_32.c
@@ -2,7 +2,6 @@
 #include <linux/kernel.h>
 #include <linux/of.h>
 #include <linux/init.h>
-#include <linux/module.h>
 #include <linux/mod_devicetable.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
diff --git a/arch/sparc/kernel/of_device_64.c b/arch/sparc/kernel/of_device_64.c
index 3bb2eace58cf..7a3be6f6737a 100644
--- a/arch/sparc/kernel/of_device_64.c
+++ b/arch/sparc/kernel/of_device_64.c
@@ -2,13 +2,14 @@
 #include <linux/kernel.h>
 #include <linux/of.h>
 #include <linux/init.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/mod_devicetable.h>
 #include <linux/slab.h>
 #include <linux/errno.h>
 #include <linux/irq.h>
 #include <linux/of_device.h>
 #include <linux/of_platform.h>
+#include <asm/spitfire.h>
 
 #include "of_device_common.h"
 
diff --git a/arch/sparc/kernel/of_device_common.c b/arch/sparc/kernel/of_device_common.c
index cb15bbf8a201..de199bf0cb05 100644
--- a/arch/sparc/kernel/of_device_common.c
+++ b/arch/sparc/kernel/of_device_common.c
@@ -2,7 +2,7 @@
 #include <linux/kernel.h>
 #include <linux/of.h>
 #include <linux/init.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/mod_devicetable.h>
 #include <linux/errno.h>
 #include <linux/irq.h>
diff --git a/arch/sparc/kernel/pci.c b/arch/sparc/kernel/pci.c
index 1e94f946570e..31111e35281e 100644
--- a/arch/sparc/kernel/pci.c
+++ b/arch/sparc/kernel/pci.c
@@ -8,7 +8,7 @@
  * with minor modifications, see there for credits.
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/kernel.h>
 #include <linux/string.h>
 #include <linux/sched.h>
@@ -230,7 +230,8 @@ static void pci_parse_of_addrs(struct platform_device *op,
                         res = &dev->resource[(i - PCI_BASE_ADDRESS_0) >> 2];
                 } else if (i == dev->rom_base_reg) {
                         res = &dev->resource[PCI_ROM_RESOURCE];
-                        flags |= IORESOURCE_READONLY | IORESOURCE_CACHEABLE;
+                        flags |= IORESOURCE_READONLY | IORESOURCE_CACHEABLE
+                              | IORESOURCE_SIZEALIGN;
                 } else {
                         printk(KERN_ERR "PCI: bad cfg reg num 0x%x\n", i);
                         continue;
diff --git a/arch/sparc/kernel/pci_fire.c b/arch/sparc/kernel/pci_fire.c
index d29a32fcc5e4..188f935276fd 100644
--- a/arch/sparc/kernel/pci_fire.c
+++ b/arch/sparc/kernel/pci_fire.c
@@ -7,6 +7,7 @@
 #include <linux/slab.h>
 #include <linux/init.h>
 #include <linux/msi.h>
+#include <linux/export.h>
 #include <linux/irq.h>
 #include <linux/of_device.h>
 
diff --git a/arch/sparc/kernel/pci_psycho.c b/arch/sparc/kernel/pci_psycho.c
index 86ae08d9b6ee..f4d29e15ce71 100644
--- a/arch/sparc/kernel/pci_psycho.c
+++ b/arch/sparc/kernel/pci_psycho.c
@@ -9,6 +9,7 @@
 #include <linux/types.h>
 #include <linux/pci.h>
 #include <linux/init.h>
+#include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/interrupt.h>
 #include <linux/of_device.h>
diff --git a/arch/sparc/kernel/pci_sabre.c b/arch/sparc/kernel/pci_sabre.c
index d1840dbdaa2f..3efaa4644d60 100644
--- a/arch/sparc/kernel/pci_sabre.c
+++ b/arch/sparc/kernel/pci_sabre.c
@@ -9,6 +9,7 @@
 #include <linux/types.h>
 #include <linux/pci.h>
 #include <linux/init.h>
+#include <linux/export.h>
 #include <linux/slab.h>
 #include <linux/interrupt.h>
 #include <linux/of_device.h>
diff --git a/arch/sparc/kernel/pci_schizo.c b/arch/sparc/kernel/pci_schizo.c
index f030b02edddd..13d4aa20b5a5 100644
--- a/arch/sparc/kernel/pci_schizo.c
+++ b/arch/sparc/kernel/pci_schizo.c
@@ -8,6 +8,7 @@
 #include <linux/pci.h>
 #include <linux/init.h>
 #include <linux/slab.h>
+#include <linux/export.h>
 #include <linux/interrupt.h>
 #include <linux/of_device.h>
 
diff --git a/arch/sparc/kernel/pci_sun4v.c b/arch/sparc/kernel/pci_sun4v.c
index b01a06e9ae4e..b272cda35a01 100644
--- a/arch/sparc/kernel/pci_sun4v.c
+++ b/arch/sparc/kernel/pci_sun4v.c
@@ -12,6 +12,7 @@
 #include <linux/percpu.h>
 #include <linux/irq.h>
 #include <linux/msi.h>
+#include <linux/export.h>
 #include <linux/log2.h>
 #include <linux/of_device.h>
 
diff --git a/arch/sparc/kernel/pcic.c b/arch/sparc/kernel/pcic.c
index 1aaf8c180be5..fcc148effaac 100644
--- a/arch/sparc/kernel/pcic.c
+++ b/arch/sparc/kernel/pcic.c
@@ -25,6 +25,7 @@
 #include <linux/time.h>
 #include <linux/timex.h>
 #include <linux/interrupt.h>
+#include <linux/export.h>
 
 #include <asm/irq.h>
 #include <asm/oplib.h>
diff --git a/arch/sparc/kernel/pcr.c b/arch/sparc/kernel/pcr.c
index 343b0f9e2e7b..a24072a49270 100644
--- a/arch/sparc/kernel/pcr.c
+++ b/arch/sparc/kernel/pcr.c
@@ -3,7 +3,7 @@
  * Copyright (C) 2009 David S. Miller (davem@davemloft.net)
  */
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/init.h>
 #include <linux/irq.h>
 
@@ -13,6 +13,7 @@
 #include <asm/pil.h>
 #include <asm/pcr.h>
 #include <asm/nmi.h>
+#include <asm/spitfire.h>
 
 /* This code is shared between various users of the performance
  * counters.  Users will be oprofile, pseudo-NMI watchdog, and the
diff --git a/arch/sparc/kernel/pmc.c b/arch/sparc/kernel/pmc.c
index 6a585d393580..0e3202239ff5 100644
--- a/arch/sparc/kernel/pmc.c
+++ b/arch/sparc/kernel/pmc.c
@@ -11,6 +11,7 @@
 #include <linux/pm.h>
 #include <linux/of.h>
 #include <linux/of_device.h>
+#include <linux/module.h>
 
 #include <asm/io.h>
 #include <asm/oplib.h>
diff --git a/arch/sparc/kernel/power.c b/arch/sparc/kernel/power.c
index cb4c0f57c024..0d39075063b2 100644
--- a/arch/sparc/kernel/power.c
+++ b/arch/sparc/kernel/power.c
@@ -4,7 +4,7 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/reboot.h>
diff --git a/arch/sparc/kernel/process_32.c b/arch/sparc/kernel/process_32.c
index c8cc461ff75f..f793742eec2b 100644
--- a/arch/sparc/kernel/process_32.c
+++ b/arch/sparc/kernel/process_32.c
@@ -380,8 +380,7 @@ void flush_thread(void)
 #endif
         }
 
-        /* Now, this task is no longer a kernel thread. */
-        current->thread.current_ds = USER_DS;
+        /* This task is no longer a kernel thread. */
         if (current->thread.flags & SPARC_FLAG_KTHREAD) {
                 current->thread.flags &= ~SPARC_FLAG_KTHREAD;
 
diff --git a/arch/sparc/kernel/process_64.c b/arch/sparc/kernel/process_64.c
index c158a95ec664..3739a06a76cb 100644
--- a/arch/sparc/kernel/process_64.c
+++ b/arch/sparc/kernel/process_64.c
@@ -12,7 +12,7 @@
 #include <stdarg.h>
 
 #include <linux/errno.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
 #include <linux/mm.h>
@@ -368,9 +368,6 @@ void flush_thread(void)
 
         /* Clear FPU register state. */
         t->fpsaved[0] = 0;
-        
-        if (get_thread_current_ds() != ASI_AIUS)
-                set_fs(USER_DS);
 }
 
 /* It's a bit more tricky when 64-bit tasks are involved... */
diff --git a/arch/sparc/kernel/prom_32.c b/arch/sparc/kernel/prom_32.c
index 5ce3d15a99b0..b51cbb9e87dc 100644
--- a/arch/sparc/kernel/prom_32.c
+++ b/arch/sparc/kernel/prom_32.c
@@ -20,7 +20,6 @@
 #include <linux/string.h>
 #include <linux/mm.h>
 #include <linux/bootmem.h>
-#include <linux/module.h>
 
 #include <asm/prom.h>
 #include <asm/oplib.h>
diff --git a/arch/sparc/kernel/prom_64.c b/arch/sparc/kernel/prom_64.c
index 86597d9867fd..340c5b976d28 100644
--- a/arch/sparc/kernel/prom_64.c
+++ b/arch/sparc/kernel/prom_64.c
@@ -19,7 +19,6 @@
 #include <linux/types.h>
 #include <linux/string.h>
 #include <linux/mm.h>
-#include <linux/module.h>
 #include <linux/memblock.h>
 #include <linux/of.h>
 
diff --git a/arch/sparc/kernel/prom_common.c b/arch/sparc/kernel/prom_common.c
index ed25834328f4..46614807a57f 100644
--- a/arch/sparc/kernel/prom_common.c
+++ b/arch/sparc/kernel/prom_common.c
@@ -15,7 +15,7 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/errno.h>
 #include <linux/mutex.h>
 #include <linux/slab.h>
diff --git a/arch/sparc/kernel/reboot.c b/arch/sparc/kernel/reboot.c
index ef89d3d69748..006a42dd2007 100644
--- a/arch/sparc/kernel/reboot.c
+++ b/arch/sparc/kernel/reboot.c
@@ -4,7 +4,7 @@
  */
 #include <linux/kernel.h>
 #include <linux/reboot.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/pm.h>
 
 #include <asm/system.h>
diff --git a/arch/sparc/kernel/sbus.c b/arch/sparc/kernel/sbus.c
index a161b9c77f05..1271b3a27d4e 100644
--- a/arch/sparc/kernel/sbus.c
+++ b/arch/sparc/kernel/sbus.c
@@ -9,6 +9,7 @@
 #include <linux/mm.h>
 #include <linux/spinlock.h>
 #include <linux/slab.h>
+#include <linux/export.h>
 #include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/of.h>
diff --git a/arch/sparc/kernel/setup_32.c b/arch/sparc/kernel/setup_32.c
index d26e1f6c717a..fe1e3fc31bc5 100644
--- a/arch/sparc/kernel/setup_32.c
+++ b/arch/sparc/kernel/setup_32.c
@@ -31,6 +31,7 @@
 #include <linux/root_dev.h>
 #include <linux/cpu.h>
 #include <linux/kdebug.h>
+#include <linux/export.h>
 
 #include <asm/system.h>
 #include <asm/io.h>
@@ -137,7 +138,7 @@ static void __init process_switch(char c)
                 prom_halt();
                 break;
         case 'p':
-                /* Just ignore, this behavior is now the default.  */
+                prom_early_console.flags &= ~CON_BOOT;
                 break;
         default:
                 printk("Unknown boot switch (-%c)\n", c);
diff --git a/arch/sparc/kernel/setup_64.c b/arch/sparc/kernel/setup_64.c
index 3e9daea1653d..a854a1c240ff 100644
--- a/arch/sparc/kernel/setup_64.c
+++ b/arch/sparc/kernel/setup_64.c
@@ -106,7 +106,7 @@ static void __init process_switch(char c)
                 prom_halt();
                 break;
         case 'p':
-                /* Just ignore, this behavior is now the default.  */
+                prom_early_console.flags &= ~CON_BOOT;
                 break;
         case 'P':
                 /* Force UltraSPARC-III P-Cache on. */
@@ -234,40 +234,50 @@ void __init per_cpu_patch(void)
         }
 }
 
-void __init sun4v_patch(void)
+void sun4v_patch_1insn_range(struct sun4v_1insn_patch_entry *start,
+                             struct sun4v_1insn_patch_entry *end)
 {
-        extern void sun4v_hvapi_init(void);
-        struct sun4v_1insn_patch_entry *p1;
-        struct sun4v_2insn_patch_entry *p2;
-
-        if (tlb_type != hypervisor)
-                return;
+        while (start < end) {
+                unsigned long addr = start->addr;
 
-        p1 = &__sun4v_1insn_patch;
-        while (p1 < &__sun4v_1insn_patch_end) {
-                unsigned long addr = p1->addr;
-
-                *(unsigned int *) (addr +  0) = p1->insn;
+                *(unsigned int *) (addr +  0) = start->insn;
                 wmb();
                 __asm__ __volatile__("flush     %0" : : "r" (addr +  0));
 
-                p1++;
+                start++;
         }
+}
 
-        p2 = &__sun4v_2insn_patch;
-        while (p2 < &__sun4v_2insn_patch_end) {
-                unsigned long addr = p2->addr;
+void sun4v_patch_2insn_range(struct sun4v_2insn_patch_entry *start,
+                             struct sun4v_2insn_patch_entry *end)
+{
+        while (start < end) {
+                unsigned long addr = start->addr;
 
-                *(unsigned int *) (addr +  0) = p2->insns[0];
+                *(unsigned int *) (addr +  0) = start->insns[0];
                 wmb();
                 __asm__ __volatile__("flush     %0" : : "r" (addr +  0));
 
-                *(unsigned int *) (addr +  4) = p2->insns[1];
+                *(unsigned int *) (addr +  4) = start->insns[1];
                 wmb();
                 __asm__ __volatile__("flush     %0" : : "r" (addr +  4));
 
-                p2++;
+                start++;
         }
+}
+
+void __init sun4v_patch(void)
+{
+        extern void sun4v_hvapi_init(void);
+
+        if (tlb_type != hypervisor)
+                return;
+
+        sun4v_patch_1insn_range(&__sun4v_1insn_patch,
+                                &__sun4v_1insn_patch_end);
+
+        sun4v_patch_2insn_range(&__sun4v_2insn_patch,
+                                &__sun4v_2insn_patch_end);
 
         sun4v_hvapi_init();
 }
@@ -425,10 +435,14 @@ static void __init init_sparc64_elf_hwcap(void)
         else if (tlb_type == hypervisor) {
                 if (sun4v_chip_type == SUN4V_CHIP_NIAGARA1 ||
                     sun4v_chip_type == SUN4V_CHIP_NIAGARA2 ||
-                    sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+                    sun4v_chip_type == SUN4V_CHIP_NIAGARA3 ||
+                    sun4v_chip_type == SUN4V_CHIP_NIAGARA4 ||
+                    sun4v_chip_type == SUN4V_CHIP_NIAGARA5)
                         cap |= HWCAP_SPARC_BLKINIT;
                 if (sun4v_chip_type == SUN4V_CHIP_NIAGARA2 ||
-                    sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+                    sun4v_chip_type == SUN4V_CHIP_NIAGARA3 ||
+                    sun4v_chip_type == SUN4V_CHIP_NIAGARA4 ||
+                    sun4v_chip_type == SUN4V_CHIP_NIAGARA5)
                         cap |= HWCAP_SPARC_N2;
         }
 
@@ -440,17 +454,27 @@ static void __init init_sparc64_elf_hwcap(void)
                         cap |= AV_SPARC_VIS;
                 if (tlb_type == cheetah || tlb_type == cheetah_plus)
                         cap |= AV_SPARC_VIS | AV_SPARC_VIS2;
-                if (tlb_type == cheetah_plus)
-                        cap |= AV_SPARC_POPC;
+                if (tlb_type == cheetah_plus) {
+                        unsigned long impl, ver;
+
+                        __asm__ __volatile__("rdpr %%ver, %0" : "=r" (ver));
+                        impl = ((ver >> 32) & 0xffff);
+                        if (impl == PANTHER_IMPL)
+                                cap |= AV_SPARC_POPC;
+                }
                 if (tlb_type == hypervisor) {
                         if (sun4v_chip_type == SUN4V_CHIP_NIAGARA1)
                                 cap |= AV_SPARC_ASI_BLK_INIT;
                         if (sun4v_chip_type == SUN4V_CHIP_NIAGARA2 ||
-                            sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+                            sun4v_chip_type == SUN4V_CHIP_NIAGARA3 ||
+                            sun4v_chip_type == SUN4V_CHIP_NIAGARA4 ||
+                            sun4v_chip_type == SUN4V_CHIP_NIAGARA5)
                                 cap |= (AV_SPARC_VIS | AV_SPARC_VIS2 |
                                         AV_SPARC_ASI_BLK_INIT |
                                         AV_SPARC_POPC);
-                        if (sun4v_chip_type == SUN4V_CHIP_NIAGARA3)
+                        if (sun4v_chip_type == SUN4V_CHIP_NIAGARA3 ||
+                            sun4v_chip_type == SUN4V_CHIP_NIAGARA4 ||
+                            sun4v_chip_type == SUN4V_CHIP_NIAGARA5)
                                 cap |= (AV_SPARC_VIS3 | AV_SPARC_HPC |
                                         AV_SPARC_FMAF);
                 }
diff --git a/arch/sparc/kernel/signal32.c b/arch/sparc/kernel/signal32.c
index 1ba95aff5d59..023b8860dc97 100644
--- a/arch/sparc/kernel/signal32.c
+++ b/arch/sparc/kernel/signal32.c
@@ -273,10 +273,7 @@ void do_sigreturn32(struct pt_regs *regs)
                 case 1: set.sig[0] = seta[0] + (((long)seta[1]) << 32);
         }
         sigdelsetmask(&set, ~_BLOCKABLE);
-        spin_lock_irq(&current->sighand->siglock);
-        current->blocked = set;
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+        set_current_blocked(&set);
         return;
 
 segv:
@@ -377,10 +374,7 @@ asmlinkage void do_rt_sigreturn32(struct pt_regs *regs)
                 case 1: set.sig[0] = seta.sig[0] + (((long)seta.sig[1]) << 32);
         }
         sigdelsetmask(&set, ~_BLOCKABLE);
-        spin_lock_irq(&current->sighand->siglock);
-        current->blocked = set;
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+        set_current_blocked(&set);
         return;
 segv:
         force_sig(SIGSEGV, current);
@@ -782,6 +776,7 @@ static inline int handle_signal32(unsigned long signr, struct k_sigaction *ka,
                                   siginfo_t *info,
                                   sigset_t *oldset, struct pt_regs *regs)
 {
+        sigset_t blocked;
         int err;
 
         if (ka->sa.sa_flags & SA_SIGINFO)
@@ -792,12 +787,10 @@ static inline int handle_signal32(unsigned long signr, struct k_sigaction *ka,
         if (err)
                 return err;
 
-        spin_lock_irq(&current->sighand->siglock);
-        sigorsets(&current->blocked,&current->blocked,&ka->sa.sa_mask);
+        sigorsets(&blocked, &current->blocked, &ka->sa.sa_mask);
         if (!(ka->sa.sa_flags & SA_NOMASK))
-                sigaddset(&current->blocked,signr);
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+                sigaddset(&blocked, signr);
+        set_current_blocked(&blocked);
 
         tracehook_signal_handler(signr, info, ka, regs, 0);
 
@@ -829,21 +822,23 @@ static inline void syscall_restart32(unsigned long orig_i0, struct pt_regs *regs
  * want to handle. Thus you cannot kill init even with a SIGKILL even by
  * mistake.
  */
-void do_signal32(sigset_t *oldset, struct pt_regs * regs,
-                 int restart_syscall, unsigned long orig_i0)
+void do_signal32(sigset_t *oldset, struct pt_regs * regs)
 {
         struct k_sigaction ka;
+        unsigned long orig_i0;
+        int restart_syscall;
         siginfo_t info;
         int signr;
         
         signr = get_signal_to_deliver(&info, &ka, regs, NULL);
 
-        /* If the debugger messes with the program counter, it clears
-         * the "in syscall" bit, directing us to not perform a syscall
-         * restart.
-         */
-        if (restart_syscall && !pt_regs_is_syscall(regs))
-                restart_syscall = 0;
+        restart_syscall = 0;
+        orig_i0 = 0;
+        if (pt_regs_is_syscall(regs) &&
+            (regs->tstate & (TSTATE_XCARRY | TSTATE_ICARRY))) {
+                restart_syscall = 1;
+                orig_i0 = regs->u_regs[UREG_G6];
+        }
 
         if (signr > 0) {
                 if (restart_syscall)
@@ -881,7 +876,7 @@ void do_signal32(sigset_t *oldset, struct pt_regs * regs,
          */
         if (current_thread_info()->status & TS_RESTORE_SIGMASK) {
                 current_thread_info()->status &= ~TS_RESTORE_SIGMASK;
-                sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
+                set_current_blocked(&current->saved_sigmask);
         }
 }
 
diff --git a/arch/sparc/kernel/signal_32.c b/arch/sparc/kernel/signal_32.c
index 04ede8f04add..d54c6e53aba0 100644
--- a/arch/sparc/kernel/signal_32.c
+++ b/arch/sparc/kernel/signal_32.c
@@ -62,12 +62,13 @@ struct rt_signal_frame {
 
 static int _sigpause_common(old_sigset_t set)
 {
-        set &= _BLOCKABLE;
-        spin_lock_irq(&current->sighand->siglock);
+        sigset_t blocked;
+
         current->saved_sigmask = current->blocked;
-        siginitset(&current->blocked, set);
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+
+        set &= _BLOCKABLE;
+        siginitset(&blocked, set);
+        set_current_blocked(&blocked);
 
         current->state = TASK_INTERRUPTIBLE;
         schedule();
@@ -139,10 +140,7 @@ asmlinkage void do_sigreturn(struct pt_regs *regs)
                 goto segv_and_exit;
 
         sigdelsetmask(&set, ~_BLOCKABLE);
-        spin_lock_irq(&current->sighand->siglock);
-        current->blocked = set;
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+        set_current_blocked(&set);
         return;
 
 segv_and_exit:
@@ -209,10 +207,7 @@ asmlinkage void do_rt_sigreturn(struct pt_regs *regs)
         }
 
         sigdelsetmask(&set, ~_BLOCKABLE);
-        spin_lock_irq(&current->sighand->siglock);
-        current->blocked = set;
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+        set_current_blocked(&set);
         return;
 segv:
         force_sig(SIGSEGV, current);
@@ -470,6 +465,7 @@ static inline int
 handle_signal(unsigned long signr, struct k_sigaction *ka,
               siginfo_t *info, sigset_t *oldset, struct pt_regs *regs)
 {
+        sigset_t blocked;
         int err;
 
         if (ka->sa.sa_flags & SA_SIGINFO)
@@ -480,12 +476,10 @@ handle_signal(unsigned long signr, struct k_sigaction *ka,
         if (err)
                 return err;
 
-        spin_lock_irq(&current->sighand->siglock);
-        sigorsets(&current->blocked,&current->blocked,&ka->sa.sa_mask);
+        sigorsets(&blocked, &current->blocked, &ka->sa.sa_mask);
         if (!(ka->sa.sa_flags & SA_NOMASK))
-                sigaddset(&current->blocked, signr);
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+                sigaddset(&blocked, signr);
+        set_current_blocked(&blocked);
 
         tracehook_signal_handler(signr, info, ka, regs, 0);
 
@@ -525,10 +519,26 @@ static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
         siginfo_t info;
         int signr;
 
+        /* It's a lot of work and synchronization to add a new ptrace
+         * register for GDB to save and restore in order to get
+         * orig_i0 correct for syscall restarts when debugging.
+         *
+         * Although it should be the case that most of the global
+         * registers are volatile across a system call, glibc already
+         * depends upon that fact that we preserve them.  So we can't
+         * just use any global register to save away the orig_i0 value.
+         *
+         * In particular %g2, %g3, %g4, and %g5 are all assumed to be
+         * preserved across a system call trap by various pieces of
+         * code in glibc.
+         *
+         * %g7 is used as the "thread register".   %g6 is not used in
+         * any fixed manner.  %g6 is used as a scratch register and
+         * a compiler temporary, but it's value is never used across
+         * a system call.  Therefore %g6 is usable for orig_i0 storage.
+         */
         if (pt_regs_is_syscall(regs) && (regs->psr & PSR_C))
-                restart_syscall = 1;
-        else
-                restart_syscall = 0;
+                regs->u_regs[UREG_G6] = orig_i0;
 
         if (test_thread_flag(TIF_RESTORE_SIGMASK))
                 oldset = &current->saved_sigmask;
@@ -541,8 +551,12 @@ static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
          * the software "in syscall" bit, directing us to not perform
          * a syscall restart.
          */
-        if (restart_syscall && !pt_regs_is_syscall(regs))
-                restart_syscall = 0;
+        restart_syscall = 0;
+        if (pt_regs_is_syscall(regs) && (regs->psr & PSR_C)) {
+                restart_syscall = 1;
+                orig_i0 = regs->u_regs[UREG_G6];
+        }
+
 
         if (signr > 0) {
                 if (restart_syscall)
@@ -581,7 +595,7 @@ static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
          */
         if (test_thread_flag(TIF_RESTORE_SIGMASK)) {
                 clear_thread_flag(TIF_RESTORE_SIGMASK);
-                sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
+                set_current_blocked(&current->saved_sigmask);
         }
 }
 
diff --git a/arch/sparc/kernel/signal_64.c b/arch/sparc/kernel/signal_64.c
index 47509df3b893..f0836cd0e2f2 100644
--- a/arch/sparc/kernel/signal_64.c
+++ b/arch/sparc/kernel/signal_64.c
@@ -70,10 +70,7 @@ asmlinkage void sparc64_set_context(struct pt_regs *regs)
                                 goto do_sigsegv;
                 }
                 sigdelsetmask(&set, ~_BLOCKABLE);
-                spin_lock_irq(&current->sighand->siglock);
-                current->blocked = set;
-                recalc_sigpending();
-                spin_unlock_irq(&current->sighand->siglock);
+                set_current_blocked(&set);
         }
         if (test_thread_flag(TIF_32BIT)) {
                 pc &= 0xffffffff;
@@ -242,12 +239,13 @@ struct rt_signal_frame {
 
 static long _sigpause_common(old_sigset_t set)
 {
-        set &= _BLOCKABLE;
-        spin_lock_irq(&current->sighand->siglock);
+        sigset_t blocked;
+
         current->saved_sigmask = current->blocked;
-        siginitset(&current->blocked, set);
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+
+        set &= _BLOCKABLE;
+        siginitset(&blocked, set);
+        set_current_blocked(&blocked);
 
         current->state = TASK_INTERRUPTIBLE;
         schedule();
@@ -327,10 +325,7 @@ void do_rt_sigreturn(struct pt_regs *regs)
         pt_regs_clear_syscall(regs);
 
         sigdelsetmask(&set, ~_BLOCKABLE);
-        spin_lock_irq(&current->sighand->siglock);
-        current->blocked = set;
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+        set_current_blocked(&set);
         return;
 segv:
         force_sig(SIGSEGV, current);
@@ -484,18 +479,17 @@ static inline int handle_signal(unsigned long signr, struct k_sigaction *ka,
                                 siginfo_t *info,
                                 sigset_t *oldset, struct pt_regs *regs)
 {
+        sigset_t blocked;
         int err;
 
         err = setup_rt_frame(ka, regs, signr, oldset,
                              (ka->sa.sa_flags & SA_SIGINFO) ? info : NULL);
         if (err)
                 return err;
-        spin_lock_irq(&current->sighand->siglock);
-        sigorsets(&current->blocked,&current->blocked,&ka->sa.sa_mask);
+        sigorsets(&blocked, &current->blocked, &ka->sa.sa_mask);
         if (!(ka->sa.sa_flags & SA_NOMASK))
-                sigaddset(&current->blocked,signr);
-        recalc_sigpending();
-        spin_unlock_irq(&current->sighand->siglock);
+                sigaddset(&blocked, signr);
+        set_current_blocked(&blocked);
 
         tracehook_signal_handler(signr, info, ka, regs, 0);
 
@@ -535,11 +529,27 @@ static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
         siginfo_t info;
         int signr;
         
+        /* It's a lot of work and synchronization to add a new ptrace
+         * register for GDB to save and restore in order to get
+         * orig_i0 correct for syscall restarts when debugging.
+         *
+         * Although it should be the case that most of the global
+         * registers are volatile across a system call, glibc already
+         * depends upon that fact that we preserve them.  So we can't
+         * just use any global register to save away the orig_i0 value.
+         *
+         * In particular %g2, %g3, %g4, and %g5 are all assumed to be
+         * preserved across a system call trap by various pieces of
+         * code in glibc.
+         *
+         * %g7 is used as the "thread register".   %g6 is not used in
+         * any fixed manner.  %g6 is used as a scratch register and
+         * a compiler temporary, but it's value is never used across
+         * a system call.  Therefore %g6 is usable for orig_i0 storage.
+         */
         if (pt_regs_is_syscall(regs) &&
-            (regs->tstate & (TSTATE_XCARRY | TSTATE_ICARRY))) {
-                restart_syscall = 1;
-        } else
-                restart_syscall = 0;
+            (regs->tstate & (TSTATE_XCARRY | TSTATE_ICARRY)))
+                regs->u_regs[UREG_G6] = orig_i0;
 
         if (current_thread_info()->status & TS_RESTORE_SIGMASK)
                 oldset = &current->saved_sigmask;
@@ -548,22 +558,20 @@ static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
 
 #ifdef CONFIG_COMPAT
         if (test_thread_flag(TIF_32BIT)) {
-                extern void do_signal32(sigset_t *, struct pt_regs *,
-                                        int restart_syscall,
-                                        unsigned long orig_i0);
-                do_signal32(oldset, regs, restart_syscall, orig_i0);
+                extern void do_signal32(sigset_t *, struct pt_regs *);
+                do_signal32(oldset, regs);
                 return;
         }
 #endif  
 
         signr = get_signal_to_deliver(&info, &ka, regs, NULL);
 
-        /* If the debugger messes with the program counter, it clears
-         * the software "in syscall" bit, directing us to not perform
-         * a syscall restart.
-         */
-        if (restart_syscall && !pt_regs_is_syscall(regs))
-                restart_syscall = 0;
+        restart_syscall = 0;
+        if (pt_regs_is_syscall(regs) &&
+            (regs->tstate & (TSTATE_XCARRY | TSTATE_ICARRY))) {
+                restart_syscall = 1;
+                orig_i0 = regs->u_regs[UREG_G6];
+        }
 
         if (signr > 0) {
                 if (restart_syscall)
@@ -601,7 +609,7 @@ static void do_signal(struct pt_regs *regs, unsigned long orig_i0)
          */
         if (current_thread_info()->status & TS_RESTORE_SIGMASK) {
                 current_thread_info()->status &= ~TS_RESTORE_SIGMASK;
-                sigprocmask(SIG_SETMASK, &current->saved_sigmask, NULL);
+                set_current_blocked(&current->saved_sigmask);
         }
 }
 
diff --git a/arch/sparc/kernel/sigutil_64.c b/arch/sparc/kernel/sigutil_64.c
index e7dc508c38eb..b19570d41a39 100644
--- a/arch/sparc/kernel/sigutil_64.c
+++ b/arch/sparc/kernel/sigutil_64.c
@@ -2,6 +2,7 @@
 #include <linux/types.h>
 #include <linux/thread_info.h>
 #include <linux/uaccess.h>
+#include <linux/errno.h>
 
 #include <asm/sigcontext.h>
 #include <asm/fpumacro.h>
diff --git a/arch/sparc/kernel/smp_64.c b/arch/sparc/kernel/smp_64.c
index 4a442c32e117..75607724d290 100644
--- a/arch/sparc/kernel/smp_64.c
+++ b/arch/sparc/kernel/smp_64.c
@@ -3,7 +3,7 @@
  * Copyright (C) 1997, 2007, 2008 David S. Miller (davem@davemloft.net)
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/kernel.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
diff --git a/arch/sparc/kernel/sparc_ksyms_64.c b/arch/sparc/kernel/sparc_ksyms_64.c
index 83b47ab02d96..12ff09824cd9 100644
--- a/arch/sparc/kernel/sparc_ksyms_64.c
+++ b/arch/sparc/kernel/sparc_ksyms_64.c
@@ -5,7 +5,7 @@
  * Copyright (C) 1999 Jakub Jelinek (jj@ultra.linux.cz)
  */
 
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/pci.h>
 #include <linux/init.h>
 #include <linux/bitops.h>
diff --git a/arch/sparc/kernel/stacktrace.c b/arch/sparc/kernel/stacktrace.c
index 3e0815349630..e78386a0029f 100644
--- a/arch/sparc/kernel/stacktrace.c
+++ b/arch/sparc/kernel/stacktrace.c
@@ -2,7 +2,7 @@
 #include <linux/stacktrace.h>
 #include <linux/thread_info.h>
 #include <linux/ftrace.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <asm/ptrace.h>
 #include <asm/stacktrace.h>
 
diff --git a/arch/sparc/kernel/sys_sparc32.c b/arch/sparc/kernel/sys_sparc32.c
index 170cd8e8eb2a..29c478ffed91 100644
--- a/arch/sparc/kernel/sys_sparc32.c
+++ b/arch/sparc/kernel/sys_sparc32.c
@@ -23,7 +23,6 @@
 #include <linux/uio.h>
 #include <linux/nfs_fs.h>
 #include <linux/quota.h>
-#include <linux/module.h>
 #include <linux/poll.h>
 #include <linux/personality.h>
 #include <linux/stat.h>
diff --git a/arch/sparc/kernel/sys_sparc_64.c b/arch/sparc/kernel/sys_sparc_64.c
index 908b47a5ee24..441521ad8a3f 100644
--- a/arch/sparc/kernel/sys_sparc_64.c
+++ b/arch/sparc/kernel/sys_sparc_64.c
@@ -23,7 +23,7 @@
 #include <linux/ipc.h>
 #include <linux/personality.h>
 #include <linux/random.h>
-#include <linux/module.h>
+#include <linux/export.h>
 
 #include <asm/uaccess.h>
 #include <asm/utrap.h>
diff --git a/arch/sparc/kernel/systbls_32.S b/arch/sparc/kernel/systbls_32.S
index 09d8ec454450..63402f9e9f51 100644
--- a/arch/sparc/kernel/systbls_32.S
+++ b/arch/sparc/kernel/systbls_32.S
@@ -84,4 +84,4 @@ sys_call_table:
 /*320*/ .long sys_dup3, sys_pipe2, sys_inotify_init1, sys_accept4, sys_preadv
 /*325*/ .long sys_pwritev, sys_rt_tgsigqueueinfo, sys_perf_event_open, sys_recvmmsg, sys_fanotify_init
 /*330*/ .long sys_fanotify_mark, sys_prlimit64, sys_name_to_handle_at, sys_open_by_handle_at, sys_clock_adjtime
-/*335*/ .long sys_syncfs, sys_sendmmsg, sys_setns
+/*335*/ .long sys_syncfs, sys_sendmmsg, sys_setns, sys_process_vm_readv, sys_process_vm_writev
diff --git a/arch/sparc/kernel/systbls_64.S b/arch/sparc/kernel/systbls_64.S
index c9296ab0b1f4..db86b1a0e9a9 100644
--- a/arch/sparc/kernel/systbls_64.S
+++ b/arch/sparc/kernel/systbls_64.S
@@ -68,7 +68,7 @@ sys_call_table32:
         .word compat_sys_fstatfs64, sys_llseek, sys_mlock, sys_munlock, sys32_mlockall
 /*240*/ .word sys_munlockall, sys32_sched_setparam, sys32_sched_getparam, sys32_sched_setscheduler, sys32_sched_getscheduler
         .word sys_sched_yield, sys32_sched_get_priority_max, sys32_sched_get_priority_min, sys32_sched_rr_get_interval, compat_sys_nanosleep
-/*250*/ .word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys32_nfsservctl
+/*250*/ .word sys_mremap, compat_sys_sysctl, sys32_getsid, sys_fdatasync, sys_nis_syscall
         .word sys32_sync_file_range, compat_sys_clock_settime, compat_sys_clock_gettime, compat_sys_clock_getres, sys32_clock_nanosleep
 /*260*/ .word compat_sys_sched_getaffinity, compat_sys_sched_setaffinity, sys32_timer_settime, compat_sys_timer_gettime, sys_timer_getoverrun
         .word sys_timer_delete, compat_sys_timer_create, sys_ni_syscall, compat_sys_io_setup, sys_io_destroy
@@ -85,7 +85,7 @@ sys_call_table32:
 /*320*/ .word sys_dup3, sys_pipe2, sys_inotify_init1, sys_accept4, compat_sys_preadv
         .word compat_sys_pwritev, compat_sys_rt_tgsigqueueinfo, sys_perf_event_open, compat_sys_recvmmsg, sys_fanotify_init
 /*330*/ .word sys32_fanotify_mark, sys_prlimit64, sys_name_to_handle_at, compat_sys_open_by_handle_at, compat_sys_clock_adjtime
-        .word sys_syncfs, compat_sys_sendmmsg, sys_setns
+        .word sys_syncfs, compat_sys_sendmmsg, sys_setns, compat_sys_process_vm_readv, compat_sys_process_vm_writev
 
 #endif /* CONFIG_COMPAT */
 
@@ -162,4 +162,4 @@ sys_call_table:
 /*320*/ .word sys_dup3, sys_pipe2, sys_inotify_init1, sys_accept4, sys_preadv
         .word sys_pwritev, sys_rt_tgsigqueueinfo, sys_perf_event_open, sys_recvmmsg, sys_fanotify_init
 /*330*/ .word sys_fanotify_mark, sys_prlimit64, sys_name_to_handle_at, sys_open_by_handle_at, sys_clock_adjtime
-        .word sys_syncfs, sys_sendmmsg, sys_setns
+        .word sys_syncfs, sys_sendmmsg, sys_setns, sys_process_vm_readv, sys_process_vm_writev
diff --git a/arch/sparc/kernel/time_64.c b/arch/sparc/kernel/time_64.c
index 1db6b18964d2..e861072b9c52 100644
--- a/arch/sparc/kernel/time_64.c
+++ b/arch/sparc/kernel/time_64.c
@@ -9,7 +9,7 @@
  */
 
 #include <linux/errno.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/sched.h>
 #include <linux/kernel.h>
 #include <linux/param.h>
diff --git a/arch/sparc/kernel/traps_32.c b/arch/sparc/kernel/traps_32.c
index c0490c7bbde0..591f20ca9e48 100644
--- a/arch/sparc/kernel/traps_32.c
+++ b/arch/sparc/kernel/traps_32.c
@@ -14,6 +14,7 @@
 #include <linux/signal.h>
 #include <linux/smp.h>
 #include <linux/kdebug.h>
+#include <linux/export.h>
 
 #include <asm/delay.h>
 #include <asm/system.h>
diff --git a/arch/sparc/kernel/unaligned_32.c b/arch/sparc/kernel/unaligned_32.c
index 7efbb2f9e77f..4d043a1b2492 100644
--- a/arch/sparc/kernel/unaligned_32.c
+++ b/arch/sparc/kernel/unaligned_32.c
@@ -10,7 +10,6 @@
 #include <linux/kernel.h>
 #include <linux/sched.h>
 #include <linux/mm.h>
-#include <linux/module.h>
 #include <asm/ptrace.h>
 #include <asm/processor.h>
 #include <asm/system.h>
diff --git a/arch/sparc/kernel/vio.c b/arch/sparc/kernel/vio.c
index 3cb1def9806c..f67e28ef598c 100644
--- a/arch/sparc/kernel/vio.c
+++ b/arch/sparc/kernel/vio.c
@@ -12,6 +12,7 @@
 #include <linux/kernel.h>
 #include <linux/slab.h>
 #include <linux/irq.h>
+#include <linux/export.h>
 #include <linux/init.h>
 
 #include <asm/mdesc.h>
diff --git a/arch/sparc/kernel/viohs.c b/arch/sparc/kernel/viohs.c
index 29348ea139c3..f8e7dd53e1c7 100644
--- a/arch/sparc/kernel/viohs.c
+++ b/arch/sparc/kernel/viohs.c
@@ -4,7 +4,7 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
+#include <linux/export.h>
 #include <linux/string.h>
 #include <linux/delay.h>
 #include <linux/sched.h>
diff --git a/arch/sparc/kernel/visemul.c b/arch/sparc/kernel/visemul.c
index 32b626c9d815..73370674ccff 100644
--- a/arch/sparc/kernel/visemul.c
+++ b/arch/sparc/kernel/visemul.c
@@ -713,17 +713,17 @@ static void pcmp(struct pt_regs *regs, unsigned int insn, unsigned int opf)
                         s16 b = (rs2 >> (i * 16)) & 0xffff;
 
                         if (a > b)
-                                rd_val |= 1 << i;
+                                rd_val |= 8 >> i;
                 }
                 break;
 
         case FCMPGT32_OPF:
                 for (i = 0; i < 2; i++) {
-                        s32 a = (rs1 >> (i * 32)) & 0xffff;
-                        s32 b = (rs2 >> (i * 32)) & 0xffff;
+                        s32 a = (rs1 >> (i * 32)) & 0xffffffff;
+                        s32 b = (rs2 >> (i * 32)) & 0xffffffff;
 
                         if (a > b)
-                                rd_val |= 1 << i;
+                                rd_val |= 2 >> i;
                 }
                 break;
 
@@ -733,17 +733,17 @@ static void pcmp(struct pt_regs *regs, unsigned int insn, unsigned int opf)
                         s16 b = (rs2 >> (i * 16)) & 0xffff;
 
                         if (a <= b)
-                                rd_val |= 1 << i;
+                                rd_val |= 8 >> i;
                 }
                 break;
 
         case FCMPLE32_OPF:
                 for (i = 0; i < 2; i++) {
-                        s32 a = (rs1 >> (i * 32)) & 0xffff;
-                        s32 b = (rs2 >> (i * 32)) & 0xffff;
+                        s32 a = (rs1 >> (i * 32)) & 0xffffffff;
+                        s32 b = (rs2 >> (i * 32)) & 0xffffffff;
 
                         if (a <= b)
-                                rd_val |= 1 << i;
+                                rd_val |= 2 >> i;
                 }
                 break;
 
@@ -753,17 +753,17 @@ static void pcmp(struct pt_regs *regs, unsigned int insn, unsigned int opf)
                         s16 b = (rs2 >> (i * 16)) & 0xffff;
 
                         if (a != b)
-                                rd_val |= 1 << i;
+                                rd_val |= 8 >> i;
                 }
                 break;
 
         case FCMPNE32_OPF:
                 for (i = 0; i < 2; i++) {
-                        s32 a = (rs1 >> (i * 32)) & 0xffff;
-                        s32 b = (rs2 >> (i * 32)) & 0xffff;
+                        s32 a = (rs1 >> (i * 32)) & 0xffffffff;
+                        s32 b = (rs2 >> (i * 32)) & 0xffffffff;
 
                         if (a != b)
-                                rd_val |= 1 << i;
+                                rd_val |= 2 >> i;
                 }
                 break;
 
@@ -773,17 +773,17 @@ static void pcmp(struct pt_regs *regs, unsigned int insn, unsigned int opf)
                         s16 b = (rs2 >> (i * 16)) & 0xffff;
 
                         if (a == b)
-                                rd_val |= 1 << i;
+                                rd_val |= 8 >> i;
                 }
                 break;
 
         case FCMPEQ32_OPF:
                 for (i = 0; i < 2; i++) {
-                        s32 a = (rs1 >> (i * 32)) & 0xffff;
-                        s32 b = (rs2 >> (i * 32)) & 0xffff;
+                        s32 a = (rs1 >> (i * 32)) & 0xffffffff;
+                        s32 b = (rs2 >> (i * 32)) & 0xffffffff;
 
                         if (a == b)
-                                rd_val |= 1 << i;
+                                rd_val |= 2 >> i;
                 }
                 break;
         }
diff --git a/arch/sparc/lib/memcpy.S b/arch/sparc/lib/memcpy.S
index 34fe65751737..4d8c497517bd 100644
--- a/arch/sparc/lib/memcpy.S
+++ b/arch/sparc/lib/memcpy.S
@@ -7,40 +7,12 @@
  * Copyright (C) 1996 Jakub Jelinek (jj@sunsite.mff.cuni.cz)
  */
 
-#ifdef __KERNEL__
-
-#define FUNC(x)                                                                                         \
+#define FUNC(x)                 \
         .globl  x;              \
         .type   x,@function;    \
-        .align  4;                                                                                      \
+        .align  4;              \
 x:
 
-#undef FASTER_REVERSE
-#undef FASTER_NONALIGNED
-#define FASTER_ALIGNED
-
-/* In kernel these functions don't return a value.
- * One should use macros in asm/string.h for that purpose.
- * We return 0, so that bugs are more apparent.
- */
-#define SETUP_RETL
-#define RETL_INSN       clr     %o0
-
-#else
-
-/* libc */
-
-#include "DEFS.h"
-
-#define FASTER_REVERSE
-#define FASTER_NONALIGNED
-#define FASTER_ALIGNED
-
-#define SETUP_RETL      mov     %o0, %g6
-#define RETL_INSN       mov     %g6, %o0
-
-#endif
-
 /* Both these macros have to start with exactly the same insn */
 #define MOVE_BIGCHUNK(src, dst, offset, t0, t1, t2, t3, t4, t5, t6, t7) \
         ldd     [%src + (offset) + 0x00], %t0; \
@@ -164,30 +136,6 @@ x:
         .text
         .align  4
 
-#ifdef FASTER_REVERSE
-
-70:     /* rdword_align */
-
-        andcc           %o1, 1, %g0
-        be              4f
-         andcc          %o1, 2, %g0
-
-        ldub            [%o1 - 1], %g2
-        sub             %o1, 1, %o1
-        stb             %g2, [%o0 - 1]
-        sub             %o2, 1, %o2
-        be              3f
-         sub            %o0, 1, %o0
-4:
-        lduh            [%o1 - 2], %g2
-        sub             %o1, 2, %o1
-        sth             %g2, [%o0 - 2]
-        sub             %o2, 2, %o2
-        b               3f
-         sub            %o0, 2, %o0
-
-#endif /* FASTER_REVERSE */
-
 0:
         retl
          nop            ! Only bcopy returns here and it retuns void...
@@ -198,7 +146,7 @@ FUNC(__memmove)
 #endif
 FUNC(memmove)
         cmp             %o0, %o1
-        SETUP_RETL
+        mov             %o0, %g7
         bleu            9f
          sub            %o0, %o1, %o4
 
@@ -207,8 +155,6 @@ FUNC(memmove)
         bleu            0f
          andcc          %o4, 3, %o5
 
-#ifndef FASTER_REVERSE
-
         add             %o1, %o2, %o1
         add             %o0, %o2, %o0
         sub             %o1, 1, %o1
@@ -224,295 +170,7 @@ FUNC(memmove)
          sub            %o0, 1, %o0
 
         retl
-         RETL_INSN
-
-#else /* FASTER_REVERSE */
-
-        add             %o1, %o2, %o1
-        add             %o0, %o2, %o0
-        bne             77f
-         cmp            %o2, 15
-        bleu            91f
-         andcc          %o1, 3, %g0
-        bne             70b
-3:
-         andcc          %o1, 4, %g0
-
-        be              2f
-         mov            %o2, %g1
-
-        ld              [%o1 - 4], %o4
-        sub             %g1, 4, %g1
-        st              %o4, [%o0 - 4]
-        sub             %o1, 4, %o1
-        sub             %o0, 4, %o0
-2:
-        andcc           %g1, 0xffffff80, %g7
-        be              3f
-         andcc          %o0, 4, %g0
-
-        be              74f + 4
-5:
-        RMOVE_BIGCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
-        RMOVE_BIGCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
-        RMOVE_BIGCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
-        RMOVE_BIGCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
-        subcc           %g7, 128, %g7
-        sub             %o1, 128, %o1
-        bne             5b
-         sub            %o0, 128, %o0
-3:
-        andcc           %g1, 0x70, %g7
-        be              72f
-         andcc          %g1, 8, %g0
-
-        sethi           %hi(72f), %o5
-        srl             %g7, 1, %o4
-        add             %g7, %o4, %o4
-        sub             %o1, %g7, %o1
-        sub             %o5, %o4, %o5
-        jmpl            %o5 + %lo(72f), %g0
-         sub            %o0, %g7, %o0
-
-71:     /* rmemcpy_table */
-        RMOVE_LASTCHUNK(o1, o0, 0x60, g2, g3, g4, g5)
-        RMOVE_LASTCHUNK(o1, o0, 0x50, g2, g3, g4, g5)
-        RMOVE_LASTCHUNK(o1, o0, 0x40, g2, g3, g4, g5)
-        RMOVE_LASTCHUNK(o1, o0, 0x30, g2, g3, g4, g5)
-        RMOVE_LASTCHUNK(o1, o0, 0x20, g2, g3, g4, g5)
-        RMOVE_LASTCHUNK(o1, o0, 0x10, g2, g3, g4, g5)
-        RMOVE_LASTCHUNK(o1, o0, 0x00, g2, g3, g4, g5)
-
-72:     /* rmemcpy_table_end */
-
-        be              73f
-         andcc          %g1, 4, %g0
-
-        ldd             [%o1 - 0x08], %g2
-        sub             %o0, 8, %o0
-        sub             %o1, 8, %o1
-        st              %g2, [%o0]
-        st              %g3, [%o0 + 0x04]
-
-73:     /* rmemcpy_last7 */
-
-        be              1f
-         andcc          %g1, 2, %g0
-
-        ld              [%o1 - 4], %g2
-        sub             %o1, 4, %o1
-        st              %g2, [%o0 - 4]
-        sub             %o0, 4, %o0
-1:
-        be              1f
-         andcc          %g1, 1, %g0
-
-        lduh            [%o1 - 2], %g2
-        sub             %o1, 2, %o1
-        sth             %g2, [%o0 - 2]
-        sub             %o0, 2, %o0
-1:
-        be              1f
-         nop
-
-        ldub            [%o1 - 1], %g2
-        stb             %g2, [%o0 - 1]
-1:
-        retl
-         RETL_INSN
-
-74:     /* rldd_std */
-        RMOVE_BIGALIGNCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
-        RMOVE_BIGALIGNCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
-        RMOVE_BIGALIGNCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
-        RMOVE_BIGALIGNCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
-        subcc           %g7, 128, %g7
-        sub             %o1, 128, %o1
-        bne             74b
-         sub            %o0, 128, %o0
-
-        andcc           %g1, 0x70, %g7
-        be              72b
-         andcc          %g1, 8, %g0
-
-        sethi           %hi(72b), %o5
-        srl             %g7, 1, %o4
-        add             %g7, %o4, %o4
-        sub             %o1, %g7, %o1
-        sub             %o5, %o4, %o5
-        jmpl            %o5 + %lo(72b), %g0
-         sub            %o0, %g7, %o0
-
-75:     /* rshort_end */
-
-        and             %o2, 0xe, %o3
-2:
-        sethi           %hi(76f), %o5
-        sll             %o3, 3, %o4
-        sub             %o0, %o3, %o0
-        sub             %o5, %o4, %o5
-        sub             %o1, %o3, %o1
-        jmpl            %o5 + %lo(76f), %g0
-         andcc          %o2, 1, %g0
-
-        RMOVE_SHORTCHUNK(o1, o0, 0x0c, g2, g3)
-        RMOVE_SHORTCHUNK(o1, o0, 0x0a, g2, g3)
-        RMOVE_SHORTCHUNK(o1, o0, 0x08, g2, g3)
-        RMOVE_SHORTCHUNK(o1, o0, 0x06, g2, g3)
-        RMOVE_SHORTCHUNK(o1, o0, 0x04, g2, g3)
-        RMOVE_SHORTCHUNK(o1, o0, 0x02, g2, g3)
-        RMOVE_SHORTCHUNK(o1, o0, 0x00, g2, g3)
-
-76:     /* rshort_table_end */
-
-        be              1f
-         nop
-        ldub            [%o1 - 1], %g2
-        stb             %g2, [%o0 - 1]
-1:
-        retl
-         RETL_INSN
-
-91:     /* rshort_aligned_end */
-
-        bne             75b
-         andcc          %o2, 8, %g0
-
-        be              1f
-         andcc          %o2, 4, %g0
-
-        ld              [%o1 - 0x08], %g2
-        ld              [%o1 - 0x04], %g3
-        sub             %o1, 8, %o1
-        st              %g2, [%o0 - 0x08]
-        st              %g3, [%o0 - 0x04]
-        sub             %o0, 8, %o0
-1:
-        b               73b
-         mov            %o2, %g1
-
-77:     /* rnon_aligned */
-        cmp             %o2, 15
-        bleu            75b
-         andcc          %o0, 3, %g0
-        be              64f
-         andcc          %o0, 1, %g0
-        be              63f
-         andcc          %o0, 2, %g0
-        ldub            [%o1 - 1], %g5
-        sub             %o1, 1, %o1
-        stb             %g5, [%o0 - 1]
-        sub             %o0, 1, %o0
-        be              64f
-         sub            %o2, 1, %o2
-63:
-        ldub            [%o1 - 1], %g5
-        sub             %o1, 2, %o1
-        stb             %g5, [%o0 - 1]
-        sub             %o0, 2, %o0
-        ldub            [%o1], %g5
-        sub             %o2, 2, %o2
-        stb             %g5, [%o0]
-64:     
-        and             %o1, 3, %g2
-        and             %o1, -4, %o1
-        and             %o2, 0xc, %g3
-        add             %o1, 4, %o1
-        cmp             %g3, 4
-        sll             %g2, 3, %g4
-        mov             32, %g2
-        be              4f
-         sub            %g2, %g4, %g7
-
-        blu             3f
-         cmp            %g3, 8
-
-        be              2f
-         srl            %o2, 2, %g3
-
-        ld              [%o1 - 4], %o3
-        add             %o0, -8, %o0
-        ld              [%o1 - 8], %o4
-        add             %o1, -16, %o1
-        b               7f
-         add            %g3, 1, %g3
-2:
-        ld              [%o1 - 4], %o4
-        add             %o0, -4, %o0
-        ld              [%o1 - 8], %g1
-        add             %o1, -12, %o1
-        b               8f
-         add            %g3, 2, %g3
-3:
-        ld              [%o1 - 4], %o5
-        add             %o0, -12, %o0
-        ld              [%o1 - 8], %o3
-        add             %o1, -20, %o1
-        b               6f
-         srl            %o2, 2, %g3
-4:
-        ld              [%o1 - 4], %g1
-        srl             %o2, 2, %g3
-        ld              [%o1 - 8], %o5
-        add             %o1, -24, %o1
-        add             %o0, -16, %o0
-        add             %g3, -1, %g3
-
-        ld              [%o1 + 12], %o3
-5:
-        sll             %o5, %g4, %g2
-        srl             %g1, %g7, %g5
-        or              %g2, %g5, %g2
-        st              %g2, [%o0 + 12]
-6:
-        ld              [%o1 + 8], %o4
-        sll             %o3, %g4, %g2
-        srl             %o5, %g7, %g5
-        or              %g2, %g5, %g2
-        st              %g2, [%o0 + 8]
-7:
-        ld              [%o1 + 4], %g1
-        sll             %o4, %g4, %g2
-        srl             %o3, %g7, %g5
-        or              %g2, %g5, %g2
-        st              %g2, [%o0 + 4]
-8:
-        ld              [%o1], %o5
-        sll             %g1, %g4, %g2
-        srl             %o4, %g7, %g5
-        addcc           %g3, -4, %g3
-        or              %g2, %g5, %g2
-        add             %o1, -16, %o1
-        st              %g2, [%o0]
-        add             %o0, -16, %o0
-        bne,a           5b      
-         ld             [%o1 + 12], %o3
-        sll             %o5, %g4, %g2
-        srl             %g1, %g7, %g5
-        srl             %g4, 3, %g3
-        or              %g2, %g5, %g2
-        add             %o1, %g3, %o1
-        andcc           %o2, 2, %g0
-        st              %g2, [%o0 + 12]
-        be              1f
-         andcc          %o2, 1, %g0
-        
-        ldub            [%o1 + 15], %g5
-        add             %o1, -2, %o1
-        stb             %g5, [%o0 + 11]
-        add             %o0, -2, %o0
-        ldub            [%o1 + 16], %g5
-        stb             %g5, [%o0 + 12]
-1:
-        be              1f
-         nop
-        ldub            [%o1 + 15], %g5
-        stb             %g5, [%o0 + 11]
-1:
-        retl
-         RETL_INSN
-
-#endif /* FASTER_REVERSE */
+         mov            %g7, %o0
 
 /* NOTE: This code is executed just for the cases,
          where %src (=%o1) & 3 is != 0.
@@ -546,7 +204,7 @@ FUNC(memmove)
 FUNC(memcpy)    /* %o0=dst %o1=src %o2=len */
 
         sub             %o0, %o1, %o4
-        SETUP_RETL
+        mov             %o0, %g7
 9:
         andcc           %o4, 3, %o5
 0:
@@ -569,7 +227,7 @@ FUNC(memcpy)	/* %o0=dst %o1=src %o2=len */
         add             %o1, 4, %o1
         add             %o0, 4, %o0
 2:
-        andcc           %g1, 0xffffff80, %g7
+        andcc           %g1, 0xffffff80, %g0
         be              3f
          andcc          %o0, 4, %g0
 
@@ -579,22 +237,23 @@ FUNC(memcpy)	/* %o0=dst %o1=src %o2=len */
         MOVE_BIGCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
         MOVE_BIGCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
         MOVE_BIGCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
-        subcc           %g7, 128, %g7
+        sub             %g1, 128, %g1
         add             %o1, 128, %o1
-        bne             5b
+        cmp             %g1, 128
+        bge             5b
          add            %o0, 128, %o0
 3:
-        andcc           %g1, 0x70, %g7
+        andcc           %g1, 0x70, %g4
         be              80f
          andcc          %g1, 8, %g0
 
         sethi           %hi(80f), %o5
-        srl             %g7, 1, %o4
-        add             %g7, %o4, %o4
-        add             %o1, %g7, %o1
+        srl             %g4, 1, %o4
+        add             %g4, %o4, %o4
+        add             %o1, %g4, %o1
         sub             %o5, %o4, %o5
         jmpl            %o5 + %lo(80f), %g0
-         add            %o0, %g7, %o0
+         add            %o0, %g4, %o0
 
 79:     /* memcpy_table */
 
@@ -641,43 +300,28 @@ FUNC(memcpy)	/* %o0=dst %o1=src %o2=len */
         stb             %g2, [%o0]
 1:
         retl
-         RETL_INSN
+         mov            %g7, %o0
 
 82:     /* ldd_std */
         MOVE_BIGALIGNCHUNK(o1, o0, 0x00, o2, o3, o4, o5, g2, g3, g4, g5)
         MOVE_BIGALIGNCHUNK(o1, o0, 0x20, o2, o3, o4, o5, g2, g3, g4, g5)
         MOVE_BIGALIGNCHUNK(o1, o0, 0x40, o2, o3, o4, o5, g2, g3, g4, g5)
         MOVE_BIGALIGNCHUNK(o1, o0, 0x60, o2, o3, o4, o5, g2, g3, g4, g5)
-        subcc           %g7, 128, %g7
+        subcc           %g1, 128, %g1
         add             %o1, 128, %o1
-        bne             82b
+        cmp             %g1, 128
+        bge             82b
          add            %o0, 128, %o0
 
-#ifndef FASTER_ALIGNED
-
-        andcc           %g1, 0x70, %g7
-        be              80b
-         andcc          %g1, 8, %g0
-
-        sethi           %hi(80b), %o5
-        srl             %g7, 1, %o4
-        add             %g7, %o4, %o4
-        add             %o1, %g7, %o1
-        sub             %o5, %o4, %o5
-        jmpl            %o5 + %lo(80b), %g0
-         add            %o0, %g7, %o0
-
-#else /* FASTER_ALIGNED */
-
-        andcc           %g1, 0x70, %g7
+        andcc           %g1, 0x70, %g4
         be              84f
          andcc          %g1, 8, %g0
 
         sethi           %hi(84f), %o5
-        add             %o1, %g7, %o1
-        sub             %o5, %g7, %o5
+        add             %o1, %g4, %o1
+        sub             %o5, %g4, %o5
         jmpl            %o5 + %lo(84f), %g0
-         add            %o0, %g7, %o0
+         add            %o0, %g4, %o0
 
 83:     /* amemcpy_table */
 
@@ -721,382 +365,132 @@ FUNC(memcpy)	/* %o0=dst %o1=src %o2=len */
         stb             %g2, [%o0]
 1:
         retl
-         RETL_INSN
-
-#endif /* FASTER_ALIGNED */
+         mov            %g7, %o0
 
 86:     /* non_aligned */
         cmp             %o2, 6
         bleu            88f
+         nop
 
-#ifdef FASTER_NONALIGNED
-
-         cmp            %o2, 256
-        bcc             87f
-
-#endif /* FASTER_NONALIGNED */
-
-         andcc          %o0, 3, %g0
+        save            %sp, -96, %sp
+        andcc           %i0, 3, %g0
         be              61f
-         andcc          %o0, 1, %g0
+         andcc          %i0, 1, %g0
         be              60f
-         andcc          %o0, 2, %g0
+         andcc          %i0, 2, %g0
 
-        ldub            [%o1], %g5
-        add             %o1, 1, %o1
-        stb             %g5, [%o0]
-        sub             %o2, 1, %o2
+        ldub            [%i1], %g5
+        add             %i1, 1, %i1
+        stb             %g5, [%i0]
+        sub             %i2, 1, %i2
         bne             61f
-         add            %o0, 1, %o0
+         add            %i0, 1, %i0
 60:
-        ldub            [%o1], %g3
-        add             %o1, 2, %o1
-        stb             %g3, [%o0]
-        sub             %o2, 2, %o2
-        ldub            [%o1 - 1], %g3
-        add             %o0, 2, %o0
-        stb             %g3, [%o0 - 1]
+        ldub            [%i1], %g3
+        add             %i1, 2, %i1
+        stb             %g3, [%i0]
+        sub             %i2, 2, %i2
+        ldub            [%i1 - 1], %g3
+        add             %i0, 2, %i0
+        stb             %g3, [%i0 - 1]
 61:
-        and             %o1, 3, %g2
-        and             %o2, 0xc, %g3
-        and             %o1, -4, %o1
+        and             %i1, 3, %g2
+        and             %i2, 0xc, %g3
+        and             %i1, -4, %i1
         cmp             %g3, 4
         sll             %g2, 3, %g4
         mov             32, %g2
         be              4f
-         sub            %g2, %g4, %g7
+         sub            %g2, %g4, %l0
         
         blu             3f
          cmp            %g3, 0x8
 
         be              2f
-         srl            %o2, 2, %g3
+         srl            %i2, 2, %g3
 
-        ld              [%o1], %o3
-        add             %o0, -8, %o0
-        ld              [%o1 + 4], %o4
+        ld              [%i1], %i3
+        add             %i0, -8, %i0
+        ld              [%i1 + 4], %i4
         b               8f
          add            %g3, 1, %g3
 2:
-        ld              [%o1], %o4
-        add             %o0, -12, %o0
-        ld              [%o1 + 4], %o5
+        ld              [%i1], %i4
+        add             %i0, -12, %i0
+        ld              [%i1 + 4], %i5
         add             %g3, 2, %g3
         b               9f
-         add            %o1, -4, %o1
+         add            %i1, -4, %i1
 3:
-        ld              [%o1], %g1
-        add             %o0, -4, %o0
-        ld              [%o1 + 4], %o3
-        srl             %o2, 2, %g3
+        ld              [%i1], %g1
+        add             %i0, -4, %i0
+        ld              [%i1 + 4], %i3
+        srl             %i2, 2, %g3
         b               7f
-         add            %o1, 4, %o1
+         add            %i1, 4, %i1
 4:
-        ld              [%o1], %o5
-        cmp             %o2, 7
-        ld              [%o1 + 4], %g1
-        srl             %o2, 2, %g3
+        ld              [%i1], %i5
+        cmp             %i2, 7
+        ld              [%i1 + 4], %g1
+        srl             %i2, 2, %g3
         bleu            10f
-         add            %o1, 8, %o1
+         add            %i1, 8, %i1
 
-        ld              [%o1], %o3
+        ld              [%i1], %i3
         add             %g3, -1, %g3
 5:
-        sll             %o5, %g4, %g2
-        srl             %g1, %g7, %g5
+        sll             %i5, %g4, %g2
+        srl             %g1, %l0, %g5
         or              %g2, %g5, %g2
-        st              %g2, [%o0]
+        st              %g2, [%i0]
 7:
-        ld              [%o1 + 4], %o4
+        ld              [%i1 + 4], %i4
         sll             %g1, %g4, %g2
-        srl             %o3, %g7, %g5
+        srl             %i3, %l0, %g5
         or              %g2, %g5, %g2
-        st              %g2, [%o0 + 4]
+        st              %g2, [%i0 + 4]
 8:
-        ld              [%o1 + 8], %o5
-        sll             %o3, %g4, %g2
-        srl             %o4, %g7, %g5
+        ld              [%i1 + 8], %i5
+        sll             %i3, %g4, %g2
+        srl             %i4, %l0, %g5
         or              %g2, %g5, %g2
-        st              %g2, [%o0 + 8]
+        st              %g2, [%i0 + 8]
 9:
-        ld              [%o1 + 12], %g1
-        sll             %o4, %g4, %g2
-        srl             %o5, %g7, %g5
+        ld              [%i1 + 12], %g1
+        sll             %i4, %g4, %g2
+        srl             %i5, %l0, %g5
         addcc           %g3, -4, %g3
         or              %g2, %g5, %g2
-        add             %o1, 16, %o1
-        st              %g2, [%o0 + 12]
-        add             %o0, 16, %o0
+        add             %i1, 16, %i1
+        st              %g2, [%i0 + 12]
+        add             %i0, 16, %i0
         bne,a           5b
-         ld             [%o1], %o3
+         ld             [%i1], %i3
 10:
-        sll             %o5, %g4, %g2
-        srl             %g1, %g7, %g5
-        srl             %g7, 3, %g3
+        sll             %i5, %g4, %g2
+        srl             %g1, %l0, %g5
+        srl             %l0, 3, %g3
         or              %g2, %g5, %g2
-        sub             %o1, %g3, %o1
-        andcc           %o2, 2, %g0
-        st              %g2, [%o0]
+        sub             %i1, %g3, %i1
+        andcc           %i2, 2, %g0
+        st              %g2, [%i0]
         be              1f
-         andcc          %o2, 1, %g0
-
-        ldub            [%o1], %g2
-        add             %o1, 2, %o1
-        stb             %g2, [%o0 + 4]
-        add             %o0, 2, %o0
-        ldub            [%o1 - 1], %g2
-        stb             %g2, [%o0 + 3]
+         andcc          %i2, 1, %g0
+
+        ldub            [%i1], %g2
+        add             %i1, 2, %i1
+        stb             %g2, [%i0 + 4]
+        add             %i0, 2, %i0
+        ldub            [%i1 - 1], %g2
+        stb             %g2, [%i0 + 3]
 1:
         be              1f
          nop
-        ldub            [%o1], %g2
-        stb             %g2, [%o0 + 4]
-1:
-        retl
-         RETL_INSN
-
-#ifdef FASTER_NONALIGNED
-
-87:     /* faster_nonaligned */
-
-        andcc           %o1, 3, %g0
-        be              3f
-         andcc          %o1, 1, %g0
-
-        be              4f
-         andcc          %o1, 2, %g0
-
-        ldub            [%o1], %g2
-        add             %o1, 1, %o1
-        stb             %g2, [%o0]
-        sub             %o2, 1, %o2
-        bne             3f
-         add            %o0, 1, %o0
-4:
-        lduh            [%o1], %g2
-        add             %o1, 2, %o1
-        srl             %g2, 8, %g3
-        sub             %o2, 2, %o2
-        stb             %g3, [%o0]
-        add             %o0, 2, %o0
-        stb             %g2, [%o0 - 1]
-3:
-         andcc          %o1, 4, %g0
-
-        bne             2f
-         cmp            %o5, 1
-
-        ld              [%o1], %o4
-        srl             %o4, 24, %g2
-        stb             %g2, [%o0]
-        srl             %o4, 16, %g3
-        stb             %g3, [%o0 + 1]
-        srl             %o4, 8, %g2
-        stb             %g2, [%o0 + 2]
-        sub             %o2, 4, %o2
-        stb             %o4, [%o0 + 3]
-        add             %o1, 4, %o1
-        add             %o0, 4, %o0
-2:
-        be              33f
-         cmp            %o5, 2
-        be              32f
-         sub            %o2, 4, %o2
-31:
-        ld              [%o1], %g2
-        add             %o1, 4, %o1
-        srl             %g2, 24, %g3
-        and             %o0, 7, %g5
-        stb             %g3, [%o0]
-        cmp             %g5, 7
-        sll             %g2, 8, %g1
-        add             %o0, 4, %o0
-        be              41f
-         and            %o2, 0xffffffc0, %o3
-        ld              [%o0 - 7], %o4
-4:
-        SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        SMOVE_CHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        SMOVE_CHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        SMOVE_CHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        subcc           %o3, 64, %o3
-        add             %o1, 64, %o1
-        bne             4b
-         add            %o0, 64, %o0
-
-        andcc           %o2, 0x30, %o3
-        be,a            1f
-         srl            %g1, 16, %g2
-4:
-        SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        subcc           %o3, 16, %o3
-        add             %o1, 16, %o1
-        bne             4b
-         add            %o0, 16, %o0
-
-        srl             %g1, 16, %g2
-1:
-        st              %o4, [%o0 - 7]
-        sth             %g2, [%o0 - 3]
-        srl             %g1, 8, %g4
-        b               88f
-         stb            %g4, [%o0 - 1]
-32:
-        ld              [%o1], %g2
-        add             %o1, 4, %o1
-        srl             %g2, 16, %g3
-        and             %o0, 7, %g5
-        sth             %g3, [%o0]
-        cmp             %g5, 6
-        sll             %g2, 16, %g1
-        add             %o0, 4, %o0
-        be              42f
-         and            %o2, 0xffffffc0, %o3
-        ld              [%o0 - 6], %o4
-4:
-        SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        SMOVE_CHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        SMOVE_CHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        SMOVE_CHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        subcc           %o3, 64, %o3
-        add             %o1, 64, %o1
-        bne             4b
-         add            %o0, 64, %o0
-
-        andcc           %o2, 0x30, %o3
-        be,a            1f
-         srl            %g1, 16, %g2
-4:
-        SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        subcc           %o3, 16, %o3
-        add             %o1, 16, %o1
-        bne             4b
-         add            %o0, 16, %o0
-
-        srl             %g1, 16, %g2
-1:
-        st              %o4, [%o0 - 6]
-        b               88f
-         sth            %g2, [%o0 - 2]
-33:
-        ld              [%o1], %g2
-        sub             %o2, 4, %o2
-        srl             %g2, 24, %g3
-        and             %o0, 7, %g5
-        stb             %g3, [%o0]
-        cmp             %g5, 5
-        srl             %g2, 8, %g4
-        sll             %g2, 24, %g1
-        sth             %g4, [%o0 + 1]
-        add             %o1, 4, %o1
-        be              43f
-         and            %o2, 0xffffffc0, %o3
-
-        ld              [%o0 - 1], %o4
-        add             %o0, 4, %o0
-4:
-        SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
-        SMOVE_CHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
-        SMOVE_CHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
-        SMOVE_CHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
-        subcc           %o3, 64, %o3
-        add             %o1, 64, %o1
-        bne             4b
-         add            %o0, 64, %o0
-
-        andcc           %o2, 0x30, %o3
-        be,a            1f
-         srl            %g1, 24, %g2
-4:
-        SMOVE_CHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, -1)
-        subcc           %o3, 16, %o3
-        add             %o1, 16, %o1
-        bne             4b
-         add            %o0, 16, %o0
-
-        srl             %g1, 24, %g2
-1:
-        st              %o4, [%o0 - 5]
-        b               88f
-         stb            %g2, [%o0 - 1]
-41:
-        SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        subcc           %o3, 64, %o3
-        add             %o1, 64, %o1
-        bne             41b
-         add            %o0, 64, %o0
-         
-        andcc           %o2, 0x30, %o3
-        be,a            1f
-         srl            %g1, 16, %g2
-4:
-        SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 8, 24, -3)
-        subcc           %o3, 16, %o3
-        add             %o1, 16, %o1
-        bne             4b
-         add            %o0, 16, %o0
-
-        srl             %g1, 16, %g2
+        ldub            [%i1], %g2
+        stb             %g2, [%i0 + 4]
 1:
-        sth             %g2, [%o0 - 3]
-        srl             %g1, 8, %g4
-        b               88f
-         stb            %g4, [%o0 - 1]
-43:
-        SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
-        subcc           %o3, 64, %o3
-        add             %o1, 64, %o1
-        bne             43b
-         add            %o0, 64, %o0
-
-        andcc           %o2, 0x30, %o3
-        be,a            1f
-         srl            %g1, 24, %g2
-4:
-        SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 24, 8, 3)
-        subcc           %o3, 16, %o3
-        add             %o1, 16, %o1
-        bne             4b
-         add            %o0, 16, %o0
-
-        srl             %g1, 24, %g2
-1:
-        stb             %g2, [%o0 + 3]
-        b               88f
-         add            %o0, 4, %o0
-42:
-        SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x10, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x20, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        SMOVE_ALIGNCHUNK(o1, o0, 0x30, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        subcc           %o3, 64, %o3
-        add             %o1, 64, %o1
-        bne             42b
-         add            %o0, 64, %o0
-         
-        andcc           %o2, 0x30, %o3
-        be,a            1f
-         srl            %g1, 16, %g2
-4:
-        SMOVE_ALIGNCHUNK(o1, o0, 0x00, g2, g3, g4, g5, o4, o5, g7, g1, 16, 16, -2)
-        subcc           %o3, 16, %o3
-        add             %o1, 16, %o1
-        bne             4b
-         add            %o0, 16, %o0
-
-        srl             %g1, 16, %g2
-1:
-        sth             %g2, [%o0 - 2]
-
-        /* Fall through */
-         
-#endif /* FASTER_NONALIGNED */
+        ret
+         restore        %g7, %g0, %o0
 
 88:     /* short_end */
 
@@ -1127,7 +521,7 @@ FUNC(memcpy)	/* %o0=dst %o1=src %o2=len */
         stb             %g2, [%o0]
 1:
         retl
-         RETL_INSN
+         mov            %g7, %o0
 
 90:     /* short_aligned_end */
         bne             88b
diff --git a/arch/sparc/mm/Makefile b/arch/sparc/mm/Makefile
index e3cda21b5ee9..301421c11291 100644
--- a/arch/sparc/mm/Makefile
+++ b/arch/sparc/mm/Makefile
@@ -8,7 +8,6 @@ obj-$(CONFIG_SPARC64)   += ultra.o tlb.o tsb.o gup.o
 obj-y                   += fault_$(BITS).o
 obj-y                   += init_$(BITS).o
 obj-$(CONFIG_SPARC32)   += loadmmu.o
-obj-y                   += generic_$(BITS).o
 obj-$(CONFIG_SPARC32)   += extable.o btfixup.o srmmu.o iommu.o io-unit.o
 obj-$(CONFIG_SPARC32)   += hypersparc.o viking.o tsunami.o swift.o
 obj-$(CONFIG_SPARC_LEON)+= leon_mm.o
diff --git a/arch/sparc/mm/fault_32.c b/arch/sparc/mm/fault_32.c
index aa1c1b1ce5cc..8023fd7e77b5 100644
--- a/arch/sparc/mm/fault_32.c
+++ b/arch/sparc/mm/fault_32.c
@@ -20,7 +20,6 @@
 #include <linux/smp.h>
 #include <linux/perf_event.h>
 #include <linux/interrupt.h>
-#include <linux/module.h>
 #include <linux/kdebug.h>
 
 #include <asm/system.h>
diff --git a/arch/sparc/mm/generic_32.c b/arch/sparc/mm/generic_32.c
deleted file mode 100644
index e6067b75f11c..000000000000
--- a/arch/sparc/mm/generic_32.c
+++ /dev/null
@@ -1,98 +0,0 @@
-/*
- * generic.c: Generic Sparc mm routines that are not dependent upon
- *            MMU type but are Sparc specific.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#include <linux/kernel.h>
-#include <linux/mm.h>
-#include <linux/swap.h>
-#include <linux/pagemap.h>
-
-#include <asm/pgalloc.h>
-#include <asm/pgtable.h>
-#include <asm/page.h>
-#include <asm/cacheflush.h>
-#include <asm/tlbflush.h>
-
-/* Remap IO memory, the same way as remap_pfn_range(), but use
- * the obio memory space.
- *
- * They use a pgprot that sets PAGE_IO and does not check the
- * mem_map table as this is independent of normal memory.
- */
-static inline void io_remap_pte_range(struct mm_struct *mm, pte_t * pte, unsigned long address, unsigned long size,
-        unsigned long offset, pgprot_t prot, int space)
-{
-        unsigned long end;
-
-        address &= ~PMD_MASK;
-        end = address + size;
-        if (end > PMD_SIZE)
-                end = PMD_SIZE;
-        do {
-                set_pte_at(mm, address, pte, mk_pte_io(offset, prot, space));
-                address += PAGE_SIZE;
-                offset += PAGE_SIZE;
-                pte++;
-        } while (address < end);
-}
-
-static inline int io_remap_pmd_range(struct mm_struct *mm, pmd_t * pmd, unsigned long address, unsigned long size,
-        unsigned long offset, pgprot_t prot, int space)
-{
-        unsigned long end;
-
-        address &= ~PGDIR_MASK;
-        end = address + size;
-        if (end > PGDIR_SIZE)
-                end = PGDIR_SIZE;
-        offset -= address;
-        do {
-                pte_t *pte = pte_alloc_map(mm, NULL, pmd, address);
-                if (!pte)
-                        return -ENOMEM;
-                io_remap_pte_range(mm, pte, address, end - address, address + offset, prot, space);
-                address = (address + PMD_SIZE) & PMD_MASK;
-                pmd++;
-        } while (address < end);
-        return 0;
-}
-
-int io_remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
-                       unsigned long pfn, unsigned long size, pgprot_t prot)
-{
-        int error = 0;
-        pgd_t * dir;
-        unsigned long beg = from;
-        unsigned long end = from + size;
-        struct mm_struct *mm = vma->vm_mm;
-        int space = GET_IOSPACE(pfn);
-        unsigned long offset = GET_PFN(pfn) << PAGE_SHIFT;
-
-        /* See comment in mm/memory.c remap_pfn_range */
-        vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;
-        vma->vm_pgoff = (offset >> PAGE_SHIFT) |
-                ((unsigned long)space << 28UL);
-
-        offset -= from;
-        dir = pgd_offset(mm, from);
-        flush_cache_range(vma, beg, end);
-
-        while (from < end) {
-                pmd_t *pmd = pmd_alloc(mm, dir, from);
-                error = -ENOMEM;
-                if (!pmd)
-                        break;
-                error = io_remap_pmd_range(mm, pmd, from, end - from, offset + from, prot, space);
-                if (error)
-                        break;
-                from = (from + PGDIR_SIZE) & PGDIR_MASK;
-                dir++;
-        }
-
-        flush_tlb_range(vma, beg, end);
-        return error;
-}
-EXPORT_SYMBOL(io_remap_pfn_range);
diff --git a/arch/sparc/mm/generic_64.c b/arch/sparc/mm/generic_64.c
deleted file mode 100644
index 3cb00dfd4bd6..000000000000
--- a/arch/sparc/mm/generic_64.c
+++ /dev/null
@@ -1,164 +0,0 @@
-/*
- * generic.c: Generic Sparc mm routines that are not dependent upon
- *            MMU type but are Sparc specific.
- *
- * Copyright (C) 1996 David S. Miller (davem@caip.rutgers.edu)
- */
-
-#include <linux/kernel.h>
-#include <linux/mm.h>
-#include <linux/swap.h>
-#include <linux/pagemap.h>
-
-#include <asm/pgalloc.h>
-#include <asm/pgtable.h>
-#include <asm/page.h>
-#include <asm/tlbflush.h>
-
-/* Remap IO memory, the same way as remap_pfn_range(), but use
- * the obio memory space.
- *
- * They use a pgprot that sets PAGE_IO and does not check the
- * mem_map table as this is independent of normal memory.
- */
-static inline void io_remap_pte_range(struct mm_struct *mm, pte_t * pte,
-                                      unsigned long address,
-                                      unsigned long size,
-                                      unsigned long offset, pgprot_t prot,
-                                      int space)
-{
-        unsigned long end;
-
-        /* clear hack bit that was used as a write_combine side-effect flag */
-        offset &= ~0x1UL;
-        address &= ~PMD_MASK;
-        end = address + size;
-        if (end > PMD_SIZE)
-                end = PMD_SIZE;
-        do {
-                pte_t entry;
-                unsigned long curend = address + PAGE_SIZE;
-                
-                entry = mk_pte_io(offset, prot, space, PAGE_SIZE);
-                if (!(address & 0xffff)) {
-                        if (PAGE_SIZE < (4 * 1024 * 1024) &&
-                            !(address & 0x3fffff) &&
-                            !(offset & 0x3ffffe) &&
-                            end >= address + 0x400000) {
-                                entry = mk_pte_io(offset, prot, space,
-                                                  4 * 1024 * 1024);
-                                curend = address + 0x400000;
-                                offset += 0x400000;
-                        } else if (PAGE_SIZE < (512 * 1024) &&
-                                   !(address & 0x7ffff) &&
-                                   !(offset & 0x7fffe) &&
-                                   end >= address + 0x80000) {
-                                entry = mk_pte_io(offset, prot, space,
-                                                  512 * 1024 * 1024);
-                                curend = address + 0x80000;
-                                offset += 0x80000;
-                        } else if (PAGE_SIZE < (64 * 1024) &&
-                                   !(offset & 0xfffe) &&
-                                   end >= address + 0x10000) {
-                                entry = mk_pte_io(offset, prot, space,
-                                                  64 * 1024);
-                                curend = address + 0x10000;
-                                offset += 0x10000;
-                        } else
-                                offset += PAGE_SIZE;
-                } else
-                        offset += PAGE_SIZE;
-
-                if (pte_write(entry))
-                        entry = pte_mkdirty(entry);
-                do {
-                        BUG_ON(!pte_none(*pte));
-                        set_pte_at(mm, address, pte, entry);
-                        address += PAGE_SIZE;
-                        pte_val(entry) += PAGE_SIZE;
-                        pte++;
-                } while (address < curend);
-        } while (address < end);
-}
-
-static inline int io_remap_pmd_range(struct mm_struct *mm, pmd_t * pmd, unsigned long address, unsigned long size,
-        unsigned long offset, pgprot_t prot, int space)
-{
-        unsigned long end;
-
-        address &= ~PGDIR_MASK;
-        end = address + size;
-        if (end > PGDIR_SIZE)
-                end = PGDIR_SIZE;
-        offset -= address;
-        do {
-                pte_t *pte = pte_alloc_map(mm, NULL, pmd, address);
-                if (!pte)
-                        return -ENOMEM;
-                io_remap_pte_range(mm, pte, address, end - address, address + offset, prot, space);
-                pte_unmap(pte);
-                address = (address + PMD_SIZE) & PMD_MASK;
-                pmd++;
-        } while (address < end);
-        return 0;
-}
-
-static inline int io_remap_pud_range(struct mm_struct *mm, pud_t * pud, unsigned long address, unsigned long size,
-        unsigned long offset, pgprot_t prot, int space)
-{
-        unsigned long end;
-
-        address &= ~PUD_MASK;
-        end = address + size;
-        if (end > PUD_SIZE)
-                end = PUD_SIZE;
-        offset -= address;
-        do {
-                pmd_t *pmd = pmd_alloc(mm, pud, address);
-                if (!pud)
-                        return -ENOMEM;
-                io_remap_pmd_range(mm, pmd, address, end - address, address + offset, prot, space);
-                address = (address + PUD_SIZE) & PUD_MASK;
-                pud++;
-        } while (address < end);
-        return 0;
-}
-
-int io_remap_pfn_range(struct vm_area_struct *vma, unsigned long from,
-                unsigned long pfn, unsigned long size, pgprot_t prot)
-{
-        int error = 0;
-        pgd_t * dir;
-        unsigned long beg = from;
-        unsigned long end = from + size;
-        struct mm_struct *mm = vma->vm_mm;
-        int space = GET_IOSPACE(pfn);
-        unsigned long offset = GET_PFN(pfn) << PAGE_SHIFT;
-        unsigned long phys_base;
-
-        phys_base = offset | (((unsigned long) space) << 32UL);
-
-        /* See comment in mm/memory.c remap_pfn_range */
-        vma->vm_flags |= VM_IO | VM_RESERVED | VM_PFNMAP;
-        vma->vm_pgoff = phys_base >> PAGE_SHIFT;
-
-        offset -= from;
-        dir = pgd_offset(mm, from);
-        flush_cache_range(vma, beg, end);
-
-        while (from < end) {
-                pud_t *pud = pud_alloc(mm, dir, from);
-                error = -ENOMEM;
-                if (!pud)
-                        break;
-                error = io_remap_pud_range(mm, pud, from, end - from, offset + from, prot, space);
-                if (error)
-                        break;
-                from = (from + PGDIR_SIZE) & PGDIR_MASK;
-                dir++;
-        }
-
-        flush_tlb_range(vma, beg, end);
-        return error;
-}
-EXPORT_SYMBOL(io_remap_pfn_range);
diff --git a/arch/sparc/mm/gup.c b/arch/sparc/mm/gup.c
index a986b5d05712..42c55df3aec3 100644
--- a/arch/sparc/mm/gup.c
+++ b/arch/sparc/mm/gup.c
@@ -56,6 +56,8 @@ static noinline int gup_pte_range(pmd_t pmd, unsigned long addr,
                         put_page(head);
                         return 0;
                 }
+                if (head != page)
+                        get_huge_page_tail(page);
 
                 pages[*nr] = page;
                 (*nr)++;
diff --git a/arch/sparc/mm/highmem.c b/arch/sparc/mm/highmem.c
index 4730eac0747b..77140a02c86a 100644
--- a/arch/sparc/mm/highmem.c
+++ b/arch/sparc/mm/highmem.c
@@ -24,6 +24,7 @@
  */
 #include <linux/mm.h>
 #include <linux/highmem.h>
+#include <linux/export.h>
 #include <asm/pgalloc.h>
 #include <asm/cacheflush.h>
 #include <asm/tlbflush.h>
diff --git a/arch/sparc/mm/hugetlbpage.c b/arch/sparc/mm/hugetlbpage.c
index f4e97646ce23..07e14535375c 100644
--- a/arch/sparc/mm/hugetlbpage.c
+++ b/arch/sparc/mm/hugetlbpage.c
@@ -5,7 +5,6 @@
  */
 
 #include <linux/init.h>
-#include <linux/module.h>
 #include <linux/fs.h>
 #include <linux/mm.h>
 #include <linux/hugetlb.h>
diff --git a/arch/sparc/mm/init_64.c b/arch/sparc/mm/init_64.c
index 581531dbc8b5..8e073d802139 100644
--- a/arch/sparc/mm/init_64.c
+++ b/arch/sparc/mm/init_64.c
@@ -511,6 +511,11 @@ static void __init read_obp_translations(void)
                 for (i = 0; i < prom_trans_ents; i++)
                         prom_trans[i].data &= ~0x0003fe0000000000UL;
         }
+
+        /* Force execute bit on.  */
+        for (i = 0; i < prom_trans_ents; i++)
+                prom_trans[i].data |= (tlb_type == hypervisor ?
+                                       _PAGE_EXEC_4V : _PAGE_EXEC_4U);
 }
 
 static void __init hypervisor_tlb_lock(unsigned long vaddr,
diff --git a/arch/sparc/mm/leon_mm.c b/arch/sparc/mm/leon_mm.c
index e485a6804998..13c2169822a8 100644
--- a/arch/sparc/mm/leon_mm.c
+++ b/arch/sparc/mm/leon_mm.c
@@ -162,7 +162,7 @@ ready:
                 printk(KERN_INFO "swprobe: padde %x\n", paddr_calc);
         if (paddr)
                 *paddr = paddr_calc;
-        return paddrbase;
+        return pte;
 }
 
 void leon_flush_icache_all(void)