22 files changed, 251 insertions, 169 deletions

diff --git a/arch/sparc64/kernel/Makefile b/arch/sparc64/kernel/Makefile
index e1eabebaed39..eff0c01d3579 100644
--- a/arch/sparc64/kernel/Makefile
+++ b/arch/sparc64/kernel/Makefile
@@ -14,6 +14,7 @@ obj-y		:= process.o setup.o cpu.o idprom.o \
                    power.o sbus.o iommu_common.o sparc64_ksyms.o chmc.o \
                    visemul.o prom.o of_device.o
 
+obj-$(CONFIG_STACKTRACE) += stacktrace.o
 obj-$(CONFIG_PCI)        += ebus.o isa.o pci_common.o pci_iommu.o \
                             pci_psycho.o pci_sabre.o pci_schizo.o \
                             pci_sun4v.o pci_sun4v_asm.o
diff --git a/arch/sparc64/kernel/binfmt_aout32.c b/arch/sparc64/kernel/binfmt_aout32.c
index d7caa60a0074..f205fc7cbcd0 100644
--- a/arch/sparc64/kernel/binfmt_aout32.c
+++ b/arch/sparc64/kernel/binfmt_aout32.c
@@ -209,7 +209,7 @@ static int load_aout32_binary(struct linux_binprm * bprm, struct pt_regs * regs)
         if ((N_MAGIC(ex) != ZMAGIC && N_MAGIC(ex) != OMAGIC &&
              N_MAGIC(ex) != QMAGIC && N_MAGIC(ex) != NMAGIC) ||
             N_TRSIZE(ex) || N_DRSIZE(ex) ||
-            bprm->file->f_dentry->d_inode->i_size < ex.a_text+ex.a_data+N_SYMSIZE(ex)+N_TXTOFF(ex)) {
+            bprm->file->f_path.dentry->d_inode->i_size < ex.a_text+ex.a_data+N_SYMSIZE(ex)+N_TXTOFF(ex)) {
                 return -ENOEXEC;
         }
 
@@ -349,7 +349,7 @@ static int load_aout32_library(struct file *file)
         int retval;
         struct exec ex;
 
-        inode = file->f_dentry->d_inode;
+        inode = file->f_path.dentry->d_inode;
 
         retval = -ENOEXEC;
         error = kernel_read(file, 0, (char *) &ex, sizeof(ex));
diff --git a/arch/sparc64/kernel/binfmt_elf32.c b/arch/sparc64/kernel/binfmt_elf32.c
index a98f3ae175a3..9ad84ff10a17 100644
--- a/arch/sparc64/kernel/binfmt_elf32.c
+++ b/arch/sparc64/kernel/binfmt_elf32.c
@@ -141,7 +141,6 @@ cputime_to_compat_timeval(const cputime_t cputime, struct compat_timeval *value)
         value->tv_sec = jiffies / HZ;
 }
 
-#define elf_addr_t      u32
 #undef start_thread
 #define start_thread start_thread32
 #define init_elf_binfmt init_elf32_binfmt
diff --git a/arch/sparc64/kernel/chmc.c b/arch/sparc64/kernel/chmc.c
index 259f37e516f5..9699abeb9907 100644
--- a/arch/sparc64/kernel/chmc.c
+++ b/arch/sparc64/kernel/chmc.c
@@ -341,7 +341,7 @@ static void fetch_decode_regs(struct mctrl_info *mp)
 
 static int init_one_mctrl(struct device_node *dp)
 {
-        struct mctrl_info *mp = kmalloc(sizeof(*mp), GFP_KERNEL);
+        struct mctrl_info *mp = kzalloc(sizeof(*mp), GFP_KERNEL);
         int portid = of_getintprop_default(dp, "portid", -1);
         struct linux_prom64_registers *regs;
         void *pval;
@@ -349,7 +349,6 @@ static int init_one_mctrl(struct device_node *dp)
 
         if (!mp)
                 return -1;
-        memset(mp, 0, sizeof(*mp));
         if (portid == -1)
                 goto fail;
 
diff --git a/arch/sparc64/kernel/entry.S b/arch/sparc64/kernel/entry.S
index 6f28bec0a9bf..c15a3edcb826 100644
--- a/arch/sparc64/kernel/entry.S
+++ b/arch/sparc64/kernel/entry.S
@@ -597,7 +597,12 @@ __spitfire_cee_trap_continue:
 1:      ba,pt           %xcc, etrap_irq
          rd             %pc, %g7
 
-2:      mov             %l4, %o1
+2:
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call    trace_hardirqs_off
+         nop
+#endif
+        mov             %l4, %o1
         mov             %l5, %o2
         call            spitfire_access_error
          add            %sp, PTREGS_OFF, %o0
@@ -824,6 +829,10 @@ do_cheetah_plus_data_parity:
         wrpr            %g0, 15, %pil
         ba,pt           %xcc, etrap_irq
          rd             %pc, %g7
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         mov             0x0, %o0
         call            cheetah_plus_parity_error
          add            %sp, PTREGS_OFF, %o1
@@ -855,6 +864,10 @@ do_cheetah_plus_insn_parity:
         wrpr            %g0, 15, %pil
         ba,pt           %xcc, etrap_irq
          rd             %pc, %g7
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         mov             0x1, %o0
         call            cheetah_plus_parity_error
          add            %sp, PTREGS_OFF, %o1
@@ -1183,6 +1196,10 @@ c_fast_ecc:
         wrpr            %g0, 15, %pil
         ba,pt           %xcc, etrap_irq
          rd             %pc, %g7
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         mov             %l4, %o1
         mov             %l5, %o2
         call            cheetah_fecc_handler
@@ -1211,6 +1228,10 @@ c_cee:
         wrpr            %g0, 15, %pil
         ba,pt           %xcc, etrap_irq
          rd             %pc, %g7
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         mov             %l4, %o1
         mov             %l5, %o2
         call            cheetah_cee_handler
@@ -1239,6 +1260,10 @@ c_deferred:
         wrpr            %g0, 15, %pil
         ba,pt           %xcc, etrap_irq
          rd             %pc, %g7
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         mov             %l4, %o1
         mov             %l5, %o2
         call            cheetah_deferred_handler
diff --git a/arch/sparc64/kernel/head.S b/arch/sparc64/kernel/head.S
index c8e9dc9d68a9..baea10a98196 100644
--- a/arch/sparc64/kernel/head.S
+++ b/arch/sparc64/kernel/head.S
@@ -78,11 +78,7 @@ sparc_ramdisk_image64:
 
         /* PROM cif handler code address is in %o4.  */
 sparc64_boot:
-1:      rd      %pc, %g7
-        set     1b, %g1
-        cmp     %g1, %g7
-        be,pn   %xcc, sparc64_boot_after_remap
-         mov    %o4, %l7
+        mov     %o4, %l7
 
         /* We need to remap the kernel.  Use position independant
          * code to remap us to KERNBASE.
@@ -295,7 +291,6 @@ is_sun4v:
 
         add     %sp, (192 + 128), %sp
 
-sparc64_boot_after_remap:
         sethi   %hi(prom_root_compatible), %g1
         or      %g1, %lo(prom_root_compatible), %g1
         sethi   %hi(prom_sun4v_name), %g7
@@ -489,6 +484,14 @@ tlb_fixup_done:
         call    __bzero
          sub    %o1, %o0, %o1
 
+#ifdef CONFIG_LOCKDEP
+        /* We have this call this super early, as even prom_init can grab
+         * spinlocks and thus call into the lockdep code.
+         */
+        call    lockdep_init
+         nop
+#endif
+
         mov     %l6, %o1                        ! OpenPROM stack
         call    prom_init
          mov    %l7, %o0                        ! OpenPROM cif handler
diff --git a/arch/sparc64/kernel/irq.c b/arch/sparc64/kernel/irq.c
index d64b1ea848de..c3d068c7a412 100644
--- a/arch/sparc64/kernel/irq.c
+++ b/arch/sparc64/kernel/irq.c
@@ -372,14 +372,14 @@ static void run_pre_handler(unsigned int virt_irq)
         }
 }
 
-static struct hw_interrupt_type sun4u_irq = {
+static struct irq_chip sun4u_irq = {
         .typename       = "sun4u",
         .enable         = sun4u_irq_enable,
         .disable        = sun4u_irq_disable,
         .end            = sun4u_irq_end,
 };
 
-static struct hw_interrupt_type sun4u_irq_ack = {
+static struct irq_chip sun4u_irq_ack = {
         .typename       = "sun4u+ack",
         .enable         = sun4u_irq_enable,
         .disable        = sun4u_irq_disable,
@@ -387,14 +387,14 @@ static struct hw_interrupt_type sun4u_irq_ack = {
         .end            = sun4u_irq_end,
 };
 
-static struct hw_interrupt_type sun4v_irq = {
+static struct irq_chip sun4v_irq = {
         .typename       = "sun4v",
         .enable         = sun4v_irq_enable,
         .disable        = sun4v_irq_disable,
         .end            = sun4v_irq_end,
 };
 
-static struct hw_interrupt_type sun4v_irq_ack = {
+static struct irq_chip sun4v_irq_ack = {
         .typename       = "sun4v+ack",
         .enable         = sun4v_irq_enable,
         .disable        = sun4v_irq_disable,
@@ -493,22 +493,6 @@ out:
         return bucket->virt_irq;
 }
 
-void hw_resend_irq(struct hw_interrupt_type *handler, unsigned int virt_irq)
-{
-        struct ino_bucket *bucket = virt_irq_to_bucket(virt_irq);
-        unsigned long pstate;
-        unsigned int *ent;
-
-        __asm__ __volatile__("rdpr %%pstate, %0" : "=r" (pstate));
-        __asm__ __volatile__("wrpr %0, %1, %%pstate"
-                             : : "r" (pstate), "i" (PSTATE_IE));
-        ent = irq_work(smp_processor_id());
-        bucket->irq_chain = *ent;
-        *ent = __irq(bucket);
-        set_softint(1 << PIL_DEVICE_IRQ);
-        __asm__ __volatile__("wrpr %0, 0x0, %%pstate" : : "r" (pstate));
-}
-
 void ack_bad_irq(unsigned int virt_irq)
 {
         struct ino_bucket *bucket = virt_irq_to_bucket(virt_irq);
diff --git a/arch/sparc64/kernel/isa.c b/arch/sparc64/kernel/isa.c
index f028e68b23f2..98721a8f8619 100644
--- a/arch/sparc64/kernel/isa.c
+++ b/arch/sparc64/kernel/isa.c
@@ -22,14 +22,15 @@ static void __init report_dev(struct sparc_isa_device *isa_dev, int child)
                 printk(" [%s", isa_dev->prom_node->name);
 }
 
-static struct linux_prom_registers * __init
-isa_dev_get_resource(struct sparc_isa_device *isa_dev)
+static void __init isa_dev_get_resource(struct sparc_isa_device *isa_dev)
 {
         struct linux_prom_registers *pregs;
         unsigned long base, len;
         int prop_len;
 
         pregs = of_get_property(isa_dev->prom_node, "reg", &prop_len);
+        if (!pregs)
+                return;
 
         /* Only the first one is interesting. */
         len = pregs[0].reg_size;
@@ -44,12 +45,9 @@ isa_dev_get_resource(struct sparc_isa_device *isa_dev)
 
         request_resource(&isa_dev->bus->parent->io_space,
                          &isa_dev->resource);
-
-        return pregs;
 }
 
-static void __init isa_dev_get_irq(struct sparc_isa_device *isa_dev,
-                                   struct linux_prom_registers *pregs)
+static void __init isa_dev_get_irq(struct sparc_isa_device *isa_dev)
 {
         struct of_device *op = of_find_device_by_node(isa_dev->prom_node);
 
@@ -69,17 +67,14 @@ static void __init isa_fill_children(struct sparc_isa_device *parent_isa_dev)
 
         printk(" ->");
         while (dp) {
-                struct linux_prom_registers *regs;
                 struct sparc_isa_device *isa_dev;
 
-                isa_dev = kmalloc(sizeof(*isa_dev), GFP_KERNEL);
+                isa_dev = kzalloc(sizeof(*isa_dev), GFP_KERNEL);
                 if (!isa_dev) {
                         fatal_err("cannot allocate child isa_dev");
                         prom_halt();
                 }
 
-                memset(isa_dev, 0, sizeof(*isa_dev));
-
                 /* Link it in to parent. */
                 isa_dev->next = parent_isa_dev->child;
                 parent_isa_dev->child = isa_dev;
@@ -87,8 +82,8 @@ static void __init isa_fill_children(struct sparc_isa_device *parent_isa_dev)
                 isa_dev->bus = parent_isa_dev->bus;
                 isa_dev->prom_node = dp;
 
-                regs = isa_dev_get_resource(isa_dev);
-                isa_dev_get_irq(isa_dev, regs);
+                isa_dev_get_resource(isa_dev);
+                isa_dev_get_irq(isa_dev);
 
                 report_dev(isa_dev, 1);
 
@@ -101,17 +96,14 @@ static void __init isa_fill_devices(struct sparc_isa_bridge *isa_br)
         struct device_node *dp = isa_br->prom_node->child;
 
         while (dp) {
-                struct linux_prom_registers *regs;
                 struct sparc_isa_device *isa_dev;
 
-                isa_dev = kmalloc(sizeof(*isa_dev), GFP_KERNEL);
+                isa_dev = kzalloc(sizeof(*isa_dev), GFP_KERNEL);
                 if (!isa_dev) {
                         printk(KERN_DEBUG "ISA: cannot allocate isa_dev");
                         return;
                 }
 
-                memset(isa_dev, 0, sizeof(*isa_dev));
-
                 isa_dev->ofdev.node = dp;
                 isa_dev->ofdev.dev.parent = &isa_br->ofdev.dev;
                 isa_dev->ofdev.dev.bus = &isa_bus_type;
@@ -141,8 +133,8 @@ static void __init isa_fill_devices(struct sparc_isa_bridge *isa_br)
                 isa_dev->bus = isa_br;
                 isa_dev->prom_node = dp;
 
-                regs = isa_dev_get_resource(isa_dev);
-                isa_dev_get_irq(isa_dev, regs);
+                isa_dev_get_resource(isa_dev);
+                isa_dev_get_irq(isa_dev);
 
                 report_dev(isa_dev, 0);
 
@@ -180,14 +172,12 @@ void __init isa_init(void)
                 pbm = pdev_cookie->pbm;
                 dp = pdev_cookie->prom_node;
 
-                isa_br = kmalloc(sizeof(*isa_br), GFP_KERNEL);
+                isa_br = kzalloc(sizeof(*isa_br), GFP_KERNEL);
                 if (!isa_br) {
                         printk(KERN_DEBUG "isa: cannot allocate sparc_isa_bridge");
                         return;
                 }
 
-                memset(isa_br, 0, sizeof(*isa_br));
-
                 isa_br->ofdev.node = dp;
                 isa_br->ofdev.dev.parent = &pdev->dev;
                 isa_br->ofdev.dev.bus = &isa_bus_type;
diff --git a/arch/sparc64/kernel/kprobes.c b/arch/sparc64/kernel/kprobes.c
index 8e75ed762fd8..ae221f0d4a6f 100644
--- a/arch/sparc64/kernel/kprobes.c
+++ b/arch/sparc64/kernel/kprobes.c
@@ -45,7 +45,11 @@ DEFINE_PER_CPU(struct kprobe_ctlblk, kprobe_ctlblk);
 int __kprobes arch_prepare_kprobe(struct kprobe *p)
 {
         p->ainsn.insn[0] = *p->addr;
+        flushi(&p->ainsn.insn[0]);
+
         p->ainsn.insn[1] = BREAKPOINT_INSTRUCTION_2;
+        flushi(&p->ainsn.insn[1]);
+
         p->opcode = *p->addr;
         return 0;
 }
@@ -185,16 +189,19 @@ no_kprobe:
 /* If INSN is a relative control transfer instruction,
  * return the corrected branch destination value.
  *
- * The original INSN location was REAL_PC, it actually
- * executed at PC and produced destination address NPC.
+ * regs->tpc and regs->tnpc still hold the values of the
+ * program counters at the time of trap due to the execution
+ * of the BREAKPOINT_INSTRUCTION_2 at p->ainsn.insn[1]
+ * 
  */
-static unsigned long __kprobes relbranch_fixup(u32 insn, unsigned long real_pc,
-                                               unsigned long pc,
-                                               unsigned long npc)
+static unsigned long __kprobes relbranch_fixup(u32 insn, struct kprobe *p,
+                                               struct pt_regs *regs)
 {
+        unsigned long real_pc = (unsigned long) p->addr;
+
         /* Branch not taken, no mods necessary.  */
-        if (npc == pc + 0x4UL)
-                return real_pc + 0x4UL;
+        if (regs->tnpc == regs->tpc + 0x4UL)
+                return real_pc + 0x8UL;
 
         /* The three cases are call, branch w/prediction,
          * and traditional branch.
@@ -202,14 +209,21 @@ static unsigned long __kprobes relbranch_fixup(u32 insn, unsigned long real_pc,
         if ((insn & 0xc0000000) == 0x40000000 ||
             (insn & 0xc1c00000) == 0x00400000 ||
             (insn & 0xc1c00000) == 0x00800000) {
+                unsigned long ainsn_addr;
+
+                ainsn_addr = (unsigned long) &p->ainsn.insn[0];
+
                 /* The instruction did all the work for us
                  * already, just apply the offset to the correct
                  * instruction location.
                  */
-                return (real_pc + (npc - pc));
+                return (real_pc + (regs->tnpc - ainsn_addr));
         }
 
-        return real_pc + 0x4UL;
+        /* It is jmpl or some other absolute PC modification instruction,
+         * leave NPC as-is.
+         */
+        return regs->tnpc;
 }
 
 /* If INSN is an instruction which writes it's PC location
@@ -220,12 +234,12 @@ static void __kprobes retpc_fixup(struct pt_regs *regs, u32 insn,
 {
         unsigned long *slot = NULL;
 
-        /* Simplest cast is call, which always uses %o7 */
+        /* Simplest case is 'call', which always uses %o7 */
         if ((insn & 0xc0000000) == 0x40000000) {
                 slot = &regs->u_regs[UREG_I7];
         }
 
-        /* Jmpl encodes the register inside of the opcode */
+        /* 'jmpl' encodes the register inside of the opcode */
         if ((insn & 0xc1f80000) == 0x81c00000) {
                 unsigned long rd = ((insn >> 25) & 0x1f);
 
@@ -247,11 +261,11 @@ static void __kprobes retpc_fixup(struct pt_regs *regs, u32 insn,
 
 /*
  * Called after single-stepping.  p->addr is the address of the
- * instruction whose first byte has been replaced by the breakpoint
+ * instruction which has been replaced by the breakpoint
  * instruction.  To avoid the SMP problems that can occur when we
  * temporarily put back the original opcode to single-step, we
  * single-stepped a copy of the instruction.  The address of this
- * copy is p->ainsn.insn.
+ * copy is &p->ainsn.insn[0].
  *
  * This function prepares to return from the post-single-step
  * breakpoint trap.
@@ -261,11 +275,11 @@ static void __kprobes resume_execution(struct kprobe *p,
 {
         u32 insn = p->ainsn.insn[0];
 
+        regs->tnpc = relbranch_fixup(insn, p, regs);
+
+        /* This assignment must occur after relbranch_fixup() */
         regs->tpc = kcb->kprobe_orig_tnpc;
-        regs->tnpc = relbranch_fixup(insn,
-                                     (unsigned long) p->addr,
-                                     (unsigned long) &p->ainsn.insn[0],
-                                     regs->tnpc);
+
         retpc_fixup(regs, insn, (unsigned long) p->addr);
 
         regs->tstate = ((regs->tstate & ~TSTATE_PIL) |
@@ -430,17 +444,8 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
         struct jprobe *jp = container_of(p, struct jprobe, kp);
         struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
 
-        kcb->jprobe_saved_regs_location = regs;
         memcpy(&(kcb->jprobe_saved_regs), regs, sizeof(*regs));
 
-        /* Save a whole stack frame, this gets arguments
-         * pushed onto the stack after using up all the
-         * arg registers.
-         */
-        memcpy(&(kcb->jprobe_saved_stack),
-               (char *) (regs->u_regs[UREG_FP] + STACK_BIAS),
-               sizeof(kcb->jprobe_saved_stack));
-
         regs->tpc  = (unsigned long) jp->entry;
         regs->tnpc = ((unsigned long) jp->entry) + 0x4UL;
         regs->tstate |= TSTATE_PIL;
@@ -450,10 +455,19 @@ int __kprobes setjmp_pre_handler(struct kprobe *p, struct pt_regs *regs)
 
 void __kprobes jprobe_return(void)
 {
-        __asm__ __volatile__(
-                ".globl jprobe_return_trap_instruction\n"
+        struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
+        register unsigned long orig_fp asm("g1");
+
+        orig_fp = kcb->jprobe_saved_regs.u_regs[UREG_FP];
+        __asm__ __volatile__("\n"
+"1:     cmp             %%sp, %0\n\t"
+        "blu,a,pt       %%xcc, 1b\n\t"
+        " restore\n\t"
+        ".globl         jprobe_return_trap_instruction\n"
 "jprobe_return_trap_instruction:\n\t"
-                "ta 0x70");
+        "ta             0x70"
+        : /* no outputs */
+        : "r" (orig_fp));
 }
 
 extern void jprobe_return_trap_instruction(void);
@@ -466,26 +480,7 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
         struct kprobe_ctlblk *kcb = get_kprobe_ctlblk();
 
         if (addr == (u32 *) jprobe_return_trap_instruction) {
-                if (kcb->jprobe_saved_regs_location != regs) {
-                        printk("JPROBE: Current regs (%p) does not match "
-                               "saved regs (%p).\n",
-                               regs, kcb->jprobe_saved_regs_location);
-                        printk("JPROBE: Saved registers\n");
-                        __show_regs(kcb->jprobe_saved_regs_location);
-                        printk("JPROBE: Current registers\n");
-                        __show_regs(regs);
-                        BUG();
-                }
-                /* Restore old register state.  Do pt_regs
-                 * first so that UREG_FP is the original one for
-                 * the stack frame restore.
-                 */
                 memcpy(regs, &(kcb->jprobe_saved_regs), sizeof(*regs));
-
-                memcpy((char *) (regs->u_regs[UREG_FP] + STACK_BIAS),
-                       &(kcb->jprobe_saved_stack),
-                       sizeof(kcb->jprobe_saved_stack));
-
                 preempt_enable_no_resched();
                 return 1;
         }
diff --git a/arch/sparc64/kernel/of_device.c b/arch/sparc64/kernel/of_device.c
index 8cc14fc6b6f1..b0f3e0082a0d 100644
--- a/arch/sparc64/kernel/of_device.c
+++ b/arch/sparc64/kernel/of_device.c
@@ -144,9 +144,12 @@ void __iomem *of_ioremap(struct resource *res, unsigned long offset, unsigned lo
 }
 EXPORT_SYMBOL(of_ioremap);
 
-void of_iounmap(void __iomem *base, unsigned long size)
+void of_iounmap(struct resource *res, void __iomem *base, unsigned long size)
 {
-        release_region((unsigned long) base, size);
+        if (res->flags & IORESOURCE_MEM)
+                release_mem_region((unsigned long) base, size);
+        else
+                release_region((unsigned long) base, size);
 }
 EXPORT_SYMBOL(of_iounmap);
 
@@ -1007,10 +1010,9 @@ struct of_device* of_platform_device_create(struct device_node *np,
 {
         struct of_device *dev;
 
-        dev = kmalloc(sizeof(*dev), GFP_KERNEL);
+        dev = kzalloc(sizeof(*dev), GFP_KERNEL);
         if (!dev)
                 return NULL;
-        memset(dev, 0, sizeof(*dev));
 
         dev->dev.parent = parent;
         dev->dev.bus = bus;
diff --git a/arch/sparc64/kernel/pci.c b/arch/sparc64/kernel/pci.c
index e02f01b644af..dfc41cd4bb5d 100644
--- a/arch/sparc64/kernel/pci.c
+++ b/arch/sparc64/kernel/pci.c
@@ -646,13 +646,4 @@ int pci_domain_nr(struct pci_bus *pbus)
 }
 EXPORT_SYMBOL(pci_domain_nr);
 
-int pcibios_prep_mwi(struct pci_dev *dev)
-{
-        /* We set correct PCI_CACHE_LINE_SIZE register values for every
-         * device probed on this platform.  So there is nothing to check
-         * and this always succeeds.
-         */
-        return 0;
-}
-
 #endif /* !(CONFIG_PCI) */
diff --git a/arch/sparc64/kernel/pci_sun4v.c b/arch/sparc64/kernel/pci_sun4v.c
index 03ad4c06758e..6b04794b7a97 100644
--- a/arch/sparc64/kernel/pci_sun4v.c
+++ b/arch/sparc64/kernel/pci_sun4v.c
@@ -798,7 +798,7 @@ static struct pci_ops pci_sun4v_ops = {
 static void pbm_scan_bus(struct pci_controller_info *p,
                          struct pci_pbm_info *pbm)
 {
-        struct pcidev_cookie *cookie = kmalloc(sizeof(*cookie), GFP_KERNEL);
+        struct pcidev_cookie *cookie = kzalloc(sizeof(*cookie), GFP_KERNEL);
 
         if (!cookie) {
                 prom_printf("%s: Critical allocation failure.\n", pbm->name);
@@ -806,7 +806,6 @@ static void pbm_scan_bus(struct pci_controller_info *p,
         }
 
         /* All we care about is the PBM. */
-        memset(cookie, 0, sizeof(*cookie));
         cookie->pbm = pbm;
 
         pbm->pci_bus = pci_scan_bus(pbm->pci_first_busno, p->pci_ops, pbm);
@@ -1048,12 +1047,11 @@ static void pci_sun4v_iommu_init(struct pci_pbm_info *pbm)
         /* Allocate and initialize the free area map.  */
         sz = num_tsb_entries / 8;
         sz = (sz + 7UL) & ~7UL;
-        iommu->arena.map = kmalloc(sz, GFP_KERNEL);
+        iommu->arena.map = kzalloc(sz, GFP_KERNEL);
         if (!iommu->arena.map) {
                 prom_printf("PCI_IOMMU: Error, kmalloc(arena.map) failed.\n");
                 prom_halt();
         }
-        memset(iommu->arena.map, 0, sz);
         iommu->arena.limit = num_tsb_entries;
 
         sz = probe_existing_entries(pbm, iommu);
@@ -1164,24 +1162,20 @@ void sun4v_pci_init(struct device_node *dp, char *model_name)
                 per_cpu(pci_iommu_batch, i).pglist = (u64 *) page;
         }
 
-        p = kmalloc(sizeof(struct pci_controller_info), GFP_ATOMIC);
+        p = kzalloc(sizeof(struct pci_controller_info), GFP_ATOMIC);
         if (!p)
                 goto fatal_memory_error;
 
-        memset(p, 0, sizeof(*p));
-
-        iommu = kmalloc(sizeof(struct pci_iommu), GFP_ATOMIC);
+        iommu = kzalloc(sizeof(struct pci_iommu), GFP_ATOMIC);
         if (!iommu)
                 goto fatal_memory_error;
 
-        memset(iommu, 0, sizeof(*iommu));
         p->pbm_A.iommu = iommu;
 
-        iommu = kmalloc(sizeof(struct pci_iommu), GFP_ATOMIC);
+        iommu = kzalloc(sizeof(struct pci_iommu), GFP_ATOMIC);
         if (!iommu)
                 goto fatal_memory_error;
 
-        memset(iommu, 0, sizeof(*iommu));
         p->pbm_B.iommu = iommu;
 
         p->next = pci_controller_root;
diff --git a/arch/sparc64/kernel/ptrace.c b/arch/sparc64/kernel/ptrace.c
index d31975e6d6f6..81111a12f0a8 100644
--- a/arch/sparc64/kernel/ptrace.c
+++ b/arch/sparc64/kernel/ptrace.c
@@ -202,7 +202,10 @@ asmlinkage void do_ptrace(struct pt_regs *regs)
 #endif
         if (request == PTRACE_TRACEME) {
                 ret = ptrace_traceme();
-                pt_succ_return(regs, 0);
+                if (ret < 0)
+                        pt_error_return(regs, -ret);
+                else
+                        pt_succ_return(regs, 0);
                 goto out;
         }
 
diff --git a/arch/sparc64/kernel/rtrap.S b/arch/sparc64/kernel/rtrap.S
index 3522cd66f3bb..079d18a11d24 100644
--- a/arch/sparc64/kernel/rtrap.S
+++ b/arch/sparc64/kernel/rtrap.S
@@ -165,14 +165,26 @@ rtrap:
 __handle_softirq_continue:
 rtrap_xcall:
                 sethi                   %hi(0xf << 20), %l4
-                andcc                   %l1, TSTATE_PRIV, %l3
                 and                     %l1, %l4, %l4
+                andn                    %l1, %l4, %l1
+                srl                     %l4, 20, %l4
+#ifdef CONFIG_TRACE_IRQFLAGS
+                brnz,pn                 %l4, rtrap_no_irq_enable
+                 nop
+                call                    trace_hardirqs_on
+                 nop
+                wrpr                    %l4, %pil
+rtrap_no_irq_enable:
+#endif
+                andcc                   %l1, TSTATE_PRIV, %l3
                 bne,pn                  %icc, to_kernel
-                 andn                   %l1, %l4, %l1
+                 nop
 
                 /* We must hold IRQs off and atomically test schedule+signal
                  * state, then hold them off all the way back to userspace.
-                 * If we are returning to kernel, none of this matters.
+                 * If we are returning to kernel, none of this matters.  Note
+                 * that we are disabling interrupts via PSTATE_IE, not using
+                 * %pil.
                  *
                  * If we do not do this, there is a window where we would do
                  * the tests, later the signal/resched event arrives but we do
@@ -256,7 +268,6 @@ rt_continue:	ldx			[%sp + PTREGS_OFF + PT_V9_G1], %g1
 
                 ld                      [%sp + PTREGS_OFF + PT_V9_Y], %o3
                 wr                      %o3, %g0, %y
-                srl                     %l4, 20, %l4
                 wrpr                    %l4, 0x0, %pil
                 wrpr                    %g0, 0x1, %tl
                 wrpr                    %l1, %g0, %tstate
@@ -374,8 +385,8 @@ to_kernel:
                  ldx                    [%g6 + TI_FLAGS], %l5
                 andcc                   %l5, _TIF_NEED_RESCHED, %g0
                 be,pt                   %xcc, kern_fpucheck
-                 srl                    %l4, 20, %l5
-                cmp                     %l5, 0
+                 nop
+                cmp                     %l4, 0
                 bne,pn                  %xcc, kern_fpucheck
                  sethi                  %hi(PREEMPT_ACTIVE), %l6
                 stw                     %l6, [%g6 + TI_PRE_COUNT]
diff --git a/arch/sparc64/kernel/smp.c b/arch/sparc64/kernel/smp.c
index cc09d8266414..fc99f7b8012f 100644
--- a/arch/sparc64/kernel/smp.c
+++ b/arch/sparc64/kernel/smp.c
@@ -1388,7 +1388,7 @@ void __devinit smp_prepare_boot_cpu(void)
 {
 }
 
-int __devinit __cpu_up(unsigned int cpu)
+int __cpuinit __cpu_up(unsigned int cpu)
 {
         int ret = smp_boot_one_cpu(cpu);
 
@@ -1447,11 +1447,8 @@ void __init setup_per_cpu_areas(void)
         char *ptr;
 
         /* Copy section for each CPU (we discard the original) */
-        goal = ALIGN(__per_cpu_end - __per_cpu_start, SMP_CACHE_BYTES);
-#ifdef CONFIG_MODULES
-        if (goal < PERCPU_ENOUGH_ROOM)
-                goal = PERCPU_ENOUGH_ROOM;
-#endif
+        goal = PERCPU_ENOUGH_ROOM;
+
         __per_cpu_shift = 0;
         for (size = 1UL; size < goal; size <<= 1UL)
                 __per_cpu_shift++;
diff --git a/arch/sparc64/kernel/stacktrace.c b/arch/sparc64/kernel/stacktrace.c
new file mode 100644
index 000000000000..c4d15f2762b9
--- /dev/null
+++ b/arch/sparc64/kernel/stacktrace.c
@@ -0,0 +1,41 @@
+#include <linux/sched.h>
+#include <linux/stacktrace.h>
+#include <linux/thread_info.h>
+#include <asm/ptrace.h>
+
+void save_stack_trace(struct stack_trace *trace, struct task_struct *task)
+{
+        unsigned long ksp, fp, thread_base;
+        struct thread_info *tp;
+
+        if (!task)
+                task = current;
+        tp = task_thread_info(task);
+        if (task == current) {
+                flushw_all();
+                __asm__ __volatile__(
+                        "mov    %%fp, %0"
+                        : "=r" (ksp)
+                );
+        } else
+                ksp = tp->ksp;
+
+        fp = ksp + STACK_BIAS;
+        thread_base = (unsigned long) tp;
+        do {
+                struct reg_window *rw;
+
+                /* Bogus frame pointer? */
+                if (fp < (thread_base + sizeof(struct thread_info)) ||
+                    fp >= (thread_base + THREAD_SIZE))
+                        break;
+
+                rw = (struct reg_window *) fp;
+                if (trace->skip > 0)
+                        trace->skip--;
+                else
+                        trace->entries[trace->nr_entries++] = rw->ins[7];
+
+                fp = rw->ins[6] + STACK_BIAS;
+        } while (trace->nr_entries < trace->max_entries);
+}
diff --git a/arch/sparc64/kernel/sun4v_ivec.S b/arch/sparc64/kernel/sun4v_ivec.S
index 49703c3c5769..405855dd886b 100644
--- a/arch/sparc64/kernel/sun4v_ivec.S
+++ b/arch/sparc64/kernel/sun4v_ivec.S
@@ -190,7 +190,10 @@ sun4v_res_mondo:
         mov     %g1, %g4
         ba,pt   %xcc, etrap_irq
          rd     %pc, %g7
-
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         /* Log the event.  */
         add     %sp, PTREGS_OFF, %o0
         call    sun4v_resum_error
@@ -216,7 +219,10 @@ sun4v_res_mondo_queue_full:
         wrpr    %g0, 15, %pil
         ba,pt   %xcc, etrap_irq
          rd     %pc, %g7
-
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         call    sun4v_resum_overflow
          add    %sp, PTREGS_OFF, %o0
 
@@ -295,7 +301,10 @@ sun4v_nonres_mondo:
         mov     %g1, %g4
         ba,pt   %xcc, etrap_irq
          rd     %pc, %g7
-
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         /* Log the event.  */
         add     %sp, PTREGS_OFF, %o0
         call    sun4v_nonresum_error
@@ -321,7 +330,10 @@ sun4v_nonres_mondo_queue_full:
         wrpr    %g0, 15, %pil
         ba,pt   %xcc, etrap_irq
          rd     %pc, %g7
-
+#ifdef CONFIG_TRACE_IRQFLAGS
+        call            trace_hardirqs_off
+         nop
+#endif
         call    sun4v_nonresum_overflow
          add    %sp, PTREGS_OFF, %o0
 
diff --git a/arch/sparc64/kernel/sys_sunos32.c b/arch/sparc64/kernel/sys_sunos32.c
index 7da72d3b322a..2ebc2c051383 100644
--- a/arch/sparc64/kernel/sys_sunos32.c
+++ b/arch/sparc64/kernel/sys_sunos32.c
@@ -83,7 +83,7 @@ asmlinkage u32 sunos_mmap(u32 addr, u32 len, u32 prot, u32 flags, u32 fd, u32 of
                 file = fget(fd);
                 if (!file)
                         goto out;
-                inode = file->f_dentry->d_inode;
+                inode = file->f_path.dentry->d_inode;
                 if (imajor(inode) == MEM_MAJOR && iminor(inode) == 5) {
                         flags |= MAP_ANONYMOUS;
                         fput(file);
@@ -615,7 +615,7 @@ sunos_nfs_get_server_fd (int fd, struct sockaddr_in *addr)
         if (!file)
                 return 0;
 
-        inode = file->f_dentry->d_inode;
+        inode = file->f_path.dentry->d_inode;
 
         socket = SOCKET_I(inode);
         local.sin_family = AF_INET;
@@ -1055,7 +1055,7 @@ asmlinkage int sunos_msgsys(int op, u32 arg1, u32 arg2, u32 arg3, u32 arg4)
                 break;
         case 2:
                 rval = -EFAULT;
-                kmbuf = (struct msgbuf *)kmalloc(sizeof(struct msgbuf) + arg3,
+                kmbuf = kmalloc(sizeof(struct msgbuf) + arg3,
                                                  GFP_KERNEL);
                 if (!kmbuf)
                         break;
@@ -1078,7 +1078,7 @@ asmlinkage int sunos_msgsys(int op, u32 arg1, u32 arg2, u32 arg3, u32 arg4)
                 break;
         case 3:
                 rval = -EFAULT;
-                kmbuf = (struct msgbuf *)kmalloc(sizeof(struct msgbuf) + arg3,
+                kmbuf = kmalloc(sizeof(struct msgbuf) + arg3,
                                                  GFP_KERNEL);
                 if (!kmbuf || sunos_msgbuf_get((struct msgbuf32 __user *)(unsigned long)arg2,
                                                kmbuf, arg3))
diff --git a/arch/sparc64/kernel/traps.c b/arch/sparc64/kernel/traps.c
index fe1796c939c3..ad67784292db 100644
--- a/arch/sparc64/kernel/traps.c
+++ b/arch/sparc64/kernel/traps.c
@@ -10,7 +10,7 @@
  */
 
 #include <linux/module.h>
-#include <linux/sched.h>  /* for jiffies */
+#include <linux/sched.h>
 #include <linux/kernel.h>
 #include <linux/kallsyms.h>
 #include <linux/signal.h>
@@ -1873,6 +1873,16 @@ void sun4v_resum_error(struct pt_regs *regs, unsigned long offset)
 
         put_cpu();
 
+        if (ent->err_type == SUN4V_ERR_TYPE_WARNING_RES) {
+                /* If err_type is 0x4, it's a powerdown request.  Do
+                 * not do the usual resumable error log because that
+                 * makes it look like some abnormal error.
+                 */
+                printk(KERN_INFO "Power down request...\n");
+                kill_cad_pid(SIGINT, 1);
+                return;
+        }
+
         sun4v_log_error(regs, &local_copy, cpu,
                         KERN_ERR "RESUMABLE ERROR",
                         &sun4v_resum_oflow_cnt);
@@ -2261,8 +2271,12 @@ void die_if_kernel(char *str, struct pt_regs *regs)
         do_exit(SIGSEGV);
 }
 
+#define VIS_OPCODE_MASK ((0x3 << 30) | (0x3f << 19))
+#define VIS_OPCODE_VAL  ((0x2 << 30) | (0x36 << 19))
+
 extern int handle_popc(u32 insn, struct pt_regs *regs);
 extern int handle_ldf_stq(u32 insn, struct pt_regs *regs);
+extern int vis_emul(struct pt_regs *, unsigned int);
 
 void do_illegal_instruction(struct pt_regs *regs)
 {
@@ -2287,10 +2301,18 @@ void do_illegal_instruction(struct pt_regs *regs)
                         if (handle_ldf_stq(insn, regs))
                                 return;
                 } else if (tlb_type == hypervisor) {
-                        extern int vis_emul(struct pt_regs *, unsigned int);
+                        if ((insn & VIS_OPCODE_MASK) == VIS_OPCODE_VAL) {
+                                if (!vis_emul(regs, insn))
+                                        return;
+                        } else {
+                                struct fpustate *f = FPUSTATE;
 
-                        if (!vis_emul(regs, insn))
-                                return;
+                                /* XXX maybe verify XFSR bits like
+                                 * XXX do_fpother() does?
+                                 */
+                                if (do_mathemu(regs, f))
+                                        return;
+                        }
                 }
         }
         info.si_signo = SIGILL;
diff --git a/arch/sparc64/kernel/unaligned.c b/arch/sparc64/kernel/unaligned.c
index a9b765271b85..bc18d480dd1c 100644
--- a/arch/sparc64/kernel/unaligned.c
+++ b/arch/sparc64/kernel/unaligned.c
@@ -243,7 +243,7 @@ static inline int ok_for_kernel(unsigned int insn)
         return !floating_point_load_or_store_p(insn);
 }
 
-static void kernel_mna_trap_fault(void)
+static void kernel_mna_trap_fault(int fixup_tstate_asi)
 {
         struct pt_regs *regs = current_thread_info()->kern_una_regs;
         unsigned int insn = current_thread_info()->kern_una_insn;
@@ -274,18 +274,15 @@ static void kernel_mna_trap_fault(void)
         regs->tpc = entry->fixup;
         regs->tnpc = regs->tpc + 4;
 
-        regs->tstate &= ~TSTATE_ASI;
-        regs->tstate |= (ASI_AIUS << 24UL);
+        if (fixup_tstate_asi) {
+                regs->tstate &= ~TSTATE_ASI;
+                regs->tstate |= (ASI_AIUS << 24UL);
+        }
 }
 
-asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
+static void log_unaligned(struct pt_regs *regs)
 {
         static unsigned long count, last_time;
-        enum direction dir = decode_direction(insn);
-        int size = decode_access_size(insn);
-
-        current_thread_info()->kern_una_regs = regs;
-        current_thread_info()->kern_una_insn = insn;
 
         if (jiffies - last_time > 5 * HZ)
                 count = 0;
@@ -295,6 +292,28 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
                 printk("Kernel unaligned access at TPC[%lx] ", regs->tpc);
                 print_symbol("%s\n", regs->tpc);
         }
+}
+
+asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
+{
+        enum direction dir = decode_direction(insn);
+        int size = decode_access_size(insn);
+        int orig_asi, asi;
+
+        current_thread_info()->kern_una_regs = regs;
+        current_thread_info()->kern_una_insn = insn;
+
+        orig_asi = asi = decode_asi(insn, regs);
+
+        /* If this is a {get,put}_user() on an unaligned userspace pointer,
+         * just signal a fault and do not log the event.
+         */
+        if (asi == ASI_AIUS) {
+                kernel_mna_trap_fault(0);
+                return;
+        }
+
+        log_unaligned(regs);
 
         if (!ok_for_kernel(insn) || dir == both) {
                 printk("Unsupported unaligned load/store trap for kernel "
@@ -302,10 +321,10 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
                 unaligned_panic("Kernel does fpu/atomic "
                                 "unaligned load/store.", regs);
 
-                kernel_mna_trap_fault();
+                kernel_mna_trap_fault(0);
         } else {
                 unsigned long addr, *reg_addr;
-                int orig_asi, asi, err;
+                int err;
 
                 addr = compute_effective_address(regs, insn,
                                                  ((insn >> 25) & 0x1f));
@@ -315,7 +334,6 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
                        regs->tpc, dirstrings[dir], addr, size,
                        regs->u_regs[UREG_RETPC]);
 #endif
-                orig_asi = asi = decode_asi(insn, regs);
                 switch (asi) {
                 case ASI_NL:
                 case ASI_AIUPL:
@@ -365,7 +383,7 @@ asmlinkage void kernel_unaligned_trap(struct pt_regs *regs, unsigned int insn)
                         /* Not reached... */
                 }
                 if (unlikely(err))
-                        kernel_mna_trap_fault();
+                        kernel_mna_trap_fault(1);
                 else
                         advance(regs);
         }
diff --git a/arch/sparc64/kernel/visemul.c b/arch/sparc64/kernel/visemul.c
index 84fedaa38aae..c3fd64706b53 100644
--- a/arch/sparc64/kernel/visemul.c
+++ b/arch/sparc64/kernel/visemul.c
@@ -128,9 +128,6 @@
 /* 001001100 - Permute bytes as specified by GSR.MASK  */
 #define BSHUFFLE_OPF    0x04c
 
-#define VIS_OPCODE_MASK ((0x3 << 30) | (0x3f << 19))
-#define VIS_OPCODE_VAL  ((0x2 << 30) | (0x36 << 19))
-
 #define VIS_OPF_SHIFT   5
 #define VIS_OPF_MASK    (0x1ff << VIS_OPF_SHIFT)
 
@@ -810,9 +807,6 @@ int vis_emul(struct pt_regs *regs, unsigned int insn)
         if (get_user(insn, (u32 __user *) pc))
                 return -EFAULT;
 
-        if ((insn & VIS_OPCODE_MASK) != VIS_OPCODE_VAL)
-                return -EINVAL;
-
         opf = (insn & VIS_OPF_MASK) >> VIS_OPF_SHIFT;
         switch (opf) {
         default:
diff --git a/arch/sparc64/kernel/vmlinux.lds.S b/arch/sparc64/kernel/vmlinux.lds.S
index bd9de8c2a2aa..4a6063f33e7a 100644
--- a/arch/sparc64/kernel/vmlinux.lds.S
+++ b/arch/sparc64/kernel/vmlinux.lds.S
@@ -13,6 +13,7 @@ SECTIONS
   . = 0x4000;
   .text 0x0000000000404000 :
   {
+    _text = .;
     *(.text)
     SCHED_TEXT
     LOCK_TEXT