8 files changed, 72 insertions, 40 deletions

diff --git a/arch/tile/kernel/compat_signal.c b/arch/tile/kernel/compat_signal.c
index 77763ccd5a7d..cdef6e5ec022 100644
--- a/arch/tile/kernel/compat_signal.c
+++ b/arch/tile/kernel/compat_signal.c
@@ -403,19 +403,17 @@ int compat_setup_rt_frame(int sig, struct k_sigaction *ka, siginfo_t *info,
          * Set up registers for signal handler.
          * Registers that we don't modify keep the value they had from
          * user-space at the time we took the signal.
+         * We always pass siginfo and mcontext, regardless of SA_SIGINFO,
+         * since some things rely on this (e.g. glibc's debug/segfault.c).
          */
         regs->pc = ptr_to_compat_reg(ka->sa.sa_handler);
         regs->ex1 = PL_ICS_EX1(USER_PL, 1); /* set crit sec in handler */
         regs->sp = ptr_to_compat_reg(frame);
         regs->lr = restorer;
         regs->regs[0] = (unsigned long) usig;
-
-        if (ka->sa.sa_flags & SA_SIGINFO) {
-                /* Need extra arguments, so mark to restore caller-saves. */
-                regs->regs[1] = ptr_to_compat_reg(&frame->info);
-                regs->regs[2] = ptr_to_compat_reg(&frame->uc);
-                regs->flags |= PT_FLAGS_CALLER_SAVES;
-        }
+        regs->regs[1] = ptr_to_compat_reg(&frame->info);
+        regs->regs[2] = ptr_to_compat_reg(&frame->uc);
+        regs->flags |= PT_FLAGS_CALLER_SAVES;
 
         /*
          * Notify any tracer that was single-stepping it.
diff --git a/arch/tile/kernel/intvec_32.S b/arch/tile/kernel/intvec_32.S
index 5d56a1ef5ba5..6943515100f8 100644
--- a/arch/tile/kernel/intvec_32.S
+++ b/arch/tile/kernel/intvec_32.S
@@ -839,6 +839,18 @@ STD_ENTRY(interrupt_return)
         FEEDBACK_REENTER(interrupt_return)
 
         /*
+         * Use r33 to hold whether we have already loaded the callee-saves
+         * into ptregs.  We don't want to do it twice in this loop, since
+         * then we'd clobber whatever changes are made by ptrace, etc.
+         * Get base of stack in r32.
+         */
+        {
+         GET_THREAD_INFO(r32)
+         movei  r33, 0
+        }
+
+.Lretry_work_pending:
+        /*
          * Disable interrupts so as to make sure we don't
          * miss an interrupt that sets any of the thread flags (like
          * need_resched or sigpending) between sampling and the iret.
@@ -848,9 +860,6 @@ STD_ENTRY(interrupt_return)
         IRQ_DISABLE(r20, r21)
         TRACE_IRQS_OFF  /* Note: clobbers registers r0-r29 */
 
-        /* Get base of stack in r32; note r30/31 are used as arguments here. */
-        GET_THREAD_INFO(r32)
-
 
         /* Check to see if there is any work to do before returning to user. */
         {
@@ -866,16 +875,18 @@ STD_ENTRY(interrupt_return)
 
         /*
          * Make sure we have all the registers saved for signal
-         * handling or single-step.  Call out to C code to figure out
-         * exactly what we need to do for each flag bit, then if
-         * necessary, reload the flags and recheck.
+         * handling, notify-resume, or single-step.  Call out to C
+         * code to figure out exactly what we need to do for each flag bit,
+         * then if necessary, reload the flags and recheck.
          */
-        push_extra_callee_saves r0
         {
          PTREGS_PTR(r0, PTREGS_OFFSET_BASE)
-         jal    do_work_pending
+         bnz    r33, 1f
         }
-        bnz     r0, .Lresume_userspace
+        push_extra_callee_saves r0
+        movei   r33, 1
+1:      jal     do_work_pending
+        bnz     r0, .Lretry_work_pending
 
         /*
          * In the NMI case we
@@ -1180,10 +1191,12 @@ handle_syscall:
         add     r20, r20, tp
         lw      r21, r20
         addi    r21, r21, 1
-        sw      r20, r21
+        {
+         sw     r20, r21
+         GET_THREAD_INFO(r31)
+        }
 
         /* Trace syscalls, if requested. */
-        GET_THREAD_INFO(r31)
         addi    r31, r31, THREAD_INFO_FLAGS_OFFSET
         lw      r30, r31
         andi    r30, r30, _TIF_SYSCALL_TRACE
@@ -1362,7 +1375,10 @@ handle_ill:
 3:
         /* set PC and continue */
         lw      r26, r24
-        sw      r28, r26
+        {
+         sw     r28, r26
+         GET_THREAD_INFO(r0)
+        }
 
         /*
          * Clear TIF_SINGLESTEP to prevent recursion if we execute an ill.
@@ -1370,7 +1386,6 @@ handle_ill:
          * need to clear it here and can't really impose on all other arches.
          * So what's another write between friends?
          */
-        GET_THREAD_INFO(r0)
 
         addi    r1, r0, THREAD_INFO_FLAGS_OFFSET
         {
diff --git a/arch/tile/kernel/intvec_64.S b/arch/tile/kernel/intvec_64.S
index 49d9d6621682..30ae76e50c44 100644
--- a/arch/tile/kernel/intvec_64.S
+++ b/arch/tile/kernel/intvec_64.S
@@ -647,6 +647,20 @@ STD_ENTRY(interrupt_return)
         FEEDBACK_REENTER(interrupt_return)
 
         /*
+         * Use r33 to hold whether we have already loaded the callee-saves
+         * into ptregs.  We don't want to do it twice in this loop, since
+         * then we'd clobber whatever changes are made by ptrace, etc.
+         */
+        {
+         movei  r33, 0
+         move   r32, sp
+        }
+
+        /* Get base of stack in r32. */
+        EXTRACT_THREAD_INFO(r32)
+
+.Lretry_work_pending:
+        /*
          * Disable interrupts so as to make sure we don't
          * miss an interrupt that sets any of the thread flags (like
          * need_resched or sigpending) between sampling and the iret.
@@ -656,9 +670,6 @@ STD_ENTRY(interrupt_return)
         IRQ_DISABLE(r20, r21)
         TRACE_IRQS_OFF  /* Note: clobbers registers r0-r29 */
 
-        /* Get base of stack in r32; note r30/31 are used as arguments here. */
-        GET_THREAD_INFO(r32)
-
 
         /* Check to see if there is any work to do before returning to user. */
         {
@@ -674,16 +685,18 @@ STD_ENTRY(interrupt_return)
 
         /*
          * Make sure we have all the registers saved for signal
-         * handling or single-step.  Call out to C code to figure out
+         * handling or notify-resume.  Call out to C code to figure out
          * exactly what we need to do for each flag bit, then if
          * necessary, reload the flags and recheck.
          */
-        push_extra_callee_saves r0
         {
          PTREGS_PTR(r0, PTREGS_OFFSET_BASE)
-         jal    do_work_pending
+         bnez   r33, 1f
         }
-        bnez    r0, .Lresume_userspace
+        push_extra_callee_saves r0
+        movei   r33, 1
+1:      jal     do_work_pending
+        bnez    r0, .Lretry_work_pending
 
         /*
          * In the NMI case we
@@ -968,11 +981,16 @@ handle_syscall:
         shl16insli r20, r20, hw0(irq_stat + IRQ_CPUSTAT_SYSCALL_COUNT_OFFSET)
         add     r20, r20, tp
         ld4s    r21, r20
-        addi    r21, r21, 1
-        st4     r20, r21
+        {
+         addi   r21, r21, 1
+         move   r31, sp
+        }
+        {
+         st4    r20, r21
+         EXTRACT_THREAD_INFO(r31)
+        }
 
         /* Trace syscalls, if requested. */
-        GET_THREAD_INFO(r31)
         addi    r31, r31, THREAD_INFO_FLAGS_OFFSET
         ld      r30, r31
         andi    r30, r30, _TIF_SYSCALL_TRACE
diff --git a/arch/tile/kernel/pci.c b/arch/tile/kernel/pci.c
index a1bb59eecc18..b56d12bf5900 100644
--- a/arch/tile/kernel/pci.c
+++ b/arch/tile/kernel/pci.c
@@ -141,7 +141,7 @@ static int __devinit tile_init_irqs(int controller_id,
  *
  * Returns the number of controllers discovered.
  */
-int __devinit tile_pci_init(void)
+int __init tile_pci_init(void)
 {
         int i;
 
@@ -287,7 +287,7 @@ static void __devinit fixup_read_and_payload_sizes(void)
  * The controllers have been set up by the time we get here, by a call to
  * tile_pci_init.
  */
-int __devinit pcibios_init(void)
+int __init pcibios_init(void)
 {
         int i;
 
diff --git a/arch/tile/kernel/proc.c b/arch/tile/kernel/proc.c
index 7a9327046404..446a7f52cc11 100644
--- a/arch/tile/kernel/proc.c
+++ b/arch/tile/kernel/proc.c
@@ -146,7 +146,6 @@ static ctl_table unaligned_table[] = {
         },
         {}
 };
-#endif
 
 static struct ctl_path tile_path[] = {
         { .procname = "tile" },
@@ -155,10 +154,9 @@ static struct ctl_path tile_path[] = {
 
 static int __init proc_sys_tile_init(void)
 {
-#ifndef __tilegx__  /* FIXME: GX: no support for unaligned access yet */
         register_sysctl_paths(tile_path, unaligned_table);
-#endif
         return 0;
 }
 
 arch_initcall(proc_sys_tile_init);
+#endif
diff --git a/arch/tile/kernel/process.c b/arch/tile/kernel/process.c
index 2d5ef617bb39..54e6c64b85cc 100644
--- a/arch/tile/kernel/process.c
+++ b/arch/tile/kernel/process.c
@@ -567,6 +567,10 @@ struct task_struct *__sched _switch_to(struct task_struct *prev,
  */
 int do_work_pending(struct pt_regs *regs, u32 thread_info_flags)
 {
+        /* If we enter in kernel mode, do nothing and exit the caller loop. */
+        if (!user_mode(regs))
+                return 0;
+
         if (thread_info_flags & _TIF_NEED_RESCHED) {
                 schedule();
                 return 1;
@@ -589,8 +593,7 @@ int do_work_pending(struct pt_regs *regs, u32 thread_info_flags)
                 return 1;
         }
         if (thread_info_flags & _TIF_SINGLESTEP) {
-                if ((regs->ex1 & SPR_EX_CONTEXT_1_1__PL_MASK) == 0)
-                        single_step_once(regs);
+                single_step_once(regs);
                 return 0;
         }
         panic("work_pending: bad flags %#x\n", thread_info_flags);
diff --git a/arch/tile/kernel/single_step.c b/arch/tile/kernel/single_step.c
index 9efbc1391b3c..89529c9f0605 100644
--- a/arch/tile/kernel/single_step.c
+++ b/arch/tile/kernel/single_step.c
@@ -346,12 +346,10 @@ void single_step_once(struct pt_regs *regs)
                 }
 
                 /* allocate a cache line of writable, executable memory */
-                down_write(&current->mm->mmap_sem);
-                buffer = (void __user *) do_mmap(NULL, 0, 64,
+                buffer = (void __user *) vm_mmap(NULL, 0, 64,
                                           PROT_EXEC | PROT_READ | PROT_WRITE,
                                           MAP_PRIVATE | MAP_ANONYMOUS,
                                           0);
-                up_write(&current->mm->mmap_sem);
 
                 if (IS_ERR((void __force *)buffer)) {
                         kfree(state);
diff --git a/arch/tile/kernel/smpboot.c b/arch/tile/kernel/smpboot.c
index b949edcec200..172aef7d3159 100644
--- a/arch/tile/kernel/smpboot.c
+++ b/arch/tile/kernel/smpboot.c
@@ -196,6 +196,8 @@ void __cpuinit online_secondary(void)
         /* This must be done before setting cpu_online_mask */
         wmb();
 
+        notify_cpu_starting(smp_processor_id());
+
         /*
          * We need to hold call_lock, so there is no inconsistency
          * between the time smp_call_function() determines number of