Merge remote-tracking branch 'origin/x86/boot' into x86/mm2

Coming patches to x86/mm2 require the changes and advanced baseline in
x86/boot.

Resolved Conflicts:
	arch/x86/kernel/setup.c
	mm/nobootmem.c

Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>

12 files changed, 227 insertions, 268 deletions

diff --git a/arch/tile/kernel/compat.c b/arch/tile/kernel/compat.c
index d67459b9ac2a..7f72401b4f45 100644
--- a/arch/tile/kernel/compat.c
+++ b/arch/tile/kernel/compat.c
@@ -76,24 +76,6 @@ long compat_sys_fallocate(int fd, int mode,
 			     ((loff_t)len_hi << 32) | len_lo);
 }
 
-
-
-long compat_sys_sched_rr_get_interval(compat_pid_t pid,
-				      struct compat_timespec __user *interval)
-{
-	struct timespec t;
-	int ret;
-	mm_segment_t old_fs = get_fs();
-
-	set_fs(KERNEL_DS);
-	ret = sys_sched_rr_get_interval(pid,
-					(struct timespec __force __user *)&t);
-	set_fs(old_fs);
-	if (put_compat_timespec(&t, interval))
-		return -EFAULT;
-	return ret;
-}
-
 /* Provide the compat syscall number to call mapping. */
 #undef __SYSCALL
 #define __SYSCALL(nr, call) [nr] = (call),
@@ -102,9 +84,7 @@ long compat_sys_sched_rr_get_interval(compat_pid_t pid,
 #define compat_sys_fadvise64_64 sys32_fadvise64_64
 #define compat_sys_readahead sys32_readahead
 
-/* Call the trampolines to manage pt_regs where necessary. */
-#define compat_sys_execve _compat_sys_execve
-#define compat_sys_sigaltstack _compat_sys_sigaltstack
+/* Call the assembly trampolines where necessary. */
 #define compat_sys_rt_sigreturn _compat_sys_rt_sigreturn
 #define sys_clone _sys_clone
 
diff --git a/arch/tile/kernel/compat_signal.c b/arch/tile/kernel/compat_signal.c
index 08b4fe1717bb..2e4cc69224a6 100644
--- a/arch/tile/kernel/compat_signal.c
+++ b/arch/tile/kernel/compat_signal.c
@@ -197,8 +197,7 @@ int copy_siginfo_from_user32(siginfo_t *to, struct compat_siginfo __user *from)
 }
 
 long compat_sys_sigaltstack(const struct compat_sigaltstack __user *uss_ptr,
-			    struct compat_sigaltstack __user *uoss_ptr,
-			    struct pt_regs *regs)
+			    struct compat_sigaltstack __user *uoss_ptr)
 {
 	stack_t uss, uoss;
 	int ret;
@@ -219,7 +218,7 @@ long compat_sys_sigaltstack(const struct compat_sigaltstack __user *uss_ptr,
 	set_fs(KERNEL_DS);
 	ret = do_sigaltstack(uss_ptr ? (stack_t __user __force *)&uss : NULL,
 			     (stack_t __user __force *)&uoss,
-			     (unsigned long)compat_ptr(regs->sp));
+			     (unsigned long)compat_ptr(current_pt_regs()->sp));
 	set_fs(seg);
 	if (ret >= 0 && uoss_ptr)  {
 		if (!access_ok(VERIFY_WRITE, uoss_ptr, sizeof(*uoss_ptr)) ||
@@ -232,8 +231,9 @@ long compat_sys_sigaltstack(const struct compat_sigaltstack __user *uss_ptr,
 }
 
 /* The assembly shim for this function arranges to ignore the return value. */
-long compat_sys_rt_sigreturn(struct pt_regs *regs)
+long compat_sys_rt_sigreturn(void)
 {
+	struct pt_regs *regs = current_pt_regs();
 	struct compat_rt_sigframe __user *frame =
 		(struct compat_rt_sigframe __user *) compat_ptr(regs->sp);
 	sigset_t set;
@@ -248,7 +248,7 @@ long compat_sys_rt_sigreturn(struct pt_regs *regs)
 	if (restore_sigcontext(regs, &frame->uc.uc_mcontext))
 		goto badframe;
 
-	if (compat_sys_sigaltstack(&frame->uc.uc_stack, NULL, regs) != 0)
+	if (compat_sys_sigaltstack(&frame->uc.uc_stack, NULL) == -EFAULT)
 		goto badframe;
 
 	return 0;
diff --git a/arch/tile/kernel/entry.S b/arch/tile/kernel/entry.S
index c31637baff28..f116cb0bce20 100644
--- a/arch/tile/kernel/entry.S
+++ b/arch/tile/kernel/entry.S
@@ -28,17 +28,6 @@ STD_ENTRY(current_text_addr)
 	STD_ENDPROC(current_text_addr)
 
 /*
- * Implement execve().  The i386 code has a note that forking from kernel
- * space results in no copy on write until the execve, so we should be
- * careful not to write to the stack here.
- */
-STD_ENTRY(kernel_execve)
-	moveli TREG_SYSCALL_NR_NAME, __NR_execve
-	swint1
-	jrp lr
-	STD_ENDPROC(kernel_execve)
-
-/*
  * We don't run this function directly, but instead copy it to a page
  * we map into every user process.  See vdso_setup().
  *
diff --git a/arch/tile/kernel/intvec_32.S b/arch/tile/kernel/intvec_32.S
index 6943515100f8..f212bf7cea86 100644
--- a/arch/tile/kernel/intvec_32.S
+++ b/arch/tile/kernel/intvec_32.S
@@ -1291,6 +1291,21 @@ STD_ENTRY(ret_from_fork)
 	}
 	STD_ENDPROC(ret_from_fork)
 
+STD_ENTRY(ret_from_kernel_thread)
+	jal     sim_notify_fork
+	jal     schedule_tail
+	FEEDBACK_REENTER(ret_from_fork)
+	{
+	 move   r0, r31
+	 jalr   r30
+	}
+	FEEDBACK_REENTER(ret_from_kernel_thread)
+	{
+	 movei  r30, 0               /* not an NMI */
+	 j      .Lresume_userspace   /* jump into middle of interrupt_return */
+	}
+	STD_ENDPROC(ret_from_kernel_thread)
+
 	/*
 	 * Code for ill interrupt.
 	 */
@@ -1437,15 +1452,6 @@ STD_ENTRY_LOCAL(bad_intr)
 	panic   "Unhandled interrupt %#x: PC %#lx"
 	STD_ENDPROC(bad_intr)
 
-/* Put address of pt_regs in reg and jump. */
-#define PTREGS_SYSCALL(x, reg)                          \
-	STD_ENTRY(_##x);                                \
-	{                                               \
-	 PTREGS_PTR(reg, PTREGS_OFFSET_BASE);           \
-	 j      x                                       \
-	};                                              \
-	STD_ENDPROC(_##x)
-
 /*
  * Special-case sigreturn to not write r0 to the stack on return.
  * This is technically more efficient, but it also avoids difficulties
@@ -1461,12 +1467,9 @@ STD_ENTRY_LOCAL(bad_intr)
 	};                                              \
 	STD_ENDPROC(_##x)
 
-PTREGS_SYSCALL(sys_execve, r3)
-PTREGS_SYSCALL(sys_sigaltstack, r2)
 PTREGS_SYSCALL_SIGRETURN(sys_rt_sigreturn, r0)
-PTREGS_SYSCALL(sys_cmpxchg_badaddr, r1)
 
-/* Save additional callee-saves to pt_regs, put address in r4 and jump. */
+/* Save additional callee-saves to pt_regs and jump to standard function. */
 STD_ENTRY(_sys_clone)
 	push_extra_callee_saves r4
 	j       sys_clone
diff --git a/arch/tile/kernel/intvec_64.S b/arch/tile/kernel/intvec_64.S
index 7c06d597ffd0..54bc9a6678e8 100644
--- a/arch/tile/kernel/intvec_64.S
+++ b/arch/tile/kernel/intvec_64.S
@@ -1150,6 +1150,21 @@ STD_ENTRY(ret_from_fork)
 	}
 	STD_ENDPROC(ret_from_fork)
 
+STD_ENTRY(ret_from_kernel_thread)
+	jal     sim_notify_fork
+	jal     schedule_tail
+	FEEDBACK_REENTER(ret_from_fork)
+	{
+	 move   r0, r31
+	 jalr   r30
+	}
+	FEEDBACK_REENTER(ret_from_kernel_thread)
+	{
+	 movei  r30, 0               /* not an NMI */
+	 j      .Lresume_userspace   /* jump into middle of interrupt_return */
+	}
+	STD_ENDPROC(ret_from_kernel_thread)
+
 /* Various stub interrupt handlers and syscall handlers */
 
 STD_ENTRY_LOCAL(_kernel_double_fault)
@@ -1166,15 +1181,6 @@ STD_ENTRY_LOCAL(bad_intr)
 	panic   "Unhandled interrupt %#x: PC %#lx"
 	STD_ENDPROC(bad_intr)
 
-/* Put address of pt_regs in reg and jump. */
-#define PTREGS_SYSCALL(x, reg)                          \
-	STD_ENTRY(_##x);                                \
-	{                                               \
-	 PTREGS_PTR(reg, PTREGS_OFFSET_BASE);           \
-	 j      x                                       \
-	};                                              \
-	STD_ENDPROC(_##x)
-
 /*
  * Special-case sigreturn to not write r0 to the stack on return.
  * This is technically more efficient, but it also avoids difficulties
@@ -1190,16 +1196,12 @@ STD_ENTRY_LOCAL(bad_intr)
 	};                                              \
 	STD_ENDPROC(_##x)
 
-PTREGS_SYSCALL(sys_execve, r3)
-PTREGS_SYSCALL(sys_sigaltstack, r2)
 PTREGS_SYSCALL_SIGRETURN(sys_rt_sigreturn, r0)
 #ifdef CONFIG_COMPAT
-PTREGS_SYSCALL(compat_sys_execve, r3)
-PTREGS_SYSCALL(compat_sys_sigaltstack, r2)
 PTREGS_SYSCALL_SIGRETURN(compat_sys_rt_sigreturn, r0)
 #endif
 
-/* Save additional callee-saves to pt_regs, put address in r4 and jump. */
+/* Save additional callee-saves to pt_regs and jump to standard function. */
 STD_ENTRY(_sys_clone)
 	push_extra_callee_saves r4
 	j       sys_clone
diff --git a/arch/tile/kernel/module.c b/arch/tile/kernel/module.c
index 243ffebe38d6..4918d91bc3a6 100644
--- a/arch/tile/kernel/module.c
+++ b/arch/tile/kernel/module.c
@@ -42,8 +42,6 @@ void *module_alloc(unsigned long size)
 	int i = 0;
 	int npages;
 
-	if (size == 0)
-		return NULL;
 	npages = (size + PAGE_SIZE - 1) / PAGE_SIZE;
 	pages = kmalloc(npages * sizeof(struct page *), GFP_KERNEL);
 	if (pages == NULL)
diff --git a/arch/tile/kernel/pci.c b/arch/tile/kernel/pci.c
index 759822687e8f..67237d34c2e2 100644
--- a/arch/tile/kernel/pci.c
+++ b/arch/tile/kernel/pci.c
@@ -81,7 +81,7 @@ EXPORT_SYMBOL(pcibios_align_resource);
  * controller_id is the controller number, config type is 0 or 1 for
  * config0 or config1 operations.
  */
-static int __devinit tile_pcie_open(int controller_id, int config_type)
+static int tile_pcie_open(int controller_id, int config_type)
 {
 	char filename[32];
 	int fd;
@@ -97,8 +97,7 @@ static int __devinit tile_pcie_open(int controller_id, int config_type)
 /*
  * Get the IRQ numbers from the HV and set up the handlers for them.
  */
-static int __devinit tile_init_irqs(int controller_id,
-				 struct pci_controller *controller)
+static int tile_init_irqs(int controller_id, struct pci_controller *controller)
 {
 	char filename[32];
 	int fd;
@@ -237,7 +236,7 @@ static int tile_map_irq(const struct pci_dev *dev, u8 slot, u8 pin)
 }
 
 
-static void __devinit fixup_read_and_payload_sizes(void)
+static void fixup_read_and_payload_sizes(void)
 {
 	struct pci_dev *dev = NULL;
 	int smallest_max_payload = 0x1; /* Tile maxes out at 256 bytes. */
@@ -245,7 +244,7 @@ static void __devinit fixup_read_and_payload_sizes(void)
 	u16 new_values;
 
 	/* Scan for the smallest maximum payload size. */
-	while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL) {
+	for_each_pci_dev(dev) {
 		u32 devcap;
 		int max_payload;
 
@@ -260,7 +259,7 @@ static void __devinit fixup_read_and_payload_sizes(void)
 
 	/* Now, set the max_payload_size for all devices to that value. */
 	new_values = (max_read_size << 12) | (smallest_max_payload << 5);
-	while ((dev = pci_get_device(PCI_ANY_ID, PCI_ANY_ID, dev)) != NULL)
+	for_each_pci_dev(dev)
 		pcie_capability_clear_and_set_word(dev, PCI_EXP_DEVCTL,
 				PCI_EXP_DEVCTL_PAYLOAD | PCI_EXP_DEVCTL_READRQ,
 				new_values);
@@ -379,7 +378,7 @@ subsys_initcall(pcibios_init);
 /*
  * No bus fixups needed.
  */
-void __devinit pcibios_fixup_bus(struct pci_bus *bus)
+void pcibios_fixup_bus(struct pci_bus *bus)
 {
 	/* Nothing needs to be done. */
 }
@@ -458,11 +457,8 @@ int pcibios_enable_device(struct pci_dev *dev, int mask)
  * specified bus & slot.
  */
 
-static int __devinit tile_cfg_read(struct pci_bus *bus,
-				   unsigned int devfn,
-				   int offset,
-				   int size,
-				   u32 *val)
+static int tile_cfg_read(struct pci_bus *bus, unsigned int devfn, int offset,
+			 int size, u32 *val)
 {
 	struct pci_controller *controller = bus->sysdata;
 	int busnum = bus->number & 0xff;
@@ -504,11 +500,8 @@ static int __devinit tile_cfg_read(struct pci_bus *bus,
  * See tile_cfg_read() for relevant comments.
  * Note that "val" is the value to write, not a pointer to that value.
  */
-static int __devinit tile_cfg_write(struct pci_bus *bus,
-				    unsigned int devfn,
-				    int offset,
-				    int size,
-				    u32 val)
+static int tile_cfg_write(struct pci_bus *bus, unsigned int devfn, int offset,
+			  int size, u32 val)
 {
 	struct pci_controller *controller = bus->sysdata;
 	int busnum = bus->number & 0xff;
diff --git a/arch/tile/kernel/pci_gx.c b/arch/tile/kernel/pci_gx.c
index 2ba6d052f85d..11425633b2d7 100644
--- a/arch/tile/kernel/pci_gx.c
+++ b/arch/tile/kernel/pci_gx.c
@@ -58,10 +58,10 @@
 #define TRACE_CFG_RD(...)
 #endif
 
-static int __devinitdata pci_probe = 1;
+static int pci_probe = 1;
 
 /* Information on the PCIe RC ports configuration. */
-static int __devinitdata pcie_rc[TILEGX_NUM_TRIO][TILEGX_TRIO_PCIES];
+static int pcie_rc[TILEGX_NUM_TRIO][TILEGX_TRIO_PCIES];
 
 /*
  * On some platforms with one or more Gx endpoint ports, we need to
@@ -72,7 +72,7 @@ static int __devinitdata pcie_rc[TILEGX_NUM_TRIO][TILEGX_TRIO_PCIES];
  * the delay in seconds. If the delay is not provided, the value
  * will be DEFAULT_RC_DELAY.
  */
-static int __devinitdata rc_delay[TILEGX_NUM_TRIO][TILEGX_TRIO_PCIES];
+static int rc_delay[TILEGX_NUM_TRIO][TILEGX_TRIO_PCIES];
 
 /* Default number of seconds that the PCIe RC port probe can be delayed. */
 #define DEFAULT_RC_DELAY	10
@@ -137,7 +137,7 @@ static int tile_irq_cpu(int irq)
 /*
  * Open a file descriptor to the TRIO shim.
  */
-static int __devinit tile_pcie_open(int trio_index)
+static int tile_pcie_open(int trio_index)
 {
 	gxio_trio_context_t *context = &trio_contexts[trio_index];
 	int ret;
@@ -265,7 +265,7 @@ trio_handle_level_irq(unsigned int irq, struct irq_desc *desc)
  * Create kernel irqs and set up the handlers for the legacy interrupts.
  * Also some minimum initialization for the MSI support.
  */
-static int __devinit tile_init_irqs(struct pci_controller *controller)
+static int tile_init_irqs(struct pci_controller *controller)
 {
 	int i;
 	int j;
@@ -459,8 +459,7 @@ static int tile_map_irq(const struct pci_dev *dev, u8 device, u8 pin)
 }
 
 
-static void __devinit fixup_read_and_payload_sizes(struct pci_controller *
-						controller)
+static void fixup_read_and_payload_sizes(struct pci_controller *controller)
 {
 	gxio_trio_context_t *trio_context = controller->trio;
 	struct pci_bus *root_bus = controller->root_bus;
@@ -541,7 +540,7 @@ static void __devinit fixup_read_and_payload_sizes(struct pci_controller *
 	}
 }
 
-static int __devinit setup_pcie_rc_delay(char *str)
+static int setup_pcie_rc_delay(char *str)
 {
 	unsigned long delay = 0;
 	unsigned long trio_index;
@@ -1016,7 +1015,7 @@ alloc_mem_map_failed:
 subsys_initcall(pcibios_init);
 
 /* Note: to be deleted after Linux 3.6 merge. */
-void __devinit pcibios_fixup_bus(struct pci_bus *bus)
+void pcibios_fixup_bus(struct pci_bus *bus)
 {
 }
 
@@ -1024,7 +1023,7 @@ void __devinit pcibios_fixup_bus(struct pci_bus *bus)
  * This can be called from the generic PCI layer, but doesn't need to
  * do anything.
  */
-char __devinit *pcibios_setup(char *str)
+char *pcibios_setup(char *str)
 {
 	if (!strcmp(str, "off")) {
 		pci_probe = 0;
@@ -1047,8 +1046,7 @@ int pcibios_enable_device(struct pci_dev *dev, int mask)
 }
 
 /* Called for each device after PCI setup is done. */
-static void __init
-pcibios_fixup_final(struct pci_dev *pdev)
+static void pcibios_fixup_final(struct pci_dev *pdev)
 {
 	set_dma_ops(&pdev->dev, gx_pci_dma_map_ops);
 	set_dma_offset(&pdev->dev, TILE_PCI_MEM_MAP_BASE_OFFSET);
@@ -1144,11 +1142,8 @@ EXPORT_SYMBOL(pci_iounmap);
  * specified bus & device.
  */
 
-static int __devinit tile_cfg_read(struct pci_bus *bus,
-				   unsigned int devfn,
-				   int offset,
-				   int size,
-				   u32 *val)
+static int tile_cfg_read(struct pci_bus *bus, unsigned int devfn, int offset,
+			 int size, u32 *val)
 {
 	struct pci_controller *controller = bus->sysdata;
 	gxio_trio_context_t *trio_context = controller->trio;
@@ -1272,11 +1267,8 @@ invalid_device:
  * See tile_cfg_read() for relevent comments.
  * Note that "val" is the value to write, not a pointer to that value.
  */
-static int __devinit tile_cfg_write(struct pci_bus *bus,
-				    unsigned int devfn,
-				    int offset,
-				    int size,
-				    u32 val)
+static int tile_cfg_write(struct pci_bus *bus, unsigned int devfn, int offset,
+			  int size, u32 val)
 {
 	struct pci_controller *controller = bus->sysdata;
 	gxio_trio_context_t *trio_context = controller->trio;
diff --git a/arch/tile/kernel/process.c b/arch/tile/kernel/process.c
index 307d010696c9..0e5661e7d00d 100644
--- a/arch/tile/kernel/process.c
+++ b/arch/tile/kernel/process.c
@@ -157,24 +157,43 @@ void arch_release_thread_info(struct thread_info *info)
 static void save_arch_state(struct thread_struct *t);
 
 int copy_thread(unsigned long clone_flags, unsigned long sp,
-		unsigned long stack_size,
-		struct task_struct *p, struct pt_regs *regs)
+		unsigned long arg, struct task_struct *p)
 {
-	struct pt_regs *childregs;
+	struct pt_regs *childregs = task_pt_regs(p), *regs = current_pt_regs();
 	unsigned long ksp;
+	unsigned long *callee_regs;
 
 	/*
-	 * When creating a new kernel thread we pass sp as zero.
-	 * Assign it to a reasonable value now that we have the stack.
+	 * Set up the stack and stack pointer appropriately for the
+	 * new child to find itself woken up in __switch_to().
+	 * The callee-saved registers must be on the stack to be read;
+	 * the new task will then jump to assembly support to handle
+	 * calling schedule_tail(), etc., and (for userspace tasks)
+	 * returning to the context set up in the pt_regs.
 	 */
-	if (sp == 0 && regs->ex1 == PL_ICS_EX1(KERNEL_PL, 0))
-		sp = KSTK_TOP(p);
+	ksp = (unsigned long) childregs;
+	ksp -= C_ABI_SAVE_AREA_SIZE;   /* interrupt-entry save area */
+	((long *)ksp)[0] = ((long *)ksp)[1] = 0;
+	ksp -= CALLEE_SAVED_REGS_COUNT * sizeof(unsigned long);
+	callee_regs = (unsigned long *)ksp;
+	ksp -= C_ABI_SAVE_AREA_SIZE;   /* __switch_to() save area */
+	((long *)ksp)[0] = ((long *)ksp)[1] = 0;
+	p->thread.ksp = ksp;
 
-	/*
-	 * Do not clone step state from the parent; each thread
-	 * must make its own lazily.
-	 */
-	task_thread_info(p)->step_state = NULL;
+	/* Record the pid of the task that created this one. */
+	p->thread.creator_pid = current->pid;
+
+	if (unlikely(p->flags & PF_KTHREAD)) {
+		/* kernel thread */
+		memset(childregs, 0, sizeof(struct pt_regs));
+		memset(&callee_regs[2], 0,
+		       (CALLEE_SAVED_REGS_COUNT - 2) * sizeof(unsigned long));
+		callee_regs[0] = sp;   /* r30 = function */
+		callee_regs[1] = arg;  /* r31 = arg */
+		childregs->ex1 = PL_ICS_EX1(KERNEL_PL, 0);
+		p->thread.pc = (unsigned long) ret_from_kernel_thread;
+		return 0;
+	}
 
 	/*
 	 * Start new thread in ret_from_fork so it schedules properly
@@ -182,46 +201,33 @@ int copy_thread(unsigned long clone_flags, unsigned long sp,
 	 */
 	p->thread.pc = (unsigned long) ret_from_fork;
 
-	/* Save user stack top pointer so we can ID the stack vm area later. */
-	p->thread.usp0 = sp;
-
-	/* Record the pid of the process that created this one. */
-	p->thread.creator_pid = current->pid;
+	/*
+	 * Do not clone step state from the parent; each thread
+	 * must make its own lazily.
+	 */
+	task_thread_info(p)->step_state = NULL;
 
 	/*
 	 * Copy the registers onto the kernel stack so the
 	 * return-from-interrupt code will reload it into registers.
 	 */
-	childregs = task_pt_regs(p);
-	*childregs = *regs;
+	*childregs = *current_pt_regs();
 	childregs->regs[0] = 0;         /* return value is zero */
-	childregs->sp = sp;  /* override with new user stack pointer */
+	if (sp)
+		childregs->sp = sp;  /* override with new user stack pointer */
+	memcpy(callee_regs, &childregs->regs[CALLEE_SAVED_FIRST_REG],
+	       CALLEE_SAVED_REGS_COUNT * sizeof(unsigned long));
+
+	/* Save user stack top pointer so we can ID the stack vm area later. */
+	p->thread.usp0 = childregs->sp;
 
 	/*
 	 * If CLONE_SETTLS is set, set "tp" in the new task to "r4",
 	 * which is passed in as arg #5 to sys_clone().
 	 */
 	if (clone_flags & CLONE_SETTLS)
-		childregs->tp = regs->regs[4];
+		childregs->tp = childregs->regs[4];
 
-	/*
-	 * Copy the callee-saved registers from the passed pt_regs struct
-	 * into the context-switch callee-saved registers area.
-	 * This way when we start the interrupt-return sequence, the
-	 * callee-save registers will be correctly in registers, which
-	 * is how we assume the compiler leaves them as we start doing
-	 * the normal return-from-interrupt path after calling C code.
-	 * Zero out the C ABI save area to mark the top of the stack.
-	 */
-	ksp = (unsigned long) childregs;
-	ksp -= C_ABI_SAVE_AREA_SIZE;   /* interrupt-entry save area */
-	((long *)ksp)[0] = ((long *)ksp)[1] = 0;
-	ksp -= CALLEE_SAVED_REGS_COUNT * sizeof(unsigned long);
-	memcpy((void *)ksp, &regs->regs[CALLEE_SAVED_FIRST_REG],
-	       CALLEE_SAVED_REGS_COUNT * sizeof(unsigned long));
-	ksp -= C_ABI_SAVE_AREA_SIZE;   /* __switch_to() save area */
-	((long *)ksp)[0] = ((long *)ksp)[1] = 0;
-	p->thread.ksp = ksp;
 
 #if CHIP_HAS_TILE_DMA()
 	/*
@@ -577,62 +583,6 @@ int do_work_pending(struct pt_regs *regs, u32 thread_info_flags)
 	panic("work_pending: bad flags %#x\n", thread_info_flags);
 }
 
-/* Note there is an implicit fifth argument if (clone_flags & CLONE_SETTLS). */
-SYSCALL_DEFINE5(clone, unsigned long, clone_flags, unsigned long, newsp,
-		void __user *, parent_tidptr, void __user *, child_tidptr,
-		struct pt_regs *, regs)
-{
-	if (!newsp)
-		newsp = regs->sp;
-	return do_fork(clone_flags, newsp, regs, 0,
-		       parent_tidptr, child_tidptr);
-}
-
-/*
- * sys_execve() executes a new program.
- */
-SYSCALL_DEFINE4(execve, const char __user *, path,
-		const char __user *const __user *, argv,
-		const char __user *const __user *, envp,
-		struct pt_regs *, regs)
-{
-	long error;
-	struct filename *filename;
-
-	filename = getname(path);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-	if (error == 0)
-		single_step_execve();
-out:
-	return error;
-}
-
-#ifdef CONFIG_COMPAT
-long compat_sys_execve(const char __user *path,
-		       compat_uptr_t __user *argv,
-		       compat_uptr_t __user *envp,
-		       struct pt_regs *regs)
-{
-	long error;
-	struct filename *filename;
-
-	filename = getname(path);
-	error = PTR_ERR(filename);
-	if (IS_ERR(filename))
-		goto out;
-	error = compat_do_execve(filename->name, argv, envp, regs);
-	putname(filename);
-	if (error == 0)
-		single_step_execve();
-out:
-	return error;
-}
-#endif
-
 unsigned long get_wchan(struct task_struct *p)
 {
 	struct KBacktraceIterator kbt;
@@ -650,37 +600,6 @@ unsigned long get_wchan(struct task_struct *p)
 	return 0;
 }
 
-/*
- * We pass in lr as zero (cleared in kernel_thread) and the caller
- * part of the backtrace ABI on the stack also zeroed (in copy_thread)
- * so that backtraces will stop with this function.
- * Note that we don't use r0, since copy_thread() clears it.
- */
-static void start_kernel_thread(int dummy, int (*fn)(int), int arg)
-{
-	do_exit(fn(arg));
-}
-
-/*
- * Create a kernel thread
- */
-int kernel_thread(int (*fn)(void *), void * arg, unsigned long flags)
-{
-	struct pt_regs regs;
-
-	memset(&regs, 0, sizeof(regs));
-	regs.ex1 = PL_ICS_EX1(KERNEL_PL, 0);  /* run at kernel PL, no ICS */
-	regs.pc = (long) start_kernel_thread;
-	regs.flags = PT_FLAGS_CALLER_SAVES;   /* need to restore r1 and r2 */
-	regs.regs[1] = (long) fn;             /* function pointer */
-	regs.regs[2] = (long) arg;            /* parameter register */
-
-	/* Ok, create the new process.. */
-	return do_fork(flags | CLONE_VM | CLONE_UNTRACED, 0, &regs,
-		       0, NULL, NULL);
-}
-EXPORT_SYMBOL(kernel_thread);
-
 /* Flush thread state. */
 void flush_thread(void)
 {
diff --git a/arch/tile/kernel/ptrace.c b/arch/tile/kernel/ptrace.c
index e92e40527d6d..9835312d5a91 100644
--- a/arch/tile/kernel/ptrace.c
+++ b/arch/tile/kernel/ptrace.c
@@ -19,7 +19,10 @@
 #include <linux/kprobes.h>
 #include <linux/compat.h>
 #include <linux/uaccess.h>
+#include <linux/regset.h>
+#include <linux/elf.h>
 #include <asm/traps.h>
+#include <arch/chip.h>
 
 void user_enable_single_step(struct task_struct *child)
 {
@@ -45,6 +48,100 @@ void ptrace_disable(struct task_struct *child)
 	clear_tsk_thread_flag(child, TIF_SYSCALL_TRACE);
 }
 
+/*
+ * Get registers from task and ready the result for userspace.
+ * Note that we localize the API issues to getregs() and putregs() at
+ * some cost in performance, e.g. we need a full pt_regs copy for
+ * PEEKUSR, and two copies for POKEUSR.  But in general we expect
+ * GETREGS/PUTREGS to be the API of choice anyway.
+ */
+static char *getregs(struct task_struct *child, struct pt_regs *uregs)
+{
+	*uregs = *task_pt_regs(child);
+
+	/* Set up flags ABI bits. */
+	uregs->flags = 0;
+#ifdef CONFIG_COMPAT
+	if (task_thread_info(child)->status & TS_COMPAT)
+		uregs->flags |= PT_FLAGS_COMPAT;
+#endif
+
+	return (char *)uregs;
+}
+
+/* Put registers back to task. */
+static void putregs(struct task_struct *child, struct pt_regs *uregs)
+{
+	struct pt_regs *regs = task_pt_regs(child);
+
+	/* Don't allow overwriting the kernel-internal flags word. */
+	uregs->flags = regs->flags;
+
+	/* Only allow setting the ICS bit in the ex1 word. */
+	uregs->ex1 = PL_ICS_EX1(USER_PL, EX1_ICS(uregs->ex1));
+
+	*regs = *uregs;
+}
+
+enum tile_regset {
+	REGSET_GPR,
+};
+
+static int tile_gpr_get(struct task_struct *target,
+			  const struct user_regset *regset,
+			  unsigned int pos, unsigned int count,
+			  void *kbuf, void __user *ubuf)
+{
+	struct pt_regs regs;
+
+	getregs(target, &regs);
+
+	return user_regset_copyout(&pos, &count, &kbuf, &ubuf, &regs, 0,
+				   sizeof(regs));
+}
+
+static int tile_gpr_set(struct task_struct *target,
+			  const struct user_regset *regset,
+			  unsigned int pos, unsigned int count,
+			  const void *kbuf, const void __user *ubuf)
+{
+	int ret;
+	struct pt_regs regs;
+
+	ret = user_regset_copyin(&pos, &count, &kbuf, &ubuf, &regs, 0,
+				 sizeof(regs));
+	if (ret)
+		return ret;
+
+	putregs(target, &regs);
+
+	return 0;
+}
+
+static const struct user_regset tile_user_regset[] = {
+	[REGSET_GPR] = {
+		.core_note_type = NT_PRSTATUS,
+		.n = ELF_NGREG,
+		.size = sizeof(elf_greg_t),
+		.align = sizeof(elf_greg_t),
+		.get = tile_gpr_get,
+		.set = tile_gpr_set,
+	},
+};
+
+static const struct user_regset_view tile_user_regset_view = {
+	.name = CHIP_ARCH_NAME,
+	.e_machine = ELF_ARCH,
+	.ei_osabi = ELF_OSABI,
+	.regsets = tile_user_regset,
+	.n = ARRAY_SIZE(tile_user_regset),
+};
+
+const struct user_regset_view *task_user_regset_view(struct task_struct *task)
+{
+	return &tile_user_regset_view;
+}
+
 long arch_ptrace(struct task_struct *child, long request,
 		 unsigned long addr, unsigned long data)
 {
@@ -53,14 +150,13 @@ long arch_ptrace(struct task_struct *child, long request,
 	long ret = -EIO;
 	char *childreg;
 	struct pt_regs copyregs;
-	int ex1_offset;
 
 	switch (request) {
 
 	case PTRACE_PEEKUSR:  /* Read register from pt_regs. */
 		if (addr >= PTREGS_SIZE)
 			break;
-		childreg = (char *)task_pt_regs(child) + addr;
+		childreg = getregs(child, &copyregs) + addr;
 #ifdef CONFIG_COMPAT
 		if (is_compat_task()) {
 			if (addr & (sizeof(compat_long_t)-1))
@@ -79,17 +175,7 @@ long arch_ptrace(struct task_struct *child, long request,
 	case PTRACE_POKEUSR:  /* Write register in pt_regs. */
 		if (addr >= PTREGS_SIZE)
 			break;
-		childreg = (char *)task_pt_regs(child) + addr;
-
-		/* Guard against overwrites of the privilege level. */
-		ex1_offset = PTREGS_OFFSET_EX1;
-#if defined(CONFIG_COMPAT) && defined(__BIG_ENDIAN)
-		if (is_compat_task())   /* point at low word */
-			ex1_offset += sizeof(compat_long_t);
-#endif
-		if (addr == ex1_offset)
-			data = PL_ICS_EX1(USER_PL, EX1_ICS(data));
-
+		childreg = getregs(child, &copyregs) + addr;
 #ifdef CONFIG_COMPAT
 		if (is_compat_task()) {
 			if (addr & (sizeof(compat_long_t)-1))
@@ -102,24 +188,20 @@ long arch_ptrace(struct task_struct *child, long request,
 				break;
 			*(long *)childreg = data;
 		}
+		putregs(child, &copyregs);
 		ret = 0;
 		break;
 
 	case PTRACE_GETREGS:  /* Get all registers from the child. */
-		if (copy_to_user(datap, task_pt_regs(child),
-				 sizeof(struct pt_regs)) == 0) {
-			ret = 0;
-		}
+		ret = copy_regset_to_user(child, &tile_user_regset_view,
+					  REGSET_GPR, 0,
+					  sizeof(struct pt_regs), datap);
 		break;
 
 	case PTRACE_SETREGS:  /* Set all registers in the child. */
-		if (copy_from_user(&copyregs, datap,
-				   sizeof(struct pt_regs)) == 0) {
-			copyregs.ex1 =
-				PL_ICS_EX1(USER_PL, EX1_ICS(copyregs.ex1));
-			*task_pt_regs(child) = copyregs;
-			ret = 0;
-		}
+		ret = copy_regset_from_user(child, &tile_user_regset_view,
+					    REGSET_GPR, 0,
+					    sizeof(struct pt_regs), datap);
 		break;
 
 	case PTRACE_GETFPREGS:  /* Get the child FPU state. */
@@ -128,12 +210,16 @@ long arch_ptrace(struct task_struct *child, long request,
 
 	case PTRACE_SETOPTIONS:
 		/* Support TILE-specific ptrace options. */
-		child->ptrace &= ~PT_TRACE_MASK_TILE;
+		BUILD_BUG_ON(PTRACE_O_MASK_TILE & PTRACE_O_MASK);
 		tmp = data & PTRACE_O_MASK_TILE;
 		data &= ~PTRACE_O_MASK_TILE;
 		ret = ptrace_request(child, request, addr, data);
-		if (tmp & PTRACE_O_TRACEMIGRATE)
-			child->ptrace |= PT_TRACE_MIGRATE;
+		if (ret == 0) {
+			unsigned int flags = child->ptrace;
+			flags &= ~(PTRACE_O_MASK_TILE << PT_OPT_FLAG_SHIFT);
+			flags |= (tmp << PT_OPT_FLAG_SHIFT);
+			child->ptrace = flags;
+		}
 		break;
 
 	default:
diff --git a/arch/tile/kernel/signal.c b/arch/tile/kernel/signal.c
index 67efb656d104..657a7ace4ab4 100644
--- a/arch/tile/kernel/signal.c
+++ b/arch/tile/kernel/signal.c
@@ -37,10 +37,10 @@
 
 #define DEBUG_SIG 0
 
-SYSCALL_DEFINE3(sigaltstack, const stack_t __user *, uss,
-		stack_t __user *, uoss, struct pt_regs *, regs)
+SYSCALL_DEFINE2(sigaltstack, const stack_t __user *, uss,
+		stack_t __user *, uoss)
 {
-	return do_sigaltstack(uss, uoss, regs->sp);
+	return do_sigaltstack(uss, uoss, current_pt_regs()->sp);
 }
 
 
@@ -83,8 +83,9 @@ void signal_fault(const char *type, struct pt_regs *regs,
 }
 
 /* The assembly shim for this function arranges to ignore the return value. */
-SYSCALL_DEFINE1(rt_sigreturn, struct pt_regs *, regs)
+SYSCALL_DEFINE0(rt_sigreturn)
 {
+	struct pt_regs *regs = current_pt_regs();
 	struct rt_sigframe __user *frame =
 		(struct rt_sigframe __user *)(regs->sp);
 	sigset_t set;
diff --git a/arch/tile/kernel/sys.c b/arch/tile/kernel/sys.c
index b08095b402d6..b881a7be24bd 100644
--- a/arch/tile/kernel/sys.c
+++ b/arch/tile/kernel/sys.c
@@ -106,14 +106,10 @@ SYSCALL_DEFINE6(mmap, unsigned long, addr, unsigned long, len,
 #define sys_readahead sys32_readahead
 #endif
 
-/* Call the trampolines to manage pt_regs where necessary. */
-#define sys_execve _sys_execve
-#define sys_sigaltstack _sys_sigaltstack
+/* Call the assembly trampolines where necessary. */
+#undef sys_rt_sigreturn
 #define sys_rt_sigreturn _sys_rt_sigreturn
 #define sys_clone _sys_clone
-#ifndef __tilegx__
-#define sys_cmpxchg_badaddr _sys_cmpxchg_badaddr
-#endif
 
 /*
  * Note that we can't include <linux/unistd.h> here since the header