Merge commit 'v2.6.36-rc5' into perf/core

Merge reason: Pick up the latest fixes in -rc5.

Signed-off-by: Ingo Molnar <mingo@elte.hu>

13 files changed, 88 insertions, 131 deletions

diff --git a/arch/alpha/include/asm/cacheflush.h b/arch/alpha/include/asm/cacheflush.h
index 01d71e1..012f124 100644
--- a/arch/alpha/include/asm/cacheflush.h
+++ b/arch/alpha/include/asm/cacheflush.h
@@ -43,6 +43,8 @@ extern void smp_imb(void);
 /* ??? Ought to use this in arch/alpha/kernel/signal.c too.  */
 
 #ifndef CONFIG_SMP
+#include <linux/sched.h>
+
 extern void __load_new_mm_context(struct mm_struct *);
 static inline void
 flush_icache_user_range(struct vm_area_struct *vma, struct page *page,
diff --git a/arch/alpha/include/asm/unistd.h b/arch/alpha/include/asm/unistd.h
index 804e531..058937b 100644
--- a/arch/alpha/include/asm/unistd.h
+++ b/arch/alpha/include/asm/unistd.h
@@ -449,10 +449,13 @@
 #define __NR_pwritev			491
 #define __NR_rt_tgsigqueueinfo		492
 #define __NR_perf_event_open		493
+#define __NR_fanotify_init		494
+#define __NR_fanotify_mark		495
+#define __NR_prlimit64			496
 
 #ifdef __KERNEL__
 
-#define NR_SYSCALLS			494
+#define NR_SYSCALLS			497
 
 #define __ARCH_WANT_IPC_PARSE_VERSION
 #define __ARCH_WANT_OLD_READDIR
@@ -463,6 +466,7 @@
 #define __ARCH_WANT_SYS_OLD_GETRLIMIT
 #define __ARCH_WANT_SYS_OLDUMOUNT
 #define __ARCH_WANT_SYS_SIGPENDING
+#define __ARCH_WANT_SYS_RT_SIGSUSPEND
 
 /* "Conditional" syscalls.  What we want is
 
diff --git a/arch/alpha/kernel/entry.S b/arch/alpha/kernel/entry.S
index b45d913..ab1ee0a 100644
--- a/arch/alpha/kernel/entry.S
+++ b/arch/alpha/kernel/entry.S
@@ -317,14 +317,14 @@ ret_from_sys_call:
 	ldq	$0, SP_OFF($sp)
 	and	$0, 8, $0
 	beq	$0, restore_all
-ret_from_reschedule:
+ret_to_user:
 	/* Make sure need_resched and sigpending don't change between
 		sampling and the rti.  */
 	lda	$16, 7
 	call_pal PAL_swpipl
 	ldl	$5, TI_FLAGS($8)
 	and	$5, _TIF_WORK_MASK, $2
-	bne	$5, work_pending
+	bne	$2, work_pending
 restore_all:
 	RESTORE_ALL
 	call_pal PAL_rti
@@ -363,7 +363,7 @@ $ret_success:
  *       $8: current.
  *      $19: The old syscall number, or zero if this is not a return
  *           from a syscall that errored and is possibly restartable.
- *      $20: Error indication.
+ *      $20: The old a3 value
  */
 
 	.align	4
@@ -392,12 +392,18 @@ $work_resched:
 
 $work_notifysig:
 	mov	$sp, $16
-	br	$1, do_switch_stack
+	bsr	$1, do_switch_stack
 	mov	$sp, $17
 	mov	$5, $18
+	mov	$19, $9		/* save old syscall number */
+	mov	$20, $10	/* save old a3 */
+	and	$5, _TIF_SIGPENDING, $2
+	cmovne	$2, 0, $9	/* we don't want double syscall restarts */
 	jsr	$26, do_notify_resume
+	mov	$9, $19
+	mov	$10, $20
 	bsr	$1, undo_switch_stack
-	br	restore_all
+	br	ret_to_user
 .end work_pending
 
 /*
@@ -430,6 +436,7 @@ strace:
 	beq	$1, 1f
 	ldq	$27, 0($2)
 1:	jsr	$26, ($27), sys_gettimeofday
+ret_from_straced:
 	ldgp	$gp, 0($26)
 
 	/* check return.. */
@@ -757,11 +764,15 @@ sys_vfork:
 	.ent	sys_sigreturn
 sys_sigreturn:
 	.prologue 0
+	lda	$9, ret_from_straced
+	cmpult	$26, $9, $9
 	mov	$sp, $17
 	lda	$18, -SWITCH_STACK_SIZE($sp)
 	lda	$sp, -SWITCH_STACK_SIZE($sp)
 	jsr	$26, do_sigreturn
-	br	$1, undo_switch_stack
+	bne	$9, 1f
+	jsr	$26, syscall_trace
+1:	br	$1, undo_switch_stack
 	br	ret_from_sys_call
 .end sys_sigreturn
 
@@ -770,47 +781,19 @@ sys_sigreturn:
 	.ent	sys_rt_sigreturn
 sys_rt_sigreturn:
 	.prologue 0
+	lda	$9, ret_from_straced
+	cmpult	$26, $9, $9
 	mov	$sp, $17
 	lda	$18, -SWITCH_STACK_SIZE($sp)
 	lda	$sp, -SWITCH_STACK_SIZE($sp)
 	jsr	$26, do_rt_sigreturn
-	br	$1, undo_switch_stack
+	bne	$9, 1f
+	jsr	$26, syscall_trace
+1:	br	$1, undo_switch_stack
 	br	ret_from_sys_call
 .end sys_rt_sigreturn
 
 	.align	4
-	.globl	sys_sigsuspend
-	.ent	sys_sigsuspend
-sys_sigsuspend:
-	.prologue 0
-	mov	$sp, $17
-	br	$1, do_switch_stack
-	mov	$sp, $18
-	subq	$sp, 16, $sp
-	stq	$26, 0($sp)
-	jsr	$26, do_sigsuspend
-	ldq	$26, 0($sp)
-	lda	$sp, SWITCH_STACK_SIZE+16($sp)
-	ret
-.end sys_sigsuspend
-
-	.align	4
-	.globl	sys_rt_sigsuspend
-	.ent	sys_rt_sigsuspend
-sys_rt_sigsuspend:
-	.prologue 0
-	mov	$sp, $18
-	br	$1, do_switch_stack
-	mov	$sp, $19
-	subq	$sp, 16, $sp
-	stq	$26, 0($sp)
-	jsr	$26, do_rt_sigsuspend
-	ldq	$26, 0($sp)
-	lda	$sp, SWITCH_STACK_SIZE+16($sp)
-	ret
-.end sys_rt_sigsuspend
-
-	.align	4
 	.globl	sys_sethae
 	.ent	sys_sethae
 sys_sethae:
diff --git a/arch/alpha/kernel/err_ev6.c b/arch/alpha/kernel/err_ev6.c
index 8ca6345..253cf1a 100644
--- a/arch/alpha/kernel/err_ev6.c
+++ b/arch/alpha/kernel/err_ev6.c
@@ -90,11 +90,13 @@ static int
 ev6_parse_cbox(u64 c_addr, u64 c1_syn, u64 c2_syn, 
 	       u64 c_stat, u64 c_sts, int print)
 {
-	char *sourcename[] = { "UNKNOWN", "UNKNOWN", "UNKNOWN",
-			       "MEMORY", "BCACHE", "DCACHE", 
-			       "BCACHE PROBE", "BCACHE PROBE" };
-	char *streamname[] = { "D", "I" };
-	char *bitsname[] = { "SINGLE", "DOUBLE" };
+	static const char * const sourcename[] = {
+		"UNKNOWN", "UNKNOWN", "UNKNOWN",
+		"MEMORY", "BCACHE", "DCACHE",
+		"BCACHE PROBE", "BCACHE PROBE"
+	};
+	static const char * const streamname[] = { "D", "I" };
+	static const char * const bitsname[] = { "SINGLE", "DOUBLE" };
 	int status = MCHK_DISPOSITION_REPORT;
 	int source = -1, stream = -1, bits = -1;
 
diff --git a/arch/alpha/kernel/err_marvel.c b/arch/alpha/kernel/err_marvel.c
index 5c905aa..648ae88 100644
--- a/arch/alpha/kernel/err_marvel.c
+++ b/arch/alpha/kernel/err_marvel.c
@@ -589,22 +589,23 @@ marvel_print_pox_spl_cmplt(u64 spl_cmplt)
 static void
 marvel_print_pox_trans_sum(u64 trans_sum)
 {
-	char *pcix_cmd[] = { "Interrupt Acknowledge",
-			     "Special Cycle",
-			     "I/O Read",
-			     "I/O Write",
-			     "Reserved",
-			     "Reserved / Device ID Message",
-			     "Memory Read",
-			     "Memory Write",
-			     "Reserved / Alias to Memory Read Block",
-			     "Reserved / Alias to Memory Write Block",
-			     "Configuration Read",
-			     "Configuration Write",
-			     "Memory Read Multiple / Split Completion",
-			     "Dual Address Cycle",
-			     "Memory Read Line / Memory Read Block",
-			     "Memory Write and Invalidate / Memory Write Block"
+	static const char * const pcix_cmd[] = {
+		"Interrupt Acknowledge",
+		"Special Cycle",
+		"I/O Read",
+		"I/O Write",
+		"Reserved",
+		"Reserved / Device ID Message",
+		"Memory Read",
+		"Memory Write",
+		"Reserved / Alias to Memory Read Block",
+		"Reserved / Alias to Memory Write Block",
+		"Configuration Read",
+		"Configuration Write",
+		"Memory Read Multiple / Split Completion",
+		"Dual Address Cycle",
+		"Memory Read Line / Memory Read Block",
+		"Memory Write and Invalidate / Memory Write Block"
 	};
 
 #define IO7__POX_TRANSUM__PCI_ADDR__S		(0)
diff --git a/arch/alpha/kernel/err_titan.c b/arch/alpha/kernel/err_titan.c
index f7ed97c..c3b3781 100644
--- a/arch/alpha/kernel/err_titan.c
+++ b/arch/alpha/kernel/err_titan.c
@@ -75,8 +75,12 @@ titan_parse_p_serror(int which, u64 serror, int print)
 	int status = MCHK_DISPOSITION_REPORT;
 
 #ifdef CONFIG_VERBOSE_MCHECK
-	char *serror_src[] = {"GPCI", "APCI", "AGP HP", "AGP LP"};
-	char *serror_cmd[] = {"DMA Read", "DMA RMW", "SGTE Read", "Reserved"};
+	static const char * const serror_src[] = {
+		"GPCI", "APCI", "AGP HP", "AGP LP"
+	};
+	static const char * const serror_cmd[] = {
+		"DMA Read", "DMA RMW", "SGTE Read", "Reserved"
+	};
 #endif /* CONFIG_VERBOSE_MCHECK */
 
 #define TITAN__PCHIP_SERROR__LOST_UECC	(1UL << 0)
@@ -140,14 +144,15 @@ titan_parse_p_perror(int which, int port, u64 perror, int print)
 	int status = MCHK_DISPOSITION_REPORT;
 
 #ifdef CONFIG_VERBOSE_MCHECK
-	char *perror_cmd[] = { "Interrupt Acknowledge", "Special Cycle",
-			       "I/O Read",	       	"I/O Write",
-			       "Reserved",	       	"Reserved",
-			       "Memory Read",		"Memory Write",
-			       "Reserved",		"Reserved",
-			       "Configuration Read",	"Configuration Write",
-			       "Memory Read Multiple",	"Dual Address Cycle",
-			       "Memory Read Line","Memory Write and Invalidate"
+	static const char * const perror_cmd[] = {
+		"Interrupt Acknowledge", "Special Cycle",
+		"I/O Read",		"I/O Write",
+		"Reserved",		"Reserved",
+		"Memory Read",		"Memory Write",
+		"Reserved",		"Reserved",
+		"Configuration Read",	"Configuration Write",
+		"Memory Read Multiple",	"Dual Address Cycle",
+		"Memory Read Line",	"Memory Write and Invalidate"
 	};
 #endif /* CONFIG_VERBOSE_MCHECK */
 
@@ -273,11 +278,11 @@ titan_parse_p_agperror(int which, u64 agperror, int print)
 	int cmd, len;
 	unsigned long addr;
 
-	char *agperror_cmd[] = { "Read (low-priority)",	"Read (high-priority)",
-				 "Write (low-priority)",
-				 "Write (high-priority)",
-				 "Reserved",		"Reserved",
-				 "Flush",		"Fence"
+	static const char * const agperror_cmd[] = {
+		"Read (low-priority)",	"Read (high-priority)",
+		"Write (low-priority)",	"Write (high-priority)",
+		"Reserved",		"Reserved",
+		"Flush",		"Fence"
 	};
 #endif /* CONFIG_VERBOSE_MCHECK */
 
diff --git a/arch/alpha/kernel/osf_sys.c b/arch/alpha/kernel/osf_sys.c
index 5d1e6d6..547e8b8 100644
--- a/arch/alpha/kernel/osf_sys.c
+++ b/arch/alpha/kernel/osf_sys.c
@@ -15,7 +15,6 @@
 #include <linux/kernel.h>
 #include <linux/mm.h>
 #include <linux/smp.h>
-#include <linux/smp_lock.h>
 #include <linux/stddef.h>
 #include <linux/syscalls.h>
 #include <linux/unistd.h>
@@ -69,7 +68,6 @@ SYSCALL_DEFINE4(osf_set_program_attributes, unsigned long, text_start,
 {
 	struct mm_struct *mm;
 
-	lock_kernel();
 	mm = current->mm;
 	mm->end_code = bss_start + bss_len;
 	mm->start_brk = bss_start + bss_len;
@@ -78,7 +76,6 @@ SYSCALL_DEFINE4(osf_set_program_attributes, unsigned long, text_start,
 	printk("set_program_attributes(%lx %lx %lx %lx)\n",
 		text_start, text_len, bss_start, bss_len);
 #endif
-	unlock_kernel();
 	return 0;
 }
 
@@ -517,7 +514,6 @@ SYSCALL_DEFINE2(osf_proplist_syscall, enum pl_code, code,
 	long error;
 	int __user *min_buf_size_ptr;
 
-	lock_kernel();
 	switch (code) {
 	case PL_SET:
 		if (get_user(error, &args->set.nbytes))
@@ -547,7 +543,6 @@ SYSCALL_DEFINE2(osf_proplist_syscall, enum pl_code, code,
 		error = -EOPNOTSUPP;
 		break;
 	};
-	unlock_kernel();
 	return error;
 }
 
@@ -594,7 +589,7 @@ SYSCALL_DEFINE2(osf_sigstack, struct sigstack __user *, uss,
 
 SYSCALL_DEFINE3(osf_sysinfo, int, command, char __user *, buf, long, count)
 {
-	char *sysinfo_table[] = {
+	const char *sysinfo_table[] = {
 		utsname()->sysname,
 		utsname()->nodename,
 		utsname()->release,
@@ -606,7 +601,7 @@ SYSCALL_DEFINE3(osf_sysinfo, int, command, char __user *, buf, long, count)
 		"dummy",	/* secure RPC domain */
 	};
 	unsigned long offset;
-	char *res;
+	const char *res;
 	long len, err = -EINVAL;
 
 	offset = command-1;
diff --git a/arch/alpha/kernel/pci-sysfs.c b/arch/alpha/kernel/pci-sysfs.c
index 738fc82..b899e95 100644
--- a/arch/alpha/kernel/pci-sysfs.c
+++ b/arch/alpha/kernel/pci-sysfs.c
@@ -66,7 +66,7 @@ static int pci_mmap_resource(struct kobject *kobj,
 {
 	struct pci_dev *pdev = to_pci_dev(container_of(kobj,
 						       struct device, kobj));
-	struct resource *res = (struct resource *)attr->private;
+	struct resource *res = attr->private;
 	enum pci_mmap_state mmap_type;
 	struct pci_bus_region bar;
 	int i;
diff --git a/arch/alpha/kernel/signal.c b/arch/alpha/kernel/signal.c
index 0932dbb..0f6b51a 100644
--- a/arch/alpha/kernel/signal.c
+++ b/arch/alpha/kernel/signal.c
@@ -144,8 +144,7 @@ SYSCALL_DEFINE5(rt_sigaction, int, sig, const struct sigaction __user *, act,
 /*
  * Atomically swap in the new signal mask, and wait for a signal.
  */
-asmlinkage int
-do_sigsuspend(old_sigset_t mask, struct pt_regs *regs, struct switch_stack *sw)
+SYSCALL_DEFINE1(sigsuspend, old_sigset_t, mask)
 {
 	mask &= _BLOCKABLE;
 	spin_lock_irq(&current->sighand->siglock);
@@ -154,41 +153,6 @@ do_sigsuspend(old_sigset_t mask, struct pt_regs *regs, struct switch_stack *sw)
 	recalc_sigpending();
 	spin_unlock_irq(&current->sighand->siglock);
 
-	/* Indicate EINTR on return from any possible signal handler,
-	   which will not come back through here, but via sigreturn.  */
-	regs->r0 = EINTR;
-	regs->r19 = 1;
-
-	current->state = TASK_INTERRUPTIBLE;
-	schedule();
-	set_thread_flag(TIF_RESTORE_SIGMASK);
-	return -ERESTARTNOHAND;
-}
-
-asmlinkage int
-do_rt_sigsuspend(sigset_t __user *uset, size_t sigsetsize,
-		 struct pt_regs *regs, struct switch_stack *sw)
-{
-	sigset_t set;
-
-	/* XXX: Don't preclude handling different sized sigset_t's.  */
-	if (sigsetsize != sizeof(sigset_t))
-		return -EINVAL;
-	if (copy_from_user(&set, uset, sizeof(set)))
-		return -EFAULT;
-
-	sigdelsetmask(&set, ~_BLOCKABLE);
-	spin_lock_irq(&current->sighand->siglock);
-	current->saved_sigmask = current->blocked;
-	current->blocked = set;
-	recalc_sigpending();
-	spin_unlock_irq(&current->sighand->siglock);
-
-	/* Indicate EINTR on return from any possible signal handler,
-	   which will not come back through here, but via sigreturn.  */
-	regs->r0 = EINTR;
-	regs->r19 = 1;
-
 	current->state = TASK_INTERRUPTIBLE;
 	schedule();
 	set_thread_flag(TIF_RESTORE_SIGMASK);
@@ -239,6 +203,8 @@ restore_sigcontext(struct sigcontext __user *sc, struct pt_regs *regs,
 	unsigned long usp;
 	long i, err = __get_user(regs->pc, &sc->sc_pc);
 
+	current_thread_info()->restart_block.fn = do_no_restart_syscall;
+
 	sw->r26 = (unsigned long) ret_from_sys_call;
 
 	err |= __get_user(regs->r0, sc->sc_regs+0);
@@ -591,7 +557,6 @@ syscall_restart(unsigned long r0, unsigned long r19,
 		regs->pc -= 4;
 		break;
 	case ERESTART_RESTARTBLOCK:
-		current_thread_info()->restart_block.fn = do_no_restart_syscall;
 		regs->r0 = EINTR;
 		break;
 	}
diff --git a/arch/alpha/kernel/srm_env.c b/arch/alpha/kernel/srm_env.c
index 4afc1a1..f0df3fb 100644
--- a/arch/alpha/kernel/srm_env.c
+++ b/arch/alpha/kernel/srm_env.c
@@ -87,7 +87,7 @@ static int srm_env_proc_show(struct seq_file *m, void *v)
 	srm_env_t	*entry;
 	char		*page;
 
-	entry = (srm_env_t *)m->private;
+	entry = m->private;
 	page = (char *)__get_free_page(GFP_USER);
 	if (!page)
 		return -ENOMEM;
diff --git a/arch/alpha/kernel/systbls.S b/arch/alpha/kernel/systbls.S
index 09acb78..ce594ef 100644
--- a/arch/alpha/kernel/systbls.S
+++ b/arch/alpha/kernel/systbls.S
@@ -512,6 +512,9 @@ sys_call_table:
 	.quad sys_pwritev
 	.quad sys_rt_tgsigqueueinfo
 	.quad sys_perf_event_open
+	.quad sys_fanotify_init
+	.quad sys_fanotify_mark				/* 495 */
+	.quad sys_prlimit64
 
 	.size sys_call_table, . - sys_call_table
 	.type sys_call_table, @object
diff --git a/arch/alpha/kernel/time.c b/arch/alpha/kernel/time.c
index eacceb2..396af17 100644
--- a/arch/alpha/kernel/time.c
+++ b/arch/alpha/kernel/time.c
@@ -191,16 +191,16 @@ irqreturn_t timer_interrupt(int irq, void *dev)
 
 	write_sequnlock(&xtime_lock);
 
-#ifndef CONFIG_SMP
-	while (nticks--)
-		update_process_times(user_mode(get_irq_regs()));
-#endif
-
 	if (test_perf_event_pending()) {
 		clear_perf_event_pending();
 		perf_event_do_pending();
 	}
 
+#ifndef CONFIG_SMP
+	while (nticks--)
+		update_process_times(user_mode(get_irq_regs()));
+#endif
+
 	return IRQ_HANDLED;
 }
 
diff --git a/arch/alpha/kernel/traps.c b/arch/alpha/kernel/traps.c
index b14f015..0414e02 100644
--- a/arch/alpha/kernel/traps.c
+++ b/arch/alpha/kernel/traps.c
@@ -13,7 +13,6 @@
 #include <linux/sched.h>
 #include <linux/tty.h>
 #include <linux/delay.h>
-#include <linux/smp_lock.h>
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/kallsyms.h>
@@ -623,7 +622,6 @@ do_entUna(void * va, unsigned long opcode, unsigned long reg,
 		return;
 	}
 
-	lock_kernel();
 	printk("Bad unaligned kernel access at %016lx: %p %lx %lu\n",
 		pc, va, opcode, reg);
 	do_exit(SIGSEGV);
@@ -646,7 +644,6 @@ got_exception:
 	 * Yikes!  No one to forward the exception to.
 	 * Since the registers are in a weird format, dump them ourselves.
  	 */
-	lock_kernel();
 
 	printk("%s(%d): unhandled unaligned exception\n",
 	       current->comm, task_pid_nr(current));