Merge branch 'percpu-for-linus' into percpu-for-next

Conflicts:
	arch/sparc/kernel/smp_64.c
	arch/x86/kernel/cpu/perf_counter.c
	arch/x86/kernel/setup_percpu.c
	drivers/cpufreq/cpufreq_ondemand.c
	mm/percpu.c

Conflicts in core and arch percpu codes are mostly from commit
ed78e1e078dd44249f88b1dd8c76dafb39567161 which substituted many
num_possible_cpus() with nr_cpu_ids.  As for-next branch has moved all
the first chunk allocators into mm/percpu.c, the changes are moved
from arch code to mm/percpu.c.

Signed-off-by: Tejun Heo <tj@kernel.org>

19 files changed, 117 insertions, 64 deletions

diff --git a/arch/s390/Kconfig b/arch/s390/Kconfig
index e577839f307..2ae5d72f47e 100644
--- a/arch/s390/Kconfig
+++ b/arch/s390/Kconfig
@@ -95,6 +95,11 @@ config S390
 	select HAVE_ARCH_TRACEHOOK
 	select INIT_ALL_POSSIBLE
 	select HAVE_PERF_COUNTERS
+	select GENERIC_ATOMIC64 if !64BIT
+
+config SCHED_OMIT_FRAME_POINTER
+	bool
+	default y
 
 source "init/Kconfig"
 
@@ -116,6 +121,9 @@ config 32BIT
 	bool
 	default y if !64BIT
 
+config KTIME_SCALAR
+	def_bool 32BIT
+
 config SMP
 	bool "Symmetric multi-processing support"
 	---help---
diff --git a/arch/s390/include/asm/atomic.h b/arch/s390/include/asm/atomic.h
index fca9dffcc66..c7d0abfb0f0 100644
--- a/arch/s390/include/asm/atomic.h
+++ b/arch/s390/include/asm/atomic.h
@@ -268,7 +268,12 @@ static __inline__ int atomic64_add_unless(atomic64_t *v,
 #define atomic64_inc_not_zero(v) atomic64_add_unless((v), 1, 0)
 
 #undef __CSG_LOOP
-#endif
+
+#else /* __s390x__ */
+
+#include <asm-generic/atomic64.h>
+
+#endif /* __s390x__ */
 
 #define smp_mb__before_atomic_dec()	smp_mb()
 #define smp_mb__after_atomic_dec()	smp_mb()
diff --git a/arch/s390/include/asm/perf_counter.h b/arch/s390/include/asm/perf_counter.h
index a7205a3828c..7015188c2cc 100644
--- a/arch/s390/include/asm/perf_counter.h
+++ b/arch/s390/include/asm/perf_counter.h
@@ -6,3 +6,5 @@
 
 static inline void set_perf_counter_pending(void) {}
 static inline void clear_perf_counter_pending(void) {}
+
+#define PERF_COUNTER_INDEX_OFFSET 0
diff --git a/arch/s390/include/asm/thread_info.h b/arch/s390/include/asm/thread_info.h
index 925bcc64903..ba1cab9fc1f 100644
--- a/arch/s390/include/asm/thread_info.h
+++ b/arch/s390/include/asm/thread_info.h
@@ -61,7 +61,7 @@ struct thread_info {
 	.exec_domain	= &default_exec_domain,	\
 	.flags		= 0,			\
 	.cpu		= 0,			\
-	.preempt_count	= 1,			\
+	.preempt_count	= INIT_PREEMPT_COUNT,	\
 	.restart_block	= {			\
 		.fn = do_no_restart_syscall,	\
 	},					\
diff --git a/arch/s390/include/asm/tlb.h b/arch/s390/include/asm/tlb.h
index 3d8a96d39d9..81150b05368 100644
--- a/arch/s390/include/asm/tlb.h
+++ b/arch/s390/include/asm/tlb.h
@@ -96,7 +96,8 @@ static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page)
  * pte_free_tlb frees a pte table and clears the CRSTE for the
  * page table from the tlb.
  */
-static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte)
+static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte,
+				unsigned long address)
 {
 	if (!tlb->fullmm) {
 		tlb->array[tlb->nr_ptes++] = pte;
@@ -113,7 +114,8 @@ static inline void pte_free_tlb(struct mmu_gather *tlb, pgtable_t pte)
  * as the pgd. pmd_free_tlb checks the asce_limit against 2GB
  * to avoid the double free of the pmd in this case.
  */
-static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
+static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd,
+				unsigned long address)
 {
 #ifdef __s390x__
 	if (tlb->mm->context.asce_limit <= (1UL << 31))
@@ -134,7 +136,8 @@ static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
  * as the pgd. pud_free_tlb checks the asce_limit against 4TB
  * to avoid the double free of the pud in this case.
  */
-static inline void pud_free_tlb(struct mmu_gather *tlb, pud_t *pud)
+static inline void pud_free_tlb(struct mmu_gather *tlb, pud_t *pud,
+				unsigned long address)
 {
 #ifdef __s390x__
 	if (tlb->mm->context.asce_limit <= (1UL << 42))
diff --git a/arch/s390/kernel/dis.c b/arch/s390/kernel/dis.c
index d2f270c995d..db943a7ec51 100644
--- a/arch/s390/kernel/dis.c
+++ b/arch/s390/kernel/dis.c
@@ -15,7 +15,6 @@
 #include <linux/timer.h>
 #include <linux/mm.h>
 #include <linux/smp.h>
-#include <linux/smp_lock.h>
 #include <linux/init.h>
 #include <linux/interrupt.h>
 #include <linux/delay.h>
diff --git a/arch/s390/kernel/early.c b/arch/s390/kernel/early.c
index f9b144049dc..cae14c49951 100644
--- a/arch/s390/kernel/early.c
+++ b/arch/s390/kernel/early.c
@@ -208,9 +208,12 @@ static noinline __init void detect_machine_type(void)
 		machine_flags |= MACHINE_FLAG_KVM;
 	else
 		machine_flags |= MACHINE_FLAG_VM;
+
+	/* Store machine flags for setting up lowcore early */
+	S390_lowcore.machine_flags = machine_flags;
 }
 
-static void early_pgm_check_handler(void)
+static __init void early_pgm_check_handler(void)
 {
 	unsigned long addr;
 	const struct exception_table_entry *fixup;
@@ -222,7 +225,7 @@ static void early_pgm_check_handler(void)
 	S390_lowcore.program_old_psw.addr = fixup->fixup | PSW_ADDR_AMODE;
 }
 
-void setup_lowcore_early(void)
+static noinline __init void setup_lowcore_early(void)
 {
 	psw_t psw;
 
diff --git a/arch/s390/kernel/ipl.c b/arch/s390/kernel/ipl.c
index b8bf4b14006..371a2d88f4a 100644
--- a/arch/s390/kernel/ipl.c
+++ b/arch/s390/kernel/ipl.c
@@ -70,6 +70,7 @@ struct shutdown_action {
 	char *name;
 	void (*fn) (struct shutdown_trigger *trigger);
 	int (*init) (void);
+	int init_rc;
 };
 
 static char *ipl_type_str(enum ipl_type type)
@@ -1486,11 +1487,13 @@ static int set_trigger(const char *buf, struct shutdown_trigger *trigger,
 	int i;
 
 	for (i = 0; i < SHUTDOWN_ACTIONS_COUNT; i++) {
-		if (!shutdown_actions_list[i])
-			continue;
 		if (sysfs_streq(buf, shutdown_actions_list[i]->name)) {
-			trigger->action = shutdown_actions_list[i];
-			return len;
+			if (shutdown_actions_list[i]->init_rc) {
+				return shutdown_actions_list[i]->init_rc;
+			} else {
+				trigger->action = shutdown_actions_list[i];
+				return len;
+			}
 		}
 	}
 	return -EINVAL;
@@ -1640,8 +1643,8 @@ static void __init shutdown_actions_init(void)
 	for (i = 0; i < SHUTDOWN_ACTIONS_COUNT; i++) {
 		if (!shutdown_actions_list[i]->init)
 			continue;
-		if (shutdown_actions_list[i]->init())
-			shutdown_actions_list[i] = NULL;
+		shutdown_actions_list[i]->init_rc =
+			shutdown_actions_list[i]->init();
 	}
 }
 
diff --git a/arch/s390/kernel/ptrace.c b/arch/s390/kernel/ptrace.c
index 490b39934d6..43acd73105b 100644
--- a/arch/s390/kernel/ptrace.c
+++ b/arch/s390/kernel/ptrace.c
@@ -26,7 +26,6 @@
 #include <linux/sched.h>
 #include <linux/mm.h>
 #include <linux/smp.h>
-#include <linux/smp_lock.h>
 #include <linux/errno.h>
 #include <linux/ptrace.h>
 #include <linux/user.h>
diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 2270730f535..be2cae08340 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -687,13 +687,14 @@ void __init smp_prepare_cpus(unsigned int max_cpus)
 #ifndef CONFIG_64BIT
 	if (MACHINE_HAS_IEEE)
 		lowcore->extended_save_area_addr = (u32) save_area;
-#else
-	if (vdso_alloc_per_cpu(smp_processor_id(), lowcore))
-		BUG();
 #endif
 	set_prefix((u32)(unsigned long) lowcore);
 	local_mcck_enable();
 	local_irq_enable();
+#ifdef CONFIG_64BIT
+	if (vdso_alloc_per_cpu(smp_processor_id(), &S390_lowcore))
+		BUG();
+#endif
 	for_each_possible_cpu(cpu)
 		if (cpu != smp_processor_id())
 			smp_create_idle(cpu);
diff --git a/arch/s390/kernel/vdso64/clock_gettime.S b/arch/s390/kernel/vdso64/clock_gettime.S
index 79dbfee831e..49106c6e6f8 100644
--- a/arch/s390/kernel/vdso64/clock_gettime.S
+++ b/arch/s390/kernel/vdso64/clock_gettime.S
@@ -88,10 +88,17 @@ __kernel_clock_gettime:
 	llilh	%r4,0x0100
 	sar	%a4,%r4
 	lghi	%r4,0
+	epsw	%r5,0
 	sacf	512				/* Magic ectg instruction */
 	.insn	ssf,0xc80100000000,__VDSO_ECTG_BASE(4),__VDSO_ECTG_USER(4),4
-	sacf	0
-	sar	%a4,%r2
+	tml	%r5,0x4000
+	jo	11f
+	tml	%r5,0x8000
+	jno	10f
+	sacf	256
+	j	11f
+10:	sacf	0
+11:	sar	%a4,%r2
 	algr	%r1,%r0				/* r1 = cputime as TOD value */
 	mghi	%r1,1000			/* convert to nanoseconds */
 	srlg	%r1,%r1,12			/* r1 = cputime in nanosec */
diff --git a/arch/s390/kvm/interrupt.c b/arch/s390/kvm/interrupt.c
index f04f5301b1b..4d613415c43 100644
--- a/arch/s390/kvm/interrupt.c
+++ b/arch/s390/kvm/interrupt.c
@@ -386,7 +386,7 @@ no_timer:
 	}
 	__unset_cpu_idle(vcpu);
 	__set_current_state(TASK_RUNNING);
-	remove_wait_queue(&vcpu->wq, &wait);
+	remove_wait_queue(&vcpu->arch.local_int.wq, &wait);
 	spin_unlock_bh(&vcpu->arch.local_int.lock);
 	spin_unlock(&vcpu->arch.local_int.float_int->lock);
 	hrtimer_try_to_cancel(&vcpu->arch.ckc_timer);
diff --git a/arch/s390/kvm/sigp.c b/arch/s390/kvm/sigp.c
index 36678835034..0ef81d6776e 100644
--- a/arch/s390/kvm/sigp.c
+++ b/arch/s390/kvm/sigp.c
@@ -169,7 +169,7 @@ static int __sigp_set_prefix(struct kvm_vcpu *vcpu, u16 cpu_addr, u32 address,
 			     unsigned long *reg)
 {
 	struct kvm_s390_float_interrupt *fi = &vcpu->kvm->arch.float_int;
-	struct kvm_s390_local_interrupt *li;
+	struct kvm_s390_local_interrupt *li = NULL;
 	struct kvm_s390_interrupt_info *inti;
 	int rc;
 	u8 tmp;
@@ -189,9 +189,10 @@ static int __sigp_set_prefix(struct kvm_vcpu *vcpu, u16 cpu_addr, u32 address,
 		return 2; /* busy */
 
 	spin_lock(&fi->lock);
-	li = fi->local_int[cpu_addr];
+	if (cpu_addr < KVM_MAX_VCPUS)
+		li = fi->local_int[cpu_addr];
 
-	if ((cpu_addr >= KVM_MAX_VCPUS) || (li == NULL)) {
+	if (li == NULL) {
 		rc = 1; /* incorrect state */
 		*reg &= SIGP_STAT_INCORRECT_STATE;
 		kfree(inti);
diff --git a/arch/s390/lib/Makefile b/arch/s390/lib/Makefile
index ab6735df2d2..97975ec7a27 100644
--- a/arch/s390/lib/Makefile
+++ b/arch/s390/lib/Makefile
@@ -3,6 +3,6 @@
 #
 
 lib-y += delay.o string.o uaccess_std.o uaccess_pt.o
-obj-$(CONFIG_32BIT) += div64.o qrnnd.o
+obj-$(CONFIG_32BIT) += div64.o qrnnd.o ucmpdi2.o
 lib-$(CONFIG_64BIT) += uaccess_mvcos.o
 lib-$(CONFIG_SMP) += spinlock.o
diff --git a/arch/s390/lib/delay.c b/arch/s390/lib/delay.c
index 3f5f680726e..97c1eca83cc 100644
--- a/arch/s390/lib/delay.c
+++ b/arch/s390/lib/delay.c
@@ -36,9 +36,11 @@ static void __udelay_disabled(unsigned long usecs)
 	cr0 = (cr0_saved & 0xffff00e0) | 0x00000800;
 	__ctl_load(cr0 , 0, 0);
 	mask = psw_kernel_bits | PSW_MASK_WAIT | PSW_MASK_EXT;
+	lockdep_off();
 	trace_hardirqs_on();
 	__load_psw_mask(mask);
 	local_irq_disable();
+	lockdep_on();
 	__ctl_load(cr0_saved, 0, 0);
 	local_tick_enable(clock_saved);
 	set_clock_comparator(S390_lowcore.clock_comparator);
diff --git a/arch/s390/lib/ucmpdi2.c b/arch/s390/lib/ucmpdi2.c
new file mode 100644
index 00000000000..3e05ff53258
--- /dev/null
+++ b/arch/s390/lib/ucmpdi2.c
@@ -0,0 +1,26 @@
+#include <linux/module.h>
+
+union ull_union {
+	unsigned long long ull;
+	struct {
+		unsigned int high;
+		unsigned int low;
+	} ui;
+};
+
+int __ucmpdi2(unsigned long long a, unsigned long long b)
+{
+	union ull_union au = {.ull = a};
+	union ull_union bu = {.ull = b};
+
+	if (au.ui.high < bu.ui.high)
+		return 0;
+	else if (au.ui.high > bu.ui.high)
+		return 2;
+	if (au.ui.low < bu.ui.low)
+		return 0;
+	else if (au.ui.low > bu.ui.low)
+		return 2;
+	return 1;
+}
+EXPORT_SYMBOL(__ucmpdi2);
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 74eb26bf197..e5e119fe03b 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -22,7 +22,6 @@
 #include <linux/compat.h>
 #include <linux/smp.h>
 #include <linux/kdebug.h>
-#include <linux/smp_lock.h>
 #include <linux/init.h>
 #include <linux/console.h>
 #include <linux/module.h>
diff --git a/arch/s390/power/swsusp.c b/arch/s390/power/swsusp.c
index e6a4fe9f5f2..bd1f5c6b0b8 100644
--- a/arch/s390/power/swsusp.c
+++ b/arch/s390/power/swsusp.c
@@ -7,24 +7,36 @@
  *
  */
 
+#include <asm/system.h>
 
-/*
- * save CPU registers before creating a hibernation image and before
- * restoring the memory state from it
- */
 void save_processor_state(void)
 {
-	/* implentation contained in the
-	 * swsusp_arch_suspend function
+	/* swsusp_arch_suspend() actually saves all cpu register contents.
+	 * Machine checks must be disabled since swsusp_arch_suspend() stores
+	 * register contents to their lowcore save areas. That's the same
+	 * place where register contents on machine checks would be saved.
+	 * To avoid register corruption disable machine checks.
+	 * We must also disable machine checks in the new psw mask for
+	 * program checks, since swsusp_arch_suspend() may generate program
+	 * checks. Disabling machine checks for all other new psw masks is
+	 * just paranoia.
 	 */
+	local_mcck_disable();
+	/* Disable lowcore protection */
+	__ctl_clear_bit(0,28);
+	S390_lowcore.external_new_psw.mask &= ~PSW_MASK_MCHECK;
+	S390_lowcore.svc_new_psw.mask &= ~PSW_MASK_MCHECK;
+	S390_lowcore.io_new_psw.mask &= ~PSW_MASK_MCHECK;
+	S390_lowcore.program_new_psw.mask &= ~PSW_MASK_MCHECK;
 }
 
-/*
- * restore the contents of CPU registers
- */
 void restore_processor_state(void)
 {
-	/* implentation contained in the
-	 * swsusp_arch_resume function
-	 */
+	S390_lowcore.external_new_psw.mask |= PSW_MASK_MCHECK;
+	S390_lowcore.svc_new_psw.mask |= PSW_MASK_MCHECK;
+	S390_lowcore.io_new_psw.mask |= PSW_MASK_MCHECK;
+	S390_lowcore.program_new_psw.mask |= PSW_MASK_MCHECK;
+	/* Enable lowcore protection */
+	__ctl_set_bit(0,28);
+	local_mcck_enable();
 }
diff --git a/arch/s390/power/swsusp_asm64.S b/arch/s390/power/swsusp_asm64.S
index 76d688da32f..b26df5c5933 100644
--- a/arch/s390/power/swsusp_asm64.S
+++ b/arch/s390/power/swsusp_asm64.S
@@ -32,19 +32,14 @@ swsusp_arch_suspend:
 	/* Deactivate DAT */
 	stnsm	__SF_EMPTY(%r15),0xfb
 
-	/* Switch off lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	ni	__SF_EMPTY+4(%r15),0xef
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Store prefix register on stack */
 	stpx	__SF_EMPTY(%r15)
 
-	/* Setup base register for lowcore (absolute 0) */
-	llgf	%r1,__SF_EMPTY(%r15)
+	/* Save prefix register contents for lowcore */
+	llgf	%r4,__SF_EMPTY(%r15)
 
 	/* Get pointer to save area */
-	aghi	%r1,0x1000
+	lghi	%r1,0x1000
 
 	/* Store registers */
 	mvc	0x318(4,%r1),__SF_EMPTY(%r15)	/* move prefix to lowcore */
@@ -79,17 +74,15 @@ swsusp_arch_suspend:
 	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
 	spx	__SF_EMPTY(%r15)
 
-	/* Setup lowcore */
-	brasl	%r14,setup_lowcore_early
+	lghi	%r2,0
+	lghi	%r3,2*PAGE_SIZE
+	lghi	%r5,2*PAGE_SIZE
+1:	mvcle	%r2,%r4,0
+	jo	1b
 
 	/* Save image */
 	brasl	%r14,swsusp_save
 
-	/* Switch on lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	oi	__SF_EMPTY+4(%r15),0x10
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Restore prefix register and return */
 	lghi	%r1,0x1000
 	spx	0x318(%r1)
@@ -117,11 +110,6 @@ swsusp_arch_resume:
 	/* Deactivate DAT */
 	stnsm	__SF_EMPTY(%r15),0xfb
 
-	/* Switch off lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	ni	__SF_EMPTY+4(%r15),0xef
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Set prefix page to zero */
 	xc	__SF_EMPTY(4,%r15),__SF_EMPTY(%r15)
 	spx	__SF_EMPTY(%r15)
@@ -175,7 +163,7 @@ swsusp_arch_resume:
 	/* Load old stack */
 	lg	%r15,0x2f8(%r13)
 
-	/* Pointer to save arae */
+	/* Pointer to save area */
 	lghi	%r13,0x1000
 
 #ifdef CONFIG_SMP
@@ -187,11 +175,6 @@ swsusp_arch_resume:
 	/* Restore prefix register */
 	spx	0x318(%r13)
 
-	/* Switch on lowcore protection */
-	stctg	%c0,%c0,__SF_EMPTY(%r15)
-	oi	__SF_EMPTY+4(%r15),0x10
-	lctlg	%c0,%c0,__SF_EMPTY(%r15)
-
 	/* Activate DAT */
 	stosm	__SF_EMPTY(%r15),0x04