4 files changed, 176 insertions, 30 deletions

diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 46ef3fd0663b..72cec9ecd96c 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -294,7 +294,7 @@ static inline int do_exception(struct pt_regs *regs, int access)
 	down_read(&mm->mmap_sem);
 
 #ifdef CONFIG_PGSTE
-	if (test_tsk_thread_flag(current, TIF_SIE) && S390_lowcore.gmap) {
+	if ((current->flags & PF_VCPU) && S390_lowcore.gmap) {
 		address = __gmap_fault(address,
 				     (struct gmap *) S390_lowcore.gmap);
 		if (address == -EFAULT) {
@@ -549,19 +549,15 @@ static void pfault_interrupt(struct ext_code ext_code,
 	if ((subcode & 0xff00) != __SUBCODE_MASK)
 		return;
 	kstat_cpu(smp_processor_id()).irqs[EXTINT_PFL]++;
-	if (subcode & 0x0080) {
-		/* Get the token (= pid of the affected task). */
-		pid = sizeof(void *) == 4 ? param32 : param64;
-		rcu_read_lock();
-		tsk = find_task_by_pid_ns(pid, &init_pid_ns);
-		if (tsk)
-			get_task_struct(tsk);
-		rcu_read_unlock();
-		if (!tsk)
-			return;
-	} else {
-		tsk = current;
-	}
+	/* Get the token (= pid of the affected task). */
+	pid = sizeof(void *) == 4 ? param32 : param64;
+	rcu_read_lock();
+	tsk = find_task_by_pid_ns(pid, &init_pid_ns);
+	if (tsk)
+		get_task_struct(tsk);
+	rcu_read_unlock();
+	if (!tsk)
+		return;
 	spin_lock(&pfault_lock);
 	if (subcode & 0x0080) {
 		/* signal bit is set -> a page has been swapped in by VM */
@@ -574,6 +570,7 @@ static void pfault_interrupt(struct ext_code ext_code,
 			tsk->thread.pfault_wait = 0;
 			list_del(&tsk->thread.list);
 			wake_up_process(tsk);
+			put_task_struct(tsk);
 		} else {
 			/* Completion interrupt was faster than initial
 			 * interrupt. Set pfault_wait to -1 so the initial
@@ -585,24 +582,35 @@ static void pfault_interrupt(struct ext_code ext_code,
 			if (tsk->state == TASK_RUNNING)
 				tsk->thread.pfault_wait = -1;
 		}
-		put_task_struct(tsk);
 	} else {
 		/* signal bit not set -> a real page is missing. */
-		if (tsk->thread.pfault_wait == -1) {
+		if (WARN_ON_ONCE(tsk != current))
+			goto out;
+		if (tsk->thread.pfault_wait == 1) {
+			/* Already on the list with a reference: put to sleep */
+			__set_task_state(tsk, TASK_UNINTERRUPTIBLE);
+			set_tsk_need_resched(tsk);
+		} else if (tsk->thread.pfault_wait == -1) {
 			/* Completion interrupt was faster than the initial
 			 * interrupt (pfault_wait == -1). Set pfault_wait
 			 * back to zero and exit. */
 			tsk->thread.pfault_wait = 0;
 		} else {
 			/* Initial interrupt arrived before completion
-			 * interrupt. Let the task sleep. */
+			 * interrupt. Let the task sleep.
+			 * An extra task reference is needed since a different
+			 * cpu may set the task state to TASK_RUNNING again
+			 * before the scheduler is reached. */
+			get_task_struct(tsk);
 			tsk->thread.pfault_wait = 1;
 			list_add(&tsk->thread.list, &pfault_list);
-			set_task_state(tsk, TASK_UNINTERRUPTIBLE);
+			__set_task_state(tsk, TASK_UNINTERRUPTIBLE);
 			set_tsk_need_resched(tsk);
 		}
 	}
+out:
 	spin_unlock(&pfault_lock);
+	put_task_struct(tsk);
 }
 
 static int __cpuinit pfault_cpu_notify(struct notifier_block *self,
@@ -620,6 +628,7 @@ static int __cpuinit pfault_cpu_notify(struct notifier_block *self,
 			list_del(&thread->list);
 			tsk = container_of(thread, struct task_struct, thread);
 			wake_up_process(tsk);
+			put_task_struct(tsk);
 		}
 		spin_unlock_irq(&pfault_lock);
 		break;
diff --git a/arch/s390/mm/hugetlbpage.c b/arch/s390/mm/hugetlbpage.c
index 597bb2d27c3c..900de2b3cf28 100644
--- a/arch/s390/mm/hugetlbpage.c
+++ b/arch/s390/mm/hugetlbpage.c
@@ -58,6 +58,8 @@ void arch_release_hugepage(struct page *page)
 	ptep = (pte_t *) page[1].index;
 	if (!ptep)
 		return;
+	clear_table((unsigned long *) ptep, _PAGE_TYPE_EMPTY,
+		    PTRS_PER_PTE * sizeof(pte_t));
 	page_table_free(&init_mm, (unsigned long *) ptep);
 	page[1].index = 0;
 }
diff --git a/arch/s390/mm/maccess.c b/arch/s390/mm/maccess.c
index 7bb15fcca75e..795a0a9bb2eb 100644
--- a/arch/s390/mm/maccess.c
+++ b/arch/s390/mm/maccess.c
@@ -12,6 +12,7 @@
 #include <linux/types.h>
 #include <linux/errno.h>
 #include <linux/gfp.h>
+#include <linux/cpu.h>
 #include <asm/ctl_reg.h>
 
 /*
@@ -61,21 +62,14 @@ long probe_kernel_write(void *dst, const void *src, size_t size)
 	return copied < 0 ? -EFAULT : 0;
 }
 
-/*
- * Copy memory in real mode (kernel to kernel)
- */
-int memcpy_real(void *dest, void *src, size_t count)
+static int __memcpy_real(void *dest, void *src, size_t count)
 {
 	register unsigned long _dest asm("2") = (unsigned long) dest;
 	register unsigned long _len1 asm("3") = (unsigned long) count;
 	register unsigned long _src  asm("4") = (unsigned long) src;
 	register unsigned long _len2 asm("5") = (unsigned long) count;
-	unsigned long flags;
 	int rc = -EFAULT;
 
-	if (!count)
-		return 0;
-	flags = __arch_local_irq_stnsm(0xf8UL);
 	asm volatile (
 		"0:	mvcle	%1,%2,0x0\n"
 		"1:	jo	0b\n"
@@ -86,7 +80,23 @@ int memcpy_real(void *dest, void *src, size_t count)
 		  "+d" (_len2), "=m" (*((long *) dest))
 		: "m" (*((long *) src))
 		: "cc", "memory");
-	arch_local_irq_restore(flags);
+	return rc;
+}
+
+/*
+ * Copy memory in real mode (kernel to kernel)
+ */
+int memcpy_real(void *dest, void *src, size_t count)
+{
+	unsigned long flags;
+	int rc;
+
+	if (!count)
+		return 0;
+	local_irq_save(flags);
+	__arch_local_irq_stnsm(0xfbUL);
+	rc = __memcpy_real(dest, src, count);
+	local_irq_restore(flags);
 	return rc;
 }
 
@@ -157,3 +167,69 @@ out:
 	free_page((unsigned long) buf);
 	return rc;
 }
+
+/*
+ * Check if physical address is within prefix or zero page
+ */
+static int is_swapped(unsigned long addr)
+{
+	unsigned long lc;
+	int cpu;
+
+	if (addr < sizeof(struct _lowcore))
+		return 1;
+	for_each_online_cpu(cpu) {
+		lc = (unsigned long) lowcore_ptr[cpu];
+		if (addr > lc + sizeof(struct _lowcore) - 1 || addr < lc)
+			continue;
+		return 1;
+	}
+	return 0;
+}
+
+/*
+ * Return swapped prefix or zero page address
+ */
+static unsigned long get_swapped(unsigned long addr)
+{
+	unsigned long prefix = store_prefix();
+
+	if (addr < sizeof(struct _lowcore))
+		return addr + prefix;
+	if (addr >= prefix && addr < prefix + sizeof(struct _lowcore))
+		return addr - prefix;
+	return addr;
+}
+
+/*
+ * Convert a physical pointer for /dev/mem access
+ *
+ * For swapped prefix pages a new buffer is returned that contains a copy of
+ * the absolute memory. The buffer size is maximum one page large.
+ */
+void *xlate_dev_mem_ptr(unsigned long addr)
+{
+	void *bounce = (void *) addr;
+	unsigned long size;
+
+	get_online_cpus();
+	preempt_disable();
+	if (is_swapped(addr)) {
+		size = PAGE_SIZE - (addr & ~PAGE_MASK);
+		bounce = (void *) __get_free_page(GFP_ATOMIC);
+		if (bounce)
+			memcpy_real(bounce, (void *) get_swapped(addr), size);
+	}
+	preempt_enable();
+	put_online_cpus();
+	return bounce;
+}
+
+/*
+ * Free converted buffer for /dev/mem access (if necessary)
+ */
+void unxlate_dev_mem_ptr(unsigned long addr, void *buf)
+{
+	if ((void *) addr != buf)
+		free_page((unsigned long) buf);
+}
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index 373adf69b01c..a3db5a3ea083 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -678,8 +678,6 @@ void page_table_free(struct mm_struct *mm, unsigned long *table)
 	}
 }
 
-#ifdef CONFIG_HAVE_RCU_TABLE_FREE
-
 static void __page_table_free_rcu(void *table, unsigned bit)
 {
 	struct page *page;
@@ -733,7 +731,66 @@ void __tlb_remove_table(void *_table)
 		free_pages((unsigned long) table, ALLOC_ORDER);
 }
 
-#endif
+static void tlb_remove_table_smp_sync(void *arg)
+{
+	/* Simply deliver the interrupt */
+}
+
+static void tlb_remove_table_one(void *table)
+{
+	/*
+	 * This isn't an RCU grace period and hence the page-tables cannot be
+	 * assumed to be actually RCU-freed.
+	 *
+	 * It is however sufficient for software page-table walkers that rely
+	 * on IRQ disabling. See the comment near struct mmu_table_batch.
+	 */
+	smp_call_function(tlb_remove_table_smp_sync, NULL, 1);
+	__tlb_remove_table(table);
+}
+
+static void tlb_remove_table_rcu(struct rcu_head *head)
+{
+	struct mmu_table_batch *batch;
+	int i;
+
+	batch = container_of(head, struct mmu_table_batch, rcu);
+
+	for (i = 0; i < batch->nr; i++)
+		__tlb_remove_table(batch->tables[i]);
+
+	free_page((unsigned long)batch);
+}
+
+void tlb_table_flush(struct mmu_gather *tlb)
+{
+	struct mmu_table_batch **batch = &tlb->batch;
+
+	if (*batch) {
+		__tlb_flush_mm(tlb->mm);
+		call_rcu_sched(&(*batch)->rcu, tlb_remove_table_rcu);
+		*batch = NULL;
+	}
+}
+
+void tlb_remove_table(struct mmu_gather *tlb, void *table)
+{
+	struct mmu_table_batch **batch = &tlb->batch;
+
+	if (*batch == NULL) {
+		*batch = (struct mmu_table_batch *)
+			__get_free_page(GFP_NOWAIT | __GFP_NOWARN);
+		if (*batch == NULL) {
+			__tlb_flush_mm(tlb->mm);
+			tlb_remove_table_one(table);
+			return;
+		}
+		(*batch)->nr = 0;
+	}
+	(*batch)->tables[(*batch)->nr++] = table;
+	if ((*batch)->nr == MAX_TABLE_BATCH)
+		tlb_table_flush(tlb);
+}
 
 /*
  * switch on pgstes for its userspace process (for kvm)
@@ -765,6 +822,8 @@ int s390_enable_sie(void)
 
 	/* we copy the mm and let dup_mm create the page tables with_pgstes */
 	tsk->mm->context.alloc_pgste = 1;
+	/* make sure that both mms have a correct rss state */
+	sync_mm_rss(tsk->mm);
 	mm = dup_mm(tsk);
 	tsk->mm->context.alloc_pgste = 0;
 	if (!mm)