Merge commit '8700c95adb03' into timers/nohz

The full dynticks tree needs the latest RCU and sched
upstream updates in order to fix some dependencies.

Merge a common upstream merge point that has these
updates.

Conflicts:
	include/linux/perf_event.h
	kernel/rcutree.h
	kernel/rcutree_plugin.h

Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>

7 files changed, 201 insertions, 137 deletions

diff --git a/arch/s390/mm/cmm.c b/arch/s390/mm/cmm.c
index 479e942..9d84a1f 100644
--- a/arch/s390/mm/cmm.c
+++ b/arch/s390/mm/cmm.c
@@ -458,12 +458,10 @@ static int __init cmm_init(void)
 	if (rc)
 		goto out_pm;
 	cmm_thread_ptr = kthread_run(cmm_thread, NULL, "cmmthread");
-	rc = IS_ERR(cmm_thread_ptr) ? PTR_ERR(cmm_thread_ptr) : 0;
-	if (rc)
-		goto out_kthread;
-	return 0;
+	if (!IS_ERR(cmm_thread_ptr))
+		return 0;
 
-out_kthread:
+	rc = PTR_ERR(cmm_thread_ptr);
 	unregister_pm_notifier(&cmm_power_notifier);
 out_pm:
 	unregister_oom_notifier(&cmm_oom_nb);
diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index 2fb9e63..047c3e4 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -395,8 +395,13 @@ void __kprobes do_protection_exception(struct pt_regs *regs)
 	int fault;
 
 	trans_exc_code = regs->int_parm_long;
-	/* Protection exception is suppressing, decrement psw address. */
-	regs->psw.addr = __rewind_psw(regs->psw, regs->int_code >> 16);
+	/*
+	 * Protection exceptions are suppressing, decrement psw address.
+	 * The exception to this rule are aborted transactions, for these
+	 * the PSW already points to the correct location.
+	 */
+	if (!(regs->int_code & 0x200))
+		regs->psw.addr = __rewind_psw(regs->psw, regs->int_code >> 16);
 	/*
 	 * Check for low-address protection.  This needs to be treated
 	 * as a special case because the translation exception code
diff --git a/arch/s390/mm/hugetlbpage.c b/arch/s390/mm/hugetlbpage.c
index 532525e..121089d 100644
--- a/arch/s390/mm/hugetlbpage.c
+++ b/arch/s390/mm/hugetlbpage.c
@@ -39,7 +39,7 @@ int arch_prepare_hugepage(struct page *page)
 	if (!ptep)
 		return -ENOMEM;
 
-	pte = mk_pte(page, PAGE_RW);
+	pte_val(pte) = addr;
 	for (i = 0; i < PTRS_PER_PTE; i++) {
 		set_pte_at(&init_mm, addr + i * PAGE_SIZE, ptep + i, pte);
 		pte_val(pte) += PAGE_SIZE;
diff --git a/arch/s390/mm/init.c b/arch/s390/mm/init.c
index 49ce6bb..0b09b23 100644
--- a/arch/s390/mm/init.c
+++ b/arch/s390/mm/init.c
@@ -42,11 +42,10 @@ pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__((__aligned__(PAGE_SIZE)));
 unsigned long empty_zero_page, zero_page_mask;
 EXPORT_SYMBOL(empty_zero_page);
 
-static unsigned long __init setup_zero_pages(void)
+static void __init setup_zero_pages(void)
 {
 	struct cpuid cpu_id;
 	unsigned int order;
-	unsigned long size;
 	struct page *page;
 	int i;
 
@@ -63,10 +62,18 @@ static unsigned long __init setup_zero_pages(void)
 		break;
 	case 0x2097:	/* z10 */
 	case 0x2098:	/* z10 */
-	default:
+	case 0x2817:	/* z196 */
+	case 0x2818:	/* z196 */
 		order = 2;
 		break;
+	case 0x2827:	/* zEC12 */
+	default:
+		order = 5;
+		break;
 	}
+	/* Limit number of empty zero pages for small memory sizes */
+	if (order > 2 && totalram_pages <= 16384)
+		order = 2;
 
 	empty_zero_page = __get_free_pages(GFP_KERNEL | __GFP_ZERO, order);
 	if (!empty_zero_page)
@@ -75,14 +82,11 @@ static unsigned long __init setup_zero_pages(void)
 	page = virt_to_page((void *) empty_zero_page);
 	split_page(page, order);
 	for (i = 1 << order; i > 0; i--) {
-		SetPageReserved(page);
+		mark_page_reserved(page);
 		page++;
 	}
 
-	size = PAGE_SIZE << order;
-	zero_page_mask = (size - 1) & PAGE_MASK;
-
-	return 1UL << order;
+	zero_page_mask = ((PAGE_SIZE << order) - 1) & PAGE_MASK;
 }
 
 /*
@@ -139,7 +143,7 @@ void __init mem_init(void)
 
 	/* this will put all low memory onto the freelists */
 	totalram_pages += free_all_bootmem();
-	totalram_pages -= setup_zero_pages();	/* Setup zeroed pages. */
+	setup_zero_pages();	/* Setup zeroed pages. */
 
 	reservedpages = 0;
 
@@ -158,34 +162,15 @@ void __init mem_init(void)
 	       PFN_ALIGN((unsigned long)&_eshared) - 1);
 }
 
-void free_init_pages(char *what, unsigned long begin, unsigned long end)
-{
-	unsigned long addr = begin;
-
-	if (begin >= end)
-		return;
-	for (; addr < end; addr += PAGE_SIZE) {
-		ClearPageReserved(virt_to_page(addr));
-		init_page_count(virt_to_page(addr));
-		memset((void *)(addr & PAGE_MASK), POISON_FREE_INITMEM,
-		       PAGE_SIZE);
-		free_page(addr);
-		totalram_pages++;
-	}
-	printk(KERN_INFO "Freeing %s: %luk freed\n", what, (end - begin) >> 10);
-}
-
 void free_initmem(void)
 {
-	free_init_pages("unused kernel memory",
-			(unsigned long)&__init_begin,
-			(unsigned long)&__init_end);
+	free_initmem_default(0);
 }
 
 #ifdef CONFIG_BLK_DEV_INITRD
 void __init free_initrd_mem(unsigned long start, unsigned long end)
 {
-	free_init_pages("initrd memory", start, end);
+	free_reserved_area(start, end, POISON_FREE_INITMEM, "initrd");
 }
 #endif
 
diff --git a/arch/s390/mm/pageattr.c b/arch/s390/mm/pageattr.c
index d21040e..80adfbf 100644
--- a/arch/s390/mm/pageattr.c
+++ b/arch/s390/mm/pageattr.c
@@ -9,31 +9,25 @@
 #include <asm/pgtable.h>
 #include <asm/page.h>
 
+static inline unsigned long sske_frame(unsigned long addr, unsigned char skey)
+{
+	asm volatile(".insn rrf,0xb22b0000,%[skey],%[addr],9,0"
+		     : [addr] "+a" (addr) : [skey] "d" (skey));
+	return addr;
+}
+
 void storage_key_init_range(unsigned long start, unsigned long end)
 {
-	unsigned long boundary, function, size;
+	unsigned long boundary, size;
 
 	while (start < end) {
-		if (MACHINE_HAS_EDAT2) {
-			/* set storage keys for a 2GB frame */
-			function = 0x22000 | PAGE_DEFAULT_KEY;
-			size = 1UL << 31;
-			boundary = (start + size) & ~(size - 1);
-			if (boundary <= end) {
-				do {
-					start = pfmf(function, start);
-				} while (start < boundary);
-				continue;
-			}
-		}
 		if (MACHINE_HAS_EDAT1) {
 			/* set storage keys for a 1MB frame */
-			function = 0x21000 | PAGE_DEFAULT_KEY;
 			size = 1UL << 20;
 			boundary = (start + size) & ~(size - 1);
 			if (boundary <= end) {
 				do {
-					start = pfmf(function, start);
+					start = sske_frame(start, PAGE_DEFAULT_KEY);
 				} while (start < boundary);
 				continue;
 			}
diff --git a/arch/s390/mm/pgtable.c b/arch/s390/mm/pgtable.c
index ae44d2a..bd954e9 100644
--- a/arch/s390/mm/pgtable.c
+++ b/arch/s390/mm/pgtable.c
@@ -379,75 +379,183 @@ out_unmap:
 }
 EXPORT_SYMBOL_GPL(gmap_map_segment);
 
-/*
- * this function is assumed to be called with mmap_sem held
- */
-unsigned long __gmap_fault(unsigned long address, struct gmap *gmap)
+static unsigned long *gmap_table_walk(unsigned long address, struct gmap *gmap)
 {
-	unsigned long *table, vmaddr, segment;
-	struct mm_struct *mm;
-	struct gmap_pgtable *mp;
-	struct gmap_rmap *rmap;
-	struct vm_area_struct *vma;
-	struct page *page;
-	pgd_t *pgd;
-	pud_t *pud;
-	pmd_t *pmd;
+	unsigned long *table;
 
-	current->thread.gmap_addr = address;
-	mm = gmap->mm;
-	/* Walk the gmap address space page table */
 	table = gmap->table + ((address >> 53) & 0x7ff);
 	if (unlikely(*table & _REGION_ENTRY_INV))
-		return -EFAULT;
+		return ERR_PTR(-EFAULT);
 	table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
 	table = table + ((address >> 42) & 0x7ff);
 	if (unlikely(*table & _REGION_ENTRY_INV))
-		return -EFAULT;
+		return ERR_PTR(-EFAULT);
 	table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
 	table = table + ((address >> 31) & 0x7ff);
 	if (unlikely(*table & _REGION_ENTRY_INV))
-		return -EFAULT;
+		return ERR_PTR(-EFAULT);
 	table = (unsigned long *)(*table & _REGION_ENTRY_ORIGIN);
 	table = table + ((address >> 20) & 0x7ff);
+	return table;
+}
+
+/**
+ * __gmap_translate - translate a guest address to a user space address
+ * @address: guest address
+ * @gmap: pointer to guest mapping meta data structure
+ *
+ * Returns user space address which corresponds to the guest address or
+ * -EFAULT if no such mapping exists.
+ * This function does not establish potentially missing page table entries.
+ * The mmap_sem of the mm that belongs to the address space must be held
+ * when this function gets called.
+ */
+unsigned long __gmap_translate(unsigned long address, struct gmap *gmap)
+{
+	unsigned long *segment_ptr, vmaddr, segment;
+	struct gmap_pgtable *mp;
+	struct page *page;
 
+	current->thread.gmap_addr = address;
+	segment_ptr = gmap_table_walk(address, gmap);
+	if (IS_ERR(segment_ptr))
+		return PTR_ERR(segment_ptr);
 	/* Convert the gmap address to an mm address. */
-	segment = *table;
-	if (likely(!(segment & _SEGMENT_ENTRY_INV))) {
+	segment = *segment_ptr;
+	if (!(segment & _SEGMENT_ENTRY_INV)) {
 		page = pfn_to_page(segment >> PAGE_SHIFT);
 		mp = (struct gmap_pgtable *) page->index;
 		return mp->vmaddr | (address & ~PMD_MASK);
 	} else if (segment & _SEGMENT_ENTRY_RO) {
 		vmaddr = segment & _SEGMENT_ENTRY_ORIGIN;
-		vma = find_vma(mm, vmaddr);
-		if (!vma || vma->vm_start > vmaddr)
-			return -EFAULT;
-
-		/* Walk the parent mm page table */
-		pgd = pgd_offset(mm, vmaddr);
-		pud = pud_alloc(mm, pgd, vmaddr);
-		if (!pud)
-			return -ENOMEM;
-		pmd = pmd_alloc(mm, pud, vmaddr);
-		if (!pmd)
-			return -ENOMEM;
-		if (!pmd_present(*pmd) &&
-		    __pte_alloc(mm, vma, pmd, vmaddr))
-			return -ENOMEM;
-		/* pmd now points to a valid segment table entry. */
-		rmap = kmalloc(sizeof(*rmap), GFP_KERNEL|__GFP_REPEAT);
-		if (!rmap)
-			return -ENOMEM;
-		/* Link gmap segment table entry location to page table. */
-		page = pmd_page(*pmd);
-		mp = (struct gmap_pgtable *) page->index;
-		rmap->entry = table;
-		spin_lock(&mm->page_table_lock);
+		return vmaddr | (address & ~PMD_MASK);
+	}
+	return -EFAULT;
+}
+EXPORT_SYMBOL_GPL(__gmap_translate);
+
+/**
+ * gmap_translate - translate a guest address to a user space address
+ * @address: guest address
+ * @gmap: pointer to guest mapping meta data structure
+ *
+ * Returns user space address which corresponds to the guest address or
+ * -EFAULT if no such mapping exists.
+ * This function does not establish potentially missing page table entries.
+ */
+unsigned long gmap_translate(unsigned long address, struct gmap *gmap)
+{
+	unsigned long rc;
+
+	down_read(&gmap->mm->mmap_sem);
+	rc = __gmap_translate(address, gmap);
+	up_read(&gmap->mm->mmap_sem);
+	return rc;
+}
+EXPORT_SYMBOL_GPL(gmap_translate);
+
+static int gmap_connect_pgtable(unsigned long segment,
+				unsigned long *segment_ptr,
+				struct gmap *gmap)
+{
+	unsigned long vmaddr;
+	struct vm_area_struct *vma;
+	struct gmap_pgtable *mp;
+	struct gmap_rmap *rmap;
+	struct mm_struct *mm;
+	struct page *page;
+	pgd_t *pgd;
+	pud_t *pud;
+	pmd_t *pmd;
+
+	mm = gmap->mm;
+	vmaddr = segment & _SEGMENT_ENTRY_ORIGIN;
+	vma = find_vma(mm, vmaddr);
+	if (!vma || vma->vm_start > vmaddr)
+		return -EFAULT;
+	/* Walk the parent mm page table */
+	pgd = pgd_offset(mm, vmaddr);
+	pud = pud_alloc(mm, pgd, vmaddr);
+	if (!pud)
+		return -ENOMEM;
+	pmd = pmd_alloc(mm, pud, vmaddr);
+	if (!pmd)
+		return -ENOMEM;
+	if (!pmd_present(*pmd) &&
+	    __pte_alloc(mm, vma, pmd, vmaddr))
+		return -ENOMEM;
+	/* pmd now points to a valid segment table entry. */
+	rmap = kmalloc(sizeof(*rmap), GFP_KERNEL|__GFP_REPEAT);
+	if (!rmap)
+		return -ENOMEM;
+	/* Link gmap segment table entry location to page table. */
+	page = pmd_page(*pmd);
+	mp = (struct gmap_pgtable *) page->index;
+	rmap->entry = segment_ptr;
+	spin_lock(&mm->page_table_lock);
+	if (*segment_ptr == segment) {
 		list_add(&rmap->list, &mp->mapper);
-		spin_unlock(&mm->page_table_lock);
 		/* Set gmap segment table entry to page table. */
-		*table = pmd_val(*pmd) & PAGE_MASK;
-		return vmaddr | (address & ~PMD_MASK);
+		*segment_ptr = pmd_val(*pmd) & PAGE_MASK;
+		rmap = NULL;
+	}
+	spin_unlock(&mm->page_table_lock);
+	kfree(rmap);
+	return 0;
+}
+
+static void gmap_disconnect_pgtable(struct mm_struct *mm, unsigned long *table)
+{
+	struct gmap_rmap *rmap, *next;
+	struct gmap_pgtable *mp;
+	struct page *page;
+	int flush;
+
+	flush = 0;
+	spin_lock(&mm->page_table_lock);
+	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
+	mp = (struct gmap_pgtable *) page->index;
+	list_for_each_entry_safe(rmap, next, &mp->mapper, list) {
+		*rmap->entry =
+			_SEGMENT_ENTRY_INV | _SEGMENT_ENTRY_RO | mp->vmaddr;
+		list_del(&rmap->list);
+		kfree(rmap);
+		flush = 1;
+	}
+	spin_unlock(&mm->page_table_lock);
+	if (flush)
+		__tlb_flush_global();
+}
+
+/*
+ * this function is assumed to be called with mmap_sem held
+ */
+unsigned long __gmap_fault(unsigned long address, struct gmap *gmap)
+{
+	unsigned long *segment_ptr, segment;
+	struct gmap_pgtable *mp;
+	struct page *page;
+	int rc;
+
+	current->thread.gmap_addr = address;
+	segment_ptr = gmap_table_walk(address, gmap);
+	if (IS_ERR(segment_ptr))
+		return -EFAULT;
+	/* Convert the gmap address to an mm address. */
+	while (1) {
+		segment = *segment_ptr;
+		if (!(segment & _SEGMENT_ENTRY_INV)) {
+			/* Page table is present */
+			page = pfn_to_page(segment >> PAGE_SHIFT);
+			mp = (struct gmap_pgtable *) page->index;
+			return mp->vmaddr | (address & ~PMD_MASK);
+		}
+		if (!(segment & _SEGMENT_ENTRY_RO))
+			/* Nothing mapped in the gmap address space. */
+			break;
+		rc = gmap_connect_pgtable(segment, segment_ptr, gmap);
+		if (rc)
+			return rc;
 	}
 	return -EFAULT;
 }
@@ -511,29 +619,6 @@ void gmap_discard(unsigned long from, unsigned long to, struct gmap *gmap)
 }
 EXPORT_SYMBOL_GPL(gmap_discard);
 
-void gmap_unmap_notifier(struct mm_struct *mm, unsigned long *table)
-{
-	struct gmap_rmap *rmap, *next;
-	struct gmap_pgtable *mp;
-	struct page *page;
-	int flush;
-
-	flush = 0;
-	spin_lock(&mm->page_table_lock);
-	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
-	mp = (struct gmap_pgtable *) page->index;
-	list_for_each_entry_safe(rmap, next, &mp->mapper, list) {
-		*rmap->entry =
-			_SEGMENT_ENTRY_INV | _SEGMENT_ENTRY_RO | mp->vmaddr;
-		list_del(&rmap->list);
-		kfree(rmap);
-		flush = 1;
-	}
-	spin_unlock(&mm->page_table_lock);
-	if (flush)
-		__tlb_flush_global();
-}
-
 static inline unsigned long *page_table_alloc_pgste(struct mm_struct *mm,
 						    unsigned long vmaddr)
 {
@@ -586,8 +671,8 @@ static inline void page_table_free_pgste(unsigned long *table)
 {
 }
 
-static inline void gmap_unmap_notifier(struct mm_struct *mm,
-					  unsigned long *table)
+static inline void gmap_disconnect_pgtable(struct mm_struct *mm,
+					   unsigned long *table)
 {
 }
 
@@ -653,7 +738,7 @@ void page_table_free(struct mm_struct *mm, unsigned long *table)
 	unsigned int bit, mask;
 
 	if (mm_has_pgste(mm)) {
-		gmap_unmap_notifier(mm, table);
+		gmap_disconnect_pgtable(mm, table);
 		return page_table_free_pgste(table);
 	}
 	/* Free 1K/2K page table fragment of a 4K page */
@@ -696,7 +781,7 @@ void page_table_free_rcu(struct mmu_gather *tlb, unsigned long *table)
 
 	mm = tlb->mm;
 	if (mm_has_pgste(mm)) {
-		gmap_unmap_notifier(mm, table);
+		gmap_disconnect_pgtable(mm, table);
 		table = (unsigned long *) (__pa(table) | FRAG_MASK);
 		tlb_remove_table(tlb, table);
 		return;
diff --git a/arch/s390/mm/vmem.c b/arch/s390/mm/vmem.c
index ffab84d..3583705 100644
--- a/arch/s390/mm/vmem.c
+++ b/arch/s390/mm/vmem.c
@@ -191,19 +191,16 @@ static void vmem_remove_range(unsigned long start, unsigned long size)
 /*
  * Add a backed mem_map array to the virtual mem_map array.
  */
-int __meminit vmemmap_populate(struct page *start, unsigned long nr, int node)
+int __meminit vmemmap_populate(unsigned long start, unsigned long end, int node)
 {
-	unsigned long address, start_addr, end_addr;
+	unsigned long address = start;
 	pgd_t *pg_dir;
 	pud_t *pu_dir;
 	pmd_t *pm_dir;
 	pte_t *pt_dir;
 	int ret = -ENOMEM;
 
-	start_addr = (unsigned long) start;
-	end_addr = (unsigned long) (start + nr);
-
-	for (address = start_addr; address < end_addr;) {
+	for (address = start; address < end;) {
 		pg_dir = pgd_offset_k(address);
 		if (pgd_none(*pg_dir)) {
 			pu_dir = vmem_pud_alloc();
@@ -262,14 +259,14 @@ int __meminit vmemmap_populate(struct page *start, unsigned long nr, int node)
 		}
 		address += PAGE_SIZE;
 	}
-	memset(start, 0, nr * sizeof(struct page));
+	memset((void *)start, 0, end - start);
 	ret = 0;
 out:
-	flush_tlb_kernel_range(start_addr, end_addr);
+	flush_tlb_kernel_range(start, end);
 	return ret;
 }
 
-void vmemmap_free(struct page *memmap, unsigned long nr_pages)
+void vmemmap_free(unsigned long start, unsigned long end)
 {
 }