Merge branch 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6

* 'for-linus' of git://git390.osdl.marist.edu/pub/scm/linux-2.6:
  [S390] 4level-fixup cleanup
  [S390] Cleanup page table definitions.
  [S390] Introduce follow_table in uaccess_pt.c
  [S390] Remove unused user_seg from thread structure.
  [S390] tlb flush fix.
  [S390] kernel: Fix dump on panic for DASDs under LPAR.
  [S390] struct class_device -> struct device conversion.
  [S390] cio: Fix incomplete commit for uevent suppression.
  [S390] cio: Use to_channelpath() for device to channel path conversion.
  [S390] Add per-cpu idle time / idle count sysfs attributes.
  [S390] Update default configuration.

8 files changed, 541 insertions, 518 deletions

diff --git a/include/asm-s390/cpu.h b/include/asm-s390/cpu.h
new file mode 100644
index 0000000..352dde1
--- /dev/null
+++ b/include/asm-s390/cpu.h
@@ -0,0 +1,25 @@
+/*
+ *  include/asm-s390/cpu.h
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Heiko Carstens <heiko.carstens@de.ibm.com>
+ */
+
+#ifndef _ASM_S390_CPU_H_
+#define _ASM_S390_CPU_H_
+
+#include <linux/types.h>
+#include <linux/percpu.h>
+#include <linux/spinlock.h>
+
+struct s390_idle_data {
+	spinlock_t lock;
+	unsigned int in_idle;
+	unsigned long long idle_count;
+	unsigned long long idle_enter;
+	unsigned long long idle_time;
+};
+
+DECLARE_PER_CPU(struct s390_idle_data, s390_idle);
+
+#endif /* _ASM_S390_CPU_H_ */
diff --git a/include/asm-s390/mmu_context.h b/include/asm-s390/mmu_context.h
index 501cb9b..05b8421 100644
--- a/include/asm-s390/mmu_context.h
+++ b/include/asm-s390/mmu_context.h
@@ -21,45 +21,43 @@
 
 #ifndef __s390x__
 #define LCTL_OPCODE "lctl"
-#define PGTABLE_BITS (_SEGMENT_TABLE|USER_STD_MASK)
 #else
 #define LCTL_OPCODE "lctlg"
-#define PGTABLE_BITS (_REGION_TABLE|USER_STD_MASK)
 #endif
 
-static inline void enter_lazy_tlb(struct mm_struct *mm,
-                                  struct task_struct *tsk)
+static inline void update_mm(struct mm_struct *mm, struct task_struct *tsk)
 {
+	pgd_t *pgd = mm->pgd;
+	unsigned long asce_bits;
+
+	/* Calculate asce bits from the first pgd table entry. */
+	asce_bits = _ASCE_TABLE_LENGTH | _ASCE_USER_BITS;
+#ifdef CONFIG_64BIT
+	asce_bits |= _ASCE_TYPE_REGION3;
+#endif
+	S390_lowcore.user_asce = asce_bits | __pa(pgd);
+	if (switch_amode) {
+		/* Load primary space page table origin. */
+		pgd_t *shadow_pgd = get_shadow_table(pgd) ? : pgd;
+		S390_lowcore.user_exec_asce = asce_bits | __pa(shadow_pgd);
+		asm volatile(LCTL_OPCODE" 1,1,%0\n"
+			     : : "m" (S390_lowcore.user_exec_asce) );
+	} else
+		/* Load home space page table origin. */
+		asm volatile(LCTL_OPCODE" 13,13,%0"
+			     : : "m" (S390_lowcore.user_asce) );
 }
 
 static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next,
 			     struct task_struct *tsk)
 {
-	pgd_t *shadow_pgd = get_shadow_pgd(next->pgd);
-
-	if (prev != next) {
-		S390_lowcore.user_asce = (__pa(next->pgd) & PAGE_MASK) |
-					 PGTABLE_BITS;
-		if (shadow_pgd) {
-			/* Load primary/secondary space page table origin. */
-			S390_lowcore.user_exec_asce =
-				(__pa(shadow_pgd) & PAGE_MASK) | PGTABLE_BITS;
-			asm volatile(LCTL_OPCODE" 1,1,%0\n"
-				     LCTL_OPCODE" 7,7,%1"
-				     : : "m" (S390_lowcore.user_exec_asce),
-					 "m" (S390_lowcore.user_asce) );
-		} else if (switch_amode) {
-			/* Load primary space page table origin. */
-			asm volatile(LCTL_OPCODE" 1,1,%0"
-				     : : "m" (S390_lowcore.user_asce) );
-		} else
-			/* Load home space page table origin. */
-			asm volatile(LCTL_OPCODE" 13,13,%0"
-				     : : "m" (S390_lowcore.user_asce) );
-	}
+	if (unlikely(prev == next))
+		return;
 	cpu_set(smp_processor_id(), next->cpu_vm_mask);
+	update_mm(next, tsk);
 }
 
+#define enter_lazy_tlb(mm,tsk)	do { } while (0)
 #define deactivate_mm(tsk,mm)	do { } while (0)
 
 static inline void activate_mm(struct mm_struct *prev,
diff --git a/include/asm-s390/page.h b/include/asm-s390/page.h
index ceec382..584d0ee 100644
--- a/include/asm-s390/page.h
+++ b/include/asm-s390/page.h
@@ -82,6 +82,7 @@ typedef struct { unsigned long pte; } pte_t;
 #ifndef __s390x__
 
 typedef struct { unsigned long pmd; } pmd_t;
+typedef struct { unsigned long pud; } pud_t;
 typedef struct {
         unsigned long pgd0;
         unsigned long pgd1;
@@ -90,6 +91,7 @@ typedef struct {
         } pgd_t;
 
 #define pmd_val(x)      ((x).pmd)
+#define pud_val(x)	((x).pud)
 #define pgd_val(x)      ((x).pgd0)
 
 #else /* __s390x__ */
@@ -98,10 +100,12 @@ typedef struct {
         unsigned long pmd0;
         unsigned long pmd1; 
         } pmd_t;
+typedef struct { unsigned long pud; } pud_t;
 typedef struct { unsigned long pgd; } pgd_t;
 
 #define pmd_val(x)      ((x).pmd0)
 #define pmd_val1(x)     ((x).pmd1)
+#define pud_val(x)	((x).pud)
 #define pgd_val(x)      ((x).pgd)
 
 #endif /* __s390x__ */
diff --git a/include/asm-s390/pgalloc.h b/include/asm-s390/pgalloc.h
index e45d3c9..709dd17 100644
--- a/include/asm-s390/pgalloc.h
+++ b/include/asm-s390/pgalloc.h
@@ -19,140 +19,115 @@
 
 #define check_pgt_cache()	do {} while (0)
 
-/*
- * Page allocation orders.
- */
-#ifndef __s390x__
-# define PTE_ALLOC_ORDER	0
-# define PMD_ALLOC_ORDER	0
-# define PGD_ALLOC_ORDER	1
-#else /* __s390x__ */
-# define PTE_ALLOC_ORDER	0
-# define PMD_ALLOC_ORDER	2
-# define PGD_ALLOC_ORDER	2
-#endif /* __s390x__ */
+unsigned long *crst_table_alloc(struct mm_struct *, int);
+void crst_table_free(unsigned long *);
 
-/*
- * Allocate and free page tables. The xxx_kernel() versions are
- * used to allocate a kernel page table - this turns on ASN bits
- * if any.
- */
+unsigned long *page_table_alloc(int);
+void page_table_free(unsigned long *);
 
-static inline pgd_t *pgd_alloc(struct mm_struct *mm)
+static inline void clear_table(unsigned long *s, unsigned long val, size_t n)
 {
-	pgd_t *pgd = (pgd_t *) __get_free_pages(GFP_KERNEL, PGD_ALLOC_ORDER);
-	int i;
-
-	if (!pgd)
-		return NULL;
-	if (s390_noexec) {
-		pgd_t *shadow_pgd = (pgd_t *)
-			__get_free_pages(GFP_KERNEL, PGD_ALLOC_ORDER);
-		struct page *page = virt_to_page(pgd);
-
-		if (!shadow_pgd) {
-			free_pages((unsigned long) pgd, PGD_ALLOC_ORDER);
-			return NULL;
-		}
-		page->lru.next = (void *) shadow_pgd;
-	}
-	for (i = 0; i < PTRS_PER_PGD; i++)
-#ifndef __s390x__
-		pmd_clear(pmd_offset(pgd + i, i*PGDIR_SIZE));
+	*s = val;
+	n = (n / 256) - 1;
+	asm volatile(
+#ifdef CONFIG_64BIT
+		"	mvc	8(248,%0),0(%0)\n"
 #else
-		pgd_clear(pgd + i);
+		"	mvc	4(252,%0),0(%0)\n"
 #endif
-	return pgd;
+		"0:	mvc	256(256,%0),0(%0)\n"
+		"	la	%0,256(%0)\n"
+		"	brct	%1,0b\n"
+		: "+a" (s), "+d" (n));
 }
 
-static inline void pgd_free(pgd_t *pgd)
+static inline void crst_table_init(unsigned long *crst, unsigned long entry)
 {
-	pgd_t *shadow_pgd = get_shadow_pgd(pgd);
-
-	if (shadow_pgd)
-		free_pages((unsigned long) shadow_pgd, PGD_ALLOC_ORDER);
-	free_pages((unsigned long) pgd, PGD_ALLOC_ORDER);
+	clear_table(crst, entry, sizeof(unsigned long)*2048);
+	crst = get_shadow_table(crst);
+	if (crst)
+		clear_table(crst, entry, sizeof(unsigned long)*2048);
 }
 
 #ifndef __s390x__
-/*
- * page middle directory allocation/free routines.
- * We use pmd cache only on s390x, so these are dummy routines. This
- * code never triggers because the pgd will always be present.
- */
-#define pmd_alloc_one(mm,address)       ({ BUG(); ((pmd_t *)2); })
-#define pmd_free(x)                     do { } while (0)
-#define __pmd_free_tlb(tlb,x)		do { } while (0)
-#define pgd_populate(mm, pmd, pte)      BUG()
-#define pgd_populate_kernel(mm, pmd, pte)	BUG()
-#else /* __s390x__ */
-static inline pmd_t * pmd_alloc_one(struct mm_struct *mm, unsigned long vmaddr)
+
+static inline unsigned long pgd_entry_type(struct mm_struct *mm)
 {
-	pmd_t *pmd = (pmd_t *) __get_free_pages(GFP_KERNEL, PMD_ALLOC_ORDER);
-	int i;
-
-	if (!pmd)
-		return NULL;
-	if (s390_noexec) {
-		pmd_t *shadow_pmd = (pmd_t *)
-			__get_free_pages(GFP_KERNEL, PMD_ALLOC_ORDER);
-		struct page *page = virt_to_page(pmd);
-
-		if (!shadow_pmd) {
-			free_pages((unsigned long) pmd, PMD_ALLOC_ORDER);
-			return NULL;
-		}
-		page->lru.next = (void *) shadow_pmd;
-	}
-	for (i=0; i < PTRS_PER_PMD; i++)
-		pmd_clear(pmd + i);
-	return pmd;
+	return _SEGMENT_ENTRY_EMPTY;
 }
 
-static inline void pmd_free (pmd_t *pmd)
+#define pud_alloc_one(mm,address)		({ BUG(); ((pud_t *)2); })
+#define pud_free(x)				do { } while (0)
+
+#define pmd_alloc_one(mm,address)		({ BUG(); ((pmd_t *)2); })
+#define pmd_free(x)				do { } while (0)
+
+#define pgd_populate(mm, pgd, pud)		BUG()
+#define pgd_populate_kernel(mm, pgd, pud)	BUG()
+
+#define pud_populate(mm, pud, pmd)		BUG()
+#define pud_populate_kernel(mm, pud, pmd)	BUG()
+
+#else /* __s390x__ */
+
+static inline unsigned long pgd_entry_type(struct mm_struct *mm)
 {
-	pmd_t *shadow_pmd = get_shadow_pmd(pmd);
+	return _REGION3_ENTRY_EMPTY;
+}
+
+#define pud_alloc_one(mm,address)		({ BUG(); ((pud_t *)2); })
+#define pud_free(x)				do { } while (0)
 
-	if (shadow_pmd)
-		free_pages((unsigned long) shadow_pmd, PMD_ALLOC_ORDER);
-	free_pages((unsigned long) pmd, PMD_ALLOC_ORDER);
+static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long vmaddr)
+{
+	unsigned long *crst = crst_table_alloc(mm, s390_noexec);
+	if (crst)
+		crst_table_init(crst, _SEGMENT_ENTRY_EMPTY);
+	return (pmd_t *) crst;
 }
+#define pmd_free(pmd) crst_table_free((unsigned long *) pmd)
 
-#define __pmd_free_tlb(tlb,pmd)			\
-	do {					\
-		tlb_flush_mmu(tlb, 0, 0);	\
-		pmd_free(pmd);			\
-	 } while (0)
+#define pgd_populate(mm, pgd, pud)		BUG()
+#define pgd_populate_kernel(mm, pgd, pud)	BUG()
 
-static inline void
-pgd_populate_kernel(struct mm_struct *mm, pgd_t *pgd, pmd_t *pmd)
+static inline void pud_populate_kernel(struct mm_struct *mm,
+				       pud_t *pud, pmd_t *pmd)
 {
-	pgd_val(*pgd) = _PGD_ENTRY | __pa(pmd);
+	pud_val(*pud) = _REGION3_ENTRY | __pa(pmd);
 }
 
-static inline void pgd_populate(struct mm_struct *mm, pgd_t *pgd, pmd_t *pmd)
+static inline void pud_populate(struct mm_struct *mm, pud_t *pud, pmd_t *pmd)
 {
-	pgd_t *shadow_pgd = get_shadow_pgd(pgd);
-	pmd_t *shadow_pmd = get_shadow_pmd(pmd);
+	pud_t *shadow_pud = get_shadow_table(pud);
+	pmd_t *shadow_pmd = get_shadow_table(pmd);
 
-	if (shadow_pgd && shadow_pmd)
-		pgd_populate_kernel(mm, shadow_pgd, shadow_pmd);
-	pgd_populate_kernel(mm, pgd, pmd);
+	if (shadow_pud && shadow_pmd)
+		pud_populate_kernel(mm, shadow_pud, shadow_pmd);
+	pud_populate_kernel(mm, pud, pmd);
 }
 
 #endif /* __s390x__ */
 
+static inline pgd_t *pgd_alloc(struct mm_struct *mm)
+{
+	unsigned long *crst = crst_table_alloc(mm, s390_noexec);
+	if (crst)
+		crst_table_init(crst, pgd_entry_type(mm));
+	return (pgd_t *) crst;
+}
+#define pgd_free(pgd) crst_table_free((unsigned long *) pgd)
+
 static inline void 
 pmd_populate_kernel(struct mm_struct *mm, pmd_t *pmd, pte_t *pte)
 {
 #ifndef __s390x__
-	pmd_val(pmd[0]) = _PAGE_TABLE + __pa(pte);
-	pmd_val(pmd[1]) = _PAGE_TABLE + __pa(pte+256);
-	pmd_val(pmd[2]) = _PAGE_TABLE + __pa(pte+512);
-	pmd_val(pmd[3]) = _PAGE_TABLE + __pa(pte+768);
+	pmd_val(pmd[0]) = _SEGMENT_ENTRY + __pa(pte);
+	pmd_val(pmd[1]) = _SEGMENT_ENTRY + __pa(pte+256);
+	pmd_val(pmd[2]) = _SEGMENT_ENTRY + __pa(pte+512);
+	pmd_val(pmd[3]) = _SEGMENT_ENTRY + __pa(pte+768);
 #else /* __s390x__ */
-	pmd_val(*pmd) = _PMD_ENTRY + __pa(pte);
-	pmd_val1(*pmd) = _PMD_ENTRY + __pa(pte+256);
+	pmd_val(*pmd) = _SEGMENT_ENTRY + __pa(pte);
+	pmd_val1(*pmd) = _SEGMENT_ENTRY + __pa(pte+256);
 #endif /* __s390x__ */
 }
 
@@ -160,7 +135,7 @@ static inline void
 pmd_populate(struct mm_struct *mm, pmd_t *pmd, struct page *page)
 {
 	pte_t *pte = (pte_t *)page_to_phys(page);
-	pmd_t *shadow_pmd = get_shadow_pmd(pmd);
+	pmd_t *shadow_pmd = get_shadow_table(pmd);
 	pte_t *shadow_pte = get_shadow_pte(pte);
 
 	pmd_populate_kernel(mm, pmd, pte);
@@ -171,67 +146,14 @@ pmd_populate(struct mm_struct *mm, pmd_t *pmd, struct page *page)
 /*
  * page table entry allocation/free routines.
  */
-static inline pte_t *
-pte_alloc_one_kernel(struct mm_struct *mm, unsigned long vmaddr)
-{
-	pte_t *pte = (pte_t *) __get_free_page(GFP_KERNEL|__GFP_REPEAT);
-	int i;
-
-	if (!pte)
-		return NULL;
-	if (s390_noexec) {
-		pte_t *shadow_pte = (pte_t *)
-			__get_free_page(GFP_KERNEL|__GFP_REPEAT);
-		struct page *page = virt_to_page(pte);
-
-		if (!shadow_pte) {
-			free_page((unsigned long) pte);
-			return NULL;
-		}
-		page->lru.next = (void *) shadow_pte;
-	}
-	for (i=0; i < PTRS_PER_PTE; i++) {
-		pte_clear(mm, vmaddr, pte + i);
-		vmaddr += PAGE_SIZE;
-	}
-	return pte;
-}
-
-static inline struct page *
-pte_alloc_one(struct mm_struct *mm, unsigned long vmaddr)
-{
-	pte_t *pte = pte_alloc_one_kernel(mm, vmaddr);
-	if (pte)
-		return virt_to_page(pte);
-	return NULL;
-}
-
-static inline void pte_free_kernel(pte_t *pte)
-{
-	pte_t *shadow_pte = get_shadow_pte(pte);
-
-	if (shadow_pte)
-		free_page((unsigned long) shadow_pte);
-	free_page((unsigned long) pte);
-}
-
-static inline void pte_free(struct page *pte)
-{
-	struct page *shadow_page = get_shadow_page(pte);
-
-	if (shadow_page)
-		__free_page(shadow_page);
-	__free_page(pte);
-}
-
-#define __pte_free_tlb(tlb, pte)					\
-({									\
-	struct mmu_gather *__tlb = (tlb);				\
-	struct page *__pte = (pte);					\
-	struct page *shadow_page = get_shadow_page(__pte);		\
-	if (shadow_page)						\
-		tlb_remove_page(__tlb, shadow_page);			\
-	tlb_remove_page(__tlb, __pte);					\
-})
+#define pte_alloc_one_kernel(mm, vmaddr) \
+	((pte_t *) page_table_alloc(s390_noexec))
+#define pte_alloc_one(mm, vmaddr) \
+	virt_to_page(page_table_alloc(s390_noexec))
+
+#define pte_free_kernel(pte) \
+	page_table_free((unsigned long *) pte)
+#define pte_free(pte) \
+	page_table_free((unsigned long *) page_to_phys((struct page *) pte))
 
 #endif /* _S390_PGALLOC_H */
diff --git a/include/asm-s390/pgtable.h b/include/asm-s390/pgtable.h
index 39bb519..f2cc25b 100644
--- a/include/asm-s390/pgtable.h
+++ b/include/asm-s390/pgtable.h
@@ -13,8 +13,6 @@
 #ifndef _ASM_S390_PGTABLE_H
 #define _ASM_S390_PGTABLE_H
 
-#include <asm-generic/4level-fixup.h>
-
 /*
  * The Linux memory management assumes a three-level page table setup. For
  * s390 31 bit we "fold" the mid level into the top-level page table, so
@@ -35,9 +33,6 @@
 #include <asm/bug.h>
 #include <asm/processor.h>
 
-struct vm_area_struct; /* forward declaration (include/linux/mm.h) */
-struct mm_struct;
-
 extern pgd_t swapper_pg_dir[] __attribute__ ((aligned (4096)));
 extern void paging_init(void);
 extern void vmem_map_init(void);
@@ -63,14 +58,18 @@ extern char empty_zero_page[PAGE_SIZE];
  */
 #ifndef __s390x__
 # define PMD_SHIFT	22
+# define PUD_SHIFT	22
 # define PGDIR_SHIFT	22
 #else /* __s390x__ */
 # define PMD_SHIFT	21
+# define PUD_SHIFT	31
 # define PGDIR_SHIFT	31
 #endif /* __s390x__ */
 
 #define PMD_SIZE        (1UL << PMD_SHIFT)
 #define PMD_MASK        (~(PMD_SIZE-1))
+#define PUD_SIZE	(1UL << PUD_SHIFT)
+#define PUD_MASK	(~(PUD_SIZE-1))
 #define PGDIR_SIZE      (1UL << PGDIR_SHIFT)
 #define PGDIR_MASK      (~(PGDIR_SIZE-1))
 
@@ -83,10 +82,12 @@ extern char empty_zero_page[PAGE_SIZE];
 #ifndef __s390x__
 # define PTRS_PER_PTE    1024
 # define PTRS_PER_PMD    1
+# define PTRS_PER_PUD	1
 # define PTRS_PER_PGD    512
 #else /* __s390x__ */
 # define PTRS_PER_PTE    512
 # define PTRS_PER_PMD    1024
+# define PTRS_PER_PUD	1
 # define PTRS_PER_PGD    2048
 #endif /* __s390x__ */
 
@@ -96,6 +97,8 @@ extern char empty_zero_page[PAGE_SIZE];
 	printk("%s:%d: bad pte %p.\n", __FILE__, __LINE__, (void *) pte_val(e))
 #define pmd_ERROR(e) \
 	printk("%s:%d: bad pmd %p.\n", __FILE__, __LINE__, (void *) pmd_val(e))
+#define pud_ERROR(e) \
+	printk("%s:%d: bad pud %p.\n", __FILE__, __LINE__, (void *) pud_val(e))
 #define pgd_ERROR(e) \
 	printk("%s:%d: bad pgd %p.\n", __FILE__, __LINE__, (void *) pgd_val(e))
 
@@ -195,7 +198,7 @@ extern unsigned long vmalloc_end;
  * I Segment-Invalid Bit:    Segment is not available for address-translation
  * TT Type 01
  * TF
- * TL Table lenght
+ * TL Table length
  *
  * The 64 bit regiontable origin of S390 has following format:
  * |      region table origon                          |       DTTL
@@ -221,6 +224,8 @@ extern unsigned long vmalloc_end;
 /* Hardware bits in the page table entry */
 #define _PAGE_RO	0x200		/* HW read-only bit  */
 #define _PAGE_INVALID	0x400		/* HW invalid bit    */
+
+/* Software bits in the page table entry */
 #define _PAGE_SWT	0x001		/* SW pte type bit t */
 #define _PAGE_SWX	0x002		/* SW pte type bit x */
 
@@ -264,60 +269,75 @@ extern unsigned long vmalloc_end;
 
 #ifndef __s390x__
 
-/* Bits in the segment table entry */
-#define _PAGE_TABLE_LEN 0xf            /* only full page-tables            */
-#define _PAGE_TABLE_COM 0x10           /* common page-table                */
-#define _PAGE_TABLE_INV 0x20           /* invalid page-table               */
-#define _SEG_PRESENT    0x001          /* Software (overlap with PTL)      */
-
-/* Bits int the storage key */
-#define _PAGE_CHANGED    0x02          /* HW changed bit                   */
-#define _PAGE_REFERENCED 0x04          /* HW referenced bit                */
-
-#define _USER_SEG_TABLE_LEN    0x7f    /* user-segment-table up to 2 GB    */
-#define _KERNEL_SEG_TABLE_LEN  0x7f    /* kernel-segment-table up to 2 GB  */
-
-/*
- * User and Kernel pagetables are identical
- */
-#define _PAGE_TABLE	_PAGE_TABLE_LEN
-#define _KERNPG_TABLE	_PAGE_TABLE_LEN
-
-/*
- * The Kernel segment-tables includes the User segment-table
- */
+/* Bits in the segment table address-space-control-element */
+#define _ASCE_SPACE_SWITCH	0x80000000UL	/* space switch event	    */
+#define _ASCE_ORIGIN_MASK	0x7ffff000UL	/* segment table origin	    */
+#define _ASCE_PRIVATE_SPACE	0x100	/* private space control	    */
+#define _ASCE_ALT_EVENT		0x80	/* storage alteration event control */
+#define _ASCE_TABLE_LENGTH	0x7f	/* 128 x 64 entries = 8k	    */
 
-#define _SEGMENT_TABLE	(_USER_SEG_TABLE_LEN|0x80000000|0x100)
-#define _KERNSEG_TABLE	_KERNEL_SEG_TABLE_LEN
+/* Bits in the segment table entry */
+#define _SEGMENT_ENTRY_ORIGIN	0x7fffffc0UL	/* page table origin	    */
+#define _SEGMENT_ENTRY_INV	0x20	/* invalid segment table entry	    */
+#define _SEGMENT_ENTRY_COMMON	0x10	/* common segment bit		    */
+#define _SEGMENT_ENTRY_PTL	0x0f	/* page table length		    */
 
-#define USER_STD_MASK	0x00000080UL
+#define _SEGMENT_ENTRY		(_SEGMENT_ENTRY_PTL)
+#define _SEGMENT_ENTRY_EMPTY	(_SEGMENT_ENTRY_INV)
 
 #else /* __s390x__ */
 
+/* Bits in the segment/region table address-space-control-element */
+#define _ASCE_ORIGIN		~0xfffUL/* segment table origin		    */
+#define _ASCE_PRIVATE_SPACE	0x100	/* private space control	    */
+#define _ASCE_ALT_EVENT		0x80	/* storage alteration event control */
+#define _ASCE_SPACE_SWITCH	0x40	/* space switch event		    */
+#define _ASCE_REAL_SPACE	0x20	/* real space control		    */
+#define _ASCE_TYPE_MASK		0x0c	/* asce table type mask		    */
+#define _ASCE_TYPE_REGION1	0x0c	/* region first table type	    */
+#define _ASCE_TYPE_REGION2	0x08	/* region second table type	    */
+#define _ASCE_TYPE_REGION3	0x04	/* region third table type	    */
+#define _ASCE_TYPE_SEGMENT	0x00	/* segment table type		    */
+#define _ASCE_TABLE_LENGTH	0x03	/* region table length		    */
+
+/* Bits in the region table entry */
+#define _REGION_ENTRY_ORIGIN	~0xfffUL/* region/segment table origin	    */
+#define _REGION_ENTRY_INV	0x20	/* invalid region table entry	    */
+#define _REGION_ENTRY_TYPE_MASK	0x0c	/* region/segment table type mask   */
+#define _REGION_ENTRY_TYPE_R1	0x0c	/* region first table type	    */
+#define _REGION_ENTRY_TYPE_R2	0x08	/* region second table type	    */
+#define _REGION_ENTRY_TYPE_R3	0x04	/* region third table type	    */
+#define _REGION_ENTRY_LENGTH	0x03	/* region third length		    */
+
+#define _REGION1_ENTRY		(_REGION_ENTRY_TYPE_R1 | _REGION_ENTRY_LENGTH)
+#define _REGION1_ENTRY_EMPTY	(_REGION_ENTRY_TYPE_R1 | _REGION_ENTRY_INV)
+#define _REGION2_ENTRY		(_REGION_ENTRY_TYPE_R2 | _REGION_ENTRY_LENGTH)
+#define _REGION2_ENTRY_EMPTY	(_REGION_ENTRY_TYPE_R2 | _REGION_ENTRY_INV)
+#define _REGION3_ENTRY		(_REGION_ENTRY_TYPE_R3 | _REGION_ENTRY_LENGTH)
+#define _REGION3_ENTRY_EMPTY	(_REGION_ENTRY_TYPE_R3 | _REGION_ENTRY_INV)
+
 /* Bits in the segment table entry */
-#define _PMD_ENTRY_INV   0x20          /* invalid segment table entry      */
-#define _PMD_ENTRY       0x00        
+#define _SEGMENT_ENTRY_ORIGIN	~0x7ffUL/* segment table origin		    */
+#define _SEGMENT_ENTRY_RO	0x200	/* page protection bit		    */
+#define _SEGMENT_ENTRY_INV	0x20	/* invalid segment table entry	    */
+
+#define _SEGMENT_ENTRY		(0)
+#define _SEGMENT_ENTRY_EMPTY	(_SEGMENT_ENTRY_INV)
 
-/* Bits in the region third table entry */
-#define _PGD_ENTRY_INV   0x20          /* invalid region table entry       */
-#define _PGD_ENTRY       0x07
+#endif /* __s390x__ */
 
 /*
- * User and kernel page directory
+ * A user page table pointer has the space-switch-event bit, the
+ * private-space-control bit and the storage-alteration-event-control
+ * bit set. A kernel page table pointer doesn't need them.
  */
-#define _REGION_THIRD       0x4
-#define _REGION_THIRD_LEN   0x3 
-#define _REGION_TABLE       (_REGION_THIRD|_REGION_THIRD_LEN|0x40|0x100)
-#define _KERN_REGION_TABLE  (_REGION_THIRD|_REGION_THIRD_LEN)
-
-#define USER_STD_MASK           0x0000000000000080UL
+#define _ASCE_USER_BITS		(_ASCE_SPACE_SWITCH | _ASCE_PRIVATE_SPACE | \
+				 _ASCE_ALT_EVENT)
 
-/* Bits in the storage key */
+/* Bits int the storage key */
 #define _PAGE_CHANGED    0x02          /* HW changed bit                   */
 #define _PAGE_REFERENCED 0x04          /* HW referenced bit                */
 
-#endif /* __s390x__ */
-
 /*
  * Page protection definitions.
  */
@@ -358,65 +378,38 @@ extern unsigned long vmalloc_end;
 #define __S111	PAGE_EX_RW
 
 #ifndef __s390x__
-# define PMD_SHADOW_SHIFT	1
-# define PGD_SHADOW_SHIFT	1
+# define PxD_SHADOW_SHIFT	1
 #else /* __s390x__ */
-# define PMD_SHADOW_SHIFT	2
-# define PGD_SHADOW_SHIFT	2
+# define PxD_SHADOW_SHIFT	2
 #endif /* __s390x__ */
 
 static inline struct page *get_shadow_page(struct page *page)
 {
-	if (s390_noexec && !list_empty(&page->lru))
-		return virt_to_page(page->lru.next);
-	return NULL;
-}
-
-static inline pte_t *get_shadow_pte(pte_t *ptep)
-{
-	unsigned long pteptr = (unsigned long) (ptep);
-
-	if (s390_noexec) {
-		unsigned long offset = pteptr & (PAGE_SIZE - 1);
-		void *addr = (void *) (pteptr ^ offset);
-		struct page *page = virt_to_page(addr);
-		if (!list_empty(&page->lru))
-			return (pte_t *) ((unsigned long) page->lru.next |
-								offset);
-	}
+	if (s390_noexec && page->index)
+		return virt_to_page((void *)(addr_t) page->index);
 	return NULL;
 }
 
-static inline pmd_t *get_shadow_pmd(pmd_t *pmdp)
+static inline void *get_shadow_pte(void *table)
 {
-	unsigned long pmdptr = (unsigned long) (pmdp);
+	unsigned long addr, offset;
+	struct page *page;
 
-	if (s390_noexec) {
-		unsigned long offset = pmdptr &
-				((PAGE_SIZE << PMD_SHADOW_SHIFT) - 1);
-		void *addr = (void *) (pmdptr ^ offset);
-		struct page *page = virt_to_page(addr);
-		if (!list_empty(&page->lru))
-			return (pmd_t *) ((unsigned long) page->lru.next |
-								offset);
-	}
-	return NULL;
+	addr = (unsigned long) table;
+	offset = addr & (PAGE_SIZE - 1);
+	page = virt_to_page((void *)(addr ^ offset));
+	return (void *)(addr_t)(page->index ? (page->index | offset) : 0UL);
 }
 
-static inline pgd_t *get_shadow_pgd(pgd_t *pgdp)
+static inline void *get_shadow_table(void *table)
 {
-	unsigned long pgdptr = (unsigned long) (pgdp);
+	unsigned long addr, offset;
+	struct page *page;
 
-	if (s390_noexec) {
-		unsigned long offset = pgdptr &
-				((PAGE_SIZE << PGD_SHADOW_SHIFT) - 1);
-		void *addr = (void *) (pgdptr ^ offset);
-		struct page *page = virt_to_page(addr);
-		if (!list_empty(&page->lru))
-			return (pgd_t *) ((unsigned long) page->lru.next |
-								offset);
-	}
-	return NULL;
+	addr = (unsigned long) table;
+	offset = addr & ((PAGE_SIZE << PxD_SHADOW_SHIFT) - 1);
+	page = virt_to_page((void *)(addr ^ offset));
+	return (void *)(addr_t)(page->index ? (page->index | offset) : 0UL);
 }
 
 /*
@@ -424,7 +417,8 @@ static inline pgd_t *get_shadow_pgd(pgd_t *pgdp)
  * within a page table are directly modified.  Thus, the following
  * hook is made available.
  */
-static inline void set_pte(pte_t *pteptr, pte_t pteval)
+static inline void set_pte_at(struct mm_struct *mm, unsigned long addr,
+			      pte_t *pteptr, pte_t pteval)
 {
 	pte_t *shadow_pte = get_shadow_pte(pteptr);
 
@@ -437,7 +431,6 @@ static inline void set_pte(pte_t *pteptr, pte_t pteval)
 			pte_val(*shadow_pte) = _PAGE_TYPE_EMPTY;
 	}
 }
-#define set_pte_at(mm,addr,ptep,pteval) set_pte(ptep,pteval)
 
 /*
  * pgd/pmd/pte query functions
@@ -448,47 +441,50 @@ static inline int pgd_present(pgd_t pgd) { return 1; }
 static inline int pgd_none(pgd_t pgd)    { return 0; }
 static inline int pgd_bad(pgd_t pgd)     { return 0; }
 
-static inline int pmd_present(pmd_t pmd) { return pmd_val(pmd) & _SEG_PRESENT; }
-static inline int pmd_none(pmd_t pmd)    { return pmd_val(pmd) & _PAGE_TABLE_INV; }
-static inline int pmd_bad(pmd_t pmd)
-{
-	return (pmd_val(pmd) & (~PAGE_MASK & ~_PAGE_TABLE_INV)) != _PAGE_TABLE;
-}
+static inline int pud_present(pud_t pud) { return 1; }
+static inline int pud_none(pud_t pud)	 { return 0; }
+static inline int pud_bad(pud_t pud)	 { return 0; }
 
 #else /* __s390x__ */
 
-static inline int pgd_present(pgd_t pgd)
+static inline int pgd_present(pgd_t pgd) { return 1; }
+static inline int pgd_none(pgd_t pgd)	 { return 0; }
+static inline int pgd_bad(pgd_t pgd)	 { return 0; }
+
+static inline int pud_present(pud_t pud)
 {
-	return (pgd_val(pgd) & ~PAGE_MASK) == _PGD_ENTRY;
+	return pud_val(pud) & _REGION_ENTRY_ORIGIN;
 }
 
-static inline int pgd_none(pgd_t pgd)
+static inline int pud_none(pud_t pud)
 {
-	return pgd_val(pgd) & _PGD_ENTRY_INV;
+	return pud_val(pud) & _REGION_ENTRY_INV;
 }
 
-static inline int pgd_bad(pgd_t pgd)
+static inline int pud_bad(pud_t pud)
 {
-	return (pgd_val(pgd) & (~PAGE_MASK & ~_PGD_ENTRY_INV)) != _PGD_ENTRY;
+	unsigned long mask = ~_REGION_ENTRY_ORIGIN & ~_REGION_ENTRY_INV;
+	return (pud_val(pud) & mask) != _REGION3_ENTRY;
 }
 
+#endif /* __s390x__ */
+
 static inline int pmd_present(pmd_t pmd)
 {
-	return (pmd_val(pmd) & ~PAGE_MASK) == _PMD_ENTRY;
+	return pmd_val(pmd) & _SEGMENT_ENTRY_ORIGIN;
 }
 
 static inline int pmd_none(pmd_t pmd)
 {
-	return pmd_val(pmd) & _PMD_ENTRY_INV;
+	return pmd_val(pmd) & _SEGMENT_ENTRY_INV;
 }
 
 static inline int pmd_bad(pmd_t pmd)
 {
-	return (pmd_val(pmd) & (~PAGE_MASK & ~_PMD_ENTRY_INV)) != _PMD_ENTRY;
+	unsigned long mask = ~_SEGMENT_ENTRY_ORIGIN & ~_SEGMENT_ENTRY_INV;
+	return (pmd_val(pmd) & mask) != _SEGMENT_ENTRY;
 }
 
-#endif /* __s390x__ */
-
 static inline int pte_none(pte_t pte)
 {
 	return (pte_val(pte) & _PAGE_INVALID) && !(pte_val(pte) & _PAGE_SWT);
@@ -508,7 +504,8 @@ static inline int pte_file(pte_t pte)
 	return (pte_val(pte) & mask) == _PAGE_TYPE_FILE;
 }
 
-#define pte_same(a,b)	(pte_val(a) == pte_val(b))
+#define __HAVE_ARCH_PTE_SAME
+#define pte_same(a,b)  (pte_val(a) == pte_val(b))
 
 /*
  * query functions pte_write/pte_dirty/pte_young only work if
@@ -543,58 +540,52 @@ static inline int pte_young(pte_t pte)
 
 #ifndef __s390x__
 
-static inline void pgd_clear(pgd_t * pgdp)      { }
+#define pgd_clear(pgd)		do { } while (0)
+#define pud_clear(pud)		do { } while (0)
 
 static inline void pmd_clear_kernel(pmd_t * pmdp)
 {
-	pmd_val(pmdp[0]) = _PAGE_TABLE_INV;
-	pmd_val(pmdp[1]) = _PAGE_TABLE_INV;
-	pmd_val(pmdp[2]) = _PAGE_TABLE_INV;
-	pmd_val(pmdp[3]) = _PAGE_TABLE_INV;
-}
-
-static inline void pmd_clear(pmd_t * pmdp)
-{
-	pmd_t *shadow_pmd = get_shadow_pmd(pmdp);
-
-	pmd_clear_kernel(pmdp);
-	if (shadow_pmd)
-		pmd_clear_kernel(shadow_pmd);
+	pmd_val(pmdp[0]) = _SEGMENT_ENTRY_EMPTY;
+	pmd_val(pmdp[1]) = _SEGMENT_ENTRY_EMPTY;
+	pmd_val(pmdp[2]) = _SEGMENT_ENTRY_EMPTY;
+	pmd_val(pmdp[3]) = _SEGMENT_ENTRY_EMPTY;
 }
 
 #else /* __s390x__ */
 
-static inline void pgd_clear_kernel(pgd_t * pgdp)
+#define pgd_clear(pgd)		do { } while (0)
+
+static inline void pud_clear_kernel(pud_t *pud)
 {
-	pgd_val(*pgdp) = _PGD_ENTRY_INV | _PGD_ENTRY;
+	pud_val(*pud) = _REGION3_ENTRY_EMPTY;
 }
 
-static inline void pgd_clear(pgd_t * pgdp)
+static inline void pud_clear(pud_t * pud)
 {
-	pgd_t *shadow_pgd = get_shadow_pgd(pgdp);
+	pud_t *shadow = get_shadow_table(pud);
 
-	pgd_clear_kernel(pgdp);
-	if (shadow_pgd)
-		pgd_clear_kernel(shadow_pgd);
+	pud_clear_kernel(pud);
+	if (shadow)
+		pud_clear_kernel(shadow);
 }
 
 static inline void pmd_clear_kernel(pmd_t * pmdp)
 {
-	pmd_val(*pmdp) = _PMD_ENTRY_INV | _PMD_ENTRY;
-	pmd_val1(*pmdp) = _PMD_ENTRY_INV | _PMD_ENTRY;
+	pmd_val(*pmdp) = _SEGMENT_ENTRY_EMPTY;
+	pmd_val1(*pmdp) = _SEGMENT_ENTRY_EMPTY;
 }
 
+#endif /* __s390x__ */
+
 static inline void pmd_clear(pmd_t * pmdp)
 {
-	pmd_t *shadow_pmd = get_shadow_pmd(pmdp);
+	pmd_t *shadow_pmd = get_shadow_table(pmdp);
 
 	pmd_clear_kernel(pmdp);
 	if (shadow_pmd)
 		pmd_clear_kernel(shadow_pmd);
 }
 
-#endif /* __s390x__ */
-
 static inline void pte_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
 {
 	pte_t *shadow_pte = get_shadow_pte(ptep);
@@ -663,24 +654,19 @@ static inline pte_t pte_mkyoung(pte_t pte)
 	return pte;
 }
 
-static inline int ptep_test_and_clear_young(struct vm_area_struct *vma, unsigned long addr, pte_t *ptep)
+#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
+static inline int ptep_test_and_clear_young(struct vm_area_struct *vma,
+					    unsigned long addr, pte_t *ptep)
 {
 	return 0;
 }
 
-static inline int
-ptep_clear_flush_young(struct vm_area_struct *vma,
-			unsigned long address, pte_t *ptep)
+#define __HAVE_ARCH_PTEP_CLEAR_YOUNG_FLUSH
+static inline int ptep_clear_flush_young(struct vm_area_struct *vma,
+					 unsigned long address, pte_t *ptep)
 {
 	/* No need to flush TLB; bits are in storage key */
-	return ptep_test_and_clear_young(vma, address, ptep);
-}
-
-static inline pte_t ptep_get_and_clear(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
-{
-	pte_t pte = *ptep;
-	pte_clear(mm, addr, ptep);
-	return pte;
+	return 0;
 }
 
 static inline void __ptep_ipte(unsigned long address, pte_t *ptep)
@@ -709,6 +695,32 @@ static inline void ptep_invalidate(unsigned long address, pte_t *ptep)
 		__ptep_ipte(address, ptep);
 }
 
+/*
+ * This is hard to understand. ptep_get_and_clear and ptep_clear_flush
+ * both clear the TLB for the unmapped pte. The reason is that
+ * ptep_get_and_clear is used in common code (e.g. change_pte_range)
+ * to modify an active pte. The sequence is
+ *   1) ptep_get_and_clear
+ *   2) set_pte_at
+ *   3) flush_tlb_range
+ * On s390 the tlb needs to get flushed with the modification of the pte
+ * if the pte is active. The only way how this can be implemented is to
+ * have ptep_get_and_clear do the tlb flush. In exchange flush_tlb_range
+ * is a nop.
+ */
+#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
+#define ptep_get_and_clear(__mm, __address, __ptep)			\
+({									\
+	pte_t __pte = *(__ptep);					\
+	if (atomic_read(&(__mm)->mm_users) > 1 ||			\
+	    (__mm) != current->active_mm)				\
+		ptep_invalidate(__address, __ptep);			\
+	else								\
+		pte_clear((__mm), (__address), (__ptep));		\
+	__pte;								\
+})
+
+#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
 static inline pte_t ptep_clear_flush(struct vm_area_struct *vma,
 				     unsigned long address, pte_t *ptep)
 {
@@ -717,12 +729,40 @@ static inline pte_t ptep_clear_flush(struct vm_area_struct *vma,
 	return pte;
 }
 
-static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr, pte_t *ptep)
+/*
+ * The batched pte unmap code uses ptep_get_and_clear_full to clear the
+ * ptes. Here an optimization is possible. tlb_gather_mmu flushes all
+ * tlbs of an mm if it can guarantee that the ptes of the mm_struct
+ * cannot be accessed while the batched unmap is running. In this case
+ * full==1 and a simple pte_clear is enough. See tlb.h.
+ */
+#define __HAVE_ARCH_PTEP_GET_AND_CLEAR_FULL
+static inline pte_t ptep_get_and_clear_full(struct mm_struct *mm,
+					    unsigned long addr,
+					    pte_t *ptep, int full)
 {
-	pte_t old_pte = *ptep;
-	set_pte_at(mm, addr, ptep, pte_wrprotect(old_pte));
+	pte_t pte = *ptep;
+
+	if (full)
+		pte_clear(mm, addr, ptep);
+	else
+		ptep_invalidate(addr, ptep);
+	return pte;
 }
 
+#define __HAVE_ARCH_PTEP_SET_WRPROTECT
+#define ptep_set_wrprotect(__mm, __addr, __ptep)			\
+({									\
+	pte_t __pte = *(__ptep);					\
+	if (pte_write(__pte)) {						\
+		if (atomic_read(&(__mm)->mm_users) > 1 ||		\
+		    (__mm) != current->active_mm)			\
+			ptep_invalidate(__addr, __ptep);		\
+		set_pte_at(__mm, __addr, __ptep, pte_wrprotect(__pte));	\
+	}								\
+})
+
+#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
 #define ptep_set_access_flags(__vma, __addr, __ptep, __entry, __dirty)	\
 ({									\
 	int __changed = !pte_same(*(__ptep), __entry);			\
@@ -740,11 +780,13 @@ static inline void ptep_set_wrprotect(struct mm_struct *mm, unsigned long addr,
  * should therefore only be called if it is not mapped in any
  * address space.
  */
+#define __HAVE_ARCH_PAGE_TEST_DIRTY
 static inline int page_test_dirty(struct page *page)
 {
 	return (page_get_storage_key(page_to_phys(page)) & _PAGE_CHANGED) != 0;
 }
 
+#define __HAVE_ARCH_PAGE_CLEAR_DIRTY
 static inline void page_clear_dirty(struct page *page)
 {
 	page_set_storage_key(page_to_phys(page), PAGE_DEFAULT_KEY);
@@ -753,6 +795,7 @@ static inline void page_clear_dirty(struct page *page)
 /*
  * Test and clear referenced bit in storage key.
  */
+#define __HAVE_ARCH_PAGE_TEST_AND_CLEAR_YOUNG
 static inline int page_test_and_clear_young(struct page *page)
 {
 	unsigned long physpage = page_to_phys(page);
@@ -784,63 +827,48 @@ static inline pte_t mk_pte(struct page *page, pgprot_t pgprot)
 	return mk_pte_phys(physpage, pgprot);
 }
 
-static inline pte_t pfn_pte(unsigned long pfn, pgprot_t pgprot)
-{
-	unsigned long physpage = __pa((pfn) << PAGE_SHIFT);
-
-	return mk_pte_phys(physpage, pgprot);
-}
-
-#ifdef __s390x__
-
-static inline pmd_t pfn_pmd(unsigned long pfn, pgprot_t pgprot)
-{
-	unsigned long physpage = __pa((pfn) << PAGE_SHIFT);
-
-	return __pmd(physpage + pgprot_val(pgprot));
-}
-
-#endif /* __s390x__ */
-
-#define pte_pfn(x) (pte_val(x) >> PAGE_SHIFT)
-#define pte_page(x) pfn_to_page(pte_pfn(x))
+#define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
+#define pud_index(address) (((address) >> PUD_SHIFT) & (PTRS_PER_PUD-1))
+#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
+#define pte_index(address) (((address) >> PAGE_SHIFT) & (PTRS_PER_PTE-1))
 
-#define pmd_page_vaddr(pmd) (pmd_val(pmd) & PAGE_MASK)
+#define pgd_offset(mm, address) ((mm)->pgd + pgd_index(address))
+#define pgd_offset_k(address) pgd_offset(&init_mm, address)
 
-#define pmd_page(pmd) pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT)
+#ifndef __s390x__
 
-#define pgd_page_vaddr(pgd) (pgd_val(pgd) & PAGE_MASK)
+#define pmd_deref(pmd) (pmd_val(pmd) & _SEGMENT_ENTRY_ORIGIN)
+#define pud_deref(pmd) ({ BUG(); 0UL; })
+#define pgd_deref(pmd) ({ BUG(); 0UL; })
 
-#define pgd_page(pgd) pfn_to_page(pgd_val(pgd) >> PAGE_SHIFT)
+#define pud_offset(pgd, address) ((pud_t *) pgd)
+#define pmd_offset(pud, address) ((pmd_t *) pud + pmd_index(address))
 
-/* to find an entry in a page-table-directory */
-#define pgd_index(address) (((address) >> PGDIR_SHIFT) & (PTRS_PER_PGD-1))
-#define pgd_offset(mm, address) ((mm)->pgd+pgd_index(address))
+#else /* __s390x__ */
 
-/* to find an entry in a kernel page-table-directory */
-#define pgd_offset_k(address) pgd_offset(&init_mm, address)
+#define pmd_deref(pmd) (pmd_val(pmd) & _SEGMENT_ENTRY_ORIGIN)
+#define pud_deref(pud) (pud_val(pud) & _REGION_ENTRY_ORIGIN)
+#define pgd_deref(pgd) ({ BUG(); 0UL; })
 
-#ifndef __s390x__
+#define pud_offset(pgd, address) ((pud_t *) pgd)
 
-/* Find an entry in the second-level page table.. */
-static inline pmd_t * pmd_offset(pgd_t * dir, unsigned long address)
+static inline pmd_t *pmd_offset(pud_t *pud, unsigned long address)
 {
-        return (pmd_t *) dir;
+	pmd_t *pmd = (pmd_t *) pud_deref(*pud);
+	return pmd + pmd_index(address);
 }
 
-#else /* __s390x__ */
+#endif /* __s390x__ */
 
-/* Find an entry in the second-level page table.. */
-#define pmd_index(address) (((address) >> PMD_SHIFT) & (PTRS_PER_PMD-1))
-#define pmd_offset(dir,addr) \
-	((pmd_t *) pgd_page_vaddr(*(dir)) + pmd_index(addr))
+#define pfn_pte(pfn,pgprot) mk_pte_phys(__pa((pfn) << PAGE_SHIFT),(pgprot))
+#define pte_pfn(x) (pte_val(x) >> PAGE_SHIFT)
+#define pte_page(x) pfn_to_page(pte_pfn(x))
 
-#endif /* __s390x__ */
+#define pmd_page(pmd) pfn_to_page(pmd_val(pmd) >> PAGE_SHIFT)
 
-/* Find an entry in the third-level page table.. */
-#define pte_index(address) (((address) >> PAGE_SHIFT) & (PTRS_PER_PTE-1))
-#define pte_offset_kernel(pmd, address) \
-	((pte_t *) pmd_page_vaddr(*(pmd)) + pte_index(address))
+/* Find an entry in the lowest level page table.. */
+#define pte_offset(pmd, addr) ((pte_t *) pmd_deref(*(pmd)) + pte_index(addr))
+#define pte_offset_kernel(pmd, address) pte_offset(pmd,address)
 #define pte_offset_map(pmd, address) pte_offset_kernel(pmd, address)
 #define pte_offset_map_nested(pmd, address) pte_offset_kernel(pmd, address)
 #define pte_unmap(pte) do { } while (0)
@@ -930,17 +958,6 @@ extern int remove_shared_memory(unsigned long start, unsigned long size);
 #define __HAVE_ARCH_MEMMAP_INIT
 extern void memmap_init(unsigned long, int, unsigned long, unsigned long);
 
-#define __HAVE_ARCH_PTEP_SET_ACCESS_FLAGS
-#define __HAVE_ARCH_PTEP_TEST_AND_CLEAR_YOUNG
-#define __HAVE_ARCH_PTEP_CLEAR_YOUNG_FLUSH
-#define __HAVE_ARCH_PTEP_GET_AND_CLEAR
-#define __HAVE_ARCH_PTEP_CLEAR_FLUSH
-#define __HAVE_ARCH_PTEP_SET_WRPROTECT
-#define __HAVE_ARCH_PTE_SAME
-#define __HAVE_ARCH_PAGE_TEST_DIRTY
-#define __HAVE_ARCH_PAGE_CLEAR_DIRTY
-#define __HAVE_ARCH_PAGE_TEST_AND_CLEAR_YOUNG
 #include <asm-generic/pgtable.h>
 
 #endif /* _S390_PAGE_H */
-
diff --git a/include/asm-s390/processor.h b/include/asm-s390/processor.h
index 3b972d4..21d40a1 100644
--- a/include/asm-s390/processor.h
+++ b/include/asm-s390/processor.h
@@ -93,7 +93,6 @@ struct thread_struct {
 	s390_fp_regs fp_regs;
 	unsigned int  acrs[NUM_ACRS];
         unsigned long ksp;              /* kernel stack pointer             */
-        unsigned long user_seg;         /* HSTD                             */
 	mm_segment_t mm_segment;
         unsigned long prot_addr;        /* address of protection-excep.     */
         unsigned int error_code;        /* error-code of last prog-excep.   */
@@ -128,22 +127,9 @@ struct stack_frame {
 
 #define ARCH_MIN_TASKALIGN	8
 
-#ifndef __s390x__
-# define __SWAPPER_PG_DIR __pa(&swapper_pg_dir[0]) + _SEGMENT_TABLE
-#else /* __s390x__ */
-# define __SWAPPER_PG_DIR __pa(&swapper_pg_dir[0]) + _REGION_TABLE
-#endif /* __s390x__ */
-
-#define INIT_THREAD {{0,{{0},{0},{0},{0},{0},{0},{0},{0},{0},{0},	       \
-			    {0},{0},{0},{0},{0},{0}}},			       \
-		     {0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0},	       \
-		     sizeof(init_stack) + (unsigned long) &init_stack,	       \
-		     __SWAPPER_PG_DIR,					       \
-		     {0},						       \
-		     0,0,0,						       \
-		     (per_struct) {{{{0,}}},0,0,0,0,{{0,}}},		       \
-		     0, 0						       \
-} 
+#define INIT_THREAD {							\
+	.ksp = sizeof(init_stack) + (unsigned long) &init_stack,	\
+}
 
 /*
  * Do necessary setup to start up a new thread.
diff --git a/include/asm-s390/tlb.h b/include/asm-s390/tlb.h
index 51bd957..618693cf 100644
--- a/include/asm-s390/tlb.h
+++ b/include/asm-s390/tlb.h
@@ -2,19 +2,130 @@
 #define _S390_TLB_H
 
 /*
- * s390 doesn't need any special per-pte or
- * per-vma handling..
+ * TLB flushing on s390 is complicated. The following requirement
+ * from the principles of operation is the most arduous:
+ *
+ * "A valid table entry must not be changed while it is attached
+ * to any CPU and may be used for translation by that CPU except to
+ * (1) invalidate the entry by using INVALIDATE PAGE TABLE ENTRY,
+ * or INVALIDATE DAT TABLE ENTRY, (2) alter bits 56-63 of a page
+ * table entry, or (3) make a change by means of a COMPARE AND SWAP
+ * AND PURGE instruction that purges the TLB."
+ *
+ * The modification of a pte of an active mm struct therefore is
+ * a two step process: i) invalidate the pte, ii) store the new pte.
+ * This is true for the page protection bit as well.
+ * The only possible optimization is to flush at the beginning of
+ * a tlb_gather_mmu cycle if the mm_struct is currently not in use.
+ *
+ * Pages used for the page tables is a different story. FIXME: more
  */
-#define tlb_start_vma(tlb, vma) do { } while (0)
-#define tlb_end_vma(tlb, vma) do { } while (0)
-#define __tlb_remove_tlb_entry(tlb, ptep, address) do { } while (0)
+
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <asm/processor.h>
+#include <asm/pgalloc.h>
+#include <asm/smp.h>
+#include <asm/tlbflush.h>
+
+#ifndef CONFIG_SMP
+#define TLB_NR_PTRS	1
+#else
+#define TLB_NR_PTRS	508
+#endif
+
+struct mmu_gather {
+	struct mm_struct *mm;
+	unsigned int fullmm;
+	unsigned int nr_ptes;
+	unsigned int nr_pmds;
+	void *array[TLB_NR_PTRS];
+};
+
+DECLARE_PER_CPU(struct mmu_gather, mmu_gathers);
+
+static inline struct mmu_gather *tlb_gather_mmu(struct mm_struct *mm,
+						unsigned int full_mm_flush)
+{
+	struct mmu_gather *tlb = &get_cpu_var(mmu_gathers);
+
+	tlb->mm = mm;
+	tlb->fullmm = full_mm_flush || (num_online_cpus() == 1) ||
+		(atomic_read(&mm->mm_users) <= 1 && mm == current->active_mm);
+	tlb->nr_ptes = 0;
+	tlb->nr_pmds = TLB_NR_PTRS;
+	if (tlb->fullmm)
+		__tlb_flush_mm(mm);
+	return tlb;
+}
+
+static inline void tlb_flush_mmu(struct mmu_gather *tlb,
+				 unsigned long start, unsigned long end)
+{
+	if (!tlb->fullmm && (tlb->nr_ptes > 0 || tlb->nr_pmds < TLB_NR_PTRS))
+		__tlb_flush_mm(tlb->mm);
+	while (tlb->nr_ptes > 0)
+		pte_free(tlb->array[--tlb->nr_ptes]);
+	while (tlb->nr_pmds < TLB_NR_PTRS)
+		pmd_free((pmd_t *) tlb->array[tlb->nr_pmds++]);
+}
+
+static inline void tlb_finish_mmu(struct mmu_gather *tlb,
+				  unsigned long start, unsigned long end)
+{
+	tlb_flush_mmu(tlb, start, end);
+
+	/* keep the page table cache within bounds */
+	check_pgt_cache();
+
+	put_cpu_var(mmu_gathers);
+}
 
 /*
- * .. because we flush the whole mm when it
- * fills up.
+ * Release the page cache reference for a pte removed by
+ * tlb_ptep_clear_flush. In both flush modes the tlb fo a page cache page
+ * has already been freed, so just do free_page_and_swap_cache.
  */
-#define tlb_flush(tlb) flush_tlb_mm((tlb)->mm)
+static inline void tlb_remove_page(struct mmu_gather *tlb, struct page *page)
+{
+	free_page_and_swap_cache(page);
+}
 
-#include <asm-generic/tlb.h>
+/*
+ * 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, struct page *page)
+{
+	if (!tlb->fullmm) {
+		tlb->array[tlb->nr_ptes++] = page;
+		if (tlb->nr_ptes >= tlb->nr_pmds)
+			tlb_flush_mmu(tlb, 0, 0);
+	} else
+		pte_free(page);
+}
 
+/*
+ * pmd_free_tlb frees a pmd table and clears the CRSTE for the
+ * segment table entry from the tlb.
+ */
+static inline void pmd_free_tlb(struct mmu_gather *tlb, pmd_t *pmd)
+{
+#ifdef __s390x__
+	if (!tlb->fullmm) {
+		tlb->array[--tlb->nr_pmds] = (struct page *) pmd;
+		if (tlb->nr_ptes >= tlb->nr_pmds)
+			tlb_flush_mmu(tlb, 0, 0);
+	} else
+		pmd_free(pmd);
 #endif
+}
+
+#define pud_free_tlb(tlb, pud)			do { } while (0)
+
+#define tlb_start_vma(tlb, vma)			do { } while (0)
+#define tlb_end_vma(tlb, vma)			do { } while (0)
+#define tlb_remove_tlb_entry(tlb, ptep, addr)	do { } while (0)
+#define tlb_migrate_finish(mm)			do { } while (0)
+
+#endif /* _S390_TLB_H */
diff --git a/include/asm-s390/tlbflush.h b/include/asm-s390/tlbflush.h
index 6de2632..a69bd24 100644
--- a/include/asm-s390/tlbflush.h
+++ b/include/asm-s390/tlbflush.h
@@ -6,68 +6,19 @@
 #include <asm/pgalloc.h>
 
 /*
- * TLB flushing:
- *
- *  - flush_tlb() flushes the current mm struct TLBs
- *  - flush_tlb_all() flushes all processes TLBs 
- *  - flush_tlb_mm(mm) flushes the specified mm context TLB's
- *  - flush_tlb_page(vma, vmaddr) flushes one page
- *  - flush_tlb_range(vma, start, end) flushes a range of pages
- *  - flush_tlb_kernel_range(start, end) flushes a range of kernel pages
- */
-
-/*
- * S/390 has three ways of flushing TLBs
- * 'ptlb' does a flush of the local processor
- * 'csp' flushes the TLBs on all PUs of a SMP
- * 'ipte' invalidates a pte in a page table and flushes that out of
- * the TLBs of all PUs of a SMP
- */
-
-#define local_flush_tlb() \
-do {  asm volatile("ptlb": : :"memory"); } while (0)
-
-#ifndef CONFIG_SMP
-
-/*
- * We always need to flush, since s390 does not flush tlb
- * on each context switch
+ * Flush all tlb entries on the local cpu.
  */
-
-static inline void flush_tlb(void)
+static inline void __tlb_flush_local(void)
 {
-	local_flush_tlb();
+	asm volatile("ptlb" : : : "memory");
 }
-static inline void flush_tlb_all(void)
-{
-	local_flush_tlb();
-}
-static inline void flush_tlb_mm(struct mm_struct *mm) 
-{
-	local_flush_tlb();
-}
-static inline void flush_tlb_page(struct vm_area_struct *vma,
-				  unsigned long addr)
-{
-	local_flush_tlb();
-}
-static inline void flush_tlb_range(struct vm_area_struct *vma,
-				   unsigned long start, unsigned long end)
-{
-	local_flush_tlb();
-}
-
-#define flush_tlb_kernel_range(start, end) \
-	local_flush_tlb();
-
-#else
 
-#include <asm/smp.h>
-
-extern void smp_ptlb_all(void);
-
-static inline void global_flush_tlb(void)
+/*
+ * Flush all tlb entries on all cpus.
+ */
+static inline void __tlb_flush_global(void)
 {
+	extern void smp_ptlb_all(void);
 	register unsigned long reg2 asm("2");
 	register unsigned long reg3 asm("3");
 	register unsigned long reg4 asm("4");
@@ -89,66 +40,75 @@ static inline void global_flush_tlb(void)
 }
 
 /*
- * We only have to do global flush of tlb if process run since last
- * flush on any other pu than current. 
- * If we have threads (mm->count > 1) we always do a global flush, 
- * since the process runs on more than one processor at the same time.
+ * Flush all tlb entries of a page table on all cpus.
  */
+static inline void __tlb_flush_idte(pgd_t *pgd)
+{
+	asm volatile(
+		"	.insn	rrf,0xb98e0000,0,%0,%1,0"
+		: : "a" (2048), "a" (__pa(pgd) & PAGE_MASK) : "cc" );
+}
 
-static inline void __flush_tlb_mm(struct mm_struct * mm)
+static inline void __tlb_flush_mm(struct mm_struct * mm)
 {
 	cpumask_t local_cpumask;
 
 	if (unlikely(cpus_empty(mm->cpu_vm_mask)))
 		return;
+	/*
+	 * If the machine has IDTE we prefer to do a per mm flush
+	 * on all cpus instead of doing a local flush if the mm
+	 * only ran on the local cpu.
+	 */
 	if (MACHINE_HAS_IDTE) {
-		pgd_t *shadow_pgd = get_shadow_pgd(mm->pgd);
+		pgd_t *shadow_pgd = get_shadow_table(mm->pgd);
 
-		if (shadow_pgd) {
-			asm volatile(
-				"	.insn	rrf,0xb98e0000,0,%0,%1,0"
-				: : "a" (2048),
-				"a" (__pa(shadow_pgd) & PAGE_MASK) : "cc" );
-		}
-		asm volatile(
-			"	.insn	rrf,0xb98e0000,0,%0,%1,0"
-			: : "a" (2048), "a" (__pa(mm->pgd)&PAGE_MASK) : "cc");
+		if (shadow_pgd)
+			__tlb_flush_idte(shadow_pgd);
+		__tlb_flush_idte(mm->pgd);
 		return;
 	}
 	preempt_disable();
+	/*
+	 * If the process only ran on the local cpu, do a local flush.
+	 */
 	local_cpumask = cpumask_of_cpu(smp_processor_id());
 	if (cpus_equal(mm->cpu_vm_mask, local_cpumask))
-		local_flush_tlb();
+		__tlb_flush_local();
 	else
-		global_flush_tlb();
+		__tlb_flush_global();
 	preempt_enable();
 }
 
-static inline void flush_tlb(void)
-{
-	__flush_tlb_mm(current->mm);
-}
-static inline void flush_tlb_all(void)
-{
-	global_flush_tlb();
-}
-static inline void flush_tlb_mm(struct mm_struct *mm) 
-{
-	__flush_tlb_mm(mm); 
-}
-static inline void flush_tlb_page(struct vm_area_struct *vma,
-				  unsigned long addr)
-{
-	__flush_tlb_mm(vma->vm_mm);
-}
-static inline void flush_tlb_range(struct vm_area_struct *vma,
-				   unsigned long start, unsigned long end)
+static inline void __tlb_flush_mm_cond(struct mm_struct * mm)
 {
-	__flush_tlb_mm(vma->vm_mm); 
+	if (atomic_read(&mm->mm_users) <= 1 && mm == current->active_mm)
+		__tlb_flush_mm(mm);
 }
 
-#define flush_tlb_kernel_range(start, end) global_flush_tlb()
+/*
+ * TLB flushing:
+ *  flush_tlb() - flushes the current mm struct TLBs
+ *  flush_tlb_all() - flushes all processes TLBs
+ *  flush_tlb_mm(mm) - flushes the specified mm context TLB's
+ *  flush_tlb_page(vma, vmaddr) - flushes one page
+ *  flush_tlb_range(vma, start, end) - flushes a range of pages
+ *  flush_tlb_kernel_range(start, end) - flushes a range of kernel pages
+ */
 
-#endif
+/*
+ * flush_tlb_mm goes together with ptep_set_wrprotect for the
+ * copy_page_range operation and flush_tlb_range is related to
+ * ptep_get_and_clear for change_protection. ptep_set_wrprotect and
+ * ptep_get_and_clear do not flush the TLBs directly if the mm has
+ * only one user. At the end of the update the flush_tlb_mm and
+ * flush_tlb_range functions need to do the flush.
+ */
+#define flush_tlb()				do { } while (0)
+#define flush_tlb_all()				do { } while (0)
+#define flush_tlb_mm(mm)			__tlb_flush_mm_cond(mm)
+#define flush_tlb_page(vma, addr)		do { } while (0)
+#define flush_tlb_range(vma, start, end)	__tlb_flush_mm_cond(mm)
+#define flush_tlb_kernel_range(start, end)	__tlb_flush_mm(&init_mm)
 
 #endif /* _S390_TLBFLUSH_H */