[S390] Cleanup page table definitions.

- De-confuse the defines for the address-space-control-elements
  and the segment/region table entries.
- Create out of line functions for page table allocation / freeing.
- Simplify get_shadow_xxx functions.
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>

arch/s390/mm/Makefile

@@ -2,6 +2,6 @@
 # Makefile for the linux s390-specific parts of the memory manager.
 #
 
-obj-y	 := init.o fault.o extmem.o mmap.o vmem.o
+obj-y	 := init.o fault.o extmem.o mmap.o vmem.o pgtable.o
 obj-$(CONFIG_CMM) += cmm.o
 

arch/s390/mm/init.c

@@ -103,32 +103,28 @@ static void __init setup_ro_region(void)
  */
 void __init paging_init(void)
 {
-	pgd_t *pg_dir;
-	int i;
-	unsigned long pgdir_k;
 	static const int ssm_mask = 0x04000000L;
 	unsigned long max_zone_pfns[MAX_NR_ZONES];
+	unsigned long pgd_type;
 
-	pg_dir = swapper_pg_dir;
-	
+	init_mm.pgd = swapper_pg_dir;
+	S390_lowcore.kernel_asce = __pa(init_mm.pgd) & PAGE_MASK;
 #ifdef CONFIG_64BIT
-	pgdir_k = (__pa(swapper_pg_dir) & PAGE_MASK) | _KERN_REGION_TABLE;
-	for (i = 0; i < PTRS_PER_PGD; i++)
-		pgd_clear_kernel(pg_dir + i);
+	S390_lowcore.kernel_asce |= _ASCE_TYPE_REGION3 | _ASCE_TABLE_LENGTH;
+	pgd_type = _REGION3_ENTRY_EMPTY;
 #else
-	pgdir_k = (__pa(swapper_pg_dir) & PAGE_MASK) | _KERNSEG_TABLE;
-	for (i = 0; i < PTRS_PER_PGD; i++)
-		pmd_clear_kernel((pmd_t *)(pg_dir + i));
+	S390_lowcore.kernel_asce |= _ASCE_TABLE_LENGTH;
+	pgd_type = _SEGMENT_ENTRY_EMPTY;
 #endif
+	clear_table((unsigned long *) init_mm.pgd, pgd_type,
+		    sizeof(unsigned long)*2048);
 	vmem_map_init();
 	setup_ro_region();
 
-	S390_lowcore.kernel_asce = pgdir_k;
-
         /* enable virtual mapping in kernel mode */
-	__ctl_load(pgdir_k, 1, 1);
-	__ctl_load(pgdir_k, 7, 7);
-	__ctl_load(pgdir_k, 13, 13);
+	__ctl_load(S390_lowcore.kernel_asce, 1, 1);
+	__ctl_load(S390_lowcore.kernel_asce, 7, 7);
+	__ctl_load(S390_lowcore.kernel_asce, 13, 13);
 	__raw_local_irq_ssm(ssm_mask);
 
 	memset(max_zone_pfns, 0, sizeof(max_zone_pfns));

arch/s390/mm/pgtable.c

+/*
+ *  arch/s390/mm/pgtable.c
+ *
+ *    Copyright IBM Corp. 2007
+ *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
+ */
+
+#include <linux/sched.h>
+#include <linux/kernel.h>
+#include <linux/errno.h>
+#include <linux/mm.h>
+#include <linux/swap.h>
+#include <linux/smp.h>
+#include <linux/highmem.h>
+#include <linux/slab.h>
+#include <linux/pagemap.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/quicklist.h>
+
+#include <asm/system.h>
+#include <asm/pgtable.h>
+#include <asm/pgalloc.h>
+#include <asm/tlb.h>
+#include <asm/tlbflush.h>
+
+#ifndef CONFIG_64BIT
+#define ALLOC_ORDER	1
+#else
+#define ALLOC_ORDER	2
+#endif
+
+unsigned long *crst_table_alloc(struct mm_struct *mm, int noexec)
+{
+	struct page *page = alloc_pages(GFP_KERNEL, ALLOC_ORDER);
+
+	if (!page)
+		return NULL;
+	page->index = 0;
+	if (noexec) {
+		struct page *shadow = alloc_pages(GFP_KERNEL, ALLOC_ORDER);
+		if (!shadow) {
+			__free_pages(page, ALLOC_ORDER);
+			return NULL;
+		}
+		page->index = page_to_phys(shadow);
+	}
+	return (unsigned long *) page_to_phys(page);
+}
+
+void crst_table_free(unsigned long *table)
+{
+	unsigned long *shadow = get_shadow_table(table);
+
+	if (shadow)
+		free_pages((unsigned long) shadow, ALLOC_ORDER);
+	free_pages((unsigned long) table, ALLOC_ORDER);
+}
+
+/*
+ * page table entry allocation/free routines.
+ */
+unsigned long *page_table_alloc(int noexec)
+{
+	struct page *page = alloc_page(GFP_KERNEL);
+	unsigned long *table;
+
+	if (!page)
+		return NULL;
+	page->index = 0;
+	if (noexec) {
+		struct page *shadow = alloc_page(GFP_KERNEL);
+		if (!shadow) {
+			__free_page(page);
+			return NULL;
+		}
+		table = (unsigned long *) page_to_phys(shadow);
+		clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
+		page->index = (addr_t) table;
+	}
+	table = (unsigned long *) page_to_phys(page);
+	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
+	return table;
+}
+
+void page_table_free(unsigned long *table)
+{
+	unsigned long *shadow = get_shadow_pte(table);
+
+	if (shadow)
+		free_page((unsigned long) shadow);
+	free_page((unsigned long) table);
+
+}

arch/s390/mm/vmem.c

@@ -75,29 +75,24 @@ static void __init_refok *vmem_alloc_pages(unsigned int order)
 
 static inline pmd_t *vmem_pmd_alloc(void)
 {
-	pmd_t *pmd;
-	int i;
+	pmd_t *pmd = NULL;
 
-	pmd = vmem_alloc_pages(PMD_ALLOC_ORDER);
+#ifdef CONFIG_64BIT
+	pmd = vmem_alloc_pages(2);
 	if (!pmd)
 		return NULL;
-	for (i = 0; i < PTRS_PER_PMD; i++)
-		pmd_clear_kernel(pmd + i);
+	clear_table((unsigned long *) pmd, _SEGMENT_ENTRY_EMPTY, PAGE_SIZE*4);
+#endif
 	return pmd;
 }
 
 static inline pte_t *vmem_pte_alloc(void)
 {
-	pte_t *pte;
-	pte_t empty_pte;
-	int i;
+	pte_t *pte = vmem_alloc_pages(0);
 
-	pte = vmem_alloc_pages(PTE_ALLOC_ORDER);
 	if (!pte)
 		return NULL;
-	pte_val(empty_pte) = _PAGE_TYPE_EMPTY;
-	for (i = 0; i < PTRS_PER_PTE; i++)
-		pte[i] = empty_pte;
+	clear_table((unsigned long *) pte, _PAGE_TYPE_EMPTY, PAGE_SIZE);
 	return pte;
 }
 

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 switch_mm(struct mm_struct *prev, struct mm_struct *next,
-			     struct task_struct *tsk)
+static inline void update_mm(struct mm_struct *mm, struct task_struct *tsk)
 {
-	pgd_t *shadow_pgd = get_shadow_pgd(next->pgd);
+	pgd_t *pgd = mm->pgd;
+	unsigned long asce_bits;
 
-	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) {
+	/* 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. */
-			asm volatile(LCTL_OPCODE" 1,1,%0"
-				     : : "m" (S390_lowcore.user_asce) );
+		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)
+{
+	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,

include/asm-s390/pgalloc.h

@@ -19,114 +19,75 @@
 
 #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.
- */
+
+static inline unsigned long pgd_entry_type(struct mm_struct *mm)
+{
+	return _SEGMENT_ENTRY_EMPTY;
+}
+
 #define pmd_alloc_one(mm,address)		({ BUG(); ((pmd_t *)2); })
 #define pmd_free(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 _REGION3_ENTRY_EMPTY;
 }
 
-static inline void pmd_free (pmd_t *pmd)
+static inline pmd_t *pmd_alloc_one(struct mm_struct *mm, unsigned long vmaddr)
 {
-	pmd_t *shadow_pmd = get_shadow_pmd(pmd);
-
-	if (shadow_pmd)
-		free_pages((unsigned long) shadow_pmd, PMD_ALLOC_ORDER);
-	free_pages((unsigned long) pmd, PMD_ALLOC_ORDER);
+	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)
 
-static inline void
-pgd_populate_kernel(struct mm_struct *mm, pgd_t *pgd, pmd_t *pmd)
+static inline void pgd_populate_kernel(struct mm_struct *mm,
+				       pgd_t *pgd, pmd_t *pmd)
 {
-	pgd_val(*pgd) = _PGD_ENTRY | __pa(pmd);
+	pgd_val(*pgd) = _REGION3_ENTRY | __pa(pmd);
 }
 
 static inline void pgd_populate(struct mm_struct *mm, pgd_t *pgd, pmd_t *pmd)
 {
-	pgd_t *shadow_pgd = get_shadow_pgd(pgd);
-	pmd_t *shadow_pmd = get_shadow_pmd(pmd);
+	pgd_t *shadow_pgd = get_shadow_table(pgd);
+	pmd_t *shadow_pmd = get_shadow_table(pmd);
 
 	if (shadow_pgd && shadow_pmd)
 		pgd_populate_kernel(mm, shadow_pgd, shadow_pmd);
@@ -135,17 +96,26 @@ static inline void pgd_populate(struct mm_struct *mm, pgd_t *pgd, pmd_t *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__ */
 }
 
@@ -153,7 +123,7 @@ 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);
@@ -164,57 +134,14 @@ 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_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 */

include/asm-s390/processor.h

@@ -127,12 +127,6 @@ 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 {							\
 	.ksp = sizeof(init_stack) + (unsigned long) &init_stack,	\
 }

include/asm-s390/tlbflush.h

@@ -61,7 +61,7 @@ static inline void __tlb_flush_mm(struct mm_struct * 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)
 			__tlb_flush_idte(shadow_pgd);