[mirror_ubuntu-artful-kernel.git] / 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>
#include <asm/mmu_context.h>

#ifndef CONFIG_64BIT
#define ALLOC_ORDER	1
#define TABLES_PER_PAGE	4
#define FRAG_MASK	15UL
#define SECOND_HALVES	10UL

void clear_table_pgstes(unsigned long *table)
{
	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
	memset(table + 256, 0, PAGE_SIZE/4);
	clear_table(table + 512, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
	memset(table + 768, 0, PAGE_SIZE/4);
}

#else
#define ALLOC_ORDER	2
#define TABLES_PER_PAGE	2
#define FRAG_MASK	3UL
#define SECOND_HALVES	2UL

void clear_table_pgstes(unsigned long *table)
{
	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/2);
	memset(table + 256, 0, PAGE_SIZE/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);
	}
	spin_lock(&mm->page_table_lock);
	list_add(&page->lru, &mm->context.crst_list);
	spin_unlock(&mm->page_table_lock);
	return (unsigned long *) page_to_phys(page);
}

void crst_table_free(struct mm_struct *mm, unsigned long *table)
{
	unsigned long *shadow = get_shadow_table(table);
	struct page *page = virt_to_page(table);

	spin_lock(&mm->page_table_lock);
	list_del(&page->lru);
	spin_unlock(&mm->page_table_lock);
	if (shadow)
		free_pages((unsigned long) shadow, ALLOC_ORDER);
	free_pages((unsigned long) table, ALLOC_ORDER);
}

#ifdef CONFIG_64BIT
int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
{
	unsigned long *table, *pgd;
	unsigned long entry;

	BUG_ON(limit > (1UL << 53));
repeat:
	table = crst_table_alloc(mm, mm->context.noexec);
	if (!table)
		return -ENOMEM;
	spin_lock(&mm->page_table_lock);
	if (mm->context.asce_limit < limit) {
		pgd = (unsigned long *) mm->pgd;
		if (mm->context.asce_limit <= (1UL << 31)) {
			entry = _REGION3_ENTRY_EMPTY;
			mm->context.asce_limit = 1UL << 42;
			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
						_ASCE_USER_BITS |
						_ASCE_TYPE_REGION3;
		} else {
			entry = _REGION2_ENTRY_EMPTY;
			mm->context.asce_limit = 1UL << 53;
			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
						_ASCE_USER_BITS |
						_ASCE_TYPE_REGION2;
		}
		crst_table_init(table, entry);
		pgd_populate(mm, (pgd_t *) table, (pud_t *) pgd);
		mm->pgd = (pgd_t *) table;
		mm->task_size = mm->context.asce_limit;
		table = NULL;
	}
	spin_unlock(&mm->page_table_lock);
	if (table)
		crst_table_free(mm, table);
	if (mm->context.asce_limit < limit)
		goto repeat;
	update_mm(mm, current);
	return 0;
}

void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
{
	pgd_t *pgd;

	if (mm->context.asce_limit <= limit)
		return;
	__tlb_flush_mm(mm);
	while (mm->context.asce_limit > limit) {
		pgd = mm->pgd;
		switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
		case _REGION_ENTRY_TYPE_R2:
			mm->context.asce_limit = 1UL << 42;
			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
						_ASCE_USER_BITS |
						_ASCE_TYPE_REGION3;
			break;
		case _REGION_ENTRY_TYPE_R3:
			mm->context.asce_limit = 1UL << 31;
			mm->context.asce_bits = _ASCE_TABLE_LENGTH |
						_ASCE_USER_BITS |
						_ASCE_TYPE_SEGMENT;
			break;
		default:
			BUG();
		}
		mm->pgd = (pgd_t *) (pgd_val(*pgd) & _REGION_ENTRY_ORIGIN);
		mm->task_size = mm->context.asce_limit;
		crst_table_free(mm, (unsigned long *) pgd);
	}
	update_mm(mm, current);
}
#endif

/*
 * page table entry allocation/free routines.
 */
unsigned long *page_table_alloc(struct mm_struct *mm)
{
	struct page *page;
	unsigned long *table;
	unsigned long bits;

	bits = (mm->context.noexec || mm->context.has_pgste) ? 3UL : 1UL;
	spin_lock(&mm->page_table_lock);
	page = NULL;
	if (!list_empty(&mm->context.pgtable_list)) {
		page = list_first_entry(&mm->context.pgtable_list,
					struct page, lru);
		if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
			page = NULL;
	}
	if (!page) {
		spin_unlock(&mm->page_table_lock);
		page = alloc_page(GFP_KERNEL|__GFP_REPEAT);
		if (!page)
			return NULL;
		pgtable_page_ctor(page);
		page->flags &= ~FRAG_MASK;
		table = (unsigned long *) page_to_phys(page);
		if (mm->context.has_pgste)
			clear_table_pgstes(table);
		else
			clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
		spin_lock(&mm->page_table_lock);
		list_add(&page->lru, &mm->context.pgtable_list);
	}
	table = (unsigned long *) page_to_phys(page);
	while (page->flags & bits) {
		table += 256;
		bits <<= 1;
	}
	page->flags |= bits;
	if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
		list_move_tail(&page->lru, &mm->context.pgtable_list);
	spin_unlock(&mm->page_table_lock);
	return table;
}

void page_table_free(struct mm_struct *mm, unsigned long *table)
{
	struct page *page;
	unsigned long bits;

	bits = (mm->context.noexec || mm->context.has_pgste) ? 3UL : 1UL;
	bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
	spin_lock(&mm->page_table_lock);
	page->flags ^= bits;
	if (page->flags & FRAG_MASK) {
		/* Page now has some free pgtable fragments. */
		list_move(&page->lru, &mm->context.pgtable_list);
		page = NULL;
	} else
		/* All fragments of the 4K page have been freed. */
		list_del(&page->lru);
	spin_unlock(&mm->page_table_lock);
	if (page) {
		pgtable_page_dtor(page);
		__free_page(page);
	}
}

void disable_noexec(struct mm_struct *mm, struct task_struct *tsk)
{
	struct page *page;

	spin_lock(&mm->page_table_lock);
	/* Free shadow region and segment tables. */
	list_for_each_entry(page, &mm->context.crst_list, lru)
		if (page->index) {
			free_pages((unsigned long) page->index, ALLOC_ORDER);
			page->index = 0;
		}
	/* "Free" second halves of page tables. */
	list_for_each_entry(page, &mm->context.pgtable_list, lru)
		page->flags &= ~SECOND_HALVES;
	spin_unlock(&mm->page_table_lock);
	mm->context.noexec = 0;
	update_mm(mm, tsk);
}

/*
 * switch on pgstes for its userspace process (for kvm)
 */
int s390_enable_sie(void)
{
	struct task_struct *tsk = current;
	struct mm_struct *mm, *old_mm;

	/* Do we have switched amode? If no, we cannot do sie */
	if (!switch_amode)
		return -EINVAL;

	/* Do we have pgstes? if yes, we are done */
	if (tsk->mm->context.has_pgste)
		return 0;

	/* lets check if we are allowed to replace the mm */
	task_lock(tsk);
	if (!tsk->mm || atomic_read(&tsk->mm->mm_users) > 1 ||
	    tsk->mm != tsk->active_mm || !hlist_empty(&tsk->mm->ioctx_list)) {
		task_unlock(tsk);
		return -EINVAL;
	}
	task_unlock(tsk);

	/* we copy the mm and let dup_mm create the page tables with_pgstes */
	tsk->mm->context.alloc_pgste = 1;
	mm = dup_mm(tsk);
	tsk->mm->context.alloc_pgste = 0;
	if (!mm)
		return -ENOMEM;

	/* Now lets check again if something happened */
	task_lock(tsk);
	if (!tsk->mm || atomic_read(&tsk->mm->mm_users) > 1 ||
	    tsk->mm != tsk->active_mm || !hlist_empty(&tsk->mm->ioctx_list)) {
		mmput(mm);
		task_unlock(tsk);
		return -EINVAL;
	}

	/* ok, we are alone. No ptrace, no threads, etc. */
	old_mm = tsk->mm;
	tsk->mm = tsk->active_mm = mm;
	preempt_disable();
	update_mm(mm, tsk);
	cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm));
	preempt_enable();
	task_unlock(tsk);
	mmput(old_mm);
	return 0;
}
EXPORT_SYMBOL_GPL(s390_enable_sie);
Commit	Line	Data
3610cce8 MS	1	/*
	2	* arch/s390/mm/pgtable.c
	3	*
	4	* Copyright IBM Corp. 2007
	5	* Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
	6	*/
	7
	8	#include <linux/sched.h>
	9	#include <linux/kernel.h>
	10	#include <linux/errno.h>
	11	#include <linux/mm.h>
	12	#include <linux/swap.h>
	13	#include <linux/smp.h>
	14	#include <linux/highmem.h>
	15	#include <linux/slab.h>
	16	#include <linux/pagemap.h>
	17	#include <linux/spinlock.h>
	18	#include <linux/module.h>
	19	#include <linux/quicklist.h>
	20
	21	#include <asm/system.h>
	22	#include <asm/pgtable.h>
	23	#include <asm/pgalloc.h>
	24	#include <asm/tlb.h>
	25	#include <asm/tlbflush.h>
6252d702	26	#include <asm/mmu_context.h>
3610cce8 MS	27
	28	#ifndef CONFIG_64BIT
	29	#define ALLOC_ORDER 1
146e4b3c MS	30	#define TABLES_PER_PAGE 4
	31	#define FRAG_MASK 15UL
	32	#define SECOND_HALVES 10UL
402b0862 CO	33
	34	void clear_table_pgstes(unsigned long *table)
	35	{
	36	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
	37	memset(table + 256, 0, PAGE_SIZE/4);
	38	clear_table(table + 512, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
	39	memset(table + 768, 0, PAGE_SIZE/4);
	40	}
	41
3610cce8 MS	42	#else
3610cce8 MS	43	#define ALLOC_ORDER 2
146e4b3c MS	44	#define TABLES_PER_PAGE 2
	45	#define FRAG_MASK 3UL
	46	#define SECOND_HALVES 2UL
402b0862 CO	47
	48	void clear_table_pgstes(unsigned long *table)
	49	{
	50	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/2);
	51	memset(table + 256, 0, PAGE_SIZE/2);
	52	}
	53
3610cce8 MS	54	#endif
	55
	56	unsigned long crst_table_alloc(struct mm_struct mm, int noexec)
	57	{
	58	struct page *page = alloc_pages(GFP_KERNEL, ALLOC_ORDER);
	59
	60	if (!page)
	61	return NULL;
	62	page->index = 0;
	63	if (noexec) {
	64	struct page *shadow = alloc_pages(GFP_KERNEL, ALLOC_ORDER);
	65	if (!shadow) {
	66	__free_pages(page, ALLOC_ORDER);
	67	return NULL;
	68	}
	69	page->index = page_to_phys(shadow);
	70	}
146e4b3c MS	71	spin_lock(&mm->page_table_lock);
	72	list_add(&page->lru, &mm->context.crst_list);
	73	spin_unlock(&mm->page_table_lock);
3610cce8 MS	74	return (unsigned long *) page_to_phys(page);
	75	}
	76
146e4b3c	77	void crst_table_free(struct mm_struct mm, unsigned long table)
3610cce8 MS	78	{
3610cce8 MS	79	unsigned long *shadow = get_shadow_table(table);
146e4b3c	80	struct page *page = virt_to_page(table);
3610cce8	81
146e4b3c MS	82	spin_lock(&mm->page_table_lock);
	83	list_del(&page->lru);
	84	spin_unlock(&mm->page_table_lock);
3610cce8 MS	85	if (shadow)
	86	free_pages((unsigned long) shadow, ALLOC_ORDER);
	87	free_pages((unsigned long) table, ALLOC_ORDER);
	88	}
	89
6252d702 MS	90	#ifdef CONFIG_64BIT
	91	int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
	92	{
	93	unsigned long table, pgd;
	94	unsigned long entry;
	95
	96	BUG_ON(limit > (1UL << 53));
	97	repeat:
	98	table = crst_table_alloc(mm, mm->context.noexec);
	99	if (!table)
	100	return -ENOMEM;
	101	spin_lock(&mm->page_table_lock);
	102	if (mm->context.asce_limit < limit) {
	103	pgd = (unsigned long *) mm->pgd;
	104	if (mm->context.asce_limit <= (1UL << 31)) {
	105	entry = _REGION3_ENTRY_EMPTY;
	106	mm->context.asce_limit = 1UL << 42;
	107	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	108	_ASCE_USER_BITS \|
	109	_ASCE_TYPE_REGION3;
	110	} else {
	111	entry = _REGION2_ENTRY_EMPTY;
	112	mm->context.asce_limit = 1UL << 53;
	113	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	114	_ASCE_USER_BITS \|
	115	_ASCE_TYPE_REGION2;
	116	}
	117	crst_table_init(table, entry);
	118	pgd_populate(mm, (pgd_t ) table, (pud_t ) pgd);
	119	mm->pgd = (pgd_t *) table;
f481bfaf	120	mm->task_size = mm->context.asce_limit;
6252d702 MS	121	table = NULL;
	122	}
	123	spin_unlock(&mm->page_table_lock);
	124	if (table)
	125	crst_table_free(mm, table);
	126	if (mm->context.asce_limit < limit)
	127	goto repeat;
	128	update_mm(mm, current);
	129	return 0;
	130	}
	131
	132	void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
	133	{
	134	pgd_t *pgd;
	135
	136	if (mm->context.asce_limit <= limit)
	137	return;
	138	__tlb_flush_mm(mm);
	139	while (mm->context.asce_limit > limit) {
	140	pgd = mm->pgd;
	141	switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
	142	case _REGION_ENTRY_TYPE_R2:
	143	mm->context.asce_limit = 1UL << 42;
	144	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	145	_ASCE_USER_BITS \|
	146	_ASCE_TYPE_REGION3;
	147	break;
	148	case _REGION_ENTRY_TYPE_R3:
	149	mm->context.asce_limit = 1UL << 31;
	150	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	151	_ASCE_USER_BITS \|
	152	_ASCE_TYPE_SEGMENT;
	153	break;
	154	default:
	155	BUG();
	156	}
	157	mm->pgd = (pgd_t ) (pgd_val(pgd) & _REGION_ENTRY_ORIGIN);
f481bfaf	158	mm->task_size = mm->context.asce_limit;
6252d702 MS	159	crst_table_free(mm, (unsigned long *) pgd);
	160	}
	161	update_mm(mm, current);
	162	}
	163	#endif
	164
3610cce8 MS	165	/*
	166	* page table entry allocation/free routines.
	167	*/
146e4b3c	168	unsigned long page_table_alloc(struct mm_struct mm)
3610cce8	169	{
146e4b3c	170	struct page *page;
3610cce8	171	unsigned long *table;
146e4b3c	172	unsigned long bits;
3610cce8	173
250cf776	174	bits = (mm->context.noexec \|\| mm->context.has_pgste) ? 3UL : 1UL;
146e4b3c MS	175	spin_lock(&mm->page_table_lock);
	176	page = NULL;
	177	if (!list_empty(&mm->context.pgtable_list)) {
	178	page = list_first_entry(&mm->context.pgtable_list,
	179	struct page, lru);
	180	if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
	181	page = NULL;
	182	}
	183	if (!page) {
	184	spin_unlock(&mm->page_table_lock);
	185	page = alloc_page(GFP_KERNEL\|__GFP_REPEAT);
	186	if (!page)
3610cce8	187	return NULL;
146e4b3c MS	188	pgtable_page_ctor(page);
	189	page->flags &= ~FRAG_MASK;
	190	table = (unsigned long *) page_to_phys(page);
250cf776	191	if (mm->context.has_pgste)
402b0862 CO	192	clear_table_pgstes(table);
	193	else
	194	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
146e4b3c MS	195	spin_lock(&mm->page_table_lock);
146e4b3c MS	196	list_add(&page->lru, &mm->context.pgtable_list);
3610cce8 MS	197	}
3610cce8 MS	198	table = (unsigned long *) page_to_phys(page);
146e4b3c MS	199	while (page->flags & bits) {
	200	table += 256;
	201	bits <<= 1;
	202	}
	203	page->flags \|= bits;
	204	if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
	205	list_move_tail(&page->lru, &mm->context.pgtable_list);
	206	spin_unlock(&mm->page_table_lock);
3610cce8 MS	207	return table;
	208	}
	209
146e4b3c	210	void page_table_free(struct mm_struct mm, unsigned long table)
3610cce8	211	{
146e4b3c MS	212	struct page *page;
146e4b3c MS	213	unsigned long bits;
3610cce8	214
250cf776	215	bits = (mm->context.noexec \|\| mm->context.has_pgste) ? 3UL : 1UL;
146e4b3c MS	216	bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
	217	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
	218	spin_lock(&mm->page_table_lock);
	219	page->flags ^= bits;
	220	if (page->flags & FRAG_MASK) {
	221	/* Page now has some free pgtable fragments. */
	222	list_move(&page->lru, &mm->context.pgtable_list);
	223	page = NULL;
	224	} else
	225	/* All fragments of the 4K page have been freed. */
	226	list_del(&page->lru);
	227	spin_unlock(&mm->page_table_lock);
	228	if (page) {
	229	pgtable_page_dtor(page);
	230	__free_page(page);
	231	}
	232	}
3610cce8	233
146e4b3c MS	234	void disable_noexec(struct mm_struct mm, struct task_struct tsk)
	235	{
	236	struct page *page;
	237
	238	spin_lock(&mm->page_table_lock);
	239	/* Free shadow region and segment tables. */
	240	list_for_each_entry(page, &mm->context.crst_list, lru)
	241	if (page->index) {
	242	free_pages((unsigned long) page->index, ALLOC_ORDER);
	243	page->index = 0;
	244	}
	245	/* "Free" second halves of page tables. */
	246	list_for_each_entry(page, &mm->context.pgtable_list, lru)
	247	page->flags &= ~SECOND_HALVES;
	248	spin_unlock(&mm->page_table_lock);
	249	mm->context.noexec = 0;
	250	update_mm(mm, tsk);
3610cce8	251	}
402b0862 CO	252
	253	/*
	254	* switch on pgstes for its userspace process (for kvm)
	255	*/
	256	int s390_enable_sie(void)
	257	{
	258	struct task_struct *tsk = current;
74b6b522	259	struct mm_struct mm, old_mm;
402b0862	260
702d9e58 CO	261	/* Do we have switched amode? If no, we cannot do sie */
	262	if (!switch_amode)
	263	return -EINVAL;
	264
74b6b522	265	/* Do we have pgstes? if yes, we are done */
250cf776	266	if (tsk->mm->context.has_pgste)
74b6b522	267	return 0;
402b0862	268
74b6b522 CB	269	/* lets check if we are allowed to replace the mm */
74b6b522 CB	270	task_lock(tsk);
402b0862	271	if (!tsk->mm \|\| atomic_read(&tsk->mm->mm_users) > 1 \|\|
abf137dd	272	tsk->mm != tsk->active_mm \|\| !hlist_empty(&tsk->mm->ioctx_list)) {
74b6b522 CB	273	task_unlock(tsk);
	274	return -EINVAL;
	275	}
	276	task_unlock(tsk);
402b0862	277
250cf776 CB	278	/* we copy the mm and let dup_mm create the page tables with_pgstes */
250cf776 CB	279	tsk->mm->context.alloc_pgste = 1;
402b0862	280	mm = dup_mm(tsk);
250cf776	281	tsk->mm->context.alloc_pgste = 0;
402b0862	282	if (!mm)
74b6b522 CB	283	return -ENOMEM;
74b6b522 CB	284
250cf776	285	/* Now lets check again if something happened */
74b6b522 CB	286	task_lock(tsk);
74b6b522 CB	287	if (!tsk->mm \|\| atomic_read(&tsk->mm->mm_users) > 1 \|\|
abf137dd	288	tsk->mm != tsk->active_mm \|\| !hlist_empty(&tsk->mm->ioctx_list)) {
74b6b522 CB	289	mmput(mm);
	290	task_unlock(tsk);
	291	return -EINVAL;
	292	}
	293
	294	/* ok, we are alone. No ptrace, no threads, etc. */
	295	old_mm = tsk->mm;
402b0862 CO	296	tsk->mm = tsk->active_mm = mm;
	297	preempt_disable();
	298	update_mm(mm, tsk);
005f8eee	299	cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm));
402b0862	300	preempt_enable();
402b0862	301	task_unlock(tsk);
74b6b522 CB	302	mmput(old_mm);
74b6b522 CB	303	return 0;
402b0862 CO	304	}
402b0862 CO	305	EXPORT_SYMBOL_GPL(s390_enable_sie);