[mirror_ubuntu-hirsute-kernel.git] / arch / s390 / mm / pgtable.c

/*
 *    Copyright IBM Corp. 2007,2009
 *    Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
 */

#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/gfp.h>
#include <linux/mm.h>
#include <linux/swap.h>
#include <linux/smp.h>
#include <linux/highmem.h>
#include <linux/pagemap.h>
#include <linux/spinlock.h>
#include <linux/module.h>
#include <linux/quicklist.h>
#include <linux/rcupdate.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>

struct rcu_table_freelist {
	struct rcu_head rcu;
	struct mm_struct *mm;
	unsigned int pgt_index;
	unsigned int crst_index;
	unsigned long *table[0];
};

#define RCU_FREELIST_SIZE \
	((PAGE_SIZE - sizeof(struct rcu_table_freelist)) \
	  / sizeof(unsigned long))

static DEFINE_PER_CPU(struct rcu_table_freelist *, rcu_table_freelist);

static void __page_table_free(struct mm_struct *mm, unsigned long *table);

static struct rcu_table_freelist *rcu_table_freelist_get(struct mm_struct *mm)
{
	struct rcu_table_freelist **batchp = &__get_cpu_var(rcu_table_freelist);
	struct rcu_table_freelist *batch = *batchp;

	if (batch)
		return batch;
	batch = (struct rcu_table_freelist *) __get_free_page(GFP_ATOMIC);
	if (batch) {
		batch->mm = mm;
		batch->pgt_index = 0;
		batch->crst_index = RCU_FREELIST_SIZE;
		*batchp = batch;
	}
	return batch;
}

static void rcu_table_freelist_callback(struct rcu_head *head)
{
	struct rcu_table_freelist *batch =
		container_of(head, struct rcu_table_freelist, rcu);

	while (batch->pgt_index > 0)
		__page_table_free(batch->mm, batch->table[--batch->pgt_index]);
	while (batch->crst_index < RCU_FREELIST_SIZE)
		crst_table_free(batch->mm, batch->table[batch->crst_index++]);
	free_page((unsigned long) batch);
}

void rcu_table_freelist_finish(void)
{
	struct rcu_table_freelist **batchp = &get_cpu_var(rcu_table_freelist);
	struct rcu_table_freelist *batch = *batchp;

	if (!batch)
		goto out;
	call_rcu(&batch->rcu, rcu_table_freelist_callback);
	*batchp = NULL;
out:
	put_cpu_var(rcu_table_freelist);
}

static void smp_sync(void *arg)
{
}

#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 VMALLOC_START = VMALLOC_END - VMALLOC_SIZE;
EXPORT_SYMBOL(VMALLOC_START);

static int __init parse_vmalloc(char *arg)
{
	if (!arg)
		return -EINVAL;
	VMALLOC_START = (VMALLOC_END - memparse(arg, &arg)) & PAGE_MASK;
	return 0;
}
early_param("vmalloc", parse_vmalloc);

unsigned long *crst_table_alloc(struct mm_struct *mm)
{
	struct page *page = alloc_pages(GFP_KERNEL, ALLOC_ORDER);

	if (!page)
		return NULL;
	return (unsigned long *) page_to_phys(page);
}

void crst_table_free(struct mm_struct *mm, unsigned long *table)
{
	free_pages((unsigned long) table, ALLOC_ORDER);
}

void crst_table_free_rcu(struct mm_struct *mm, unsigned long *table)
{
	struct rcu_table_freelist *batch;

	preempt_disable();
	if (atomic_read(&mm->mm_users) < 2 &&
	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
		crst_table_free(mm, table);
		goto out;
	}
	batch = rcu_table_freelist_get(mm);
	if (!batch) {
		smp_call_function(smp_sync, NULL, 1);
		crst_table_free(mm, table);
		goto out;
	}
	batch->table[--batch->crst_index] = table;
	if (batch->pgt_index >= batch->crst_index)
		rcu_table_freelist_finish();
out:
	preempt_enable();
}

#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);
	if (!table)
		return -ENOMEM;
	spin_lock_bh(&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_bh(&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.has_pgste) ? 3UL : 1UL;
	spin_lock_bh(&mm->context.list_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_bh(&mm->context.list_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_bh(&mm->context.list_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_bh(&mm->context.list_lock);
	return table;
}

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

	bits = ((unsigned long) table) & 15;
	table = (unsigned long *)(((unsigned long) table) ^ bits);
	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
	page->flags ^= bits;
	if (!(page->flags & FRAG_MASK)) {
		pgtable_page_dtor(page);
		__free_page(page);
	}
}

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

	bits = (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_bh(&mm->context.list_lock);
	page->flags ^= bits;
	if (page->flags & FRAG_MASK) {
		/* Page now has some free pgtable fragments. */
		if (!list_empty(&page->lru))
			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_bh(&mm->context.list_lock);
	if (page) {
		pgtable_page_dtor(page);
		__free_page(page);
	}
}

void page_table_free_rcu(struct mm_struct *mm, unsigned long *table)
{
	struct rcu_table_freelist *batch;
	struct page *page;
	unsigned long bits;

	preempt_disable();
	if (atomic_read(&mm->mm_users) < 2 &&
	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
		page_table_free(mm, table);
		goto out;
	}
	batch = rcu_table_freelist_get(mm);
	if (!batch) {
		smp_call_function(smp_sync, NULL, 1);
		page_table_free(mm, table);
		goto out;
	}
	bits = (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_bh(&mm->context.list_lock);
	/* Delayed freeing with rcu prevents reuse of pgtable fragments */
	list_del_init(&page->lru);
	spin_unlock_bh(&mm->context.list_lock);
	table = (unsigned long *)(((unsigned long) table) | bits);
	batch->table[batch->pgt_index++] = table;
	if (batch->pgt_index >= batch->crst_index)
		rcu_table_freelist_finish();
out:
	preempt_enable();
}

/*
 * 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 (user_mode == HOME_SPACE_MODE)
		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 ||
#ifdef CONFIG_AIO
	    !hlist_empty(&tsk->mm->ioctx_list) ||
#endif
	    tsk->mm != tsk->active_mm) {
		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 ||
#ifdef CONFIG_AIO
	    !hlist_empty(&tsk->mm->ioctx_list) ||
#endif
	    tsk->mm != tsk->active_mm) {
		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);
	atomic_inc(&mm->context.attach_count);
	atomic_dec(&old_mm->context.attach_count);
	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);

#if defined(CONFIG_DEBUG_PAGEALLOC) && defined(CONFIG_HIBERNATION)
bool kernel_page_present(struct page *page)
{
	unsigned long addr;
	int cc;

	addr = page_to_phys(page);
	asm volatile(
		"	lra	%1,0(%1)\n"
		"	ipm	%0\n"
		"	srl	%0,28"
		: "=d" (cc), "+a" (addr) : : "cc");
	return cc == 0;
}
#endif /* CONFIG_HIBERNATION && CONFIG_DEBUG_PAGEALLOC */
Commit	Line	Data
3610cce8	1	/*
239a6425	2	* Copyright IBM Corp. 2007,2009
3610cce8 MS	3	* Author(s): Martin Schwidefsky <schwidefsky@de.ibm.com>
	4	*/
	5
	6	#include <linux/sched.h>
	7	#include <linux/kernel.h>
	8	#include <linux/errno.h>
5a0e3ad6	9	#include <linux/gfp.h>
3610cce8 MS	10	#include <linux/mm.h>
	11	#include <linux/swap.h>
	12	#include <linux/smp.h>
	13	#include <linux/highmem.h>
3610cce8 MS	14	#include <linux/pagemap.h>
	15	#include <linux/spinlock.h>
	16	#include <linux/module.h>
	17	#include <linux/quicklist.h>
80217147	18	#include <linux/rcupdate.h>
3610cce8 MS	19
	20	#include <asm/system.h>
	21	#include <asm/pgtable.h>
	22	#include <asm/pgalloc.h>
	23	#include <asm/tlb.h>
	24	#include <asm/tlbflush.h>
6252d702	25	#include <asm/mmu_context.h>
3610cce8	26
80217147 MS	27	struct rcu_table_freelist {
	28	struct rcu_head rcu;
	29	struct mm_struct *mm;
	30	unsigned int pgt_index;
	31	unsigned int crst_index;
	32	unsigned long *table[0];
	33	};
	34
	35	#define RCU_FREELIST_SIZE \
	36	((PAGE_SIZE - sizeof(struct rcu_table_freelist)) \
	37	/ sizeof(unsigned long))
	38
80217147 MS	39	static DEFINE_PER_CPU(struct rcu_table_freelist *, rcu_table_freelist);
	40
	41	static void __page_table_free(struct mm_struct mm, unsigned long table);
80217147 MS	42
	43	static struct rcu_table_freelist rcu_table_freelist_get(struct mm_struct mm)
	44	{
	45	struct rcu_table_freelist **batchp = &__get_cpu_var(rcu_table_freelist);
	46	struct rcu_table_freelist batch = batchp;
	47
	48	if (batch)
	49	return batch;
	50	batch = (struct rcu_table_freelist *) __get_free_page(GFP_ATOMIC);
	51	if (batch) {
	52	batch->mm = mm;
	53	batch->pgt_index = 0;
	54	batch->crst_index = RCU_FREELIST_SIZE;
	55	*batchp = batch;
	56	}
	57	return batch;
	58	}
	59
	60	static void rcu_table_freelist_callback(struct rcu_head *head)
	61	{
	62	struct rcu_table_freelist *batch =
	63	container_of(head, struct rcu_table_freelist, rcu);
	64
	65	while (batch->pgt_index > 0)
	66	__page_table_free(batch->mm, batch->table[--batch->pgt_index]);
	67	while (batch->crst_index < RCU_FREELIST_SIZE)
043d0708	68	crst_table_free(batch->mm, batch->table[batch->crst_index++]);
80217147 MS	69	free_page((unsigned long) batch);
	70	}
	71
	72	void rcu_table_freelist_finish(void)
	73	{
3c5cffb6 HC	74	struct rcu_table_freelist **batchp = &get_cpu_var(rcu_table_freelist);
3c5cffb6 HC	75	struct rcu_table_freelist batch = batchp;
80217147 MS	76
80217147 MS	77	if (!batch)
3c5cffb6	78	goto out;
80217147	79	call_rcu(&batch->rcu, rcu_table_freelist_callback);
3c5cffb6 HC	80	*batchp = NULL;
	81	out:
	82	put_cpu_var(rcu_table_freelist);
80217147 MS	83	}
	84
	85	static void smp_sync(void *arg)
	86	{
	87	}
	88
3610cce8 MS	89	#ifndef CONFIG_64BIT
3610cce8 MS	90	#define ALLOC_ORDER 1
146e4b3c MS	91	#define TABLES_PER_PAGE 4
	92	#define FRAG_MASK 15UL
	93	#define SECOND_HALVES 10UL
402b0862 CO	94
	95	void clear_table_pgstes(unsigned long *table)
	96	{
	97	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
	98	memset(table + 256, 0, PAGE_SIZE/4);
	99	clear_table(table + 512, _PAGE_TYPE_EMPTY, PAGE_SIZE/4);
	100	memset(table + 768, 0, PAGE_SIZE/4);
	101	}
	102
3610cce8 MS	103	#else
3610cce8 MS	104	#define ALLOC_ORDER 2
146e4b3c MS	105	#define TABLES_PER_PAGE 2
	106	#define FRAG_MASK 3UL
	107	#define SECOND_HALVES 2UL
402b0862 CO	108
	109	void clear_table_pgstes(unsigned long *table)
	110	{
	111	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE/2);
	112	memset(table + 256, 0, PAGE_SIZE/2);
	113	}
	114
3610cce8 MS	115	#endif
3610cce8 MS	116
239a6425 HC	117	unsigned long VMALLOC_START = VMALLOC_END - VMALLOC_SIZE;
	118	EXPORT_SYMBOL(VMALLOC_START);
	119
	120	static int __init parse_vmalloc(char *arg)
	121	{
	122	if (!arg)
	123	return -EINVAL;
	124	VMALLOC_START = (VMALLOC_END - memparse(arg, &arg)) & PAGE_MASK;
	125	return 0;
	126	}
	127	early_param("vmalloc", parse_vmalloc);
	128
043d0708	129	unsigned long crst_table_alloc(struct mm_struct mm)
3610cce8 MS	130	{
	131	struct page *page = alloc_pages(GFP_KERNEL, ALLOC_ORDER);
	132
	133	if (!page)
	134	return NULL;
3610cce8 MS	135	return (unsigned long *) page_to_phys(page);
	136	}
	137
80217147 MS	138	void crst_table_free(struct mm_struct mm, unsigned long table)
80217147 MS	139	{
043d0708	140	free_pages((unsigned long) table, ALLOC_ORDER);
80217147 MS	141	}
	142
	143	void crst_table_free_rcu(struct mm_struct mm, unsigned long table)
	144	{
	145	struct rcu_table_freelist *batch;
80217147	146
3c5cffb6	147	preempt_disable();
80217147 MS	148	if (atomic_read(&mm->mm_users) < 2 &&
80217147 MS	149	cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
043d0708	150	crst_table_free(mm, table);
3c5cffb6	151	goto out;
80217147 MS	152	}
	153	batch = rcu_table_freelist_get(mm);
	154	if (!batch) {
	155	smp_call_function(smp_sync, NULL, 1);
043d0708	156	crst_table_free(mm, table);
3c5cffb6	157	goto out;
80217147 MS	158	}
	159	batch->table[--batch->crst_index] = table;
	160	if (batch->pgt_index >= batch->crst_index)
	161	rcu_table_freelist_finish();
3c5cffb6 HC	162	out:
3c5cffb6 HC	163	preempt_enable();
80217147 MS	164	}
80217147 MS	165
6252d702 MS	166	#ifdef CONFIG_64BIT
	167	int crst_table_upgrade(struct mm_struct *mm, unsigned long limit)
	168	{
	169	unsigned long table, pgd;
	170	unsigned long entry;
	171
	172	BUG_ON(limit > (1UL << 53));
	173	repeat:
043d0708	174	table = crst_table_alloc(mm);
6252d702 MS	175	if (!table)
6252d702 MS	176	return -ENOMEM;
80217147	177	spin_lock_bh(&mm->page_table_lock);
6252d702 MS	178	if (mm->context.asce_limit < limit) {
	179	pgd = (unsigned long *) mm->pgd;
	180	if (mm->context.asce_limit <= (1UL << 31)) {
	181	entry = _REGION3_ENTRY_EMPTY;
	182	mm->context.asce_limit = 1UL << 42;
	183	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	184	_ASCE_USER_BITS \|
	185	_ASCE_TYPE_REGION3;
	186	} else {
	187	entry = _REGION2_ENTRY_EMPTY;
	188	mm->context.asce_limit = 1UL << 53;
	189	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	190	_ASCE_USER_BITS \|
	191	_ASCE_TYPE_REGION2;
	192	}
	193	crst_table_init(table, entry);
	194	pgd_populate(mm, (pgd_t ) table, (pud_t ) pgd);
	195	mm->pgd = (pgd_t *) table;
f481bfaf	196	mm->task_size = mm->context.asce_limit;
6252d702 MS	197	table = NULL;
6252d702 MS	198	}
80217147	199	spin_unlock_bh(&mm->page_table_lock);
6252d702 MS	200	if (table)
	201	crst_table_free(mm, table);
	202	if (mm->context.asce_limit < limit)
	203	goto repeat;
	204	update_mm(mm, current);
	205	return 0;
	206	}
	207
	208	void crst_table_downgrade(struct mm_struct *mm, unsigned long limit)
	209	{
	210	pgd_t *pgd;
	211
	212	if (mm->context.asce_limit <= limit)
	213	return;
	214	__tlb_flush_mm(mm);
	215	while (mm->context.asce_limit > limit) {
	216	pgd = mm->pgd;
	217	switch (pgd_val(*pgd) & _REGION_ENTRY_TYPE_MASK) {
	218	case _REGION_ENTRY_TYPE_R2:
	219	mm->context.asce_limit = 1UL << 42;
	220	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	221	_ASCE_USER_BITS \|
	222	_ASCE_TYPE_REGION3;
	223	break;
	224	case _REGION_ENTRY_TYPE_R3:
	225	mm->context.asce_limit = 1UL << 31;
	226	mm->context.asce_bits = _ASCE_TABLE_LENGTH \|
	227	_ASCE_USER_BITS \|
	228	_ASCE_TYPE_SEGMENT;
	229	break;
	230	default:
	231	BUG();
	232	}
	233	mm->pgd = (pgd_t ) (pgd_val(pgd) & _REGION_ENTRY_ORIGIN);
f481bfaf	234	mm->task_size = mm->context.asce_limit;
6252d702 MS	235	crst_table_free(mm, (unsigned long *) pgd);
	236	}
	237	update_mm(mm, current);
	238	}
	239	#endif
	240
3610cce8 MS	241	/*
	242	* page table entry allocation/free routines.
	243	*/
146e4b3c	244	unsigned long page_table_alloc(struct mm_struct mm)
3610cce8	245	{
146e4b3c	246	struct page *page;
3610cce8	247	unsigned long *table;
146e4b3c	248	unsigned long bits;
3610cce8	249
043d0708	250	bits = (mm->context.has_pgste) ? 3UL : 1UL;
80217147	251	spin_lock_bh(&mm->context.list_lock);
146e4b3c MS	252	page = NULL;
	253	if (!list_empty(&mm->context.pgtable_list)) {
	254	page = list_first_entry(&mm->context.pgtable_list,
	255	struct page, lru);
	256	if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
	257	page = NULL;
	258	}
	259	if (!page) {
80217147	260	spin_unlock_bh(&mm->context.list_lock);
146e4b3c MS	261	page = alloc_page(GFP_KERNEL\|__GFP_REPEAT);
146e4b3c MS	262	if (!page)
3610cce8	263	return NULL;
146e4b3c MS	264	pgtable_page_ctor(page);
	265	page->flags &= ~FRAG_MASK;
	266	table = (unsigned long *) page_to_phys(page);
250cf776	267	if (mm->context.has_pgste)
402b0862 CO	268	clear_table_pgstes(table);
	269	else
	270	clear_table(table, _PAGE_TYPE_EMPTY, PAGE_SIZE);
80217147	271	spin_lock_bh(&mm->context.list_lock);
146e4b3c	272	list_add(&page->lru, &mm->context.pgtable_list);
3610cce8 MS	273	}
3610cce8 MS	274	table = (unsigned long *) page_to_phys(page);
146e4b3c MS	275	while (page->flags & bits) {
	276	table += 256;
	277	bits <<= 1;
	278	}
	279	page->flags \|= bits;
	280	if ((page->flags & FRAG_MASK) == ((1UL << TABLES_PER_PAGE) - 1))
	281	list_move_tail(&page->lru, &mm->context.pgtable_list);
80217147	282	spin_unlock_bh(&mm->context.list_lock);
3610cce8 MS	283	return table;
	284	}
	285
80217147 MS	286	static void __page_table_free(struct mm_struct mm, unsigned long table)
	287	{
	288	struct page *page;
	289	unsigned long bits;
	290
	291	bits = ((unsigned long) table) & 15;
	292	table = (unsigned long *)(((unsigned long) table) ^ bits);
	293	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
	294	page->flags ^= bits;
	295	if (!(page->flags & FRAG_MASK)) {
	296	pgtable_page_dtor(page);
	297	__free_page(page);
	298	}
	299	}
	300
146e4b3c	301	void page_table_free(struct mm_struct mm, unsigned long table)
3610cce8	302	{
146e4b3c MS	303	struct page *page;
146e4b3c MS	304	unsigned long bits;
3610cce8	305
043d0708	306	bits = (mm->context.has_pgste) ? 3UL : 1UL;
146e4b3c MS	307	bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
146e4b3c MS	308	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
80217147	309	spin_lock_bh(&mm->context.list_lock);
146e4b3c MS	310	page->flags ^= bits;
	311	if (page->flags & FRAG_MASK) {
	312	/* Page now has some free pgtable fragments. */
f1be77bb MS	313	if (!list_empty(&page->lru))
f1be77bb MS	314	list_move(&page->lru, &mm->context.pgtable_list);
146e4b3c MS	315	page = NULL;
	316	} else
	317	/* All fragments of the 4K page have been freed. */
	318	list_del(&page->lru);
80217147	319	spin_unlock_bh(&mm->context.list_lock);
146e4b3c MS	320	if (page) {
	321	pgtable_page_dtor(page);
	322	__free_page(page);
	323	}
	324	}
3610cce8	325
80217147 MS	326	void page_table_free_rcu(struct mm_struct mm, unsigned long table)
	327	{
	328	struct rcu_table_freelist *batch;
	329	struct page *page;
	330	unsigned long bits;
	331
3c5cffb6	332	preempt_disable();
80217147 MS	333	if (atomic_read(&mm->mm_users) < 2 &&
	334	cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
	335	page_table_free(mm, table);
3c5cffb6	336	goto out;
80217147 MS	337	}
	338	batch = rcu_table_freelist_get(mm);
	339	if (!batch) {
	340	smp_call_function(smp_sync, NULL, 1);
	341	page_table_free(mm, table);
3c5cffb6	342	goto out;
80217147	343	}
043d0708	344	bits = (mm->context.has_pgste) ? 3UL : 1UL;
80217147 MS	345	bits <<= (__pa(table) & (PAGE_SIZE - 1)) / 256 / sizeof(unsigned long);
	346	page = pfn_to_page(__pa(table) >> PAGE_SHIFT);
	347	spin_lock_bh(&mm->context.list_lock);
	348	/* Delayed freeing with rcu prevents reuse of pgtable fragments */
	349	list_del_init(&page->lru);
	350	spin_unlock_bh(&mm->context.list_lock);
	351	table = (unsigned long *)(((unsigned long) table) \| bits);
	352	batch->table[batch->pgt_index++] = table;
	353	if (batch->pgt_index >= batch->crst_index)
	354	rcu_table_freelist_finish();
3c5cffb6 HC	355	out:
3c5cffb6 HC	356	preempt_enable();
80217147 MS	357	}
80217147 MS	358
402b0862 CO	359	/*
	360	* switch on pgstes for its userspace process (for kvm)
	361	*/
	362	int s390_enable_sie(void)
	363	{
	364	struct task_struct *tsk = current;
74b6b522	365	struct mm_struct mm, old_mm;
402b0862	366
702d9e58	367	/* Do we have switched amode? If no, we cannot do sie */
b11b5334	368	if (user_mode == HOME_SPACE_MODE)
702d9e58 CO	369	return -EINVAL;
702d9e58 CO	370
74b6b522	371	/* Do we have pgstes? if yes, we are done */
250cf776	372	if (tsk->mm->context.has_pgste)
74b6b522	373	return 0;
402b0862	374
74b6b522 CB	375	/* lets check if we are allowed to replace the mm */
74b6b522 CB	376	task_lock(tsk);
402b0862	377	if (!tsk->mm \|\| atomic_read(&tsk->mm->mm_users) > 1 \|\|
52a21f2c MS	378	#ifdef CONFIG_AIO
	379	!hlist_empty(&tsk->mm->ioctx_list) \|\|
	380	#endif
	381	tsk->mm != tsk->active_mm) {
74b6b522 CB	382	task_unlock(tsk);
	383	return -EINVAL;
	384	}
	385	task_unlock(tsk);
402b0862	386
250cf776 CB	387	/* we copy the mm and let dup_mm create the page tables with_pgstes */
250cf776 CB	388	tsk->mm->context.alloc_pgste = 1;
402b0862	389	mm = dup_mm(tsk);
250cf776	390	tsk->mm->context.alloc_pgste = 0;
402b0862	391	if (!mm)
74b6b522 CB	392	return -ENOMEM;
74b6b522 CB	393
250cf776	394	/* Now lets check again if something happened */
74b6b522 CB	395	task_lock(tsk);
74b6b522 CB	396	if (!tsk->mm \|\| atomic_read(&tsk->mm->mm_users) > 1 \|\|
52a21f2c MS	397	#ifdef CONFIG_AIO
	398	!hlist_empty(&tsk->mm->ioctx_list) \|\|
	399	#endif
	400	tsk->mm != tsk->active_mm) {
74b6b522 CB	401	mmput(mm);
	402	task_unlock(tsk);
	403	return -EINVAL;
	404	}
	405
	406	/* ok, we are alone. No ptrace, no threads, etc. */
	407	old_mm = tsk->mm;
402b0862 CO	408	tsk->mm = tsk->active_mm = mm;
	409	preempt_disable();
	410	update_mm(mm, tsk);
e05ef9bd CB	411	atomic_inc(&mm->context.attach_count);
e05ef9bd CB	412	atomic_dec(&old_mm->context.attach_count);
005f8eee	413	cpumask_set_cpu(smp_processor_id(), mm_cpumask(mm));
402b0862	414	preempt_enable();
402b0862	415	task_unlock(tsk);
74b6b522 CB	416	mmput(old_mm);
74b6b522 CB	417	return 0;
402b0862 CO	418	}
402b0862 CO	419	EXPORT_SYMBOL_GPL(s390_enable_sie);
7db11a36	420
87458ff4	421	#if defined(CONFIG_DEBUG_PAGEALLOC) && defined(CONFIG_HIBERNATION)
7db11a36 HJP	422	bool kernel_page_present(struct page *page)
	423	{
	424	unsigned long addr;
	425	int cc;
	426
	427	addr = page_to_phys(page);
87458ff4 HC	428	asm volatile(
	429	" lra %1,0(%1)\n"
	430	" ipm %0\n"
	431	" srl %0,28"
	432	: "=d" (cc), "+a" (addr) : : "cc");
7db11a36 HJP	433	return cc == 0;
7db11a36 HJP	434	}
87458ff4	435	#endif /* CONFIG_HIBERNATION && CONFIG_DEBUG_PAGEALLOC */