[GitHub/mt8127/android_kernel_alcatel_ttab.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 *batch = __get_cpu_var(rcu_table_freelist);

	if (!batch)
		return;
	call_rcu(&batch->rcu, rcu_table_freelist_callback);
	__get_cpu_var(rcu_table_freelist) = NULL;
}

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;

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

#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;

	if (atomic_read(&mm->mm_users) < 2 &&
	    cpumask_equal(mm_cpumask(mm), cpumask_of(smp_processor_id()))) {
		page_table_free(mm, table);
		return;
	}
	batch = rcu_table_freelist_get(mm);
	if (!batch) {
		smp_call_function(smp_sync, NULL, 1);
		page_table_free(mm, table);
		return;
	}
	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();
}

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