[linux-block.git] / arch / i386 / mm / pgtable.c

/*
 *  linux/arch/i386/mm/pgtable.c
 */

#include <linux/config.h>
#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 <asm/system.h>
#include <asm/pgtable.h>
#include <asm/pgalloc.h>
#include <asm/fixmap.h>
#include <asm/e820.h>
#include <asm/tlb.h>
#include <asm/tlbflush.h>

void show_mem(void)
{
	int total = 0, reserved = 0;
	int shared = 0, cached = 0;
	int highmem = 0;
	struct page *page;
	pg_data_t *pgdat;
	unsigned long i;
	struct page_state ps;
	unsigned long flags;

	printk(KERN_INFO "Mem-info:\n");
	show_free_areas();
	printk(KERN_INFO "Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
	for_each_online_pgdat(pgdat) {
		pgdat_resize_lock(pgdat, &flags);
		for (i = 0; i < pgdat->node_spanned_pages; ++i) {
			page = pgdat_page_nr(pgdat, i);
			total++;
			if (PageHighMem(page))
				highmem++;
			if (PageReserved(page))
				reserved++;
			else if (PageSwapCache(page))
				cached++;
			else if (page_count(page))
				shared += page_count(page) - 1;
		}
		pgdat_resize_unlock(pgdat, &flags);
	}
	printk(KERN_INFO "%d pages of RAM\n", total);
	printk(KERN_INFO "%d pages of HIGHMEM\n", highmem);
	printk(KERN_INFO "%d reserved pages\n", reserved);
	printk(KERN_INFO "%d pages shared\n", shared);
	printk(KERN_INFO "%d pages swap cached\n", cached);

	get_page_state(&ps);
	printk(KERN_INFO "%lu pages dirty\n", ps.nr_dirty);
	printk(KERN_INFO "%lu pages writeback\n", ps.nr_writeback);
	printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED));
	printk(KERN_INFO "%lu pages slab\n", ps.nr_slab);
	printk(KERN_INFO "%lu pages pagetables\n", ps.nr_page_table_pages);
}

/*
 * Associate a virtual page frame with a given physical page frame 
 * and protection flags for that frame.
 */ 
static void set_pte_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *pte;

	pgd = swapper_pg_dir + pgd_index(vaddr);
	if (pgd_none(*pgd)) {
		BUG();
		return;
	}
	pud = pud_offset(pgd, vaddr);
	if (pud_none(*pud)) {
		BUG();
		return;
	}
	pmd = pmd_offset(pud, vaddr);
	if (pmd_none(*pmd)) {
		BUG();
		return;
	}
	pte = pte_offset_kernel(pmd, vaddr);
	/* <pfn,flags> stored as-is, to permit clearing entries */
	set_pte(pte, pfn_pte(pfn, flags));

	/*
	 * It's enough to flush this one mapping.
	 * (PGE mappings get flushed as well)
	 */
	__flush_tlb_one(vaddr);
}

/*
 * Associate a large virtual page frame with a given physical page frame 
 * and protection flags for that frame. pfn is for the base of the page,
 * vaddr is what the page gets mapped to - both must be properly aligned. 
 * The pmd must already be instantiated. Assumes PAE mode.
 */ 
void set_pmd_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
{
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;

	if (vaddr & (PMD_SIZE-1)) {		/* vaddr is misaligned */
		printk(KERN_WARNING "set_pmd_pfn: vaddr misaligned\n");
		return; /* BUG(); */
	}
	if (pfn & (PTRS_PER_PTE-1)) {		/* pfn is misaligned */
		printk(KERN_WARNING "set_pmd_pfn: pfn misaligned\n");
		return; /* BUG(); */
	}
	pgd = swapper_pg_dir + pgd_index(vaddr);
	if (pgd_none(*pgd)) {
		printk(KERN_WARNING "set_pmd_pfn: pgd_none\n");
		return; /* BUG(); */
	}
	pud = pud_offset(pgd, vaddr);
	pmd = pmd_offset(pud, vaddr);
	set_pmd(pmd, pfn_pmd(pfn, flags));
	/*
	 * It's enough to flush this one mapping.
	 * (PGE mappings get flushed as well)
	 */
	__flush_tlb_one(vaddr);
}

void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
{
	unsigned long address = __fix_to_virt(idx);

	if (idx >= __end_of_fixed_addresses) {
		BUG();
		return;
	}
	set_pte_pfn(address, phys >> PAGE_SHIFT, flags);
}

pte_t *pte_alloc_one_kernel(struct mm_struct *mm, unsigned long address)
{
	return (pte_t *)__get_free_page(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO);
}

struct page *pte_alloc_one(struct mm_struct *mm, unsigned long address)
{
	struct page *pte;

#ifdef CONFIG_HIGHPTE
	pte = alloc_pages(GFP_KERNEL|__GFP_HIGHMEM|__GFP_REPEAT|__GFP_ZERO, 0);
#else
	pte = alloc_pages(GFP_KERNEL|__GFP_REPEAT|__GFP_ZERO, 0);
#endif
	return pte;
}

void pmd_ctor(void *pmd, kmem_cache_t *cache, unsigned long flags)
{
	memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
}

/*
 * List of all pgd's needed for non-PAE so it can invalidate entries
 * in both cached and uncached pgd's; not needed for PAE since the
 * kernel pmd is shared. If PAE were not to share the pmd a similar
 * tactic would be needed. This is essentially codepath-based locking
 * against pageattr.c; it is the unique case in which a valid change
 * of kernel pagetables can't be lazily synchronized by vmalloc faults.
 * vmalloc faults work because attached pagetables are never freed.
 * The locking scheme was chosen on the basis of manfred's
 * recommendations and having no core impact whatsoever.
 * -- wli
 */
DEFINE_SPINLOCK(pgd_lock);
struct page *pgd_list;

static inline void pgd_list_add(pgd_t *pgd)
{
	struct page *page = virt_to_page(pgd);
	page->index = (unsigned long)pgd_list;
	if (pgd_list)
		set_page_private(pgd_list, (unsigned long)&page->index);
	pgd_list = page;
	set_page_private(page, (unsigned long)&pgd_list);
}

static inline void pgd_list_del(pgd_t *pgd)
{
	struct page *next, **pprev, *page = virt_to_page(pgd);
	next = (struct page *)page->index;
	pprev = (struct page **)page_private(page);
	*pprev = next;
	if (next)
		set_page_private(next, (unsigned long)pprev);
}

void pgd_ctor(void *pgd, kmem_cache_t *cache, unsigned long unused)
{
	unsigned long flags;

	if (PTRS_PER_PMD == 1) {
		memset(pgd, 0, USER_PTRS_PER_PGD*sizeof(pgd_t));
		spin_lock_irqsave(&pgd_lock, flags);
	}

	clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
			swapper_pg_dir + USER_PTRS_PER_PGD,
			KERNEL_PGD_PTRS);
	if (PTRS_PER_PMD > 1)
		return;

	pgd_list_add(pgd);
	spin_unlock_irqrestore(&pgd_lock, flags);
}

/* never called when PTRS_PER_PMD > 1 */
void pgd_dtor(void *pgd, kmem_cache_t *cache, unsigned long unused)
{
	unsigned long flags; /* can be called from interrupt context */

	spin_lock_irqsave(&pgd_lock, flags);
	pgd_list_del(pgd);
	spin_unlock_irqrestore(&pgd_lock, flags);
}

pgd_t *pgd_alloc(struct mm_struct *mm)
{
	int i;
	pgd_t *pgd = kmem_cache_alloc(pgd_cache, GFP_KERNEL);

	if (PTRS_PER_PMD == 1 || !pgd)
		return pgd;

	for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
		pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
		if (!pmd)
			goto out_oom;
		set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
	}
	return pgd;

out_oom:
	for (i--; i >= 0; i--)
		kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
	kmem_cache_free(pgd_cache, pgd);
	return NULL;
}

void pgd_free(pgd_t *pgd)
{
	int i;

	/* in the PAE case user pgd entries are overwritten before usage */
	if (PTRS_PER_PMD > 1)
		for (i = 0; i < USER_PTRS_PER_PGD; ++i)
			kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
	/* in the non-PAE case, free_pgtables() clears user pgd entries */
	kmem_cache_free(pgd_cache, pgd);
}
Commit	Line	Data
1da177e4 LT	1	/*
	2	* linux/arch/i386/mm/pgtable.c
	3	*/
	4
	5	#include <linux/config.h>
	6	#include <linux/sched.h>
	7	#include <linux/kernel.h>
	8	#include <linux/errno.h>
	9	#include <linux/mm.h>
	10	#include <linux/swap.h>
	11	#include <linux/smp.h>
	12	#include <linux/highmem.h>
	13	#include <linux/slab.h>
	14	#include <linux/pagemap.h>
	15	#include <linux/spinlock.h>
	16
	17	#include <asm/system.h>
	18	#include <asm/pgtable.h>
	19	#include <asm/pgalloc.h>
	20	#include <asm/fixmap.h>
	21	#include <asm/e820.h>
	22	#include <asm/tlb.h>
	23	#include <asm/tlbflush.h>
	24
	25	void show_mem(void)
	26	{
	27	int total = 0, reserved = 0;
	28	int shared = 0, cached = 0;
	29	int highmem = 0;
	30	struct page *page;
	31	pg_data_t *pgdat;
	32	unsigned long i;
6f4e1e50	33	struct page_state ps;
208d54e5	34	unsigned long flags;
1da177e4	35
f90e7185	36	printk(KERN_INFO "Mem-info:\n");
1da177e4	37	show_free_areas();
f90e7185	38	printk(KERN_INFO "Free swap: %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
ec936fc5	39	for_each_online_pgdat(pgdat) {
208d54e5	40	pgdat_resize_lock(pgdat, &flags);
1da177e4	41	for (i = 0; i < pgdat->node_spanned_pages; ++i) {
408fde81	42	page = pgdat_page_nr(pgdat, i);
1da177e4 LT	43	total++;
	44	if (PageHighMem(page))
	45	highmem++;
	46	if (PageReserved(page))
	47	reserved++;
	48	else if (PageSwapCache(page))
	49	cached++;
	50	else if (page_count(page))
	51	shared += page_count(page) - 1;
	52	}
208d54e5	53	pgdat_resize_unlock(pgdat, &flags);
1da177e4	54	}
f90e7185 CL	55	printk(KERN_INFO "%d pages of RAM\n", total);
	56	printk(KERN_INFO "%d pages of HIGHMEM\n", highmem);
	57	printk(KERN_INFO "%d reserved pages\n", reserved);
	58	printk(KERN_INFO "%d pages shared\n", shared);
	59	printk(KERN_INFO "%d pages swap cached\n", cached);
6f4e1e50 MB	60
6f4e1e50 MB	61	get_page_state(&ps);
f90e7185 CL	62	printk(KERN_INFO "%lu pages dirty\n", ps.nr_dirty);
f90e7185 CL	63	printk(KERN_INFO "%lu pages writeback\n", ps.nr_writeback);
65ba55f5	64	printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED));
f90e7185 CL	65	printk(KERN_INFO "%lu pages slab\n", ps.nr_slab);
f90e7185 CL	66	printk(KERN_INFO "%lu pages pagetables\n", ps.nr_page_table_pages);
1da177e4 LT	67	}
	68
	69	/*
	70	* Associate a virtual page frame with a given physical page frame
	71	* and protection flags for that frame.
	72	*/
	73	static void set_pte_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
	74	{
	75	pgd_t *pgd;
	76	pud_t *pud;
	77	pmd_t *pmd;
	78	pte_t *pte;
	79
	80	pgd = swapper_pg_dir + pgd_index(vaddr);
	81	if (pgd_none(*pgd)) {
	82	BUG();
	83	return;
	84	}
	85	pud = pud_offset(pgd, vaddr);
	86	if (pud_none(*pud)) {
	87	BUG();
	88	return;
	89	}
	90	pmd = pmd_offset(pud, vaddr);
	91	if (pmd_none(*pmd)) {
	92	BUG();
	93	return;
	94	}
	95	pte = pte_offset_kernel(pmd, vaddr);
	96	/* <pfn,flags> stored as-is, to permit clearing entries */
	97	set_pte(pte, pfn_pte(pfn, flags));
	98
	99	/*
	100	* It's enough to flush this one mapping.
	101	* (PGE mappings get flushed as well)
	102	*/
	103	__flush_tlb_one(vaddr);
	104	}
	105
	106	/*
	107	* Associate a large virtual page frame with a given physical page frame
	108	* and protection flags for that frame. pfn is for the base of the page,
	109	* vaddr is what the page gets mapped to - both must be properly aligned.
	110	* The pmd must already be instantiated. Assumes PAE mode.
	111	*/
	112	void set_pmd_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
	113	{
	114	pgd_t *pgd;
	115	pud_t *pud;
	116	pmd_t *pmd;
	117
	118	if (vaddr & (PMD_SIZE-1)) { /* vaddr is misaligned */
f90e7185	119	printk(KERN_WARNING "set_pmd_pfn: vaddr misaligned\n");
1da177e4 LT	120	return; /* BUG(); */
	121	}
	122	if (pfn & (PTRS_PER_PTE-1)) { /* pfn is misaligned */
f90e7185	123	printk(KERN_WARNING "set_pmd_pfn: pfn misaligned\n");
1da177e4 LT	124	return; /* BUG(); */
	125	}
	126	pgd = swapper_pg_dir + pgd_index(vaddr);
	127	if (pgd_none(*pgd)) {
f90e7185	128	printk(KERN_WARNING "set_pmd_pfn: pgd_none\n");
1da177e4 LT	129	return; /* BUG(); */
	130	}
	131	pud = pud_offset(pgd, vaddr);
	132	pmd = pmd_offset(pud, vaddr);
	133	set_pmd(pmd, pfn_pmd(pfn, flags));
	134	/*
	135	* It's enough to flush this one mapping.
	136	* (PGE mappings get flushed as well)
	137	*/
	138	__flush_tlb_one(vaddr);
	139	}
	140
	141	void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
	142	{
	143	unsigned long address = __fix_to_virt(idx);
	144
	145	if (idx >= __end_of_fixed_addresses) {
	146	BUG();
	147	return;
	148	}
	149	set_pte_pfn(address, phys >> PAGE_SHIFT, flags);
	150	}
	151
	152	pte_t pte_alloc_one_kernel(struct mm_struct mm, unsigned long address)
	153	{
	154	return (pte_t *)__get_free_page(GFP_KERNEL\|__GFP_REPEAT\|__GFP_ZERO);
	155	}
	156
	157	struct page pte_alloc_one(struct mm_struct mm, unsigned long address)
	158	{
	159	struct page *pte;
	160
	161	#ifdef CONFIG_HIGHPTE
	162	pte = alloc_pages(GFP_KERNEL\|__GFP_HIGHMEM\|__GFP_REPEAT\|__GFP_ZERO, 0);
	163	#else
	164	pte = alloc_pages(GFP_KERNEL\|__GFP_REPEAT\|__GFP_ZERO, 0);
	165	#endif
	166	return pte;
	167	}
	168
	169	void pmd_ctor(void pmd, kmem_cache_t cache, unsigned long flags)
	170	{
	171	memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
	172	}
	173
	174	/*
	175	* List of all pgd's needed for non-PAE so it can invalidate entries
	176	* in both cached and uncached pgd's; not needed for PAE since the
	177	* kernel pmd is shared. If PAE were not to share the pmd a similar
	178	* tactic would be needed. This is essentially codepath-based locking
	179	* against pageattr.c; it is the unique case in which a valid change
	180	* of kernel pagetables can't be lazily synchronized by vmalloc faults.
	181	* vmalloc faults work because attached pagetables are never freed.
	182	* The locking scheme was chosen on the basis of manfred's
	183	* recommendations and having no core impact whatsoever.
	184	* -- wli
	185	*/
	186	DEFINE_SPINLOCK(pgd_lock);
	187	struct page *pgd_list;
	188
	189	static inline void pgd_list_add(pgd_t *pgd)
	190	{
	191	struct page *page = virt_to_page(pgd);
	192	page->index = (unsigned long)pgd_list;
193	if (pgd_list)
4c21e2f2	194	set_page_private(pgd_list, (unsigned long)&page->index);
1da177e4	195	pgd_list = page;
4c21e2f2	196	set_page_private(page, (unsigned long)&pgd_list);
1da177e4 LT	197	}
	198
	199	static inline void pgd_list_del(pgd_t *pgd)
	200	{
	201	struct page next, pprev, page = virt_to_page(pgd);
	202	next = (struct page *)page->index;
4c21e2f2	203	pprev = (struct page **)page_private(page);
1da177e4 LT	204	*pprev = next;
1da177e4 LT	205	if (next)
4c21e2f2	206	set_page_private(next, (unsigned long)pprev);
1da177e4 LT	207	}
	208
	209	void pgd_ctor(void pgd, kmem_cache_t cache, unsigned long unused)
	210	{
	211	unsigned long flags;
	212
d7271b14 ZA	213	if (PTRS_PER_PMD == 1) {
d7271b14 ZA	214	memset(pgd, 0, USER_PTRS_PER_PGD*sizeof(pgd_t));
1da177e4	215	spin_lock_irqsave(&pgd_lock, flags);
d7271b14	216	}
1da177e4	217
d7271b14	218	clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
1da177e4	219	swapper_pg_dir + USER_PTRS_PER_PGD,
d7271b14	220	KERNEL_PGD_PTRS);
1da177e4 LT	221	if (PTRS_PER_PMD > 1)
	222	return;
	223
	224	pgd_list_add(pgd);
	225	spin_unlock_irqrestore(&pgd_lock, flags);
1da177e4 LT	226	}
	227
	228	/* never called when PTRS_PER_PMD > 1 */
	229	void pgd_dtor(void pgd, kmem_cache_t cache, unsigned long unused)
	230	{
	231	unsigned long flags; /* can be called from interrupt context */
	232
	233	spin_lock_irqsave(&pgd_lock, flags);
	234	pgd_list_del(pgd);
	235	spin_unlock_irqrestore(&pgd_lock, flags);
	236	}
	237
	238	pgd_t pgd_alloc(struct mm_struct mm)
	239	{
	240	int i;
	241	pgd_t *pgd = kmem_cache_alloc(pgd_cache, GFP_KERNEL);
	242
	243	if (PTRS_PER_PMD == 1 \|\| !pgd)
	244	return pgd;
	245
	246	for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
	247	pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
	248	if (!pmd)
	249	goto out_oom;
	250	set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
	251	}
	252	return pgd;
	253
	254	out_oom:
	255	for (i--; i >= 0; i--)
	256	kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
	257	kmem_cache_free(pgd_cache, pgd);
	258	return NULL;
	259	}
	260
	261	void pgd_free(pgd_t *pgd)
	262	{
	263	int i;
	264
	265	/* in the PAE case user pgd entries are overwritten before usage */
	266	if (PTRS_PER_PMD > 1)
	267	for (i = 0; i < USER_PTRS_PER_PGD; ++i)
	268	kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
e0da382c	269	/* in the non-PAE case, free_pgtables() clears user pgd entries */
1da177e4 LT	270	kmem_cache_free(pgd_cache, pgd);
1da177e4 LT	271	}