[net-next-2.6.git] / arch / i386 / mm / pgtable.c

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

#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 <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;
	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);

	printk(KERN_INFO "%lu pages dirty\n", global_page_state(NR_FILE_DIRTY));
	printk(KERN_INFO "%lu pages writeback\n",
					global_page_state(NR_WRITEBACK));
	printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED));
	printk(KERN_INFO "%lu pages slab\n",
		global_page_state(NR_SLAB_RECLAIMABLE) +
		global_page_state(NR_SLAB_UNRECLAIMABLE));
	printk(KERN_INFO "%lu pages pagetables\n",
					global_page_state(NR_PAGETABLE));
}

/*
 * 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);
}

static int fixmaps;
#ifndef CONFIG_COMPAT_VDSO
unsigned long __FIXADDR_TOP = 0xfffff000;
EXPORT_SYMBOL(__FIXADDR_TOP);
#endif

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);
	fixmaps++;
}

/**
 * reserve_top_address - reserves a hole in the top of kernel address space
 * @reserve - size of hole to reserve
 *
 * Can be used to relocate the fixmap area and poke a hole in the top
 * of kernel address space to make room for a hypervisor.
 */
void reserve_top_address(unsigned long reserve)
{
	BUG_ON(fixmaps > 0);
#ifdef CONFIG_COMPAT_VDSO
	BUG_ON(reserve != 0);
#else
	__FIXADDR_TOP = -reserve - PAGE_SIZE;
	__VMALLOC_RESERVE += reserve;
#endif
}

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
1da177e4 LT	5	#include <linux/sched.h>
	6	#include <linux/kernel.h>
	7	#include <linux/errno.h>
	8	#include <linux/mm.h>
	9	#include <linux/swap.h>
	10	#include <linux/smp.h>
	11	#include <linux/highmem.h>
	12	#include <linux/slab.h>
	13	#include <linux/pagemap.h>
	14	#include <linux/spinlock.h>
052e7994	15	#include <linux/module.h>
1da177e4 LT	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;
208d54e5	33	unsigned long flags;
1da177e4	34
f90e7185	35	printk(KERN_INFO "Mem-info:\n");
1da177e4	36	show_free_areas();
f90e7185	37	printk(KERN_INFO "Free swap: %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
ec936fc5	38	for_each_online_pgdat(pgdat) {
208d54e5	39	pgdat_resize_lock(pgdat, &flags);
1da177e4	40	for (i = 0; i < pgdat->node_spanned_pages; ++i) {
408fde81	41	page = pgdat_page_nr(pgdat, i);
1da177e4 LT	42	total++;
	43	if (PageHighMem(page))
	44	highmem++;
	45	if (PageReserved(page))
	46	reserved++;
	47	else if (PageSwapCache(page))
	48	cached++;
	49	else if (page_count(page))
	50	shared += page_count(page) - 1;
	51	}
208d54e5	52	pgdat_resize_unlock(pgdat, &flags);
1da177e4	53	}
f90e7185 CL	54	printk(KERN_INFO "%d pages of RAM\n", total);
	55	printk(KERN_INFO "%d pages of HIGHMEM\n", highmem);
	56	printk(KERN_INFO "%d reserved pages\n", reserved);
	57	printk(KERN_INFO "%d pages shared\n", shared);
	58	printk(KERN_INFO "%d pages swap cached\n", cached);
6f4e1e50	59
b1e7a8fd	60	printk(KERN_INFO "%lu pages dirty\n", global_page_state(NR_FILE_DIRTY));
ce866b34 CL	61	printk(KERN_INFO "%lu pages writeback\n",
ce866b34 CL	62	global_page_state(NR_WRITEBACK));
65ba55f5	63	printk(KERN_INFO "%lu pages mapped\n", global_page_state(NR_FILE_MAPPED));
972d1a7b CL	64	printk(KERN_INFO "%lu pages slab\n",
	65	global_page_state(NR_SLAB_RECLAIMABLE) +
	66	global_page_state(NR_SLAB_UNRECLAIMABLE));
df849a15 CL	67	printk(KERN_INFO "%lu pages pagetables\n",
df849a15 CL	68	global_page_state(NR_PAGETABLE));
1da177e4 LT	69	}
	70
	71	/*
	72	* Associate a virtual page frame with a given physical page frame
	73	* and protection flags for that frame.
	74	*/
	75	static void set_pte_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
	76	{
	77	pgd_t *pgd;
	78	pud_t *pud;
	79	pmd_t *pmd;
	80	pte_t *pte;
	81
	82	pgd = swapper_pg_dir + pgd_index(vaddr);
	83	if (pgd_none(*pgd)) {
	84	BUG();
	85	return;
	86	}
	87	pud = pud_offset(pgd, vaddr);
	88	if (pud_none(*pud)) {
	89	BUG();
	90	return;
	91	}
	92	pmd = pmd_offset(pud, vaddr);
	93	if (pmd_none(*pmd)) {
	94	BUG();
	95	return;
	96	}
	97	pte = pte_offset_kernel(pmd, vaddr);
	98	/* <pfn,flags> stored as-is, to permit clearing entries */
	99	set_pte(pte, pfn_pte(pfn, flags));
	100
	101	/*
	102	* It's enough to flush this one mapping.
	103	* (PGE mappings get flushed as well)
	104	*/
	105	__flush_tlb_one(vaddr);
	106	}
	107
	108	/*
	109	* Associate a large virtual page frame with a given physical page frame
	110	* and protection flags for that frame. pfn is for the base of the page,
	111	* vaddr is what the page gets mapped to - both must be properly aligned.
	112	* The pmd must already be instantiated. Assumes PAE mode.
	113	*/
	114	void set_pmd_pfn(unsigned long vaddr, unsigned long pfn, pgprot_t flags)
	115	{
	116	pgd_t *pgd;
	117	pud_t *pud;
	118	pmd_t *pmd;
	119
	120	if (vaddr & (PMD_SIZE-1)) { /* vaddr is misaligned */
f90e7185	121	printk(KERN_WARNING "set_pmd_pfn: vaddr misaligned\n");
1da177e4 LT	122	return; /* BUG(); */
	123	}
	124	if (pfn & (PTRS_PER_PTE-1)) { /* pfn is misaligned */
f90e7185	125	printk(KERN_WARNING "set_pmd_pfn: pfn misaligned\n");
1da177e4 LT	126	return; /* BUG(); */
	127	}
	128	pgd = swapper_pg_dir + pgd_index(vaddr);
	129	if (pgd_none(*pgd)) {
f90e7185	130	printk(KERN_WARNING "set_pmd_pfn: pgd_none\n");
1da177e4 LT	131	return; /* BUG(); */
	132	}
	133	pud = pud_offset(pgd, vaddr);
	134	pmd = pmd_offset(pud, vaddr);
	135	set_pmd(pmd, pfn_pmd(pfn, flags));
	136	/*
	137	* It's enough to flush this one mapping.
	138	* (PGE mappings get flushed as well)
	139	*/
	140	__flush_tlb_one(vaddr);
	141	}
	142
052e7994 JF	143	static int fixmaps;
	144	#ifndef CONFIG_COMPAT_VDSO
	145	unsigned long __FIXADDR_TOP = 0xfffff000;
	146	EXPORT_SYMBOL(__FIXADDR_TOP);
	147	#endif
	148
1da177e4 LT	149	void __set_fixmap (enum fixed_addresses idx, unsigned long phys, pgprot_t flags)
	150	{
	151	unsigned long address = __fix_to_virt(idx);
	152
	153	if (idx >= __end_of_fixed_addresses) {
	154	BUG();
	155	return;
	156	}
	157	set_pte_pfn(address, phys >> PAGE_SHIFT, flags);
052e7994 JF	158	fixmaps++;
	159	}
	160
	161	/**
	162	* reserve_top_address - reserves a hole in the top of kernel address space
	163	* @reserve - size of hole to reserve
	164	*
	165	* Can be used to relocate the fixmap area and poke a hole in the top
	166	* of kernel address space to make room for a hypervisor.
	167	*/
	168	void reserve_top_address(unsigned long reserve)
	169	{
	170	BUG_ON(fixmaps > 0);
	171	#ifdef CONFIG_COMPAT_VDSO
	172	BUG_ON(reserve != 0);
	173	#else
	174	__FIXADDR_TOP = -reserve - PAGE_SIZE;
	175	__VMALLOC_RESERVE += reserve;
	176	#endif
1da177e4 LT	177	}
	178
	179	pte_t pte_alloc_one_kernel(struct mm_struct mm, unsigned long address)
	180	{
	181	return (pte_t *)__get_free_page(GFP_KERNEL\|__GFP_REPEAT\|__GFP_ZERO);
	182	}
	183
	184	struct page pte_alloc_one(struct mm_struct mm, unsigned long address)
	185	{
	186	struct page *pte;
	187
	188	#ifdef CONFIG_HIGHPTE
	189	pte = alloc_pages(GFP_KERNEL\|__GFP_HIGHMEM\|__GFP_REPEAT\|__GFP_ZERO, 0);
	190	#else
	191	pte = alloc_pages(GFP_KERNEL\|__GFP_REPEAT\|__GFP_ZERO, 0);
	192	#endif
	193	return pte;
	194	}
	195
	196	void pmd_ctor(void pmd, kmem_cache_t cache, unsigned long flags)
	197	{
	198	memset(pmd, 0, PTRS_PER_PMD*sizeof(pmd_t));
	199	}
	200
	201	/*
	202	* List of all pgd's needed for non-PAE so it can invalidate entries
	203	* in both cached and uncached pgd's; not needed for PAE since the
	204	* kernel pmd is shared. If PAE were not to share the pmd a similar
	205	* tactic would be needed. This is essentially codepath-based locking
	206	* against pageattr.c; it is the unique case in which a valid change
	207	* of kernel pagetables can't be lazily synchronized by vmalloc faults.
	208	* vmalloc faults work because attached pagetables are never freed.
	209	* The locking scheme was chosen on the basis of manfred's
	210	* recommendations and having no core impact whatsoever.
	211	* -- wli
	212	*/
	213	DEFINE_SPINLOCK(pgd_lock);
	214	struct page *pgd_list;
	215
	216	static inline void pgd_list_add(pgd_t *pgd)
	217	{
	218	struct page *page = virt_to_page(pgd);
	219	page->index = (unsigned long)pgd_list;
	220	if (pgd_list)
4c21e2f2	221	set_page_private(pgd_list, (unsigned long)&page->index);
1da177e4	222	pgd_list = page;
4c21e2f2	223	set_page_private(page, (unsigned long)&pgd_list);
1da177e4 LT	224	}
	225
	226	static inline void pgd_list_del(pgd_t *pgd)
	227	{
	228	struct page next, pprev, page = virt_to_page(pgd);
	229	next = (struct page *)page->index;
4c21e2f2	230	pprev = (struct page **)page_private(page);
1da177e4 LT	231	*pprev = next;
1da177e4 LT	232	if (next)
4c21e2f2	233	set_page_private(next, (unsigned long)pprev);
1da177e4 LT	234	}
	235
	236	void pgd_ctor(void pgd, kmem_cache_t cache, unsigned long unused)
	237	{
	238	unsigned long flags;
	239
d7271b14 ZA	240	if (PTRS_PER_PMD == 1) {
d7271b14 ZA	241	memset(pgd, 0, USER_PTRS_PER_PGD*sizeof(pgd_t));
1da177e4	242	spin_lock_irqsave(&pgd_lock, flags);
d7271b14	243	}
1da177e4	244
d7271b14	245	clone_pgd_range((pgd_t *)pgd + USER_PTRS_PER_PGD,
1da177e4	246	swapper_pg_dir + USER_PTRS_PER_PGD,
d7271b14	247	KERNEL_PGD_PTRS);
1da177e4 LT	248	if (PTRS_PER_PMD > 1)
	249	return;
	250
	251	pgd_list_add(pgd);
	252	spin_unlock_irqrestore(&pgd_lock, flags);
1da177e4 LT	253	}
	254
	255	/* never called when PTRS_PER_PMD > 1 */
	256	void pgd_dtor(void pgd, kmem_cache_t cache, unsigned long unused)
	257	{
	258	unsigned long flags; /* can be called from interrupt context */
	259
	260	spin_lock_irqsave(&pgd_lock, flags);
	261	pgd_list_del(pgd);
	262	spin_unlock_irqrestore(&pgd_lock, flags);
	263	}
	264
	265	pgd_t pgd_alloc(struct mm_struct mm)
	266	{
	267	int i;
	268	pgd_t *pgd = kmem_cache_alloc(pgd_cache, GFP_KERNEL);
	269
	270	if (PTRS_PER_PMD == 1 \|\| !pgd)
	271	return pgd;
	272
	273	for (i = 0; i < USER_PTRS_PER_PGD; ++i) {
	274	pmd_t *pmd = kmem_cache_alloc(pmd_cache, GFP_KERNEL);
	275	if (!pmd)
	276	goto out_oom;
	277	set_pgd(&pgd[i], __pgd(1 + __pa(pmd)));
	278	}
	279	return pgd;
	280
	281	out_oom:
	282	for (i--; i >= 0; i--)
	283	kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
	284	kmem_cache_free(pgd_cache, pgd);
	285	return NULL;
	286	}
	287
	288	void pgd_free(pgd_t *pgd)
	289	{
	290	int i;
	291
	292	/* in the PAE case user pgd entries are overwritten before usage */
	293	if (PTRS_PER_PMD > 1)
	294	for (i = 0; i < USER_PTRS_PER_PGD; ++i)
	295	kmem_cache_free(pmd_cache, (void *)__va(pgd_val(pgd[i])-1));
e0da382c	296	/* in the non-PAE case, free_pgtables() clears user pgd entries */
1da177e4 LT	297	kmem_cache_free(pgd_cache, pgd);
1da177e4 LT	298	}