[net-next-2.6.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;

	printk(KERN_INFO "Mem-info:\n");
	show_free_areas();
	printk(KERN_INFO "Free swap:       %6ldkB\n", nr_swap_pages<<(PAGE_SHIFT-10));
	for_each_pgdat(pgdat) {
		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;
		}
	}
	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", ps.nr_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)
		pgd_list->private = (unsigned long)&page->index;
	pgd_list = page;
	page->private = (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;
	*pprev = next;
	if (next)
		next->private = (unsigned long)pprev;
}

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

	if (PTRS_PER_PMD == 1)
		spin_lock_irqsave(&pgd_lock, flags);

	memcpy((pgd_t *)pgd + USER_PTRS_PER_PGD,
			swapper_pg_dir + USER_PTRS_PER_PGD,
			(PTRS_PER_PGD - USER_PTRS_PER_PGD) * sizeof(pgd_t));

	if (PTRS_PER_PMD > 1)
		return;

	pgd_list_add(pgd);
	spin_unlock_irqrestore(&pgd_lock, flags);
	memset(pgd, 0, USER_PTRS_PER_PGD*sizeof(pgd_t));
}

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