summaryrefslogtreecommitdiffstats
path: root/mm/highmem.c
blob: 3be601c6f266cbbcc2edfc503c3fb0f6b2ff4267 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
/*
 * High memory handling common code and variables.
 *
 * (C) 1999 Andrea Arcangeli, SuSE GmbH, andrea@suse.de
 *          Gerhard Wichert, Siemens AG, Gerhard.Wichert@pdb.siemens.de
 *
 *
 * Redesigned the x86 32-bit VM architecture to deal with
 * 64-bit physical space. With current x86 CPUs this
 * means up to 64 Gigabytes physical RAM.
 *
 * Rewrote high memory support to move the page cache into
 * high memory. Implemented permanent (schedulable) kmaps
 * based on Linus' idea.
 *
 * Copyright (C) 1999 Ingo Molnar <mingo@redhat.com>
 */

#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/highmem.h>
#include <linux/swap.h>
#include <linux/slab.h>

/*
 * Take one locked page, return another low-memory locked page.
 */
struct page * prepare_highmem_swapout(struct page * page)
{
	struct page *new_page;
	unsigned long regular_page;
	unsigned long vaddr;
	/*
	 * If this is a highmem page so it can't be swapped out directly
	 * otherwise the b_data buffer addresses will break
	 * the lowlevel device drivers.
	 */
	if (!PageHighMem(page))
		return page;

	/*
	 * Here we break the page lock, and we split the
	 * dirty page into two. We can unlock the old page,
	 * and we'll now have two of them. Too bad, it would
	 * have been nice to continue to potentially share
	 * across a fork().
	 */
	UnlockPage(page);
	regular_page = __get_free_page(GFP_ATOMIC);
	if (!regular_page)
		return NULL;

	vaddr = kmap(page);
	copy_page((void *)regular_page, (void *)vaddr);
	kunmap(page);

	/*
	 * ok, we can just forget about our highmem page since 
	 * we stored its data into the new regular_page.
	 */
	page_cache_release(page);
	new_page = virt_to_page(regular_page);
	LockPage(new_page);
	return new_page;
}

struct page * replace_with_highmem(struct page * page)
{
	struct page *highpage;
	unsigned long vaddr;

	if (PageHighMem(page) || !nr_free_highpages())
		return page;

	highpage = alloc_page(GFP_ATOMIC|__GFP_HIGHMEM);
	if (!highpage)
		return page;
	if (!PageHighMem(highpage)) {
		page_cache_release(highpage);
		return page;
	}

	vaddr = kmap(highpage);
	copy_page((void *)vaddr, page_address(page));
	kunmap(highpage);

	if (page->mapping)
		BUG();

	/*
	 * We can just forget the old page since 
	 * we stored its data into the new highmem-page.
	 */
	page_cache_release(page);

	return highpage;
}

/*
 * Virtual_count is not a pure "count".
 *  0 means that it is not mapped, and has not been mapped
 *    since a TLB flush - it is usable.
 *  1 means that there are no users, but it has been mapped
 *    since the last TLB flush - so we can't use it.
 *  n means that there are (n-1) current users of it.
 */
static int pkmap_count[LAST_PKMAP];
static unsigned int last_pkmap_nr;
static spinlock_t kmap_lock = SPIN_LOCK_UNLOCKED;

pte_t * pkmap_page_table;

static DECLARE_WAIT_QUEUE_HEAD(pkmap_map_wait);

static void flush_all_zero_pkmaps(void)
{
	int i;

	flush_cache_all();

	for (i = 0; i < LAST_PKMAP; i++) {
		struct page *page;
		pte_t pte;
		/*
		 * zero means we don't have anything to do,
		 * >1 means that it is still in use. Only
		 * a count of 1 means that it is free but
		 * needs to be unmapped
		 */
		if (pkmap_count[i] != 1)
			continue;
		pkmap_count[i] = 0;
		pte = pkmap_page_table[i];
		if (pte_none(pte))
			BUG();
		pte_clear(pkmap_page_table+i);
		page = pte_page(pte);
		page->virtual = NULL;
	}
	flush_tlb_all();
}

static inline unsigned long map_new_virtual(struct page *page)
{
	unsigned long vaddr;
	int count;

start:
	count = LAST_PKMAP;
	/* Find an empty entry */
	for (;;) {
		last_pkmap_nr = (last_pkmap_nr + 1) & LAST_PKMAP_MASK;
		if (!last_pkmap_nr) {
			flush_all_zero_pkmaps();
			count = LAST_PKMAP;
		}
		if (!pkmap_count[last_pkmap_nr])
			break;	/* Found a usable entry */
		if (--count)
			continue;

		/*
		 * Sleep for somebody else to unmap their entries
		 */
		{
			DECLARE_WAITQUEUE(wait, current);

			current->state = TASK_UNINTERRUPTIBLE;
			add_wait_queue(&pkmap_map_wait, &wait);
			spin_unlock(&kmap_lock);
			schedule();
			remove_wait_queue(&pkmap_map_wait, &wait);
			spin_lock(&kmap_lock);

			/* Somebody else might have mapped it while we slept */
			if (page->virtual)
				return (unsigned long) page->virtual;

			/* Re-start */
			goto start;
		}
	}
	vaddr = PKMAP_ADDR(last_pkmap_nr);
	set_pte(&(pkmap_page_table[last_pkmap_nr]), mk_pte(page, kmap_prot));

	pkmap_count[last_pkmap_nr] = 1;
	page->virtual = (void *) vaddr;

	return vaddr;
}

unsigned long kmap_high(struct page *page)
{
	unsigned long vaddr;

	/*
	 * For highmem pages, we can't trust "virtual" until
	 * after we have the lock.
	 *
	 * We cannot call this from interrupts, as it may block
	 */
	spin_lock(&kmap_lock);
	vaddr = (unsigned long) page->virtual;
	if (!vaddr)
		vaddr = map_new_virtual(page);
	pkmap_count[PKMAP_NR(vaddr)]++;
	if (pkmap_count[PKMAP_NR(vaddr)] < 2)
		BUG();
	spin_unlock(&kmap_lock);
	return vaddr;
}

void kunmap_high(struct page *page)
{
	unsigned long vaddr;
	unsigned long nr;

	spin_lock(&kmap_lock);
	vaddr = (unsigned long) page->virtual;
	if (!vaddr)
		BUG();
	nr = PKMAP_NR(vaddr);

	/*
	 * A count must never go down to zero
	 * without a TLB flush!
	 */
	switch (--pkmap_count[nr]) {
	case 0:
		BUG();
	case 1:
		wake_up(&pkmap_map_wait);
	}
	spin_unlock(&kmap_lock);
}

/*
 * Simple bounce buffer support for highmem pages.
 * This will be moved to the block layer in 2.5.
 */

static inline void copy_from_high_bh (struct buffer_head *to,
			 struct buffer_head *from)
{
	struct page *p_from;
	unsigned long vfrom;
	unsigned long flags;

	p_from = from->b_page;

	/*
	 * Since this can be executed from IRQ context, reentrance
	 * on the same CPU must be avoided:
	 */
	__save_flags(flags);
	__cli();
	vfrom = kmap_atomic(p_from, KM_BOUNCE_WRITE);
	memcpy(to->b_data, (char *)vfrom + bh_offset(from), to->b_size);
	kunmap_atomic(vfrom, KM_BOUNCE_WRITE);
	__restore_flags(flags);
}

static inline void copy_to_high_bh_irq (struct buffer_head *to,
			 struct buffer_head *from)
{
	struct page *p_to;
	unsigned long vto;
	unsigned long flags;

	p_to = to->b_page;
	__save_flags(flags);
	__cli();
	vto = kmap_atomic(p_to, KM_BOUNCE_READ);
	memcpy((char *)vto + bh_offset(to), from->b_data, to->b_size);
	kunmap_atomic(vto, KM_BOUNCE_READ);
	__restore_flags(flags);
}

static inline void bounce_end_io (struct buffer_head *bh, int uptodate)
{
	struct buffer_head *bh_orig = (struct buffer_head *)(bh->b_private);

	bh_orig->b_end_io(bh_orig, uptodate);
	__free_page(bh->b_page);
	kmem_cache_free(bh_cachep, bh);
}

static void bounce_end_io_write (struct buffer_head *bh, int uptodate)
{
	bounce_end_io(bh, uptodate);
}

static void bounce_end_io_read (struct buffer_head *bh, int uptodate)
{
	struct buffer_head *bh_orig = (struct buffer_head *)(bh->b_private);

	if (uptodate)
		copy_to_high_bh_irq(bh_orig, bh);
	bounce_end_io(bh, uptodate);
}

struct buffer_head * create_bounce(int rw, struct buffer_head * bh_orig)
{
	struct page *page;
	struct buffer_head *bh;

	if (!PageHighMem(bh_orig->b_page))
		return bh_orig;

repeat_bh:
	bh = kmem_cache_alloc(bh_cachep, SLAB_BUFFER);
	if (!bh) {
		wakeup_bdflush(1);
		current->policy |= SCHED_YIELD;
		schedule();
		goto repeat_bh;
	}
	/*
	 * This is wasteful for 1k buffers, but this is a stopgap measure
	 * and we are being ineffective anyway. This approach simplifies
	 * things immensly. On boxes with more than 4GB RAM this should
	 * not be an issue anyway.
	 */
repeat_page:
	page = alloc_page(GFP_BUFFER);
	if (!page) {
		wakeup_bdflush(1);
		current->policy |= SCHED_YIELD;
		schedule();
		goto repeat_page;
	}
	set_bh_page(bh, page, 0);

	bh->b_next = NULL;
	bh->b_blocknr = bh_orig->b_blocknr;
	bh->b_size = bh_orig->b_size;
	bh->b_list = -1;
	bh->b_dev = bh_orig->b_dev;
	bh->b_count = bh_orig->b_count;
	bh->b_rdev = bh_orig->b_rdev;
	bh->b_state = bh_orig->b_state;
	bh->b_flushtime = jiffies;
	bh->b_next_free = NULL;
	bh->b_prev_free = NULL;
	/* bh->b_this_page */
	bh->b_reqnext = NULL;
	bh->b_pprev = NULL;
	/* bh->b_page */
	if (rw == WRITE) {
		bh->b_end_io = bounce_end_io_write;
		copy_from_high_bh(bh, bh_orig);
	} else
		bh->b_end_io = bounce_end_io_read;
	bh->b_private = (void *)bh_orig;
	bh->b_rsector = bh_orig->b_rsector;
	memset(&bh->b_wait, -1, sizeof(bh->b_wait));

	return bh;
}