summaryrefslogtreecommitdiffstats
path: root/drivers/gpu/drm/gma500/mmu.c
blob: 49bac41beefb3251ecc6a63f04e86d0d36de9974 (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
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
/**************************************************************************
 * Copyright (c) 2007, Intel Corporation.
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms and conditions of the GNU General Public License,
 * version 2, as published by the Free Software Foundation.
 *
 * This program is distributed in the hope it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
 * more details.
 *
 * You should have received a copy of the GNU General Public License along with
 * this program; if not, write to the Free Software Foundation, Inc.,
 * 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
 *
 **************************************************************************/
#include <drm/drmP.h>
#include "psb_drv.h"
#include "psb_reg.h"

/*
 * Code for the SGX MMU:
 */

/*
 * clflush on one processor only:
 * clflush should apparently flush the cache line on all processors in an
 * SMP system.
 */

/*
 * kmap atomic:
 * The usage of the slots must be completely encapsulated within a spinlock, and
 * no other functions that may be using the locks for other purposed may be
 * called from within the locked region.
 * Since the slots are per processor, this will guarantee that we are the only
 * user.
 */

/*
 * TODO: Inserting ptes from an interrupt handler:
 * This may be desirable for some SGX functionality where the GPU can fault in
 * needed pages. For that, we need to make an atomic insert_pages function, that
 * may fail.
 * If it fails, the caller need to insert the page using a workqueue function,
 * but on average it should be fast.
 */

struct psb_mmu_driver {
	/* protects driver- and pd structures. Always take in read mode
	 * before taking the page table spinlock.
	 */
	struct rw_semaphore sem;

	/* protects page tables, directory tables and pt tables.
	 * and pt structures.
	 */
	spinlock_t lock;

	atomic_t needs_tlbflush;

	uint8_t __iomem *register_map;
	struct psb_mmu_pd *default_pd;
	/*uint32_t bif_ctrl;*/
	int has_clflush;
	int clflush_add;
	unsigned long clflush_mask;

	struct drm_psb_private *dev_priv;
};

struct psb_mmu_pd;

struct psb_mmu_pt {
	struct psb_mmu_pd *pd;
	uint32_t index;
	uint32_t count;
	struct page *p;
	uint32_t *v;
};

struct psb_mmu_pd {
	struct psb_mmu_driver *driver;
	int hw_context;
	struct psb_mmu_pt **tables;
	struct page *p;
	struct page *dummy_pt;
	struct page *dummy_page;
	uint32_t pd_mask;
	uint32_t invalid_pde;
	uint32_t invalid_pte;
};

static inline uint32_t psb_mmu_pt_index(uint32_t offset)
{
	return (offset >> PSB_PTE_SHIFT) & 0x3FF;
}

static inline uint32_t psb_mmu_pd_index(uint32_t offset)
{
	return offset >> PSB_PDE_SHIFT;
}

static inline void psb_clflush(void *addr)
{
	__asm__ __volatile__("clflush (%0)\n" : : "r"(addr) : "memory");
}

static inline void psb_mmu_clflush(struct psb_mmu_driver *driver,
				   void *addr)
{
	if (!driver->has_clflush)
		return;

	mb();
	psb_clflush(addr);
	mb();
}

static void psb_page_clflush(struct psb_mmu_driver *driver, struct page* page)
{
	uint32_t clflush_add = driver->clflush_add >> PAGE_SHIFT;
	uint32_t clflush_count = PAGE_SIZE / clflush_add;
	int i;
	uint8_t *clf;

	clf = kmap_atomic(page);
	mb();
	for (i = 0; i < clflush_count; ++i) {
		psb_clflush(clf);
		clf += clflush_add;
	}
	mb();
	kunmap_atomic(clf);
}

static void psb_pages_clflush(struct psb_mmu_driver *driver,
				struct page *page[], unsigned long num_pages)
{
	int i;

	if (!driver->has_clflush)
		return ;

	for (i = 0; i < num_pages; i++)
		psb_page_clflush(driver, *page++);
}

static void psb_mmu_flush_pd_locked(struct psb_mmu_driver *driver,
				    int force)
{
	atomic_set(&driver->needs_tlbflush, 0);
}

static void psb_mmu_flush_pd(struct psb_mmu_driver *driver, int force)
{
	down_write(&driver->sem);
	psb_mmu_flush_pd_locked(driver, force);
	up_write(&driver->sem);
}

void psb_mmu_flush(struct psb_mmu_driver *driver, int rc_prot)
{
	if (rc_prot)
		down_write(&driver->sem);
	if (rc_prot)
		up_write(&driver->sem);
}

void psb_mmu_set_pd_context(struct psb_mmu_pd *pd, int hw_context)
{
	/*ttm_tt_cache_flush(&pd->p, 1);*/
	psb_pages_clflush(pd->driver, &pd->p, 1);
	down_write(&pd->driver->sem);
	wmb();
	psb_mmu_flush_pd_locked(pd->driver, 1);
	pd->hw_context = hw_context;
	up_write(&pd->driver->sem);

}

static inline unsigned long psb_pd_addr_end(unsigned long addr,
					    unsigned long end)
{

	addr = (addr + PSB_PDE_MASK + 1) & ~PSB_PDE_MASK;
	return (addr < end) ? addr : end;
}

static inline uint32_t psb_mmu_mask_pte(uint32_t pfn, int type)
{
	uint32_t mask = PSB_PTE_VALID;

	if (type & PSB_MMU_CACHED_MEMORY)
		mask |= PSB_PTE_CACHED;
	if (type & PSB_MMU_RO_MEMORY)
		mask |= PSB_PTE_RO;
	if (type & PSB_MMU_WO_MEMORY)
		mask |= PSB_PTE_WO;

	return (pfn << PAGE_SHIFT) | mask;
}

struct psb_mmu_pd *psb_mmu_alloc_pd(struct psb_mmu_driver *driver,
				    int trap_pagefaults, int invalid_type)
{
	struct psb_mmu_pd *pd = kmalloc(sizeof(*pd), GFP_KERNEL);
	uint32_t *v;
	int i;

	if (!pd)
		return NULL;

	pd->p = alloc_page(GFP_DMA32);
	if (!pd->p)
		goto out_err1;
	pd->dummy_pt = alloc_page(GFP_DMA32);
	if (!pd->dummy_pt)
		goto out_err2;
	pd->dummy_page = alloc_page(GFP_DMA32);
	if (!pd->dummy_page)
		goto out_err3;

	if (!trap_pagefaults) {
		pd->invalid_pde =
		    psb_mmu_mask_pte(page_to_pfn(pd->dummy_pt),
				     invalid_type);
		pd->invalid_pte =
		    psb_mmu_mask_pte(page_to_pfn(pd->dummy_page),
				     invalid_type);
	} else {
		pd->invalid_pde = 0;
		pd->invalid_pte = 0;
	}

	v = kmap(pd->dummy_pt);
	for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i)
		v[i] = pd->invalid_pte;

	kunmap(pd->dummy_pt);

	v = kmap(pd->p);
	for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i)
		v[i] = pd->invalid_pde;

	kunmap(pd->p);

	clear_page(kmap(pd->dummy_page));
	kunmap(pd->dummy_page);

	pd->tables = vmalloc_user(sizeof(struct psb_mmu_pt *) * 1024);
	if (!pd->tables)
		goto out_err4;

	pd->hw_context = -1;
	pd->pd_mask = PSB_PTE_VALID;
	pd->driver = driver;

	return pd;

out_err4:
	__free_page(pd->dummy_page);
out_err3:
	__free_page(pd->dummy_pt);
out_err2:
	__free_page(pd->p);
out_err1:
	kfree(pd);
	return NULL;
}

static void psb_mmu_free_pt(struct psb_mmu_pt *pt)
{
	__free_page(pt->p);
	kfree(pt);
}

void psb_mmu_free_pagedir(struct psb_mmu_pd *pd)
{
	struct psb_mmu_driver *driver = pd->driver;
	struct psb_mmu_pt *pt;
	int i;

	down_write(&driver->sem);
	if (pd->hw_context != -1)
		psb_mmu_flush_pd_locked(driver, 1);

	/* Should take the spinlock here, but we don't need to do that
	   since we have the semaphore in write mode. */

	for (i = 0; i < 1024; ++i) {
		pt = pd->tables[i];
		if (pt)
			psb_mmu_free_pt(pt);
	}

	vfree(pd->tables);
	__free_page(pd->dummy_page);
	__free_page(pd->dummy_pt);
	__free_page(pd->p);
	kfree(pd);
	up_write(&driver->sem);
}

static struct psb_mmu_pt *psb_mmu_alloc_pt(struct psb_mmu_pd *pd)
{
	struct psb_mmu_pt *pt = kmalloc(sizeof(*pt), GFP_KERNEL);
	void *v;
	uint32_t clflush_add = pd->driver->clflush_add >> PAGE_SHIFT;
	uint32_t clflush_count = PAGE_SIZE / clflush_add;
	spinlock_t *lock = &pd->driver->lock;
	uint8_t *clf;
	uint32_t *ptes;
	int i;

	if (!pt)
		return NULL;

	pt->p = alloc_page(GFP_DMA32);
	if (!pt->p) {
		kfree(pt);
		return NULL;
	}

	spin_lock(lock);

	v = kmap_atomic(pt->p);
	clf = (uint8_t *) v;
	ptes = (uint32_t *) v;
	for (i = 0; i < (PAGE_SIZE / sizeof(uint32_t)); ++i)
		*ptes++ = pd->invalid_pte;


	if (pd->driver->has_clflush && pd->hw_context != -1) {
		mb();
		for (i = 0; i < clflush_count; ++i) {
			psb_clflush(clf);
			clf += clflush_add;
		}
		mb();
	}

	kunmap_atomic(v);
	spin_unlock(lock);

	pt->count = 0;
	pt->pd = pd;
	pt->index = 0;

	return pt;
}

static struct psb_mmu_pt *psb_mmu_pt_alloc_map_lock(struct psb_mmu_pd *pd,
					     unsigned long addr)
{
	uint32_t index = psb_mmu_pd_index(addr);
	struct psb_mmu_pt *pt;
	uint32_t *v;
	spinlock_t *lock = &pd->driver->lock;

	spin_lock(lock);
	pt = pd->tables[index];
	while (!pt) {
		spin_unlock(lock);
		pt = psb_mmu_alloc_pt(pd);
		if (!pt)
			return NULL;
		spin_lock(lock);

		if (pd->tables[index]) {
			spin_unlock(lock);
			psb_mmu_free_pt(pt);
			spin_lock(lock);
			pt = pd->tables[index];
			continue;
		}

		v = kmap_atomic(pd->p);
		pd->tables[index] = pt;
		v[index] = (page_to_pfn(pt->p) << 12) | pd->pd_mask;
		pt->index = index;
		kunmap_atomic((void *) v);

		if (pd->hw_context != -1) {
			psb_mmu_clflush(pd->driver, (void *) &v[index]);
			atomic_set(&pd->driver->needs_tlbflush, 1);
		}
	}
	pt->v = kmap_atomic(pt->p);
	return pt;
}

static struct psb_mmu_pt *psb_mmu_pt_map_lock(struct psb_mmu_pd *pd,
					      unsigned long addr)
{
	uint32_t index = psb_mmu_pd_index(addr);
	struct psb_mmu_pt *pt;
	spinlock_t *lock = &pd->driver->lock;

	spin_lock(lock);
	pt = pd->tables[index];
	if (!pt) {
		spin_unlock(lock);
		return NULL;
	}
	pt->v = kmap_atomic(pt->p);
	return pt;
}

static void psb_mmu_pt_unmap_unlock(struct psb_mmu_pt *pt)
{
	struct psb_mmu_pd *pd = pt->pd;
	uint32_t *v;

	kunmap_atomic(pt->v);
	if (pt->count == 0) {
		v = kmap_atomic(pd->p);
		v[pt->index] = pd->invalid_pde;
		pd->tables[pt->index] = NULL;

		if (pd->hw_context != -1) {
			psb_mmu_clflush(pd->driver,
					(void *) &v[pt->index]);
			atomic_set(&pd->driver->needs_tlbflush, 1);
		}
		kunmap_atomic(pt->v);
		spin_unlock(&pd->driver->lock);
		psb_mmu_free_pt(pt);
		return;
	}
	spin_unlock(&pd->driver->lock);
}

static inline void psb_mmu_set_pte(struct psb_mmu_pt *pt,
				   unsigned long addr, uint32_t pte)
{
	pt->v[psb_mmu_pt_index(addr)] = pte;
}

static inline void psb_mmu_invalidate_pte(struct psb_mmu_pt *pt,
					  unsigned long addr)
{
	pt->v[psb_mmu_pt_index(addr)] = pt->pd->invalid_pte;
}


void psb_mmu_mirror_gtt(struct psb_mmu_pd *pd,
			uint32_t mmu_offset, uint32_t gtt_start,
			uint32_t gtt_pages)
{
	uint32_t *v;
	uint32_t start = psb_mmu_pd_index(mmu_offset);
	struct psb_mmu_driver *driver = pd->driver;
	int num_pages = gtt_pages;

	down_read(&driver->sem);
	spin_lock(&driver->lock);

	v = kmap_atomic(pd->p);
	v += start;

	while (gtt_pages--) {
		*v++ = gtt_start | pd->pd_mask;
		gtt_start += PAGE_SIZE;
	}

	/*ttm_tt_cache_flush(&pd->p, num_pages);*/
	psb_pages_clflush(pd->driver, &pd->p, num_pages);
	kunmap_atomic(v);
	spin_unlock(&driver->lock);

	if (pd->hw_context != -1)
		atomic_set(&pd->driver->needs_tlbflush, 1);

	up_read(&pd->driver->sem);
	psb_mmu_flush_pd(pd->driver, 0);
}

struct psb_mmu_pd *psb_mmu_get_default_pd(struct psb_mmu_driver *driver)
{
	struct psb_mmu_pd *pd;

	/* down_read(&driver->sem); */
	pd = driver->default_pd;
	/* up_read(&driver->sem); */

	return pd;
}

void psb_mmu_driver_takedown(struct psb_mmu_driver *driver)
{
	psb_mmu_free_pagedir(driver->default_pd);
	kfree(driver);
}

struct psb_mmu_driver *psb_mmu_driver_init(uint8_t __iomem * registers,
					int trap_pagefaults,
					int invalid_type,
					struct drm_psb_private *dev_priv)
{
	struct psb_mmu_driver *driver;

	driver = kmalloc(sizeof(*driver), GFP_KERNEL);

	if (!driver)
		return NULL;
	driver->dev_priv = dev_priv;

	driver->default_pd = psb_mmu_alloc_pd(driver, trap_pagefaults,
					      invalid_type);
	if (!driver->default_pd)
		goto out_err1;

	spin_lock_init(&driver->lock);
	init_rwsem(&driver->sem);
	down_write(&driver->sem);
	driver->register_map = registers;
	atomic_set(&driver->needs_tlbflush, 1);

	driver->has_clflush = 0;

	if (boot_cpu_has(X86_FEATURE_CLFLSH)) {
		uint32_t tfms, misc, cap0, cap4, clflush_size;

		/*
		 * clflush size is determined at kernel setup for x86_64
		 *  but not for i386. We have to do it here.
		 */

		cpuid(0x00000001, &tfms, &misc, &cap0, &cap4);
		clflush_size = ((misc >> 8) & 0xff) * 8;
		driver->has_clflush = 1;
		driver->clflush_add =
		    PAGE_SIZE * clflush_size / sizeof(uint32_t);
		driver->clflush_mask = driver->clflush_add - 1;
		driver->clflush_mask = ~driver->clflush_mask;
	}

	up_write(&driver->sem);
	return driver;

out_err1:
	kfree(driver);
	return NULL;
}

static void psb_mmu_flush_ptes(struct psb_mmu_pd *pd,
			       unsigned long address, uint32_t num_pages,
			       uint32_t desired_tile_stride,
			       uint32_t hw_tile_stride)
{
	struct psb_mmu_pt *pt;
	uint32_t rows = 1;
	uint32_t i;
	unsigned long addr;
	unsigned long end;
	unsigned long next;
	unsigned long add;
	unsigned long row_add;
	unsigned long clflush_add = pd->driver->clflush_add;
	unsigned long clflush_mask = pd->driver->clflush_mask;

	if (!pd->driver->has_clflush) {
		/*ttm_tt_cache_flush(&pd->p, num_pages);*/
		psb_pages_clflush(pd->driver, &pd->p, num_pages);
		return;
	}

	if (hw_tile_stride)
		rows = num_pages / desired_tile_stride;
	else
		desired_tile_stride = num_pages;

	add = desired_tile_stride << PAGE_SHIFT;
	row_add = hw_tile_stride << PAGE_SHIFT;
	mb();
	for (i = 0; i < rows; ++i) {

		addr = address;
		end = addr + add;

		do {
			next = psb_pd_addr_end(addr, end);
			pt = psb_mmu_pt_map_lock(pd, addr);
			if (!pt)
				continue;
			do {
				psb_clflush(&pt->v
					    [psb_mmu_pt_index(addr)]);
			} while (addr +=
				 clflush_add,
				 (addr & clflush_mask) < next);

			psb_mmu_pt_unmap_unlock(pt);
		} while (addr = next, next != end);
		address += row_add;
	}
	mb();
}

void psb_mmu_remove_pfn_sequence(struct psb_mmu_pd *pd,
				 unsigned long address, uint32_t num_pages)
{
	struct psb_mmu_pt *pt;
	unsigned long addr;
	unsigned long end;
	unsigned long next;
	unsigned long f_address = address;

	down_read(&pd->driver->sem);

	addr = address;
	end = addr + (num_pages << PAGE_SHIFT);

	do {
		next = psb_pd_addr_end(addr, end);
		pt = psb_mmu_pt_alloc_map_lock(pd, addr);
		if (!pt)
			goto out;
		do {
			psb_mmu_invalidate_pte(pt, addr);
			--pt->count;
		} while (addr += PAGE_SIZE, addr < next);
		psb_mmu_pt_unmap_unlock(pt);

	} while (addr = next, next != end);

out:
	if (pd->hw_context != -1)
		psb_mmu_flush_ptes(pd, f_address, num_pages, 1, 1);

	up_read(&pd->driver->sem);

	if (pd->hw_context != -1)
		psb_mmu_flush(pd->driver, 0);

	return;
}

void psb_mmu_remove_pages(struct psb_mmu_pd *pd, unsigned long address,
			  uint32_t num_pages, uint32_t desired_tile_stride,
			  uint32_t hw_tile_stride)
{
	struct psb_mmu_pt *pt;
	uint32_t rows = 1;
	uint32_t i;
	unsigned long addr;
	unsigned long end;
	unsigned long next;
	unsigned long add;
	unsigned long row_add;
	unsigned long f_address = address;

	if (hw_tile_stride)
		rows = num_pages / desired_tile_stride;
	else
		desired_tile_stride = num_pages;

	add = desired_tile_stride << PAGE_SHIFT;
	row_add = hw_tile_stride << PAGE_SHIFT;

	/* down_read(&pd->driver->sem); */

	/* Make sure we only need to flush this processor's cache */

	for (i = 0; i < rows; ++i) {

		addr = address;
		end = addr + add;

		do {
			next = psb_pd_addr_end(addr, end);
			pt = psb_mmu_pt_map_lock(pd, addr);
			if (!pt)
				continue;
			do {
				psb_mmu_invalidate_pte(pt, addr);
				--pt->count;

			} while (addr += PAGE_SIZE, addr < next);
			psb_mmu_pt_unmap_unlock(pt);

		} while (addr = next, next != end);
		address += row_add;
	}
	if (pd->hw_context != -1)
		psb_mmu_flush_ptes(pd, f_address, num_pages,
				   desired_tile_stride, hw_tile_stride);

	/* up_read(&pd->driver->sem); */

	if (pd->hw_context != -1)
		psb_mmu_flush(pd->driver, 0);
}

int psb_mmu_insert_pfn_sequence(struct psb_mmu_pd *pd, uint32_t start_pfn,
				unsigned long address, uint32_t num_pages,
				int type)
{
	struct psb_mmu_pt *pt;
	uint32_t pte;
	unsigned long addr;
	unsigned long end;
	unsigned long next;
	unsigned long f_address = address;
	int ret = 0;

	down_read(&pd->driver->sem);

	addr = address;
	end = addr + (num_pages << PAGE_SHIFT);

	do {
		next = psb_pd_addr_end(addr, end);
		pt = psb_mmu_pt_alloc_map_lock(pd, addr);
		if (!pt) {
			ret = -ENOMEM;
			goto out;
		}
		do {
			pte = psb_mmu_mask_pte(start_pfn++, type);
			psb_mmu_set_pte(pt, addr, pte);
			pt->count++;
		} while (addr += PAGE_SIZE, addr < next);
		psb_mmu_pt_unmap_unlock(pt);

	} while (addr = next, next != end);

out:
	if (pd->hw_context != -1)
		psb_mmu_flush_ptes(pd, f_address, num_pages, 1, 1);

	up_read(&pd->driver->sem);

	if (pd->hw_context != -1)
		psb_mmu_flush(pd->driver, 1);

	return ret;
}

int psb_mmu_insert_pages(struct psb_mmu_pd *pd, struct page **pages,
			 unsigned long address, uint32_t num_pages,
			 uint32_t desired_tile_stride,
			 uint32_t hw_tile_stride, int type)
{
	struct psb_mmu_pt *pt;
	uint32_t rows = 1;
	uint32_t i;
	uint32_t pte;
	unsigned long addr;
	unsigned long end;
	unsigned long next;
	unsigned long add;
	unsigned long row_add;
	unsigned long f_address = address;
	int ret = 0;

	if (hw_tile_stride) {
		if (num_pages % desired_tile_stride != 0)
			return -EINVAL;
		rows = num_pages / desired_tile_stride;
	} else {
		desired_tile_stride = num_pages;
	}

	add = desired_tile_stride << PAGE_SHIFT;
	row_add = hw_tile_stride << PAGE_SHIFT;

	down_read(&pd->driver->sem);

	for (i = 0; i < rows; ++i) {

		addr = address;
		end = addr + add;

		do {
			next = psb_pd_addr_end(addr, end);
			pt = psb_mmu_pt_alloc_map_lock(pd, addr);
			if (!pt) {
				ret = -ENOMEM;
				goto out;
			}
			do {
				pte =
				    psb_mmu_mask_pte(page_to_pfn(*pages++),
						     type);
				psb_mmu_set_pte(pt, addr, pte);
				pt->count++;
			} while (addr += PAGE_SIZE, addr < next);
			psb_mmu_pt_unmap_unlock(pt);

		} while (addr = next, next != end);

		address += row_add;
	}
out:
	if (pd->hw_context != -1)
		psb_mmu_flush_ptes(pd, f_address, num_pages,
				   desired_tile_stride, hw_tile_stride);

	up_read(&pd->driver->sem);

	if (pd->hw_context != -1)
		psb_mmu_flush(pd->driver, 1);

	return ret;
}

int psb_mmu_virtual_to_pfn(struct psb_mmu_pd *pd, uint32_t virtual,
			   unsigned long *pfn)
{
	int ret;
	struct psb_mmu_pt *pt;
	uint32_t tmp;
	spinlock_t *lock = &pd->driver->lock;

	down_read(&pd->driver->sem);
	pt = psb_mmu_pt_map_lock(pd, virtual);
	if (!pt) {
		uint32_t *v;

		spin_lock(lock);
		v = kmap_atomic(pd->p);
		tmp = v[psb_mmu_pd_index(virtual)];
		kunmap_atomic(v);
		spin_unlock(lock);

		if (tmp != pd->invalid_pde || !(tmp & PSB_PTE_VALID) ||
		    !(pd->invalid_pte & PSB_PTE_VALID)) {
			ret = -EINVAL;
			goto out;
		}
		ret = 0;
		*pfn = pd->invalid_pte >> PAGE_SHIFT;
		goto out;
	}
	tmp = pt->v[psb_mmu_pt_index(virtual)];
	if (!(tmp & PSB_PTE_VALID)) {
		ret = -EINVAL;
	} else {
		ret = 0;
		*pfn = tmp >> PAGE_SHIFT;
	}
	psb_mmu_pt_unmap_unlock(pt);
out:
	up_read(&pd->driver->sem);
	return ret;
}