summaryrefslogtreecommitdiff
path: root/drivers/video/fbdev/vermilion/vermilion.c
blob: 840ead69654b85ea2e2927fc315dc3e7b90cec0d (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
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * Copyright (c) Intel Corp. 2007.
 * All Rights Reserved.
 *
 * Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
 * develop this driver.
 *
 * This file is part of the Vermilion Range fb driver.
 *
 * Authors:
 *   Thomas Hellström <thomas-at-tungstengraphics-dot-com>
 *   Michel Dänzer <michel-at-tungstengraphics-dot-com>
 *   Alan Hourihane <alanh-at-tungstengraphics-dot-com>
 */

#include <linux/aperture.h>
#include <linux/module.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/delay.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/fb.h>
#include <linux/pci.h>
#include <asm/set_memory.h>
#include <asm/tlbflush.h>
#include <linux/mmzone.h>

/* #define VERMILION_DEBUG */

#include "vermilion.h"

#define MODULE_NAME "vmlfb"

#define VML_TOHW(_val, _width) ((((_val) << (_width)) + 0x7FFF - (_val)) >> 16)

static struct mutex vml_mutex;
static struct list_head global_no_mode;
static struct list_head global_has_mode;
static struct fb_ops vmlfb_ops;
static struct vml_sys *subsys = NULL;
static char *vml_default_mode = "1024x768@60";
static const struct fb_videomode defaultmode = {
	NULL, 60, 1024, 768, 12896, 144, 24, 29, 3, 136, 6,
	0, FB_VMODE_NONINTERLACED
};

static u32 vml_mem_requested = (10 * 1024 * 1024);
static u32 vml_mem_contig = (4 * 1024 * 1024);
static u32 vml_mem_min = (4 * 1024 * 1024);

static u32 vml_clocks[] = {
	6750,
	13500,
	27000,
	29700,
	37125,
	54000,
	59400,
	74250,
	120000,
	148500
};

static u32 vml_num_clocks = ARRAY_SIZE(vml_clocks);

/*
 * Allocate a contiguous vram area and make its linear kernel map
 * uncached.
 */

static int vmlfb_alloc_vram_area(struct vram_area *va, unsigned max_order,
				 unsigned min_order)
{
	gfp_t flags;
	unsigned long i;

	max_order++;
	do {
		/*
		 * Really try hard to get the needed memory.
		 * We need memory below the first 32MB, so we
		 * add the __GFP_DMA flag that guarantees that we are
		 * below the first 16MB.
		 */

		flags = __GFP_DMA | __GFP_HIGH | __GFP_KSWAPD_RECLAIM;
		va->logical =
			 __get_free_pages(flags, --max_order);
	} while (va->logical == 0 && max_order > min_order);

	if (!va->logical)
		return -ENOMEM;

	va->phys = virt_to_phys((void *)va->logical);
	va->size = PAGE_SIZE << max_order;
	va->order = max_order;

	/*
	 * It seems like __get_free_pages only ups the usage count
	 * of the first page. This doesn't work with fault mapping, so
	 * up the usage count once more (XXX: should use split_page or
	 * compound page).
	 */

	memset((void *)va->logical, 0x00, va->size);
	for (i = va->logical; i < va->logical + va->size; i += PAGE_SIZE) {
		get_page(virt_to_page(i));
	}

	/*
	 * Change caching policy of the linear kernel map to avoid
	 * mapping type conflicts with user-space mappings.
	 */
	set_pages_uc(virt_to_page(va->logical), va->size >> PAGE_SHIFT);

	printk(KERN_DEBUG MODULE_NAME
	       ": Allocated %ld bytes vram area at 0x%08lx\n",
	       va->size, va->phys);

	return 0;
}

/*
 * Free a contiguous vram area and reset its linear kernel map
 * mapping type.
 */

static void vmlfb_free_vram_area(struct vram_area *va)
{
	unsigned long j;

	if (va->logical) {

		/*
		 * Reset the linear kernel map caching policy.
		 */

		set_pages_wb(virt_to_page(va->logical),
				 va->size >> PAGE_SHIFT);

		/*
		 * Decrease the usage count on the pages we've used
		 * to compensate for upping when allocating.
		 */

		for (j = va->logical; j < va->logical + va->size;
		     j += PAGE_SIZE) {
			(void)put_page_testzero(virt_to_page(j));
		}

		printk(KERN_DEBUG MODULE_NAME
		       ": Freeing %ld bytes vram area at 0x%08lx\n",
		       va->size, va->phys);
		free_pages(va->logical, va->order);

		va->logical = 0;
	}
}

/*
 * Free allocated vram.
 */

static void vmlfb_free_vram(struct vml_info *vinfo)
{
	int i;

	for (i = 0; i < vinfo->num_areas; ++i) {
		vmlfb_free_vram_area(&vinfo->vram[i]);
	}
	vinfo->num_areas = 0;
}

/*
 * Allocate vram. Currently we try to allocate contiguous areas from the
 * __GFP_DMA zone and puzzle them together. A better approach would be to
 * allocate one contiguous area for scanout and use one-page allocations for
 * offscreen areas. This requires user-space and GPU virtual mappings.
 */

static int vmlfb_alloc_vram(struct vml_info *vinfo,
			    size_t requested,
			    size_t min_total, size_t min_contig)
{
	int i, j;
	int order;
	int contiguous;
	int err;
	struct vram_area *va;
	struct vram_area *va2;

	vinfo->num_areas = 0;
	for (i = 0; i < VML_VRAM_AREAS; ++i) {
		va = &vinfo->vram[i];
		order = 0;

		while (requested > (PAGE_SIZE << order) && order <= MAX_ORDER)
			order++;

		err = vmlfb_alloc_vram_area(va, order, 0);

		if (err)
			break;

		if (i == 0) {
			vinfo->vram_start = va->phys;
			vinfo->vram_logical = (void __iomem *) va->logical;
			vinfo->vram_contig_size = va->size;
			vinfo->num_areas = 1;
		} else {
			contiguous = 0;

			for (j = 0; j < i; ++j) {
				va2 = &vinfo->vram[j];
				if (va->phys + va->size == va2->phys ||
				    va2->phys + va2->size == va->phys) {
					contiguous = 1;
					break;
				}
			}

			if (contiguous) {
				vinfo->num_areas++;
				if (va->phys < vinfo->vram_start) {
					vinfo->vram_start = va->phys;
					vinfo->vram_logical =
						(void __iomem *)va->logical;
				}
				vinfo->vram_contig_size += va->size;
			} else {
				vmlfb_free_vram_area(va);
				break;
			}
		}

		if (requested < va->size)
			break;
		else
			requested -= va->size;
	}

	if (vinfo->vram_contig_size > min_total &&
	    vinfo->vram_contig_size > min_contig) {

		printk(KERN_DEBUG MODULE_NAME
		       ": Contiguous vram: %ld bytes at physical 0x%08lx.\n",
		       (unsigned long)vinfo->vram_contig_size,
		       (unsigned long)vinfo->vram_start);

		return 0;
	}

	printk(KERN_ERR MODULE_NAME
	       ": Could not allocate requested minimal amount of vram.\n");

	vmlfb_free_vram(vinfo);

	return -ENOMEM;
}

/*
 * Find the GPU to use with our display controller.
 */

static int vmlfb_get_gpu(struct vml_par *par)
{
	mutex_lock(&vml_mutex);

	par->gpu = pci_get_device(PCI_VENDOR_ID_INTEL, VML_DEVICE_GPU, NULL);

	if (!par->gpu) {
		mutex_unlock(&vml_mutex);
		return -ENODEV;
	}

	mutex_unlock(&vml_mutex);

	if (pci_enable_device(par->gpu) < 0) {
		pci_dev_put(par->gpu);
		return -ENODEV;
	}

	return 0;
}

/*
 * Find a contiguous vram area that contains a given offset from vram start.
 */
static int vmlfb_vram_offset(struct vml_info *vinfo, unsigned long offset)
{
	unsigned long aoffset;
	unsigned i;

	for (i = 0; i < vinfo->num_areas; ++i) {
		aoffset = offset - (vinfo->vram[i].phys - vinfo->vram_start);

		if (aoffset < vinfo->vram[i].size) {
			return 0;
		}
	}

	return -EINVAL;
}

/*
 * Remap the MMIO register spaces of the VDC and the GPU.
 */

static int vmlfb_enable_mmio(struct vml_par *par)
{
	int err;

	par->vdc_mem_base = pci_resource_start(par->vdc, 0);
	par->vdc_mem_size = pci_resource_len(par->vdc, 0);
	if (!request_mem_region(par->vdc_mem_base, par->vdc_mem_size, "vmlfb")) {
		printk(KERN_ERR MODULE_NAME
		       ": Could not claim display controller MMIO.\n");
		return -EBUSY;
	}
	par->vdc_mem = ioremap(par->vdc_mem_base, par->vdc_mem_size);
	if (par->vdc_mem == NULL) {
		printk(KERN_ERR MODULE_NAME
		       ": Could not map display controller MMIO.\n");
		err = -ENOMEM;
		goto out_err_0;
	}

	par->gpu_mem_base = pci_resource_start(par->gpu, 0);
	par->gpu_mem_size = pci_resource_len(par->gpu, 0);
	if (!request_mem_region(par->gpu_mem_base, par->gpu_mem_size, "vmlfb")) {
		printk(KERN_ERR MODULE_NAME ": Could not claim GPU MMIO.\n");
		err = -EBUSY;
		goto out_err_1;
	}
	par->gpu_mem = ioremap(par->gpu_mem_base, par->gpu_mem_size);
	if (par->gpu_mem == NULL) {
		printk(KERN_ERR MODULE_NAME ": Could not map GPU MMIO.\n");
		err = -ENOMEM;
		goto out_err_2;
	}

	return 0;

out_err_2:
	release_mem_region(par->gpu_mem_base, par->gpu_mem_size);
out_err_1:
	iounmap(par->vdc_mem);
out_err_0:
	release_mem_region(par->vdc_mem_base, par->vdc_mem_size);
	return err;
}

/*
 * Unmap the VDC and GPU register spaces.
 */

static void vmlfb_disable_mmio(struct vml_par *par)
{
	iounmap(par->gpu_mem);
	release_mem_region(par->gpu_mem_base, par->gpu_mem_size);
	iounmap(par->vdc_mem);
	release_mem_region(par->vdc_mem_base, par->vdc_mem_size);
}

/*
 * Release and uninit the VDC and GPU.
 */

static void vmlfb_release_devices(struct vml_par *par)
{
	if (atomic_dec_and_test(&par->refcount)) {
		pci_disable_device(par->gpu);
		pci_disable_device(par->vdc);
	}
}

/*
 * Free up allocated resources for a device.
 */

static void vml_pci_remove(struct pci_dev *dev)
{
	struct fb_info *info;
	struct vml_info *vinfo;
	struct vml_par *par;

	info = pci_get_drvdata(dev);
	if (info) {
		vinfo = container_of(info, struct vml_info, info);
		par = vinfo->par;
		mutex_lock(&vml_mutex);
		unregister_framebuffer(info);
		fb_dealloc_cmap(&info->cmap);
		vmlfb_free_vram(vinfo);
		vmlfb_disable_mmio(par);
		vmlfb_release_devices(par);
		kfree(vinfo);
		kfree(par);
		mutex_unlock(&vml_mutex);
	}
}

static void vmlfb_set_pref_pixel_format(struct fb_var_screeninfo *var)
{
	switch (var->bits_per_pixel) {
	case 16:
		var->blue.offset = 0;
		var->blue.length = 5;
		var->green.offset = 5;
		var->green.length = 5;
		var->red.offset = 10;
		var->red.length = 5;
		var->transp.offset = 15;
		var->transp.length = 1;
		break;
	case 32:
		var->blue.offset = 0;
		var->blue.length = 8;
		var->green.offset = 8;
		var->green.length = 8;
		var->red.offset = 16;
		var->red.length = 8;
		var->transp.offset = 24;
		var->transp.length = 0;
		break;
	default:
		break;
	}

	var->blue.msb_right = var->green.msb_right =
	    var->red.msb_right = var->transp.msb_right = 0;
}

/*
 * Device initialization.
 * We initialize one vml_par struct per device and one vml_info
 * struct per pipe. Currently we have only one pipe.
 */

static int vml_pci_probe(struct pci_dev *dev, const struct pci_device_id *id)
{
	struct vml_info *vinfo;
	struct fb_info *info;
	struct vml_par *par;
	int err;

	err = aperture_remove_conflicting_pci_devices(dev, "vmlfb");
	if (err)
		return err;

	par = kzalloc(sizeof(*par), GFP_KERNEL);
	if (par == NULL)
		return -ENOMEM;

	vinfo = kzalloc(sizeof(*vinfo), GFP_KERNEL);
	if (vinfo == NULL) {
		err = -ENOMEM;
		goto out_err_0;
	}

	vinfo->par = par;
	par->vdc = dev;
	atomic_set(&par->refcount, 1);

	switch (id->device) {
	case VML_DEVICE_VDC:
		if ((err = vmlfb_get_gpu(par)))
			goto out_err_1;
		pci_set_drvdata(dev, &vinfo->info);
		break;
	default:
		err = -ENODEV;
		goto out_err_1;
	}

	info = &vinfo->info;
	info->flags = FBINFO_PARTIAL_PAN_OK;

	err = vmlfb_enable_mmio(par);
	if (err)
		goto out_err_2;

	err = vmlfb_alloc_vram(vinfo, vml_mem_requested,
			       vml_mem_contig, vml_mem_min);
	if (err)
		goto out_err_3;

	strcpy(info->fix.id, "Vermilion Range");
	info->fix.mmio_start = 0;
	info->fix.mmio_len = 0;
	info->fix.smem_start = vinfo->vram_start;
	info->fix.smem_len = vinfo->vram_contig_size;
	info->fix.type = FB_TYPE_PACKED_PIXELS;
	info->fix.visual = FB_VISUAL_TRUECOLOR;
	info->fix.ypanstep = 1;
	info->fix.xpanstep = 1;
	info->fix.ywrapstep = 0;
	info->fix.accel = FB_ACCEL_NONE;
	info->screen_base = vinfo->vram_logical;
	info->pseudo_palette = vinfo->pseudo_palette;
	info->par = par;
	info->fbops = &vmlfb_ops;
	info->device = &dev->dev;

	INIT_LIST_HEAD(&vinfo->head);
	vinfo->pipe_disabled = 1;
	vinfo->cur_blank_mode = FB_BLANK_UNBLANK;

	info->var.grayscale = 0;
	info->var.bits_per_pixel = 16;
	vmlfb_set_pref_pixel_format(&info->var);

	if (!fb_find_mode
	    (&info->var, info, vml_default_mode, NULL, 0, &defaultmode, 16)) {
		printk(KERN_ERR MODULE_NAME ": Could not find initial mode\n");
	}

	if (fb_alloc_cmap(&info->cmap, 256, 1) < 0) {
		err = -ENOMEM;
		goto out_err_4;
	}

	err = register_framebuffer(info);
	if (err) {
		printk(KERN_ERR MODULE_NAME ": Register framebuffer error.\n");
		goto out_err_5;
	}

	printk("Initialized vmlfb\n");

	return 0;

out_err_5:
	fb_dealloc_cmap(&info->cmap);
out_err_4:
	vmlfb_free_vram(vinfo);
out_err_3:
	vmlfb_disable_mmio(par);
out_err_2:
	vmlfb_release_devices(par);
out_err_1:
	kfree(vinfo);
out_err_0:
	kfree(par);
	return err;
}

static int vmlfb_open(struct fb_info *info, int user)
{
	/*
	 * Save registers here?
	 */
	return 0;
}

static int vmlfb_release(struct fb_info *info, int user)
{
	/*
	 * Restore registers here.
	 */

	return 0;
}

static int vml_nearest_clock(int clock)
{

	int i;
	int cur_index;
	int cur_diff;
	int diff;

	cur_index = 0;
	cur_diff = clock - vml_clocks[0];
	cur_diff = (cur_diff < 0) ? -cur_diff : cur_diff;
	for (i = 1; i < vml_num_clocks; ++i) {
		diff = clock - vml_clocks[i];
		diff = (diff < 0) ? -diff : diff;
		if (diff < cur_diff) {
			cur_index = i;
			cur_diff = diff;
		}
	}
	return vml_clocks[cur_index];
}

static int vmlfb_check_var_locked(struct fb_var_screeninfo *var,
				  struct vml_info *vinfo)
{
	u32 pitch;
	u64 mem;
	int nearest_clock;
	int clock;
	int clock_diff;
	struct fb_var_screeninfo v;

	v = *var;
	clock = PICOS2KHZ(var->pixclock);

	if (subsys && subsys->nearest_clock) {
		nearest_clock = subsys->nearest_clock(subsys, clock);
	} else {
		nearest_clock = vml_nearest_clock(clock);
	}

	/*
	 * Accept a 20% diff.
	 */

	clock_diff = nearest_clock - clock;
	clock_diff = (clock_diff < 0) ? -clock_diff : clock_diff;
	if (clock_diff > clock / 5) {
#if 0
		printk(KERN_DEBUG MODULE_NAME ": Diff failure. %d %d\n",clock_diff,clock);
#endif
		return -EINVAL;
	}

	v.pixclock = KHZ2PICOS(nearest_clock);

	if (var->xres > VML_MAX_XRES || var->yres > VML_MAX_YRES) {
		printk(KERN_DEBUG MODULE_NAME ": Resolution failure.\n");
		return -EINVAL;
	}
	if (var->xres_virtual > VML_MAX_XRES_VIRTUAL) {
		printk(KERN_DEBUG MODULE_NAME
		       ": Virtual resolution failure.\n");
		return -EINVAL;
	}
	switch (v.bits_per_pixel) {
	case 0 ... 16:
		v.bits_per_pixel = 16;
		break;
	case 17 ... 32:
		v.bits_per_pixel = 32;
		break;
	default:
		printk(KERN_DEBUG MODULE_NAME ": Invalid bpp: %d.\n",
		       var->bits_per_pixel);
		return -EINVAL;
	}

	pitch = ALIGN((var->xres * var->bits_per_pixel) >> 3, 0x40);
	mem = (u64)pitch * var->yres_virtual;
	if (mem > vinfo->vram_contig_size) {
		return -ENOMEM;
	}

	switch (v.bits_per_pixel) {
	case 16:
		if (var->blue.offset != 0 ||
		    var->blue.length != 5 ||
		    var->green.offset != 5 ||
		    var->green.length != 5 ||
		    var->red.offset != 10 ||
		    var->red.length != 5 ||
		    var->transp.offset != 15 || var->transp.length != 1) {
			vmlfb_set_pref_pixel_format(&v);
		}
		break;
	case 32:
		if (var->blue.offset != 0 ||
		    var->blue.length != 8 ||
		    var->green.offset != 8 ||
		    var->green.length != 8 ||
		    var->red.offset != 16 ||
		    var->red.length != 8 ||
		    (var->transp.length != 0 && var->transp.length != 8) ||
		    (var->transp.length == 8 && var->transp.offset != 24)) {
			vmlfb_set_pref_pixel_format(&v);
		}
		break;
	default:
		return -EINVAL;
	}

	*var = v;

	return 0;
}

static int vmlfb_check_var(struct fb_var_screeninfo *var, struct fb_info *info)
{
	struct vml_info *vinfo = container_of(info, struct vml_info, info);
	int ret;

	mutex_lock(&vml_mutex);
	ret = vmlfb_check_var_locked(var, vinfo);
	mutex_unlock(&vml_mutex);

	return ret;
}

static void vml_wait_vblank(struct vml_info *vinfo)
{
	/* Wait for vblank. For now, just wait for a 50Hz cycle (20ms)) */
	mdelay(20);
}

static void vmlfb_disable_pipe(struct vml_info *vinfo)
{
	struct vml_par *par = vinfo->par;

	/* Disable the MDVO pad */
	VML_WRITE32(par, VML_RCOMPSTAT, 0);
	while (!(VML_READ32(par, VML_RCOMPSTAT) & VML_MDVO_VDC_I_RCOMP)) ;

	/* Disable display planes */
	VML_WRITE32(par, VML_DSPCCNTR,
		    VML_READ32(par, VML_DSPCCNTR) & ~VML_GFX_ENABLE);
	(void)VML_READ32(par, VML_DSPCCNTR);
	/* Wait for vblank for the disable to take effect */
	vml_wait_vblank(vinfo);

	/* Next, disable display pipes */
	VML_WRITE32(par, VML_PIPEACONF, 0);
	(void)VML_READ32(par, VML_PIPEACONF);

	vinfo->pipe_disabled = 1;
}

#ifdef VERMILION_DEBUG
static void vml_dump_regs(struct vml_info *vinfo)
{
	struct vml_par *par = vinfo->par;

	printk(KERN_DEBUG MODULE_NAME ": Modesetting register dump:\n");
	printk(KERN_DEBUG MODULE_NAME ": \tHTOTAL_A         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_HTOTAL_A));
	printk(KERN_DEBUG MODULE_NAME ": \tHBLANK_A         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_HBLANK_A));
	printk(KERN_DEBUG MODULE_NAME ": \tHSYNC_A          : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_HSYNC_A));
	printk(KERN_DEBUG MODULE_NAME ": \tVTOTAL_A         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_VTOTAL_A));
	printk(KERN_DEBUG MODULE_NAME ": \tVBLANK_A         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_VBLANK_A));
	printk(KERN_DEBUG MODULE_NAME ": \tVSYNC_A          : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_VSYNC_A));
	printk(KERN_DEBUG MODULE_NAME ": \tDSPCSTRIDE       : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_DSPCSTRIDE));
	printk(KERN_DEBUG MODULE_NAME ": \tDSPCSIZE         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_DSPCSIZE));
	printk(KERN_DEBUG MODULE_NAME ": \tDSPCPOS          : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_DSPCPOS));
	printk(KERN_DEBUG MODULE_NAME ": \tDSPARB           : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_DSPARB));
	printk(KERN_DEBUG MODULE_NAME ": \tDSPCADDR         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_DSPCADDR));
	printk(KERN_DEBUG MODULE_NAME ": \tBCLRPAT_A        : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_BCLRPAT_A));
	printk(KERN_DEBUG MODULE_NAME ": \tCANVSCLR_A       : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_CANVSCLR_A));
	printk(KERN_DEBUG MODULE_NAME ": \tPIPEASRC         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_PIPEASRC));
	printk(KERN_DEBUG MODULE_NAME ": \tPIPEACONF        : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_PIPEACONF));
	printk(KERN_DEBUG MODULE_NAME ": \tDSPCCNTR         : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_DSPCCNTR));
	printk(KERN_DEBUG MODULE_NAME ": \tRCOMPSTAT        : 0x%08x\n",
	       (unsigned)VML_READ32(par, VML_RCOMPSTAT));
	printk(KERN_DEBUG MODULE_NAME ": End of modesetting register dump.\n");
}
#endif

static int vmlfb_set_par_locked(struct vml_info *vinfo)
{
	struct vml_par *par = vinfo->par;
	struct fb_info *info = &vinfo->info;
	struct fb_var_screeninfo *var = &info->var;
	u32 htotal, hactive, hblank_start, hblank_end, hsync_start, hsync_end;
	u32 vtotal, vactive, vblank_start, vblank_end, vsync_start, vsync_end;
	u32 dspcntr;
	int clock;

	vinfo->bytes_per_pixel = var->bits_per_pixel >> 3;
	vinfo->stride = ALIGN(var->xres_virtual * vinfo->bytes_per_pixel, 0x40);
	info->fix.line_length = vinfo->stride;

	if (!subsys)
		return 0;

	htotal =
	    var->xres + var->right_margin + var->hsync_len + var->left_margin;
	hactive = var->xres;
	hblank_start = var->xres;
	hblank_end = htotal;
	hsync_start = hactive + var->right_margin;
	hsync_end = hsync_start + var->hsync_len;

	vtotal =
	    var->yres + var->lower_margin + var->vsync_len + var->upper_margin;
	vactive = var->yres;
	vblank_start = var->yres;
	vblank_end = vtotal;
	vsync_start = vactive + var->lower_margin;
	vsync_end = vsync_start + var->vsync_len;

	dspcntr = VML_GFX_ENABLE | VML_GFX_GAMMABYPASS;
	clock = PICOS2KHZ(var->pixclock);

	if (subsys->nearest_clock) {
		clock = subsys->nearest_clock(subsys, clock);
	} else {
		clock = vml_nearest_clock(clock);
	}
	printk(KERN_DEBUG MODULE_NAME
	       ": Set mode Hfreq : %d kHz, Vfreq : %d Hz.\n", clock / htotal,
	       ((clock / htotal) * 1000) / vtotal);

	switch (var->bits_per_pixel) {
	case 16:
		dspcntr |= VML_GFX_ARGB1555;
		break;
	case 32:
		if (var->transp.length == 8)
			dspcntr |= VML_GFX_ARGB8888 | VML_GFX_ALPHAMULT;
		else
			dspcntr |= VML_GFX_RGB0888;
		break;
	default:
		return -EINVAL;
	}

	vmlfb_disable_pipe(vinfo);
	mb();

	if (subsys->set_clock)
		subsys->set_clock(subsys, clock);
	else
		return -EINVAL;

	VML_WRITE32(par, VML_HTOTAL_A, ((htotal - 1) << 16) | (hactive - 1));
	VML_WRITE32(par, VML_HBLANK_A,
		    ((hblank_end - 1) << 16) | (hblank_start - 1));
	VML_WRITE32(par, VML_HSYNC_A,
		    ((hsync_end - 1) << 16) | (hsync_start - 1));
	VML_WRITE32(par, VML_VTOTAL_A, ((vtotal - 1) << 16) | (vactive - 1));
	VML_WRITE32(par, VML_VBLANK_A,
		    ((vblank_end - 1) << 16) | (vblank_start - 1));
	VML_WRITE32(par, VML_VSYNC_A,
		    ((vsync_end - 1) << 16) | (vsync_start - 1));
	VML_WRITE32(par, VML_DSPCSTRIDE, vinfo->stride);
	VML_WRITE32(par, VML_DSPCSIZE,
		    ((var->yres - 1) << 16) | (var->xres - 1));
	VML_WRITE32(par, VML_DSPCPOS, 0x00000000);
	VML_WRITE32(par, VML_DSPARB, VML_FIFO_DEFAULT);
	VML_WRITE32(par, VML_BCLRPAT_A, 0x00000000);
	VML_WRITE32(par, VML_CANVSCLR_A, 0x00000000);
	VML_WRITE32(par, VML_PIPEASRC,
		    ((var->xres - 1) << 16) | (var->yres - 1));

	wmb();
	VML_WRITE32(par, VML_PIPEACONF, VML_PIPE_ENABLE);
	wmb();
	VML_WRITE32(par, VML_DSPCCNTR, dspcntr);
	wmb();
	VML_WRITE32(par, VML_DSPCADDR, (u32) vinfo->vram_start +
		    var->yoffset * vinfo->stride +
		    var->xoffset * vinfo->bytes_per_pixel);

	VML_WRITE32(par, VML_RCOMPSTAT, VML_MDVO_PAD_ENABLE);

	while (!(VML_READ32(par, VML_RCOMPSTAT) &
		 (VML_MDVO_VDC_I_RCOMP | VML_MDVO_PAD_ENABLE))) ;

	vinfo->pipe_disabled = 0;
#ifdef VERMILION_DEBUG
	vml_dump_regs(vinfo);
#endif

	return 0;
}

static int vmlfb_set_par(struct fb_info *info)
{
	struct vml_info *vinfo = container_of(info, struct vml_info, info);
	int ret;

	mutex_lock(&vml_mutex);
	list_move(&vinfo->head, (subsys) ? &global_has_mode : &global_no_mode);
	ret = vmlfb_set_par_locked(vinfo);

	mutex_unlock(&vml_mutex);
	return ret;
}

static int vmlfb_blank_locked(struct vml_info *vinfo)
{
	struct vml_par *par = vinfo->par;
	u32 cur = VML_READ32(par, VML_PIPEACONF);

	switch (vinfo->cur_blank_mode) {
	case FB_BLANK_UNBLANK:
		if (vinfo->pipe_disabled) {
			vmlfb_set_par_locked(vinfo);
		}
		VML_WRITE32(par, VML_PIPEACONF, cur & ~VML_PIPE_FORCE_BORDER);
		(void)VML_READ32(par, VML_PIPEACONF);
		break;
	case FB_BLANK_NORMAL:
		if (vinfo->pipe_disabled) {
			vmlfb_set_par_locked(vinfo);
		}
		VML_WRITE32(par, VML_PIPEACONF, cur | VML_PIPE_FORCE_BORDER);
		(void)VML_READ32(par, VML_PIPEACONF);
		break;
	case FB_BLANK_VSYNC_SUSPEND:
	case FB_BLANK_HSYNC_SUSPEND:
		if (!vinfo->pipe_disabled) {
			vmlfb_disable_pipe(vinfo);
		}
		break;
	case FB_BLANK_POWERDOWN:
		if (!vinfo->pipe_disabled) {
			vmlfb_disable_pipe(vinfo);
		}
		break;
	default:
		return -EINVAL;
	}

	return 0;
}

static int vmlfb_blank(int blank_mode, struct fb_info *info)
{
	struct vml_info *vinfo = container_of(info, struct vml_info, info);
	int ret;

	mutex_lock(&vml_mutex);
	vinfo->cur_blank_mode = blank_mode;
	ret = vmlfb_blank_locked(vinfo);
	mutex_unlock(&vml_mutex);
	return ret;
}

static int vmlfb_pan_display(struct fb_var_screeninfo *var,
			     struct fb_info *info)
{
	struct vml_info *vinfo = container_of(info, struct vml_info, info);
	struct vml_par *par = vinfo->par;

	mutex_lock(&vml_mutex);
	VML_WRITE32(par, VML_DSPCADDR, (u32) vinfo->vram_start +
		    var->yoffset * vinfo->stride +
		    var->xoffset * vinfo->bytes_per_pixel);
	(void)VML_READ32(par, VML_DSPCADDR);
	mutex_unlock(&vml_mutex);

	return 0;
}

static int vmlfb_setcolreg(u_int regno, u_int red, u_int green, u_int blue,
			   u_int transp, struct fb_info *info)
{
	u32 v;

	if (regno >= 16)
		return -EINVAL;

	if (info->var.grayscale) {
		red = green = blue = (red * 77 + green * 151 + blue * 28) >> 8;
	}

	if (info->fix.visual != FB_VISUAL_TRUECOLOR)
		return -EINVAL;

	red = VML_TOHW(red, info->var.red.length);
	blue = VML_TOHW(blue, info->var.blue.length);
	green = VML_TOHW(green, info->var.green.length);
	transp = VML_TOHW(transp, info->var.transp.length);

	v = (red << info->var.red.offset) |
	    (green << info->var.green.offset) |
	    (blue << info->var.blue.offset) |
	    (transp << info->var.transp.offset);

	switch (info->var.bits_per_pixel) {
	case 16:
		((u32 *) info->pseudo_palette)[regno] = v;
		break;
	case 24:
	case 32:
		((u32 *) info->pseudo_palette)[regno] = v;
		break;
	}
	return 0;
}

static int vmlfb_mmap(struct fb_info *info, struct vm_area_struct *vma)
{
	struct vml_info *vinfo = container_of(info, struct vml_info, info);
	unsigned long offset = vma->vm_pgoff << PAGE_SHIFT;
	int ret;
	unsigned long prot;

	ret = vmlfb_vram_offset(vinfo, offset);
	if (ret)
		return -EINVAL;

	prot = pgprot_val(vma->vm_page_prot) & ~_PAGE_CACHE_MASK;
	pgprot_val(vma->vm_page_prot) =
		prot | cachemode2protval(_PAGE_CACHE_MODE_UC_MINUS);

	return vm_iomap_memory(vma, vinfo->vram_start,
			vinfo->vram_contig_size);
}

static int vmlfb_sync(struct fb_info *info)
{
	return 0;
}

static int vmlfb_cursor(struct fb_info *info, struct fb_cursor *cursor)
{
	return -EINVAL;	/* just to force soft_cursor() call */
}

static struct fb_ops vmlfb_ops = {
	.owner = THIS_MODULE,
	.fb_open = vmlfb_open,
	.fb_release = vmlfb_release,
	__FB_DEFAULT_IOMEM_OPS_RDWR,
	.fb_check_var = vmlfb_check_var,
	.fb_set_par = vmlfb_set_par,
	.fb_blank = vmlfb_blank,
	.fb_pan_display = vmlfb_pan_display,
	__FB_DEFAULT_IOMEM_OPS_DRAW,
	.fb_cursor = vmlfb_cursor,
	.fb_sync = vmlfb_sync,
	.fb_mmap = vmlfb_mmap,
	.fb_setcolreg = vmlfb_setcolreg
};

static const struct pci_device_id vml_ids[] = {
	{PCI_DEVICE(PCI_VENDOR_ID_INTEL, VML_DEVICE_VDC)},
	{0}
};

static struct pci_driver vmlfb_pci_driver = {
	.name = "vmlfb",
	.id_table = vml_ids,
	.probe = vml_pci_probe,
	.remove = vml_pci_remove,
};

static void __exit vmlfb_cleanup(void)
{
	pci_unregister_driver(&vmlfb_pci_driver);
}

static int __init vmlfb_init(void)
{

#ifndef MODULE
	char *option = NULL;
#endif

	if (fb_modesetting_disabled("vmlfb"))
		return -ENODEV;

#ifndef MODULE
	if (fb_get_options(MODULE_NAME, &option))
		return -ENODEV;
#endif

	printk(KERN_DEBUG MODULE_NAME ": initializing\n");
	mutex_init(&vml_mutex);
	INIT_LIST_HEAD(&global_no_mode);
	INIT_LIST_HEAD(&global_has_mode);

	return pci_register_driver(&vmlfb_pci_driver);
}

int vmlfb_register_subsys(struct vml_sys *sys)
{
	struct vml_info *entry;
	struct list_head *list;
	u32 save_activate;

	mutex_lock(&vml_mutex);
	if (subsys != NULL) {
		subsys->restore(subsys);
	}
	subsys = sys;
	subsys->save(subsys);

	/*
	 * We need to restart list traversal for each item, since we
	 * release the list mutex in the loop.
	 */

	list = global_no_mode.next;
	while (list != &global_no_mode) {
		list_del_init(list);
		entry = list_entry(list, struct vml_info, head);

		/*
		 * First, try the current mode which might not be
		 * completely validated with respect to the pixel clock.
		 */

		if (!vmlfb_check_var_locked(&entry->info.var, entry)) {
			vmlfb_set_par_locked(entry);
			list_add_tail(list, &global_has_mode);
		} else {

			/*
			 * Didn't work. Try to find another mode,
			 * that matches this subsys.
			 */

			mutex_unlock(&vml_mutex);
			save_activate = entry->info.var.activate;
			entry->info.var.bits_per_pixel = 16;
			vmlfb_set_pref_pixel_format(&entry->info.var);
			if (fb_find_mode(&entry->info.var,
					 &entry->info,
					 vml_default_mode, NULL, 0, NULL, 16)) {
				entry->info.var.activate |=
				    FB_ACTIVATE_FORCE | FB_ACTIVATE_NOW;
				fb_set_var(&entry->info, &entry->info.var);
			} else {
				printk(KERN_ERR MODULE_NAME
				       ": Sorry. no mode found for this subsys.\n");
			}
			entry->info.var.activate = save_activate;
			mutex_lock(&vml_mutex);
		}
		vmlfb_blank_locked(entry);
		list = global_no_mode.next;
	}
	mutex_unlock(&vml_mutex);

	printk(KERN_DEBUG MODULE_NAME ": Registered %s subsystem.\n",
				subsys->name ? subsys->name : "unknown");
	return 0;
}

EXPORT_SYMBOL_GPL(vmlfb_register_subsys);

void vmlfb_unregister_subsys(struct vml_sys *sys)
{
	struct vml_info *entry, *next;

	mutex_lock(&vml_mutex);
	if (subsys != sys) {
		mutex_unlock(&vml_mutex);
		return;
	}
	subsys->restore(subsys);
	subsys = NULL;
	list_for_each_entry_safe(entry, next, &global_has_mode, head) {
		printk(KERN_DEBUG MODULE_NAME ": subsys disable pipe\n");
		vmlfb_disable_pipe(entry);
		list_move_tail(&entry->head, &global_no_mode);
	}
	mutex_unlock(&vml_mutex);
}

EXPORT_SYMBOL_GPL(vmlfb_unregister_subsys);

module_init(vmlfb_init);
module_exit(vmlfb_cleanup);

MODULE_AUTHOR("Tungsten Graphics");
MODULE_DESCRIPTION("Initialization of the Vermilion display devices");
MODULE_VERSION("1.0.0");
MODULE_LICENSE("GPL");