1// SPDX-License-Identifier: GPL-2.0-only
2#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
3
4#include <linux/kernel.h>
5#include <linux/export.h>
6#include <media/drv-intf/saa7146_vv.h>
7
8static void calculate_output_format_register(struct saa7146_dev* saa, u32 palette, u32* clip_format)
9{
10	/* clear out the necessary bits */
11	*clip_format &= 0x0000ffff;
12	/* set these bits new */
13	*clip_format |=  (( ((palette&0xf00)>>8) << 30) | ((palette&0x00f) << 24) | (((palette&0x0f0)>>4) << 16));
14}
15
16static void calculate_hps_source_and_sync(struct saa7146_dev *dev, int source, int sync, u32* hps_ctrl)
17{
18	*hps_ctrl &= ~(MASK_30 | MASK_31 | MASK_28);
19	*hps_ctrl |= (source << 30) | (sync << 28);
20}
21
22static void calculate_hxo_and_hyo(struct saa7146_vv *vv, u32* hps_h_scale, u32* hps_ctrl)
23{
24	int hyo = 0, hxo = 0;
25
26	hyo = vv->standard->v_offset;
27	hxo = vv->standard->h_offset;
28
29	*hps_h_scale	&= ~(MASK_B0 | 0xf00);
30	*hps_h_scale	|= (hxo <<  0);
31
32	*hps_ctrl	&= ~(MASK_W0 | MASK_B2);
33	*hps_ctrl	|= (hyo << 12);
34}
35
36/* helper functions for the calculation of the horizontal- and vertical
37   scaling registers, clip-format-register etc ...
38   these functions take pointers to the (most-likely read-out
39   original-values) and manipulate them according to the requested
40   changes.
41*/
42
43/* hps_coeff used for CXY and CXUV; scale 1/1 -> scale 1/64 */
44static struct {
45	u16 hps_coeff;
46	u16 weight_sum;
47} hps_h_coeff_tab [] = {
48	{0x00,   2}, {0x02,   4}, {0x00,   4}, {0x06,   8}, {0x02,   8},
49	{0x08,   8}, {0x00,   8}, {0x1E,  16}, {0x0E,   8}, {0x26,   8},
50	{0x06,   8}, {0x42,   8}, {0x02,   8}, {0x80,   8}, {0x00,   8},
51	{0xFE,  16}, {0xFE,   8}, {0x7E,   8}, {0x7E,   8}, {0x3E,   8},
52	{0x3E,   8}, {0x1E,   8}, {0x1E,   8}, {0x0E,   8}, {0x0E,   8},
53	{0x06,   8}, {0x06,   8}, {0x02,   8}, {0x02,   8}, {0x00,   8},
54	{0x00,   8}, {0xFE,  16}, {0xFE,   8}, {0xFE,   8}, {0xFE,   8},
55	{0xFE,   8}, {0xFE,   8}, {0xFE,   8}, {0xFE,   8}, {0xFE,   8},
56	{0xFE,   8}, {0xFE,   8}, {0xFE,   8}, {0xFE,   8}, {0xFE,   8},
57	{0xFE,   8}, {0xFE,   8}, {0xFE,   8}, {0xFE,   8}, {0x7E,   8},
58	{0x7E,   8}, {0x3E,   8}, {0x3E,   8}, {0x1E,   8}, {0x1E,   8},
59	{0x0E,   8}, {0x0E,   8}, {0x06,   8}, {0x06,   8}, {0x02,   8},
60	{0x02,   8}, {0x00,   8}, {0x00,   8}, {0xFE,  16}
61};
62
63/* table of attenuation values for horizontal scaling */
64static u8 h_attenuation[] = { 1, 2, 4, 8, 2, 4, 8, 16, 0};
65
66/* calculate horizontal scale registers */
67static int calculate_h_scale_registers(struct saa7146_dev *dev,
68	int in_x, int out_x, int flip_lr,
69	u32* hps_ctrl, u32* hps_v_gain, u32* hps_h_prescale, u32* hps_h_scale)
70{
71	/* horizontal prescaler */
72	u32 dcgx = 0, xpsc = 0, xacm = 0, cxy = 0, cxuv = 0;
73	/* horizontal scaler */
74	u32 xim = 0, xp = 0, xsci =0;
75	/* vertical scale & gain */
76	u32 pfuv = 0;
77
78	/* helper variables */
79	u32 h_atten = 0, i = 0;
80
81	if ( 0 == out_x ) {
82		return -EINVAL;
83	}
84
85	/* mask out vanity-bit */
86	*hps_ctrl &= ~MASK_29;
87
88	/* calculate prescale-(xspc)-value:	[n   .. 1/2) : 1
89						[1/2 .. 1/3) : 2
90						[1/3 .. 1/4) : 3
91						...		*/
92	if (in_x > out_x) {
93		xpsc = in_x / out_x;
94	}
95	else {
96		/* zooming */
97		xpsc = 1;
98	}
99
100	/* if flip_lr-bit is set, number of pixels after
101	   horizontal prescaling must be < 384 */
102	if ( 0 != flip_lr ) {
103
104		/* set vanity bit */
105		*hps_ctrl |= MASK_29;
106
107		while (in_x / xpsc >= 384 )
108			xpsc++;
109	}
110	/* if zooming is wanted, number of pixels after
111	   horizontal prescaling must be < 768 */
112	else {
113		while ( in_x / xpsc >= 768 )
114			xpsc++;
115	}
116
117	/* maximum prescale is 64 (p.69) */
118	if ( xpsc > 64 )
119		xpsc = 64;
120
121	/* keep xacm clear*/
122	xacm = 0;
123
124	/* set horizontal filter parameters (CXY = CXUV) */
125	cxy = hps_h_coeff_tab[( (xpsc - 1) < 63 ? (xpsc - 1) : 63 )].hps_coeff;
126	cxuv = cxy;
127
128	/* calculate and set horizontal fine scale (xsci) */
129
130	/* bypass the horizontal scaler ? */
131	if ( (in_x == out_x) && ( 1 == xpsc ) )
132		xsci = 0x400;
133	else
134		xsci = ( (1024 * in_x) / (out_x * xpsc) ) + xpsc;
135
136	/* set start phase for horizontal fine scale (xp) to 0 */
137	xp = 0;
138
139	/* set xim, if we bypass the horizontal scaler */
140	if ( 0x400 == xsci )
141		xim = 1;
142	else
143		xim = 0;
144
145	/* if the prescaler is bypassed, enable horizontal
146	   accumulation mode (xacm) and clear dcgx */
147	if( 1 == xpsc ) {
148		xacm = 1;
149		dcgx = 0;
150	} else {
151		xacm = 0;
152		/* get best match in the table of attenuations
153		   for horizontal scaling */
154		h_atten = hps_h_coeff_tab[( (xpsc - 1) < 63 ? (xpsc - 1) : 63 )].weight_sum;
155
156		for (i = 0; h_attenuation[i] != 0; i++) {
157			if (h_attenuation[i] >= h_atten)
158				break;
159		}
160
161		dcgx = i;
162	}
163
164	/* the horizontal scaling increment controls the UV filter
165	   to reduce the bandwidth to improve the display quality,
166	   so set it ... */
167	if ( xsci == 0x400)
168		pfuv = 0x00;
169	else if ( xsci < 0x600)
170		pfuv = 0x01;
171	else if ( xsci < 0x680)
172		pfuv = 0x11;
173	else if ( xsci < 0x700)
174		pfuv = 0x22;
175	else
176		pfuv = 0x33;
177
178
179	*hps_v_gain  &= MASK_W0|MASK_B2;
180	*hps_v_gain  |= (pfuv << 24);
181
182	*hps_h_scale	&= ~(MASK_W1 | 0xf000);
183	*hps_h_scale	|= (xim << 31) | (xp << 24) | (xsci << 12);
184
185	*hps_h_prescale	|= (dcgx << 27) | ((xpsc-1) << 18) | (xacm << 17) | (cxy << 8) | (cxuv << 0);
186
187	return 0;
188}
189
190static struct {
191	u16 hps_coeff;
192	u16 weight_sum;
193} hps_v_coeff_tab [] = {
194 {0x0100,   2},  {0x0102,   4},  {0x0300,   4},  {0x0106,   8},  {0x0502,   8},
195 {0x0708,   8},  {0x0F00,   8},  {0x011E,  16},  {0x110E,  16},  {0x1926,  16},
196 {0x3906,  16},  {0x3D42,  16},  {0x7D02,  16},  {0x7F80,  16},  {0xFF00,  16},
197 {0x01FE,  32},  {0x01FE,  32},  {0x817E,  32},  {0x817E,  32},  {0xC13E,  32},
198 {0xC13E,  32},  {0xE11E,  32},  {0xE11E,  32},  {0xF10E,  32},  {0xF10E,  32},
199 {0xF906,  32},  {0xF906,  32},  {0xFD02,  32},  {0xFD02,  32},  {0xFF00,  32},
200 {0xFF00,  32},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},
201 {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},
202 {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},
203 {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x01FE,  64},  {0x817E,  64},
204 {0x817E,  64},  {0xC13E,  64},  {0xC13E,  64},  {0xE11E,  64},  {0xE11E,  64},
205 {0xF10E,  64},  {0xF10E,  64},  {0xF906,  64},  {0xF906,  64},  {0xFD02,  64},
206 {0xFD02,  64},  {0xFF00,  64},  {0xFF00,  64},  {0x01FE, 128}
207};
208
209/* table of attenuation values for vertical scaling */
210static u16 v_attenuation[] = { 2, 4, 8, 16, 32, 64, 128, 256, 0};
211
212/* calculate vertical scale registers */
213static int calculate_v_scale_registers(struct saa7146_dev *dev, enum v4l2_field field,
214	int in_y, int out_y, u32* hps_v_scale, u32* hps_v_gain)
215{
216	int lpi = 0;
217
218	/* vertical scaling */
219	u32 yacm = 0, ysci = 0, yacl = 0, ypo = 0, ype = 0;
220	/* vertical scale & gain */
221	u32 dcgy = 0, cya_cyb = 0;
222
223	/* helper variables */
224	u32 v_atten = 0, i = 0;
225
226	/* error, if vertical zooming */
227	if ( in_y < out_y ) {
228		return -EINVAL;
229	}
230
231	/* linear phase interpolation may be used
232	   if scaling is between 1 and 1/2 (both fields used)
233	   or scaling is between 1/2 and 1/4 (if only one field is used) */
234
235	if (V4L2_FIELD_HAS_BOTH(field)) {
236		if( 2*out_y >= in_y) {
237			lpi = 1;
238		}
239	} else if (field == V4L2_FIELD_TOP
240		|| field == V4L2_FIELD_ALTERNATE
241		|| field == V4L2_FIELD_BOTTOM) {
242		if( 4*out_y >= in_y ) {
243			lpi = 1;
244		}
245		out_y *= 2;
246	}
247	if( 0 != lpi ) {
248
249		yacm = 0;
250		yacl = 0;
251		cya_cyb = 0x00ff;
252
253		/* calculate scaling increment */
254		if ( in_y > out_y )
255			ysci = ((1024 * in_y) / (out_y + 1)) - 1024;
256		else
257			ysci = 0;
258
259		dcgy = 0;
260
261		/* calculate ype and ypo */
262		ype = ysci / 16;
263		ypo = ype + (ysci / 64);
264
265	} else {
266		yacm = 1;
267
268		/* calculate scaling increment */
269		ysci = (((10 * 1024 * (in_y - out_y - 1)) / in_y) + 9) / 10;
270
271		/* calculate ype and ypo */
272		ypo = ype = ((ysci + 15) / 16);
273
274		/* the sequence length interval (yacl) has to be set according
275		   to the prescale value, e.g.	[n   .. 1/2) : 0
276						[1/2 .. 1/3) : 1
277						[1/3 .. 1/4) : 2
278						... */
279		if ( ysci < 512) {
280			yacl = 0;
281		} else {
282			yacl = ( ysci / (1024 - ysci) );
283		}
284
285		/* get filter coefficients for cya, cyb from table hps_v_coeff_tab */
286		cya_cyb = hps_v_coeff_tab[ (yacl < 63 ? yacl : 63 ) ].hps_coeff;
287
288		/* get best match in the table of attenuations for vertical scaling */
289		v_atten = hps_v_coeff_tab[ (yacl < 63 ? yacl : 63 ) ].weight_sum;
290
291		for (i = 0; v_attenuation[i] != 0; i++) {
292			if (v_attenuation[i] >= v_atten)
293				break;
294		}
295
296		dcgy = i;
297	}
298
299	/* ypo and ype swapped in spec ? */
300	*hps_v_scale	|= (yacm << 31) | (ysci << 21) | (yacl << 15) | (ypo << 8 ) | (ype << 1);
301
302	*hps_v_gain	&= ~(MASK_W0|MASK_B2);
303	*hps_v_gain	|= (dcgy << 16) | (cya_cyb << 0);
304
305	return 0;
306}
307
308/* simple bubble-sort algorithm with duplicate elimination */
309static int sort_and_eliminate(u32* values, int* count)
310{
311	int low = 0, high = 0, top = 0;
312	int cur = 0, next = 0;
313
314	/* sanity checks */
315	if( (0 > *count) || (NULL == values) ) {
316		return -EINVAL;
317	}
318
319	/* bubble sort the first @count items of the array @values */
320	for( top = *count; top > 0; top--) {
321		for( low = 0, high = 1; high < top; low++, high++) {
322			if( values[low] > values[high] )
323				swap(values[low], values[high]);
324		}
325	}
326
327	/* remove duplicate items */
328	for( cur = 0, next = 1; next < *count; next++) {
329		if( values[cur] != values[next])
330			values[++cur] = values[next];
331	}
332
333	*count = cur + 1;
334
335	return 0;
336}
337
338static void calculate_clipping_registers_rect(struct saa7146_dev *dev, struct saa7146_fh *fh,
339	struct saa7146_video_dma *vdma2, u32* clip_format, u32* arbtr_ctrl, enum v4l2_field field)
340{
341	struct saa7146_vv *vv = dev->vv_data;
342	__le32 *clipping = vv->d_clipping.cpu_addr;
343
344	int width = vv->ov.win.w.width;
345	int height =  vv->ov.win.w.height;
346	int clipcount = vv->ov.nclips;
347
348	u32 line_list[32];
349	u32 pixel_list[32];
350	int numdwords = 0;
351
352	int i = 0, j = 0;
353	int cnt_line = 0, cnt_pixel = 0;
354
355	int x[32], y[32], w[32], h[32];
356
357	/* clear out memory */
358	memset(&line_list[0],  0x00, sizeof(u32)*32);
359	memset(&pixel_list[0], 0x00, sizeof(u32)*32);
360	memset(clipping,  0x00, SAA7146_CLIPPING_MEM);
361
362	/* fill the line and pixel-lists */
363	for(i = 0; i < clipcount; i++) {
364		int l = 0, r = 0, t = 0, b = 0;
365
366		x[i] = vv->ov.clips[i].c.left;
367		y[i] = vv->ov.clips[i].c.top;
368		w[i] = vv->ov.clips[i].c.width;
369		h[i] = vv->ov.clips[i].c.height;
370
371		if( w[i] < 0) {
372			x[i] += w[i]; w[i] = -w[i];
373		}
374		if( h[i] < 0) {
375			y[i] += h[i]; h[i] = -h[i];
376		}
377		if( x[i] < 0) {
378			w[i] += x[i]; x[i] = 0;
379		}
380		if( y[i] < 0) {
381			h[i] += y[i]; y[i] = 0;
382		}
383		if( 0 != vv->vflip ) {
384			y[i] = height - y[i] - h[i];
385		}
386
387		l = x[i];
388		r = x[i]+w[i];
389		t = y[i];
390		b = y[i]+h[i];
391
392		/* insert left/right coordinates */
393		pixel_list[ 2*i   ] = min_t(int, l, width);
394		pixel_list[(2*i)+1] = min_t(int, r, width);
395		/* insert top/bottom coordinates */
396		line_list[ 2*i   ] = min_t(int, t, height);
397		line_list[(2*i)+1] = min_t(int, b, height);
398	}
399
400	/* sort and eliminate lists */
401	cnt_line = cnt_pixel = 2*clipcount;
402	sort_and_eliminate( &pixel_list[0], &cnt_pixel );
403	sort_and_eliminate( &line_list[0], &cnt_line );
404
405	/* calculate the number of used u32s */
406	numdwords = max_t(int, (cnt_line+1), (cnt_pixel+1))*2;
407	numdwords = max_t(int, 4, numdwords);
408	numdwords = min_t(int, 64, numdwords);
409
410	/* fill up cliptable */
411	for(i = 0; i < cnt_pixel; i++) {
412		clipping[2*i] |= cpu_to_le32(pixel_list[i] << 16);
413	}
414	for(i = 0; i < cnt_line; i++) {
415		clipping[(2*i)+1] |= cpu_to_le32(line_list[i] << 16);
416	}
417
418	/* fill up cliptable with the display infos */
419	for(j = 0; j < clipcount; j++) {
420
421		for(i = 0; i < cnt_pixel; i++) {
422
423			if( x[j] < 0)
424				x[j] = 0;
425
426			if( pixel_list[i] < (x[j] + w[j])) {
427
428				if ( pixel_list[i] >= x[j] ) {
429					clipping[2*i] |= cpu_to_le32(1 << j);
430				}
431			}
432		}
433		for(i = 0; i < cnt_line; i++) {
434
435			if( y[j] < 0)
436				y[j] = 0;
437
438			if( line_list[i] < (y[j] + h[j]) ) {
439
440				if( line_list[i] >= y[j] ) {
441					clipping[(2*i)+1] |= cpu_to_le32(1 << j);
442				}
443			}
444		}
445	}
446
447	/* adjust arbitration control register */
448	*arbtr_ctrl &= 0xffff00ff;
449	*arbtr_ctrl |= 0x00001c00;
450
451	vdma2->base_even	= vv->d_clipping.dma_handle;
452	vdma2->base_odd		= vv->d_clipping.dma_handle;
453	vdma2->prot_addr	= vv->d_clipping.dma_handle+((sizeof(u32))*(numdwords));
454	vdma2->base_page	= 0x04;
455	vdma2->pitch		= 0x00;
456	vdma2->num_line_byte	= (0 << 16 | (sizeof(u32))*(numdwords-1) );
457
458	/* set clipping-mode. this depends on the field(s) used */
459	*clip_format &= 0xfffffff7;
460	if (V4L2_FIELD_HAS_BOTH(field)) {
461		*clip_format |= 0x00000008;
462	} else {
463		*clip_format |= 0x00000000;
464	}
465}
466
467/* disable clipping */
468static void saa7146_disable_clipping(struct saa7146_dev *dev)
469{
470	u32 clip_format	= saa7146_read(dev, CLIP_FORMAT_CTRL);
471
472	/* mask out relevant bits (=lower word)*/
473	clip_format &= MASK_W1;
474
475	/* upload clipping-registers*/
476	saa7146_write(dev, CLIP_FORMAT_CTRL,clip_format);
477	saa7146_write(dev, MC2, (MASK_05 | MASK_21));
478
479	/* disable video dma2 */
480	saa7146_write(dev, MC1, MASK_21);
481}
482
483static void saa7146_set_clipping_rect(struct saa7146_fh *fh)
484{
485	struct saa7146_dev *dev = fh->dev;
486	struct saa7146_vv *vv = dev->vv_data;
487	enum v4l2_field field = vv->ov.win.field;
488	struct	saa7146_video_dma vdma2;
489	u32 clip_format;
490	u32 arbtr_ctrl;
491
492	/* check clipcount, disable clipping if clipcount == 0*/
493	if (vv->ov.nclips == 0) {
494		saa7146_disable_clipping(dev);
495		return;
496	}
497
498	clip_format = saa7146_read(dev, CLIP_FORMAT_CTRL);
499	arbtr_ctrl = saa7146_read(dev, PCI_BT_V1);
500
501	calculate_clipping_registers_rect(dev, fh, &vdma2, &clip_format, &arbtr_ctrl, field);
502
503	/* set clipping format */
504	clip_format &= 0xffff0008;
505	clip_format |= (SAA7146_CLIPPING_RECT << 4);
506
507	/* prepare video dma2 */
508	saa7146_write(dev, BASE_EVEN2,		vdma2.base_even);
509	saa7146_write(dev, BASE_ODD2,		vdma2.base_odd);
510	saa7146_write(dev, PROT_ADDR2,		vdma2.prot_addr);
511	saa7146_write(dev, BASE_PAGE2,		vdma2.base_page);
512	saa7146_write(dev, PITCH2,		vdma2.pitch);
513	saa7146_write(dev, NUM_LINE_BYTE2,	vdma2.num_line_byte);
514
515	/* prepare the rest */
516	saa7146_write(dev, CLIP_FORMAT_CTRL,clip_format);
517	saa7146_write(dev, PCI_BT_V1, arbtr_ctrl);
518
519	/* upload clip_control-register, clipping-registers, enable video dma2 */
520	saa7146_write(dev, MC2, (MASK_05 | MASK_21 | MASK_03 | MASK_19));
521	saa7146_write(dev, MC1, (MASK_05 | MASK_21));
522}
523
524static void saa7146_set_window(struct saa7146_dev *dev, int width, int height, enum v4l2_field field)
525{
526	struct saa7146_vv *vv = dev->vv_data;
527
528	int source = vv->current_hps_source;
529	int sync = vv->current_hps_sync;
530
531	u32 hps_v_scale = 0, hps_v_gain  = 0, hps_ctrl = 0, hps_h_prescale = 0, hps_h_scale = 0;
532
533	/* set vertical scale */
534	hps_v_scale = 0; /* all bits get set by the function-call */
535	hps_v_gain  = 0; /* fixme: saa7146_read(dev, HPS_V_GAIN);*/
536	calculate_v_scale_registers(dev, field, vv->standard->v_field*2, height, &hps_v_scale, &hps_v_gain);
537
538	/* set horizontal scale */
539	hps_ctrl	= 0;
540	hps_h_prescale	= 0; /* all bits get set in the function */
541	hps_h_scale	= 0;
542	calculate_h_scale_registers(dev, vv->standard->h_pixels, width, vv->hflip, &hps_ctrl, &hps_v_gain, &hps_h_prescale, &hps_h_scale);
543
544	/* set hyo and hxo */
545	calculate_hxo_and_hyo(vv, &hps_h_scale, &hps_ctrl);
546	calculate_hps_source_and_sync(dev, source, sync, &hps_ctrl);
547
548	/* write out new register contents */
549	saa7146_write(dev, HPS_V_SCALE,	hps_v_scale);
550	saa7146_write(dev, HPS_V_GAIN,	hps_v_gain);
551	saa7146_write(dev, HPS_CTRL,	hps_ctrl);
552	saa7146_write(dev, HPS_H_PRESCALE,hps_h_prescale);
553	saa7146_write(dev, HPS_H_SCALE,	hps_h_scale);
554
555	/* upload shadow-ram registers */
556	saa7146_write(dev, MC2, (MASK_05 | MASK_06 | MASK_21 | MASK_22) );
557}
558
559/* calculate the new memory offsets for a desired position */
560static void saa7146_set_position(struct saa7146_dev *dev, int w_x, int w_y, int w_height, enum v4l2_field field, u32 pixelformat)
561{
562	struct saa7146_vv *vv = dev->vv_data;
563	struct saa7146_format *sfmt = saa7146_format_by_fourcc(dev, pixelformat);
564
565	int b_depth = vv->ov_fmt->depth;
566	int b_bpl = vv->ov_fb.fmt.bytesperline;
567	/* The unsigned long cast is to remove a 64-bit compile warning since
568	   it looks like a 64-bit address is cast to a 32-bit value, even
569	   though the base pointer is really a 32-bit physical address that
570	   goes into a 32-bit DMA register.
571	   FIXME: might not work on some 64-bit platforms, but see the FIXME
572	   in struct v4l2_framebuffer (videodev2.h) for that.
573	 */
574	u32 base = (u32)(unsigned long)vv->ov_fb.base;
575
576	struct	saa7146_video_dma vdma1;
577
578	/* calculate memory offsets for picture, look if we shall top-down-flip */
579	vdma1.pitch	= 2*b_bpl;
580	if ( 0 == vv->vflip ) {
581		vdma1.base_even = base + (w_y * (vdma1.pitch/2)) + (w_x * (b_depth / 8));
582		vdma1.base_odd  = vdma1.base_even + (vdma1.pitch / 2);
583		vdma1.prot_addr = vdma1.base_even + (w_height * (vdma1.pitch / 2));
584	}
585	else {
586		vdma1.base_even = base + ((w_y+w_height) * (vdma1.pitch/2)) + (w_x * (b_depth / 8));
587		vdma1.base_odd  = vdma1.base_even - (vdma1.pitch / 2);
588		vdma1.prot_addr = vdma1.base_odd - (w_height * (vdma1.pitch / 2));
589	}
590
591	if (V4L2_FIELD_HAS_BOTH(field)) {
592	} else if (field == V4L2_FIELD_ALTERNATE) {
593		/* fixme */
594		vdma1.base_odd = vdma1.prot_addr;
595		vdma1.pitch /= 2;
596	} else if (field == V4L2_FIELD_TOP) {
597		vdma1.base_odd = vdma1.prot_addr;
598		vdma1.pitch /= 2;
599	} else if (field == V4L2_FIELD_BOTTOM) {
600		vdma1.base_odd = vdma1.base_even;
601		vdma1.base_even = vdma1.prot_addr;
602		vdma1.pitch /= 2;
603	}
604
605	if ( 0 != vv->vflip ) {
606		vdma1.pitch *= -1;
607	}
608
609	vdma1.base_page = sfmt->swap;
610	vdma1.num_line_byte = (vv->standard->v_field<<16)+vv->standard->h_pixels;
611
612	saa7146_write_out_dma(dev, 1, &vdma1);
613}
614
615static void saa7146_set_output_format(struct saa7146_dev *dev, unsigned long palette)
616{
617	u32 clip_format = saa7146_read(dev, CLIP_FORMAT_CTRL);
618
619	/* call helper function */
620	calculate_output_format_register(dev,palette,&clip_format);
621
622	/* update the hps registers */
623	saa7146_write(dev, CLIP_FORMAT_CTRL, clip_format);
624	saa7146_write(dev, MC2, (MASK_05 | MASK_21));
625}
626
627/* select input-source */
628void saa7146_set_hps_source_and_sync(struct saa7146_dev *dev, int source, int sync)
629{
630	struct saa7146_vv *vv = dev->vv_data;
631	u32 hps_ctrl = 0;
632
633	/* read old state */
634	hps_ctrl = saa7146_read(dev, HPS_CTRL);
635
636	hps_ctrl &= ~( MASK_31 | MASK_30 | MASK_28 );
637	hps_ctrl |= (source << 30) | (sync << 28);
638
639	/* write back & upload register */
640	saa7146_write(dev, HPS_CTRL, hps_ctrl);
641	saa7146_write(dev, MC2, (MASK_05 | MASK_21));
642
643	vv->current_hps_source = source;
644	vv->current_hps_sync = sync;
645}
646EXPORT_SYMBOL_GPL(saa7146_set_hps_source_and_sync);
647
648int saa7146_enable_overlay(struct saa7146_fh *fh)
649{
650	struct saa7146_dev *dev = fh->dev;
651	struct saa7146_vv *vv = dev->vv_data;
652
653	saa7146_set_window(dev, vv->ov.win.w.width, vv->ov.win.w.height, vv->ov.win.field);
654	saa7146_set_position(dev, vv->ov.win.w.left, vv->ov.win.w.top, vv->ov.win.w.height, vv->ov.win.field, vv->ov_fmt->pixelformat);
655	saa7146_set_output_format(dev, vv->ov_fmt->trans);
656	saa7146_set_clipping_rect(fh);
657
658	/* enable video dma1 */
659	saa7146_write(dev, MC1, (MASK_06 | MASK_22));
660	return 0;
661}
662
663void saa7146_disable_overlay(struct saa7146_fh *fh)
664{
665	struct saa7146_dev *dev = fh->dev;
666
667	/* disable clipping + video dma1 */
668	saa7146_disable_clipping(dev);
669	saa7146_write(dev, MC1, MASK_22);
670}
671
672void saa7146_write_out_dma(struct saa7146_dev* dev, int which, struct saa7146_video_dma* vdma)
673{
674	int where = 0;
675
676	if( which < 1 || which > 3) {
677		return;
678	}
679
680	/* calculate starting address */
681	where  = (which-1)*0x18;
682
683	saa7146_write(dev, where,	vdma->base_odd);
684	saa7146_write(dev, where+0x04,	vdma->base_even);
685	saa7146_write(dev, where+0x08,	vdma->prot_addr);
686	saa7146_write(dev, where+0x0c,	vdma->pitch);
687	saa7146_write(dev, where+0x10,	vdma->base_page);
688	saa7146_write(dev, where+0x14,	vdma->num_line_byte);
689
690	/* upload */
691	saa7146_write(dev, MC2, (MASK_02<<(which-1))|(MASK_18<<(which-1)));
692/*
693	printk("vdma%d.base_even:     0x%08x\n", which,vdma->base_even);
694	printk("vdma%d.base_odd:      0x%08x\n", which,vdma->base_odd);
695	printk("vdma%d.prot_addr:     0x%08x\n", which,vdma->prot_addr);
696	printk("vdma%d.base_page:     0x%08x\n", which,vdma->base_page);
697	printk("vdma%d.pitch:         0x%08x\n", which,vdma->pitch);
698	printk("vdma%d.num_line_byte: 0x%08x\n", which,vdma->num_line_byte);
699*/
700}
701
702static int calculate_video_dma_grab_packed(struct saa7146_dev* dev, struct saa7146_buf *buf)
703{
704	struct saa7146_vv *vv = dev->vv_data;
705	struct saa7146_video_dma vdma1;
706
707	struct saa7146_format *sfmt = saa7146_format_by_fourcc(dev,buf->fmt->pixelformat);
708
709	int width = buf->fmt->width;
710	int height = buf->fmt->height;
711	int bytesperline = buf->fmt->bytesperline;
712	enum v4l2_field field = buf->fmt->field;
713
714	int depth = sfmt->depth;
715
716	DEB_CAP("[size=%dx%d,fields=%s]\n",
717		width, height, v4l2_field_names[field]);
718
719	if( bytesperline != 0) {
720		vdma1.pitch = bytesperline*2;
721	} else {
722		vdma1.pitch = (width*depth*2)/8;
723	}
724	vdma1.num_line_byte	= ((vv->standard->v_field<<16) + vv->standard->h_pixels);
725	vdma1.base_page		= buf->pt[0].dma | ME1 | sfmt->swap;
726
727	if( 0 != vv->vflip ) {
728		vdma1.prot_addr	= buf->pt[0].offset;
729		vdma1.base_even	= buf->pt[0].offset+(vdma1.pitch/2)*height;
730		vdma1.base_odd	= vdma1.base_even - (vdma1.pitch/2);
731	} else {
732		vdma1.base_even	= buf->pt[0].offset;
733		vdma1.base_odd	= vdma1.base_even + (vdma1.pitch/2);
734		vdma1.prot_addr	= buf->pt[0].offset+(vdma1.pitch/2)*height;
735	}
736
737	if (V4L2_FIELD_HAS_BOTH(field)) {
738	} else if (field == V4L2_FIELD_ALTERNATE) {
739		/* fixme */
740		if ( vv->last_field == V4L2_FIELD_TOP ) {
741			vdma1.base_odd	= vdma1.prot_addr;
742			vdma1.pitch /= 2;
743		} else if ( vv->last_field == V4L2_FIELD_BOTTOM ) {
744			vdma1.base_odd	= vdma1.base_even;
745			vdma1.base_even = vdma1.prot_addr;
746			vdma1.pitch /= 2;
747		}
748	} else if (field == V4L2_FIELD_TOP) {
749		vdma1.base_odd	= vdma1.prot_addr;
750		vdma1.pitch /= 2;
751	} else if (field == V4L2_FIELD_BOTTOM) {
752		vdma1.base_odd	= vdma1.base_even;
753		vdma1.base_even = vdma1.prot_addr;
754		vdma1.pitch /= 2;
755	}
756
757	if( 0 != vv->vflip ) {
758		vdma1.pitch *= -1;
759	}
760
761	saa7146_write_out_dma(dev, 1, &vdma1);
762	return 0;
763}
764
765static int calc_planar_422(struct saa7146_vv *vv, struct saa7146_buf *buf, struct saa7146_video_dma *vdma2, struct saa7146_video_dma *vdma3)
766{
767	int height = buf->fmt->height;
768	int width = buf->fmt->width;
769
770	vdma2->pitch	= width;
771	vdma3->pitch	= width;
772
773	/* fixme: look at bytesperline! */
774
775	if( 0 != vv->vflip ) {
776		vdma2->prot_addr	= buf->pt[1].offset;
777		vdma2->base_even	= ((vdma2->pitch/2)*height)+buf->pt[1].offset;
778		vdma2->base_odd		= vdma2->base_even - (vdma2->pitch/2);
779
780		vdma3->prot_addr	= buf->pt[2].offset;
781		vdma3->base_even	= ((vdma3->pitch/2)*height)+buf->pt[2].offset;
782		vdma3->base_odd		= vdma3->base_even - (vdma3->pitch/2);
783	} else {
784		vdma3->base_even	= buf->pt[2].offset;
785		vdma3->base_odd		= vdma3->base_even + (vdma3->pitch/2);
786		vdma3->prot_addr	= (vdma3->pitch/2)*height+buf->pt[2].offset;
787
788		vdma2->base_even	= buf->pt[1].offset;
789		vdma2->base_odd		= vdma2->base_even + (vdma2->pitch/2);
790		vdma2->prot_addr	= (vdma2->pitch/2)*height+buf->pt[1].offset;
791	}
792
793	return 0;
794}
795
796static int calc_planar_420(struct saa7146_vv *vv, struct saa7146_buf *buf, struct saa7146_video_dma *vdma2, struct saa7146_video_dma *vdma3)
797{
798	int height = buf->fmt->height;
799	int width = buf->fmt->width;
800
801	vdma2->pitch	= width/2;
802	vdma3->pitch	= width/2;
803
804	if( 0 != vv->vflip ) {
805		vdma2->prot_addr	= buf->pt[2].offset;
806		vdma2->base_even	= ((vdma2->pitch/2)*height)+buf->pt[2].offset;
807		vdma2->base_odd		= vdma2->base_even - (vdma2->pitch/2);
808
809		vdma3->prot_addr	= buf->pt[1].offset;
810		vdma3->base_even	= ((vdma3->pitch/2)*height)+buf->pt[1].offset;
811		vdma3->base_odd		= vdma3->base_even - (vdma3->pitch/2);
812
813	} else {
814		vdma3->base_even	= buf->pt[2].offset;
815		vdma3->base_odd		= vdma3->base_even + (vdma3->pitch);
816		vdma3->prot_addr	= (vdma3->pitch/2)*height+buf->pt[2].offset;
817
818		vdma2->base_even	= buf->pt[1].offset;
819		vdma2->base_odd		= vdma2->base_even + (vdma2->pitch);
820		vdma2->prot_addr	= (vdma2->pitch/2)*height+buf->pt[1].offset;
821	}
822	return 0;
823}
824
825static int calculate_video_dma_grab_planar(struct saa7146_dev* dev, struct saa7146_buf *buf)
826{
827	struct saa7146_vv *vv = dev->vv_data;
828	struct saa7146_video_dma vdma1;
829	struct saa7146_video_dma vdma2;
830	struct saa7146_video_dma vdma3;
831
832	struct saa7146_format *sfmt = saa7146_format_by_fourcc(dev,buf->fmt->pixelformat);
833
834	int width = buf->fmt->width;
835	int height = buf->fmt->height;
836	enum v4l2_field field = buf->fmt->field;
837
838	BUG_ON(0 == buf->pt[0].dma);
839	BUG_ON(0 == buf->pt[1].dma);
840	BUG_ON(0 == buf->pt[2].dma);
841
842	DEB_CAP("[size=%dx%d,fields=%s]\n",
843		width, height, v4l2_field_names[field]);
844
845	/* fixme: look at bytesperline! */
846
847	/* fixme: what happens for user space buffers here?. The offsets are
848	   most likely wrong, this version here only works for page-aligned
849	   buffers, modifications to the pagetable-functions are necessary...*/
850
851	vdma1.pitch		= width*2;
852	vdma1.num_line_byte	= ((vv->standard->v_field<<16) + vv->standard->h_pixels);
853	vdma1.base_page		= buf->pt[0].dma | ME1;
854
855	if( 0 != vv->vflip ) {
856		vdma1.prot_addr	= buf->pt[0].offset;
857		vdma1.base_even	= ((vdma1.pitch/2)*height)+buf->pt[0].offset;
858		vdma1.base_odd	= vdma1.base_even - (vdma1.pitch/2);
859	} else {
860		vdma1.base_even	= buf->pt[0].offset;
861		vdma1.base_odd	= vdma1.base_even + (vdma1.pitch/2);
862		vdma1.prot_addr	= (vdma1.pitch/2)*height+buf->pt[0].offset;
863	}
864
865	vdma2.num_line_byte	= 0; /* unused */
866	vdma2.base_page		= buf->pt[1].dma | ME1;
867
868	vdma3.num_line_byte	= 0; /* unused */
869	vdma3.base_page		= buf->pt[2].dma | ME1;
870
871	switch( sfmt->depth ) {
872		case 12: {
873			calc_planar_420(vv,buf,&vdma2,&vdma3);
874			break;
875		}
876		case 16: {
877			calc_planar_422(vv,buf,&vdma2,&vdma3);
878			break;
879		}
880		default: {
881			return -1;
882		}
883	}
884
885	if (V4L2_FIELD_HAS_BOTH(field)) {
886	} else if (field == V4L2_FIELD_ALTERNATE) {
887		/* fixme */
888		vdma1.base_odd	= vdma1.prot_addr;
889		vdma1.pitch /= 2;
890		vdma2.base_odd	= vdma2.prot_addr;
891		vdma2.pitch /= 2;
892		vdma3.base_odd	= vdma3.prot_addr;
893		vdma3.pitch /= 2;
894	} else if (field == V4L2_FIELD_TOP) {
895		vdma1.base_odd	= vdma1.prot_addr;
896		vdma1.pitch /= 2;
897		vdma2.base_odd	= vdma2.prot_addr;
898		vdma2.pitch /= 2;
899		vdma3.base_odd	= vdma3.prot_addr;
900		vdma3.pitch /= 2;
901	} else if (field == V4L2_FIELD_BOTTOM) {
902		vdma1.base_odd	= vdma1.base_even;
903		vdma1.base_even = vdma1.prot_addr;
904		vdma1.pitch /= 2;
905		vdma2.base_odd	= vdma2.base_even;
906		vdma2.base_even = vdma2.prot_addr;
907		vdma2.pitch /= 2;
908		vdma3.base_odd	= vdma3.base_even;
909		vdma3.base_even = vdma3.prot_addr;
910		vdma3.pitch /= 2;
911	}
912
913	if( 0 != vv->vflip ) {
914		vdma1.pitch *= -1;
915		vdma2.pitch *= -1;
916		vdma3.pitch *= -1;
917	}
918
919	saa7146_write_out_dma(dev, 1, &vdma1);
920	if( (sfmt->flags & FORMAT_BYTE_SWAP) != 0 ) {
921		saa7146_write_out_dma(dev, 3, &vdma2);
922		saa7146_write_out_dma(dev, 2, &vdma3);
923	} else {
924		saa7146_write_out_dma(dev, 2, &vdma2);
925		saa7146_write_out_dma(dev, 3, &vdma3);
926	}
927	return 0;
928}
929
930static void program_capture_engine(struct saa7146_dev *dev, int planar)
931{
932	struct saa7146_vv *vv = dev->vv_data;
933	int count = 0;
934
935	unsigned long e_wait = vv->current_hps_sync == SAA7146_HPS_SYNC_PORT_A ? CMD_E_FID_A : CMD_E_FID_B;
936	unsigned long o_wait = vv->current_hps_sync == SAA7146_HPS_SYNC_PORT_A ? CMD_O_FID_A : CMD_O_FID_B;
937
938	/* wait for o_fid_a/b / e_fid_a/b toggle only if rps register 0 is not set*/
939	WRITE_RPS0(CMD_PAUSE | CMD_OAN | CMD_SIG0 | o_wait);
940	WRITE_RPS0(CMD_PAUSE | CMD_OAN | CMD_SIG0 | e_wait);
941
942	/* set rps register 0 */
943	WRITE_RPS0(CMD_WR_REG | (1 << 8) | (MC2/4));
944	WRITE_RPS0(MASK_27 | MASK_11);
945
946	/* turn on video-dma1 */
947	WRITE_RPS0(CMD_WR_REG_MASK | (MC1/4));
948	WRITE_RPS0(MASK_06 | MASK_22);			/* => mask */
949	WRITE_RPS0(MASK_06 | MASK_22);			/* => values */
950	if( 0 != planar ) {
951		/* turn on video-dma2 */
952		WRITE_RPS0(CMD_WR_REG_MASK | (MC1/4));
953		WRITE_RPS0(MASK_05 | MASK_21);			/* => mask */
954		WRITE_RPS0(MASK_05 | MASK_21);			/* => values */
955
956		/* turn on video-dma3 */
957		WRITE_RPS0(CMD_WR_REG_MASK | (MC1/4));
958		WRITE_RPS0(MASK_04 | MASK_20);			/* => mask */
959		WRITE_RPS0(MASK_04 | MASK_20);			/* => values */
960	}
961
962	/* wait for o_fid_a/b / e_fid_a/b toggle */
963	if ( vv->last_field == V4L2_FIELD_INTERLACED ) {
964		WRITE_RPS0(CMD_PAUSE | o_wait);
965		WRITE_RPS0(CMD_PAUSE | e_wait);
966	} else if ( vv->last_field == V4L2_FIELD_TOP ) {
967		WRITE_RPS0(CMD_PAUSE | (vv->current_hps_sync == SAA7146_HPS_SYNC_PORT_A ? MASK_10 : MASK_09));
968		WRITE_RPS0(CMD_PAUSE | o_wait);
969	} else if ( vv->last_field == V4L2_FIELD_BOTTOM ) {
970		WRITE_RPS0(CMD_PAUSE | (vv->current_hps_sync == SAA7146_HPS_SYNC_PORT_A ? MASK_10 : MASK_09));
971		WRITE_RPS0(CMD_PAUSE | e_wait);
972	}
973
974	/* turn off video-dma1 */
975	WRITE_RPS0(CMD_WR_REG_MASK | (MC1/4));
976	WRITE_RPS0(MASK_22 | MASK_06);			/* => mask */
977	WRITE_RPS0(MASK_22);				/* => values */
978	if( 0 != planar ) {
979		/* turn off video-dma2 */
980		WRITE_RPS0(CMD_WR_REG_MASK | (MC1/4));
981		WRITE_RPS0(MASK_05 | MASK_21);			/* => mask */
982		WRITE_RPS0(MASK_21);				/* => values */
983
984		/* turn off video-dma3 */
985		WRITE_RPS0(CMD_WR_REG_MASK | (MC1/4));
986		WRITE_RPS0(MASK_04 | MASK_20);			/* => mask */
987		WRITE_RPS0(MASK_20);				/* => values */
988	}
989
990	/* generate interrupt */
991	WRITE_RPS0(CMD_INTERRUPT);
992
993	/* stop */
994	WRITE_RPS0(CMD_STOP);
995}
996
997void saa7146_set_capture(struct saa7146_dev *dev, struct saa7146_buf *buf, struct saa7146_buf *next)
998{
999	struct saa7146_format *sfmt = saa7146_format_by_fourcc(dev,buf->fmt->pixelformat);
1000	struct saa7146_vv *vv = dev->vv_data;
1001	u32 vdma1_prot_addr;
1002
1003	DEB_CAP("buf:%p, next:%p\n", buf, next);
1004
1005	vdma1_prot_addr = saa7146_read(dev, PROT_ADDR1);
1006	if( 0 == vdma1_prot_addr ) {
1007		/* clear out beginning of streaming bit (rps register 0)*/
1008		DEB_CAP("forcing sync to new frame\n");
1009		saa7146_write(dev, MC2, MASK_27 );
1010	}
1011
1012	saa7146_set_window(dev, buf->fmt->width, buf->fmt->height, buf->fmt->field);
1013	saa7146_set_output_format(dev, sfmt->trans);
1014	saa7146_disable_clipping(dev);
1015
1016	if ( vv->last_field == V4L2_FIELD_INTERLACED ) {
1017	} else if ( vv->last_field == V4L2_FIELD_TOP ) {
1018		vv->last_field = V4L2_FIELD_BOTTOM;
1019	} else if ( vv->last_field == V4L2_FIELD_BOTTOM ) {
1020		vv->last_field = V4L2_FIELD_TOP;
1021	}
1022
1023	if( 0 != IS_PLANAR(sfmt->trans)) {
1024		calculate_video_dma_grab_planar(dev, buf);
1025		program_capture_engine(dev,1);
1026	} else {
1027		calculate_video_dma_grab_packed(dev, buf);
1028		program_capture_engine(dev,0);
1029	}
1030
1031/*
1032	printk("vdma%d.base_even:     0x%08x\n", 1,saa7146_read(dev,BASE_EVEN1));
1033	printk("vdma%d.base_odd:      0x%08x\n", 1,saa7146_read(dev,BASE_ODD1));
1034	printk("vdma%d.prot_addr:     0x%08x\n", 1,saa7146_read(dev,PROT_ADDR1));
1035	printk("vdma%d.base_page:     0x%08x\n", 1,saa7146_read(dev,BASE_PAGE1));
1036	printk("vdma%d.pitch:         0x%08x\n", 1,saa7146_read(dev,PITCH1));
1037	printk("vdma%d.num_line_byte: 0x%08x\n", 1,saa7146_read(dev,NUM_LINE_BYTE1));
1038	printk("vdma%d => vptr      : 0x%08x\n", 1,saa7146_read(dev,PCI_VDP1));
1039*/
1040
1041	/* write the address of the rps-program */
1042	saa7146_write(dev, RPS_ADDR0, dev->d_rps0.dma_handle);
1043
1044	/* turn on rps */
1045	saa7146_write(dev, MC1, (MASK_12 | MASK_28));
1046}
1047