[linux-2.6-block.git] / drivers / video / cfbimgblt.c

/*
 *  Generic BitBLT function for frame buffer with packed pixels of any depth.
 *
 *      Copyright (C)  June 1999 James Simmons
 *
 *  This file is subject to the terms and conditions of the GNU General Public
 *  License.  See the file COPYING in the main directory of this archive for
 *  more details.
 *
 * NOTES:
 *
 *    This function copys a image from system memory to video memory. The
 *  image can be a bitmap where each 0 represents the background color and
 *  each 1 represents the foreground color. Great for font handling. It can
 *  also be a color image. This is determined by image_depth. The color image
 *  must be laid out exactly in the same format as the framebuffer. Yes I know
 *  their are cards with hardware that coverts images of various depths to the
 *  framebuffer depth. But not every card has this. All images must be rounded
 *  up to the nearest byte. For example a bitmap 12 bits wide must be two 
 *  bytes width. 
 *
 *  Tony: 
 *  Incorporate mask tables similar to fbcon-cfb*.c in 2.4 API.  This speeds 
 *  up the code significantly.
 *  
 *  Code for depths not multiples of BITS_PER_LONG is still kludgy, which is
 *  still processed a bit at a time.   
 *
 *  Also need to add code to deal with cards endians that are different than
 *  the native cpu endians. I also need to deal with MSB position in the word.
 */
#include <linux/module.h>
#include <linux/string.h>
#include <linux/fb.h>
#include <asm/types.h>
#include "fb_draw.h"

#define DEBUG

#ifdef DEBUG
#define DPRINTK(fmt, args...) printk(KERN_DEBUG "%s: " fmt,__FUNCTION__,## args)
#else
#define DPRINTK(fmt, args...)
#endif

static const u32 cfb_tab8[] = {
#if defined(__BIG_ENDIAN)
    0x00000000,0x000000ff,0x0000ff00,0x0000ffff,
    0x00ff0000,0x00ff00ff,0x00ffff00,0x00ffffff,
    0xff000000,0xff0000ff,0xff00ff00,0xff00ffff,
    0xffff0000,0xffff00ff,0xffffff00,0xffffffff
#elif defined(__LITTLE_ENDIAN)
    0x00000000,0xff000000,0x00ff0000,0xffff0000,
    0x0000ff00,0xff00ff00,0x00ffff00,0xffffff00,
    0x000000ff,0xff0000ff,0x00ff00ff,0xffff00ff,
    0x0000ffff,0xff00ffff,0x00ffffff,0xffffffff
#else
#error FIXME: No endianness??
#endif
};

static const u32 cfb_tab16[] = {
#if defined(__BIG_ENDIAN)
    0x00000000, 0x0000ffff, 0xffff0000, 0xffffffff
#elif defined(__LITTLE_ENDIAN)
    0x00000000, 0xffff0000, 0x0000ffff, 0xffffffff
#else
#error FIXME: No endianness??
#endif
};

static const u32 cfb_tab32[] = {
	0x00000000, 0xffffffff
};

#define FB_WRITEL fb_writel
#define FB_READL  fb_readl

static inline void color_imageblit(const struct fb_image *image, 
				   struct fb_info *p, u8 __iomem *dst1, 
				   u32 start_index,
				   u32 pitch_index)
{
	/* Draw the penguin */
	u32 __iomem *dst, *dst2;
	u32 color = 0, val, shift;
	int i, n, bpp = p->var.bits_per_pixel;
	u32 null_bits = 32 - bpp;
	u32 *palette = (u32 *) p->pseudo_palette;
	const u8 *src = image->data;
	u32 bswapmask = fb_compute_bswapmask(p);

	dst2 = (u32 __iomem *) dst1;
	for (i = image->height; i--; ) {
		n = image->width;
		dst = (u32 __iomem *) dst1;
		shift = 0;
		val = 0;
		
		if (start_index) {
			u32 start_mask = ~fb_shifted_pixels_mask_u32(start_index, bswapmask);
			val = FB_READL(dst) & start_mask;
			shift = start_index;
		}
		while (n--) {
			if (p->fix.visual == FB_VISUAL_TRUECOLOR ||
			    p->fix.visual == FB_VISUAL_DIRECTCOLOR )
				color = palette[*src];
			else
				color = *src;
			color <<= FB_LEFT_POS(bpp);
			val |= FB_SHIFT_HIGH(color, shift ^ bswapmask);
			if (shift >= null_bits) {
				FB_WRITEL(val, dst++);
	
				val = (shift == null_bits) ? 0 : 
					FB_SHIFT_LOW(color, 32 - shift);
			}
			shift += bpp;
			shift &= (32 - 1);
			src++;
		}
		if (shift) {
			u32 end_mask = fb_shifted_pixels_mask_u32(shift, bswapmask);

			FB_WRITEL((FB_READL(dst) & end_mask) | val, dst);
		}
		dst1 += p->fix.line_length;
		if (pitch_index) {
			dst2 += p->fix.line_length;
			dst1 = (u8 __iomem *)((long __force)dst2 & ~(sizeof(u32) - 1));

			start_index += pitch_index;
			start_index &= 32 - 1;
		}
	}
}

static inline void slow_imageblit(const struct fb_image *image, struct fb_info *p, 
				  u8 __iomem *dst1, u32 fgcolor,
				  u32 bgcolor, 
				  u32 start_index,
				  u32 pitch_index)
{
	u32 shift, color = 0, bpp = p->var.bits_per_pixel;
	u32 __iomem *dst, *dst2;
	u32 val, pitch = p->fix.line_length;
	u32 null_bits = 32 - bpp;
	u32 spitch = (image->width+7)/8;
	const u8 *src = image->data, *s;
	u32 i, j, l;
	u32 bswapmask = fb_compute_bswapmask(p);

	dst2 = (u32 __iomem *) dst1;
	fgcolor <<= FB_LEFT_POS(bpp);
	bgcolor <<= FB_LEFT_POS(bpp);

	for (i = image->height; i--; ) {
		shift = val = 0;
		l = 8;
		j = image->width;
		dst = (u32 __iomem *) dst1;
		s = src;

		/* write leading bits */
		if (start_index) {
			u32 start_mask = ~fb_shifted_pixels_mask_u32(start_index, bswapmask);
			val = FB_READL(dst) & start_mask;
			shift = start_index;
		}

		while (j--) {
			l--;
			color = (*s & (1 << l)) ? fgcolor : bgcolor;
			val |= FB_SHIFT_HIGH(color, shift ^ bswapmask);
			
			/* Did the bitshift spill bits to the next long? */
			if (shift >= null_bits) {
				FB_WRITEL(val, dst++);
				val = (shift == null_bits) ? 0 :
					FB_SHIFT_LOW(color,32 - shift);
			}
			shift += bpp;
			shift &= (32 - 1);
			if (!l) { l = 8; s++; };
		}

		/* write trailing bits */
 		if (shift) {
			u32 end_mask = fb_shifted_pixels_mask_u32(shift, bswapmask);

			FB_WRITEL((FB_READL(dst) & end_mask) | val, dst);
		}
		
		dst1 += pitch;
		src += spitch;	
		if (pitch_index) {
			dst2 += pitch;
			dst1 = (u8 __iomem *)((long __force)dst2 & ~(sizeof(u32) - 1));
			start_index += pitch_index;
			start_index &= 32 - 1;
		}
		
	}
}

/*
 * fast_imageblit - optimized monochrome color expansion
 *
 * Only if:  bits_per_pixel == 8, 16, or 32
 *           image->width is divisible by pixel/dword (ppw);
 *           fix->line_legth is divisible by 4;
 *           beginning and end of a scanline is dword aligned
 */
static inline void fast_imageblit(const struct fb_image *image, struct fb_info *p, 
				  u8 __iomem *dst1, u32 fgcolor, 
				  u32 bgcolor) 
{
	u32 fgx = fgcolor, bgx = bgcolor, bpp = p->var.bits_per_pixel;
	u32 ppw = 32/bpp, spitch = (image->width + 7)/8;
	u32 bit_mask, end_mask, eorx, shift;
	const char *s = image->data, *src;
	u32 __iomem *dst;
	const u32 *tab = NULL;
	int i, j, k;
		
	switch (bpp) {
	case 8:
		tab = cfb_tab8;
		break;
	case 16:
		tab = cfb_tab16;
		break;
	case 32:
	default:
		tab = cfb_tab32;
		break;
	}

	for (i = ppw-1; i--; ) {
		fgx <<= bpp;
		bgx <<= bpp;
		fgx |= fgcolor;
		bgx |= bgcolor;
	}
	
	bit_mask = (1 << ppw) - 1;
	eorx = fgx ^ bgx;
	k = image->width/ppw;

	for (i = image->height; i--; ) {
		dst = (u32 __iomem *) dst1, shift = 8; src = s;
		
		for (j = k; j--; ) {
			shift -= ppw;
			end_mask = tab[(*src >> shift) & bit_mask];
			FB_WRITEL((end_mask & eorx)^bgx, dst++);
			if (!shift) { shift = 8; src++; }		
		}
		dst1 += p->fix.line_length;
		s += spitch;
	}
}	
	
void cfb_imageblit(struct fb_info *p, const struct fb_image *image)
{
	u32 fgcolor, bgcolor, start_index, bitstart, pitch_index = 0;
	u32 bpl = sizeof(u32), bpp = p->var.bits_per_pixel;
	u32 width = image->width;
	u32 dx = image->dx, dy = image->dy;
	u8 __iomem *dst1;

	if (p->state != FBINFO_STATE_RUNNING)
		return;

	bitstart = (dy * p->fix.line_length * 8) + (dx * bpp);
	start_index = bitstart & (32 - 1);
	pitch_index = (p->fix.line_length & (bpl - 1)) * 8;

	bitstart /= 8;
	bitstart &= ~(bpl - 1);
	dst1 = p->screen_base + bitstart;

	if (p->fbops->fb_sync)
		p->fbops->fb_sync(p);

	if (image->depth == 1) {
		if (p->fix.visual == FB_VISUAL_TRUECOLOR ||
		    p->fix.visual == FB_VISUAL_DIRECTCOLOR) {
			fgcolor = ((u32*)(p->pseudo_palette))[image->fg_color];
			bgcolor = ((u32*)(p->pseudo_palette))[image->bg_color];
		} else {
			fgcolor = image->fg_color;
			bgcolor = image->bg_color;
		}	
		
		if (32 % bpp == 0 && !start_index && !pitch_index && 
		    ((width & (32/bpp-1)) == 0) &&
		    bpp >= 8 && bpp <= 32) 			
			fast_imageblit(image, p, dst1, fgcolor, bgcolor);
		else 
			slow_imageblit(image, p, dst1, fgcolor, bgcolor,
					start_index, pitch_index);
	} else
		color_imageblit(image, p, dst1, start_index, pitch_index);
}

EXPORT_SYMBOL(cfb_imageblit);

MODULE_AUTHOR("James Simmons <jsimmons@users.sf.net>");
MODULE_DESCRIPTION("Generic software accelerated imaging drawing");
MODULE_LICENSE("GPL");
Commit	Line	Data
1da177e4 LT	1	/*
	2	* Generic BitBLT function for frame buffer with packed pixels of any depth.
	3	*
	4	* Copyright (C) June 1999 James Simmons
	5	*
	6	* This file is subject to the terms and conditions of the GNU General Public
	7	* License. See the file COPYING in the main directory of this archive for
	8	* more details.
	9	*
	10	* NOTES:
	11	*
	12	* This function copys a image from system memory to video memory. The
	13	* image can be a bitmap where each 0 represents the background color and
	14	* each 1 represents the foreground color. Great for font handling. It can
	15	* also be a color image. This is determined by image_depth. The color image
	16	* must be laid out exactly in the same format as the framebuffer. Yes I know
	17	* their are cards with hardware that coverts images of various depths to the
	18	* framebuffer depth. But not every card has this. All images must be rounded
	19	* up to the nearest byte. For example a bitmap 12 bits wide must be two
	20	* bytes width.
	21	*
	22	* Tony:
	23	* Incorporate mask tables similar to fbcon-cfb*.c in 2.4 API. This speeds
	24	* up the code significantly.
	25	*
	26	* Code for depths not multiples of BITS_PER_LONG is still kludgy, which is
	27	* still processed a bit at a time.
	28	*
	29	* Also need to add code to deal with cards endians that are different than
	30	* the native cpu endians. I also need to deal with MSB position in the word.
	31	*/
1da177e4 LT	32	#include <linux/module.h>
	33	#include <linux/string.h>
	34	#include <linux/fb.h>
	35	#include <asm/types.h>
779121e9	36	#include "fb_draw.h"
1da177e4 LT	37
	38	#define DEBUG
	39
	40	#ifdef DEBUG
	41	#define DPRINTK(fmt, args...) printk(KERN_DEBUG "%s: " fmt,__FUNCTION__,## args)
	42	#else
	43	#define DPRINTK(fmt, args...)
	44	#endif
	45
d95159cf	46	static const u32 cfb_tab8[] = {
1da177e4 LT	47	#if defined(__BIG_ENDIAN)
	48	0x00000000,0x000000ff,0x0000ff00,0x0000ffff,
	49	0x00ff0000,0x00ff00ff,0x00ffff00,0x00ffffff,
	50	0xff000000,0xff0000ff,0xff00ff00,0xff00ffff,
	51	0xffff0000,0xffff00ff,0xffffff00,0xffffffff
	52	#elif defined(__LITTLE_ENDIAN)
	53	0x00000000,0xff000000,0x00ff0000,0xffff0000,
	54	0x0000ff00,0xff00ff00,0x00ffff00,0xffffff00,
	55	0x000000ff,0xff0000ff,0x00ff00ff,0xffff00ff,
	56	0x0000ffff,0xff00ffff,0x00ffffff,0xffffffff
	57	#else
	58	#error FIXME: No endianness??
	59	#endif
	60	};
	61
d95159cf	62	static const u32 cfb_tab16[] = {
1da177e4 LT	63	#if defined(__BIG_ENDIAN)
	64	0x00000000, 0x0000ffff, 0xffff0000, 0xffffffff
	65	#elif defined(__LITTLE_ENDIAN)
	66	0x00000000, 0xffff0000, 0x0000ffff, 0xffffffff
	67	#else
	68	#error FIXME: No endianness??
	69	#endif
	70	};
	71
d95159cf	72	static const u32 cfb_tab32[] = {
1da177e4 LT	73	0x00000000, 0xffffffff
	74	};
	75
	76	#define FB_WRITEL fb_writel
	77	#define FB_READL fb_readl
	78
1da177e4 LT	79	static inline void color_imageblit(const struct fb_image *image,
	80	struct fb_info p, u8 __iomem dst1,
	81	u32 start_index,
	82	u32 pitch_index)
	83	{
	84	/* Draw the penguin */
	85	u32 __iomem dst, dst2;
	86	u32 color = 0, val, shift;
	87	int i, n, bpp = p->var.bits_per_pixel;
	88	u32 null_bits = 32 - bpp;
	89	u32 palette = (u32 ) p->pseudo_palette;
	90	const u8 *src = image->data;
779121e9	91	u32 bswapmask = fb_compute_bswapmask(p);
1da177e4 LT	92
	93	dst2 = (u32 __iomem *) dst1;
	94	for (i = image->height; i--; ) {
	95	n = image->width;
	96	dst = (u32 __iomem *) dst1;
	97	shift = 0;
	98	val = 0;
	99
	100	if (start_index) {
779121e9	101	u32 start_mask = ~fb_shifted_pixels_mask_u32(start_index, bswapmask);
1da177e4 LT	102	val = FB_READL(dst) & start_mask;
	103	shift = start_index;
	104	}
	105	while (n--) {
	106	if (p->fix.visual == FB_VISUAL_TRUECOLOR \|\|
	107	p->fix.visual == FB_VISUAL_DIRECTCOLOR )
	108	color = palette[*src];
	109	else
	110	color = *src;
be0d9b6c	111	color <<= FB_LEFT_POS(bpp);
779121e9	112	val \|= FB_SHIFT_HIGH(color, shift ^ bswapmask);
1da177e4 LT	113	if (shift >= null_bits) {
	114	FB_WRITEL(val, dst++);
	115
	116	val = (shift == null_bits) ? 0 :
be0d9b6c	117	FB_SHIFT_LOW(color, 32 - shift);
1da177e4 LT	118	}
	119	shift += bpp;
	120	shift &= (32 - 1);
	121	src++;
	122	}
	123	if (shift) {
779121e9	124	u32 end_mask = fb_shifted_pixels_mask_u32(shift, bswapmask);
1da177e4 LT	125
	126	FB_WRITEL((FB_READL(dst) & end_mask) \| val, dst);
	127	}
	128	dst1 += p->fix.line_length;
	129	if (pitch_index) {
	130	dst2 += p->fix.line_length;
	131	dst1 = (u8 __iomem *)((long __force)dst2 & ~(sizeof(u32) - 1));
	132
	133	start_index += pitch_index;
	134	start_index &= 32 - 1;
	135	}
	136	}
	137	}
	138
	139	static inline void slow_imageblit(const struct fb_image image, struct fb_info p,
	140	u8 __iomem *dst1, u32 fgcolor,
	141	u32 bgcolor,
	142	u32 start_index,
	143	u32 pitch_index)
	144	{
	145	u32 shift, color = 0, bpp = p->var.bits_per_pixel;
	146	u32 __iomem dst, dst2;
	147	u32 val, pitch = p->fix.line_length;
	148	u32 null_bits = 32 - bpp;
	149	u32 spitch = (image->width+7)/8;
	150	const u8 src = image->data, s;
	151	u32 i, j, l;
779121e9 PP	152	u32 bswapmask = fb_compute_bswapmask(p);
779121e9 PP	153
1da177e4	154	dst2 = (u32 __iomem *) dst1;
be0d9b6c AD	155	fgcolor <<= FB_LEFT_POS(bpp);
be0d9b6c AD	156	bgcolor <<= FB_LEFT_POS(bpp);
1da177e4 LT	157
	158	for (i = image->height; i--; ) {
	159	shift = val = 0;
	160	l = 8;
	161	j = image->width;
	162	dst = (u32 __iomem *) dst1;
	163	s = src;
	164
	165	/* write leading bits */
	166	if (start_index) {
779121e9	167	u32 start_mask = ~fb_shifted_pixels_mask_u32(start_index, bswapmask);
1da177e4 LT	168	val = FB_READL(dst) & start_mask;
	169	shift = start_index;
	170	}
	171
	172	while (j--) {
	173	l--;
a536093a	174	color = (*s & (1 << l)) ? fgcolor : bgcolor;
779121e9	175	val \|= FB_SHIFT_HIGH(color, shift ^ bswapmask);
1da177e4 LT	176
	177	/* Did the bitshift spill bits to the next long? */
	178	if (shift >= null_bits) {
	179	FB_WRITEL(val, dst++);
	180	val = (shift == null_bits) ? 0 :
be0d9b6c	181	FB_SHIFT_LOW(color,32 - shift);
1da177e4 LT	182	}
	183	shift += bpp;
	184	shift &= (32 - 1);
	185	if (!l) { l = 8; s++; };
	186	}
	187
	188	/* write trailing bits */
	189	if (shift) {
779121e9	190	u32 end_mask = fb_shifted_pixels_mask_u32(shift, bswapmask);
1da177e4 LT	191
	192	FB_WRITEL((FB_READL(dst) & end_mask) \| val, dst);
	193	}
	194
	195	dst1 += pitch;
	196	src += spitch;
	197	if (pitch_index) {
	198	dst2 += pitch;
	199	dst1 = (u8 __iomem *)((long __force)dst2 & ~(sizeof(u32) - 1));
	200	start_index += pitch_index;
	201	start_index &= 32 - 1;
	202	}
	203
	204	}
	205	}
	206
	207	/*
	208	* fast_imageblit - optimized monochrome color expansion
	209	*
	210	* Only if: bits_per_pixel == 8, 16, or 32
	211	* image->width is divisible by pixel/dword (ppw);
	212	* fix->line_legth is divisible by 4;
	213	* beginning and end of a scanline is dword aligned
	214	*/
	215	static inline void fast_imageblit(const struct fb_image image, struct fb_info p,
	216	u8 __iomem *dst1, u32 fgcolor,
	217	u32 bgcolor)
	218	{
	219	u32 fgx = fgcolor, bgx = bgcolor, bpp = p->var.bits_per_pixel;
	220	u32 ppw = 32/bpp, spitch = (image->width + 7)/8;
	221	u32 bit_mask, end_mask, eorx, shift;
	222	const char s = image->data, src;
	223	u32 __iomem *dst;
d95159cf	224	const u32 *tab = NULL;
1da177e4 LT	225	int i, j, k;
	226
	227	switch (bpp) {
	228	case 8:
	229	tab = cfb_tab8;
	230	break;
	231	case 16:
	232	tab = cfb_tab16;
	233	break;
	234	case 32:
8126a044	235	default:
1da177e4 LT	236	tab = cfb_tab32;
	237	break;
	238	}
	239
	240	for (i = ppw-1; i--; ) {
	241	fgx <<= bpp;
	242	bgx <<= bpp;
	243	fgx \|= fgcolor;
	244	bgx \|= bgcolor;
	245	}
	246
	247	bit_mask = (1 << ppw) - 1;
	248	eorx = fgx ^ bgx;
	249	k = image->width/ppw;
	250
	251	for (i = image->height; i--; ) {
	252	dst = (u32 __iomem *) dst1, shift = 8; src = s;
	253
	254	for (j = k; j--; ) {
	255	shift -= ppw;
	256	end_mask = tab[(*src >> shift) & bit_mask];
	257	FB_WRITEL((end_mask & eorx)^bgx, dst++);
	258	if (!shift) { shift = 8; src++; }
	259	}
	260	dst1 += p->fix.line_length;
	261	s += spitch;
	262	}
	263	}
	264
	265	void cfb_imageblit(struct fb_info p, const struct fb_image image)
	266	{
	267	u32 fgcolor, bgcolor, start_index, bitstart, pitch_index = 0;
	268	u32 bpl = sizeof(u32), bpp = p->var.bits_per_pixel;
b4d8aea6	269	u32 width = image->width;
1da177e4	270	u32 dx = image->dx, dy = image->dy;
1da177e4 LT	271	u8 __iomem *dst1;
	272
	273	if (p->state != FBINFO_STATE_RUNNING)
	274	return;
	275
1da177e4 LT	276	bitstart = (dy * p->fix.line_length * 8) + (dx * bpp);
	277	start_index = bitstart & (32 - 1);
	278	pitch_index = (p->fix.line_length & (bpl - 1)) * 8;
	279
	280	bitstart /= 8;
	281	bitstart &= ~(bpl - 1);
	282	dst1 = p->screen_base + bitstart;
	283
	284	if (p->fbops->fb_sync)
	285	p->fbops->fb_sync(p);
	286
	287	if (image->depth == 1) {
	288	if (p->fix.visual == FB_VISUAL_TRUECOLOR \|\|
	289	p->fix.visual == FB_VISUAL_DIRECTCOLOR) {
	290	fgcolor = ((u32*)(p->pseudo_palette))[image->fg_color];
	291	bgcolor = ((u32*)(p->pseudo_palette))[image->bg_color];
	292	} else {
	293	fgcolor = image->fg_color;
	294	bgcolor = image->bg_color;
	295	}
	296
	297	if (32 % bpp == 0 && !start_index && !pitch_index &&
	298	((width & (32/bpp-1)) == 0) &&
	299	bpp >= 8 && bpp <= 32)
	300	fast_imageblit(image, p, dst1, fgcolor, bgcolor);
	301	else
	302	slow_imageblit(image, p, dst1, fgcolor, bgcolor,
	303	start_index, pitch_index);
	304	} else
	305	color_imageblit(image, p, dst1, start_index, pitch_index);
	306	}
	307
	308	EXPORT_SYMBOL(cfb_imageblit);
	309
	310	MODULE_AUTHOR("James Simmons <jsimmons@users.sf.net>");
	311	MODULE_DESCRIPTION("Generic software accelerated imaging drawing");
	312	MODULE_LICENSE("GPL");
	313