* [PATCH/v2] Atari frame buffer fixes
@ 2009-01-04 11:46 Geert Uytterhoeven
2009-01-04 11:46 ` [PATCH 1/7] fbdev: atafb - Fix line length handling Geert Uytterhoeven
0 siblings, 1 reply; 8+ messages in thread
From: Geert Uytterhoeven @ 2009-01-04 11:46 UTC (permalink / raw)
To: Michael Schmitz, linux-fbdev-devel; +Cc: linux-m68k, linux-kernel
This patchsets contains several fixes for the Atari frame buffer device, and
includes a refactoring of the c2p (Chunky-to-Planar) code:
[1] fbdev: atafb - Fix line length handling
[2] fbdev: atafb - Fix 16 bpp console
[3] fbdev: c2p - Correct indentation
[4] fbdev: c2p - Cleanups
[5] fbdev: c2p - Extract common c2p core to c2p_core.h
[6] fbdev: c2p/atafb - Add support for Atari interleaved bitplanes
[7] fbdev: c2p - Rename c2p to c2p_planar
Changes since previous submission:
- Convert macros to static inline functions
- Make sure we get a build-time error (undefined reference to
'c2p_unsupported') in case of future misuse
- Replace `unsigned long' by `u32' in comp(), as that's what all callers use
- Use {get,put}_unaligned_be32() in store_{iplan2,planar}{,_masked}()
If nobody objects, I'd like to push this to Linus via the m68k tree, as the
c2p code is used by m68k drivers (atafb and amifb) only.
drivers/video/Makefile | 4 +-
drivers/video/amifb.c | 6 +-
drivers/video/atafb.c | 101 ++++++++-----------
drivers/video/c2p.c | 232 --------------------------------------------
drivers/video/c2p.h | 11 ++-
drivers/video/c2p_core.h | 153 +++++++++++++++++++++++++++++
drivers/video/c2p_iplan2.c | 153 +++++++++++++++++++++++++++++
drivers/video/c2p_planar.c | 156 +++++++++++++++++++++++++++++
8 files changed, 515 insertions(+), 301 deletions(-)
^ permalink raw reply [flat|nested] 8+ messages in thread* [PATCH 1/7] fbdev: atafb - Fix line length handling 2009-01-04 11:46 [PATCH/v2] Atari frame buffer fixes Geert Uytterhoeven @ 2009-01-04 11:46 ` Geert Uytterhoeven 2009-01-04 11:46 ` [PATCH 2/7] fbdev: atafb - Fix 16 bpp console Geert Uytterhoeven 0 siblings, 1 reply; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:46 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven - Make sure par->next_line is always set (this was done for Falcon only), as all the text console drawing operations need a valid par->next_line, - Make sure fix->line_length is always set, as some userspace applications need it because they don't have fallback code for the case where it's zero. Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/atafb.c | 12 ++++++++---- 1 files changed, 8 insertions(+), 4 deletions(-) diff --git a/drivers/video/atafb.c b/drivers/video/atafb.c index 77eb8b3..b7646ad 100644 --- a/drivers/video/atafb.c +++ b/drivers/video/atafb.c @@ -614,7 +614,7 @@ static int tt_encode_fix(struct fb_fix_screeninfo *fix, struct atafb_par *par) fix->xpanstep = 0; fix->ypanstep = 1; fix->ywrapstep = 0; - fix->line_length = 0; + fix->line_length = par->next_line; fix->accel = FB_ACCEL_ATARIBLITT; return 0; } @@ -691,6 +691,7 @@ static int tt_decode_var(struct fb_var_screeninfo *var, struct atafb_par *par) return -EINVAL; par->yres_virtual = yres_virtual; par->screen_base = screen_base + var->yoffset * linelen; + par->next_line = linelen; return 0; } @@ -918,7 +919,7 @@ static int falcon_encode_fix(struct fb_fix_screeninfo *fix, fix->visual = FB_VISUAL_TRUECOLOR; fix->xpanstep = 2; } - fix->line_length = 0; + fix->line_length = par->next_line; fix->accel = FB_ACCEL_ATARIBLITT; return 0; } @@ -1852,7 +1853,7 @@ static int stste_encode_fix(struct fb_fix_screeninfo *fix, fix->ypanstep = 0; } fix->ywrapstep = 0; - fix->line_length = 0; + fix->line_length = par->next_line; fix->accel = FB_ACCEL_ATARIBLITT; return 0; } @@ -1910,6 +1911,7 @@ static int stste_decode_var(struct fb_var_screeninfo *var, return -EINVAL; par->yres_virtual = yres_virtual; par->screen_base = screen_base + var->yoffset * linelen; + par->next_line = linelen; return 0; } @@ -2169,7 +2171,7 @@ static int ext_encode_fix(struct fb_fix_screeninfo *fix, struct atafb_par *par) fix->xpanstep = 0; fix->ypanstep = 0; fix->ywrapstep = 0; - fix->line_length = 0; + fix->line_length = par->next_line; return 0; } @@ -2184,6 +2186,8 @@ static int ext_decode_var(struct fb_var_screeninfo *var, struct atafb_par *par) var->xoffset > 0 || var->yoffset > 0) return -EINVAL; + + par->next_line = external_xres_virtual * external_depth / 8; return 0; } -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 2/7] fbdev: atafb - Fix 16 bpp console 2009-01-04 11:46 ` [PATCH 1/7] fbdev: atafb - Fix line length handling Geert Uytterhoeven @ 2009-01-04 11:46 ` Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 3/7] fbdev: c2p - Correct indentation Geert Uytterhoeven 0 siblings, 1 reply; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:46 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven - 16 bpp must use the cfb_*() ops - 16 bpp needs to set up info->pseudo_palette[] (was fbcon_cfb16_cmap[] in 2.4.x) - Kill commented out 2.4.x fbcon remnants Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/atafb.c | 74 ++++++++++++++++++++----------------------------- 1 files changed, 30 insertions(+), 44 deletions(-) diff --git a/drivers/video/atafb.c b/drivers/video/atafb.c index b7646ad..f932c13 100644 --- a/drivers/video/atafb.c +++ b/drivers/video/atafb.c @@ -149,6 +149,7 @@ static struct atafb_par { short mono; short ste_mode; short bpp; + u32 pseudo_palette[16]; } falcon; #endif /* Nothing needed for external mode */ @@ -885,10 +886,6 @@ static int vdl_prescale[4][3] = { /* Default hsync timing [mon_type] in picoseconds */ static long h_syncs[4] = { 3000000, 4875000, 4000000, 4875000 }; -#ifdef FBCON_HAS_CFB16 -static u16 fbcon_cfb16_cmap[16]; -#endif - static inline int hxx_prescale(struct falcon_hw *hw) { return hw->ste_mode ? 16 @@ -1736,10 +1733,10 @@ static int falcon_setcolreg(unsigned int regno, unsigned int red, (((red & 0xe000) >> 13) | ((red & 0x1000) >> 12) << 8) | (((green & 0xe000) >> 13) | ((green & 0x1000) >> 12) << 4) | ((blue & 0xe000) >> 13) | ((blue & 0x1000) >> 12); -#ifdef FBCON_HAS_CFB16 - fbcon_cfb16_cmap[regno] = ((red & 0xf800) | - ((green & 0xfc00) >> 5) | - ((blue & 0xf800) >> 11)); +#ifdef ATAFB_FALCON + ((u32 *)info->pseudo_palette)[regno] = ((red & 0xf800) | + ((green & 0xfc00) >> 5) | + ((blue & 0xf800) >> 11)); #endif } return 0; @@ -2447,42 +2444,6 @@ static void atafb_set_disp(struct fb_info *info) atafb_get_fix(&info->fix, info); info->screen_base = (void *)info->fix.smem_start; - - switch (info->fix.type) { - case FB_TYPE_INTERLEAVED_PLANES: - switch (info->var.bits_per_pixel) { - case 2: - // display->dispsw = &fbcon_iplan2p2; - break; - case 4: - // display->dispsw = &fbcon_iplan2p4; - break; - case 8: - // display->dispsw = &fbcon_iplan2p8; - break; - } - break; - case FB_TYPE_PACKED_PIXELS: - switch (info->var.bits_per_pixel) { -#ifdef FBCON_HAS_MFB - case 1: - // display->dispsw = &fbcon_mfb; - break; -#endif -#ifdef FBCON_HAS_CFB8 - case 8: - // display->dispsw = &fbcon_cfb8; - break; -#endif -#ifdef FBCON_HAS_CFB16 - case 16: - // display->dispsw = &fbcon_cfb16; - // display->dispsw_data = fbcon_cfb16_cmap; - break; -#endif - } - break; - } } static int atafb_setcolreg(u_int regno, u_int red, u_int green, u_int blue, @@ -2553,6 +2514,13 @@ static void atafb_fillrect(struct fb_info *info, const struct fb_fillrect *rect) if (!rect->width || !rect->height) return; +#ifdef ATAFB_FALCON + if (info->var.bits_per_pixel == 16) { + cfb_fillrect(info, rect); + return; + } +#endif + /* * We could use hardware clipping but on many cards you get around * hardware clipping by writing to framebuffer directly. @@ -2587,6 +2555,13 @@ static void atafb_copyarea(struct fb_info *info, const struct fb_copyarea *area) u32 dx, dy, sx, sy, width, height; int rev_copy = 0; +#ifdef ATAFB_FALCON + if (info->var.bits_per_pixel == 16) { + cfb_copyarea(info, area); + return; + } +#endif + /* clip the destination */ x2 = area->dx + area->width; y2 = area->dy + area->height; @@ -2636,6 +2611,13 @@ static void atafb_imageblit(struct fb_info *info, const struct fb_image *image) const char *src; u32 dx, dy, width, height, pitch; +#ifdef ATAFB_FALCON + if (info->var.bits_per_pixel == 16) { + cfb_imageblit(info, image); + return; + } +#endif + /* * We could use hardware clipping but on many cards you get around * hardware clipping by writing to framebuffer directly like we are @@ -3229,6 +3211,10 @@ int __init atafb_init(void) // tries to read from HW which may not be initialized yet // so set sane var first, then call atafb_set_par atafb_get_var(&fb_info.var, &fb_info); + +#ifdef ATAFB_FALCON + fb_info.pseudo_palette = current_par.hw.falcon.pseudo_palette; +#endif fb_info.flags = FBINFO_FLAG_DEFAULT; if (!fb_find_mode(&fb_info.var, &fb_info, mode_option, atafb_modedb, -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 3/7] fbdev: c2p - Correct indentation 2009-01-04 11:46 ` [PATCH 2/7] fbdev: atafb - Fix 16 bpp console Geert Uytterhoeven @ 2009-01-04 11:47 ` Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 4/7] fbdev: c2p - Cleanups Geert Uytterhoeven 0 siblings, 1 reply; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:47 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/c2p.c | 238 ++++++++++++++++++++++++++------------------------- 1 files changed, 120 insertions(+), 118 deletions(-) diff --git a/drivers/video/c2p.c b/drivers/video/c2p.c index 376bc07..f102b57 100644 --- a/drivers/video/c2p.c +++ b/drivers/video/c2p.c @@ -21,71 +21,71 @@ * Basic transpose step */ -#define _transp(d, i1, i2, shift, mask) \ - do { \ - u32 t = (d[i1] ^ (d[i2] >> shift)) & mask; \ - d[i1] ^= t; \ - d[i2] ^= t << shift; \ - } while (0) +#define _transp(d, i1, i2, shift, mask) \ + do { \ + u32 t = (d[i1] ^ (d[i2] >> shift)) & mask; \ + d[i1] ^= t; \ + d[i2] ^= t << shift; \ + } while (0) static inline u32 get_mask(int n) { - switch (n) { + switch (n) { case 1: - return 0x55555555; - break; + return 0x55555555; + break; case 2: - return 0x33333333; - break; + return 0x33333333; + break; case 4: - return 0x0f0f0f0f; - break; + return 0x0f0f0f0f; + break; case 8: - return 0x00ff00ff; - break; + return 0x00ff00ff; + break; case 16: - return 0x0000ffff; - break; - } - return 0; + return 0x0000ffff; + break; + } + return 0; } -#define transp_nx1(d, n) \ - do { \ - u32 mask = get_mask(n); \ - /* First block */ \ - _transp(d, 0, 1, n, mask); \ - /* Second block */ \ - _transp(d, 2, 3, n, mask); \ - /* Third block */ \ - _transp(d, 4, 5, n, mask); \ - /* Fourth block */ \ - _transp(d, 6, 7, n, mask); \ - } while (0) - -#define transp_nx2(d, n) \ - do { \ - u32 mask = get_mask(n); \ - /* First block */ \ - _transp(d, 0, 2, n, mask); \ - _transp(d, 1, 3, n, mask); \ - /* Second block */ \ - _transp(d, 4, 6, n, mask); \ - _transp(d, 5, 7, n, mask); \ - } while (0) - -#define transp_nx4(d, n) \ - do { \ - u32 mask = get_mask(n); \ - _transp(d, 0, 4, n, mask); \ - _transp(d, 1, 5, n, mask); \ - _transp(d, 2, 6, n, mask); \ - _transp(d, 3, 7, n, mask); \ - } while (0) +#define transp_nx1(d, n) \ + do { \ + u32 mask = get_mask(n); \ + /* First block */ \ + _transp(d, 0, 1, n, mask); \ + /* Second block */ \ + _transp(d, 2, 3, n, mask); \ + /* Third block */ \ + _transp(d, 4, 5, n, mask); \ + /* Fourth block */ \ + _transp(d, 6, 7, n, mask); \ + } while (0) + +#define transp_nx2(d, n) \ + do { \ + u32 mask = get_mask(n); \ + /* First block */ \ + _transp(d, 0, 2, n, mask); \ + _transp(d, 1, 3, n, mask); \ + /* Second block */ \ + _transp(d, 4, 6, n, mask); \ + _transp(d, 5, 7, n, mask); \ + } while (0) + +#define transp_nx4(d, n) \ + do { \ + u32 mask = get_mask(n); \ + _transp(d, 0, 4, n, mask); \ + _transp(d, 1, 5, n, mask); \ + _transp(d, 2, 6, n, mask); \ + _transp(d, 3, 7, n, mask); \ + } while (0) #define transp(d, n, m) transp_nx ## m(d, n) @@ -99,11 +99,11 @@ static inline u32 get_mask(int n) static void c2p_8bpp(u32 d[8]) { - transp(d, 16, 4); - transp(d, 8, 2); - transp(d, 4, 1); - transp(d, 2, 4); - transp(d, 1, 2); + transp(d, 16, 4); + transp(d, 8, 2); + transp(d, 4, 1); + transp(d, 2, 4); + transp(d, 1, 2); } @@ -132,10 +132,10 @@ static inline unsigned long comp(unsigned long a, unsigned long b, static inline void store_planar(char *dst, u32 dst_inc, u32 bpp, u32 d[8]) { - int i; + int i; - for (i = 0; i < bpp; i++, dst += dst_inc) - *(u32 *)dst = d[perm_c2p_8bpp[i]]; + for (i = 0; i < bpp; i++, dst += dst_inc) + *(u32 *)dst = d[perm_c2p_8bpp[i]]; } @@ -146,10 +146,10 @@ static inline void store_planar(char *dst, u32 dst_inc, u32 bpp, u32 d[8]) static inline void store_planar_masked(char *dst, u32 dst_inc, u32 bpp, u32 d[8], u32 mask) { - int i; + int i; - for (i = 0; i < bpp; i++, dst += dst_inc) - *(u32 *)dst = comp(d[perm_c2p_8bpp[i]], *(u32 *)dst, mask); + for (i = 0; i < bpp; i++, dst += dst_inc) + *(u32 *)dst = comp(d[perm_c2p_8bpp[i]], *(u32 *)dst, mask); } @@ -169,63 +169,65 @@ static inline void store_planar_masked(char *dst, u32 dst_inc, u32 bpp, void c2p(u8 *dst, const u8 *src, u32 dx, u32 dy, u32 width, u32 height, u32 dst_nextline, u32 dst_nextplane, u32 src_nextline, u32 bpp) { - int dst_idx; - u32 d[8], first, last, w; - const u8 *c; - u8 *p; - - dst += dy*dst_nextline+(dx & ~31); - dst_idx = dx % 32; - first = ~0UL >> dst_idx; - last = ~(~0UL >> ((dst_idx+width) % 32)); - while (height--) { - c = src; - p = dst; - w = width; - if (dst_idx+width <= 32) { - /* Single destination word */ - first &= last; - memset(d, 0, sizeof(d)); - memcpy((u8 *)d+dst_idx, c, width); - c += width; - c2p_8bpp(d); - store_planar_masked(p, dst_nextplane, bpp, d, first); - p += 4; - } else { - /* Multiple destination words */ - w = width; - /* Leading bits */ - if (dst_idx) { - w = 32 - dst_idx; - memset(d, 0, dst_idx); - memcpy((u8 *)d+dst_idx, c, w); - c += w; - c2p_8bpp(d); - store_planar_masked(p, dst_nextplane, bpp, d, first); - p += 4; - w = width-w; - } - /* Main chunk */ - while (w >= 32) { - memcpy(d, c, 32); - c += 32; - c2p_8bpp(d); - store_planar(p, dst_nextplane, bpp, d); - p += 4; - w -= 32; - } - /* Trailing bits */ - w %= 32; - if (w > 0) { - memcpy(d, c, w); - memset((u8 *)d+w, 0, 32-w); - c2p_8bpp(d); - store_planar_masked(p, dst_nextplane, bpp, d, last); - } + int dst_idx; + u32 d[8], first, last, w; + const u8 *c; + u8 *p; + + dst += dy*dst_nextline+(dx & ~31); + dst_idx = dx % 32; + first = ~0UL >> dst_idx; + last = ~(~0UL >> ((dst_idx+width) % 32)); + while (height--) { + c = src; + p = dst; + w = width; + if (dst_idx+width <= 32) { + /* Single destination word */ + first &= last; + memset(d, 0, sizeof(d)); + memcpy((u8 *)d+dst_idx, c, width); + c += width; + c2p_8bpp(d); + store_planar_masked(p, dst_nextplane, bpp, d, first); + p += 4; + } else { + /* Multiple destination words */ + w = width; + /* Leading bits */ + if (dst_idx) { + w = 32 - dst_idx; + memset(d, 0, dst_idx); + memcpy((u8 *)d+dst_idx, c, w); + c += w; + c2p_8bpp(d); + store_planar_masked(p, dst_nextplane, bpp, d, + first); + p += 4; + w = width-w; + } + /* Main chunk */ + while (w >= 32) { + memcpy(d, c, 32); + c += 32; + c2p_8bpp(d); + store_planar(p, dst_nextplane, bpp, d); + p += 4; + w -= 32; + } + /* Trailing bits */ + w %= 32; + if (w > 0) { + memcpy(d, c, w); + memset((u8 *)d+w, 0, 32-w); + c2p_8bpp(d); + store_planar_masked(p, dst_nextplane, bpp, d, + last); + } + } + src += src_nextline; + dst += dst_nextline; } - src += src_nextline; - dst += dst_nextline; - } } EXPORT_SYMBOL_GPL(c2p); -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 4/7] fbdev: c2p - Cleanups 2009-01-04 11:47 ` [PATCH 3/7] fbdev: c2p - Correct indentation Geert Uytterhoeven @ 2009-01-04 11:47 ` Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 5/7] fbdev: c2p - Extract common c2p core to c2p_core.h Geert Uytterhoeven 0 siblings, 1 reply; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:47 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven - Improve comments and naming - Convert macros to static inline functions - Remove superfluous `break' after `return' - Make sure we get a build-time error (undefined reference to 'c2p_unsupported') in case of future misuse - Replace `unsigned long' by `u32' in comp(), as that's what all callers use - Use {get,put}_unaligned_be32() in store_planar{,_masked}() - Use void * for arbitrary pointers - Use a union to represent pixels/words, to avoid casts Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/c2p.c | 181 +++++++++++++++++++++++++++------------------------ drivers/video/c2p.h | 6 +- 2 files changed, 99 insertions(+), 88 deletions(-) diff --git a/drivers/video/c2p.c b/drivers/video/c2p.c index f102b57..c170fff 100644 --- a/drivers/video/c2p.c +++ b/drivers/video/c2p.c @@ -1,7 +1,7 @@ /* * Fast C2P (Chunky-to-Planar) Conversion * - * Copyright (C) 2003 Geert Uytterhoeven + * Copyright (C) 2003-2008 Geert Uytterhoeven * * NOTES: * - This code was inspired by Scout's C2P tutorial @@ -14,6 +14,9 @@ #include <linux/module.h> #include <linux/string.h> + +#include <asm/unaligned.h> + #include "c2p.h" @@ -21,97 +24,100 @@ * Basic transpose step */ -#define _transp(d, i1, i2, shift, mask) \ - do { \ - u32 t = (d[i1] ^ (d[i2] >> shift)) & mask; \ - d[i1] ^= t; \ - d[i2] ^= t << shift; \ - } while (0) +static inline void _transp(u32 d[], unsigned int i1, unsigned int i2, + unsigned int shift, u32 mask) +{ + u32 t = (d[i1] ^ (d[i2] >> shift)) & mask; + + d[i1] ^= t; + d[i2] ^= t << shift; +} -static inline u32 get_mask(int n) +extern void c2p_unsupported(void); + +static inline u32 get_mask(unsigned int n) { switch (n) { case 1: return 0x55555555; - break; case 2: return 0x33333333; - break; case 4: return 0x0f0f0f0f; - break; case 8: return 0x00ff00ff; - break; case 16: return 0x0000ffff; - break; } + + c2p_unsupported(); return 0; } -#define transp_nx1(d, n) \ - do { \ - u32 mask = get_mask(n); \ - /* First block */ \ - _transp(d, 0, 1, n, mask); \ - /* Second block */ \ - _transp(d, 2, 3, n, mask); \ - /* Third block */ \ - _transp(d, 4, 5, n, mask); \ - /* Fourth block */ \ - _transp(d, 6, 7, n, mask); \ - } while (0) - -#define transp_nx2(d, n) \ - do { \ - u32 mask = get_mask(n); \ - /* First block */ \ - _transp(d, 0, 2, n, mask); \ - _transp(d, 1, 3, n, mask); \ - /* Second block */ \ - _transp(d, 4, 6, n, mask); \ - _transp(d, 5, 7, n, mask); \ - } while (0) - -#define transp_nx4(d, n) \ - do { \ - u32 mask = get_mask(n); \ - _transp(d, 0, 4, n, mask); \ - _transp(d, 1, 5, n, mask); \ - _transp(d, 2, 6, n, mask); \ - _transp(d, 3, 7, n, mask); \ - } while (0) - -#define transp(d, n, m) transp_nx ## m(d, n) +static inline void transp8(u32 d[], unsigned int n, unsigned int m) +{ + u32 mask = get_mask(n); + + switch (m) { + case 1: + /* First n x 1 block */ + _transp(d, 0, 1, n, mask); + /* Second n x 1 block */ + _transp(d, 2, 3, n, mask); + /* Third n x 1 block */ + _transp(d, 4, 5, n, mask); + /* Fourth n x 1 block */ + _transp(d, 6, 7, n, mask); + return; + + case 2: + /* First n x 2 block */ + _transp(d, 0, 2, n, mask); + _transp(d, 1, 3, n, mask); + /* Second n x 2 block */ + _transp(d, 4, 6, n, mask); + _transp(d, 5, 7, n, mask); + return; + + case 4: + /* Single n x 4 block */ + _transp(d, 0, 4, n, mask); + _transp(d, 1, 5, n, mask); + _transp(d, 2, 6, n, mask); + _transp(d, 3, 7, n, mask); + return; + } + + c2p_unsupported(); +} /* * Perform a full C2P step on 32 8-bit pixels, stored in 8 32-bit words * containing * - 32 8-bit chunky pixels on input - * - permuted planar data on output + * - permutated planar data (1 plane per 32-bit word) on output */ -static void c2p_8bpp(u32 d[8]) +static void c2p_32x8(u32 d[8]) { - transp(d, 16, 4); - transp(d, 8, 2); - transp(d, 4, 1); - transp(d, 2, 4); - transp(d, 1, 2); + transp8(d, 16, 4); + transp8(d, 8, 2); + transp8(d, 4, 1); + transp8(d, 2, 4); + transp8(d, 1, 2); } /* - * Array containing the permution indices of the planar data after c2p + * Array containing the permutation indices of the planar data after c2p */ -static const int perm_c2p_8bpp[8] = { 7, 5, 3, 1, 6, 4, 2, 0 }; +static const int perm_c2p_32x8[8] = { 7, 5, 3, 1, 6, 4, 2, 0 }; /* @@ -119,8 +125,7 @@ static const int perm_c2p_8bpp[8] = { 7, 5, 3, 1, 6, 4, 2, 0 }; * This is equivalent to (a & mask) | (b & ~mask) */ -static inline unsigned long comp(unsigned long a, unsigned long b, - unsigned long mask) +static inline u32 comp(u32 a, u32 b, u32 mask) { return ((a ^ b) & mask) ^ b; } @@ -130,12 +135,12 @@ static inline unsigned long comp(unsigned long a, unsigned long b, * Store a full block of planar data after c2p conversion */ -static inline void store_planar(char *dst, u32 dst_inc, u32 bpp, u32 d[8]) +static inline void store_planar(void *dst, u32 dst_inc, u32 bpp, u32 d[8]) { int i; for (i = 0; i < bpp; i++, dst += dst_inc) - *(u32 *)dst = d[perm_c2p_8bpp[i]]; + put_unaligned_be32(d[perm_c2p_32x8[i]], dst); } @@ -143,13 +148,15 @@ static inline void store_planar(char *dst, u32 dst_inc, u32 bpp, u32 d[8]) * Store a partial block of planar data after c2p conversion */ -static inline void store_planar_masked(char *dst, u32 dst_inc, u32 bpp, +static inline void store_planar_masked(void *dst, u32 dst_inc, u32 bpp, u32 d[8], u32 mask) { int i; for (i = 0; i < bpp; i++, dst += dst_inc) - *(u32 *)dst = comp(d[perm_c2p_8bpp[i]], *(u32 *)dst, mask); + put_unaligned_be32(comp(d[perm_c2p_32x8[i]], + get_unaligned_be32(dst), mask), + dst); } @@ -166,18 +173,21 @@ static inline void store_planar_masked(char *dst, u32 dst_inc, u32 bpp, * @bpp: Bits per pixel of the planar frame buffer (1-8) */ -void c2p(u8 *dst, const u8 *src, u32 dx, u32 dy, u32 width, u32 height, +void c2p(void *dst, const void *src, u32 dx, u32 dy, u32 width, u32 height, u32 dst_nextline, u32 dst_nextplane, u32 src_nextline, u32 bpp) { - int dst_idx; - u32 d[8], first, last, w; + union { + u8 pixels[32]; + u32 words[8]; + } d; + u32 dst_idx, first, last, w; const u8 *c; - u8 *p; + void *p; dst += dy*dst_nextline+(dx & ~31); dst_idx = dx % 32; - first = ~0UL >> dst_idx; - last = ~(~0UL >> ((dst_idx+width) % 32)); + first = 0xffffffffU >> dst_idx; + last = ~(0xffffffffU >> ((dst_idx+width) % 32)); while (height--) { c = src; p = dst; @@ -185,11 +195,12 @@ void c2p(u8 *dst, const u8 *src, u32 dx, u32 dy, u32 width, u32 height, if (dst_idx+width <= 32) { /* Single destination word */ first &= last; - memset(d, 0, sizeof(d)); - memcpy((u8 *)d+dst_idx, c, width); + memset(d.pixels, 0, sizeof(d)); + memcpy(d.pixels+dst_idx, c, width); c += width; - c2p_8bpp(d); - store_planar_masked(p, dst_nextplane, bpp, d, first); + c2p_32x8(d.words); + store_planar_masked(p, dst_nextplane, bpp, d.words, + first); p += 4; } else { /* Multiple destination words */ @@ -197,32 +208,32 @@ void c2p(u8 *dst, const u8 *src, u32 dx, u32 dy, u32 width, u32 height, /* Leading bits */ if (dst_idx) { w = 32 - dst_idx; - memset(d, 0, dst_idx); - memcpy((u8 *)d+dst_idx, c, w); + memset(d.pixels, 0, dst_idx); + memcpy(d.pixels+dst_idx, c, w); c += w; - c2p_8bpp(d); - store_planar_masked(p, dst_nextplane, bpp, d, - first); + c2p_32x8(d.words); + store_planar_masked(p, dst_nextplane, bpp, + d.words, first); p += 4; w = width-w; } /* Main chunk */ while (w >= 32) { - memcpy(d, c, 32); + memcpy(d.pixels, c, 32); c += 32; - c2p_8bpp(d); - store_planar(p, dst_nextplane, bpp, d); + c2p_32x8(d.words); + store_planar(p, dst_nextplane, bpp, d.words); p += 4; w -= 32; } /* Trailing bits */ w %= 32; if (w > 0) { - memcpy(d, c, w); - memset((u8 *)d+w, 0, 32-w); - c2p_8bpp(d); - store_planar_masked(p, dst_nextplane, bpp, d, - last); + memcpy(d.pixels, c, w); + memset(d.pixels+w, 0, 32-w); + c2p_32x8(d.words); + store_planar_masked(p, dst_nextplane, bpp, + d.words, last); } } src += src_nextline; diff --git a/drivers/video/c2p.h b/drivers/video/c2p.h index c77cbf1..daafd87 100644 --- a/drivers/video/c2p.h +++ b/drivers/video/c2p.h @@ -10,7 +10,7 @@ #include <linux/types.h> -extern void c2p(u8 *dst, const u8 *src, u32 dx, u32 dy, u32 width, u32 height, - u32 dst_nextline, u32 dst_nextplane, u32 src_nextline, - u32 bpp); +extern void c2p(void *dst, const void *src, u32 dx, u32 dy, u32 width, + u32 height, u32 dst_nextline, u32 dst_nextplane, + u32 src_nextline, u32 bpp); -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 5/7] fbdev: c2p - Extract common c2p core to c2p_core.h 2009-01-04 11:47 ` [PATCH 4/7] fbdev: c2p - Cleanups Geert Uytterhoeven @ 2009-01-04 11:47 ` Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 6/7] fbdev: c2p/atafb - Add support for Atari interleaved bitplanes Geert Uytterhoeven 0 siblings, 1 reply; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:47 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/c2p.c | 92 +--------------------------------------- drivers/video/c2p_core.h | 106 ++++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 107 insertions(+), 91 deletions(-) create mode 100644 drivers/video/c2p_core.h diff --git a/drivers/video/c2p.c b/drivers/video/c2p.c index c170fff..1238c8b 100644 --- a/drivers/video/c2p.c +++ b/drivers/video/c2p.c @@ -3,10 +3,6 @@ * * Copyright (C) 2003-2008 Geert Uytterhoeven * - * NOTES: - * - This code was inspired by Scout's C2P tutorial - * - It assumes to run on a big endian system - * * 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. @@ -18,82 +14,7 @@ #include <asm/unaligned.h> #include "c2p.h" - - - /* - * Basic transpose step - */ - -static inline void _transp(u32 d[], unsigned int i1, unsigned int i2, - unsigned int shift, u32 mask) -{ - u32 t = (d[i1] ^ (d[i2] >> shift)) & mask; - - d[i1] ^= t; - d[i2] ^= t << shift; -} - -extern void c2p_unsupported(void); - -static inline u32 get_mask(unsigned int n) -{ - switch (n) { - case 1: - return 0x55555555; - - case 2: - return 0x33333333; - - case 4: - return 0x0f0f0f0f; - - case 8: - return 0x00ff00ff; - - case 16: - return 0x0000ffff; - } - - c2p_unsupported(); - return 0; -} - -static inline void transp8(u32 d[], unsigned int n, unsigned int m) -{ - u32 mask = get_mask(n); - - switch (m) { - case 1: - /* First n x 1 block */ - _transp(d, 0, 1, n, mask); - /* Second n x 1 block */ - _transp(d, 2, 3, n, mask); - /* Third n x 1 block */ - _transp(d, 4, 5, n, mask); - /* Fourth n x 1 block */ - _transp(d, 6, 7, n, mask); - return; - - case 2: - /* First n x 2 block */ - _transp(d, 0, 2, n, mask); - _transp(d, 1, 3, n, mask); - /* Second n x 2 block */ - _transp(d, 4, 6, n, mask); - _transp(d, 5, 7, n, mask); - return; - - case 4: - /* Single n x 4 block */ - _transp(d, 0, 4, n, mask); - _transp(d, 1, 5, n, mask); - _transp(d, 2, 6, n, mask); - _transp(d, 3, 7, n, mask); - return; - } - - c2p_unsupported(); -} +#include "c2p_core.h" /* @@ -121,17 +42,6 @@ static const int perm_c2p_32x8[8] = { 7, 5, 3, 1, 6, 4, 2, 0 }; /* - * Compose two values, using a bitmask as decision value - * This is equivalent to (a & mask) | (b & ~mask) - */ - -static inline u32 comp(u32 a, u32 b, u32 mask) -{ - return ((a ^ b) & mask) ^ b; -} - - - /* * Store a full block of planar data after c2p conversion */ diff --git a/drivers/video/c2p_core.h b/drivers/video/c2p_core.h new file mode 100644 index 0000000..3573cf7 --- /dev/null +++ b/drivers/video/c2p_core.h @@ -0,0 +1,106 @@ +/* + * Fast C2P (Chunky-to-Planar) Conversion + * + * Copyright (C) 2003-2008 Geert Uytterhoeven + * + * NOTES: + * - This code was inspired by Scout's C2P tutorial + * - It assumes to run on a big endian system + * + * 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. + */ + + + /* + * Basic transpose step + */ + +static inline void _transp(u32 d[], unsigned int i1, unsigned int i2, + unsigned int shift, u32 mask) +{ + u32 t = (d[i1] ^ (d[i2] >> shift)) & mask; + + d[i1] ^= t; + d[i2] ^= t << shift; +} + + +extern void c2p_unsupported(void); + +static inline u32 get_mask(unsigned int n) +{ + switch (n) { + case 1: + return 0x55555555; + + case 2: + return 0x33333333; + + case 4: + return 0x0f0f0f0f; + + case 8: + return 0x00ff00ff; + + case 16: + return 0x0000ffff; + } + + c2p_unsupported(); + return 0; +} + + + /* + * Transpose operations on 8 32-bit words + */ + +static inline void transp8(u32 d[], unsigned int n, unsigned int m) +{ + u32 mask = get_mask(n); + + switch (m) { + case 1: + /* First n x 1 block */ + _transp(d, 0, 1, n, mask); + /* Second n x 1 block */ + _transp(d, 2, 3, n, mask); + /* Third n x 1 block */ + _transp(d, 4, 5, n, mask); + /* Fourth n x 1 block */ + _transp(d, 6, 7, n, mask); + return; + + case 2: + /* First n x 2 block */ + _transp(d, 0, 2, n, mask); + _transp(d, 1, 3, n, mask); + /* Second n x 2 block */ + _transp(d, 4, 6, n, mask); + _transp(d, 5, 7, n, mask); + return; + + case 4: + /* Single n x 4 block */ + _transp(d, 0, 4, n, mask); + _transp(d, 1, 5, n, mask); + _transp(d, 2, 6, n, mask); + _transp(d, 3, 7, n, mask); + return; + } + + c2p_unsupported(); +} + + + /* + * Compose two values, using a bitmask as decision value + * This is equivalent to (a & mask) | (b & ~mask) + */ + +static inline u32 comp(u32 a, u32 b, u32 mask) +{ + return ((a ^ b) & mask) ^ b; +} -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 6/7] fbdev: c2p/atafb - Add support for Atari interleaved bitplanes 2009-01-04 11:47 ` [PATCH 5/7] fbdev: c2p - Extract common c2p core to c2p_core.h Geert Uytterhoeven @ 2009-01-04 11:47 ` Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 7/7] fbdev: c2p - Rename c2p to c2p_planar Geert Uytterhoeven 0 siblings, 1 reply; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:47 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven The c2p() for normal bitplanes is not suitable for interleaved bitplanes with 2 bytes of interleave, causing a garbled penguin logo. Add c2p_iplan2(). Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/Makefile | 2 +- drivers/video/atafb.c | 15 +---- drivers/video/c2p.h | 5 +- drivers/video/c2p_core.h | 47 ++++++++++++++ drivers/video/c2p_iplan2.c | 153 ++++++++++++++++++++++++++++++++++++++++++++ 5 files changed, 208 insertions(+), 14 deletions(-) create mode 100644 drivers/video/c2p_iplan2.c diff --git a/drivers/video/Makefile b/drivers/video/Makefile index e39e33e..4fc29ea 100644 --- a/drivers/video/Makefile +++ b/drivers/video/Makefile @@ -72,7 +72,7 @@ obj-$(CONFIG_FB_TCX) += tcx.o sbuslib.o obj-$(CONFIG_FB_LEO) += leo.o sbuslib.o obj-$(CONFIG_FB_SGIVW) += sgivwfb.o obj-$(CONFIG_FB_ACORN) += acornfb.o -obj-$(CONFIG_FB_ATARI) += atafb.o c2p.o atafb_mfb.o \ +obj-$(CONFIG_FB_ATARI) += atafb.o c2p_iplan2.o atafb_mfb.o \ atafb_iplan2p2.o atafb_iplan2p4.o atafb_iplan2p8.o obj-$(CONFIG_FB_MAC) += macfb.o obj-$(CONFIG_FB_HECUBA) += hecubafb.o diff --git a/drivers/video/atafb.c b/drivers/video/atafb.c index f932c13..477ce55 100644 --- a/drivers/video/atafb.c +++ b/drivers/video/atafb.c @@ -122,7 +122,6 @@ static struct atafb_par { void *screen_base; int yres_virtual; u_long next_line; - u_long next_plane; #if defined ATAFB_TT || defined ATAFB_STE union { struct { @@ -1392,14 +1391,7 @@ set_screen_base: par->screen_base = screen_base + var->yoffset * linelen; par->hw.falcon.xoffset = 0; - // FIXME!!! sort of works, no crash - //par->next_line = linelen; - //par->next_plane = yres_virtual * linelen; par->next_line = linelen; - par->next_plane = 2; - // crashes - //par->next_plane = linelen; - //par->next_line = yres_virtual * linelen; return 0; } @@ -2662,10 +2654,9 @@ static void atafb_imageblit(struct fb_info *info, const struct fb_image *image) src += pitch; } } else { - // only used for logo; broken - c2p(info->screen_base, image->data, dx, dy, width, height, - par->next_line, par->next_plane, image->width, - info->var.bits_per_pixel); + c2p_iplan2(info->screen_base, image->data, dx, dy, width, + height, par->next_line, image->width, + info->var.bits_per_pixel); } } diff --git a/drivers/video/c2p.h b/drivers/video/c2p.h index daafd87..5364129 100644 --- a/drivers/video/c2p.h +++ b/drivers/video/c2p.h @@ -1,7 +1,7 @@ /* * Fast C2P (Chunky-to-Planar) Conversion * - * Copyright (C) 2003 Geert Uytterhoeven + * Copyright (C) 2003-2008 Geert Uytterhoeven * * 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 @@ -14,3 +14,6 @@ extern void c2p(void *dst, const void *src, u32 dx, u32 dy, u32 width, u32 height, u32 dst_nextline, u32 dst_nextplane, u32 src_nextline, u32 bpp); +extern void c2p_iplan2(void *dst, const void *src, u32 dx, u32 dy, u32 width, + u32 height, u32 dst_nextline, u32 src_nextline, + u32 bpp); diff --git a/drivers/video/c2p_core.h b/drivers/video/c2p_core.h index 3573cf7..e1035a8 100644 --- a/drivers/video/c2p_core.h +++ b/drivers/video/c2p_core.h @@ -96,6 +96,53 @@ static inline void transp8(u32 d[], unsigned int n, unsigned int m) /* + * Transpose operations on 4 32-bit words + */ + +static inline void transp4(u32 d[], unsigned int n, unsigned int m) +{ + u32 mask = get_mask(n); + + switch (m) { + case 1: + /* First n x 1 block */ + _transp(d, 0, 1, n, mask); + /* Second n x 1 block */ + _transp(d, 2, 3, n, mask); + return; + + case 2: + /* Single n x 2 block */ + _transp(d, 0, 2, n, mask); + _transp(d, 1, 3, n, mask); + return; + } + + c2p_unsupported(); +} + + + /* + * Transpose operations on 4 32-bit words (reverse order) + */ + +static inline void transp4x(u32 d[], unsigned int n, unsigned int m) +{ + u32 mask = get_mask(n); + + switch (m) { + case 2: + /* Single n x 2 block */ + _transp(d, 2, 0, n, mask); + _transp(d, 3, 1, n, mask); + return; + } + + c2p_unsupported(); +} + + + /* * Compose two values, using a bitmask as decision value * This is equivalent to (a & mask) | (b & ~mask) */ diff --git a/drivers/video/c2p_iplan2.c b/drivers/video/c2p_iplan2.c new file mode 100644 index 0000000..19156dc --- /dev/null +++ b/drivers/video/c2p_iplan2.c @@ -0,0 +1,153 @@ +/* + * Fast C2P (Chunky-to-Planar) Conversion + * + * Copyright (C) 2003-2008 Geert Uytterhoeven + * + * 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. + */ + +#include <linux/module.h> +#include <linux/string.h> + +#include <asm/unaligned.h> + +#include "c2p.h" +#include "c2p_core.h" + + + /* + * Perform a full C2P step on 16 8-bit pixels, stored in 4 32-bit words + * containing + * - 16 8-bit chunky pixels on input + * - permutated planar data (2 planes per 32-bit word) on output + */ + +static void c2p_16x8(u32 d[4]) +{ + transp4(d, 8, 2); + transp4(d, 1, 2); + transp4x(d, 16, 2); + transp4x(d, 2, 2); + transp4(d, 4, 1); +} + + + /* + * Array containing the permutation indices of the planar data after c2p + */ + +static const int perm_c2p_16x8[4] = { 1, 3, 0, 2 }; + + + /* + * Store a full block of iplan2 data after c2p conversion + */ + +static inline void store_iplan2(void *dst, u32 bpp, u32 d[4]) +{ + int i; + + for (i = 0; i < bpp/2; i++, dst += 4) + put_unaligned_be32(d[perm_c2p_16x8[i]], dst); +} + + + /* + * Store a partial block of iplan2 data after c2p conversion + */ + +static inline void store_iplan2_masked(void *dst, u32 bpp, u32 d[4], u32 mask) +{ + int i; + + for (i = 0; i < bpp/2; i++, dst += 4) + put_unaligned_be32(comp(d[perm_c2p_16x8[i]], + get_unaligned_be32(dst), mask), + dst); +} + + + /* + * c2p_iplan2 - Copy 8-bit chunky image data to an interleaved planar + * frame buffer with 2 bytes of interleave + * @dst: Starting address of the planar frame buffer + * @dx: Horizontal destination offset (in pixels) + * @dy: Vertical destination offset (in pixels) + * @width: Image width (in pixels) + * @height: Image height (in pixels) + * @dst_nextline: Frame buffer offset to the next line (in bytes) + * @src_nextline: Image offset to the next line (in bytes) + * @bpp: Bits per pixel of the planar frame buffer (2, 4, or 8) + */ + +void c2p_iplan2(void *dst, const void *src, u32 dx, u32 dy, u32 width, + u32 height, u32 dst_nextline, u32 src_nextline, u32 bpp) +{ + union { + u8 pixels[16]; + u32 words[4]; + } d; + u32 dst_idx, first, last, w; + const u8 *c; + void *p; + + dst += dy*dst_nextline+(dx & ~15)*bpp; + dst_idx = dx % 16; + first = 0xffffU >> dst_idx; + first |= first << 16; + last = 0xffffU ^ (0xffffU >> ((dst_idx+width) % 16)); + last |= last << 16; + while (height--) { + c = src; + p = dst; + w = width; + if (dst_idx+width <= 16) { + /* Single destination word */ + first &= last; + memset(d.pixels, 0, sizeof(d)); + memcpy(d.pixels+dst_idx, c, width); + c += width; + c2p_16x8(d.words); + store_iplan2_masked(p, bpp, d.words, first); + p += bpp*2; + } else { + /* Multiple destination words */ + w = width; + /* Leading bits */ + if (dst_idx) { + w = 16 - dst_idx; + memset(d.pixels, 0, dst_idx); + memcpy(d.pixels+dst_idx, c, w); + c += w; + c2p_16x8(d.words); + store_iplan2_masked(p, bpp, d.words, first); + p += bpp*2; + w = width-w; + } + /* Main chunk */ + while (w >= 16) { + memcpy(d.pixels, c, 16); + c += 16; + c2p_16x8(d.words); + store_iplan2(p, bpp, d.words); + p += bpp*2; + w -= 16; + } + /* Trailing bits */ + w %= 16; + if (w > 0) { + memcpy(d.pixels, c, w); + memset(d.pixels+w, 0, 16-w); + c2p_16x8(d.words); + store_iplan2_masked(p, bpp, d.words, last); + } + } + src += src_nextline; + dst += dst_nextline; + } +} +EXPORT_SYMBOL_GPL(c2p_iplan2); + +MODULE_LICENSE("GPL"); -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* [PATCH 7/7] fbdev: c2p - Rename c2p to c2p_planar 2009-01-04 11:47 ` [PATCH 6/7] fbdev: c2p/atafb - Add support for Atari interleaved bitplanes Geert Uytterhoeven @ 2009-01-04 11:47 ` Geert Uytterhoeven 0 siblings, 0 replies; 8+ messages in thread From: Geert Uytterhoeven @ 2009-01-04 11:47 UTC (permalink / raw) To: Michael Schmitz, linux-fbdev-devel Cc: linux-m68k, linux-kernel, Geert Uytterhoeven Signed-off-by: Geert Uytterhoeven <geert@linux-m68k.org> --- drivers/video/Makefile | 2 +- drivers/video/amifb.c | 6 +- drivers/video/c2p.c | 155 ------------------------------------------- drivers/video/c2p.h | 6 +- drivers/video/c2p_planar.c | 156 ++++++++++++++++++++++++++++++++++++++++++++ 5 files changed, 163 insertions(+), 162 deletions(-) delete mode 100644 drivers/video/c2p.c create mode 100644 drivers/video/c2p_planar.c diff --git a/drivers/video/Makefile b/drivers/video/Makefile index 4fc29ea..be2b657 100644 --- a/drivers/video/Makefile +++ b/drivers/video/Makefile @@ -28,7 +28,7 @@ obj-$(CONFIG_FB_DDC) += fb_ddc.o obj-$(CONFIG_FB_DEFERRED_IO) += fb_defio.o # Hardware specific drivers go first -obj-$(CONFIG_FB_AMIGA) += amifb.o c2p.o +obj-$(CONFIG_FB_AMIGA) += amifb.o c2p_planar.o obj-$(CONFIG_FB_ARC) += arcfb.o obj-$(CONFIG_FB_CLPS711X) += clps711xfb.o obj-$(CONFIG_FB_CYBER2000) += cyber2000fb.o diff --git a/drivers/video/amifb.c b/drivers/video/amifb.c index b8e9a86..100f236 100644 --- a/drivers/video/amifb.c +++ b/drivers/video/amifb.c @@ -2159,9 +2159,9 @@ static void amifb_imageblit(struct fb_info *info, const struct fb_image *image) src += pitch; } } else { - c2p(info->screen_base, image->data, dx, dy, width, height, - par->next_line, par->next_plane, image->width, - info->var.bits_per_pixel); + c2p_planar(info->screen_base, image->data, dx, dy, width, + height, par->next_line, par->next_plane, + image->width, info->var.bits_per_pixel); } } diff --git a/drivers/video/c2p.c b/drivers/video/c2p.c deleted file mode 100644 index 1238c8b..0000000 --- a/drivers/video/c2p.c +++ /dev/null @@ -1,155 +0,0 @@ -/* - * Fast C2P (Chunky-to-Planar) Conversion - * - * Copyright (C) 2003-2008 Geert Uytterhoeven - * - * 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. - */ - -#include <linux/module.h> -#include <linux/string.h> - -#include <asm/unaligned.h> - -#include "c2p.h" -#include "c2p_core.h" - - - /* - * Perform a full C2P step on 32 8-bit pixels, stored in 8 32-bit words - * containing - * - 32 8-bit chunky pixels on input - * - permutated planar data (1 plane per 32-bit word) on output - */ - -static void c2p_32x8(u32 d[8]) -{ - transp8(d, 16, 4); - transp8(d, 8, 2); - transp8(d, 4, 1); - transp8(d, 2, 4); - transp8(d, 1, 2); -} - - - /* - * Array containing the permutation indices of the planar data after c2p - */ - -static const int perm_c2p_32x8[8] = { 7, 5, 3, 1, 6, 4, 2, 0 }; - - - /* - * Store a full block of planar data after c2p conversion - */ - -static inline void store_planar(void *dst, u32 dst_inc, u32 bpp, u32 d[8]) -{ - int i; - - for (i = 0; i < bpp; i++, dst += dst_inc) - put_unaligned_be32(d[perm_c2p_32x8[i]], dst); -} - - - /* - * Store a partial block of planar data after c2p conversion - */ - -static inline void store_planar_masked(void *dst, u32 dst_inc, u32 bpp, - u32 d[8], u32 mask) -{ - int i; - - for (i = 0; i < bpp; i++, dst += dst_inc) - put_unaligned_be32(comp(d[perm_c2p_32x8[i]], - get_unaligned_be32(dst), mask), - dst); -} - - - /* - * c2p - Copy 8-bit chunky image data to a planar frame buffer - * @dst: Starting address of the planar frame buffer - * @dx: Horizontal destination offset (in pixels) - * @dy: Vertical destination offset (in pixels) - * @width: Image width (in pixels) - * @height: Image height (in pixels) - * @dst_nextline: Frame buffer offset to the next line (in bytes) - * @dst_nextplane: Frame buffer offset to the next plane (in bytes) - * @src_nextline: Image offset to the next line (in bytes) - * @bpp: Bits per pixel of the planar frame buffer (1-8) - */ - -void c2p(void *dst, const void *src, u32 dx, u32 dy, u32 width, u32 height, - u32 dst_nextline, u32 dst_nextplane, u32 src_nextline, u32 bpp) -{ - union { - u8 pixels[32]; - u32 words[8]; - } d; - u32 dst_idx, first, last, w; - const u8 *c; - void *p; - - dst += dy*dst_nextline+(dx & ~31); - dst_idx = dx % 32; - first = 0xffffffffU >> dst_idx; - last = ~(0xffffffffU >> ((dst_idx+width) % 32)); - while (height--) { - c = src; - p = dst; - w = width; - if (dst_idx+width <= 32) { - /* Single destination word */ - first &= last; - memset(d.pixels, 0, sizeof(d)); - memcpy(d.pixels+dst_idx, c, width); - c += width; - c2p_32x8(d.words); - store_planar_masked(p, dst_nextplane, bpp, d.words, - first); - p += 4; - } else { - /* Multiple destination words */ - w = width; - /* Leading bits */ - if (dst_idx) { - w = 32 - dst_idx; - memset(d.pixels, 0, dst_idx); - memcpy(d.pixels+dst_idx, c, w); - c += w; - c2p_32x8(d.words); - store_planar_masked(p, dst_nextplane, bpp, - d.words, first); - p += 4; - w = width-w; - } - /* Main chunk */ - while (w >= 32) { - memcpy(d.pixels, c, 32); - c += 32; - c2p_32x8(d.words); - store_planar(p, dst_nextplane, bpp, d.words); - p += 4; - w -= 32; - } - /* Trailing bits */ - w %= 32; - if (w > 0) { - memcpy(d.pixels, c, w); - memset(d.pixels+w, 0, 32-w); - c2p_32x8(d.words); - store_planar_masked(p, dst_nextplane, bpp, - d.words, last); - } - } - src += src_nextline; - dst += dst_nextline; - } -} -EXPORT_SYMBOL_GPL(c2p); - -MODULE_LICENSE("GPL"); diff --git a/drivers/video/c2p.h b/drivers/video/c2p.h index 5364129..6c38d40 100644 --- a/drivers/video/c2p.h +++ b/drivers/video/c2p.h @@ -10,9 +10,9 @@ #include <linux/types.h> -extern void c2p(void *dst, const void *src, u32 dx, u32 dy, u32 width, - u32 height, u32 dst_nextline, u32 dst_nextplane, - u32 src_nextline, u32 bpp); +extern void c2p_planar(void *dst, const void *src, u32 dx, u32 dy, u32 width, + u32 height, u32 dst_nextline, u32 dst_nextplane, + u32 src_nextline, u32 bpp); extern void c2p_iplan2(void *dst, const void *src, u32 dx, u32 dy, u32 width, u32 height, u32 dst_nextline, u32 src_nextline, diff --git a/drivers/video/c2p_planar.c b/drivers/video/c2p_planar.c new file mode 100644 index 0000000..ec7ac85 --- /dev/null +++ b/drivers/video/c2p_planar.c @@ -0,0 +1,156 @@ +/* + * Fast C2P (Chunky-to-Planar) Conversion + * + * Copyright (C) 2003-2008 Geert Uytterhoeven + * + * 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. + */ + +#include <linux/module.h> +#include <linux/string.h> + +#include <asm/unaligned.h> + +#include "c2p.h" +#include "c2p_core.h" + + + /* + * Perform a full C2P step on 32 8-bit pixels, stored in 8 32-bit words + * containing + * - 32 8-bit chunky pixels on input + * - permutated planar data (1 plane per 32-bit word) on output + */ + +static void c2p_32x8(u32 d[8]) +{ + transp8(d, 16, 4); + transp8(d, 8, 2); + transp8(d, 4, 1); + transp8(d, 2, 4); + transp8(d, 1, 2); +} + + + /* + * Array containing the permutation indices of the planar data after c2p + */ + +static const int perm_c2p_32x8[8] = { 7, 5, 3, 1, 6, 4, 2, 0 }; + + + /* + * Store a full block of planar data after c2p conversion + */ + +static inline void store_planar(void *dst, u32 dst_inc, u32 bpp, u32 d[8]) +{ + int i; + + for (i = 0; i < bpp; i++, dst += dst_inc) + put_unaligned_be32(d[perm_c2p_32x8[i]], dst); +} + + + /* + * Store a partial block of planar data after c2p conversion + */ + +static inline void store_planar_masked(void *dst, u32 dst_inc, u32 bpp, + u32 d[8], u32 mask) +{ + int i; + + for (i = 0; i < bpp; i++, dst += dst_inc) + put_unaligned_be32(comp(d[perm_c2p_32x8[i]], + get_unaligned_be32(dst), mask), + dst); +} + + + /* + * c2p_planar - Copy 8-bit chunky image data to a planar frame buffer + * @dst: Starting address of the planar frame buffer + * @dx: Horizontal destination offset (in pixels) + * @dy: Vertical destination offset (in pixels) + * @width: Image width (in pixels) + * @height: Image height (in pixels) + * @dst_nextline: Frame buffer offset to the next line (in bytes) + * @dst_nextplane: Frame buffer offset to the next plane (in bytes) + * @src_nextline: Image offset to the next line (in bytes) + * @bpp: Bits per pixel of the planar frame buffer (1-8) + */ + +void c2p_planar(void *dst, const void *src, u32 dx, u32 dy, u32 width, + u32 height, u32 dst_nextline, u32 dst_nextplane, + u32 src_nextline, u32 bpp) +{ + union { + u8 pixels[32]; + u32 words[8]; + } d; + u32 dst_idx, first, last, w; + const u8 *c; + void *p; + + dst += dy*dst_nextline+(dx & ~31); + dst_idx = dx % 32; + first = 0xffffffffU >> dst_idx; + last = ~(0xffffffffU >> ((dst_idx+width) % 32)); + while (height--) { + c = src; + p = dst; + w = width; + if (dst_idx+width <= 32) { + /* Single destination word */ + first &= last; + memset(d.pixels, 0, sizeof(d)); + memcpy(d.pixels+dst_idx, c, width); + c += width; + c2p_32x8(d.words); + store_planar_masked(p, dst_nextplane, bpp, d.words, + first); + p += 4; + } else { + /* Multiple destination words */ + w = width; + /* Leading bits */ + if (dst_idx) { + w = 32 - dst_idx; + memset(d.pixels, 0, dst_idx); + memcpy(d.pixels+dst_idx, c, w); + c += w; + c2p_32x8(d.words); + store_planar_masked(p, dst_nextplane, bpp, + d.words, first); + p += 4; + w = width-w; + } + /* Main chunk */ + while (w >= 32) { + memcpy(d.pixels, c, 32); + c += 32; + c2p_32x8(d.words); + store_planar(p, dst_nextplane, bpp, d.words); + p += 4; + w -= 32; + } + /* Trailing bits */ + w %= 32; + if (w > 0) { + memcpy(d.pixels, c, w); + memset(d.pixels+w, 0, 32-w); + c2p_32x8(d.words); + store_planar_masked(p, dst_nextplane, bpp, + d.words, last); + } + } + src += src_nextline; + dst += dst_nextline; + } +} +EXPORT_SYMBOL_GPL(c2p_planar); + +MODULE_LICENSE("GPL"); -- 1.5.6.5 ^ permalink raw reply related [flat|nested] 8+ messages in thread
end of thread, other threads:[~2009-01-04 11:47 UTC | newest] Thread overview: 8+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2009-01-04 11:46 [PATCH/v2] Atari frame buffer fixes Geert Uytterhoeven 2009-01-04 11:46 ` [PATCH 1/7] fbdev: atafb - Fix line length handling Geert Uytterhoeven 2009-01-04 11:46 ` [PATCH 2/7] fbdev: atafb - Fix 16 bpp console Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 3/7] fbdev: c2p - Correct indentation Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 4/7] fbdev: c2p - Cleanups Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 5/7] fbdev: c2p - Extract common c2p core to c2p_core.h Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 6/7] fbdev: c2p/atafb - Add support for Atari interleaved bitplanes Geert Uytterhoeven 2009-01-04 11:47 ` [PATCH 7/7] fbdev: c2p - Rename c2p to c2p_planar Geert Uytterhoeven
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox