cachepc-linux

Fork of AMDESE/linux with modifications for CachePC side-channel attack
git clone https://git.sinitax.com/sinitax/cachepc-linux
Log | Files | Refs | README | LICENSE | sfeed.txt

c2p_planar.c (3697B)

      1/*
      2 *  Fast C2P (Chunky-to-Planar) Conversion
      3 *
      4 *  Copyright (C) 2003-2008 Geert Uytterhoeven
      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
      8 *  for more details.
      9 */
     10
     11#include <linux/module.h>
     12#include <linux/string.h>
     13
     14#include <asm/unaligned.h>
     15
     16#include "c2p.h"
     17#include "c2p_core.h"
     18
     19
     20    /*
     21     *  Perform a full C2P step on 32 8-bit pixels, stored in 8 32-bit words
     22     *  containing
     23     *    - 32 8-bit chunky pixels on input
     24     *    - permutated planar data (1 plane per 32-bit word) on output
     25     */
     26
     27static void c2p_32x8(u32 d[8])
     28{
     29	transp8(d, 16, 4);
     30	transp8(d, 8, 2);
     31	transp8(d, 4, 1);
     32	transp8(d, 2, 4);
     33	transp8(d, 1, 2);
     34}
     35
     36
     37    /*
     38     *  Array containing the permutation indices of the planar data after c2p
     39     */
     40
     41static const int perm_c2p_32x8[8] = { 7, 5, 3, 1, 6, 4, 2, 0 };
     42
     43
     44    /*
     45     *  Store a full block of planar data after c2p conversion
     46     */
     47
     48static inline void store_planar(void *dst, u32 dst_inc, u32 bpp, u32 d[8])
     49{
     50	int i;
     51
     52	for (i = 0; i < bpp; i++, dst += dst_inc)
     53		put_unaligned_be32(d[perm_c2p_32x8[i]], dst);
     54}
     55
     56
     57    /*
     58     *  Store a partial block of planar data after c2p conversion
     59     */
     60
     61static inline void store_planar_masked(void *dst, u32 dst_inc, u32 bpp,
     62				       u32 d[8], u32 mask)
     63{
     64	int i;
     65
     66	for (i = 0; i < bpp; i++, dst += dst_inc)
     67		put_unaligned_be32(comp(d[perm_c2p_32x8[i]],
     68					get_unaligned_be32(dst), mask),
     69				   dst);
     70}
     71
     72
     73    /*
     74     *  c2p_planar - Copy 8-bit chunky image data to a planar frame buffer
     75     *  @dst: Starting address of the planar frame buffer
     76     *  @dx: Horizontal destination offset (in pixels)
     77     *  @dy: Vertical destination offset (in pixels)
     78     *  @width: Image width (in pixels)
     79     *  @height: Image height (in pixels)
     80     *  @dst_nextline: Frame buffer offset to the next line (in bytes)
     81     *  @dst_nextplane: Frame buffer offset to the next plane (in bytes)
     82     *  @src_nextline: Image offset to the next line (in bytes)
     83     *  @bpp: Bits per pixel of the planar frame buffer (1-8)
     84     */
     85
     86void c2p_planar(void *dst, const void *src, u32 dx, u32 dy, u32 width,
     87		u32 height, u32 dst_nextline, u32 dst_nextplane,
     88		u32 src_nextline, u32 bpp)
     89{
     90	union {
     91		u8 pixels[32];
     92		u32 words[8];
     93	} d;
     94	u32 dst_idx, first, last, w;
     95	const u8 *c;
     96	void *p;
     97
     98	dst += dy*dst_nextline+(dx & ~31);
     99	dst_idx = dx % 32;
    100	first = 0xffffffffU >> dst_idx;
    101	last = ~(0xffffffffU >> ((dst_idx+width) % 32));
    102	while (height--) {
    103		c = src;
    104		p = dst;
    105		w = width;
    106		if (dst_idx+width <= 32) {
    107			/* Single destination word */
    108			first &= last;
    109			memset(d.pixels, 0, sizeof(d));
    110			memcpy(d.pixels+dst_idx, c, width);
    111			c += width;
    112			c2p_32x8(d.words);
    113			store_planar_masked(p, dst_nextplane, bpp, d.words,
    114					    first);
    115			p += 4;
    116		} else {
    117			/* Multiple destination words */
    118			w = width;
    119			/* Leading bits */
    120			if (dst_idx) {
    121				w = 32 - dst_idx;
    122				memset(d.pixels, 0, dst_idx);
    123				memcpy(d.pixels+dst_idx, c, w);
    124				c += w;
    125				c2p_32x8(d.words);
    126				store_planar_masked(p, dst_nextplane, bpp,
    127						    d.words, first);
    128				p += 4;
    129				w = width-w;
    130			}
    131			/* Main chunk */
    132			while (w >= 32) {
    133				memcpy(d.pixels, c, 32);
    134				c += 32;
    135				c2p_32x8(d.words);
    136				store_planar(p, dst_nextplane, bpp, d.words);
    137				p += 4;
    138				w -= 32;
    139			}
    140			/* Trailing bits */
    141			w %= 32;
    142			if (w > 0) {
    143				memcpy(d.pixels, c, w);
    144				memset(d.pixels+w, 0, 32-w);
    145				c2p_32x8(d.words);
    146				store_planar_masked(p, dst_nextplane, bpp,
    147						    d.words, last);
    148			}
    149		}
    150		src += src_nextline;
    151		dst += dst_nextline;
    152	}
    153}
    154EXPORT_SYMBOL_GPL(c2p_planar);
    155
    156MODULE_LICENSE("GPL");