X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fcavsdsp.c;h=cc7898968fb846614163471d08ff08f8059c1326;hb=e10b7ef2fe56603fb1baac6b20fd6bd0a3fdd0d0;hp=d2517d9e9ded49616bc031bb9a97b5c80ff3be6f;hpb=595e7bd94029713d87e3e4943b40676df1b2d0a0;p=ffmpeg

diff --git a/libavcodec/cavsdsp.c b/libavcodec/cavsdsp.c
index d2517d9e9de..cc7898968fb 100644
--- a/libavcodec/cavsdsp.c
+++ b/libavcodec/cavsdsp.c
@@ -5,23 +5,29 @@
  *
  * Copyright (c) 2006  Stefan Gehrer <stefan.gehrer@gmx.de>
  *
- * This library is free software; you can redistribute it and/or
+ * This file is part of Libav.
+ *
+ * Libav is free software; you can redistribute it and/or
  * modify it under the terms of the GNU Lesser General Public
  * License as published by the Free Software Foundation; either
- * version 2 of the License, or (at your option) any later version.
+ * version 2.1 of the License, or (at your option) any later version.
  *
- * This library is distributed in the hope that it will be useful,
+ * Libav is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  * Lesser General Public License for more details.
  *
  * You should have received a copy of the GNU Lesser General Public
- * License along with this library; if not, write to the Free Software
- * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ * License along with Libav; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
 #include <stdio.h>
-#include "dsputil.h"
+
+#include "idctdsp.h"
+#include "mathops.h"
+#include "cavsdsp.h"
+#include "libavutil/common.h"
 
 /*****************************************************************************
  *
@@ -61,16 +67,16 @@ static inline void loop_filter_l1(uint8_t *p0_p, int stride, int alpha, int beta
     int q0 = Q0;
 
     if(abs(p0-q0)<alpha && abs(P1-p0)<beta && abs(Q1-q0)<beta) {
-        int delta = clip(((q0-p0)*3+P1-Q1+4)>>3,-tc, tc);
-        P0 = clip_uint8(p0+delta);
-        Q0 = clip_uint8(q0-delta);
+        int delta = av_clip(((q0-p0)*3+P1-Q1+4)>>3,-tc, tc);
+        P0 = av_clip_uint8(p0+delta);
+        Q0 = av_clip_uint8(q0-delta);
         if(abs(P2-p0)<beta) {
-            delta = clip(((P0-P1)*3+P2-Q0+4)>>3, -tc, tc);
-            P1 = clip_uint8(P1+delta);
+            delta = av_clip(((P0-P1)*3+P2-Q0+4)>>3, -tc, tc);
+            P1 = av_clip_uint8(P1+delta);
         }
         if(abs(Q2-q0)<beta) {
-            delta = clip(((Q1-Q0)*3+P0-Q2+4)>>3, -tc, tc);
-            Q1 = clip_uint8(Q1-delta);
+            delta = av_clip(((Q1-Q0)*3+P0-Q2+4)>>3, -tc, tc);
+            Q1 = av_clip_uint8(Q1-delta);
         }
     }
 }
@@ -96,9 +102,9 @@ static inline void loop_filter_c2(uint8_t *p0_p,int stride,int alpha, int beta)
 static inline void loop_filter_c1(uint8_t *p0_p,int stride,int alpha, int beta,
                                   int tc) {
     if(abs(P0-Q0)<alpha && abs(P1-P0)<beta && abs(Q1-Q0)<beta) {
-        int delta = clip(((Q0-P0)*3+P1-Q1+4)>>3, -tc, tc);
-        P0 = clip_uint8(P0+delta);
-        Q0 = clip_uint8(Q0-delta);
+        int delta = av_clip(((Q0-P0)*3+P1-Q1+4)>>3, -tc, tc);
+        P0 = av_clip_uint8(P0+delta);
+        Q0 = av_clip_uint8(Q0-delta);
     }
 }
 
@@ -179,10 +185,10 @@ static void cavs_filter_ch_c(uint8_t *d, int stride, int alpha, int beta, int tc
  *
  ****************************************************************************/
 
-static void cavs_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride) {
+static void cavs_idct8_add_c(uint8_t *dst, int16_t *block, int stride) {
     int i;
-    DCTELEM (*src)[8] = (DCTELEM(*)[8])block;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    int16_t (*src)[8] = (int16_t(*)[8])block;
+    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;
 
     src[0][0] += 8;
 
@@ -246,7 +252,6 @@ static void cavs_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride) {
         dst[i + 6*stride] = cm[ dst[i + 6*stride] + ((b1 - b5) >> 7)];
         dst[i + 7*stride] = cm[ dst[i + 7*stride] + ((b0 - b4) >> 7)];
     }
-    memset(block,0,64*sizeof(DCTELEM));
 }
 
 /*****************************************************************************
@@ -256,9 +261,9 @@ static void cavs_idct8_add_c(uint8_t *dst, DCTELEM *block, int stride) {
  ****************************************************************************/
 
 #define CAVS_SUBPIX(OPNAME, OP, NAME, A, B, C, D, E, F) \
-static void OPNAME ## cavs_filt8_h_ ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+static void OPNAME ## cavs_filt8_h_ ## NAME(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride){\
     const int h=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<h; i++)\
     {\
@@ -275,9 +280,9 @@ static void OPNAME ## cavs_filt8_h_ ## NAME(uint8_t *dst, uint8_t *src, int dstS
     }\
 }\
 \
-static void OPNAME ## cavs_filt8_v_  ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+static void OPNAME ## cavs_filt8_v_  ## NAME(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride){\
     const int w=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<w; i++)\
     {\
@@ -307,7 +312,7 @@ static void OPNAME ## cavs_filt8_v_  ## NAME(uint8_t *dst, uint8_t *src, int dst
     }\
 }\
 \
-static void OPNAME ## cavs_filt16_v_ ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+static void OPNAME ## cavs_filt16_v_ ## NAME(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride){\
     OPNAME ## cavs_filt8_v_ ## NAME(dst  , src  , dstStride, srcStride);\
     OPNAME ## cavs_filt8_v_ ## NAME(dst+8, src+8, dstStride, srcStride);\
     src += 8*srcStride;\
@@ -316,7 +321,7 @@ static void OPNAME ## cavs_filt16_v_ ## NAME(uint8_t *dst, uint8_t *src, int dst
     OPNAME ## cavs_filt8_v_ ## NAME(dst+8, src+8, dstStride, srcStride);\
 }\
 \
-static void OPNAME ## cavs_filt16_h_ ## NAME(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+static void OPNAME ## cavs_filt16_h_ ## NAME(uint8_t *dst, const uint8_t *src, int dstStride, int srcStride){\
     OPNAME ## cavs_filt8_h_ ## NAME(dst  , src  , dstStride, srcStride);\
     OPNAME ## cavs_filt8_h_ ## NAME(dst+8, src+8, dstStride, srcStride);\
     src += 8*srcStride;\
@@ -326,12 +331,12 @@ static void OPNAME ## cavs_filt16_h_ ## NAME(uint8_t *dst, uint8_t *src, int dst
 }\
 
 #define CAVS_SUBPIX_HV(OPNAME, OP, NAME, AH, BH, CH, DH, EH, FH, AV, BV, CV, DV, EV, FV, FULL) \
-static void OPNAME ## cavs_filt8_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8_t *src2, int dstStride, int srcStride){\
+static void OPNAME ## cavs_filt8_hv_ ## NAME(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dstStride, int srcStride){\
     int16_t temp[8*(8+5)];\
     int16_t *tmp = temp;\
     const int h=8;\
     const int w=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    const uint8_t *cm = ff_crop_tab + MAX_NEG_CROP;\
     int i;\
     src1 -= 2*srcStride;\
     for(i=0; i<h+5; i++)\
@@ -407,7 +412,7 @@ static void OPNAME ## cavs_filt8_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8_
     }\
 }\
 \
-static void OPNAME ## cavs_filt16_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8_t *src2, int dstStride, int srcStride){ \
+static void OPNAME ## cavs_filt16_hv_ ## NAME(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dstStride, int srcStride){ \
     OPNAME ## cavs_filt8_hv_ ## NAME(dst  , src1,   src2  , dstStride, srcStride); \
     OPNAME ## cavs_filt8_hv_ ## NAME(dst+8, src1+8, src2+8, dstStride, srcStride); \
     src1 += 8*srcStride;\
@@ -418,63 +423,78 @@ static void OPNAME ## cavs_filt16_hv_ ## NAME(uint8_t *dst, uint8_t *src1, uint8
 }\
 
 #define CAVS_MC(OPNAME, SIZE) \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc10_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
     OPNAME ## cavs_filt ## SIZE ## _h_qpel_l(dst, src, stride, stride);\
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc20_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
     OPNAME ## cavs_filt ## SIZE ## _h_hpel(dst, src, stride, stride);\
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc30_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
     OPNAME ## cavs_filt ## SIZE ## _h_qpel_r(dst, src, stride, stride);\
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc01_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
     OPNAME ## cavs_filt ## SIZE ## _v_qpel_l(dst, src, stride, stride);\
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc02_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
     OPNAME ## cavs_filt ## SIZE ## _v_hpel(dst, src, stride, stride);\
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc03_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
     OPNAME ## cavs_filt ## SIZE ## _v_qpel_r(dst, src, stride, stride);\
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc22_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_jj(dst, src, NULL, stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc11_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src, stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc13_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+stride, stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc31_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+1, stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc33_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_egpr(dst, src, src+stride+1,stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc21_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_ff(dst, src, src+stride+1,stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc12_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_ii(dst, src, src+stride+1,stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc32_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_kk(dst, src, src+stride+1,stride, stride); \
 }\
 \
-static void ff_ ## OPNAME ## cavs_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){\
+static void OPNAME ## cavs_qpel ## SIZE ## _mc23_c(uint8_t *dst, const uint8_t *src, ptrdiff_t stride)\
+{\
   OPNAME ## cavs_filt ## SIZE ## _hv_qq(dst, src, src+stride+1,stride, stride); \
 }\
 
@@ -509,29 +529,29 @@ CAVS_MC(put_, 16)
 CAVS_MC(avg_, 8)
 CAVS_MC(avg_, 16)
 
-void ff_put_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-void ff_avg_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-void ff_put_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride);
-void ff_avg_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride);
+#define put_cavs_qpel8_mc00_c  ff_put_pixels8x8_c
+#define avg_cavs_qpel8_mc00_c  ff_avg_pixels8x8_c
+#define put_cavs_qpel16_mc00_c ff_put_pixels16x16_c
+#define avg_cavs_qpel16_mc00_c ff_avg_pixels16x16_c
 
-void ff_cavsdsp_init(DSPContext* c, AVCodecContext *avctx) {
+av_cold void ff_cavsdsp_init(CAVSDSPContext* c, AVCodecContext *avctx) {
 #define dspfunc(PFX, IDX, NUM) \
-    c->PFX ## _pixels_tab[IDX][ 0] = ff_ ## PFX ## NUM ## _mc00_c; \
-    c->PFX ## _pixels_tab[IDX][ 1] = ff_ ## PFX ## NUM ## _mc10_c; \
-    c->PFX ## _pixels_tab[IDX][ 2] = ff_ ## PFX ## NUM ## _mc20_c; \
-    c->PFX ## _pixels_tab[IDX][ 3] = ff_ ## PFX ## NUM ## _mc30_c; \
-    c->PFX ## _pixels_tab[IDX][ 4] = ff_ ## PFX ## NUM ## _mc01_c; \
-    c->PFX ## _pixels_tab[IDX][ 5] = ff_ ## PFX ## NUM ## _mc11_c; \
-    c->PFX ## _pixels_tab[IDX][ 6] = ff_ ## PFX ## NUM ## _mc21_c; \
-    c->PFX ## _pixels_tab[IDX][ 7] = ff_ ## PFX ## NUM ## _mc31_c; \
-    c->PFX ## _pixels_tab[IDX][ 8] = ff_ ## PFX ## NUM ## _mc02_c; \
-    c->PFX ## _pixels_tab[IDX][ 9] = ff_ ## PFX ## NUM ## _mc12_c; \
-    c->PFX ## _pixels_tab[IDX][10] = ff_ ## PFX ## NUM ## _mc22_c; \
-    c->PFX ## _pixels_tab[IDX][11] = ff_ ## PFX ## NUM ## _mc32_c; \
-    c->PFX ## _pixels_tab[IDX][12] = ff_ ## PFX ## NUM ## _mc03_c; \
-    c->PFX ## _pixels_tab[IDX][13] = ff_ ## PFX ## NUM ## _mc13_c; \
-    c->PFX ## _pixels_tab[IDX][14] = ff_ ## PFX ## NUM ## _mc23_c; \
-    c->PFX ## _pixels_tab[IDX][15] = ff_ ## PFX ## NUM ## _mc33_c
+    c->PFX ## _pixels_tab[IDX][ 0] = PFX ## NUM ## _mc00_c; \
+    c->PFX ## _pixels_tab[IDX][ 1] = PFX ## NUM ## _mc10_c; \
+    c->PFX ## _pixels_tab[IDX][ 2] = PFX ## NUM ## _mc20_c; \
+    c->PFX ## _pixels_tab[IDX][ 3] = PFX ## NUM ## _mc30_c; \
+    c->PFX ## _pixels_tab[IDX][ 4] = PFX ## NUM ## _mc01_c; \
+    c->PFX ## _pixels_tab[IDX][ 5] = PFX ## NUM ## _mc11_c; \
+    c->PFX ## _pixels_tab[IDX][ 6] = PFX ## NUM ## _mc21_c; \
+    c->PFX ## _pixels_tab[IDX][ 7] = PFX ## NUM ## _mc31_c; \
+    c->PFX ## _pixels_tab[IDX][ 8] = PFX ## NUM ## _mc02_c; \
+    c->PFX ## _pixels_tab[IDX][ 9] = PFX ## NUM ## _mc12_c; \
+    c->PFX ## _pixels_tab[IDX][10] = PFX ## NUM ## _mc22_c; \
+    c->PFX ## _pixels_tab[IDX][11] = PFX ## NUM ## _mc32_c; \
+    c->PFX ## _pixels_tab[IDX][12] = PFX ## NUM ## _mc03_c; \
+    c->PFX ## _pixels_tab[IDX][13] = PFX ## NUM ## _mc13_c; \
+    c->PFX ## _pixels_tab[IDX][14] = PFX ## NUM ## _mc23_c; \
+    c->PFX ## _pixels_tab[IDX][15] = PFX ## NUM ## _mc33_c
     dspfunc(put_cavs_qpel, 0, 16);
     dspfunc(put_cavs_qpel, 1, 8);
     dspfunc(avg_cavs_qpel, 0, 16);
@@ -541,4 +561,8 @@ void ff_cavsdsp_init(DSPContext* c, AVCodecContext *avctx) {
     c->cavs_filter_cv = cavs_filter_cv_c;
     c->cavs_filter_ch = cavs_filter_ch_c;
     c->cavs_idct8_add = cavs_idct8_add_c;
+    c->idct_perm = FF_IDCT_PERM_NONE;
+
+    if (ARCH_X86)
+        ff_cavsdsp_init_x86(c, avctx);
 }