summaryrefslogtreecommitdiff
path: root/src/libffmpeg/libavcodec/ppc
diff options
context:
space:
mode:
Diffstat (limited to 'src/libffmpeg/libavcodec/ppc')
-rw-r--r--src/libffmpeg/libavcodec/ppc/Makefile.am3
-rw-r--r--src/libffmpeg/libavcodec/ppc/dsputil_altivec.c8
-rw-r--r--src/libffmpeg/libavcodec/ppc/dsputil_altivec.h10
-rw-r--r--src/libffmpeg/libavcodec/ppc/dsputil_ppc.c17
-rw-r--r--src/libffmpeg/libavcodec/ppc/dsputil_ppc.h123
-rw-r--r--src/libffmpeg/libavcodec/ppc/gmc_altivec.c169
-rw-r--r--src/libffmpeg/libavcodec/ppc/idct_altivec.c8
-rw-r--r--src/libffmpeg/libavcodec/ppc/mpegvideo_altivec.c11
-rw-r--r--src/libffmpeg/libavcodec/ppc/mpegvideo_ppc.c12
9 files changed, 334 insertions, 27 deletions
diff --git a/src/libffmpeg/libavcodec/ppc/Makefile.am b/src/libffmpeg/libavcodec/ppc/Makefile.am
index a623a96f3..bb4618713 100644
--- a/src/libffmpeg/libavcodec/ppc/Makefile.am
+++ b/src/libffmpeg/libavcodec/ppc/Makefile.am
@@ -16,6 +16,7 @@ libavcodec_ppc_src = dsputil_altivec.c \
dsputil_ppc.c \
fft_altivec.c \
idct_altivec.c \
+ gmc_altivec.c \
mpegvideo_altivec.c \
mpegvideo_ppc.c
libavcodec_ppc_dummy = libavcodec_ppc_dummy.c
@@ -29,7 +30,7 @@ endif
libavcodec_ppc_la_SOURCES = $(ppc_modules) $(libavcodec_ppc_dummy)
-noinst_HEADERS = dsputil_altivec.h
+noinst_HEADERS = dsputil_altivec.h dsputil_ppc.h
.s.lo:
$(ASCOMPILE) -o $@ `test -f $< || echo '$(srcdir)/'`$<
diff --git a/src/libffmpeg/libavcodec/ppc/dsputil_altivec.c b/src/libffmpeg/libavcodec/ppc/dsputil_altivec.c
index dc62e70f4..d48332ba9 100644
--- a/src/libffmpeg/libavcodec/ppc/dsputil_altivec.c
+++ b/src/libffmpeg/libavcodec/ppc/dsputil_altivec.c
@@ -480,7 +480,7 @@ int sse16_altivec(void *v, uint8_t *pix1, uint8_t *pix2, int line_size)
return s;
}
-int pix_sum_altivec(UINT8 * pix, int line_size)
+int pix_sum_altivec(uint8_t * pix, int line_size)
{
const vector unsigned int zero = (const vector unsigned int)vec_splat_u32(0);
vector unsigned char perm, *pixv;
@@ -513,7 +513,7 @@ int pix_sum_altivec(UINT8 * pix, int line_size)
return s;
}
-void get_pixels_altivec(DCTELEM *restrict block, const UINT8 *pixels, int line_size)
+void get_pixels_altivec(DCTELEM *restrict block, const uint8_t *pixels, int line_size)
{
int i;
vector unsigned char perm, bytes, *pixv;
@@ -539,8 +539,8 @@ void get_pixels_altivec(DCTELEM *restrict block, const UINT8 *pixels, int line_s
}
}
-void diff_pixels_altivec(DCTELEM *restrict block, const UINT8 *s1,
- const UINT8 *s2, int stride)
+void diff_pixels_altivec(DCTELEM *restrict block, const uint8_t *s1,
+ const uint8_t *s2, int stride)
{
int i;
vector unsigned char perm, bytes, *pixv;
diff --git a/src/libffmpeg/libavcodec/ppc/dsputil_altivec.h b/src/libffmpeg/libavcodec/ppc/dsputil_altivec.h
index 61dbec548..f04496d62 100644
--- a/src/libffmpeg/libavcodec/ppc/dsputil_altivec.h
+++ b/src/libffmpeg/libavcodec/ppc/dsputil_altivec.h
@@ -34,12 +34,12 @@ extern int sad8x8_altivec(void *s, uint8_t *a, uint8_t *b, int stride);
extern int pix_norm1_altivec(uint8_t *pix, int line_size);
extern int sse8_altivec(void *v, uint8_t *pix1, uint8_t *pix2, int line_size);
extern int sse16_altivec(void *v, uint8_t *pix1, uint8_t *pix2, int line_size);
-extern int pix_sum_altivec(UINT8 * pix, int line_size);
-extern void diff_pixels_altivec(DCTELEM* block, const UINT8* s1, const UINT8* s2, int stride);
-extern void get_pixels_altivec(DCTELEM* block, const UINT8 * pixels, int line_size);
+extern int pix_sum_altivec(uint8_t * pix, int line_size);
+extern void diff_pixels_altivec(DCTELEM* block, const uint8_t* s1, const uint8_t* s2, int stride);
+extern void get_pixels_altivec(DCTELEM* block, const uint8_t * pixels, int line_size);
extern void add_bytes_altivec(uint8_t *dst, uint8_t *src, int w);
-extern void put_pixels_clamped_altivec(const DCTELEM *block, UINT8 *restrict pixels, int line_size);
+extern void put_pixels_clamped_altivec(const DCTELEM *block, uint8_t *restrict pixels, int line_size);
extern void put_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h);
extern void avg_pixels16_altivec(uint8_t *block, const uint8_t *pixels, int line_size, int h);
extern void avg_pixels8_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h);
@@ -48,7 +48,7 @@ extern void put_no_rnd_pixels8_xy2_altivec(uint8_t *block, const uint8_t *pixels
extern void put_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h);
extern void put_no_rnd_pixels16_xy2_altivec(uint8_t * block, const uint8_t * pixels, int line_size, int h);
-extern void gmc1_altivec(UINT8 *dst, UINT8 *src, int stride, int h, int x16, int y16, int rounder);
+extern void gmc1_altivec(uint8_t *dst, uint8_t *src, int stride, int h, int x16, int y16, int rounder);
extern int has_altivec(void);
diff --git a/src/libffmpeg/libavcodec/ppc/dsputil_ppc.c b/src/libffmpeg/libavcodec/ppc/dsputil_ppc.c
index c502f5819..374309e37 100644
--- a/src/libffmpeg/libavcodec/ppc/dsputil_ppc.c
+++ b/src/libffmpeg/libavcodec/ppc/dsputil_ppc.c
@@ -25,6 +25,9 @@
#include "dsputil_altivec.h"
#endif
+extern void idct_put_altivec(uint8_t *dest, int line_size, int16_t *block);
+extern void idct_add_altivec(uint8_t *dest, int line_size, int16_t *block);
+
int mm_flags = 0;
int mm_support(void)
@@ -169,7 +172,7 @@ long check_dcbz_effect(void)
return count;
}
-void dsputil_init_ppc(DSPContext* c, unsigned mask)
+void dsputil_init_ppc(DSPContext* c, AVCodecContext *avctx)
{
// Common optimisations whether Altivec or not
@@ -215,6 +218,18 @@ void dsputil_init_ppc(DSPContext* c, unsigned mask)
c->put_no_rnd_pixels_tab[0][3] = put_no_rnd_pixels16_xy2_altivec;
c->gmc1 = gmc1_altivec;
+
+ if ((avctx->idct_algo == FF_IDCT_AUTO) ||
+ (avctx->idct_algo == FF_IDCT_ALTIVEC))
+ {
+ c->idct_put = idct_put_altivec;
+ c->idct_add = idct_add_altivec;
+#ifndef ALTIVEC_USE_REFERENCE_C_CODE
+ c->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;
+#else /* ALTIVEC_USE_REFERENCE_C_CODE */
+ c->idct_permutation_type = FF_NO_IDCT_PERM;
+#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
+ }
#ifdef POWERPC_TBL_PERFORMANCE_REPORT
{
diff --git a/src/libffmpeg/libavcodec/ppc/dsputil_ppc.h b/src/libffmpeg/libavcodec/ppc/dsputil_ppc.h
new file mode 100644
index 000000000..ef1481a99
--- /dev/null
+++ b/src/libffmpeg/libavcodec/ppc/dsputil_ppc.h
@@ -0,0 +1,123 @@
+/*
+ * Copyright (c) 2003 Romain Dolbeau <romain@dolbeau.org>
+ *
+ * This library 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.
+ *
+ * This library 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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#ifndef _DSPUTIL_PPC_
+#define _DSPUTIL_PPC_
+
+#ifdef POWERPC_TBL_PERFORMANCE_REPORT
+void powerpc_display_perf_report(void);
+/* if you add to the enum below, also add to the perfname array
+ in dsputil_ppc.c */
+enum powerpc_perf_index {
+ altivec_fft_num = 0,
+ altivec_gmc1_num,
+ altivec_dct_unquantize_h263_num,
+ altivec_idct_add_num,
+ altivec_idct_put_num,
+ altivec_put_pixels16_num,
+ altivec_avg_pixels16_num,
+ altivec_avg_pixels8_num,
+ altivec_put_pixels8_xy2_num,
+ altivec_put_no_rnd_pixels8_xy2_num,
+ altivec_put_pixels16_xy2_num,
+ altivec_put_no_rnd_pixels16_xy2_num,
+ powerpc_clear_blocks_dcbz32,
+ powerpc_perf_total
+};
+enum powerpc_data_index {
+ powerpc_data_min = 0,
+ powerpc_data_max,
+ powerpc_data_sum,
+ powerpc_data_num,
+ powerpc_data_total
+};
+extern unsigned long long perfdata[powerpc_perf_total][powerpc_data_total];
+#ifdef POWERPC_PERF_USE_PMC
+extern unsigned long long perfdata_miss[powerpc_perf_total][powerpc_data_total];
+#endif
+
+#ifndef POWERPC_PERF_USE_PMC
+#define POWERPC_GET_CYCLES(a) asm volatile("mftb %0" : "=r" (a))
+#define POWERPC_TBL_DECLARE(a, cond) register unsigned long tbl_start, tbl_stop
+#define POWERPC_TBL_START_COUNT(a, cond) do { POWERPC_GET_CYCLES(tbl_start); } while (0)
+#define POWERPC_TBL_STOP_COUNT(a, cond) do { \
+ POWERPC_GET_CYCLES(tbl_stop); \
+ if (tbl_stop > tbl_start) \
+ { \
+ unsigned long diff = tbl_stop - tbl_start; \
+ if (cond) \
+ { \
+ if (diff < perfdata[a][powerpc_data_min]) \
+ perfdata[a][powerpc_data_min] = diff; \
+ if (diff > perfdata[a][powerpc_data_max]) \
+ perfdata[a][powerpc_data_max] = diff; \
+ perfdata[a][powerpc_data_sum] += diff; \
+ perfdata[a][powerpc_data_num] ++; \
+ } \
+ } \
+} while (0)
+
+#else /* POWERPC_PERF_USE_PMC */
+#define POWERPC_GET_CYCLES(a) asm volatile("mfspr %0, 937" : "=r" (a))
+#define POWERPC_GET_MISS(a) asm volatile("mfspr %0, 938" : "=r" (a))
+#define POWERPC_TBL_DECLARE(a, cond) register unsigned long cycles_start, cycles_stop, miss_start, miss_stop
+#define POWERPC_TBL_START_COUNT(a, cond) do { POWERPC_GET_MISS(miss_start); POWERPC_GET_CYCLES(cycles_start); } while (0)
+#define POWERPC_TBL_STOP_COUNT(a, cond) do { \
+ POWERPC_GET_CYCLES(cycles_stop); \
+ POWERPC_GET_MISS(miss_stop); \
+ if (cycles_stop >= cycles_start) \
+ { \
+ unsigned long diff = \
+ cycles_stop - cycles_start; \
+ if (cond) \
+ { \
+ if (diff < perfdata[a][powerpc_data_min]) \
+ perfdata[a][powerpc_data_min] = diff; \
+ if (diff > perfdata[a][powerpc_data_max]) \
+ perfdata[a][powerpc_data_max] = diff; \
+ perfdata[a][powerpc_data_sum] += diff; \
+ perfdata[a][powerpc_data_num] ++; \
+ } \
+ } \
+ if (miss_stop >= miss_start) \
+ { \
+ unsigned long diff = \
+ miss_stop - miss_start; \
+ if (cond) \
+ { \
+ if (diff < perfdata_miss[a][powerpc_data_min]) \
+ perfdata_miss[a][powerpc_data_min] = diff; \
+ if (diff > perfdata_miss[a][powerpc_data_max]) \
+ perfdata_miss[a][powerpc_data_max] = diff; \
+ perfdata_miss[a][powerpc_data_sum] += diff; \
+ perfdata_miss[a][powerpc_data_num] ++; \
+ } \
+ } \
+} while (0)
+
+#endif /* POWERPC_PERF_USE_PMC */
+
+
+#else /* POWERPC_TBL_PERFORMANCE_REPORT */
+// those are needed to avoid empty statements.
+#define POWERPC_TBL_DECLARE(a, cond) int altivec_placeholder __attribute__ ((unused))
+#define POWERPC_TBL_START_COUNT(a, cond) do {} while (0)
+#define POWERPC_TBL_STOP_COUNT(a, cond) do {} while (0)
+#endif /* POWERPC_TBL_PERFORMANCE_REPORT */
+
+#endif /* _DSPUTIL_PPC_ */
diff --git a/src/libffmpeg/libavcodec/ppc/gmc_altivec.c b/src/libffmpeg/libavcodec/ppc/gmc_altivec.c
new file mode 100644
index 000000000..515a766d7
--- /dev/null
+++ b/src/libffmpeg/libavcodec/ppc/gmc_altivec.c
@@ -0,0 +1,169 @@
+/*
+ * GMC (Global Motion Compensation)
+ * AltiVec-enabled
+ * Copyright (c) 2003 Romain Dolbeau <romain@dolbeau.org>
+ *
+ * This library 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.
+ *
+ * This library 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., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#include "../dsputil.h"
+
+#include "dsputil_altivec.h"
+
+/*
+ altivec-enhanced gmc1. ATM this code assume stride is a multiple of 8,
+ to preserve proper dst alignement.
+*/
+void gmc1_altivec(uint8_t *dst /* align 8 */, uint8_t *src /* align1 */, int stride, int h, int x16, int y16, int rounder)
+{
+POWERPC_TBL_DECLARE(altivec_gmc1_num, h == 8);
+#ifdef ALTIVEC_USE_REFERENCE_C_CODE
+ const int A=(16-x16)*(16-y16);
+ const int B=( x16)*(16-y16);
+ const int C=(16-x16)*( y16);
+ const int D=( x16)*( y16);
+ int i;
+
+POWERPC_TBL_START_COUNT(altivec_gmc1_num, h == 8);
+
+ for(i=0; i<h; i++)
+ {
+ dst[0]= (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1] + rounder)>>8;
+ dst[1]= (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2] + rounder)>>8;
+ dst[2]= (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3] + rounder)>>8;
+ dst[3]= (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4] + rounder)>>8;
+ dst[4]= (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5] + rounder)>>8;
+ dst[5]= (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6] + rounder)>>8;
+ dst[6]= (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7] + rounder)>>8;
+ dst[7]= (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8] + rounder)>>8;
+ dst+= stride;
+ src+= stride;
+ }
+
+POWERPC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8);
+
+#else /* ALTIVEC_USE_REFERENCE_C_CODE */
+ const unsigned short __attribute__ ((aligned(16))) rounder_a[8] =
+ {rounder, rounder, rounder, rounder,
+ rounder, rounder, rounder, rounder};
+ const unsigned short __attribute__ ((aligned(16))) ABCD[8] =
+ {
+ (16-x16)*(16-y16), /* A */
+ ( x16)*(16-y16), /* B */
+ (16-x16)*( y16), /* C */
+ ( x16)*( y16), /* D */
+ 0, 0, 0, 0 /* padding */
+ };
+ register const vector unsigned char vczero = (const vector unsigned char)vec_splat_u8(0);
+ register const vector unsigned short vcsr8 = (const vector unsigned short)vec_splat_u16(8);
+ register vector unsigned char dstv, dstv2, src_0, src_1, srcvA, srcvB, srcvC, srcvD;
+ register vector unsigned short Av, Bv, Cv, Dv, rounderV, tempA, tempB, tempC, tempD;
+ int i;
+ unsigned long dst_odd = (unsigned long)dst & 0x0000000F;
+ unsigned long src_really_odd = (unsigned long)src & 0x0000000F;
+
+
+POWERPC_TBL_START_COUNT(altivec_gmc1_num, h == 8);
+
+ tempA = vec_ld(0, (unsigned short*)ABCD);
+ Av = vec_splat(tempA, 0);
+ Bv = vec_splat(tempA, 1);
+ Cv = vec_splat(tempA, 2);
+ Dv = vec_splat(tempA, 3);
+
+ rounderV = vec_ld(0, (unsigned short*)rounder_a);
+
+ // we'll be able to pick-up our 9 char elements
+ // at src from those 32 bytes
+ // we load the first batch here, as inside the loop
+ // we can re-use 'src+stride' from one iteration
+ // as the 'src' of the next.
+ src_0 = vec_ld(0, src);
+ src_1 = vec_ld(16, src);
+ srcvA = vec_perm(src_0, src_1, vec_lvsl(0, src));
+
+ if (src_really_odd != 0x0000000F)
+ { // if src & 0xF == 0xF, then (src+1) is properly aligned on the second vector.
+ srcvB = vec_perm(src_0, src_1, vec_lvsl(1, src));
+ }
+ else
+ {
+ srcvB = src_1;
+ }
+ srcvA = vec_mergeh(vczero, srcvA);
+ srcvB = vec_mergeh(vczero, srcvB);
+
+ for(i=0; i<h; i++)
+ {
+ dst_odd = (unsigned long)dst & 0x0000000F;
+ src_really_odd = (((unsigned long)src) + stride) & 0x0000000F;
+
+ dstv = vec_ld(0, dst);
+
+ // we we'll be able to pick-up our 9 char elements
+ // at src + stride from those 32 bytes
+ // then reuse the resulting 2 vectors srvcC and srcvD
+ // as the next srcvA and srcvB
+ src_0 = vec_ld(stride + 0, src);
+ src_1 = vec_ld(stride + 16, src);
+ srcvC = vec_perm(src_0, src_1, vec_lvsl(stride + 0, src));
+
+ if (src_really_odd != 0x0000000F)
+ { // if src & 0xF == 0xF, then (src+1) is properly aligned on the second vector.
+ srcvD = vec_perm(src_0, src_1, vec_lvsl(stride + 1, src));
+ }
+ else
+ {
+ srcvD = src_1;
+ }
+
+ srcvC = vec_mergeh(vczero, srcvC);
+ srcvD = vec_mergeh(vczero, srcvD);
+
+
+ // OK, now we (finally) do the math :-)
+ // those four instructions replaces 32 int muls & 32 int adds.
+ // isn't AltiVec nice ?
+ tempA = vec_mladd((vector unsigned short)srcvA, Av, rounderV);
+ tempB = vec_mladd((vector unsigned short)srcvB, Bv, tempA);
+ tempC = vec_mladd((vector unsigned short)srcvC, Cv, tempB);
+ tempD = vec_mladd((vector unsigned short)srcvD, Dv, tempC);
+
+ srcvA = srcvC;
+ srcvB = srcvD;
+
+ tempD = vec_sr(tempD, vcsr8);
+
+ dstv2 = vec_pack(tempD, (vector unsigned short)vczero);
+
+ if (dst_odd)
+ {
+ dstv2 = vec_perm(dstv, dstv2, vcprm(0,1,s0,s1));
+ }
+ else
+ {
+ dstv2 = vec_perm(dstv, dstv2, vcprm(s0,s1,2,3));
+ }
+
+ vec_st(dstv2, 0, dst);
+
+ dst += stride;
+ src += stride;
+ }
+
+POWERPC_TBL_STOP_COUNT(altivec_gmc1_num, h == 8);
+
+#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
+}
diff --git a/src/libffmpeg/libavcodec/ppc/idct_altivec.c b/src/libffmpeg/libavcodec/ppc/idct_altivec.c
index 1619f1731..13df78f32 100644
--- a/src/libffmpeg/libavcodec/ppc/idct_altivec.c
+++ b/src/libffmpeg/libavcodec/ppc/idct_altivec.c
@@ -176,8 +176,8 @@ void idct_put_altivec(uint8_t* dest, int stride, vector_s16_t* block)
POWERPC_TBL_DECLARE(altivec_idct_put_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
POWERPC_TBL_START_COUNT(altivec_idct_put_num, 1);
- void simple_idct_put(UINT8 *dest, int line_size, INT16 *block);
- simple_idct_put(dest, stride, (INT16*)block);
+ void simple_idct_put(uint8_t *dest, int line_size, int16_t *block);
+ simple_idct_put(dest, stride, (int16_t*)block);
POWERPC_TBL_STOP_COUNT(altivec_idct_put_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
vector_u8_t tmp;
@@ -209,8 +209,8 @@ void idct_add_altivec(uint8_t* dest, int stride, vector_s16_t* block)
POWERPC_TBL_DECLARE(altivec_idct_add_num, 1);
#ifdef ALTIVEC_USE_REFERENCE_C_CODE
POWERPC_TBL_START_COUNT(altivec_idct_add_num, 1);
- void simple_idct_add(UINT8 *dest, int line_size, INT16 *block);
- simple_idct_add(dest, stride, (INT16*)block);
+ void simple_idct_add(uint8_t *dest, int line_size, int16_t *block);
+ simple_idct_add(dest, stride, (int16_t*)block);
POWERPC_TBL_STOP_COUNT(altivec_idct_add_num, 1);
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
vector_u8_t tmp;
diff --git a/src/libffmpeg/libavcodec/ppc/mpegvideo_altivec.c b/src/libffmpeg/libavcodec/ppc/mpegvideo_altivec.c
index 6c3be0e77..95558a6be 100644
--- a/src/libffmpeg/libavcodec/ppc/mpegvideo_altivec.c
+++ b/src/libffmpeg/libavcodec/ppc/mpegvideo_altivec.c
@@ -21,7 +21,6 @@
#include "../dsputil.h"
#include "../mpegvideo.h"
#include "dsputil_altivec.h"
-#include "xineutils.h"
// Swaps two variables (used for altivec registers)
#define SWAP(a,b) \
@@ -469,7 +468,7 @@ int dct_quantize_altivec(MpegEncContext* s,
// and handle it using the vector unit if we can. This is the permute used
// by the altivec idct, so it is common when using the altivec dct.
- if ((lastNonZero > 0) && (s->idct_permutation_type == FF_TRANSPOSE_IDCT_PERM))
+ if ((lastNonZero > 0) && (s->dsp.idct_permutation_type == FF_TRANSPOSE_IDCT_PERM))
{
TRANSPOSE8(data0, data1, data2, data3, data4, data5, data6, data7);
}
@@ -502,10 +501,10 @@ int dct_quantize_altivec(MpegEncContext* s,
// We handled the tranpose permutation above and we don't
// need to permute the "no" permutation case.
if ((lastNonZero > 0) &&
- (s->idct_permutation_type != FF_TRANSPOSE_IDCT_PERM) &&
- (s->idct_permutation_type != FF_NO_IDCT_PERM))
+ (s->dsp.idct_permutation_type != FF_TRANSPOSE_IDCT_PERM) &&
+ (s->dsp.idct_permutation_type != FF_NO_IDCT_PERM))
{
- ff_block_permute(data, s->idct_permutation,
+ ff_block_permute(data, s->dsp.idct_permutation,
s->intra_scantable.scantable, lastNonZero);
}
@@ -524,7 +523,7 @@ POWERPC_TBL_DECLARE(altivec_dct_unquantize_h263_num, 1);
int i, level, qmul, qadd;
int nCoeffs;
- XINE_ASSERT(s->block_last_index[n]>=0, "s->block_last_index[%d] < 0", n);
+ assert(s->block_last_index[n]>=0);
POWERPC_TBL_START_COUNT(altivec_dct_unquantize_h263_num, 1);
diff --git a/src/libffmpeg/libavcodec/ppc/mpegvideo_ppc.c b/src/libffmpeg/libavcodec/ppc/mpegvideo_ppc.c
index 9757f5f39..c290dde16 100644
--- a/src/libffmpeg/libavcodec/ppc/mpegvideo_ppc.c
+++ b/src/libffmpeg/libavcodec/ppc/mpegvideo_ppc.c
@@ -30,8 +30,8 @@ extern int dct_quantize_altivec(MpegEncContext *s,
extern void dct_unquantize_h263_altivec(MpegEncContext *s,
DCTELEM *block, int n, int qscale);
-extern void idct_put_altivec(UINT8 *dest, int line_size, INT16 *block);
-extern void idct_add_altivec(UINT8 *dest, int line_size, INT16 *block);
+extern void idct_put_altivec(uint8_t *dest, int line_size, int16_t *block);
+extern void idct_add_altivec(uint8_t *dest, int line_size, int16_t *block);
void MPV_common_init_ppc(MpegEncContext *s)
@@ -42,12 +42,12 @@ void MPV_common_init_ppc(MpegEncContext *s)
if ((s->avctx->idct_algo == FF_IDCT_AUTO) ||
(s->avctx->idct_algo == FF_IDCT_ALTIVEC))
{
- s->idct_put = idct_put_altivec;
- s->idct_add = idct_add_altivec;
+ s->dsp.idct_put = idct_put_altivec;
+ s->dsp.idct_add = idct_add_altivec;
#ifndef ALTIVEC_USE_REFERENCE_C_CODE
- s->idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;
+ s->dsp.idct_permutation_type = FF_TRANSPOSE_IDCT_PERM;
#else /* ALTIVEC_USE_REFERENCE_C_CODE */
- s->idct_permutation_type = FF_NO_IDCT_PERM;
+ s->dsp.idct_permutation_type = FF_NO_IDCT_PERM;
#endif /* ALTIVEC_USE_REFERENCE_C_CODE */
}