removed the idct rtcd idct calls

More cleanup to do after this,  but this is a good chunk of removing rtcd.

Change-Id: I551db75e341a0a85c3ad650df1e9a60dc305681a

test/fdct4x4_test.cc

@@ -15,7 +15,6 @@
 #include "third_party/googletest/src/include/gtest/gtest.h"
 
 extern "C" {
-#include "vp9/common/idct.h"
 #include "vp9_rtcd.h"
 }
 

test/fdct8x8_test.cc

@@ -15,7 +15,6 @@
 #include "third_party/googletest/src/include/gtest/gtest.h"
 
 extern "C" {
-#include "vp9/common/idct.h"
 #include "vp9_rtcd.h"
 }
 

test/idct8x8_test.cc

@@ -15,7 +15,6 @@
 #include "third_party/googletest/src/include/gtest/gtest.h"
 
 extern "C" {
-#include "vp9/common/idct.h"
 #include "vp9_rtcd.h"
 }
 

vp8/common/ppc/systemdependent.c

@@ -12,13 +12,8 @@
 #include "subpixel.h"
 #include "loopfilter.h"
 #include "recon.h"
-#include "idct.h"
 #include "onyxc_int.h"
 
-void (*vp8_short_idct4x4)(short *input, short *output, int pitch);
-void (*vp8_short_idct4x4_1)(short *input, short *output, int pitch);
-void (*vp8_dc_only_idct)(short input_dc, short *output, int pitch);
-
 extern void (*vp8_post_proc_down_and_across_mb_row)(
     unsigned char *src_ptr,
     unsigned char *dst_ptr,

vp9/common/arm/arm_systemdependent.c

@@ -15,7 +15,6 @@
 #include "vp9/common/subpixel.h"
 #include "vp9/common/loopfilter.h"
 #include "vp9/common/recon.h"
-#include "vp9/common/idct.h"
 #include "vp9/common/onyxc_int.h"
 
 void vp9_arch_arm_common_init(VP9_COMMON *ctx) {

vp9/common/blockd.h

@@ -145,6 +145,11 @@ typedef enum {
 
 #define VP9_MVREFS (1 + SPLITMV - NEARESTMV)
 
+#if CONFIG_LOSSLESS
+#define WHT_UPSCALE_FACTOR 3
+#define Y2_WHT_UPSCALE_FACTOR 2
+#endif
+
 typedef enum {
   B_DC_PRED,          /* average of above and left pixels */
   B_TM_PRED,
@@ -370,6 +375,14 @@ typedef struct macroblockd {
 
   unsigned int frames_since_golden;
   unsigned int frames_till_alt_ref_frame;
+
+  /* Inverse transform function pointers. */
+  void (*inv_xform4x4_1_x8)(short *input, short *output, int pitch);
+  void (*inv_xform4x4_x8)(short *input, short *output, int pitch);
+  void (*inv_walsh4x4_1)(short *in, short *out);
+  void (*inv_walsh4x4_lossless)(short *in, short *out);
+
+
   vp9_subpix_fn_t  subpixel_predict;
   vp9_subpix_fn_t  subpixel_predict8x4;
   vp9_subpix_fn_t  subpixel_predict8x8;

vp9/common/generic/systemdependent.c

@@ -13,7 +13,6 @@
 #include "vp9_rtcd.h"
 #include "vp9/common/subpixel.h"
 #include "vp9/common/loopfilter.h"
-#include "vp9/common/idct.h"
 #include "vp9/common/onyxc_int.h"
 
 extern void vp9_arch_x86_common_init(VP9_COMMON *ctx);
@@ -23,18 +22,6 @@ void vp9_machine_specific_config(VP9_COMMON *ctx) {
 #if CONFIG_RUNTIME_CPU_DETECT
   VP9_COMMON_RTCD *rtcd = &ctx->rtcd;
 
-  rtcd->idct.idct1        = vp9_short_idct4x4llm_1_c;
-  rtcd->idct.idct16       = vp9_short_idct4x4llm_c;
-  rtcd->idct.idct1_scalar_add = vp9_dc_only_idct_add_c;
-  rtcd->idct.iwalsh1      = vp9_short_inv_walsh4x4_1_c;
-  rtcd->idct.iwalsh16     = vp9_short_inv_walsh4x4_c;
-  rtcd->idct.idct8        = vp9_short_idct8x8_c;
-  rtcd->idct.idct10_8     = vp9_short_idct10_8x8_c;
-  rtcd->idct.idct1_scalar_add_8x8 = vp9_dc_only_idct_add_8x8_c;
-  rtcd->idct.ihaar2       = vp9_short_ihaar2x2_c;
-  rtcd->idct.idct16x16    = vp9_short_idct16x16_c;
-  rtcd->idct.idct10_16x16 = vp9_short_idct10_16x16_c;
-
 #if CONFIG_POSTPROC || (CONFIG_VP9_ENCODER && CONFIG_INTERNAL_STATS)
   rtcd->postproc.down             = vp9_mbpost_proc_down_c;
   rtcd->postproc.across           = vp9_mbpost_proc_across_ip_c;

vp9/common/idct.h

-/*
- *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
- *
- *  Use of this source code is governed by a BSD-style license
- *  that can be found in the LICENSE file in the root of the source
- *  tree. An additional intellectual property rights grant can be found
- *  in the file PATENTS.  All contributing project authors may
- *  be found in the AUTHORS file in the root of the source tree.
- */
-
-
-#ifndef __INC_IDCT_H
-#define __INC_IDCT_H
-
-#include "vp9/common/blockd.h"
-
-#define prototype_second_order(sym) \
-  void sym(short *input, short *output)
-
-#define prototype_idct(sym) \
-  void sym(short *input, short *output, int pitch)
-
-#define prototype_idct_scalar_add(sym) \
-  void sym(short input, \
-           unsigned char *pred, unsigned char *output, \
-           int pitch, int stride)
-
-#if ARCH_X86 || ARCH_X86_64
-#include "x86/idct_x86.h"
-#endif
-
-#ifdef _MSC_VER
-/* TODO: remove these after integer implmementations are done */
-#define M_PI       3.14159265358979323846
-#define round(x) (((x)>0)? floor((x)+0.5): ceil((x)-0.5))
-#endif
-
-
-#if ARCH_ARM
-#include "arm/idct_arm.h"
-#endif
-
-#if CONFIG_LOSSLESS
-#define WHT_UPSCALE_FACTOR 3
-#define Y2_WHT_UPSCALE_FACTOR 2
-#endif
-
-#ifndef vp9_idct_idct16x16
-#define vp9_idct_idct16x16 vp9_short_idct16x16_c
-#endif
-extern prototype_idct(vp9_idct_idct16x16);
-
-#ifndef vp9_idct_idct10_16x16
-#define vp9_idct_idct10_16x16 vp9_short_idct10_16x16_c
-#endif
-extern prototype_idct(vp9_idct_idct10_16x16);
-
-#ifndef vp9_idct_idct8
-#define vp9_idct_idct8 vp9_short_idct8x8_c
-#endif
-extern prototype_idct(vp9_idct_idct8);
-
-#ifndef vp9_idct_idct10_8
-#define vp9_idct_idct10_8 vp9_short_idct10_8x8_c
-#endif
-extern prototype_idct(vp9_idct_idct10_8);
-
-#ifndef vp9_idct_idct8_1
-#define vp9_idct_idct8_1 vp9_short_idct8x8_1_c
-#endif
-extern prototype_idct(vp9_idct_idct8_1);
-
-#ifndef vp9_idct_ihaar2
-#define vp9_idct_ihaar2 vp9_short_ihaar2x2_c
-#endif
-extern prototype_idct(vp9_idct_ihaar2);
-
-#ifndef vp9_idct_ihaar2_1
-#define vp9_idct_ihaar2_1 vp9_short_ihaar2x2_1_c
-#endif
-extern prototype_idct(vp9_idct_ihaar2_1);
-
-#ifndef vp9_idct_idct1_scalar_add_8x8
-#define vp9_idct_idct1_scalar_add_8x8 vp9_dc_only_idct_add_8x8_c
-#endif
-extern prototype_idct_scalar_add(vp9_idct_idct1_scalar_add_8x8);
-
-
-
-#ifndef vp9_idct_idct1
-#define vp9_idct_idct1 vp9_short_idct4x4llm_1_c
-#endif
-extern prototype_idct(vp9_idct_idct1);
-
-#ifndef vp9_idct_idct16
-#define vp9_idct_idct16 vp9_short_idct4x4llm_c
-#endif
-extern prototype_idct(vp9_idct_idct16);
-
-#ifndef vp9_idct_idct1_scalar_add
-#define vp9_idct_idct1_scalar_add vp9_dc_only_idct_add_c
-#endif
-extern prototype_idct_scalar_add(vp9_idct_idct1_scalar_add);
-
-
-#ifndef vp9_idct_iwalsh1
-#define vp9_idct_iwalsh1 vp9_short_inv_walsh4x4_1_c
-#endif
-extern prototype_second_order(vp9_idct_iwalsh1);
-
-#ifndef vp9_idct_iwalsh16
-#define vp9_idct_iwalsh16 vp9_short_inv_walsh4x4_c
-#endif
-extern prototype_second_order(vp9_idct_iwalsh16);
-
-#if CONFIG_LOSSLESS
-extern prototype_idct(vp9_short_inv_walsh4x4_x8_c);
-extern prototype_idct(vp9_short_inv_walsh4x4_1_x8_c);
-extern prototype_idct_scalar_add(vp9_dc_only_inv_walsh_add_c);
-extern prototype_second_order(vp9_short_inv_walsh4x4_lossless_c);
-extern prototype_second_order(vp9_short_inv_walsh4x4_1_lossless_c);
-#endif
-
-void vp9_ihtllm_c(const int16_t *input, int16_t *output, int pitch,
-                  TX_TYPE tx_type, int tx_dim);
-
-typedef prototype_idct((*vp9_idct_fn_t));
-typedef prototype_idct_scalar_add((*vp9_idct_scalar_add_fn_t));
-typedef prototype_second_order((*vp9_second_order_fn_t));
-
-typedef struct {
-  vp9_idct_fn_t            idct1;
-  vp9_idct_fn_t            idct16;
-  vp9_idct_scalar_add_fn_t idct1_scalar_add;
-
-  vp9_second_order_fn_t iwalsh1;
-  vp9_second_order_fn_t iwalsh16;
-
-  vp9_idct_fn_t            idct8;
-  vp9_idct_fn_t            idct10_8;
-  vp9_idct_fn_t            idct8_1;
-  vp9_idct_scalar_add_fn_t idct1_scalar_add_8x8;
-  vp9_idct_fn_t ihaar2;
-  vp9_idct_fn_t ihaar2_1;
-
-  vp9_idct_fn_t            idct16x16;
-  vp9_idct_fn_t            idct10_16x16;
-} vp9_idct_rtcd_vtable_t;
-
-#if CONFIG_RUNTIME_CPU_DETECT
-#define IDCT_INVOKE(ctx,fn) (ctx)->fn
-#else
-#define IDCT_INVOKE(ctx,fn) vp9_idct_##fn
-#endif
-
-#endif

vp9/common/idctllm.c

@@ -25,7 +25,6 @@
 #include <assert.h>
 #include <math.h>
 #include "vpx_ports/config.h"
-#include "vp9/common/idct.h"
 #include "vp9/common/systemdependent.h"
 
 #include "vp9/common/blockd.h"

vp9/common/invtrans.c

@@ -9,6 +9,7 @@
  */
 
 #include "invtrans.h"
+#include "./vp9_rtcd.h"
 
 static void recon_dcblock(MACROBLOCKD *xd) {
   BLOCKD *b = &xd->block[24];
@@ -28,108 +29,92 @@ static void recon_dcblock_8x8(MACROBLOCKD *xd) {
   xd->block[12].dqcoeff[0] = b->diff[8];
 }
 
-void vp9_inverse_transform_b_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                 BLOCKD *b, int pitch) {
+void vp9_inverse_transform_b_4x4(MACROBLOCKD *xd, int block, int pitch) {
+  BLOCKD *b = &xd->block[block];
   if (b->eob <= 1)
-    IDCT_INVOKE(rtcd, idct1)(b->dqcoeff, b->diff, pitch);
+    xd->inv_xform4x4_1_x8(b->dqcoeff, b->diff, pitch);
   else
-    IDCT_INVOKE(rtcd, idct16)(b->dqcoeff, b->diff, pitch);
+    xd->inv_xform4x4_x8(b->dqcoeff, b->diff, pitch);
 }
 
-void vp9_inverse_transform_mby_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                   MACROBLOCKD *xd) {
+void vp9_inverse_transform_mby_4x4(MACROBLOCKD *xd) {
   int i;
   BLOCKD *blockd = xd->block;
 
   if (xd->mode_info_context->mbmi.mode != SPLITMV) {
     /* do 2nd order transform on the dc block */
-    IDCT_INVOKE(rtcd, iwalsh16)(blockd[24].dqcoeff, blockd[24].diff);
+    vp9_short_inv_walsh4x4(blockd[24].dqcoeff, blockd[24].diff);
     recon_dcblock(xd);
   }
 
   for (i = 0; i < 16; i++) {
-    vp9_inverse_transform_b_4x4(rtcd, &blockd[i], 32);
+    vp9_inverse_transform_b_4x4(xd, i, 32);
   }
 }
 
-void vp9_inverse_transform_mbuv_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                    MACROBLOCKD *xd) {
+void vp9_inverse_transform_mbuv_4x4(MACROBLOCKD *xd) {
   int i;
-  BLOCKD *blockd = xd->block;
-
   for (i = 16; i < 24; i++) {
-    vp9_inverse_transform_b_4x4(rtcd, &blockd[i], 16);
+    vp9_inverse_transform_b_4x4(xd, i, 16);
   }
 }
 
-void vp9_inverse_transform_mb_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                  MACROBLOCKD *xd) {
-  vp9_inverse_transform_mby_4x4(rtcd, xd);
-  vp9_inverse_transform_mbuv_4x4(rtcd, xd);
+void vp9_inverse_transform_mb_4x4(MACROBLOCKD *xd) {
+  vp9_inverse_transform_mby_4x4(xd);
+  vp9_inverse_transform_mbuv_4x4(xd);
 }
 
-void vp9_inverse_transform_b_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                 short *input_dqcoeff, short *output_coeff,
+void vp9_inverse_transform_b_8x8(short *input_dqcoeff, short *output_coeff,
                                  int pitch) {
-  // int b,i;
-  // if (b->eob > 1)
-  IDCT_INVOKE(rtcd, idct8)(input_dqcoeff, output_coeff, pitch);
-  // else
-  // IDCT_INVOKE(rtcd, idct8_1)(b->dqcoeff, b->diff, pitch);//pitch
+  vp9_short_idct8x8(input_dqcoeff, output_coeff, pitch);
 }
 
-void vp9_inverse_transform_mby_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                   MACROBLOCKD *xd) {
+void vp9_inverse_transform_mby_8x8(MACROBLOCKD *xd) {
   int i;
   BLOCKD *blockd = xd->block;
 
   if (xd->mode_info_context->mbmi.mode != SPLITMV) {
     // do 2nd order transform on the dc block
-    IDCT_INVOKE(rtcd, ihaar2)(blockd[24].dqcoeff, blockd[24].diff, 8);
+    vp9_short_ihaar2x2(blockd[24].dqcoeff, blockd[24].diff, 8);
     recon_dcblock_8x8(xd); // need to change for 8x8
   }
 
   for (i = 0; i < 9; i += 8) {
-    vp9_inverse_transform_b_8x8(rtcd, &blockd[i].dqcoeff[0],
+    vp9_inverse_transform_b_8x8(&blockd[i].dqcoeff[0],
                                 &blockd[i].diff[0], 32);
   }
   for (i = 2; i < 11; i += 8) {
-    vp9_inverse_transform_b_8x8(rtcd, &blockd[i + 2].dqcoeff[0],
+    vp9_inverse_transform_b_8x8(&blockd[i + 2].dqcoeff[0],
                                 &blockd[i].diff[0], 32);
   }
 }
 
-void vp9_inverse_transform_mbuv_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                    MACROBLOCKD *xd) {
+void vp9_inverse_transform_mbuv_8x8(MACROBLOCKD *xd) {
   int i;
   BLOCKD *blockd = xd->block;
 
   for (i = 16; i < 24; i += 4) {
-    vp9_inverse_transform_b_8x8(rtcd, &blockd[i].dqcoeff[0],
+    vp9_inverse_transform_b_8x8(&blockd[i].dqcoeff[0],
                                 &blockd[i].diff[0], 16);
   }
 }
 
-void vp9_inverse_transform_mb_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                  MACROBLOCKD *xd) {
-  vp9_inverse_transform_mby_8x8(rtcd, xd);
-  vp9_inverse_transform_mbuv_8x8(rtcd, xd);
+void vp9_inverse_transform_mb_8x8(MACROBLOCKD *xd) {
+  vp9_inverse_transform_mby_8x8(xd);
+  vp9_inverse_transform_mbuv_8x8(xd);
 }
 
-void vp9_inverse_transform_b_16x16(const vp9_idct_rtcd_vtable_t *rtcd,
-                                   short *input_dqcoeff,
+void vp9_inverse_transform_b_16x16(short *input_dqcoeff,
                                    short *output_coeff, int pitch) {
-  IDCT_INVOKE(rtcd, idct16x16)(input_dqcoeff, output_coeff, pitch);
+  vp9_short_idct16x16(input_dqcoeff, output_coeff, pitch);
 }
 
-void vp9_inverse_transform_mby_16x16(const vp9_idct_rtcd_vtable_t *rtcd,
-                                     MACROBLOCKD *xd) {
-  vp9_inverse_transform_b_16x16(rtcd, &xd->block[0].dqcoeff[0],
+void vp9_inverse_transform_mby_16x16(MACROBLOCKD *xd) {
+  vp9_inverse_transform_b_16x16(&xd->block[0].dqcoeff[0],
                                 &xd->block[0].diff[0], 32);
 }
 
-void vp9_inverse_transform_mb_16x16(const vp9_idct_rtcd_vtable_t *rtcd,
-                                    MACROBLOCKD *xd) {
-  vp9_inverse_transform_mby_16x16(rtcd, xd);
-  vp9_inverse_transform_mbuv_8x8(rtcd, xd);
+void vp9_inverse_transform_mb_16x16(MACROBLOCKD *xd) {
+  vp9_inverse_transform_mby_16x16(xd);
+  vp9_inverse_transform_mbuv_8x8(xd);
 }

vp9/common/invtrans.h

@@ -12,42 +12,30 @@
 #define __INC_INVTRANS_H
 
 #include "vpx_ports/config.h"
-#include "idct.h"
 #include "blockd.h"
 
-extern void vp9_inverse_transform_b_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                        BLOCKD *b, int pitch);
+extern void vp9_inverse_transform_b_4x4(MACROBLOCKD *xd, int block, int pitch);
 
-extern void vp9_inverse_transform_mb_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                         MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mb_4x4(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_mby_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                          MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mby_4x4(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_mbuv_4x4(const vp9_idct_rtcd_vtable_t *rtcd,
-                                           MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mbuv_4x4(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_b_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                        short *input_dqcoeff,
+extern void vp9_inverse_transform_b_8x8(short *input_dqcoeff,
                                         short *output_coeff, int pitch);
 
-extern void vp9_inverse_transform_mb_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                         MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mb_8x8(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_mby_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                          MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mby_8x8(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_mbuv_8x8(const vp9_idct_rtcd_vtable_t *rtcd,
-                                           MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mbuv_8x8(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_b_16x16(const vp9_idct_rtcd_vtable_t *rtcd,
-                                          short *input_dqcoeff,
+extern void vp9_inverse_transform_b_16x16(short *input_dqcoeff,
                                           short *output_coeff, int pitch);
 
-extern void vp9_inverse_transform_mb_16x16(const vp9_idct_rtcd_vtable_t *rtcd,
-                                           MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mb_16x16(MACROBLOCKD *xd);
 
-extern void vp9_inverse_transform_mby_16x16(const vp9_idct_rtcd_vtable_t *rtcd,
-                                            MACROBLOCKD *xd);
+extern void vp9_inverse_transform_mby_16x16(MACROBLOCKD *xd);
 
 #endif  // __INC_INVTRANS_H

vp9/common/onyxc_int.h

@@ -19,7 +19,6 @@
 #include "entropymv.h"
 #include "entropy.h"
 #include "entropymode.h"
-#include "idct.h"
 #if CONFIG_POSTPROC
 #include "postproc.h"
 #endif
@@ -148,7 +147,6 @@ typedef enum {
 
 typedef struct VP9_COMMON_RTCD {
 #if CONFIG_RUNTIME_CPU_DETECT
-  vp9_idct_rtcd_vtable_t        idct;
 #if CONFIG_POSTPROC
   vp9_postproc_rtcd_vtable_t    postproc;
 #endif

vp9/common/ppc/systemdependent.c

@@ -11,7 +11,6 @@
 #include "subpixel.h"
 #include "loopfilter.h"
 #include "recon.h"
-#include "idct.h"
 #include "onyxc_int.h"
 
 void (*vp8_short_idct4x4)(short *input, short *output, int pitch);

vp9/common/rtcd_defs.sh

@@ -227,6 +227,9 @@ specialize vp9_sad16x3 sse2
 prototype unsigned int vp9_sad3x16 "const unsigned char *src_ptr, int  src_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"
 specialize vp9_sad3x16 sse2
 
+#
+# Sub Pixel Filters
+#
 prototype void vp9_eighttap_predict16x16 "unsigned char *src_ptr, int  src_pixels_per_line, int  xoffset, int  yoffset, unsigned char *dst_ptr, int  dst_pitch"
 specialize vp9_eighttap_predict16x16
 
@@ -311,6 +314,57 @@ specialize vp9_bilinear_predict4x4 mmx
 prototype void vp9_bilinear_predict_avg4x4 "unsigned char *src_ptr, int  src_pixels_per_line, int  xoffset, int  yoffset, unsigned char *dst_ptr, int  dst_pitch"
 specialize vp9_bilinear_predict_avg4x4
 
+#
+# dct
+#
+prototype void vp9_short_idct4x4llm_1 "short *input, short *output, int pitch"
+specialize vp9_short_idct4x4llm_1 mmx
+
+prototype void vp9_short_idct4x4llm "short *input, short *output, int pitch"
+specialize vp9_short_idct4x4llm mmx
+
+prototype void vp9_short_idct8x8 "short *input, short *output, int pitch"
+specialize vp9_short_idct8x8
+
+prototype void vp9_short_idct10_8x8 "short *input, short *output, int pitch"
+specialize vp9_short_idct10_8x8
+
+prototype void vp9_short_ihaar2x2 "short *input, short *output, int pitch"
+specialize vp9_short_ihaar2x2
+
+prototype void vp9_short_idct16x16 "short *input, short *output, int pitch"
+specialize vp9_short_idct16x16
+
+prototype void vp9_short_idct10_16x16 "short *input, short *output, int pitch"
+specialize vp9_short_idct10_16x16
+
+#
+# 2nd order
+#
+prototype void vp9_short_inv_walsh4x4_1 "short *in, short *out"
+specialize vp9_short_inv_walsh4x4_1
+
+prototype void vp9_short_inv_walsh4x4 "short *in, short *out"
+specialize vp9_short_inv_walsh4x4_
+
+
+# dct and add
+prototype void vp9_dc_only_idct_add_8x8 "short input_dc, unsigned char *pred_ptr, unsigned char *dst_ptr, int pitch, int stride"
+specialize vp9_dc_only_idct_add_8x8
+
+prototype void vp9_dc_only_idct_add "short input_dc, unsigned char *pred_ptr, unsigned char *dst_ptr, int pitch, int stride"
+specialize vp9_dc_only_idct_add
+
+if [ "$CONFIG_LOSSLESS" = "yes" ]; then
+prototype void vp9_short_inv_walsh4x4_1_x8 "short *input, short *output, int pitch"
+prototype void vp9_short_inv_walsh4x4_x8 "short *input, short *output, int pitch"
+prototype void vp9_dc_only_inv_walsh_add "short input_dc, unsigned char *pred_ptr, unsigned char *dst_ptr, int pitch, int stride"
+prototype void vp9_short_inv_walsh4x4_1_lossless "short *in, short *out"
+prototype void vp9_short_inv_walsh4x4_lossless "short *in, short *out"
+fi
+
+
+
 if [ "$CONFIG_SUPERBLOCKS" = "yes" ]; then
 
 prototype unsigned int vp9_sad32x3 "const unsigned char *src_ptr, int  src_stride, const unsigned char *ref_ptr, int ref_stride, int max_sad"

vp9/common/x86/x86_systemdependent.c

@@ -11,7 +11,6 @@
 #include "vpx_config.h"
 #include "vpx_ports/x86.h"
 #include "vp9/common/loopfilter.h"
-#include "vp9/common/idct.h"
 #include "vp9/common/pragmas.h"
 #include "vp9/common/onyxc_int.h"
 
@@ -31,11 +30,6 @@ void vp9_arch_x86_common_init(VP9_COMMON *ctx) {
 #if HAVE_MMX
 // The commented functions need to be re-written for vpx.
   if (flags & HAS_MMX) {
-    rtcd->idct.idct1        = vp9_short_idct4x4llm_1_mmx;
-    rtcd->idct.idct16       = vp9_short_idct4x4llm_mmx;
-    rtcd->idct.idct1_scalar_add = vp9_dc_only_idct_add_mmx;
-    // rtcd->idct.iwalsh16     = vp9_short_inv_walsh4x4_mmx;
-    // rtcd->idct.iwalsh1     = vp9_short_inv_walsh4x4_1_mmx;
 
 #if CONFIG_POSTPROC
     rtcd->postproc.down        = vp9_mbpost_proc_down_mmx;

vp9/decoder/arm/armv6/idct_blk_v6.c

@@ -9,7 +9,7 @@
  */
 
 #include "vpx_ports/config.h"
-#include "vp9/common/idct.h"
+#include "vp9/common/blockd.h"
 #include "vp9/decoder/dequantize.h"
 
 void vp8_dequant_dc_idct_add_y_block_v6(short *q, short *dq,

vp9/decoder/arm/dequantize_arm.c

@@ -11,7 +11,7 @@
 
 #include "vpx_ports/config.h"
 #include "vp9/decoder/dequantize.h"
-#include "vp9/common/idct.h"
+#include "vp9/common/blockd.h"
 #include "vpx_mem/vpx_mem.h"
 
 #if HAVE_ARMV7

vp9/decoder/arm/neon/idct_blk_neon.c

@@ -9,7 +9,7 @@
  */
 
 #include "vpx_ports/config.h"
-#include "vp9/common/idct.h"
+#include "vp9/common/blockd.h"
 #include "vp9/decoder/dequantize.h"
 
 /* place these declarations here because we don't want to maintain them

vp9/decoder/decodframe.c

@@ -27,7 +27,6 @@
 #include "vp9/common/extend.h"
 #include "vp9/common/modecont.h"
 #include "vpx_mem/vpx_mem.h"