libvpx: Pull from upstream

source/libvpx/vp9/encoder/vp9_dct.c

Index: source/libvpx/vp9/encoder/vp9_dct.c
===================================================================
--- source/libvpx/vp9/encoder/vp9_dct.c	(revision 278778)
+++ source/libvpx/vp9/encoder/vp9_dct.c	(working copy)
@@ -43,6 +43,17 @@
   output[3] = fdct_round_shift(temp2);
 }
 
+void vp9_fdct4x4_1_c(const int16_t *input, int16_t *output, int stride) {
+  int r, c;
+  int16_t sum = 0;
+  for (r = 0; r < 4; ++r)
+    for (c = 0; c < 4; ++c)
+      sum += input[r * stride + c];
+
+  output[0] = sum << 1;
+  output[1] = 0;
+}
+
 void vp9_fdct4x4_c(const int16_t *input, int16_t *output, int stride) {
   // The 2D transform is done with two passes which are actually pretty
   // similar. In the first one, we transform the columns and transpose
@@ -240,6 +251,17 @@
   output[7] = fdct_round_shift(t3);
 }
 
+void vp9_fdct8x8_1_c(const int16_t *input, int16_t *output, int stride) {
+  int r, c;
+  int16_t sum = 0;
+  for (r = 0; r < 8; ++r)
+    for (c = 0; c < 8; ++c)
+      sum += input[r * stride + c];
+
+  output[0] = sum;
+  output[1] = 0;
+}
+
 void vp9_fdct8x8_c(const int16_t *input, int16_t *final_output, int stride) {
   int i, j;
   int16_t intermediate[64];
@@ -311,6 +333,17 @@
   }
 }
 
+void vp9_fdct16x16_1_c(const int16_t *input, int16_t *output, int stride) {
+  int r, c;
+  int16_t sum = 0;
+  for (r = 0; r < 16; ++r)
+    for (c = 0; c < 16; ++c)
+      sum += input[r * stride + c];
+
+  output[0] = sum >> 1;
+  output[1] = 0;
+}
+
 void vp9_fdct16x16_c(const int16_t *input, int16_t *output, int stride) {
   // The 2D transform is done with two passes which are actually pretty
   // similar. In the first one, we transform the columns and transpose
@@ -445,20 +478,20 @@
         step3[7] = step1[7] + step2[4];
         // step 4
         temp1 = step3[1] *  -cospi_8_64 + step3[6] * cospi_24_64;
-        temp2 = step3[2] * -cospi_24_64 - step3[5] *  cospi_8_64;
+        temp2 = step3[2] * cospi_24_64 + step3[5] *  cospi_8_64;
         step2[1] = fdct_round_shift(temp1);
         step2[2] = fdct_round_shift(temp2);
-        temp1 = step3[2] * -cospi_8_64 + step3[5] * cospi_24_64;
+        temp1 = step3[2] * cospi_8_64 - step3[5] * cospi_24_64;
         temp2 = step3[1] * cospi_24_64 + step3[6] *  cospi_8_64;
         step2[5] = fdct_round_shift(temp1);
         step2[6] = fdct_round_shift(temp2);
         // step 5
         step1[0] = step3[0] + step2[1];
         step1[1] = step3[0] - step2[1];
-        step1[2] = step3[3] - step2[2];
-        step1[3] = step3[3] + step2[2];
-        step1[4] = step3[4] + step2[5];
-        step1[5] = step3[4] - step2[5];
+        step1[2] = step3[3] + step2[2];
+        step1[3] = step3[3] - step2[2];
+        step1[4] = step3[4] - step2[5];
+        step1[5] = step3[4] + step2[5];
         step1[6] = step3[7] - step2[6];
         step1[7] = step3[7] + step2[6];
         // step 6
@@ -755,10 +788,10 @@
 
   // step 4
   temp1 = step3[1] *  -cospi_8_64 + step3[6] * cospi_24_64;
-  temp2 = step3[2] * -cospi_24_64 - step3[5] *  cospi_8_64;
+  temp2 = step3[2] * cospi_24_64 + step3[5] *  cospi_8_64;
   step2[1] = fdct_round_shift(temp1);
   step2[2] = fdct_round_shift(temp2);
-  temp1 = step3[2] * -cospi_8_64 + step3[5] * cospi_24_64;
+  temp1 = step3[2] * cospi_8_64 - step3[5] * cospi_24_64;
   temp2 = step3[1] * cospi_24_64 + step3[6] *  cospi_8_64;
   step2[5] = fdct_round_shift(temp1);
   step2[6] = fdct_round_shift(temp2);
@@ -766,10 +799,10 @@
   // step 5
   step1[0] = step3[0] + step2[1];
   step1[1] = step3[0] - step2[1];
-  step1[2] = step3[3] - step2[2];
-  step1[3] = step3[3] + step2[2];
-  step1[4] = step3[4] + step2[5];
-  step1[5] = step3[4] - step2[5];
+  step1[2] = step3[3] + step2[2];
+  step1[3] = step3[3] - step2[2];
+  step1[4] = step3[4] - step2[5];
+  step1[5] = step3[4] + step2[5];
   step1[6] = step3[7] - step2[6];
   step1[7] = step3[7] + step2[6];
 
@@ -1329,6 +1362,17 @@
   output[31] = dct_32_round(step[31] * cospi_31_64 + step[16] * -cospi_1_64);
 }
 
+void vp9_fdct32x32_1_c(const int16_t *input, int16_t *output, int stride) {
+  int r, c;
+  int16_t sum = 0;
+  for (r = 0; r < 32; ++r)
+    for (c = 0; c < 32; ++c)
+      sum += input[r * stride + c];
+
+  output[0] = sum >> 3;
+  output[1] = 0;
+}
+
 void vp9_fdct32x32_c(const int16_t *input, int16_t *out, int stride) {
   int i, j;
   int output[32 * 32];