--- mem_transfer.c	2003/11/09 20:47:14	1.9.2.3
+++ mem_transfer.c	2005/09/13 12:12:15	1.16
@@ -19,7 +19,7 @@
  *  along with this program ; if not, write to the Free Software
  *  Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307 USA
  *
- * $Id: mem_transfer.c,v 1.9.2.3 2003/11/09 20:47:14 edgomez Exp $
+ * $Id: mem_transfer.c,v 1.16 2005/09/13 12:12:15 suxen_drol Exp $
  *
  ****************************************************************************/
 
@@ -34,10 +34,13 @@
 TRANSFER_8TO16SUB_PTR  transfer_8to16sub;
 TRANSFER_8TO16SUBRO_PTR  transfer_8to16subro;
 TRANSFER_8TO16SUB2_PTR transfer_8to16sub2;
+TRANSFER_8TO16SUB2RO_PTR transfer_8to16sub2ro;
 TRANSFER_16TO8ADD_PTR  transfer_16to8add;
 
 TRANSFER8X8_COPY_PTR transfer8x8_copy;
+TRANSFER8X4_COPY_PTR transfer8x4_copy;
 
+#define USE_REFERENCE_C
 
 /*****************************************************************************
  *
@@ -64,8 +67,7 @@
 					 const uint8_t * const src,
 					 uint32_t stride)
 {
-	uint32_t i, j;
-
+	int i, j;
 	for (j = 0; j < 8; j++) {
 		for (i = 0; i < 8; i++) {
 			dst[j * 8 + i] = (int16_t) src[j * stride + i];
@@ -87,10 +89,11 @@
 					 const int16_t * const src,
 					 uint32_t stride)
 {
-	uint32_t i, j;
+	int i, j;
 
 	for (j = 0; j < 8; j++) {
 		for (i = 0; i < 8; i++) {
+#ifdef USE_REFERENCE_C
 			int16_t pixel = src[j * 8 + i];
 
 			if (pixel < 0) {
@@ -99,7 +102,12 @@
 				pixel = 255;
 			}
 			dst[j * stride + i] = (uint8_t) pixel;
-		}
+#else
+			const int16_t pixel = src[j * 8 + i];
+			const uint8_t value = (uint8_t)( (pixel&~255) ? (-pixel)>>(8*sizeof(pixel)-1) : pixel );
+			dst[j*stride + i] = value;
+#endif
+    }
 	}
 }
 
@@ -123,12 +131,12 @@
 					const uint8_t * ref,
 					const uint32_t stride)
 {
-	uint32_t i, j;
+	int i, j;
 
 	for (j = 0; j < 8; j++) {
 		for (i = 0; i < 8; i++) {
-			uint8_t c = cur[j * stride + i];
-			uint8_t r = ref[j * stride + i];
+			const uint8_t c = cur[j * stride + i];
+			const uint8_t r = ref[j * stride + i];
 
 			cur[j * stride + i] = r;
 			dct[j * 8 + i] = (int16_t) c - (int16_t) r;
@@ -143,12 +151,12 @@
 					const uint8_t * ref,
 					const uint32_t stride)
 {
-	uint32_t i, j;
+	int i, j;
 
 	for (j = 0; j < 8; j++) {
 		for (i = 0; i < 8; i++) {
-			uint8_t c = cur[j * stride + i];
-			uint8_t r = ref[j * stride + i];
+			const uint8_t c = cur[j * stride + i];
+			const uint8_t r = ref[j * stride + i];
 			dct[j * 8 + i] = (int16_t) c - (int16_t) r;
 		}
 	}
@@ -181,18 +189,32 @@
 
 	for (j = 0; j < 8; j++) {
 		for (i = 0; i < 8; i++) {
-			uint8_t c = cur[j * stride + i];
-			int r = (ref1[j * stride + i] + ref2[j * stride + i] + 1) / 2;
-
-			if (r > 255) {
-				r = 255;
-			}
+			const uint8_t c = cur[j * stride + i];
+			const uint8_t r = (ref1[j * stride + i] + ref2[j * stride + i] + 1) >> 1;
 			cur[j * stride + i] = r;
 			dct[j * 8 + i] = (int16_t) c - (int16_t) r;
 		}
 	}
 }
 
+void
+transfer_8to16sub2ro_c(int16_t * const dct,
+					 const uint8_t * const cur,
+					 const uint8_t * ref1,
+					 const uint8_t * ref2,
+					 const uint32_t stride)
+{
+	uint32_t i, j;
+
+	for (j = 0; j < 8; j++) {
+		for (i = 0; i < 8; i++) {
+			const uint8_t c = cur[j * stride + i];
+			const uint8_t r = (ref1[j * stride + i] + ref2[j * stride + i] + 1) >> 1;
+			dct[j * 8 + i] = (int16_t) c - (int16_t) r;
+		}
+	}
+}
+
 
 /*
  * SRC - the source buffer
@@ -208,10 +230,11 @@
 					const int16_t * const src,
 					uint32_t stride)
 {
-	uint32_t i, j;
+	int i, j;
 
 	for (j = 0; j < 8; j++) {
 		for (i = 0; i < 8; i++) {
+#ifdef USE_REFERENCE_C
 			int16_t pixel = (int16_t) dst[j * stride + i] + src[j * 8 + i];
 
 			if (pixel < 0) {
@@ -220,6 +243,12 @@
 				pixel = 255;
 			}
 			dst[j * stride + i] = (uint8_t) pixel;
+#else
+      const int16_t pixel = (int16_t) dst[j * stride + i] + src[j * 8 + i];
+			const uint8_t value = (uint8_t)( (pixel&~255) ? (-pixel)>>(8*sizeof(pixel)-1) : pixel );
+			dst[j*stride + i] = value;
+#endif
+
 		}
 	}
 }
@@ -238,11 +267,39 @@
 				   const uint8_t * const src,
 				   const uint32_t stride)
 {
-	uint32_t i, j;
+	int j, i;
 
-	for (j = 0; j < 8; j++) {
-		for (i = 0; i < 8; i++) {
-			dst[j * stride + i] = src[j * stride + i];
+	for (j = 0; j < 8; ++j) {
+	    uint8_t *d = dst + j*stride;
+		const uint8_t *s = src + j*stride;
+
+		for (i = 0; i < 8; ++i)
+		{
+			*d++ = *s++;
 		}
 	}
 }
+
+/*
+ * SRC - the source buffer
+ * DST - the destination buffer
+ *
+ * Then the function does the 8->8 bit transfer and this serie of operations :
+ *
+ *    SRC (8bit) = SRC
+ *    DST (8bit) = SRC
+ */
+void
+transfer8x4_copy_c(uint8_t * const dst,
+				   const uint8_t * const src,
+				   const uint32_t stride)
+{
+	uint32_t j;
+
+	for (j = 0; j < 4; j++) {
+		uint32_t *d= (uint32_t*)(dst + j*stride);
+		const uint32_t *s = (const uint32_t*)(src + j*stride);
+		*(d+0) = *(s+0);
+		*(d+1) = *(s+1);
+	}
+}