multi-thread decoding: ~25-30% faster

To be enabled with the flag WEBP_USE_THREAD.
For now it's only available on unix (pthread), when using Makefile.unix
Will be switched on more generally later.

In-loop filtering and output (=rescaling/yuv->rgb conversion)
is done in parallel to bitstream decoding, lagging 1 row behind.

Example:
examples/dwebp bryce.webp -v
Time to decode picture: 0.680s

examples/dwebp bryce.webp -v -mt
Time to decode picture: 0.515s

Change-Id: Ic30a897423137a3bdace9c4e30465ef758fe53f2

Android.mk

@@ -15,6 +15,7 @@ LOCAL_SRC_FILES := \
 	src/dec/io.c \
 	src/dec/buffer.c \
 	src/dec/yuv.c \
+	src/dec/thread.c \
 	src/enc/alpha.c \
 	src/enc/analysis.c \
 	src/enc/bit_writer.c \
@@ -30,7 +31,7 @@ LOCAL_SRC_FILES := \
 	src/enc/tree.c \
 	src/enc/webpenc.c
 
-LOCAL_CFLAGS := -Wall -DANDROID -DHAVE_MALLOC_H -DHAVE_PTHREAD \
+LOCAL_CFLAGS := -Wall -DANDROID -DHAVE_MALLOC_H -DHAVE_PTHREAD -DWEBP_USE_THREAD \
                 -finline-functions -frename-registers -ffast-math \
                 -s -fomit-frame-pointer -Isrc/webp