avcodec/get_bits: add cached bitstream reader

Signed-off-by: Paul B Mahol <onemda@gmail.com>

2 files changed, 367 insertions, 2 deletions

diff --git a/libavcodec/get_bits.h b/libavcodec/get_bits.h
index 56ef5f0cbe..58ebb64656 100644
--- a/libavcodec/get_bits.h
+++ b/libavcodec/get_bits.h
@@ -1,5 +1,6 @@
 /*
- * copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (c) 2004 Michael Niedermayer <michaelni@gmx.at>
+ * Copyright (c) 2016 Alexandra Hájková
  *
  * This file is part of FFmpeg.
  *
@@ -55,11 +56,19 @@
 
 typedef struct GetBitContext {
     const uint8_t *buffer, *buffer_end;
+#ifdef CACHED_BITSTREAM_READER
+    uint64_t cache;
+    unsigned bits_left;
+#endif
     int index;
     int size_in_bits;
     int size_in_bits_plus8;
 } GetBitContext;
 
+static inline unsigned int get_bits(GetBitContext *s, int n);
+static inline void skip_bits(GetBitContext *s, int n);
+static inline unsigned int show_bits(GetBitContext *s, int n);
+
 /* Bitstream reader API docs:
  * name
  *   arbitrary name which is used as prefix for the internal variables
@@ -107,12 +116,16 @@ typedef struct GetBitContext {
  * For examples see get_bits, show_bits, skip_bits, get_vlc.
  */
 
-#ifdef LONG_BITSTREAM_READER
+#ifdef CACHED_BITSTREAM_READER
+#   define MIN_CACHE_BITS 64
+#elif defined LONG_BITSTREAM_READER
 #   define MIN_CACHE_BITS 32
 #else
 #   define MIN_CACHE_BITS 25
 #endif
 
+#ifndef CACHED_BITSTREAM_READER
+
 #define OPEN_READER_NOSIZE(name, gb)            \
     unsigned int name ## _index = (gb)->index;  \
     unsigned int av_unused name ## _cache
@@ -197,11 +210,75 @@ typedef struct GetBitContext {
 
 #define GET_CACHE(name, gb) ((uint32_t) name ## _cache)
 
+#endif
+
 static inline int get_bits_count(const GetBitContext *s)
 {
+#ifdef CACHED_BITSTREAM_READER
+    return s->index - s->bits_left;
+#else
     return s->index;
+#endif
+}
+
+#ifdef CACHED_BITSTREAM_READER
+static inline void refill_32(GetBitContext *s)
+{
+#if !UNCHECKED_BITSTREAM_READER
+    if (s->index >> 3 >= s->buffer_end - s->buffer)
+        return;
+#endif
+
+#ifdef BITSTREAM_READER_LE
+    s->cache       = (uint64_t)AV_RL32(s->buffer + (s->index >> 3)) << s->bits_left | s->cache;
+#else
+    s->cache       = s->cache | (uint64_t)AV_RB32(s->buffer + (s->index >> 3)) << (32 - s->bits_left);
+#endif
+    s->index     += 32;
+    s->bits_left += 32;
 }
 
+static inline void refill_64(GetBitContext *s)
+{
+#if !UNCHECKED_BITSTREAM_READER
+    if (s->index >> 3 >= s->buffer_end - s->buffer)
+        return;
+#endif
+
+#ifdef BITSTREAM_READER_LE
+    s->cache = AV_RL64(s->buffer + (s->index >> 3));
+#else
+    s->cache = AV_RB64(s->buffer + (s->index >> 3));
+#endif
+    s->index += 64;
+    s->bits_left = 64;
+}
+
+static inline uint64_t get_val(GetBitContext *s, unsigned n, int is_le)
+{
+    uint64_t ret;
+    av_assert2(n>0 && n<=63);
+    if (is_le) {
+        ret = s->cache & ((UINT64_C(1) << n) - 1);
+        s->cache >>= n;
+    } else {
+        ret = s->cache >> (64 - n);
+        s->cache <<= n;
+    }
+    s->bits_left -= n;
+    return ret;
+}
+
+static inline unsigned show_val(const GetBitContext *s, unsigned n)
+{
+#ifdef BITSTREAM_READER_LE
+    return s->cache & ((UINT64_C(1) << n) - 1);
+#else
+    return s->cache >> (64 - n);
+#endif
+}
+#endif
+
 /**
  * Skips the specified number of bits.
  * @param n the number of bits to skip,
@@ -211,13 +288,29 @@ static inline int get_bits_count(const GetBitContext *s)
  */
 static inline void skip_bits_long(GetBitContext *s, int n)
 {
+#ifdef CACHED_BITSTREAM_READER
+    skip_bits(s, n);
+#else
 #if UNCHECKED_BITSTREAM_READER
     s->index += n;
 #else
     s->index += av_clip(n, -s->index, s->size_in_bits_plus8 - s->index);
 #endif
+#endif
 }
 
+#ifdef CACHED_BITSTREAM_READER
+static inline void skip_remaining(GetBitContext *s, unsigned n)
+{
+#ifdef BITSTREAM_READER_LE
+    s->cache >>= n;
+#else
+    s->cache <<= n;
+#endif
+    s->bits_left -= n;
+}
+#endif
+
 /**
  * Read MPEG-1 dc-style VLC (sign bit + mantissa with no MSB).
  * if MSB not set it is negative
@@ -225,6 +318,13 @@ static inline void skip_bits_long(GetBitContext *s, int n)
  */
 static inline int get_xbits(GetBitContext *s, int n)
 {
+#ifdef CACHED_BITSTREAM_READER
+    int32_t cache = show_bits(s, 32);
+    int sign = ~cache >> 31;
+    skip_remaining(s, n);
+
+    return ((((uint32_t)(sign ^ cache)) >> (32 - n)) ^ sign) - sign;
+#else
     register int sign;
     register int32_t cache;
     OPEN_READER(re, s);
@@ -235,8 +335,10 @@ static inline int get_xbits(GetBitContext *s, int n)
     LAST_SKIP_BITS(re, s, n);
     CLOSE_READER(re, s);
     return (NEG_USR32(sign ^ cache, n) ^ sign) - sign;
+#endif
 }
 
+#ifndef CACHED_BITSTREAM_READER
 static inline int get_xbits_le(GetBitContext *s, int n)
 {
     register int sign;
@@ -250,16 +352,22 @@ static inline int get_xbits_le(GetBitContext *s, int n)
     CLOSE_READER(re, s);
     return (zero_extend(sign ^ cache, n) ^ sign) - sign;
 }
+#endif
 
 static inline int get_sbits(GetBitContext *s, int n)
 {
     register int tmp;
+#ifdef CACHED_BITSTREAM_READER
+    av_assert2(n>0 && n<=25);
+    tmp = sign_extend(get_bits(s, n), n);
+#else
     OPEN_READER(re, s);
     av_assert2(n>0 && n<=25);
     UPDATE_CACHE(re, s);
     tmp = SHOW_SBITS(re, s, n);
     LAST_SKIP_BITS(re, s, n);
     CLOSE_READER(re, s);
+#endif
     return tmp;
 }
 
@@ -269,12 +377,28 @@ static inline int get_sbits(GetBitContext *s, int n)
 static inline unsigned int get_bits(GetBitContext *s, int n)
 {
     register int tmp;
+#ifdef CACHED_BITSTREAM_READER
+
+    av_assert2(n>0 && n<=32);
+    if (n > s->bits_left) {
+        refill_32(s);
+        if (s->bits_left < 32)
+            s->bits_left = n;
+    }
+
+#ifdef BITSTREAM_READER_LE
+    tmp = get_val(s, n, 1);
+#else
+    tmp = get_val(s, n, 0);
+#endif
+#else
     OPEN_READER(re, s);
     av_assert2(n>0 && n<=25);
     UPDATE_CACHE(re, s);
     tmp = SHOW_UBITS(re, s, n);
     LAST_SKIP_BITS(re, s, n);
     CLOSE_READER(re, s);
+#endif
     return tmp;
 }
 
@@ -288,6 +412,16 @@ static av_always_inline int get_bitsz(GetBitContext *s, int n)
 
 static inline unsigned int get_bits_le(GetBitContext *s, int n)
 {
+#ifdef CACHED_BITSTREAM_READER
+    av_assert2(n>0 && n<=32);
+    if (n > s->bits_left) {
+        refill_32(s);
+        if (s->bits_left < 32)
+            s->bits_left = n;
+    }
+
+    return get_val(s, n, 1);
+#else
     register int tmp;
     OPEN_READER(re, s);
     av_assert2(n>0 && n<=25);
@@ -296,6 +430,7 @@ static inline unsigned int get_bits_le(GetBitContext *s, int n)
     LAST_SKIP_BITS(re, s, n);
     CLOSE_READER(re, s);
     return tmp;
+#endif
 }
 
 /**
@@ -304,22 +439,59 @@ static inline unsigned int get_bits_le(GetBitContext *s, int n)
 static inline unsigned int show_bits(GetBitContext *s, int n)
 {
     register int tmp;
+#ifdef CACHED_BITSTREAM_READER
+    if (n > s->bits_left)
+        refill_32(s);
+
+    tmp = show_val(s, n);
+#else
     OPEN_READER_NOSIZE(re, s);
     av_assert2(n>0 && n<=25);
     UPDATE_CACHE(re, s);
     tmp = SHOW_UBITS(re, s, n);
+#endif
     return tmp;
 }
 
 static inline void skip_bits(GetBitContext *s, int n)
 {
+#ifdef CACHED_BITSTREAM_READER
+    if (n < s->bits_left)
+        skip_remaining(s, n);
+    else {
+        n -= s->bits_left;
+        s->cache = 0;
+        s->bits_left = 0;
+
+        if (n >= 64) {
+            unsigned skip = (n / 8) * 8;
+
+            n -= skip;
+            s->index += skip;
+        }
+        refill_64(s);
+        if (n)
+            skip_remaining(s, n);
+    }
+#else
     OPEN_READER(re, s);
     LAST_SKIP_BITS(re, s, n);
     CLOSE_READER(re, s);
+#endif
 }
 
 static inline unsigned int get_bits1(GetBitContext *s)
 {
+#ifdef CACHED_BITSTREAM_READER
+    if (!s->bits_left)
+        refill_64(s);
+
+#ifdef BITSTREAM_READER_LE
+    return get_val(s, 1, 1);
+#else
+    return get_val(s, 1, 0);
+#endif
+#else
     unsigned int index = s->index;
     uint8_t result     = s->buffer[index >> 3];
 #ifdef BITSTREAM_READER_LE
@@ -336,6 +508,7 @@ static inline unsigned int get_bits1(GetBitContext *s)
     s->index = index;
 
     return result;
+#endif
 }
 
 static inline unsigned int show_bits1(GetBitContext *s)
@@ -356,6 +529,10 @@ static inline unsigned int get_bits_long(GetBitContext *s, int n)
     av_assert2(n>=0 && n<=32);
     if (!n) {
         return 0;
+#ifdef CACHED_BITSTREAM_READER
+    }
+    return get_bits(s, n);
+#else
     } else if (n <= MIN_CACHE_BITS) {
         return get_bits(s, n);
     } else {
@@ -367,6 +544,7 @@ static inline unsigned int get_bits_long(GetBitContext *s, int n)
         return ret | get_bits(s, n - 16);
 #endif
     }
+#endif
 }
 
 /**
@@ -450,6 +628,10 @@ static inline int init_get_bits(GetBitContext *s, const uint8_t *buffer,
     s->buffer_end         = buffer + buffer_size;
     s->index              = 0;
 
+#ifdef CACHED_BITSTREAM_READER
+    refill_64(s);
+#endif
+
     return ret;
 }
 
@@ -551,6 +733,19 @@ static inline const uint8_t *align_get_bits(GetBitContext *s)
         SKIP_BITS(name, gb, n);                                 \
     } while (0)
 
+/* Return the LUT element for the given bitstream configuration. */
+static inline int set_idx(GetBitContext *s, int code, int *n, int *nb_bits,
+                          VLC_TYPE (*table)[2])
+{
+    unsigned idx;
+
+    *nb_bits = -*n;
+    idx = show_bits(s, *nb_bits) + code;
+    *n = table[idx][1];
+
+    return table[idx][0];
+}
+
 /**
  * Parse a vlc code.
  * @param bits is the number of bits which will be read at once, must be
@@ -563,6 +758,24 @@ static inline const uint8_t *align_get_bits(GetBitContext *s)
 static av_always_inline int get_vlc2(GetBitContext *s, VLC_TYPE (*table)[2],
                                      int bits, int max_depth)
 {
+#ifdef CACHED_BITSTREAM_READER
+    int nb_bits;
+    unsigned idx = show_bits(s, bits);
+    int code = table[idx][0];
+    int n    = table[idx][1];
+
+    if (max_depth > 1 && n < 0) {
+        skip_remaining(s, bits);
+        code = set_idx(s, code, &n, &nb_bits, table);
+        if (max_depth > 2 && n < 0) {
+            skip_remaining(s, nb_bits);
+            code = set_idx(s, code, &n, &nb_bits, table);
+        }
+    }
+    skip_remaining(s, n);
+
+    return code;
+#else
     int code;
 
     OPEN_READER(re, s);
@@ -573,6 +786,7 @@ static av_always_inline int get_vlc2(GetBitContext *s, VLC_TYPE (*table)[2],
     CLOSE_READER(re, s);
 
     return code;
+#endif
 }
 
 static inline int decode012(GetBitContext *gb)
diff --git a/libavcodec/golomb.h b/libavcodec/golomb.h
index efb1eff8aa..e66675b292 100644
--- a/libavcodec/golomb.h
+++ b/libavcodec/golomb.h
@@ -54,6 +54,23 @@ static inline int get_ue_golomb(GetBitContext *gb)
 {
     unsigned int buf;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    if (buf >= (1 << 27)) {
+        buf >>= 32 - 9;
+        skip_bits_long(gb, ff_golomb_vlc_len[buf]);
+
+        return ff_ue_golomb_vlc_code[buf];
+    } else {
+        int log = 2 * av_log2(buf) - 31;
+        buf >>= log;
+        buf--;
+        skip_bits_long(gb, 32 - log);
+
+        return buf;
+    }
+#else
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -77,6 +94,7 @@ static inline int get_ue_golomb(GetBitContext *gb)
 
         return buf;
     }
+#endif
 }
 
 /**
@@ -101,6 +119,13 @@ static inline int get_ue_golomb_31(GetBitContext *gb)
 {
     unsigned int buf;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    buf >>= 32 - 9;
+    skip_bits_long(gb, ff_golomb_vlc_len[buf]);
+#else
+
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -108,6 +133,7 @@ static inline int get_ue_golomb_31(GetBitContext *gb)
     buf >>= 32 - 9;
     LAST_SKIP_BITS(re, gb, ff_golomb_vlc_len[buf]);
     CLOSE_READER(re, gb);
+#endif
 
     return ff_ue_golomb_vlc_code[buf];
 }
@@ -116,6 +142,33 @@ static inline unsigned get_interleaved_ue_golomb(GetBitContext *gb)
 {
     uint32_t buf;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    if (buf & 0xAA800000) {
+        buf >>= 32 - 8;
+        skip_bits_long(gb, ff_interleaved_golomb_vlc_len[buf]);
+
+        return ff_interleaved_ue_golomb_vlc_code[buf];
+    } else {
+        unsigned ret = 1;
+
+        do {
+            buf >>= 32 - 8;
+            skip_bits_long(gb, FFMIN(ff_interleaved_golomb_vlc_len[buf], 8));
+
+            if (ff_interleaved_golomb_vlc_len[buf] != 9) {
+                ret <<= (ff_interleaved_golomb_vlc_len[buf] - 1) >> 1;
+                ret  |= ff_interleaved_dirac_golomb_vlc_code[buf];
+                break;
+            }
+            ret = (ret << 4) | ff_interleaved_dirac_golomb_vlc_code[buf];
+            buf = show_bits_long(gb, 32);
+        } while (get_bits_left(gb) > 0);
+
+        return ret - 1;
+    }
+#else
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -147,6 +200,7 @@ static inline unsigned get_interleaved_ue_golomb(GetBitContext *gb)
         CLOSE_READER(re, gb);
         return ret - 1;
     }
+#endif
 }
 
 /**
@@ -184,6 +238,28 @@ static inline int get_se_golomb(GetBitContext *gb)
 {
     unsigned int buf;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    if (buf >= (1 << 27)) {
+        buf >>= 32 - 9;
+        skip_bits_long(gb, ff_golomb_vlc_len[buf]);
+
+        return ff_se_golomb_vlc_code[buf];
+    } else {
+        int log = 2 * av_log2(buf) - 31;
+        buf >>= log;
+
+        skip_bits_long(gb, 32 - log);
+
+        if (buf & 1)
+            buf = -(buf >> 1);
+        else
+            buf = (buf >> 1);
+
+        return buf;
+    }
+#else
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -210,6 +286,7 @@ static inline int get_se_golomb(GetBitContext *gb)
 
         return buf;
     }
+#endif
 }
 
 static inline int get_se_golomb_long(GetBitContext *gb)
@@ -223,6 +300,30 @@ static inline int get_interleaved_se_golomb(GetBitContext *gb)
 {
     unsigned int buf;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    if (buf & 0xAA800000) {
+        buf >>= 32 - 8;
+        skip_bits_long(gb, ff_interleaved_golomb_vlc_len[buf]);
+
+        return ff_interleaved_se_golomb_vlc_code[buf];
+    } else {
+        int log;
+        skip_bits(gb, 8);
+        buf |= 1 | show_bits_long(gb, 24);
+
+        if ((buf & 0xAAAAAAAA) == 0)
+            return INVALID_VLC;
+
+        for (log = 31; (buf & 0x80000000) == 0; log--)
+            buf = (buf << 2) - ((buf << log) >> (log - 1)) + (buf >> 30);
+
+        skip_bits_long(gb, 63 - 2 * log - 8);
+
+        return (signed) (((((buf << log) >> log) - 1) ^ -(buf & 0x1)) + 1) >> 1;
+    }
+#else
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -250,6 +351,7 @@ static inline int get_interleaved_se_golomb(GetBitContext *gb)
 
         return (signed) (((((buf << log) >> log) - 1) ^ -(buf & 0x1)) + 1) >> 1;
     }
+#endif
 }
 
 static inline int dirac_get_se_golomb(GetBitContext *gb)
@@ -273,6 +375,24 @@ static inline int get_ur_golomb(GetBitContext *gb, int k, int limit,
     unsigned int buf;
     int log;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    log = av_log2(buf);
+
+    if (log > 31 - limit) {
+        buf >>= log - k;
+        buf  += (30 - log) << k;
+        skip_bits_long(gb, 32 + k - log);
+
+        return buf;
+    } else {
+        skip_bits_long(gb, limit);
+        buf = get_bits_long(gb, esc_len);
+
+        return buf + limit - 1;
+    }
+#else
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -297,6 +417,7 @@ static inline int get_ur_golomb(GetBitContext *gb, int k, int limit,
 
         return buf + limit - 1;
     }
+#endif
 }
 
 /**
@@ -308,6 +429,35 @@ static inline int get_ur_golomb_jpegls(GetBitContext *gb, int k, int limit,
     unsigned int buf;
     int log;
 
+#ifdef CACHED_BITSTREAM_READER
+    buf = show_bits_long(gb, 32);
+
+    log = av_log2(buf);
+
+    if (log - k >= 1 && 32 - log < limit) {
+        buf >>= log - k;
+        buf  += (30 - log) << k;
+        skip_bits_long(gb, 32 + k - log);
+
+        return buf;
+    } else {
+        int i;
+        for (i = 0;
+             i < limit && get_bits1(gb) == 0 && get_bits_left(gb) > 0;
+             i++);
+
+        if (i < limit - 1) {
+            buf = get_bits_long(gb, k);
+
+            return buf + (i << k);
+        } else if (i == limit - 1) {
+            buf = get_bits_long(gb, esc_len);
+
+            return buf + 1;
+        } else
+            return -1;
+    }
+#else
     OPEN_READER(re, gb);
     UPDATE_CACHE(re, gb);
     buf = GET_CACHE(re, gb);
@@ -364,6 +514,7 @@ static inline int get_ur_golomb_jpegls(GetBitContext *gb, int k, int limit,
         CLOSE_READER(re, gb);
         return buf;
     }
+#endif
 }
 
 /**