zlib源码分析—输出Buffer机制

旧城等待， 2022-11-20 09:46 189阅读 0赞

## 双缓冲概述 ##

为了输出压缩数据，zlib使用两个缓冲区：一个pending缓冲区和一个输出缓冲区。初始化后，zlib将创建一个pending缓冲区，并创建一个输出缓冲区。输出数据首先被存储在pending缓冲区中，然后被复制到输出缓冲区中，最后被写入输出压缩的zip或gz文件中。数据在函数flush\_pending中从pending缓冲区复制到输出缓冲区。当文字缓冲区已满时，即表示已处理数据块时，将调用此函数。在某些其他情况下，当需要刷新块时也调用它。复制到输出缓冲区的数据长度受输出缓冲区中的可用空间限制。当输出缓冲区已满或发出刷新信号时，zlib将输出缓冲区写入zip或gz文件。  
![在这里插入图片描述][watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2FzbWFydGtpbGxlcg_size_16_color_FFFFFF_t_70_pic_center]

### 向bi\_buf输入数据 ###

send\_bits函数用于向bi\_buf缓存中放入比特数据，首先需要判断bi\_buf中能否放得下length比特的数据。如果放不下（s->bi\_valid > (int)Buf\_size - len），将输入数据的低Buf\_size-bi\_valid比特放在bi\_buf的高位，将bi\_buf刷写到pending\_buf中，并将输入数据剩余的位放在bi\_buf中，更新bi\_valid有效数据为bi\_valid+len-Buf\_size。如果放的下，直接将value移位bi\_valid，并设置bi\_valid。

#define send_bits(s, value, length) \ { int len = length;\ if (s->bi_valid > (int)Buf_size - len) {\ int val = (int)value;\ s->bi_buf |= (ush)val << s->bi_valid;\ put_short(s, s->bi_buf);\ s->bi_buf = (ush)val >> (Buf_size - s->bi_valid);\ s->bi_valid += len - Buf_size;\ } else {\ s->bi_buf |= (ush)(value) << s->bi_valid;\ s->bi_valid += len;\ }\ }

### 将bi\_buf数据输出 ###

bi\_windup函数以字节为边界输出bi\_buf中的比特数据。如果bi\_buf中的有效比特大于8，将整个bi\_buf全部输出；如果bi\_valid小于等于8，大于0，则将包含有效比特的8比特数据输出到pending\_buf中。

/* =========================================================================== * Flush the bit buffer and align the output on a byte boundary */
    local void bi_windup(deflate_state *s) { 
        if (s->bi_valid > 8) { 
            put_short(s, s->bi_buf);
        } else if (s->bi_valid > 0) { 
            put_byte(s, (Byte)s->bi_buf);
        }
        s->bi_buf = 0;
        s->bi_valid = 0;
    }

bi\_flush函数将bi\_buf中的比特数据刷写到pending\_buf中，当bi\_valid等于16时，将全部数据输出；当bi\_valid大于等于8时，将低8比特输出，也就是最多留7比特在bi\_buf中。

/* =========================================================================== * Flush the bits in the bit buffer to pending output (leaves at most 7 bits) */
    void ZLIB_INTERNAL _tr_flush_bits(deflate_state *s) { 
        bi_flush(s);
    }
    /* =========================================================================== * Flush the bit buffer, keeping at most 7 bits in it. */
    local void bi_flush(deflate_state *s){ 
        if (s->bi_valid == 16) { 
            put_short(s, s->bi_buf);
            s->bi_buf = 0;
            s->bi_valid = 0;
        } else if (s->bi_valid >= 8) { 
            put_byte(s, (Byte)s->bi_buf);
            s->bi_buf >>= 8;
            s->bi_valid -= 8;
        }
    }

### 向pending\_buf输入数据 ###

向pending\_buf中输出单字节数据，pending用于指示pending\_buf中数据的数量（以字节为单位）。

/* Output a byte on the stream. * IN assertion: there is enough room in pending_buf. */
    #define put_byte(s, c) {s->pending_buf[s->pending++] = (Bytef)(c);}

put\_short函数向pending\_buf中输出2字节数据，先输出低8位，后输入高8位，即LSB。putShortMSB函数向pending\_buf中输出2字节数据，先输出高8位，后输入低8位，即MSB。

/* =========================================================================== * Output a short LSB first on the stream. * IN assertion: there is enough room in pendingBuf. */
    #define put_short(s, w) { \ put_byte(s, (uch)((w) & 0xff)); \ put_byte(s, (uch)((ush)(w) >> 8)); \ }
    /* ========================================================================= * Put a short in the pending buffer. The 16-bit value is put in MSB order. * IN assertion: the stream state is correct and there is enough room in * pending_buf. */
    local void putShortMSB (deflate_state *s, uInt b){ 
        put_byte(s, (Byte)(b >> 8));
        put_byte(s, (Byte)(b & 0xff));
    }

### 将pending\_buf数据输出 ###

除了deflate\_stored函数的处理数据输出，其他在deflate函数中输出的数据都需要通过该函数将数据刷写到输出。\_tr\_flush\_bits将bi\_buf中的有效比特输出到pending\_buf中。pending指的是pending\_buf中未刷写的数据的大小，avail\_out指可获得的输出空间大小，这两者取小值来确定调用刷出的数据量。拷贝之后，需要将next\_out指针向前移动len，pending\_out指针向前移动len，给total\_out增加len，给avail\_out减少len，并减少pending的大小。最后判定，如果没有可以刷写的数据，直接将pending\_out指针归零（pending\_buf）。

/* ========================================================================= * Flush as much pending output as possible. All deflate() output, except for * some deflate_stored() output, goes through this function so some * applications may wish to modify it to avoid allocating a large * strm->next_out buffer and copying into it. (See also read_buf()). */
    local void flush_pending(z_streamp strm)
    { 
        unsigned len;
        deflate_state *s = strm->state;
        _tr_flush_bits(s);
        len = s->pending;
        if (len > strm->avail_out) len = strm->avail_out;
        if (len == 0) return;
        zmemcpy(strm->next_out, s->pending_out, len);
        strm->next_out  += len;
        s->pending_out  += len;
        strm->total_out += len;
        strm->avail_out -= len;
        s->pending      -= len;
        if (s->pending == 0) { 
            s->pending_out = s->pending_buf;
        }
    }

[watermark_type_ZmFuZ3poZW5naGVpdGk_shadow_10_text_aHR0cHM6Ly9ibG9nLmNzZG4ubmV0L2FzbWFydGtpbGxlcg_size_16_color_FFFFFF_t_70_pic_center]: /images/20221120/a3bb71dda09f4881b0ea0e755388f644.png