2 * LiZaRd -- Fast compression method based on Lempel-Ziv 77
4 * (c) 2004, Robert Spalek <robert@ucw.cz>
6 * This software may be freely distributed and used according to the terms
7 * of the GNU Lesser General Public License.
13 #ifdef CONFIG_UCW_CLEAN_ABI
14 #define adler32_update ucw_adler32_update
15 #define lizard_alloc ucw_lizard_alloc
16 #define lizard_compress ucw_lizard_compress
17 #define lizard_decompress ucw_lizard_decompress
18 #define lizard_decompress_safe ucw_lizard_decompress_safe
19 #define lizard_free ucw_lizard_free
29 * The compression routine needs input buffer 8 characters longer, because it
30 * does not check the input bounds all the time.
32 #define LIZARD_NEEDS_CHARS 8
34 #define LIZARD_MAX_MULTIPLY 23./22
35 #define LIZARD_MAX_ADD 4
36 /* In the worst case, the compressed file will not be longer than its
37 * original length * 23/22 + 4.
39 * The additive constant is for EOF and the header of the file.
41 * The multiplicative constant comes from 19-byte incompressible string
42 * followed by a 3-sequence that can be compressed into 2-byte link. This
43 * breaks the copy-mode and it needs to be restarted with a new header. The
44 * total length is 2(header) + 19(string) + 2(link) = 23.
48 * The compressed data will not be longer than `LIZARD_MAX_LEN(input_length)`.
49 * Note that `LIZARD_MAX_LEN(length) > length` (this is not a problem of the algorithm,
50 * every lossless compression algorithm must have an input for which it produces a larger
53 * Use this to compute the size of @out parameter of @lizard_compress().
55 #define LIZARD_MAX_LEN(LENGTH) ((LENGTH) * LIZARD_MAX_MULTIPLY + LIZARD_MAX_ADD)
60 * Compress data provided in @in.
61 * The input buffer must be at last `@in_len + <<def_LIZARD_NEEDS_CHARS,LIZARD_NEEDS_CHARS>>`
62 * long (the compression algorithm does not check the bounds all the time).
64 * The output will be stored in @out. The @out buffer must be at last <<def_LIZARD_LEN,`LIZARD_LEN(@in_len)`>>
65 * bytes long for the output to fit in for sure.
67 * The function returns number of bytes actually needed (the size of output).
69 * Use @lizard_decompress() to get the original data.
71 int lizard_compress(const byte *in, uint in_len, byte *out);
74 * Decompress data previously compressed by @lizard_compress().
75 * Input is taken from @in and the result stored in @out.
76 * The size of output is returned.
78 * Note that you need to know the maximal possible size of the output to
79 * allocate enough memory.
81 * See also <<safe,safe decompression>>.
83 int lizard_decompress(const byte *in, byte *out);
92 * You can use safe decompression, when you want to make sure you got the
93 * length right and when you want to reuse the buffer for output.
96 struct lizard_buffer; /** Type of the output buffer for @lizard_decompress_safe(). **/
98 struct lizard_buffer *lizard_alloc(void); /** Get me a new <<struct_lizard_buffer,`lizard_buffer`>>. **/
100 * Return memory used by a <<struct_lizard_buffer,`lizard_buffer`>>.
101 * It frees even the data stored in it (the result of
102 * @lizard_decompress_safe() call that used this buffer).
104 void lizard_free(struct lizard_buffer *buf);
107 * This one acts much like @lizard_decompress(). The difference is it
108 * checks the data to be of correct length (therefore it will not
109 * crash on invalid data).
111 * It decompresses data provided by @in. The @buf is used to get the
112 * memory for output (you get one by @lizard_alloc()).
114 * The pointer to decompressed data is returned. To free it, free the
115 * buffer by @lizard_free().
117 * In the case of error, NULL is returned. In that case, `errno` is
118 * set either to `EINVAL` (expected_length does not match) or to
119 * `EFAULT` (a segfault has been caught while decompressing -- it
120 * probably means expected_length was set way too low). Both cases
121 * suggest either wrongly computed length or data corruption.
123 * The @buf argument may be reused for multiple decompresses. However,
124 * the data will be overwritten by the next call.
126 * Beware this function is not thread-safe and is not even reentrant
127 * (because of internal segfault handling).
129 byte *lizard_decompress_safe(const byte *in, struct lizard_buffer *buf, uint expected_length);
138 * This is here because it is commonly used to check data compressed by LiZaRd.
139 * However, it could also belong to <<hash,hashing routines>>.
143 * Update the Adler-32 checksum with more data.
144 * @adler is the old value, @byte points to @len bytes of data to update with.
145 * Result is returned.
147 uint adler32_update(uint adler, const byte *ptr, uint len);
150 * Compute the Adler-32 checksum of a block of data.
152 static inline uint adler32(const byte *buf, uint len)
154 return adler32_update(1, buf, len);