For very short input data (0 - 1 bytes), lzo-rle was not behaving correctly. Fix this behaviour and update documentation accordingly. For zero-length input, lzo v0 outputs an end-of-stream marker only, which was misinterpreted by lzo-rle as a bitstream version number. Ensure bitstream versions > 0 require a minimum stream length of 5. Also fixes a bug in handling the tail for very short inputs when a bitstream version is present.
Unified · Split
| @@ -102,9 +102,11 @@ Byte sequences · 101 unmodified lines | ||
| 102 | 102 | dictionary which is empty, and that it will always be |
| 103 | 103 | invalid at this place. |
| 104 | 104 | |
| 105 | − 17 : bitstream version. If the first byte is 17, the next byte | |
| 106 | − gives the bitstream version (version 1 only). If the first byte | |
| 107 | − is not 17, the bitstream version is 0. | |
| 105 | + 17 : bitstream version. If the first byte is 17, and compressed | |
| 106 | + stream length is at least 5 bytes (length of shortest possible | |
| 107 | + versioned bitstream), the next byte gives the bitstream version | |
| 108 | + (version 1 only). | |
| 109 | + Otherwise, the bitstream version is 0. | |
| 108 | 110 | |
| 109 | 111 | 18..21 : copy 0..3 literals |
| 110 | 112 | state = (byte - 17) = 0..3 [ copy <state> literals ] |
| @@ -291,13 +291,14 @@ int lzogeneric1x_1_compress(const unsigned char *in, size_t in_len, · 290 unmodified lines | ||
| 291 | 291 | { |
| 292 | 292 | const unsigned char *ip = in; |
| 293 | 293 | unsigned char *op = out; |
| 294 | + unsigned char *data_start; | |
| 294 | 295 | size_t l = in_len; |
| 295 | 296 | size_t t = 0; |
| 296 | 297 | signed char state_offset = -2; |
| 297 | 298 | unsigned int m4_max_offset; |
| 298 | 299 | |
| 299 | − // LZO v0 will never write 17 as first byte, | |
| 300 | − // so this is used to version the bitstream | |
| 300 | + // LZO v0 will never write 17 as first byte (except for zero-length | |
| 301 | + // input), so this is used to version the bitstream | |
| 301 | 302 | if (bitstream_version > 0) { |
| 302 | 303 | *op++ = 17; |
| 303 | 304 | *op++ = bitstream_version; |
| @@ -306,6 +307,8 @@ int lzogeneric1x_1_compress(const unsigned char *in, size_t in_len, · 2 unmodified lines | ||
| 306 | 307 | m4_max_offset = M4_MAX_OFFSET_V0; |
| 307 | 308 | } |
| 308 | 309 | |
| 310 | + data_start = op; | |
| 311 | + | |
| 309 | 312 | while (l > 20) { |
| 310 | 313 | size_t ll = l <= (m4_max_offset + 1) ? l : (m4_max_offset + 1); |
| 311 | 314 | uintptr_t ll_end = (uintptr_t) ip + ll; |
| @@ -324,7 +327,7 @@ int lzogeneric1x_1_compress(const unsigned char *in, size_t in_len, · 12 unmodified lines | ||
| 324 | 327 | if (t > 0) { |
| 325 | 328 | const unsigned char *ii = in + in_len - t; |
| 326 | 329 | |
| 327 | − if (op == out && t <= 238) { | |
| 330 | + if (op == data_start && t <= 238) { | |
| 328 | 331 | *op++ = (17 + t); |
| 329 | 332 | } else if (t <= 3) { |
| 330 | 333 | op[state_offset] |= t; |
| @@ -54,11 +54,9 @@ int lzo1x_decompress_safe(const unsigned char *in, size_t in_len, · 53 unmodified lines | ||
| 54 | 54 | if (unlikely(in_len < 3)) |
| 55 | 55 | goto input_overrun; |
| 56 | 56 | |
| 57 | − if (likely(*ip == 17)) { | |
| 57 | + if (likely(in_len >= 5) && likely(*ip == 17)) { | |
| 58 | 58 | bitstream_version = ip[1]; |
| 59 | 59 | ip += 2; |
| 60 | − if (unlikely(in_len < 5)) | |
| 61 | − goto input_overrun; | |
| 62 | 60 | } else { |
| 63 | 61 | bitstream_version = 0; |
| 64 | 62 | } |