source: branches/stable/mindi-busybox/archival/libunarchive/decompress_bunzip2.c @ 1770

Last change on this file since 1770 was 1770, checked in by Bruno Cornec, 13 years ago
  • Better output for mindi-busybox revision
  • Remove dummy file created on NFS - report from Arnaud Tiger <arnaud.tiger_at_hp.com>
  • strace useful for debug
  • fix new versions for pb (2.0.0 for mindi and 1.7.2 for mindi-busybox)
  • fix build process for mindi-busybox + options used in that version (dd for label-partitions-as-necessary)
  • fix typo in label-partitions-as-necessary which doesn't seem to work
  • Update to busybox 1.7.2
  • perl is now required at restore time to support uuid swap partitions (and will be used for many other thigs

in the future for sure)

  • next mindi version will be 2.0.0 due to all the changes made in it (udev may break working distros)
  • small optimization in mindi on keyboard handling (one single find instead of multiple)
  • better interaction for USB device when launching mindi manually
  • attempt to automatically guess block disk size for ramdisk
  • fix typos in bkphw
  • Fix the remaining problem with UUID support for swap partitions
  • Updates mondoarchive man page for USB support
  • Adds preliminary Hardware support to mindi (Proliant SSSTK)
  • Tries to add udev support also for rhel4
  • Fix UUID support which was still broken.
  • Be conservative in test for the start-nfs script
  • Update config file for mindi-busybox for 1.7.2 migration
  • Try to run around a busybox bug (1.2.2 pb on inexistant links)
  • Add build content for mindi-busybox in pb
  • Remove distributions content for mindi-busybox
  • Fix a warning on inexistant raidtab
  • Solve problem on tmpfs in restore init (Problem of inexistant symlink and busybox)
  • Create MONDO_CACHE and use it everywhere + creation at start
  • Really never try to eject a USB device
  • Fix a issue with &> usage (replaced with 1> and 2>)
  • Adds magic file to depllist in order to have file working + ldd which helps for debugging issues
  • tty modes correct to avoid sh error messages
  • Use ext3 normally and not ext2 instead
  • USB device should be corrected after reading (take 1st part)
  • Adds a mount_USB_here function derived from mount_CDROM_here
  • usb detection place before /dev detection in device name at restore time
  • Fix when restoring from USB: media is asked in interactive mode
  • Adds USB support for mondorestore
  • mount_cdrom => mount_media
  • elilo.efi is now searched throughout /boot/efi and not in a fixed place as there is no standard
  • untar-and-softlink => untar (+ interface change)
  • suppress useless softlinks creation/removal in boot process
  • avoids udevd messages on groups
  • Increase # of disks to 99 as in mindi at restore time (should be a conf file parameter)
  • skip existing big file creation
  • seems to work correctly for USB mindi boot
  • Adds group and tty link to udev conf
  • Always load usb-torage (even 2.6) to initiate USB bus discovery
  • Better printing of messages
  • Attempt to fix a bug in supporting OpenSusE 10.3 kernel for initramfs (mindi may now use multiple regex for kernel initrd detection)
  • Links were not correctly done as non relative for modules in mindi
  • exclusion of modules denied now works
  • Also create modules in their ordinary place, so that classical modprobe works + copy modules.dep
  • Fix bugs for DENY_MODS handling
  • Add device /dev/console for udev
  • ide-generic should now really be excluded
  • Fix a bug in major number for tty
  • If udev then adds modprobe/insmod to rootfs
  • tty0 is also cretaed with udev
  • ide-generic put rather in DENY_MODS
  • udevd remove from deplist s handled in mindi directly
  • better default for mindi when using --usb
  • Handles dynamically linked busybox (in case we want to use it soon ;-)
  • Adds fixed devices to create for udev
  • ide-generic should not be part of the initrd when using libata v2
  • support a dynamically linked udev (case on Ubuntu 7.10 and Mandriva 2008.0 so should be quite generic) This will give incitation to move to dyn. linked binaries in the initrd which will help for other tasks (ia6 4)
  • Improvement in udev support (do not use cl options not available in busybox)
  • Udev in mindi
    • auto creation of the right links at boot time with udev-links.conf(from Mandriva 2008.0)
    • rework startup of udev as current makes kernel crash (from Mandriva 2008.0)
    • add support for 64 bits udev
  • Try to render MyInsmod? silent at boot time
  • Adds udev support (mandatory for newest distributions to avoid remapping of devices in a different way as on the original system)
  • We also need vaft format support for USB boot
  • Adds libusual support (Ubuntu 7.10 needs it for USB)
  • Improve Ubuntu/Debian? keyboard detection and support
  • pbinit adapted to new pb (0.8.10). Filtering of docs done in it
  • Suppress some mondo warnings and errors on USB again
  • Tries to fix lack of files in deb mindi package
  • Verify should now work for USB devices
  • More log/mesages improvement for USB support
  • - Supress g_erase_tmpdir_and_scratchdir
  • Improve some log messages for USB support
  • Try to improve install in mindi to avoid issues with isolinux.cfg not installed vene if in the pkg :-(
  • Improve mindi-busybox build
  • In conformity with pb 0.8.9
  • Add support for Ubuntu 7.10 in build process
  • Add USB Key button to Menu UI (CD streamer removed)
  • Attempt to fix error messages on tmp/scratch files at the end by removing those dir at the latest possible.
  • Fix a bug linked to the size of the -E param which could be used (Arnaud Tiger/René? Ribaud).
  • Integrate ~/.pbrc content into mondorescue.pb (required project-builder >= 0.8.7)
  • Put mondorescue in conformity with new pb filtering rules
  • Add USB support at restore time (no test done yet). New start-usb script PB varibale added where useful
  • Unmounting USB device before removal of temporary scratchdir
  • Stil refining USB copy back to mondo (one command was not executed)
  • No need to have the image subdor in the csratchdir when USB.
  • umount the USB partition before attempting to use it
  • Remove useless copy from mindi to mondo at end of USB handling

(risky merge, we are raising the limits of 2 diverging branches. The status of stable is not completely sure as such. Will need lots of tests, but it's not yet done :-()
(merge -r1692:1769 $SVN_M/branches/2.2.5)

File size: 23.1 KB
Line 
1/* vi: set sw=4 ts=4: */
2/* Small bzip2 deflate implementation, by Rob Landley (rob@landley.net).
3
4   Based on bzip2 decompression code by Julian R Seward (jseward@acm.org),
5   which also acknowledges contributions by Mike Burrows, David Wheeler,
6   Peter Fenwick, Alistair Moffat, Radford Neal, Ian H. Witten,
7   Robert Sedgewick, and Jon L. Bentley.
8
9   Licensed under GPLv2 or later, see file LICENSE in this tarball for details.
10*/
11
12/*
13    Size and speed optimizations by Manuel Novoa III  (mjn3@codepoet.org).
14
15    More efficient reading of Huffman codes, a streamlined read_bunzip()
16    function, and various other tweaks.  In (limited) tests, approximately
17    20% faster than bzcat on x86 and about 10% faster on arm.
18
19    Note that about 2/3 of the time is spent in read_unzip() reversing
20    the Burrows-Wheeler transformation.  Much of that time is delay
21    resulting from cache misses.
22
23    I would ask that anyone benefiting from this work, especially those
24    using it in commercial products, consider making a donation to my local
25    non-profit hospice organization (www.hospiceacadiana.com) in the name of
26    the woman I loved, Toni W. Hagan, who passed away Feb. 12, 2003.
27
28    Manuel
29 */
30
31#include "libbb.h"
32#include "unarchive.h"
33
34/* Constants for Huffman coding */
35#define MAX_GROUPS          6
36#define GROUP_SIZE          50      /* 64 would have been more efficient */
37#define MAX_HUFCODE_BITS    20      /* Longest Huffman code allowed */
38#define MAX_SYMBOLS         258     /* 256 literals + RUNA + RUNB */
39#define SYMBOL_RUNA         0
40#define SYMBOL_RUNB         1
41
42/* Status return values */
43#define RETVAL_OK                       0
44#define RETVAL_LAST_BLOCK               (-1)
45#define RETVAL_NOT_BZIP_DATA            (-2)
46#define RETVAL_UNEXPECTED_INPUT_EOF     (-3)
47#define RETVAL_UNEXPECTED_OUTPUT_EOF    (-4)
48#define RETVAL_DATA_ERROR               (-5)
49#define RETVAL_OUT_OF_MEMORY            (-6)
50#define RETVAL_OBSOLETE_INPUT           (-7)
51
52/* Other housekeeping constants */
53#define IOBUF_SIZE          4096
54
55/* This is what we know about each Huffman coding group */
56struct group_data {
57    /* We have an extra slot at the end of limit[] for a sentinal value. */
58    int limit[MAX_HUFCODE_BITS+1], base[MAX_HUFCODE_BITS], permute[MAX_SYMBOLS];
59    int minLen, maxLen;
60};
61
62/* Structure holding all the housekeeping data, including IO buffers and
63   memory that persists between calls to bunzip */
64
65struct bunzip_data {
66    /* State for interrupting output loop */
67    int writeCopies, writePos, writeRunCountdown, writeCount, writeCurrent;
68
69    /* I/O tracking data (file handles, buffers, positions, etc.) */
70    int in_fd, out_fd, inbufCount, inbufPos /*, outbufPos*/;
71    unsigned char *inbuf /*,*outbuf*/;
72    unsigned inbufBitCount, inbufBits;
73
74    /* The CRC values stored in the block header and calculated from the data */
75    uint32_t headerCRC, totalCRC, writeCRC;
76
77    /* Intermediate buffer and its size (in bytes) */
78    unsigned *dbuf, dbufSize;
79
80    /* For I/O error handling */
81    jmp_buf jmpbuf;
82
83    /* Big things go last (register-relative addressing can be larger for big offsets */
84    uint32_t crc32Table[256];
85    unsigned char selectors[32768];         /* nSelectors=15 bits */
86    struct group_data groups[MAX_GROUPS];   /* Huffman coding tables */
87};
88/* typedef struct bunzip_data bunzip_data; -- done in .h file */
89
90
91/* Return the next nnn bits of input.  All reads from the compressed input
92   are done through this function.  All reads are big endian */
93
94static unsigned get_bits(bunzip_data *bd, char bits_wanted)
95{
96    unsigned bits = 0;
97
98    /* If we need to get more data from the byte buffer, do so.  (Loop getting
99       one byte at a time to enforce endianness and avoid unaligned access.) */
100
101    while (bd->inbufBitCount < bits_wanted) {
102
103        /* If we need to read more data from file into byte buffer, do so */
104
105        if (bd->inbufPos == bd->inbufCount) {
106            /* if "no input fd" case: in_fd == -1, read fails, we jump */
107            bd->inbufCount = read(bd->in_fd, bd->inbuf, IOBUF_SIZE);
108            if (bd->inbufCount <= 0)
109                longjmp(bd->jmpbuf, RETVAL_UNEXPECTED_INPUT_EOF);
110            bd->inbufPos = 0;
111        }
112
113        /* Avoid 32-bit overflow (dump bit buffer to top of output) */
114
115        if (bd->inbufBitCount >= 24) {
116            bits = bd->inbufBits & ((1 << bd->inbufBitCount) - 1);
117            bits_wanted -= bd->inbufBitCount;
118            bits <<= bits_wanted;
119            bd->inbufBitCount = 0;
120        }
121
122        /* Grab next 8 bits of input from buffer. */
123
124        bd->inbufBits = (bd->inbufBits<<8) | bd->inbuf[bd->inbufPos++];
125        bd->inbufBitCount += 8;
126    }
127
128    /* Calculate result */
129
130    bd->inbufBitCount -= bits_wanted;
131    bits |= (bd->inbufBits >> bd->inbufBitCount) & ((1 << bits_wanted) - 1);
132
133    return bits;
134}
135
136/* Unpacks the next block and sets up for the inverse burrows-wheeler step. */
137
138static int get_next_block(bunzip_data *bd)
139{
140    struct group_data *hufGroup;
141    int dbufCount, nextSym, dbufSize, groupCount, *base, *limit, selector,
142        i, j, k, t, runPos, symCount, symTotal, nSelectors, byteCount[256];
143    unsigned char uc, symToByte[256], mtfSymbol[256], *selectors;
144    unsigned *dbuf, origPtr;
145
146    dbuf = bd->dbuf;
147    dbufSize = bd->dbufSize;
148    selectors = bd->selectors;
149
150    /* Reset longjmp I/O error handling */
151
152    i = setjmp(bd->jmpbuf);
153    if (i) return i;
154
155    /* Read in header signature and CRC, then validate signature.
156       (last block signature means CRC is for whole file, return now) */
157
158    i = get_bits(bd, 24);
159    j = get_bits(bd, 24);
160    bd->headerCRC = get_bits(bd, 32);
161    if ((i == 0x177245) && (j == 0x385090)) return RETVAL_LAST_BLOCK;
162    if ((i != 0x314159) || (j != 0x265359)) return RETVAL_NOT_BZIP_DATA;
163
164    /* We can add support for blockRandomised if anybody complains.  There was
165       some code for this in busybox 1.0.0-pre3, but nobody ever noticed that
166       it didn't actually work. */
167
168    if (get_bits(bd, 1)) return RETVAL_OBSOLETE_INPUT;
169    origPtr = get_bits(bd, 24);
170    if (origPtr > dbufSize) return RETVAL_DATA_ERROR;
171
172    /* mapping table: if some byte values are never used (encoding things
173       like ascii text), the compression code removes the gaps to have fewer
174       symbols to deal with, and writes a sparse bitfield indicating which
175       values were present.  We make a translation table to convert the symbols
176       back to the corresponding bytes. */
177
178    t = get_bits(bd, 16);
179    symTotal = 0;
180    for (i = 0; i < 16; i++) {
181        if (t & (1 << (15-i))) {
182            k = get_bits(bd, 16);
183            for (j = 0; j < 16; j++)
184                if (k & (1 << (15-j)))
185                    symToByte[symTotal++] = (16*i) + j;
186        }
187    }
188
189    /* How many different Huffman coding groups does this block use? */
190
191    groupCount = get_bits(bd, 3);
192    if (groupCount < 2 || groupCount > MAX_GROUPS)
193        return RETVAL_DATA_ERROR;
194
195    /* nSelectors: Every GROUP_SIZE many symbols we select a new Huffman coding
196       group.  Read in the group selector list, which is stored as MTF encoded
197       bit runs.  (MTF=Move To Front, as each value is used it's moved to the
198       start of the list.) */
199
200    nSelectors = get_bits(bd, 15);
201    if (!nSelectors) return RETVAL_DATA_ERROR;
202    for (i = 0; i < groupCount; i++) mtfSymbol[i] = i;
203    for (i = 0; i < nSelectors; i++) {
204
205        /* Get next value */
206
207        for (j = 0; get_bits(bd, 1); j++)
208            if (j>=groupCount) return RETVAL_DATA_ERROR;
209
210        /* Decode MTF to get the next selector */
211
212        uc = mtfSymbol[j];
213        for (;j;j--) mtfSymbol[j] = mtfSymbol[j-1];
214        mtfSymbol[0] = selectors[i] = uc;
215    }
216
217    /* Read the Huffman coding tables for each group, which code for symTotal
218       literal symbols, plus two run symbols (RUNA, RUNB) */
219
220    symCount = symTotal + 2;
221    for (j = 0; j < groupCount; j++) {
222        unsigned char length[MAX_SYMBOLS], temp[MAX_HUFCODE_BITS+1];
223        int minLen, maxLen, pp;
224
225        /* Read Huffman code lengths for each symbol.  They're stored in
226           a way similar to mtf; record a starting value for the first symbol,
227           and an offset from the previous value for everys symbol after that.
228           (Subtracting 1 before the loop and then adding it back at the end is
229           an optimization that makes the test inside the loop simpler: symbol
230           length 0 becomes negative, so an unsigned inequality catches it.) */
231
232        t = get_bits(bd, 5) - 1;
233        for (i = 0; i < symCount; i++) {
234            for (;;) {
235                if ((unsigned)t > (MAX_HUFCODE_BITS-1))
236                    return RETVAL_DATA_ERROR;
237
238                /* If first bit is 0, stop.  Else second bit indicates whether
239                   to increment or decrement the value.  Optimization: grab 2
240                   bits and unget the second if the first was 0. */
241
242                k = get_bits(bd, 2);
243                if (k < 2) {
244                    bd->inbufBitCount++;
245                    break;
246                }
247
248                /* Add one if second bit 1, else subtract 1.  Avoids if/else */
249
250                t += (((k+1) & 2) - 1);
251            }
252
253            /* Correct for the initial -1, to get the final symbol length */
254
255            length[i] = t + 1;
256        }
257
258        /* Find largest and smallest lengths in this group */
259
260        minLen = maxLen = length[0];
261        for (i = 1; i < symCount; i++) {
262            if (length[i] > maxLen) maxLen = length[i];
263            else if (length[i] < minLen) minLen = length[i];
264        }
265
266        /* Calculate permute[], base[], and limit[] tables from length[].
267         *
268         * permute[] is the lookup table for converting Huffman coded symbols
269         * into decoded symbols.  base[] is the amount to subtract from the
270         * value of a Huffman symbol of a given length when using permute[].
271         *
272         * limit[] indicates the largest numerical value a symbol with a given
273         * number of bits can have.  This is how the Huffman codes can vary in
274         * length: each code with a value>limit[length] needs another bit.
275         */
276
277        hufGroup = bd->groups + j;
278        hufGroup->minLen = minLen;
279        hufGroup->maxLen = maxLen;
280
281        /* Note that minLen can't be smaller than 1, so we adjust the base
282           and limit array pointers so we're not always wasting the first
283           entry.  We do this again when using them (during symbol decoding).*/
284
285        base = hufGroup->base - 1;
286        limit = hufGroup->limit - 1;
287
288        /* Calculate permute[].  Concurently, initialize temp[] and limit[]. */
289
290        pp = 0;
291        for (i = minLen; i <= maxLen; i++) {
292            temp[i] = limit[i] = 0;
293            for (t = 0; t < symCount; t++)
294                if (length[t] == i)
295                    hufGroup->permute[pp++] = t;
296        }
297
298        /* Count symbols coded for at each bit length */
299
300        for (i = 0; i < symCount; i++) temp[length[i]]++;
301
302        /* Calculate limit[] (the largest symbol-coding value at each bit
303         * length, which is (previous limit<<1)+symbols at this level), and
304         * base[] (number of symbols to ignore at each bit length, which is
305         * limit minus the cumulative count of symbols coded for already). */
306
307        pp = t = 0;
308        for (i = minLen; i < maxLen; i++) {
309            pp += temp[i];
310
311            /* We read the largest possible symbol size and then unget bits
312               after determining how many we need, and those extra bits could
313               be set to anything.  (They're noise from future symbols.)  At
314               each level we're really only interested in the first few bits,
315               so here we set all the trailing to-be-ignored bits to 1 so they
316               don't affect the value>limit[length] comparison. */
317
318            limit[i] = (pp << (maxLen - i)) - 1;
319            pp <<= 1;
320            t += temp[i];
321            base[i+1] = pp - t;
322        }
323        limit[maxLen+1] = INT_MAX; /* Sentinal value for reading next sym. */
324        limit[maxLen] = pp + temp[maxLen] - 1;
325        base[minLen] = 0;
326    }
327
328    /* We've finished reading and digesting the block header.  Now read this
329       block's Huffman coded symbols from the file and undo the Huffman coding
330       and run length encoding, saving the result into dbuf[dbufCount++]=uc */
331
332    /* Initialize symbol occurrence counters and symbol Move To Front table */
333
334    for (i = 0; i < 256; i++) {
335        byteCount[i] = 0;
336        mtfSymbol[i] = (unsigned char)i;
337    }
338
339    /* Loop through compressed symbols. */
340
341    runPos = dbufCount = selector = 0;
342    for (;;) {
343
344        /* fetch next Huffman coding group from list. */
345
346        symCount = GROUP_SIZE - 1;
347        if (selector >= nSelectors) return RETVAL_DATA_ERROR;
348        hufGroup = bd->groups + selectors[selector++];
349        base = hufGroup->base - 1;
350        limit = hufGroup->limit - 1;
351 continue_this_group:
352
353        /* Read next Huffman-coded symbol. */
354
355        /* Note: It is far cheaper to read maxLen bits and back up than it is
356           to read minLen bits and then an additional bit at a time, testing
357           as we go.  Because there is a trailing last block (with file CRC),
358           there is no danger of the overread causing an unexpected EOF for a
359           valid compressed file.  As a further optimization, we do the read
360           inline (falling back to a call to get_bits if the buffer runs
361           dry).  The following (up to got_huff_bits:) is equivalent to
362           j = get_bits(bd, hufGroup->maxLen);
363         */
364
365        while (bd->inbufBitCount < hufGroup->maxLen) {
366            if (bd->inbufPos == bd->inbufCount) {
367                j = get_bits(bd, hufGroup->maxLen);
368                goto got_huff_bits;
369            }
370            bd->inbufBits = (bd->inbufBits << 8) | bd->inbuf[bd->inbufPos++];
371            bd->inbufBitCount += 8;
372        };
373        bd->inbufBitCount -= hufGroup->maxLen;
374        j = (bd->inbufBits >> bd->inbufBitCount) & ((1 << hufGroup->maxLen) - 1);
375
376 got_huff_bits:
377
378        /* Figure how how many bits are in next symbol and unget extras */
379
380        i = hufGroup->minLen;
381        while (j > limit[i]) ++i;
382        bd->inbufBitCount += (hufGroup->maxLen - i);
383
384        /* Huffman decode value to get nextSym (with bounds checking) */
385
386        if (i > hufGroup->maxLen)
387            return RETVAL_DATA_ERROR;
388        j = (j >> (hufGroup->maxLen - i)) - base[i];
389        if ((unsigned)j >= MAX_SYMBOLS)
390            return RETVAL_DATA_ERROR;
391        nextSym = hufGroup->permute[j];
392
393        /* We have now decoded the symbol, which indicates either a new literal
394           byte, or a repeated run of the most recent literal byte.  First,
395           check if nextSym indicates a repeated run, and if so loop collecting
396           how many times to repeat the last literal. */
397
398        if ((unsigned)nextSym <= SYMBOL_RUNB) { /* RUNA or RUNB */
399
400            /* If this is the start of a new run, zero out counter */
401
402            if (!runPos) {
403                runPos = 1;
404                t = 0;
405            }
406
407            /* Neat trick that saves 1 symbol: instead of or-ing 0 or 1 at
408               each bit position, add 1 or 2 instead.  For example,
409               1011 is 1<<0 + 1<<1 + 2<<2.  1010 is 2<<0 + 2<<1 + 1<<2.
410               You can make any bit pattern that way using 1 less symbol than
411               the basic or 0/1 method (except all bits 0, which would use no
412               symbols, but a run of length 0 doesn't mean anything in this
413               context).  Thus space is saved. */
414
415            t += (runPos << nextSym); /* +runPos if RUNA; +2*runPos if RUNB */
416            if (runPos < dbufSize) runPos <<= 1;
417            goto end_of_huffman_loop;
418        }
419
420        /* When we hit the first non-run symbol after a run, we now know
421           how many times to repeat the last literal, so append that many
422           copies to our buffer of decoded symbols (dbuf) now.  (The last
423           literal used is the one at the head of the mtfSymbol array.) */
424
425        if (runPos) {
426            runPos = 0;
427            if (dbufCount + t >= dbufSize) return RETVAL_DATA_ERROR;
428
429            uc = symToByte[mtfSymbol[0]];
430            byteCount[uc] += t;
431            while (t--) dbuf[dbufCount++] = uc;
432        }
433
434        /* Is this the terminating symbol? */
435
436        if (nextSym > symTotal) break;
437
438        /* At this point, nextSym indicates a new literal character.  Subtract
439           one to get the position in the MTF array at which this literal is
440           currently to be found.  (Note that the result can't be -1 or 0,
441           because 0 and 1 are RUNA and RUNB.  But another instance of the
442           first symbol in the mtf array, position 0, would have been handled
443           as part of a run above.  Therefore 1 unused mtf position minus
444           2 non-literal nextSym values equals -1.) */
445
446        if (dbufCount >= dbufSize) return RETVAL_DATA_ERROR;
447        i = nextSym - 1;
448        uc = mtfSymbol[i];
449
450        /* Adjust the MTF array.  Since we typically expect to move only a
451         * small number of symbols, and are bound by 256 in any case, using
452         * memmove here would typically be bigger and slower due to function
453         * call overhead and other assorted setup costs. */
454
455        do {
456            mtfSymbol[i] = mtfSymbol[i-1];
457        } while (--i);
458        mtfSymbol[0] = uc;
459        uc = symToByte[uc];
460
461        /* We have our literal byte.  Save it into dbuf. */
462
463        byteCount[uc]++;
464        dbuf[dbufCount++] = (unsigned)uc;
465
466        /* Skip group initialization if we're not done with this group.  Done
467         * this way to avoid compiler warning. */
468
469 end_of_huffman_loop:
470        if (symCount--) goto continue_this_group;
471    }
472
473    /* At this point, we've read all the Huffman-coded symbols (and repeated
474       runs) for this block from the input stream, and decoded them into the
475       intermediate buffer.  There are dbufCount many decoded bytes in dbuf[].
476       Now undo the Burrows-Wheeler transform on dbuf.
477       See http://dogma.net/markn/articles/bwt/bwt.htm
478     */
479
480    /* Turn byteCount into cumulative occurrence counts of 0 to n-1. */
481
482    j = 0;
483    for (i = 0; i < 256; i++) {
484        k = j + byteCount[i];
485        byteCount[i] = j;
486        j = k;
487    }
488
489    /* Figure out what order dbuf would be in if we sorted it. */
490
491    for (i = 0; i < dbufCount; i++) {
492        uc = (unsigned char)(dbuf[i] & 0xff);
493        dbuf[byteCount[uc]] |= (i << 8);
494        byteCount[uc]++;
495    }
496
497    /* Decode first byte by hand to initialize "previous" byte.  Note that it
498       doesn't get output, and if the first three characters are identical
499       it doesn't qualify as a run (hence writeRunCountdown=5). */
500
501    if (dbufCount) {
502        if (origPtr >= dbufCount) return RETVAL_DATA_ERROR;
503        bd->writePos = dbuf[origPtr];
504        bd->writeCurrent = (unsigned char)(bd->writePos & 0xff);
505        bd->writePos >>= 8;
506        bd->writeRunCountdown = 5;
507    }
508    bd->writeCount = dbufCount;
509
510    return RETVAL_OK;
511}
512
513/* Undo burrows-wheeler transform on intermediate buffer to produce output.
514   If start_bunzip was initialized with out_fd=-1, then up to len bytes of
515   data are written to outbuf.  Return value is number of bytes written or
516   error (all errors are negative numbers).  If out_fd!=-1, outbuf and len
517   are ignored, data is written to out_fd and return is RETVAL_OK or error.
518*/
519
520int read_bunzip(bunzip_data *bd, char *outbuf, int len)
521{
522    const unsigned *dbuf;
523    int pos, current, previous, gotcount;
524
525    /* If last read was short due to end of file, return last block now */
526    if (bd->writeCount < 0) return bd->writeCount;
527
528    gotcount = 0;
529    dbuf = bd->dbuf;
530    pos = bd->writePos;
531    current = bd->writeCurrent;
532
533    /* We will always have pending decoded data to write into the output
534       buffer unless this is the very first call (in which case we haven't
535       Huffman-decoded a block into the intermediate buffer yet). */
536
537    if (bd->writeCopies) {
538
539        /* Inside the loop, writeCopies means extra copies (beyond 1) */
540
541        --bd->writeCopies;
542
543        /* Loop outputting bytes */
544
545        for (;;) {
546
547            /* If the output buffer is full, snapshot state and return */
548
549            if (gotcount >= len) {
550                bd->writePos  =pos;
551                bd->writeCurrent = current;
552                bd->writeCopies++;
553                return len;
554            }
555
556            /* Write next byte into output buffer, updating CRC */
557
558            outbuf[gotcount++] = current;
559            bd->writeCRC = (bd->writeCRC << 8)
560                          ^ bd->crc32Table[(bd->writeCRC >> 24) ^ current];
561
562            /* Loop now if we're outputting multiple copies of this byte */
563
564            if (bd->writeCopies) {
565                --bd->writeCopies;
566                continue;
567            }
568 decode_next_byte:
569            if (!bd->writeCount--) break;
570            /* Follow sequence vector to undo Burrows-Wheeler transform */
571            previous = current;
572            pos = dbuf[pos];
573            current = pos & 0xff;
574            pos >>= 8;
575
576            /* After 3 consecutive copies of the same byte, the 4th is a repeat
577               count.  We count down from 4 instead
578             * of counting up because testing for non-zero is faster */
579
580            if (--bd->writeRunCountdown) {
581                if (current != previous)
582                    bd->writeRunCountdown = 4;
583            } else {
584
585                /* We have a repeated run, this byte indicates the count */
586
587                bd->writeCopies = current;
588                current = previous;
589                bd->writeRunCountdown = 5;
590
591                /* Sometimes there are just 3 bytes (run length 0) */
592
593                if (!bd->writeCopies) goto decode_next_byte;
594
595                /* Subtract the 1 copy we'd output anyway to get extras */
596
597                --bd->writeCopies;
598            }
599        }
600
601        /* Decompression of this block completed successfully */
602
603        bd->writeCRC = ~bd->writeCRC;
604        bd->totalCRC = ((bd->totalCRC << 1) | (bd->totalCRC >> 31)) ^ bd->writeCRC;
605
606        /* If this block had a CRC error, force file level CRC error. */
607
608        if (bd->writeCRC != bd->headerCRC) {
609            bd->totalCRC = bd->headerCRC+1;
610            return RETVAL_LAST_BLOCK;
611        }
612    }
613
614    /* Refill the intermediate buffer by Huffman-decoding next block of input */
615    /* (previous is just a convenient unused temp variable here) */
616
617    previous = get_next_block(bd);
618    if (previous) {
619        bd->writeCount = previous;
620        return (previous != RETVAL_LAST_BLOCK) ? previous : gotcount;
621    }
622    bd->writeCRC = ~0;
623    pos = bd->writePos;
624    current = bd->writeCurrent;
625    goto decode_next_byte;
626}
627
628
629/* Allocate the structure, read file header.  If in_fd==-1, inbuf must contain
630   a complete bunzip file (len bytes long).  If in_fd!=-1, inbuf and len are
631   ignored, and data is read from file handle into temporary buffer. */
632
633/* Because bunzip2 is used for help text unpacking, and because bb_show_usage()
634   should work for NOFORK applets too, we must be extremely careful to not leak
635   any allocations! */
636
637int start_bunzip(bunzip_data **bdp, int in_fd, const unsigned char *inbuf,
638                        int len)
639{
640    bunzip_data *bd;
641    unsigned i;
642    enum {
643        BZh0 = ('B' << 24) + ('Z' << 16) + ('h' << 8) + '0'
644    };
645
646    /* Figure out how much data to allocate */
647
648    i = sizeof(bunzip_data);
649    if (in_fd != -1) i += IOBUF_SIZE;
650
651    /* Allocate bunzip_data.  Most fields initialize to zero. */
652
653    bd = *bdp = xzalloc(i);
654
655    /* Setup input buffer */
656
657    bd->in_fd = in_fd;
658    if (-1 == in_fd) {
659        /* in this case, bd->inbuf is read-only */
660        bd->inbuf = (void*)inbuf; /* cast away const-ness */
661        bd->inbufCount = len;
662    } else
663        bd->inbuf = (unsigned char *)(bd + 1);
664
665    /* Init the CRC32 table (big endian) */
666
667    crc32_filltable(bd->crc32Table, 1);
668
669    /* Setup for I/O error handling via longjmp */
670
671    i = setjmp(bd->jmpbuf);
672    if (i) return i;
673
674    /* Ensure that file starts with "BZh['1'-'9']." */
675
676    i = get_bits(bd, 32);
677    if ((unsigned)(i - BZh0 - 1) >= 9) return RETVAL_NOT_BZIP_DATA;
678
679    /* Fourth byte (ascii '1'-'9'), indicates block size in units of 100k of
680       uncompressed data.  Allocate intermediate buffer for block. */
681
682    bd->dbufSize = 100000 * (i - BZh0);
683
684    /* Cannot use xmalloc - may leak bd in NOFORK case! */
685    bd->dbuf = malloc_or_warn(bd->dbufSize * sizeof(int));
686    if (!bd->dbuf) {
687        free(bd);
688        xfunc_die();
689    }
690    return RETVAL_OK;
691}
692
693void dealloc_bunzip(bunzip_data *bd)
694{
695    free(bd->dbuf);
696    free(bd);
697}
698
699
700/* Decompress src_fd to dst_fd.  Stops at end of bzip data, not end of file. */
701
702USE_DESKTOP(long long) int
703unpack_bz2_stream(int src_fd, int dst_fd)
704{
705    USE_DESKTOP(long long total_written = 0;)
706    char *outbuf;
707    bunzip_data *bd;
708    int i;
709
710    outbuf = xmalloc(IOBUF_SIZE);
711    i = start_bunzip(&bd, src_fd, NULL, 0);
712    if (!i) {
713        for (;;) {
714            i = read_bunzip(bd, outbuf, IOBUF_SIZE);
715            if (i <= 0) break;
716            if (i != safe_write(dst_fd, outbuf, i)) {
717                i = RETVAL_UNEXPECTED_OUTPUT_EOF;
718                break;
719            }
720            USE_DESKTOP(total_written += i;)
721        }
722    }
723
724    /* Check CRC and release memory */
725
726    if (i == RETVAL_LAST_BLOCK) {
727        if (bd->headerCRC != bd->totalCRC) {
728            bb_error_msg("data integrity error when decompressing");
729        } else {
730            i = RETVAL_OK;
731        }
732    } else if (i == RETVAL_UNEXPECTED_OUTPUT_EOF) {
733        bb_error_msg("compressed file ends unexpectedly");
734    } else {
735        bb_error_msg("decompression failed");
736    }
737    dealloc_bunzip(bd);
738    free(outbuf);
739
740    return i ? i : USE_DESKTOP(total_written) + 0;
741}
742
743#ifdef TESTING
744
745static char *const bunzip_errors[] = {
746    NULL, "Bad file checksum", "Not bzip data",
747    "Unexpected input EOF", "Unexpected output EOF", "Data error",
748    "Out of memory", "Obsolete (pre 0.9.5) bzip format not supported"
749};
750
751/* Dumb little test thing, decompress stdin to stdout */
752int main(int argc, char **argv)
753{
754    int i = unpack_bz2_stream(0, 1);
755    char c;
756
757    if (i < 0)
758        fprintf(stderr,"%s\n", bunzip_errors[-i]);
759    else if (read(0, &c, 1))
760        fprintf(stderr,"Trailing garbage ignored\n");
761    return -i;
762}
763#endif
Note: See TracBrowser for help on using the repository browser.