gzipvfs.cpp@ 37043

Last change on this file since 37043 was 34049, checked in by vboxsync, 14 years ago
iprt/tarvfs: Working on validating and parsing the tar headers.
Property svn:eol-style set to `native` Property svn:keywords set to `Author Date Id Revision`
File size: 17.6 KB

Line
1	/* $Id: gzipvfs.cpp 34049 2010-11-13 01:31:07Z vboxsync $ */
2	/** @file
3	* IPRT - GZIP Compressor and Decompressor I/O Stream.
4	*/
5
6	/*
7	* Copyright (C) 2010 Oracle Corporation
8	*
9	* This file is part of VirtualBox Open Source Edition (OSE), as
10	* available from http://www.virtualbox.org. This file is free software;
11	* you can redistribute it and/or modify it under the terms of the GNU
12	* General Public License (GPL) as published by the Free Software
13	* Foundation, in version 2 as it comes in the "COPYING" file of the
14	* VirtualBox OSE distribution. VirtualBox OSE is distributed in the
15	* hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
16	*
17	* The contents of this file may alternatively be used under the terms
18	* of the Common Development and Distribution License Version 1.0
19	* (CDDL) only, as it comes in the "COPYING.CDDL" file of the
20	* VirtualBox OSE distribution, in which case the provisions of the
21	* CDDL are applicable instead of those of the GPL.
22	*
23	* You may elect to license modified versions of this file under the
24	* terms and conditions of either the GPL or the CDDL or both.
25	*/
26
27
28	/*******************************************************************************
29	* Defined Constants And Macros *
30	*******************************************************************************/
31	#include "internal/iprt.h"
32	#include <iprt/zip.h>
33
34	#include <iprt/assert.h>
35	#include <iprt/file.h>
36	#include <iprt/err.h>
37	#include <iprt/poll.h>
38	#include <iprt/string.h>
39	#include <iprt/vfslowlevel.h>
40
41	#include <zlib.h>
42
43
44	/*******************************************************************************
45	* Structures and Typedefs *
46	*******************************************************************************/
47	#pragma pack(1)
48	typedef struct RTZIPGZIPHDR
49	{
50	/** RTZIPGZIPHDR_ID1. */
51	uint8_t bId1;
52	/** RTZIPGZIPHDR_ID2. */
53	uint8_t bId2;
54	/** CM - The compression method. */
55	uint8_t bCompressionMethod;
56	/** FLG - Flags. */
57	uint8_t fFlags;
58	/** Modification time of the source file or the timestamp at the time the
59	* compression took place. Can also be zero. Is the number of seconds since
60	* unix epoch. */
61	uint32_t u32ModTime;
62	/** Flags specific to the compression method. */
63	uint8_t bXtraFlags;
64	/** An ID indicating which OS or FS gzip ran on. */
65	uint8_t bOS;
66	} RTZIPGZIPHDR;
67	#pragma pack()
68	AssertCompileSize(RTZIPGZIPHDR, 10);
69	/** Pointer to a const gzip header. */
70	typedef RTZIPGZIPHDR const *PCRTZIPGZIPHDR;
71
72	/** gzip header identification no 1. */
73	#define RTZIPGZIPHDR_ID1 0x1f
74	/** gzip header identification no 2. */
75	#define RTZIPGZIPHDR_ID2 0x8b
76	/** gzip deflate compression method. */
77	#define RTZIPGZIPHDR_CM_DEFLATE 8
78
79	/** @name gzip header flags
80	* @{ */
81	/** Probably a text file */
82	#define RTZIPGZIPHDR_FLG_TEXT UINT8_C(0x01)
83	/** Header CRC present (crc32 of header cast to uint16_t). */
84	#define RTZIPGZIPHDR_FLG_HDR_CRC UINT8_C(0x02)
85	/** Length prefixed xtra field is present. */
86	#define RTZIPGZIPHDR_FLG_EXTRA UINT8_C(0x04)
87	/** A name field is present (latin-1). */
88	#define RTZIPGZIPHDR_FLG_NAME UINT8_C(0x08)
89	/** A comment field is present (latin-1). */
90	#define RTZIPGZIPHDR_FLG_COMMENT UINT8_C(0x10)
91	/** Mask of valid flags. */
92	#define RTZIPGZIPHDR_FLG_VALID_MASK UINT8_C(0x1f)
93	/** @} */
94
95	/** @name gzip default xtra flag values
96	* @{ */
97	#define RTZIPGZIPHDR_XFL_DEFLATE_MAX UINT8_C(0x02)
98	#define RTZIPGZIPHDR_XFL_DEFLATE_FASTEST UINT8_C(0x04)
99	/** @} */
100
101	/** @name Operating system / Filesystem IDs
102	* @{ */
103	#define RTZIPGZIPHDR_OS_FAT UINT8_C(0x00)
104	#define RTZIPGZIPHDR_OS_AMIGA UINT8_C(0x01)
105	#define RTZIPGZIPHDR_OS_VMS UINT8_C(0x02)
106	#define RTZIPGZIPHDR_OS_UNIX UINT8_C(0x03)
107	#define RTZIPGZIPHDR_OS_VM_CMS UINT8_C(0x04)
108	#define RTZIPGZIPHDR_OS_ATARIS_TOS UINT8_C(0x05)
109	#define RTZIPGZIPHDR_OS_HPFS UINT8_C(0x06)
110	#define RTZIPGZIPHDR_OS_MACINTOSH UINT8_C(0x07)
111	#define RTZIPGZIPHDR_OS_Z_SYSTEM UINT8_C(0x08)
112	#define RTZIPGZIPHDR_OS_CPM UINT8_C(0x09)
113	#define RTZIPGZIPHDR_OS_TOPS_20 UINT8_C(0x0a)
114	#define RTZIPGZIPHDR_OS_NTFS UINT8_C(0x0b)
115	#define RTZIPGZIPHDR_OS_QDOS UINT8_C(0x0c)
116	#define RTZIPGZIPHDR_OS_ACORN_RISCOS UINT8_C(0x0d)
117	#define RTZIPGZIPHDR_OS_UNKNOWN UINT8_C(0xff)
118	/** @} */
119
120
121	/**
122	* The internal data of a GZIP I/O stream.
123	*/
124	typedef struct RTZIPGZIPSTREAM
125	{
126	/** The stream we're reading or writing the compressed data from or to. */
127	RTVFSIOSTREAM hVfsIos;
128	/** Set if it's a decompressor, clear if it's a compressor. */
129	bool fDecompress;
130	/** Set if zlib reported a fatal error. */
131	bool fFatalError;
132	/** Set if we've reached the end of the zlib stream. */
133	bool fEndOfStream;
134	/** The stream offset for pfnTell. */
135	RTFOFF offStream;
136	/** The zlib stream. */
137	z_stream Zlib;
138	/** The data buffer. */
139	uint8_t abBuffer[_64K];
140	/** Scatter gather segment describing abBuffer. */
141	RTSGSEG SgSeg;
142	/** Scatter gather buffer describing abBuffer. */
143	RTSGBUF SgBuf;
144	/** The original file name (decompressor only). */
145	char *pszOrgName;
146	/** The comment (decompressor only). */
147	char *pszComment;
148	/** The gzip header. */
149	RTZIPGZIPHDR Hdr;
150	} RTZIPGZIPSTREAM;
151	/** Pointer to a the internal data of a GZIP I/O stream. */
152	typedef RTZIPGZIPSTREAM *PRTZIPGZIPSTREAM;
153
154
155	/**
156	* Convert from zlib to IPRT status codes.
157	*
158	* This will also set the fFatalError flag when appropriate.
159	*
160	* @returns IPRT status code.
161	* @param pThis The gzip I/O stream instance data.
162	* @param rc Zlib error code.
163	*/
164	static int rtZipGzipConvertErrFromZlib(PRTZIPGZIPSTREAM pThis, int rc)
165	{
166	switch (rc)
167	{
168	case Z_OK:
169	return VINF_SUCCESS;
170
171	case Z_BUF_ERROR:
172	/* This isn't fatal. */
173	return VINF_SUCCESS;
174
175	case Z_ERRNO:
176	case Z_STREAM_ERROR:
177	case Z_DATA_ERROR:
178	case Z_MEM_ERROR:
179	case Z_VERSION_ERROR:
180	pThis->fFatalError = true;
181	switch (rc)
182	{
183	case Z_ERRNO: return VERR_INTERNAL_ERROR_5;
184	case Z_STREAM_ERROR: return VERR_INTERNAL_ERROR_3;
185	case Z_DATA_ERROR: return VERR_ZIP_ERROR;
186	case Z_MEM_ERROR: return VERR_ZIP_NO_MEMORY;
187	case Z_VERSION_ERROR: return VERR_ZIP_UNSUPPORTED_VERSION;
188	}
189	/* not reached */
190
191	default:
192	if (rc >= 0)
193	return VINF_SUCCESS;
194	pThis->fFatalError = true;
195	return VERR_ZIP_ERROR;
196	}
197	}
198
199
200	/**
201	* @interface_method_impl{RTVFSOBJOPS,pfnClose}
202	*/
203	static DECLCALLBACK(int) rtZipGzip_Close(void *pvThis)
204	{
205	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
206
207	int rc;
208	if (pThis->fDecompress)
209	rc = inflateEnd(&pThis->Zlib);
210	else
211	rc = deflateEnd(&pThis->Zlib);
212	if (rc != Z_OK)
213	rc = rtZipGzipConvertErrFromZlib(pThis, rc);
214
215	RTVfsIoStrmRelease(pThis->hVfsIos);
216	pThis->hVfsIos = NIL_RTVFSIOSTREAM;
217	RTStrFree(pThis->pszOrgName);
218	pThis->pszOrgName = NULL;
219	RTStrFree(pThis->pszComment);
220	pThis->pszComment = NULL;
221
222	return rc;
223	}
224
225
226	/**
227	* @interface_method_impl{RTVFSOBJOPS,pfnQueryInfo}
228	*/
229	static DECLCALLBACK(int) rtZipGzip_QueryInfo(void *pvThis, PRTFSOBJINFO pObjInfo, RTFSOBJATTRADD enmAddAttr)
230	{
231	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
232	return RTVfsIoStrmQueryInfo(pThis->hVfsIos, pObjInfo, enmAddAttr);
233	}
234
235
236	/**
237	* Reads one segment.
238	*
239	* @returns IPRT status code.
240	* @param pThis The gzip I/O stream instance data.
241	* @param pvBuf Where to put the read bytes.
242	* @param cbToRead The number of bytes to read.
243	* @param fBlocking Whether to block or not.
244	* @param pcbRead Where to store the number of bytes actually read.
245	*/
246	static int rtZipGzip_ReadOneSeg(PRTZIPGZIPSTREAM pThis, void pvBuf, size_t cbToRead, bool fBlocking, size_t pcbRead)
247	{
248	/*
249	* This simplifies life a wee bit below.
250	*/
251	if (pThis->fEndOfStream)
252	return pcbRead ? VINF_EOF : VERR_EOF;
253
254	/*
255	* Set up the output buffer.
256	*/
257	pThis->Zlib.next_out = (Bytef *)pvBuf;
258	pThis->Zlib.avail_out = (uInt)cbToRead;
259	AssertReturn(pThis->Zlib.avail_out == cbToRead, VERR_OUT_OF_RANGE);
260
261	/*
262	* Be greedy reading input, even if no output buffer is left. It's possible
263	* that it's just the end of stream marker which needs to be read. Happens
264	* for incompressible blocks just larger than the input buffer size.
265	*/
266	int rc = VINF_SUCCESS;
267	while ( pThis->Zlib.avail_out > 0
268	\|\| pThis->Zlib.avail_in == 0 /* greedy */)
269	{
270	/*
271	* Read more input?
272	*
273	* N.B. The assertions here validate the RTVfsIoStrmSgRead behavior
274	* since the API is new and untested. They could be removed later
275	* but, better leaving them in.
276	*/
277	if (pThis->Zlib.avail_in == 0)
278	{
279	size_t cbReadIn = ~(size_t)0;
280	rc = RTVfsIoStrmSgRead(pThis->hVfsIos, &pThis->SgBuf, fBlocking, &cbReadIn);
281	if (rc != VINF_SUCCESS)
282	{
283	AssertMsg(RT_FAILURE(rc) \|\| rc == VINF_TRY_AGAIN \|\| rc == VINF_EOF, ("%Rrc\n", rc));
284	if (rc == VERR_INTERRUPTED)
285	{
286	Assert(cbReadIn == 0);
287	continue;
288	}
289	if (RT_FAILURE(rc) \|\| rc == VINF_TRY_AGAIN \|\| cbReadIn == 0)
290	{
291	Assert(cbReadIn == 0);
292	break;
293	}
294	AssertMsg(rc == VINF_EOF, ("%Rrc\n", rc));
295	}
296	AssertMsgBreakStmt(cbReadIn > 0 && cbReadIn <= sizeof(pThis->abBuffer), ("%zu %Rrc\n", cbReadIn, rc),
297	rc = VERR_INTERNAL_ERROR_4);
298
299	pThis->Zlib.avail_in = (uInt)cbReadIn;
300	pThis->Zlib.next_in = &pThis->abBuffer[0];
301	}
302
303	/*
304	* Pass it on to zlib.
305	*/
306	rc = inflate(&pThis->Zlib, Z_NO_FLUSH);
307	if (rc != Z_OK && rc != Z_BUF_ERROR)
308	{
309	if (rc == Z_STREAM_END)
310	{
311	pThis->fEndOfStream = true;
312	if (pThis->Zlib.avail_out == 0)
313	rc = VINF_SUCCESS;
314	else
315	rc = pcbRead ? VINF_EOF : VERR_EOF;
316	}
317	else
318	rc = rtZipGzipConvertErrFromZlib(pThis, rc);
319	break;
320	}
321	rc = VINF_SUCCESS;
322	}
323
324	/*
325	* Update the read counters before returning.
326	*/
327	size_t const cbRead = cbToRead - pThis->Zlib.avail_out;
328	pThis->offStream += cbRead;
329	if (pcbRead)
330	*pcbRead = cbRead;
331
332	return rc;
333	}
334
335	/**
336	* @interface_method_impl{RTVFSIOSTREAMOPS,pfnRead}
337	*/
338	static DECLCALLBACK(int) rtZipGzip_Read(void pvThis, RTFOFF off, PCRTSGBUF pSgBuf, bool fBlocking, size_t pcbRead)
339	{
340	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
341	int rc;
342
343	if (!pThis->fDecompress)
344	return VERR_ACCESS_DENIED;
345
346	if (pSgBuf->cSegs == 1)
347	rc = rtZipGzip_ReadOneSeg(pThis, pSgBuf->paSegs[0].pvSeg, pSgBuf->paSegs[0].cbSeg, fBlocking, pcbRead);
348	else
349	{
350	rc = VINF_SUCCESS;
351	size_t cbRead = 0;
352	size_t cbReadSeg;
353	size_t *pcbReadSeg = pcbRead ? &cbReadSeg : NULL;
354	for (uint32_t iSeg = 0; iSeg < pSgBuf->cSegs; iSeg++)
355	{
356	cbReadSeg = 0;
357	rc = rtZipGzip_ReadOneSeg(pThis, pSgBuf->paSegs[iSeg].pvSeg, pSgBuf->paSegs[iSeg].cbSeg, fBlocking, pcbReadSeg);
358	if (RT_FAILURE(rc))
359	break;
360	if (pcbRead)
361	{
362	cbRead += cbReadSeg;
363	if (cbReadSeg != pSgBuf->paSegs[iSeg].cbSeg)
364	break;
365	}
366	}
367	if (pcbRead)
368	*pcbRead = cbRead;
369	}
370
371	return rc;
372	}
373
374
375	/**
376	* @interface_method_impl{RTVFSIOSTREAMOPS,pfnWrite}
377	*/
378	static DECLCALLBACK(int) rtZipGzip_Write(void pvThis, RTFOFF off, PCRTSGBUF pSgBuf, bool fBlocking, size_t pcbWritten)
379	{
380	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
381	//int rc;
382
383	NOREF(fBlocking);
384	if (pThis->fDecompress)
385	return VERR_ACCESS_DENIED;
386
387	/** @todo implement compression. */
388	return VERR_NOT_IMPLEMENTED;
389	}
390
391
392	/**
393	* @interface_method_impl{RTVFSIOSTREAMOPS,pfnFlush}
394	*/
395	static DECLCALLBACK(int) rtZipGzip_Flush(void *pvThis)
396	{
397	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
398	return RTVfsIoStrmFlush(pThis->hVfsIos);
399	}
400
401
402	/**
403	* @interface_method_impl{RTVFSIOSTREAMOPS,pfnPollOne}
404	*/
405	static DECLCALLBACK(int) rtZipGzip_PollOne(void *pvThis, uint32_t fEvents, RTMSINTERVAL cMillies, bool fIntr,
406	uint32_t *pfRetEvents)
407	{
408	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
409
410	/*
411	* Collect our own events first and see if that satisfies the request. If
412	* not forward the call to the compressed stream.
413	*/
414	uint32_t fRetEvents = 0;
415	if (pThis->fFatalError)
416	fRetEvents \|= RTPOLL_EVT_ERROR;
417	if (pThis->fDecompress)
418	{
419	fEvents &= ~RTPOLL_EVT_WRITE;
420	if (pThis->Zlib.avail_in > 0)
421	fRetEvents = RTPOLL_EVT_READ;
422	}
423	else
424	{
425	fEvents &= ~RTPOLL_EVT_READ;
426	if (pThis->Zlib.avail_out > 0)
427	fRetEvents = RTPOLL_EVT_WRITE;
428	}
429
430	int rc = VINF_SUCCESS;
431	fRetEvents &= fEvents;
432	if (!fRetEvents)
433	rc = RTVfsIoStrmPoll(pThis->hVfsIos, fEvents, cMillies, fIntr, pfRetEvents);
434	return rc;
435	}
436
437
438	/**
439	* @interface_method_impl{RTVFSIOSTREAMOPS,pfnTell}
440	*/
441	static DECLCALLBACK(int) rtZipGzip_Tell(void *pvThis, PRTFOFF poffActual)
442	{
443	PRTZIPGZIPSTREAM pThis = (PRTZIPGZIPSTREAM)pvThis;
444	*poffActual = pThis->offStream;
445	return VINF_SUCCESS;
446	}
447
448
449	/**
450	* The GZIP I/O stream vtable.
451	*/
452	static RTVFSIOSTREAMOPS g_rtZipGzipOps =
453	{
454	{ /* Obj */
455	RTVFSOBJOPS_VERSION,
456	RTVFSOBJTYPE_IO_STREAM,
457	"gzip",
458	rtZipGzip_Close,
459	rtZipGzip_QueryInfo,
460	RTVFSOBJOPS_VERSION
461	},
462	RTVFSIOSTREAMOPS_VERSION,
463	0,
464	rtZipGzip_Read,
465	rtZipGzip_Write,
466	rtZipGzip_Flush,
467	rtZipGzip_PollOne,
468	rtZipGzip_Tell,
469	NULL /* Skip */,
470	NULL /ZeroFill/,
471	RTVFSIOSTREAMOPS_VERSION,
472	};
473
474
475	RTDECL(int) RTZipGzipDecompressIoStream(RTVFSIOSTREAM hVfsIosIn, uint32_t fFlags, PRTVFSIOSTREAM phVfsIosOut)
476	{
477	AssertPtrReturn(hVfsIosIn, VERR_INVALID_HANDLE);
478	AssertReturn(!fFlags, VERR_INVALID_PARAMETER);
479	AssertPtrReturn(phVfsIosOut, VERR_INVALID_POINTER);
480
481	uint32_t cRefs = RTVfsIoStrmRetain(hVfsIosIn);
482	AssertReturn(cRefs != UINT32_MAX, VERR_INVALID_HANDLE);
483
484	/*
485	* Create the decompression I/O stream.
486	*/
487	RTVFSIOSTREAM hVfsIos;
488	PRTZIPGZIPSTREAM pThis;
489	int rc = RTVfsNewIoStream(&g_rtZipGzipOps, sizeof(RTZIPGZIPSTREAM), RTFILE_O_READ, NIL_RTVFS, NIL_RTVFSLOCK,
490	&hVfsIos, (void **)&pThis);
491	if (RT_SUCCESS(rc))
492	{
493	pThis->hVfsIos = hVfsIosIn;
494	pThis->offStream = 0;
495	pThis->fDecompress = true;
496	pThis->SgSeg.pvSeg = &pThis->abBuffer[0];
497	pThis->SgSeg.cbSeg = sizeof(pThis->abBuffer);
498	RTSgBufInit(&pThis->SgBuf, &pThis->SgSeg, 1);
499
500	memset(&pThis->Zlib, 0, sizeof(pThis->Zlib));
501	pThis->Zlib.opaque = pThis;
502	rc = inflateInit2(&pThis->Zlib, MAX_WBITS + 16 /* autodetect gzip header */);
503	if (rc >= 0)
504	{
505	/*
506	* Read the gzip header from the input stream to check that it's
507	* a gzip stream.
508	*
509	* Note!. Since we've told zlib to check for the gzip header, we
510	* prebuffer what we read in the input buffer so it can
511	* be handed on to zlib later on.
512	*/
513	size_t cbRead = 0;
514	rc = RTVfsIoStrmRead(pThis->hVfsIos, pThis->abBuffer, sizeof(RTZIPGZIPHDR), true /fBlocking/, NULL /pcbRead/);
515	if (RT_SUCCESS(rc))
516	{
517	/* Validate the header and make a copy of it. */
518	PCRTZIPGZIPHDR pHdr = (PCRTZIPGZIPHDR)pThis->abBuffer;
519	if ( pHdr->bId1 != RTZIPGZIPHDR_ID1
520	\|\| pHdr->bId2 != RTZIPGZIPHDR_ID2
521	\|\| pHdr->fFlags & ~RTZIPGZIPHDR_FLG_VALID_MASK)
522	rc = VERR_ZIP_BAD_HEADER;
523	else if (pHdr->bCompressionMethod != RTZIPGZIPHDR_CM_DEFLATE)
524	rc = VERR_ZIP_UNSUPPORTED_METHOD;
525	else
526	{
527	pThis->Hdr = *pHdr;
528	pThis->Zlib.avail_in = sizeof(RTZIPGZIPHDR);
529	pThis->Zlib.next_in = &pThis->abBuffer[0];
530
531	/* Parse on if there are names or comments. */
532	if (pHdr->fFlags & (RTZIPGZIPHDR_FLG_NAME \| RTZIPGZIPHDR_FLG_COMMENT))
533	{
534	/** @todo Can implement this when someone needs the
535	* name or comment for something useful. */
536	}
537	if (RT_SUCCESS(rc))
538	{
539	*phVfsIosOut = hVfsIos;
540	return VINF_SUCCESS;
541	}
542	}
543	}
544	}
545	else
546	rc = rtZipGzipConvertErrFromZlib(pThis, rc); /** @todo cleaning up in this situation is going to go wrong. */
547	RTVfsIoStrmRelease(hVfsIos);
548	}
549	else
550	RTVfsIoStrmRelease(hVfsIosIn);
551	return rc;
552	}
553

Note: See TracBrowser for help on using the repository browser.

source: vbox/trunk/src/VBox/Runtime/common/zip/gzipvfs.cpp@ 37043

Download in other formats: