xref: /freebsd/sys/kern/subr_compressor.c (revision f18976136625a7d016e97bfd9eabddf640b3e06d) 
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
3  *
4  * Copyright (c) 2014, 2017 Mark Johnston <markj@FreeBSD.org>
5  * Copyright (c) 2017 Conrad Meyer <cem@FreeBSD.org>
6  *
7  * Redistribution and use in source and binary forms, with or without
8  * modification, are permitted provided that the following conditions are
9  * met:
10  * 1. Redistributions of source code must retain the above copyright
11  *    notice, this list of conditions and the following disclaimer.
12  * 2. Redistributions in binary form must reproduce the above copyright
13  *    notice, this list of conditions and the following disclaimer in
14  *    the documentation and/or other materials provided with the
15  *    distribution.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  */
29 
30 /*
31  * Subroutines used for writing compressed user process and kernel core dumps.
32  */
33 
34 #include <sys/cdefs.h>
35 __FBSDID("$FreeBSD$");
36 
37 #include "opt_gzio.h"
38 #include "opt_zstdio.h"
39 
40 #include <sys/param.h>
41 #include <sys/systm.h>
42 
43 #include <sys/compressor.h>
44 #include <sys/kernel.h>
45 #include <sys/linker_set.h>
46 #include <sys/malloc.h>
47 
48 MALLOC_DEFINE(M_COMPRESS, "compressor", "kernel compression subroutines");
49 
50 struct compressor_methods {
51 	int format;
52 	void *(* const init)(size_t, int);
53 	void (* const reset)(void *);
54 	int (* const write)(void *, void *, size_t, compressor_cb_t, void *);
55 	void (* const fini)(void *);
56 };
57 
58 struct compressor {
59 	const struct compressor_methods *methods;
60 	compressor_cb_t cb;
61 	void *priv;
62 	void *arg;
63 };
64 
65 SET_DECLARE(compressors, struct compressor_methods);
66 
67 #ifdef GZIO
68 
69 #include <contrib/zlib/zutil.h>
70 
71 struct gz_stream {
72 	uint8_t		*gz_buffer;	/* output buffer */
73 	size_t		gz_bufsz;	/* output buffer size */
74 	off_t		gz_off;		/* offset into the output stream */
75 	uint32_t	gz_crc;		/* stream CRC32 */
76 	z_stream	gz_stream;	/* zlib state */
77 };
78 
79 static void 	*gz_init(size_t maxiosize, int level);
80 static void	gz_reset(void *stream);
81 static int	gz_write(void *stream, void *data, size_t len, compressor_cb_t,
82 		    void *);
83 static void	gz_fini(void *stream);
84 
85 static void *
86 gz_alloc(void *arg __unused, u_int n, u_int sz)
87 {
88 
89 	/*
90 	 * Memory for zlib state is allocated using M_NODUMP since it may be
91 	 * used to compress a kernel dump, and we don't want zlib to attempt to
92 	 * compress its own state.
93 	 */
94 	return (malloc(n * sz, M_COMPRESS, M_WAITOK | M_ZERO | M_NODUMP));
95 }
96 
97 static void
98 gz_free(void *arg __unused, void *ptr)
99 {
100 
101 	free(ptr, M_COMPRESS);
102 }
103 
104 static void *
105 gz_init(size_t maxiosize, int level)
106 {
107 	struct gz_stream *s;
108 	int error;
109 
110 	s = gz_alloc(NULL, 1, roundup2(sizeof(*s), PAGE_SIZE));
111 	s->gz_buffer = gz_alloc(NULL, 1, maxiosize);
112 	s->gz_bufsz = maxiosize;
113 
114 	s->gz_stream.zalloc = gz_alloc;
115 	s->gz_stream.zfree = gz_free;
116 	s->gz_stream.opaque = NULL;
117 	s->gz_stream.next_in = Z_NULL;
118 	s->gz_stream.avail_in = 0;
119 
120 	error = deflateInit2(&s->gz_stream, level, Z_DEFLATED, -MAX_WBITS,
121 	    DEF_MEM_LEVEL, Z_DEFAULT_STRATEGY);
122 	if (error != 0)
123 		goto fail;
124 
125 	gz_reset(s);
126 
127 	return (s);
128 
129 fail:
130 	gz_free(NULL, s);
131 	return (NULL);
132 }
133 
134 static void
135 gz_reset(void *stream)
136 {
137 	struct gz_stream *s;
138 	uint8_t *hdr;
139 	const size_t hdrlen = 10;
140 
141 	s = stream;
142 	s->gz_off = 0;
143 	s->gz_crc = crc32(0L, Z_NULL, 0);
144 
145 	(void)deflateReset(&s->gz_stream);
146 	s->gz_stream.avail_out = s->gz_bufsz;
147 	s->gz_stream.next_out = s->gz_buffer;
148 
149 	/* Write the gzip header to the output buffer. */
150 	hdr = s->gz_buffer;
151 	memset(hdr, 0, hdrlen);
152 	hdr[0] = 0x1f;
153 	hdr[1] = 0x8b;
154 	hdr[2] = Z_DEFLATED;
155 	hdr[9] = OS_CODE;
156 	s->gz_stream.next_out += hdrlen;
157 	s->gz_stream.avail_out -= hdrlen;
158 }
159 
160 static int
161 gz_write(void *stream, void *data, size_t len, compressor_cb_t cb,
162     void *arg)
163 {
164 	struct gz_stream *s;
165 	uint8_t trailer[8];
166 	size_t room;
167 	int error, zerror, zflag;
168 
169 	s = stream;
170 	zflag = data == NULL ? Z_FINISH : Z_NO_FLUSH;
171 
172 	if (len > 0) {
173 		s->gz_stream.avail_in = len;
174 		s->gz_stream.next_in = data;
175 		s->gz_crc = crc32(s->gz_crc, data, len);
176 	}
177 
178 	error = 0;
179 	do {
180 		zerror = deflate(&s->gz_stream, zflag);
181 		if (zerror != Z_OK && zerror != Z_STREAM_END) {
182 			error = EIO;
183 			break;
184 		}
185 
186 		if (s->gz_stream.avail_out == 0 || zerror == Z_STREAM_END) {
187 			/*
188 			 * Our output buffer is full or there's nothing left
189 			 * to produce, so we're flushing the buffer.
190 			 */
191 			len = s->gz_bufsz - s->gz_stream.avail_out;
192 			if (zerror == Z_STREAM_END) {
193 				/*
194 				 * Try to pack as much of the trailer into the
195 				 * output buffer as we can.
196 				 */
197 				((uint32_t *)trailer)[0] = s->gz_crc;
198 				((uint32_t *)trailer)[1] =
199 				    s->gz_stream.total_in;
200 				room = MIN(sizeof(trailer),
201 				    s->gz_bufsz - len);
202 				memcpy(s->gz_buffer + len, trailer, room);
203 				len += room;
204 			}
205 
206 			error = cb(s->gz_buffer, len, s->gz_off, arg);
207 			if (error != 0)
208 				break;
209 
210 			s->gz_off += len;
211 			s->gz_stream.next_out = s->gz_buffer;
212 			s->gz_stream.avail_out = s->gz_bufsz;
213 
214 			/*
215 			 * If we couldn't pack the trailer into the output
216 			 * buffer, write it out now.
217 			 */
218 			if (zerror == Z_STREAM_END && room < sizeof(trailer))
219 				error = cb(trailer + room,
220 				    sizeof(trailer) - room, s->gz_off, arg);
221 		}
222 	} while (zerror != Z_STREAM_END &&
223 	    (zflag == Z_FINISH || s->gz_stream.avail_in > 0));
224 
225 	return (error);
226 }
227 
228 static void
229 gz_fini(void *stream)
230 {
231 	struct gz_stream *s;
232 
233 	s = stream;
234 	(void)deflateEnd(&s->gz_stream);
235 	gz_free(NULL, s->gz_buffer);
236 	gz_free(NULL, s);
237 }
238 
239 struct compressor_methods gzip_methods = {
240 	.format = COMPRESS_GZIP,
241 	.init = gz_init,
242 	.reset = gz_reset,
243 	.write = gz_write,
244 	.fini = gz_fini,
245 };
246 DATA_SET(compressors, gzip_methods);
247 
248 #endif /* GZIO */
249 
250 #ifdef ZSTDIO
251 
252 #define	ZSTD_STATIC_LINKING_ONLY
253 #include <contrib/zstd/lib/zstd.h>
254 
255 struct zstdio_stream {
256 	ZSTD_CCtx	*zst_stream;
257 	ZSTD_inBuffer	zst_inbuffer;
258 	ZSTD_outBuffer	zst_outbuffer;
259 	uint8_t *	zst_buffer;	/* output buffer */
260 	size_t		zst_maxiosz;	/* Max output IO size */
261 	off_t		zst_off;	/* offset into the output stream */
262 	void *		zst_static_wkspc;
263 };
264 
265 static void 	*zstdio_init(size_t maxiosize, int level);
266 static void	zstdio_reset(void *stream);
267 static int	zstdio_write(void *stream, void *data, size_t len,
268 		    compressor_cb_t, void *);
269 static void	zstdio_fini(void *stream);
270 
271 static void *
272 zstdio_init(size_t maxiosize, int level)
273 {
274 	ZSTD_CCtx *dump_compressor;
275 	struct zstdio_stream *s;
276 	void *wkspc, *owkspc, *buffer;
277 	size_t wkspc_size, buf_size, rc;
278 
279 	s = NULL;
280 	wkspc_size = ZSTD_estimateCStreamSize(level);
281 	owkspc = wkspc = malloc(wkspc_size + 8, M_COMPRESS,
282 	    M_WAITOK | M_NODUMP);
283 	/* Zstd API requires 8-byte alignment. */
284 	if ((uintptr_t)wkspc % 8 != 0)
285 		wkspc = (void *)roundup2((uintptr_t)wkspc, 8);
286 
287 	dump_compressor = ZSTD_initStaticCCtx(wkspc, wkspc_size);
288 	if (dump_compressor == NULL) {
289 		printf("%s: workspace too small.\n", __func__);
290 		goto out;
291 	}
292 
293 	rc = ZSTD_CCtx_setParameter(dump_compressor, ZSTD_c_checksumFlag, 1);
294 	if (ZSTD_isError(rc)) {
295 		printf("%s: error setting checksumFlag: %s\n", __func__,
296 		    ZSTD_getErrorName(rc));
297 		goto out;
298 	}
299 	rc = ZSTD_CCtx_setParameter(dump_compressor, ZSTD_c_compressionLevel,
300 	    level);
301 	if (ZSTD_isError(rc)) {
302 		printf("%s: error setting compressLevel: %s\n", __func__,
303 		    ZSTD_getErrorName(rc));
304 		goto out;
305 	}
306 
307 	buf_size = ZSTD_CStreamOutSize() * 2;
308 	buffer = malloc(buf_size, M_COMPRESS, M_WAITOK | M_NODUMP);
309 
310 	s = malloc(sizeof(*s), M_COMPRESS, M_NODUMP | M_WAITOK);
311 	s->zst_buffer = buffer;
312 	s->zst_outbuffer.dst = buffer;
313 	s->zst_outbuffer.size = buf_size;
314 	s->zst_maxiosz = maxiosize;
315 	s->zst_stream = dump_compressor;
316 	s->zst_static_wkspc = owkspc;
317 
318 	zstdio_reset(s);
319 
320 out:
321 	if (s == NULL)
322 		free(owkspc, M_COMPRESS);
323 	return (s);
324 }
325 
326 static void
327 zstdio_reset(void *stream)
328 {
329 	struct zstdio_stream *s;
330 	size_t res;
331 
332 	s = stream;
333 	res = ZSTD_resetCStream(s->zst_stream, 0);
334 	if (ZSTD_isError(res))
335 		panic("%s: could not reset stream %p: %s\n", __func__, s,
336 		    ZSTD_getErrorName(res));
337 
338 	s->zst_off = 0;
339 	s->zst_inbuffer.src = NULL;
340 	s->zst_inbuffer.size = 0;
341 	s->zst_inbuffer.pos = 0;
342 	s->zst_outbuffer.pos = 0;
343 }
344 
345 static int
346 zst_flush_intermediate(struct zstdio_stream *s, compressor_cb_t cb, void *arg)
347 {
348 	size_t bytes_to_dump;
349 	int error;
350 
351 	/* Flush as many full output blocks as possible. */
352 	/* XXX: 4096 is arbitrary safe HDD block size for kernel dumps */
353 	while (s->zst_outbuffer.pos >= 4096) {
354 		bytes_to_dump = rounddown(s->zst_outbuffer.pos, 4096);
355 
356 		if (bytes_to_dump > s->zst_maxiosz)
357 			bytes_to_dump = s->zst_maxiosz;
358 
359 		error = cb(s->zst_buffer, bytes_to_dump, s->zst_off, arg);
360 		if (error != 0)
361 			return (error);
362 
363 		/*
364 		 * Shift any non-full blocks up to the front of the output
365 		 * buffer.
366 		 */
367 		s->zst_outbuffer.pos -= bytes_to_dump;
368 		memmove(s->zst_outbuffer.dst,
369 		    (char *)s->zst_outbuffer.dst + bytes_to_dump,
370 		    s->zst_outbuffer.pos);
371 		s->zst_off += bytes_to_dump;
372 	}
373 	return (0);
374 }
375 
376 static int
377 zstdio_flush(struct zstdio_stream *s, compressor_cb_t cb, void *arg)
378 {
379 	size_t rc, lastpos;
380 	int error;
381 
382 	/*
383 	 * Positive return indicates unflushed data remaining; need to call
384 	 * endStream again after clearing out room in output buffer.
385 	 */
386 	rc = 1;
387 	lastpos = s->zst_outbuffer.pos;
388 	while (rc > 0) {
389 		rc = ZSTD_endStream(s->zst_stream, &s->zst_outbuffer);
390 		if (ZSTD_isError(rc)) {
391 			printf("%s: ZSTD_endStream failed (%s)\n", __func__,
392 			    ZSTD_getErrorName(rc));
393 			return (EIO);
394 		}
395 		if (lastpos == s->zst_outbuffer.pos) {
396 			printf("%s: did not make forward progress endStream %zu\n",
397 			    __func__, lastpos);
398 			return (EIO);
399 		}
400 
401 		error = zst_flush_intermediate(s, cb, arg);
402 		if (error != 0)
403 			return (error);
404 
405 		lastpos = s->zst_outbuffer.pos;
406 	}
407 
408 	/*
409 	 * We've already done an intermediate flush, so all full blocks have
410 	 * been written.  Only a partial block remains.  Padding happens in a
411 	 * higher layer.
412 	 */
413 	if (s->zst_outbuffer.pos != 0) {
414 		error = cb(s->zst_buffer, s->zst_outbuffer.pos, s->zst_off,
415 		    arg);
416 		if (error != 0)
417 			return (error);
418 	}
419 
420 	return (0);
421 }
422 
423 static int
424 zstdio_write(void *stream, void *data, size_t len, compressor_cb_t cb,
425     void *arg)
426 {
427 	struct zstdio_stream *s;
428 	size_t lastpos, rc;
429 	int error;
430 
431 	s = stream;
432 	if (data == NULL)
433 		return (zstdio_flush(s, cb, arg));
434 
435 	s->zst_inbuffer.src = data;
436 	s->zst_inbuffer.size = len;
437 	s->zst_inbuffer.pos = 0;
438 	lastpos = 0;
439 
440 	while (s->zst_inbuffer.pos < s->zst_inbuffer.size) {
441 		rc = ZSTD_compressStream(s->zst_stream, &s->zst_outbuffer,
442 		    &s->zst_inbuffer);
443 		if (ZSTD_isError(rc)) {
444 			printf("%s: Compress failed on %p! (%s)\n",
445 			    __func__, data, ZSTD_getErrorName(rc));
446 			return (EIO);
447 		}
448 
449 		if (lastpos == s->zst_inbuffer.pos) {
450 			/*
451 			 * XXX: May need flushStream to make forward progress
452 			 */
453 			printf("ZSTD: did not make forward progress @pos %zu\n",
454 			    lastpos);
455 			return (EIO);
456 		}
457 		lastpos = s->zst_inbuffer.pos;
458 
459 		error = zst_flush_intermediate(s, cb, arg);
460 		if (error != 0)
461 			return (error);
462 	}
463 	return (0);
464 }
465 
466 static void
467 zstdio_fini(void *stream)
468 {
469 	struct zstdio_stream *s;
470 
471 	s = stream;
472 	if (s->zst_static_wkspc != NULL)
473 		free(s->zst_static_wkspc, M_COMPRESS);
474 	else
475 		ZSTD_freeCCtx(s->zst_stream);
476 	free(s->zst_buffer, M_COMPRESS);
477 	free(s, M_COMPRESS);
478 }
479 
480 static struct compressor_methods zstd_methods = {
481 	.format = COMPRESS_ZSTD,
482 	.init = zstdio_init,
483 	.reset = zstdio_reset,
484 	.write = zstdio_write,
485 	.fini = zstdio_fini,
486 };
487 DATA_SET(compressors, zstd_methods);
488 
489 #endif /* ZSTDIO */
490 
491 bool
492 compressor_avail(int format)
493 {
494 	struct compressor_methods **iter;
495 
496 	SET_FOREACH(iter, compressors) {
497 		if ((*iter)->format == format)
498 			return (true);
499 	}
500 	return (false);
501 }
502 
503 struct compressor *
504 compressor_init(compressor_cb_t cb, int format, size_t maxiosize, int level,
505     void *arg)
506 {
507 	struct compressor_methods **iter;
508 	struct compressor *s;
509 	void *priv;
510 
511 	SET_FOREACH(iter, compressors) {
512 		if ((*iter)->format == format)
513 			break;
514 	}
515 	if (iter == SET_LIMIT(compressors))
516 		return (NULL);
517 
518 	priv = (*iter)->init(maxiosize, level);
519 	if (priv == NULL)
520 		return (NULL);
521 
522 	s = malloc(sizeof(*s), M_COMPRESS, M_WAITOK | M_ZERO);
523 	s->methods = (*iter);
524 	s->priv = priv;
525 	s->cb = cb;
526 	s->arg = arg;
527 	return (s);
528 }
529 
530 void
531 compressor_reset(struct compressor *stream)
532 {
533 
534 	stream->methods->reset(stream->priv);
535 }
536 
537 int
538 compressor_write(struct compressor *stream, void *data, size_t len)
539 {
540 
541 	return (stream->methods->write(stream->priv, data, len, stream->cb,
542 	    stream->arg));
543 }
544 
545 int
546 compressor_flush(struct compressor *stream)
547 {
548 
549 	return (stream->methods->write(stream->priv, NULL, 0, stream->cb,
550 	    stream->arg));
551 }
552 
553 void
554 compressor_fini(struct compressor *stream)
555 {
556 
557 	stream->methods->fini(stream->priv);
558 }
559