cksum,df,digest: prefer signed types
[coreutils.git] / src / dd.c
blob21f7a4edea6ec6e371d1b08334f4e296f38d6330
1 /* dd -- convert a file while copying it.
2 Copyright (C) 1985-2023 Free Software Foundation, Inc.
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 3 of the License, or
7 (at your option) any later version.
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <https://www.gnu.org/licenses/>. */
17 /* Written by Paul Rubin, David MacKenzie, and Stuart Kemp. */
19 #include <config.h>
21 #include <sys/types.h>
22 #include <signal.h>
23 #include <stdckdint.h>
25 #include "system.h"
26 #include "alignalloc.h"
27 #include "close-stream.h"
28 #include "fd-reopen.h"
29 #include "gethrxtime.h"
30 #include "human.h"
31 #include "ioblksize.h"
32 #include "long-options.h"
33 #include "quote.h"
34 #include "verror.h"
35 #include "xstrtol.h"
36 #include "xtime.h"
38 /* The official name of this program (e.g., no 'g' prefix). */
39 #define PROGRAM_NAME "dd"
41 #define AUTHORS \
42 proper_name ("Paul Rubin"), \
43 proper_name ("David MacKenzie"), \
44 proper_name ("Stuart Kemp")
46 /* Use SA_NOCLDSTOP as a proxy for whether the sigaction machinery is
47 present. */
48 #ifndef SA_NOCLDSTOP
49 # define SA_NOCLDSTOP 0
50 # define sigprocmask(How, Set, Oset) /* empty */
51 # define sigset_t int
52 # if ! HAVE_SIGINTERRUPT
53 # define siginterrupt(sig, flag) /* empty */
54 # endif
55 #endif
57 /* NonStop circa 2011 lacks SA_RESETHAND; see Bug#9076. */
58 #ifndef SA_RESETHAND
59 # define SA_RESETHAND 0
60 #endif
62 #ifndef SIGINFO
63 # define SIGINFO SIGUSR1
64 #endif
66 /* This may belong in GNULIB's fcntl module instead.
67 Define O_CIO to 0 if it is not supported by this OS. */
68 #ifndef O_CIO
69 # define O_CIO 0
70 #endif
72 /* On AIX 5.1 and AIX 5.2, O_NOCACHE is defined via <fcntl.h>
73 and would interfere with our use of that name, below. */
74 #undef O_NOCACHE
76 #if ! HAVE_FDATASYNC
77 # define fdatasync(fd) (errno = ENOSYS, -1)
78 #endif
80 #define output_char(c) \
81 do \
82 { \
83 obuf[oc++] = (c); \
84 if (oc >= output_blocksize) \
85 write_output (); \
86 } \
87 while (0)
89 /* Default input and output blocksize. */
90 #define DEFAULT_BLOCKSIZE 512
92 /* Conversions bit masks. */
93 enum
95 C_ASCII = 01,
97 C_EBCDIC = 02,
98 C_IBM = 04,
99 C_BLOCK = 010,
100 C_UNBLOCK = 020,
101 C_LCASE = 040,
102 C_UCASE = 0100,
103 C_SWAB = 0200,
104 C_NOERROR = 0400,
105 C_NOTRUNC = 01000,
106 C_SYNC = 02000,
108 /* Use separate input and output buffers, and combine partial
109 input blocks. */
110 C_TWOBUFS = 04000,
112 C_NOCREAT = 010000,
113 C_EXCL = 020000,
114 C_FDATASYNC = 040000,
115 C_FSYNC = 0100000,
117 C_SPARSE = 0200000
120 /* Status levels. */
121 enum
123 STATUS_NONE = 1,
124 STATUS_NOXFER = 2,
125 STATUS_DEFAULT = 3,
126 STATUS_PROGRESS = 4
129 /* The name of the input file, or nullptr for the standard input. */
130 static char const *input_file = nullptr;
132 /* The name of the output file, or nullptr for the standard output. */
133 static char const *output_file = nullptr;
135 /* The page size on this host. */
136 static idx_t page_size;
138 /* The number of bytes in which atomic reads are done. */
139 static idx_t input_blocksize = 0;
141 /* The number of bytes in which atomic writes are done. */
142 static idx_t output_blocksize = 0;
144 /* Conversion buffer size, in bytes. 0 prevents conversions. */
145 static idx_t conversion_blocksize = 0;
147 /* Skip this many records of 'input_blocksize' bytes before input. */
148 static intmax_t skip_records = 0;
150 /* Skip this many bytes before input in addition of 'skip_records'
151 records. */
152 static idx_t skip_bytes = 0;
154 /* Skip this many records of 'output_blocksize' bytes before output. */
155 static intmax_t seek_records = 0;
157 /* Skip this many bytes in addition to 'seek_records' records before
158 output. */
159 static intmax_t seek_bytes = 0;
161 /* Whether the final output was done with a seek (rather than a write). */
162 static bool final_op_was_seek;
164 /* Copy only this many records. The default is effectively infinity. */
165 static intmax_t max_records = INTMAX_MAX;
167 /* Copy this many bytes in addition to 'max_records' records. */
168 static idx_t max_bytes = 0;
170 /* Bit vector of conversions to apply. */
171 static int conversions_mask = 0;
173 /* Open flags for the input and output files. */
174 static int input_flags = 0;
175 static int output_flags = 0;
177 /* Status flags for what is printed to stderr. */
178 static int status_level = STATUS_DEFAULT;
180 /* If nonzero, filter characters through the translation table. */
181 static bool translation_needed = false;
183 /* Number of partial blocks written. */
184 static intmax_t w_partial = 0;
186 /* Number of full blocks written. */
187 static intmax_t w_full = 0;
189 /* Number of partial blocks read. */
190 static intmax_t r_partial = 0;
192 /* Number of full blocks read. */
193 static intmax_t r_full = 0;
195 /* Number of bytes written. */
196 static intmax_t w_bytes = 0;
198 /* Last-reported number of bytes written, or negative if never reported. */
199 static intmax_t reported_w_bytes = -1;
201 /* Time that dd started. */
202 static xtime_t start_time;
204 /* Next time to report periodic progress. */
205 static xtime_t next_time;
207 /* If positive, the number of bytes output in the current progress line. */
208 static int progress_len;
210 /* True if input is seekable. */
211 static bool input_seekable;
213 /* Error number corresponding to initial attempt to lseek input.
214 If ESPIPE, do not issue any more diagnostics about it. */
215 static int input_seek_errno;
217 /* File offset of the input, in bytes, or -1 if it overflowed. */
218 static off_t input_offset;
220 /* True if a partial read should be diagnosed. */
221 static bool warn_partial_read;
223 /* Records truncated by conv=block. */
224 static intmax_t r_truncate = 0;
226 /* Output representation of newline and space characters.
227 They change if we're converting to EBCDIC. */
228 static char newline_character = '\n';
229 static char space_character = ' ';
231 /* I/O buffers. */
232 static char *ibuf;
233 static char *obuf;
235 /* Current index into 'obuf'. */
236 static idx_t oc = 0;
238 /* Index into current line, for 'conv=block' and 'conv=unblock'. */
239 static idx_t col = 0;
241 /* The set of signals that are caught. */
242 static sigset_t caught_signals;
244 /* If nonzero, the value of the pending fatal signal. */
245 static sig_atomic_t volatile interrupt_signal;
247 /* A count of the number of pending info signals that have been received. */
248 static sig_atomic_t volatile info_signal_count;
250 /* Whether to discard cache for input or output. */
251 static bool i_nocache, o_nocache;
253 /* Whether to instruct the kernel to discard the complete file. */
254 static bool i_nocache_eof, o_nocache_eof;
256 /* Function used for read (to handle iflag=fullblock parameter). */
257 static ssize_t (*iread_fnc) (int fd, char *buf, idx_t size);
259 /* A longest symbol in the struct symbol_values tables below. */
260 #define LONGEST_SYMBOL "count_bytes"
262 /* A symbol and the corresponding integer value. */
263 struct symbol_value
265 char symbol[sizeof LONGEST_SYMBOL];
266 int value;
269 /* Conversion symbols, for conv="...". */
270 static struct symbol_value const conversions[] =
272 {"ascii", C_ASCII | C_UNBLOCK | C_TWOBUFS}, /* EBCDIC to ASCII. */
273 {"ebcdic", C_EBCDIC | C_BLOCK | C_TWOBUFS}, /* ASCII to EBCDIC. */
274 {"ibm", C_IBM | C_BLOCK | C_TWOBUFS}, /* Different ASCII to EBCDIC. */
275 {"block", C_BLOCK | C_TWOBUFS}, /* Variable to fixed length records. */
276 {"unblock", C_UNBLOCK | C_TWOBUFS}, /* Fixed to variable length records. */
277 {"lcase", C_LCASE | C_TWOBUFS}, /* Translate upper to lower case. */
278 {"ucase", C_UCASE | C_TWOBUFS}, /* Translate lower to upper case. */
279 {"sparse", C_SPARSE}, /* Try to sparsely write output. */
280 {"swab", C_SWAB | C_TWOBUFS}, /* Swap bytes of input. */
281 {"noerror", C_NOERROR}, /* Ignore i/o errors. */
282 {"nocreat", C_NOCREAT}, /* Do not create output file. */
283 {"excl", C_EXCL}, /* Fail if the output file already exists. */
284 {"notrunc", C_NOTRUNC}, /* Do not truncate output file. */
285 {"sync", C_SYNC}, /* Pad input records to ibs with NULs. */
286 {"fdatasync", C_FDATASYNC}, /* Synchronize output data before finishing. */
287 {"fsync", C_FSYNC}, /* Also synchronize output metadata. */
288 {"", 0}
291 #define FFS_MASK(x) ((x) ^ ((x) & ((x) - 1)))
292 enum
294 /* Compute a value that's bitwise disjoint from the union
295 of all O_ values. */
296 v = ~(0
297 | O_APPEND
298 | O_BINARY
299 | O_CIO
300 | O_DIRECT
301 | O_DIRECTORY
302 | O_DSYNC
303 | O_NOATIME
304 | O_NOCTTY
305 | O_NOFOLLOW
306 | O_NOLINKS
307 | O_NONBLOCK
308 | O_SYNC
309 | O_TEXT
312 /* Use its lowest bits for private flags. */
313 O_FULLBLOCK = FFS_MASK (v),
314 v2 = v ^ O_FULLBLOCK,
316 O_NOCACHE = FFS_MASK (v2),
317 v3 = v2 ^ O_NOCACHE,
319 O_COUNT_BYTES = FFS_MASK (v3),
320 v4 = v3 ^ O_COUNT_BYTES,
322 O_SKIP_BYTES = FFS_MASK (v4),
323 v5 = v4 ^ O_SKIP_BYTES,
325 O_SEEK_BYTES = FFS_MASK (v5)
328 /* Ensure that we got something. */
329 static_assert (O_FULLBLOCK != 0);
330 static_assert (O_NOCACHE != 0);
331 static_assert (O_COUNT_BYTES != 0);
332 static_assert (O_SKIP_BYTES != 0);
333 static_assert (O_SEEK_BYTES != 0);
335 #define MULTIPLE_BITS_SET(i) (((i) & ((i) - 1)) != 0)
337 /* Ensure that this is a single-bit value. */
338 static_assert ( ! MULTIPLE_BITS_SET (O_FULLBLOCK));
339 static_assert ( ! MULTIPLE_BITS_SET (O_NOCACHE));
340 static_assert ( ! MULTIPLE_BITS_SET (O_COUNT_BYTES));
341 static_assert ( ! MULTIPLE_BITS_SET (O_SKIP_BYTES));
342 static_assert ( ! MULTIPLE_BITS_SET (O_SEEK_BYTES));
344 /* Flags, for iflag="..." and oflag="...". */
345 static struct symbol_value const flags[] =
347 {"append", O_APPEND},
348 {"binary", O_BINARY},
349 {"cio", O_CIO},
350 {"direct", O_DIRECT},
351 {"directory", O_DIRECTORY},
352 {"dsync", O_DSYNC},
353 {"noatime", O_NOATIME},
354 {"nocache", O_NOCACHE}, /* Discard cache. */
355 {"noctty", O_NOCTTY},
356 {"nofollow", HAVE_WORKING_O_NOFOLLOW ? O_NOFOLLOW : 0},
357 {"nolinks", O_NOLINKS},
358 {"nonblock", O_NONBLOCK},
359 {"sync", O_SYNC},
360 {"text", O_TEXT},
361 {"fullblock", O_FULLBLOCK}, /* Accumulate full blocks from input. */
362 {"count_bytes", O_COUNT_BYTES},
363 {"skip_bytes", O_SKIP_BYTES},
364 {"seek_bytes", O_SEEK_BYTES},
365 {"", 0}
368 /* Status, for status="...". */
369 static struct symbol_value const statuses[] =
371 {"none", STATUS_NONE},
372 {"noxfer", STATUS_NOXFER},
373 {"progress", STATUS_PROGRESS},
374 {"", 0}
377 /* Translation table formed by applying successive transformations. */
378 static unsigned char trans_table[256];
380 /* Standard translation tables, taken from POSIX 1003.1-2013.
381 Beware of imitations; there are lots of ASCII<->EBCDIC tables
382 floating around the net, perhaps valid for some applications but
383 not correct here. */
385 static char const ascii_to_ebcdic[] =
387 '\000', '\001', '\002', '\003', '\067', '\055', '\056', '\057',
388 '\026', '\005', '\045', '\013', '\014', '\015', '\016', '\017',
389 '\020', '\021', '\022', '\023', '\074', '\075', '\062', '\046',
390 '\030', '\031', '\077', '\047', '\034', '\035', '\036', '\037',
391 '\100', '\132', '\177', '\173', '\133', '\154', '\120', '\175',
392 '\115', '\135', '\134', '\116', '\153', '\140', '\113', '\141',
393 '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
394 '\370', '\371', '\172', '\136', '\114', '\176', '\156', '\157',
395 '\174', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
396 '\310', '\311', '\321', '\322', '\323', '\324', '\325', '\326',
397 '\327', '\330', '\331', '\342', '\343', '\344', '\345', '\346',
398 '\347', '\350', '\351', '\255', '\340', '\275', '\232', '\155',
399 '\171', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
400 '\210', '\211', '\221', '\222', '\223', '\224', '\225', '\226',
401 '\227', '\230', '\231', '\242', '\243', '\244', '\245', '\246',
402 '\247', '\250', '\251', '\300', '\117', '\320', '\137', '\007',
403 '\040', '\041', '\042', '\043', '\044', '\025', '\006', '\027',
404 '\050', '\051', '\052', '\053', '\054', '\011', '\012', '\033',
405 '\060', '\061', '\032', '\063', '\064', '\065', '\066', '\010',
406 '\070', '\071', '\072', '\073', '\004', '\024', '\076', '\341',
407 '\101', '\102', '\103', '\104', '\105', '\106', '\107', '\110',
408 '\111', '\121', '\122', '\123', '\124', '\125', '\126', '\127',
409 '\130', '\131', '\142', '\143', '\144', '\145', '\146', '\147',
410 '\150', '\151', '\160', '\161', '\162', '\163', '\164', '\165',
411 '\166', '\167', '\170', '\200', '\212', '\213', '\214', '\215',
412 '\216', '\217', '\220', '\152', '\233', '\234', '\235', '\236',
413 '\237', '\240', '\252', '\253', '\254', '\112', '\256', '\257',
414 '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
415 '\270', '\271', '\272', '\273', '\274', '\241', '\276', '\277',
416 '\312', '\313', '\314', '\315', '\316', '\317', '\332', '\333',
417 '\334', '\335', '\336', '\337', '\352', '\353', '\354', '\355',
418 '\356', '\357', '\372', '\373', '\374', '\375', '\376', '\377'
421 static char const ascii_to_ibm[] =
423 '\000', '\001', '\002', '\003', '\067', '\055', '\056', '\057',
424 '\026', '\005', '\045', '\013', '\014', '\015', '\016', '\017',
425 '\020', '\021', '\022', '\023', '\074', '\075', '\062', '\046',
426 '\030', '\031', '\077', '\047', '\034', '\035', '\036', '\037',
427 '\100', '\132', '\177', '\173', '\133', '\154', '\120', '\175',
428 '\115', '\135', '\134', '\116', '\153', '\140', '\113', '\141',
429 '\360', '\361', '\362', '\363', '\364', '\365', '\366', '\367',
430 '\370', '\371', '\172', '\136', '\114', '\176', '\156', '\157',
431 '\174', '\301', '\302', '\303', '\304', '\305', '\306', '\307',
432 '\310', '\311', '\321', '\322', '\323', '\324', '\325', '\326',
433 '\327', '\330', '\331', '\342', '\343', '\344', '\345', '\346',
434 '\347', '\350', '\351', '\255', '\340', '\275', '\137', '\155',
435 '\171', '\201', '\202', '\203', '\204', '\205', '\206', '\207',
436 '\210', '\211', '\221', '\222', '\223', '\224', '\225', '\226',
437 '\227', '\230', '\231', '\242', '\243', '\244', '\245', '\246',
438 '\247', '\250', '\251', '\300', '\117', '\320', '\241', '\007',
439 '\040', '\041', '\042', '\043', '\044', '\025', '\006', '\027',
440 '\050', '\051', '\052', '\053', '\054', '\011', '\012', '\033',
441 '\060', '\061', '\032', '\063', '\064', '\065', '\066', '\010',
442 '\070', '\071', '\072', '\073', '\004', '\024', '\076', '\341',
443 '\101', '\102', '\103', '\104', '\105', '\106', '\107', '\110',
444 '\111', '\121', '\122', '\123', '\124', '\125', '\126', '\127',
445 '\130', '\131', '\142', '\143', '\144', '\145', '\146', '\147',
446 '\150', '\151', '\160', '\161', '\162', '\163', '\164', '\165',
447 '\166', '\167', '\170', '\200', '\212', '\213', '\214', '\215',
448 '\216', '\217', '\220', '\232', '\233', '\234', '\235', '\236',
449 '\237', '\240', '\252', '\253', '\254', '\255', '\256', '\257',
450 '\260', '\261', '\262', '\263', '\264', '\265', '\266', '\267',
451 '\270', '\271', '\272', '\273', '\274', '\275', '\276', '\277',
452 '\312', '\313', '\314', '\315', '\316', '\317', '\332', '\333',
453 '\334', '\335', '\336', '\337', '\352', '\353', '\354', '\355',
454 '\356', '\357', '\372', '\373', '\374', '\375', '\376', '\377'
457 static char const ebcdic_to_ascii[] =
459 '\000', '\001', '\002', '\003', '\234', '\011', '\206', '\177',
460 '\227', '\215', '\216', '\013', '\014', '\015', '\016', '\017',
461 '\020', '\021', '\022', '\023', '\235', '\205', '\010', '\207',
462 '\030', '\031', '\222', '\217', '\034', '\035', '\036', '\037',
463 '\200', '\201', '\202', '\203', '\204', '\012', '\027', '\033',
464 '\210', '\211', '\212', '\213', '\214', '\005', '\006', '\007',
465 '\220', '\221', '\026', '\223', '\224', '\225', '\226', '\004',
466 '\230', '\231', '\232', '\233', '\024', '\025', '\236', '\032',
467 '\040', '\240', '\241', '\242', '\243', '\244', '\245', '\246',
468 '\247', '\250', '\325', '\056', '\074', '\050', '\053', '\174',
469 '\046', '\251', '\252', '\253', '\254', '\255', '\256', '\257',
470 '\260', '\261', '\041', '\044', '\052', '\051', '\073', '\176',
471 '\055', '\057', '\262', '\263', '\264', '\265', '\266', '\267',
472 '\270', '\271', '\313', '\054', '\045', '\137', '\076', '\077',
473 '\272', '\273', '\274', '\275', '\276', '\277', '\300', '\301',
474 '\302', '\140', '\072', '\043', '\100', '\047', '\075', '\042',
475 '\303', '\141', '\142', '\143', '\144', '\145', '\146', '\147',
476 '\150', '\151', '\304', '\305', '\306', '\307', '\310', '\311',
477 '\312', '\152', '\153', '\154', '\155', '\156', '\157', '\160',
478 '\161', '\162', '\136', '\314', '\315', '\316', '\317', '\320',
479 '\321', '\345', '\163', '\164', '\165', '\166', '\167', '\170',
480 '\171', '\172', '\322', '\323', '\324', '\133', '\326', '\327',
481 '\330', '\331', '\332', '\333', '\334', '\335', '\336', '\337',
482 '\340', '\341', '\342', '\343', '\344', '\135', '\346', '\347',
483 '\173', '\101', '\102', '\103', '\104', '\105', '\106', '\107',
484 '\110', '\111', '\350', '\351', '\352', '\353', '\354', '\355',
485 '\175', '\112', '\113', '\114', '\115', '\116', '\117', '\120',
486 '\121', '\122', '\356', '\357', '\360', '\361', '\362', '\363',
487 '\134', '\237', '\123', '\124', '\125', '\126', '\127', '\130',
488 '\131', '\132', '\364', '\365', '\366', '\367', '\370', '\371',
489 '\060', '\061', '\062', '\063', '\064', '\065', '\066', '\067',
490 '\070', '\071', '\372', '\373', '\374', '\375', '\376', '\377'
493 /* True if we need to close the standard output *stream*. */
494 static bool close_stdout_required = true;
496 /* The only reason to close the standard output *stream* is if
497 parse_long_options fails (as it does for --help or --version).
498 In any other case, dd uses only the STDOUT_FILENO file descriptor,
499 and the "cleanup" function calls "close (STDOUT_FILENO)".
500 Closing the file descriptor and then letting the usual atexit-run
501 close_stdout function call "fclose (stdout)" would result in a
502 harmless failure of the close syscall (with errno EBADF).
503 This function serves solely to avoid the unnecessary close_stdout
504 call, once parse_long_options has succeeded.
505 Meanwhile, we guarantee that the standard error stream is flushed,
506 by inlining the last half of close_stdout as needed. */
507 static void
508 maybe_close_stdout (void)
510 if (close_stdout_required)
511 close_stdout ();
512 else if (close_stream (stderr) != 0)
513 _exit (EXIT_FAILURE);
516 /* Like the 'error' function but handle any pending newline,
517 and do not exit. */
519 ATTRIBUTE_FORMAT ((__printf__, 2, 3))
520 static void
521 diagnose (int errnum, char const *fmt, ...)
523 if (0 < progress_len)
525 fputc ('\n', stderr);
526 progress_len = 0;
529 va_list ap;
530 va_start (ap, fmt);
531 verror (0, errnum, fmt, ap);
532 va_end (ap);
535 void
536 usage (int status)
538 if (status != EXIT_SUCCESS)
539 emit_try_help ();
540 else
542 printf (_("\
543 Usage: %s [OPERAND]...\n\
544 or: %s OPTION\n\
546 program_name, program_name);
547 fputs (_("\
548 Copy a file, converting and formatting according to the operands.\n\
550 bs=BYTES read and write up to BYTES bytes at a time (default: 512);\n\
551 overrides ibs and obs\n\
552 cbs=BYTES convert BYTES bytes at a time\n\
553 conv=CONVS convert the file as per the comma separated symbol list\n\
554 count=N copy only N input blocks\n\
555 ibs=BYTES read up to BYTES bytes at a time (default: 512)\n\
556 "), stdout);
557 fputs (_("\
558 if=FILE read from FILE instead of stdin\n\
559 iflag=FLAGS read as per the comma separated symbol list\n\
560 obs=BYTES write BYTES bytes at a time (default: 512)\n\
561 of=FILE write to FILE instead of stdout\n\
562 oflag=FLAGS write as per the comma separated symbol list\n\
563 seek=N (or oseek=N) skip N obs-sized output blocks\n\
564 skip=N (or iseek=N) skip N ibs-sized input blocks\n\
565 status=LEVEL The LEVEL of information to print to stderr;\n\
566 'none' suppresses everything but error messages,\n\
567 'noxfer' suppresses the final transfer statistics,\n\
568 'progress' shows periodic transfer statistics\n\
569 "), stdout);
570 fputs (_("\
572 N and BYTES may be followed by the following multiplicative suffixes:\n\
573 c=1, w=2, b=512, kB=1000, K=1024, MB=1000*1000, M=1024*1024, xM=M,\n\
574 GB=1000*1000*1000, G=1024*1024*1024, and so on for T, P, E, Z, Y, R, Q.\n\
575 Binary prefixes can be used, too: KiB=K, MiB=M, and so on.\n\
576 If N ends in 'B', it counts bytes not blocks.\n\
578 Each CONV symbol may be:\n\
580 "), stdout);
581 fputs (_("\
582 ascii from EBCDIC to ASCII\n\
583 ebcdic from ASCII to EBCDIC\n\
584 ibm from ASCII to alternate EBCDIC\n\
585 block pad newline-terminated records with spaces to cbs-size\n\
586 unblock replace trailing spaces in cbs-size records with newline\n\
587 lcase change upper case to lower case\n\
588 ucase change lower case to upper case\n\
589 sparse try to seek rather than write all-NUL output blocks\n\
590 swab swap every pair of input bytes\n\
591 sync pad every input block with NULs to ibs-size; when used\n\
592 with block or unblock, pad with spaces rather than NULs\n\
593 "), stdout);
594 fputs (_("\
595 excl fail if the output file already exists\n\
596 nocreat do not create the output file\n\
597 notrunc do not truncate the output file\n\
598 noerror continue after read errors\n\
599 fdatasync physically write output file data before finishing\n\
600 fsync likewise, but also write metadata\n\
601 "), stdout);
602 fputs (_("\
604 Each FLAG symbol may be:\n\
606 append append mode (makes sense only for output; conv=notrunc suggested)\n\
607 "), stdout);
608 if (O_CIO)
609 fputs (_(" cio use concurrent I/O for data\n"), stdout);
610 if (O_DIRECT)
611 fputs (_(" direct use direct I/O for data\n"), stdout);
612 if (O_DIRECTORY)
613 fputs (_(" directory fail unless a directory\n"), stdout);
614 if (O_DSYNC)
615 fputs (_(" dsync use synchronized I/O for data\n"), stdout);
616 if (O_SYNC)
617 fputs (_(" sync likewise, but also for metadata\n"), stdout);
618 fputs (_(" fullblock accumulate full blocks of input (iflag only)\n"),
619 stdout);
620 if (O_NONBLOCK)
621 fputs (_(" nonblock use non-blocking I/O\n"), stdout);
622 if (O_NOATIME)
623 fputs (_(" noatime do not update access time\n"), stdout);
624 #if HAVE_POSIX_FADVISE
625 if (O_NOCACHE)
626 fputs (_(" nocache Request to drop cache. See also oflag=sync\n"),
627 stdout);
628 #endif
629 if (O_NOCTTY)
630 fputs (_(" noctty do not assign controlling terminal from file\n"),
631 stdout);
632 if (HAVE_WORKING_O_NOFOLLOW)
633 fputs (_(" nofollow do not follow symlinks\n"), stdout);
634 if (O_NOLINKS)
635 fputs (_(" nolinks fail if multiply-linked\n"), stdout);
636 if (O_BINARY)
637 fputs (_(" binary use binary I/O for data\n"), stdout);
638 if (O_TEXT)
639 fputs (_(" text use text I/O for data\n"), stdout);
642 printf (_("\
644 Sending a %s signal to a running 'dd' process makes it\n\
645 print I/O statistics to standard error and then resume copying.\n\
647 Options are:\n\
649 "), SIGINFO == SIGUSR1 ? "USR1" : "INFO");
652 fputs (HELP_OPTION_DESCRIPTION, stdout);
653 fputs (VERSION_OPTION_DESCRIPTION, stdout);
654 emit_ancillary_info (PROGRAM_NAME);
656 exit (status);
659 /* Common options to use when displaying sizes and rates. */
661 enum { human_opts = (human_autoscale | human_round_to_nearest
662 | human_space_before_unit | human_SI | human_B) };
664 /* Ensure input buffer IBUF is allocated. */
666 static void
667 alloc_ibuf (void)
669 if (ibuf)
670 return;
672 bool extra_byte_for_swab = !!(conversions_mask & C_SWAB);
673 ibuf = alignalloc (page_size, input_blocksize + extra_byte_for_swab);
674 if (!ibuf)
676 char hbuf[LONGEST_HUMAN_READABLE + 1];
677 error (EXIT_FAILURE, 0,
678 _("memory exhausted by input buffer of size %td bytes (%s)"),
679 input_blocksize,
680 human_readable (input_blocksize, hbuf,
681 human_opts | human_base_1024, 1, 1));
685 /* Ensure output buffer OBUF is allocated/initialized. */
687 static void
688 alloc_obuf (void)
690 if (obuf)
691 return;
693 if (conversions_mask & C_TWOBUFS)
695 obuf = alignalloc (page_size, output_blocksize);
696 if (!obuf)
698 char hbuf[LONGEST_HUMAN_READABLE + 1];
699 error (EXIT_FAILURE, 0,
700 _("memory exhausted by output buffer of size %td"
701 " bytes (%s)"),
702 output_blocksize,
703 human_readable (output_blocksize, hbuf,
704 human_opts | human_base_1024, 1, 1));
707 else
709 alloc_ibuf ();
710 obuf = ibuf;
714 static void
715 translate_charset (char const *new_trans)
717 for (int i = 0; i < 256; i++)
718 trans_table[i] = new_trans[trans_table[i]];
719 translation_needed = true;
722 /* Return true if I has more than one bit set. I must be nonnegative. */
724 static inline bool
725 multiple_bits_set (int i)
727 return MULTIPLE_BITS_SET (i);
730 static bool
731 abbreviation_lacks_prefix (char const *message)
733 return message[strlen (message) - 2] == ' ';
736 /* Print transfer statistics. */
738 static void
739 print_xfer_stats (xtime_t progress_time)
741 xtime_t now = progress_time ? progress_time : gethrxtime ();
742 static char const slash_s[] = "/s";
743 char hbuf[3][LONGEST_HUMAN_READABLE + sizeof slash_s];
744 double delta_s;
745 char const *bytes_per_second;
746 char const *si = human_readable (w_bytes, hbuf[0], human_opts, 1, 1);
747 char const *iec = human_readable (w_bytes, hbuf[1],
748 human_opts | human_base_1024, 1, 1);
750 /* Use integer arithmetic to compute the transfer rate,
751 since that makes it easy to use SI abbreviations. */
752 char *bpsbuf = hbuf[2];
753 int bpsbufsize = sizeof hbuf[2];
754 if (start_time < now)
756 double XTIME_PRECISIONe0 = XTIME_PRECISION;
757 xtime_t delta_xtime = now - start_time;
758 delta_s = delta_xtime / XTIME_PRECISIONe0;
759 bytes_per_second = human_readable (w_bytes, bpsbuf, human_opts,
760 XTIME_PRECISION, delta_xtime);
761 strcat (bytes_per_second - bpsbuf + bpsbuf, slash_s);
763 else
765 delta_s = 0;
766 snprintf (bpsbuf, bpsbufsize, "%s B/s", _("Infinity"));
767 bytes_per_second = bpsbuf;
770 if (progress_time)
771 fputc ('\r', stderr);
773 /* Use full seconds when printing progress, since the progress
774 report is output once per second and there is little point
775 displaying any subsecond jitter. Use default precision with %g
776 otherwise, as this provides more-useful output then. With long
777 transfers %g can generate a number with an exponent; that is OK. */
778 char delta_s_buf[24];
779 snprintf (delta_s_buf, sizeof delta_s_buf,
780 progress_time ? "%.0f s" : "%g s", delta_s);
782 int stats_len
783 = (abbreviation_lacks_prefix (si)
784 ? fprintf (stderr,
785 ngettext ("%"PRIdMAX" byte copied, %s, %s",
786 "%"PRIdMAX" bytes copied, %s, %s",
787 select_plural (w_bytes)),
788 w_bytes, delta_s_buf, bytes_per_second)
789 : abbreviation_lacks_prefix (iec)
790 ? fprintf (stderr,
791 _("%"PRIdMAX" bytes (%s) copied, %s, %s"),
792 w_bytes, si, delta_s_buf, bytes_per_second)
793 : fprintf (stderr,
794 _("%"PRIdMAX" bytes (%s, %s) copied, %s, %s"),
795 w_bytes, si, iec, delta_s_buf, bytes_per_second));
797 if (progress_time)
799 /* Erase any trailing junk on the output line by outputting
800 spaces. In theory this could glitch the display because the
801 formatted translation of a line describing a larger file
802 could consume fewer screen columns than the strlen difference
803 from the previously formatted translation. In practice this
804 does not seem to be a problem. */
805 if (0 <= stats_len && stats_len < progress_len)
806 fprintf (stderr, "%*s", progress_len - stats_len, "");
807 progress_len = stats_len;
809 else
810 fputc ('\n', stderr);
812 reported_w_bytes = w_bytes;
815 static void
816 print_stats (void)
818 if (status_level == STATUS_NONE)
819 return;
821 if (0 < progress_len)
823 fputc ('\n', stderr);
824 progress_len = 0;
827 fprintf (stderr,
828 _("%"PRIdMAX"+%"PRIdMAX" records in\n"
829 "%"PRIdMAX"+%"PRIdMAX" records out\n"),
830 r_full, r_partial, w_full, w_partial);
832 if (r_truncate != 0)
833 fprintf (stderr,
834 ngettext ("%"PRIdMAX" truncated record\n",
835 "%"PRIdMAX" truncated records\n",
836 select_plural (r_truncate)),
837 r_truncate);
839 if (status_level == STATUS_NOXFER)
840 return;
842 print_xfer_stats (0);
845 /* An ordinary signal was received; arrange for the program to exit. */
847 static void
848 interrupt_handler (int sig)
850 if (! SA_RESETHAND)
851 signal (sig, SIG_DFL);
852 interrupt_signal = sig;
855 /* An info signal was received; arrange for the program to print status. */
857 static void
858 siginfo_handler (int sig)
860 if (! SA_NOCLDSTOP)
861 signal (sig, siginfo_handler);
862 info_signal_count++;
865 /* Install the signal handlers. */
867 static void
868 install_signal_handlers (void)
870 bool catch_siginfo = ! (SIGINFO == SIGUSR1 && getenv ("POSIXLY_CORRECT"));
872 #if SA_NOCLDSTOP
874 struct sigaction act;
875 sigemptyset (&caught_signals);
876 if (catch_siginfo)
877 sigaddset (&caught_signals, SIGINFO);
878 sigaction (SIGINT, nullptr, &act);
879 if (act.sa_handler != SIG_IGN)
880 sigaddset (&caught_signals, SIGINT);
881 act.sa_mask = caught_signals;
883 if (sigismember (&caught_signals, SIGINFO))
885 act.sa_handler = siginfo_handler;
886 /* Note we don't use SA_RESTART here and instead
887 handle EINTR explicitly in iftruncate etc.
888 to avoid blocking on noncommitted read/write calls. */
889 act.sa_flags = 0;
890 sigaction (SIGINFO, &act, nullptr);
893 if (sigismember (&caught_signals, SIGINT))
895 act.sa_handler = interrupt_handler;
896 act.sa_flags = SA_NODEFER | SA_RESETHAND;
897 sigaction (SIGINT, &act, nullptr);
900 #else
902 if (catch_siginfo)
904 signal (SIGINFO, siginfo_handler);
905 siginterrupt (SIGINFO, 1);
907 if (signal (SIGINT, SIG_IGN) != SIG_IGN)
909 signal (SIGINT, interrupt_handler);
910 siginterrupt (SIGINT, 1);
912 #endif
915 /* Close FD. Return 0 if successful, -1 (setting errno) otherwise.
916 If close fails with errno == EINTR, POSIX says the file descriptor
917 is in an unspecified state, so keep trying to close FD but do not
918 consider EBADF to be an error. Do not process signals. This all
919 differs somewhat from functions like ifdatasync and ifsync. */
920 static int
921 iclose (int fd)
923 if (close (fd) != 0)
925 if (errno != EINTR)
926 return -1;
927 while (close (fd) != 0 && errno != EBADF);
929 return 0;
932 static int synchronize_output (void);
934 static void
935 cleanup (void)
937 if (!interrupt_signal)
939 int sync_status = synchronize_output ();
940 if (sync_status)
941 exit (sync_status);
944 if (iclose (STDIN_FILENO) != 0)
945 error (EXIT_FAILURE, errno, _("closing input file %s"),
946 quoteaf (input_file));
948 /* Don't remove this call to close, even though close_stdout
949 closes standard output. This close is necessary when cleanup
950 is called as a consequence of signal handling. */
951 if (iclose (STDOUT_FILENO) != 0)
952 error (EXIT_FAILURE, errno,
953 _("closing output file %s"), quoteaf (output_file));
956 /* Process any pending signals. If signals are caught, this function
957 should be called periodically. Ideally there should never be an
958 unbounded amount of time when signals are not being processed. */
960 static void
961 process_signals (void)
963 while (interrupt_signal || info_signal_count)
965 int interrupt;
966 int infos;
967 sigset_t oldset;
969 sigprocmask (SIG_BLOCK, &caught_signals, &oldset);
971 /* Reload interrupt_signal and info_signal_count, in case a new
972 signal was handled before sigprocmask took effect. */
973 interrupt = interrupt_signal;
974 infos = info_signal_count;
976 if (infos)
977 info_signal_count = infos - 1;
979 sigprocmask (SIG_SETMASK, &oldset, nullptr);
981 if (interrupt)
982 cleanup ();
983 print_stats ();
984 if (interrupt)
985 raise (interrupt);
989 static void
990 finish_up (void)
992 /* Process signals first, so that cleanup is called at most once. */
993 process_signals ();
994 cleanup ();
995 print_stats ();
998 static void
999 quit (int code)
1001 finish_up ();
1002 exit (code);
1005 /* Return LEN rounded down to a multiple of IO_BUFSIZE
1006 (to minimize calls to the expensive posix_fadvise (,POSIX_FADV_DONTNEED),
1007 while storing the remainder internally per FD.
1008 Pass LEN == 0 to get the current remainder. */
1010 static off_t
1011 cache_round (int fd, off_t len)
1013 static off_t i_pending, o_pending;
1014 off_t *pending = (fd == STDIN_FILENO ? &i_pending : &o_pending);
1016 if (len)
1018 intmax_t c_pending;
1019 if (ckd_add (&c_pending, *pending, len))
1020 c_pending = INTMAX_MAX;
1021 *pending = c_pending % IO_BUFSIZE;
1022 if (c_pending > *pending)
1023 len = c_pending - *pending;
1024 else
1025 len = 0;
1027 else
1028 len = *pending;
1030 return len;
1033 /* Discard the cache from the current offset of either
1034 STDIN_FILENO or STDOUT_FILENO.
1035 Return true on success. */
1037 static bool
1038 invalidate_cache (int fd, off_t len)
1040 int adv_ret = -1;
1041 off_t offset;
1042 bool nocache_eof = (fd == STDIN_FILENO ? i_nocache_eof : o_nocache_eof);
1044 /* Minimize syscalls. */
1045 off_t clen = cache_round (fd, len);
1046 if (len && !clen)
1047 return true; /* Don't advise this time. */
1048 else if (! len && ! clen && ! nocache_eof)
1049 return true;
1050 off_t pending = len ? cache_round (fd, 0) : 0;
1052 if (fd == STDIN_FILENO)
1054 if (input_seekable)
1055 offset = input_offset;
1056 else
1058 offset = -1;
1059 errno = ESPIPE;
1062 else
1064 static off_t output_offset = -2;
1066 if (output_offset != -1)
1068 if (output_offset < 0)
1069 output_offset = lseek (fd, 0, SEEK_CUR);
1070 else if (len)
1071 output_offset += clen + pending;
1074 offset = output_offset;
1077 if (0 <= offset)
1079 if (! len && clen && nocache_eof)
1081 pending = clen;
1082 clen = 0;
1085 /* Note we're being careful here to only invalidate what
1086 we've read, so as not to dump any read ahead cache.
1087 Note also the kernel is conservative and only invalidates
1088 full pages in the specified range. */
1089 #if HAVE_POSIX_FADVISE
1090 offset = offset - clen - pending;
1091 /* ensure full page specified when invalidating to eof. */
1092 if (clen == 0)
1093 offset -= offset % page_size;
1094 adv_ret = posix_fadvise (fd, offset, clen, POSIX_FADV_DONTNEED);
1095 #else
1096 errno = ENOTSUP;
1097 #endif
1100 return adv_ret != -1 ? true : false;
1103 /* Read from FD into the buffer BUF of size SIZE, processing any
1104 signals that arrive before bytes are read. Return the number of
1105 bytes read if successful, -1 (setting errno) on failure. */
1107 static ssize_t
1108 iread (int fd, char *buf, idx_t size)
1110 ssize_t nread;
1111 static ssize_t prev_nread;
1115 process_signals ();
1116 nread = read (fd, buf, size);
1117 /* Ignore final read error with iflag=direct as that
1118 returns EINVAL due to the non aligned file offset. */
1119 if (nread == -1 && errno == EINVAL
1120 && 0 < prev_nread && prev_nread < size
1121 && (input_flags & O_DIRECT))
1123 errno = 0;
1124 nread = 0;
1127 while (nread < 0 && errno == EINTR);
1129 /* Short read may be due to received signal. */
1130 if (0 < nread && nread < size)
1131 process_signals ();
1133 if (0 < nread && warn_partial_read)
1135 if (0 < prev_nread && prev_nread < size)
1137 idx_t prev = prev_nread;
1138 if (status_level != STATUS_NONE)
1139 diagnose (0, ngettext (("warning: partial read (%td byte); "
1140 "suggest iflag=fullblock"),
1141 ("warning: partial read (%td bytes); "
1142 "suggest iflag=fullblock"),
1143 select_plural (prev)),
1144 prev);
1145 warn_partial_read = false;
1149 prev_nread = nread;
1150 return nread;
1153 /* Wrapper around iread function to accumulate full blocks. */
1154 static ssize_t
1155 iread_fullblock (int fd, char *buf, idx_t size)
1157 ssize_t nread = 0;
1159 while (0 < size)
1161 ssize_t ncurr = iread (fd, buf, size);
1162 if (ncurr < 0)
1163 return ncurr;
1164 if (ncurr == 0)
1165 break;
1166 nread += ncurr;
1167 buf += ncurr;
1168 size -= ncurr;
1171 return nread;
1174 /* Write to FD the buffer BUF of size SIZE, processing any signals
1175 that arrive. Return the number of bytes written, setting errno if
1176 this is less than SIZE. Keep trying if there are partial
1177 writes. */
1179 static idx_t
1180 iwrite (int fd, char const *buf, idx_t size)
1182 idx_t total_written = 0;
1184 if ((output_flags & O_DIRECT) && size < output_blocksize)
1186 int old_flags = fcntl (STDOUT_FILENO, F_GETFL);
1187 if (fcntl (STDOUT_FILENO, F_SETFL, old_flags & ~O_DIRECT) != 0
1188 && status_level != STATUS_NONE)
1189 diagnose (errno, _("failed to turn off O_DIRECT: %s"),
1190 quotef (output_file));
1192 /* Since we have just turned off O_DIRECT for the final write,
1193 we try to preserve some of its semantics. */
1195 /* Call invalidate_cache to setup the appropriate offsets
1196 for subsequent calls. */
1197 o_nocache_eof = true;
1198 invalidate_cache (STDOUT_FILENO, 0);
1200 /* Attempt to ensure that that final block is committed
1201 to stable storage as quickly as possible. */
1202 conversions_mask |= C_FSYNC;
1204 /* After the subsequent fsync we'll call invalidate_cache
1205 to attempt to clear all data from the page cache. */
1208 while (total_written < size)
1210 ssize_t nwritten = 0;
1211 process_signals ();
1213 /* Perform a seek for a NUL block if sparse output is enabled. */
1214 final_op_was_seek = false;
1215 if ((conversions_mask & C_SPARSE) && is_nul (buf, size))
1217 if (lseek (fd, size, SEEK_CUR) < 0)
1219 conversions_mask &= ~C_SPARSE;
1220 /* Don't warn about the advisory sparse request. */
1222 else
1224 final_op_was_seek = true;
1225 nwritten = size;
1229 if (!nwritten)
1230 nwritten = write (fd, buf + total_written, size - total_written);
1232 if (nwritten < 0)
1234 if (errno != EINTR)
1235 break;
1237 else if (nwritten == 0)
1239 /* Some buggy drivers return 0 when one tries to write beyond
1240 a device's end. (Example: Linux kernel 1.2.13 on /dev/fd0.)
1241 Set errno to ENOSPC so they get a sensible diagnostic. */
1242 errno = ENOSPC;
1243 break;
1245 else
1246 total_written += nwritten;
1249 if (o_nocache && total_written)
1250 invalidate_cache (fd, total_written);
1252 return total_written;
1255 /* Write, then empty, the output buffer 'obuf'. */
1257 static void
1258 write_output (void)
1260 idx_t nwritten = iwrite (STDOUT_FILENO, obuf, output_blocksize);
1261 w_bytes += nwritten;
1262 if (nwritten != output_blocksize)
1264 diagnose (errno, _("writing to %s"), quoteaf (output_file));
1265 if (nwritten != 0)
1266 w_partial++;
1267 quit (EXIT_FAILURE);
1269 else
1270 w_full++;
1271 oc = 0;
1274 /* Restart on EINTR from fdatasync. */
1276 static int
1277 ifdatasync (int fd)
1279 int ret;
1283 process_signals ();
1284 ret = fdatasync (fd);
1286 while (ret < 0 && errno == EINTR);
1288 return ret;
1291 /* Restart on EINTR from fd_reopen. */
1293 static int
1294 ifd_reopen (int desired_fd, char const *file, int flag, mode_t mode)
1296 int ret;
1300 process_signals ();
1301 ret = fd_reopen (desired_fd, file, flag, mode);
1303 while (ret < 0 && errno == EINTR);
1305 return ret;
1308 /* Restart on EINTR from fstat. */
1310 static int
1311 ifstat (int fd, struct stat *st)
1313 int ret;
1317 process_signals ();
1318 ret = fstat (fd, st);
1320 while (ret < 0 && errno == EINTR);
1322 return ret;
1325 /* Restart on EINTR from fsync. */
1327 static int
1328 ifsync (int fd)
1330 int ret;
1334 process_signals ();
1335 ret = fsync (fd);
1337 while (ret < 0 && errno == EINTR);
1339 return ret;
1342 /* Restart on EINTR from ftruncate. */
1344 static int
1345 iftruncate (int fd, off_t length)
1347 int ret;
1351 process_signals ();
1352 ret = ftruncate (fd, length);
1354 while (ret < 0 && errno == EINTR);
1356 return ret;
1359 /* Return true if STR is of the form "PATTERN" or "PATTERNDELIM...". */
1361 ATTRIBUTE_PURE
1362 static bool
1363 operand_matches (char const *str, char const *pattern, char delim)
1365 while (*pattern)
1366 if (*str++ != *pattern++)
1367 return false;
1368 return !*str || *str == delim;
1371 /* Interpret one "conv=..." or similar operand STR according to the
1372 symbols in TABLE, returning the flags specified. If the operand
1373 cannot be parsed, use ERROR_MSGID to generate a diagnostic. */
1375 static int
1376 parse_symbols (char const *str, struct symbol_value const *table,
1377 bool exclusive, char const *error_msgid)
1379 int value = 0;
1381 while (true)
1383 char const *strcomma = strchr (str, ',');
1384 struct symbol_value const *entry;
1386 for (entry = table;
1387 ! (operand_matches (str, entry->symbol, ',') && entry->value);
1388 entry++)
1390 if (! entry->symbol[0])
1392 idx_t slen = strcomma ? strcomma - str : strlen (str);
1393 diagnose (0, "%s: %s", _(error_msgid),
1394 quotearg_n_style_mem (0, locale_quoting_style,
1395 str, slen));
1396 usage (EXIT_FAILURE);
1400 if (exclusive)
1401 value = entry->value;
1402 else
1403 value |= entry->value;
1404 if (!strcomma)
1405 break;
1406 str = strcomma + 1;
1409 return value;
1412 /* Return the value of STR, interpreted as a non-negative decimal integer,
1413 optionally multiplied by various values.
1414 Set *INVALID to an appropriate error value and return INTMAX_MAX if
1415 it is an overflow, an indeterminate value if some other error occurred. */
1417 static intmax_t
1418 parse_integer (char const *str, strtol_error *invalid)
1420 /* Call xstrtoumax, not xstrtoimax, since we don't want to
1421 allow strings like " -0". Initialize N to an interminate value;
1422 calling code should not rely on this function returning 0
1423 when *INVALID represents a non-overflow error. */
1424 int indeterminate = 0;
1425 uintmax_t n = indeterminate;
1426 char *suffix;
1427 static char const suffixes[] = "bcEGkKMPQRTwYZ0";
1428 strtol_error e = xstrtoumax (str, &suffix, 10, &n, suffixes);
1429 intmax_t result;
1431 if ((e & ~LONGINT_OVERFLOW) == LONGINT_INVALID_SUFFIX_CHAR
1432 && *suffix == 'B' && str < suffix && suffix[-1] != 'B')
1434 suffix++;
1435 if (!*suffix)
1436 e &= ~LONGINT_INVALID_SUFFIX_CHAR;
1439 if ((e & ~LONGINT_OVERFLOW) == LONGINT_INVALID_SUFFIX_CHAR
1440 && *suffix == 'x')
1442 strtol_error f = LONGINT_OK;
1443 intmax_t o = parse_integer (suffix + 1, &f);
1444 if ((f & ~LONGINT_OVERFLOW) != LONGINT_OK)
1446 e = f;
1447 result = indeterminate;
1449 else if (ckd_mul (&result, n, o)
1450 || (result != 0 && ((e | f) & LONGINT_OVERFLOW)))
1452 e = LONGINT_OVERFLOW;
1453 result = INTMAX_MAX;
1455 else
1457 if (result == 0 && STRPREFIX (str, "0x"))
1458 diagnose (0, _("warning: %s is a zero multiplier; "
1459 "use %s if that is intended"),
1460 quote_n (0, "0x"), quote_n (1, "00x"));
1461 e = LONGINT_OK;
1464 else if (n <= INTMAX_MAX)
1465 result = n;
1466 else
1468 e = LONGINT_OVERFLOW;
1469 result = INTMAX_MAX;
1472 *invalid = e;
1473 return result;
1476 /* OPERAND is of the form "X=...". Return true if X is NAME. */
1478 ATTRIBUTE_PURE
1479 static bool
1480 operand_is (char const *operand, char const *name)
1482 return operand_matches (operand, name, '=');
1485 static void
1486 scanargs (int argc, char *const *argv)
1488 idx_t blocksize = 0;
1489 intmax_t count = INTMAX_MAX;
1490 intmax_t skip = 0;
1491 intmax_t seek = 0;
1492 bool count_B = false, skip_B = false, seek_B = false;
1494 for (int i = optind; i < argc; i++)
1496 char const *name = argv[i];
1497 char const *val = strchr (name, '=');
1499 if (val == nullptr)
1501 diagnose (0, _("unrecognized operand %s"), quoteaf (name));
1502 usage (EXIT_FAILURE);
1504 val++;
1506 if (operand_is (name, "if"))
1507 input_file = val;
1508 else if (operand_is (name, "of"))
1509 output_file = val;
1510 else if (operand_is (name, "conv"))
1511 conversions_mask |= parse_symbols (val, conversions, false,
1512 N_("invalid conversion"));
1513 else if (operand_is (name, "iflag"))
1514 input_flags |= parse_symbols (val, flags, false,
1515 N_("invalid input flag"));
1516 else if (operand_is (name, "oflag"))
1517 output_flags |= parse_symbols (val, flags, false,
1518 N_("invalid output flag"));
1519 else if (operand_is (name, "status"))
1520 status_level = parse_symbols (val, statuses, true,
1521 N_("invalid status level"));
1522 else
1524 strtol_error invalid = LONGINT_OK;
1525 intmax_t n = parse_integer (val, &invalid);
1526 bool has_B = !!strchr (val, 'B');
1527 intmax_t n_min = 0;
1528 intmax_t n_max = INTMAX_MAX;
1529 idx_t *converted_idx = nullptr;
1531 /* Maximum blocksize. Keep it smaller than IDX_MAX, so that
1532 it fits into blocksize vars even if 1 is added for conv=swab.
1533 Do not exceed SSIZE_MAX, for the benefit of system calls
1534 like "read". And do not exceed OFF_T_MAX, for the
1535 benefit of the large-offset seek code. */
1536 idx_t max_blocksize = MIN (IDX_MAX - 1, MIN (SSIZE_MAX, OFF_T_MAX));
1538 if (operand_is (name, "ibs"))
1540 n_min = 1;
1541 n_max = max_blocksize;
1542 converted_idx = &input_blocksize;
1544 else if (operand_is (name, "obs"))
1546 n_min = 1;
1547 n_max = max_blocksize;
1548 converted_idx = &output_blocksize;
1550 else if (operand_is (name, "bs"))
1552 n_min = 1;
1553 n_max = max_blocksize;
1554 converted_idx = &blocksize;
1556 else if (operand_is (name, "cbs"))
1558 n_min = 1;
1559 n_max = MIN (SIZE_MAX, IDX_MAX);
1560 converted_idx = &conversion_blocksize;
1562 else if (operand_is (name, "skip") || operand_is (name, "iseek"))
1564 skip = n;
1565 skip_B = has_B;
1567 else if (operand_is (name + (*name == 'o'), "seek"))
1569 seek = n;
1570 seek_B = has_B;
1572 else if (operand_is (name, "count"))
1574 count = n;
1575 count_B = has_B;
1577 else
1579 diagnose (0, _("unrecognized operand %s"), quoteaf (name));
1580 usage (EXIT_FAILURE);
1583 if (n < n_min)
1584 invalid = LONGINT_INVALID;
1585 else if (n_max < n)
1586 invalid = LONGINT_OVERFLOW;
1588 if (invalid != LONGINT_OK)
1589 error (EXIT_FAILURE, invalid == LONGINT_OVERFLOW ? EOVERFLOW : 0,
1590 "%s: %s", _("invalid number"), quoteaf (val));
1591 else if (converted_idx)
1592 *converted_idx = n;
1596 if (blocksize)
1597 input_blocksize = output_blocksize = blocksize;
1598 else
1600 /* POSIX says dd aggregates partial reads into
1601 output_blocksize if bs= is not specified. */
1602 conversions_mask |= C_TWOBUFS;
1605 if (input_blocksize == 0)
1606 input_blocksize = DEFAULT_BLOCKSIZE;
1607 if (output_blocksize == 0)
1608 output_blocksize = DEFAULT_BLOCKSIZE;
1609 if (conversion_blocksize == 0)
1610 conversions_mask &= ~(C_BLOCK | C_UNBLOCK);
1612 if (input_flags & (O_DSYNC | O_SYNC))
1613 input_flags |= O_RSYNC;
1615 if (output_flags & O_FULLBLOCK)
1617 diagnose (0, "%s: %s", _("invalid output flag"), quote ("fullblock"));
1618 usage (EXIT_FAILURE);
1621 if (skip_B)
1622 input_flags |= O_SKIP_BYTES;
1623 if (input_flags & O_SKIP_BYTES && skip != 0)
1625 skip_records = skip / input_blocksize;
1626 skip_bytes = skip % input_blocksize;
1628 else if (skip != 0)
1629 skip_records = skip;
1631 if (count_B)
1632 input_flags |= O_COUNT_BYTES;
1633 if (input_flags & O_COUNT_BYTES && count != INTMAX_MAX)
1635 max_records = count / input_blocksize;
1636 max_bytes = count % input_blocksize;
1638 else if (count != INTMAX_MAX)
1639 max_records = count;
1641 if (seek_B)
1642 output_flags |= O_SEEK_BYTES;
1643 if (output_flags & O_SEEK_BYTES && seek != 0)
1645 seek_records = seek / output_blocksize;
1646 seek_bytes = seek % output_blocksize;
1648 else if (seek != 0)
1649 seek_records = seek;
1651 /* Warn about partial reads if bs=SIZE is given and iflag=fullblock
1652 is not, and if counting or skipping bytes or using direct I/O.
1653 This helps to avoid confusion with miscounts, and to avoid issues
1654 with direct I/O on GNU/Linux. */
1655 warn_partial_read =
1656 (! (conversions_mask & C_TWOBUFS) && ! (input_flags & O_FULLBLOCK)
1657 && (skip_records
1658 || (0 < max_records && max_records < INTMAX_MAX)
1659 || (input_flags | output_flags) & O_DIRECT));
1661 iread_fnc = ((input_flags & O_FULLBLOCK)
1662 ? iread_fullblock
1663 : iread);
1664 input_flags &= ~O_FULLBLOCK;
1666 if (multiple_bits_set (conversions_mask & (C_ASCII | C_EBCDIC | C_IBM)))
1667 error (EXIT_FAILURE, 0, _("cannot combine any two of {ascii,ebcdic,ibm}"));
1668 if (multiple_bits_set (conversions_mask & (C_BLOCK | C_UNBLOCK)))
1669 error (EXIT_FAILURE, 0, _("cannot combine block and unblock"));
1670 if (multiple_bits_set (conversions_mask & (C_LCASE | C_UCASE)))
1671 error (EXIT_FAILURE, 0, _("cannot combine lcase and ucase"));
1672 if (multiple_bits_set (conversions_mask & (C_EXCL | C_NOCREAT)))
1673 error (EXIT_FAILURE, 0, _("cannot combine excl and nocreat"));
1674 if (multiple_bits_set (input_flags & (O_DIRECT | O_NOCACHE))
1675 || multiple_bits_set (output_flags & (O_DIRECT | O_NOCACHE)))
1676 error (EXIT_FAILURE, 0, _("cannot combine direct and nocache"));
1678 if (input_flags & O_NOCACHE)
1680 i_nocache = true;
1681 i_nocache_eof = (max_records == 0 && max_bytes == 0);
1682 input_flags &= ~O_NOCACHE;
1684 if (output_flags & O_NOCACHE)
1686 o_nocache = true;
1687 o_nocache_eof = (max_records == 0 && max_bytes == 0);
1688 output_flags &= ~O_NOCACHE;
1692 /* Fix up translation table. */
1694 static void
1695 apply_translations (void)
1697 int i;
1699 if (conversions_mask & C_ASCII)
1700 translate_charset (ebcdic_to_ascii);
1702 if (conversions_mask & C_UCASE)
1704 for (i = 0; i < 256; i++)
1705 trans_table[i] = toupper (trans_table[i]);
1706 translation_needed = true;
1708 else if (conversions_mask & C_LCASE)
1710 for (i = 0; i < 256; i++)
1711 trans_table[i] = tolower (trans_table[i]);
1712 translation_needed = true;
1715 if (conversions_mask & C_EBCDIC)
1717 translate_charset (ascii_to_ebcdic);
1718 newline_character = ascii_to_ebcdic['\n'];
1719 space_character = ascii_to_ebcdic[' '];
1721 else if (conversions_mask & C_IBM)
1723 translate_charset (ascii_to_ibm);
1724 newline_character = ascii_to_ibm['\n'];
1725 space_character = ascii_to_ibm[' '];
1729 /* Apply the character-set translations specified by the user
1730 to the NREAD bytes in BUF. */
1732 static void
1733 translate_buffer (char *buf, idx_t nread)
1735 idx_t i;
1736 char *cp;
1737 for (i = nread, cp = buf; i; i--, cp++)
1738 *cp = trans_table[to_uchar (*cp)];
1741 /* Swap *NREAD bytes in BUF, which should have room for an extra byte
1742 after the end because the swapping is not in-place. If *SAVED_BYTE
1743 is nonnegative, also swap that initial byte from the previous call.
1744 Save the last byte into into *SAVED_BYTE if needed to make the
1745 resulting *NREAD even, and set *SAVED_BYTE to -1 otherwise.
1746 Return the buffer's adjusted start, either BUF or BUF + 1. */
1748 static char *
1749 swab_buffer (char *buf, idx_t *nread, int *saved_byte)
1751 if (*nread == 0)
1752 return buf;
1754 /* Update *SAVED_BYTE, and set PREV_SAVED to its old value. */
1755 int prev_saved = *saved_byte;
1756 if ((prev_saved < 0) == (*nread & 1))
1758 unsigned char c = buf[--*nread];
1759 *saved_byte = c;
1761 else
1762 *saved_byte = -1;
1764 /* Do the byte-swapping by moving every other byte two
1765 positions toward the end, working from the end of the buffer
1766 toward the beginning. This way we move only half the data. */
1767 for (idx_t i = *nread; 1 < i; i -= 2)
1768 buf[i] = buf[i - 2];
1770 if (prev_saved < 0)
1771 return buf + 1;
1773 buf[1] = prev_saved;
1774 ++*nread;
1775 return buf;
1778 /* Add OFFSET to the input offset, setting the overflow flag if
1779 necessary. */
1781 static void
1782 advance_input_offset (intmax_t offset)
1784 if (0 <= input_offset && ckd_add (&input_offset, input_offset, offset))
1785 input_offset = -1;
1788 /* Throw away RECORDS blocks of BLOCKSIZE bytes plus BYTES bytes on
1789 file descriptor FDESC, which is open with read permission for FILE.
1790 Store up to BLOCKSIZE bytes of the data at a time in IBUF or OBUF, if
1791 necessary. RECORDS or BYTES must be nonzero. If FDESC is
1792 STDIN_FILENO, advance the input offset. Return the number of
1793 records remaining, i.e., that were not skipped because EOF was
1794 reached. If FDESC is STDOUT_FILENO, on return, BYTES is the
1795 remaining bytes in addition to the remaining records. */
1797 static intmax_t
1798 skip (int fdesc, char const *file, intmax_t records, idx_t blocksize,
1799 idx_t *bytes)
1801 /* Try lseek and if an error indicates it was an inappropriate operation --
1802 or if the file offset is not representable as an off_t --
1803 fall back on using read. */
1805 errno = 0;
1806 off_t offset;
1807 if (! ckd_mul (&offset, records, blocksize)
1808 && ! ckd_add (&offset, offset, *bytes)
1809 && 0 <= lseek (fdesc, offset, SEEK_CUR))
1811 if (fdesc == STDIN_FILENO)
1813 struct stat st;
1814 if (ifstat (STDIN_FILENO, &st) != 0)
1815 error (EXIT_FAILURE, errno, _("cannot fstat %s"), quoteaf (file));
1816 if (usable_st_size (&st) && 0 <= input_offset
1817 && st.st_size - input_offset < offset)
1819 /* When skipping past EOF, return the number of _full_ blocks
1820 * that are not skipped, and set offset to EOF, so the caller
1821 * can determine the requested skip was not satisfied. */
1822 records = ( offset - st.st_size ) / blocksize;
1823 offset = st.st_size - input_offset;
1825 else
1826 records = 0;
1827 advance_input_offset (offset);
1829 else
1831 records = 0;
1832 *bytes = 0;
1834 return records;
1836 else
1838 int lseek_errno = errno;
1840 /* The seek request may have failed above if it was too big
1841 (> device size, > max file size, etc.)
1842 Or it may not have been done at all (> OFF_T_MAX).
1843 Therefore try to seek to the end of the file,
1844 to avoid redundant reading. */
1845 if (lseek (fdesc, 0, SEEK_END) >= 0)
1847 /* File is seekable, and we're at the end of it, and
1848 size <= OFF_T_MAX. So there's no point using read to advance. */
1850 if (!lseek_errno)
1852 /* The original seek was not attempted as offset > OFF_T_MAX.
1853 We should error for write as can't get to the desired
1854 location, even if OFF_T_MAX < max file size.
1855 For read we're not going to read any data anyway,
1856 so we should error for consistency.
1857 It would be nice to not error for /dev/{zero,null}
1858 for any offset, but that's not a significant issue. */
1859 lseek_errno = EOVERFLOW;
1862 diagnose (lseek_errno,
1863 gettext (fdesc == STDIN_FILENO
1864 ? N_("%s: cannot skip")
1865 : N_("%s: cannot seek")),
1866 quotef (file));
1867 /* If the file has a specific size and we've asked
1868 to skip/seek beyond the max allowable, then quit. */
1869 quit (EXIT_FAILURE);
1871 /* else file_size && offset > OFF_T_MAX or file ! seekable */
1873 char *buf;
1874 if (fdesc == STDIN_FILENO)
1876 alloc_ibuf ();
1877 buf = ibuf;
1879 else
1881 alloc_obuf ();
1882 buf = obuf;
1887 ssize_t nread = iread_fnc (fdesc, buf, records ? blocksize : *bytes);
1888 if (nread < 0)
1890 if (fdesc == STDIN_FILENO)
1892 diagnose (errno, _("error reading %s"), quoteaf (file));
1893 if (conversions_mask & C_NOERROR)
1894 print_stats ();
1896 else
1897 diagnose (lseek_errno, _("%s: cannot seek"), quotef (file));
1898 quit (EXIT_FAILURE);
1900 else if (nread == 0)
1901 break;
1902 else if (fdesc == STDIN_FILENO)
1903 advance_input_offset (nread);
1905 if (records != 0)
1906 records--;
1907 else
1908 *bytes = 0;
1910 while (records || *bytes);
1912 return records;
1916 /* Advance the input by NBYTES if possible, after a read error.
1917 The input file offset may or may not have advanced after the failed
1918 read; adjust it to point just after the bad record regardless.
1919 Return true if successful, or if the input is already known to not
1920 be seekable. */
1922 static bool
1923 advance_input_after_read_error (idx_t nbytes)
1925 if (! input_seekable)
1927 if (input_seek_errno == ESPIPE)
1928 return true;
1929 errno = input_seek_errno;
1931 else
1933 off_t offset;
1934 advance_input_offset (nbytes);
1935 if (input_offset < 0)
1937 diagnose (0, _("offset overflow while reading file %s"),
1938 quoteaf (input_file));
1939 return false;
1941 offset = lseek (STDIN_FILENO, 0, SEEK_CUR);
1942 if (0 <= offset)
1944 off_t diff;
1945 if (offset == input_offset)
1946 return true;
1947 diff = input_offset - offset;
1948 if (! (0 <= diff && diff <= nbytes) && status_level != STATUS_NONE)
1949 diagnose (0, _("warning: invalid file offset after failed read"));
1950 if (0 <= lseek (STDIN_FILENO, diff, SEEK_CUR))
1951 return true;
1952 if (errno == 0)
1953 diagnose (0, _("cannot work around kernel bug after all"));
1957 diagnose (errno, _("%s: cannot seek"), quotef (input_file));
1958 return false;
1961 /* Copy NREAD bytes of BUF, with no conversions. */
1963 static void
1964 copy_simple (char const *buf, idx_t nread)
1966 char const *start = buf; /* First uncopied char in BUF. */
1970 idx_t nfree = MIN (nread, output_blocksize - oc);
1972 memcpy (obuf + oc, start, nfree);
1974 nread -= nfree; /* Update the number of bytes left to copy. */
1975 start += nfree;
1976 oc += nfree;
1977 if (oc >= output_blocksize)
1978 write_output ();
1980 while (nread != 0);
1983 /* Copy NREAD bytes of BUF, doing conv=block
1984 (pad newline-terminated records to 'conversion_blocksize',
1985 replacing the newline with trailing spaces). */
1987 static void
1988 copy_with_block (char const *buf, idx_t nread)
1990 for (idx_t i = nread; i; i--, buf++)
1992 if (*buf == newline_character)
1994 if (col < conversion_blocksize)
1996 idx_t j;
1997 for (j = col; j < conversion_blocksize; j++)
1998 output_char (space_character);
2000 col = 0;
2002 else
2004 if (col == conversion_blocksize)
2005 r_truncate++;
2006 else if (col < conversion_blocksize)
2007 output_char (*buf);
2008 col++;
2013 /* Copy NREAD bytes of BUF, doing conv=unblock
2014 (replace trailing spaces in 'conversion_blocksize'-sized records
2015 with a newline). */
2017 static void
2018 copy_with_unblock (char const *buf, idx_t nread)
2020 static idx_t pending_spaces = 0;
2022 for (idx_t i = 0; i < nread; i++)
2024 char c = buf[i];
2026 if (col++ >= conversion_blocksize)
2028 col = pending_spaces = 0; /* Wipe out any pending spaces. */
2029 i--; /* Push the char back; get it later. */
2030 output_char (newline_character);
2032 else if (c == space_character)
2033 pending_spaces++;
2034 else
2036 /* 'c' is the character after a run of spaces that were not
2037 at the end of the conversion buffer. Output them. */
2038 while (pending_spaces)
2040 output_char (space_character);
2041 --pending_spaces;
2043 output_char (c);
2048 /* Set the file descriptor flags for FD that correspond to the nonzero bits
2049 in ADD_FLAGS. The file's name is NAME. */
2051 static void
2052 set_fd_flags (int fd, int add_flags, char const *name)
2054 /* Ignore file creation flags that are no-ops on file descriptors. */
2055 add_flags &= ~ (O_NOCTTY | O_NOFOLLOW);
2057 if (add_flags)
2059 int old_flags = fcntl (fd, F_GETFL);
2060 int new_flags = old_flags | add_flags;
2061 bool ok = true;
2062 if (old_flags < 0)
2063 ok = false;
2064 else if (old_flags != new_flags)
2066 if (new_flags & (O_DIRECTORY | O_NOLINKS))
2068 /* NEW_FLAGS contains at least one file creation flag that
2069 requires some checking of the open file descriptor. */
2070 struct stat st;
2071 if (ifstat (fd, &st) != 0)
2072 ok = false;
2073 else if ((new_flags & O_DIRECTORY) && ! S_ISDIR (st.st_mode))
2075 errno = ENOTDIR;
2076 ok = false;
2078 else if ((new_flags & O_NOLINKS) && 1 < st.st_nlink)
2080 errno = EMLINK;
2081 ok = false;
2083 new_flags &= ~ (O_DIRECTORY | O_NOLINKS);
2086 if (ok && old_flags != new_flags
2087 && fcntl (fd, F_SETFL, new_flags) == -1)
2088 ok = false;
2091 if (!ok)
2092 error (EXIT_FAILURE, errno, _("setting flags for %s"), quoteaf (name));
2096 /* The main loop. */
2098 static int
2099 dd_copy (void)
2101 char *bufstart; /* Input buffer. */
2102 ssize_t nread; /* Bytes read in the current block. */
2104 /* If nonzero, then the previously read block was partial and
2105 PARTREAD was its size. */
2106 idx_t partread = 0;
2108 int exit_status = EXIT_SUCCESS;
2109 idx_t n_bytes_read;
2111 if (skip_records != 0 || skip_bytes != 0)
2113 intmax_t us_bytes;
2114 bool us_bytes_overflow =
2115 (ckd_mul (&us_bytes, skip_records, input_blocksize)
2116 || ckd_add (&us_bytes, skip_bytes, us_bytes));
2117 off_t input_offset0 = input_offset;
2118 intmax_t us_blocks = skip (STDIN_FILENO, input_file,
2119 skip_records, input_blocksize, &skip_bytes);
2121 /* POSIX doesn't say what to do when dd detects it has been
2122 asked to skip past EOF, so I assume it's non-fatal.
2123 There are 3 reasons why there might be unskipped blocks/bytes:
2124 1. file is too small
2125 2. pipe has not enough data
2126 3. partial reads */
2127 if ((us_blocks
2128 || (0 <= input_offset
2129 && (us_bytes_overflow
2130 || us_bytes != input_offset - input_offset0)))
2131 && status_level != STATUS_NONE)
2133 diagnose (0, _("%s: cannot skip to specified offset"),
2134 quotef (input_file));
2138 if (seek_records != 0 || seek_bytes != 0)
2140 idx_t bytes = seek_bytes;
2141 intmax_t write_records = skip (STDOUT_FILENO, output_file,
2142 seek_records, output_blocksize, &bytes);
2144 if (write_records != 0 || bytes != 0)
2146 memset (obuf, 0, write_records ? output_blocksize : bytes);
2150 idx_t size = write_records ? output_blocksize : bytes;
2151 if (iwrite (STDOUT_FILENO, obuf, size) != size)
2153 diagnose (errno, _("writing to %s"), quoteaf (output_file));
2154 quit (EXIT_FAILURE);
2157 if (write_records != 0)
2158 write_records--;
2159 else
2160 bytes = 0;
2162 while (write_records || bytes);
2166 if (max_records == 0 && max_bytes == 0)
2167 return exit_status;
2169 alloc_ibuf ();
2170 alloc_obuf ();
2171 int saved_byte = -1;
2173 while (true)
2175 if (status_level == STATUS_PROGRESS)
2177 xtime_t progress_time = gethrxtime ();
2178 if (next_time <= progress_time)
2180 print_xfer_stats (progress_time);
2181 next_time += XTIME_PRECISION;
2185 if (r_partial + r_full >= max_records + !!max_bytes)
2186 break;
2188 /* Zero the buffer before reading, so that if we get a read error,
2189 whatever data we are able to read is followed by zeros.
2190 This minimizes data loss. */
2191 if ((conversions_mask & C_SYNC) && (conversions_mask & C_NOERROR))
2192 memset (ibuf,
2193 (conversions_mask & (C_BLOCK | C_UNBLOCK)) ? ' ' : '\0',
2194 input_blocksize);
2196 if (r_partial + r_full >= max_records)
2197 nread = iread_fnc (STDIN_FILENO, ibuf, max_bytes);
2198 else
2199 nread = iread_fnc (STDIN_FILENO, ibuf, input_blocksize);
2201 if (nread > 0)
2203 advance_input_offset (nread);
2204 if (i_nocache)
2205 invalidate_cache (STDIN_FILENO, nread);
2207 else if (nread == 0)
2209 i_nocache_eof |= i_nocache;
2210 o_nocache_eof |= o_nocache && ! (conversions_mask & C_NOTRUNC);
2211 break; /* EOF. */
2213 else
2215 if (!(conversions_mask & C_NOERROR) || status_level != STATUS_NONE)
2216 diagnose (errno, _("error reading %s"), quoteaf (input_file));
2218 if (conversions_mask & C_NOERROR)
2220 print_stats ();
2221 idx_t bad_portion = input_blocksize - partread;
2223 /* We already know this data is not cached,
2224 but call this so that correct offsets are maintained. */
2225 invalidate_cache (STDIN_FILENO, bad_portion);
2227 /* Seek past the bad block if possible. */
2228 if (!advance_input_after_read_error (bad_portion))
2230 exit_status = EXIT_FAILURE;
2232 /* Suppress duplicate diagnostics. */
2233 input_seekable = false;
2234 input_seek_errno = ESPIPE;
2236 if ((conversions_mask & C_SYNC) && !partread)
2237 /* Replace the missing input with null bytes and
2238 proceed normally. */
2239 nread = 0;
2240 else
2241 continue;
2243 else
2245 /* Write any partial block. */
2246 exit_status = EXIT_FAILURE;
2247 break;
2251 n_bytes_read = nread;
2253 if (n_bytes_read < input_blocksize)
2255 r_partial++;
2256 partread = n_bytes_read;
2257 if (conversions_mask & C_SYNC)
2259 if (!(conversions_mask & C_NOERROR))
2260 /* If C_NOERROR, we zeroed the block before reading. */
2261 memset (ibuf + n_bytes_read,
2262 (conversions_mask & (C_BLOCK | C_UNBLOCK)) ? ' ' : '\0',
2263 input_blocksize - n_bytes_read);
2264 n_bytes_read = input_blocksize;
2267 else
2269 r_full++;
2270 partread = 0;
2273 if (ibuf == obuf) /* If not C_TWOBUFS. */
2275 idx_t nwritten = iwrite (STDOUT_FILENO, obuf, n_bytes_read);
2276 w_bytes += nwritten;
2277 if (nwritten != n_bytes_read)
2279 diagnose (errno, _("error writing %s"), quoteaf (output_file));
2280 return EXIT_FAILURE;
2282 else if (n_bytes_read == input_blocksize)
2283 w_full++;
2284 else
2285 w_partial++;
2286 continue;
2289 /* Do any translations on the whole buffer at once. */
2291 if (translation_needed)
2292 translate_buffer (ibuf, n_bytes_read);
2294 if (conversions_mask & C_SWAB)
2295 bufstart = swab_buffer (ibuf, &n_bytes_read, &saved_byte);
2296 else
2297 bufstart = ibuf;
2299 if (conversions_mask & C_BLOCK)
2300 copy_with_block (bufstart, n_bytes_read);
2301 else if (conversions_mask & C_UNBLOCK)
2302 copy_with_unblock (bufstart, n_bytes_read);
2303 else
2304 copy_simple (bufstart, n_bytes_read);
2307 /* If we have a char left as a result of conv=swab, output it. */
2308 if (0 <= saved_byte)
2310 char saved_char = saved_byte;
2311 if (conversions_mask & C_BLOCK)
2312 copy_with_block (&saved_char, 1);
2313 else if (conversions_mask & C_UNBLOCK)
2314 copy_with_unblock (&saved_char, 1);
2315 else
2316 output_char (saved_char);
2319 if ((conversions_mask & C_BLOCK) && col > 0)
2321 /* If the final input line didn't end with a '\n', pad
2322 the output block to 'conversion_blocksize' chars. */
2323 for (idx_t i = col; i < conversion_blocksize; i++)
2324 output_char (space_character);
2327 if (col && (conversions_mask & C_UNBLOCK))
2329 /* If there was any output, add a final '\n'. */
2330 output_char (newline_character);
2333 /* Write out the last block. */
2334 if (oc != 0)
2336 idx_t nwritten = iwrite (STDOUT_FILENO, obuf, oc);
2337 w_bytes += nwritten;
2338 if (nwritten != 0)
2339 w_partial++;
2340 if (nwritten != oc)
2342 diagnose (errno, _("error writing %s"), quoteaf (output_file));
2343 return EXIT_FAILURE;
2347 /* If the last write was converted to a seek, then for a regular file
2348 or shared memory object, ftruncate to extend the size. */
2349 if (final_op_was_seek)
2351 struct stat stdout_stat;
2352 if (ifstat (STDOUT_FILENO, &stdout_stat) != 0)
2354 diagnose (errno, _("cannot fstat %s"), quoteaf (output_file));
2355 return EXIT_FAILURE;
2357 if (S_ISREG (stdout_stat.st_mode) || S_TYPEISSHM (&stdout_stat))
2359 off_t output_offset = lseek (STDOUT_FILENO, 0, SEEK_CUR);
2360 if (0 <= output_offset && stdout_stat.st_size < output_offset)
2362 if (iftruncate (STDOUT_FILENO, output_offset) != 0)
2364 diagnose (errno, _("failed to truncate to %" PRIdMAX " bytes"
2365 " in output file %s"),
2366 (intmax_t) output_offset, quoteaf (output_file));
2367 return EXIT_FAILURE;
2373 /* fdatasync/fsync can take a long time, so issue a final progress
2374 indication now if progress has been made since the previous indication. */
2375 if (conversions_mask & (C_FDATASYNC | C_FSYNC)
2376 && status_level == STATUS_PROGRESS
2377 && 0 <= reported_w_bytes && reported_w_bytes < w_bytes)
2378 print_xfer_stats (0);
2380 return exit_status;
2383 /* Synchronize output according to conversions_mask.
2384 Do this even if w_bytes is zero, as fsync and fdatasync
2385 flush out write requests from other processes too.
2386 Clear bits in conversions_mask so that synchronization is done only once.
2387 Return zero if successful, an exit status otherwise. */
2389 static int
2390 synchronize_output (void)
2392 int exit_status = 0;
2393 int mask = conversions_mask;
2394 conversions_mask &= ~ (C_FDATASYNC | C_FSYNC);
2396 if ((mask & C_FDATASYNC) && ifdatasync (STDOUT_FILENO) != 0)
2398 if (errno != ENOSYS && errno != EINVAL)
2400 diagnose (errno, _("fdatasync failed for %s"), quoteaf (output_file));
2401 exit_status = EXIT_FAILURE;
2403 mask |= C_FSYNC;
2406 if ((mask & C_FSYNC) && ifsync (STDOUT_FILENO) != 0)
2408 diagnose (errno, _("fsync failed for %s"), quoteaf (output_file));
2409 return EXIT_FAILURE;
2412 return exit_status;
2416 main (int argc, char **argv)
2418 int i;
2419 int exit_status;
2420 off_t offset;
2422 install_signal_handlers ();
2424 initialize_main (&argc, &argv);
2425 set_program_name (argv[0]);
2426 setlocale (LC_ALL, "");
2427 bindtextdomain (PACKAGE, LOCALEDIR);
2428 textdomain (PACKAGE);
2430 /* Arrange to close stdout if parse_long_options exits. */
2431 atexit (maybe_close_stdout);
2433 page_size = getpagesize ();
2435 parse_gnu_standard_options_only (argc, argv, PROGRAM_NAME, PACKAGE, Version,
2436 true, usage, AUTHORS,
2437 (char const *) nullptr);
2438 close_stdout_required = false;
2440 /* Initialize translation table to identity translation. */
2441 for (i = 0; i < 256; i++)
2442 trans_table[i] = i;
2444 /* Decode arguments. */
2445 scanargs (argc, argv);
2447 apply_translations ();
2449 if (input_file == nullptr)
2451 input_file = _("standard input");
2452 set_fd_flags (STDIN_FILENO, input_flags, input_file);
2454 else
2456 if (ifd_reopen (STDIN_FILENO, input_file, O_RDONLY | input_flags, 0) < 0)
2457 error (EXIT_FAILURE, errno, _("failed to open %s"),
2458 quoteaf (input_file));
2461 offset = lseek (STDIN_FILENO, 0, SEEK_CUR);
2462 input_seekable = (0 <= offset);
2463 input_offset = MAX (0, offset);
2464 input_seek_errno = errno;
2466 if (output_file == nullptr)
2468 output_file = _("standard output");
2469 set_fd_flags (STDOUT_FILENO, output_flags, output_file);
2471 else
2473 mode_t perms = MODE_RW_UGO;
2474 int opts
2475 = (output_flags
2476 | (conversions_mask & C_NOCREAT ? 0 : O_CREAT)
2477 | (conversions_mask & C_EXCL ? O_EXCL : 0)
2478 | (seek_records || (conversions_mask & C_NOTRUNC) ? 0 : O_TRUNC));
2480 off_t size;
2481 if ((ckd_mul (&size, seek_records, output_blocksize)
2482 || ckd_add (&size, seek_bytes, size))
2483 && !(conversions_mask & C_NOTRUNC))
2484 error (EXIT_FAILURE, 0,
2485 _("offset too large: "
2486 "cannot truncate to a length of seek=%"PRIdMAX""
2487 " (%td-byte) blocks"),
2488 seek_records, output_blocksize);
2490 /* Open the output file with *read* access only if we might
2491 need to read to satisfy a 'seek=' request. If we can't read
2492 the file, go ahead with write-only access; it might work. */
2493 if ((! seek_records
2494 || ifd_reopen (STDOUT_FILENO, output_file, O_RDWR | opts, perms) < 0)
2495 && (ifd_reopen (STDOUT_FILENO, output_file, O_WRONLY | opts, perms)
2496 < 0))
2497 error (EXIT_FAILURE, errno, _("failed to open %s"),
2498 quoteaf (output_file));
2500 if (seek_records != 0 && !(conversions_mask & C_NOTRUNC))
2502 if (iftruncate (STDOUT_FILENO, size) != 0)
2504 /* Complain only when ftruncate fails on a regular file, a
2505 directory, or a shared memory object, as POSIX 1003.1-2004
2506 specifies ftruncate's behavior only for these file types.
2507 For example, do not complain when Linux kernel 2.4 ftruncate
2508 fails on /dev/fd0. */
2509 int ftruncate_errno = errno;
2510 struct stat stdout_stat;
2511 if (ifstat (STDOUT_FILENO, &stdout_stat) != 0)
2513 diagnose (errno, _("cannot fstat %s"), quoteaf (output_file));
2514 exit_status = EXIT_FAILURE;
2516 else if (S_ISREG (stdout_stat.st_mode)
2517 || S_ISDIR (stdout_stat.st_mode)
2518 || S_TYPEISSHM (&stdout_stat))
2520 intmax_t isize = size;
2521 diagnose (ftruncate_errno,
2522 _("failed to truncate to %"PRIdMAX" bytes"
2523 " in output file %s"),
2524 isize, quoteaf (output_file));
2525 exit_status = EXIT_FAILURE;
2531 start_time = gethrxtime ();
2532 next_time = start_time + XTIME_PRECISION;
2534 exit_status = dd_copy ();
2536 int sync_status = synchronize_output ();
2537 if (sync_status)
2538 exit_status = sync_status;
2540 if (max_records == 0 && max_bytes == 0)
2542 /* Special case to invalidate cache to end of file. */
2543 if (i_nocache && !invalidate_cache (STDIN_FILENO, 0))
2545 diagnose (errno, _("failed to discard cache for: %s"),
2546 quotef (input_file));
2547 exit_status = EXIT_FAILURE;
2549 if (o_nocache && !invalidate_cache (STDOUT_FILENO, 0))
2551 diagnose (errno, _("failed to discard cache for: %s"),
2552 quotef (output_file));
2553 exit_status = EXIT_FAILURE;
2556 else
2558 /* Invalidate any pending region or to EOF if appropriate. */
2559 if (i_nocache || i_nocache_eof)
2560 invalidate_cache (STDIN_FILENO, 0);
2561 if (o_nocache || o_nocache_eof)
2562 invalidate_cache (STDOUT_FILENO, 0);
2565 finish_up ();
2566 main_exit (exit_status);