man5/core.5

   1 .\" Copyright (c) 2006, 2008 by Michael Kerrisk <mtk.manpages@gmail.com>
   2 .\"
   3 .\" Permission is granted to make and distribute verbatim copies of this
   4 .\" manual provided the copyright notice and this permission notice are
   5 .\" preserved on all copies.
   6 .\"
   7 .\" Permission is granted to copy and distribute modified versions of this
   8 .\" manual under the conditions for verbatim copying, provided that the
   9 .\" entire resulting derived work is distributed under the terms of a
  10 .\" permission notice identical to this one.
  11 .\"
  12 .\" Since the Linux kernel and libraries are constantly changing, this
  13 .\" manual page may be incorrect or out-of-date.  The author(s) assume no
  14 .\" responsibility for errors or omissions, or for damages resulting from
  15 .\" the use of the information contained herein.  The author(s) may not
  16 .\" have taken the same level of care in the production of this manual,
  17 .\" which is licensed free of charge, as they might when working
  18 .\" professionally.
  19 .\"
  20 .\" Formatted or processed versions of this manual, if unaccompanied by
  21 .\" the source, must acknowledge the copyright and authors of this work.
  22 .\"
  23 .TH CORE 5 2012-07-16 "Linux" "Linux Programmer's Manual"
  24 .SH NAME
  25 core \- core dump file
  26 .SH DESCRIPTION
  27 The default action of certain signals is to cause a process to terminate
  28 and produce a
  29 .IR "core dump file" ,
  30 a disk file containing an image of the process's memory at
  31 the time of termination.
  32 This image can be used in a debugger (e.g.,
  33 .BR gdb (1))
  34 to inspect the state of the program at the time that it terminated.
  35 A list of the signals which cause a process to dump core can be found in
  36 .BR signal (7).
  37
  38 A process can set its soft
  39 .B RLIMIT_CORE
  40 resource limit to place an upper limit on the size of the core dump file
  41 that will be produced if it receives a "core dump" signal; see
  42 .BR getrlimit (2)
  43 for details.
  44
  45 There are various circumstances in which a core dump file is
  46 not produced:
  47 .IP * 3
  48 The process does not have permission to write the core file.
  49 (By default the core file is called
  50 .IR core ,
  51 and is created in the current working directory.
  52 See below for details on naming.)
  53 Writing the core file will fail if the directory in which
  54 it is to be created is nonwritable,
  55 or if a file with the same name exists and
  56 is not writable
  57 or is not a regular file
  58 (e.g., it is a directory or a symbolic link).
  59 .IP *
  60 A (writable, regular) file with the same name as would be used for the
  61 core dump already exists, but there is more than one hard link to that
  62 file.
  63 .IP *
  64 The file system where the core dump file would be created is full;
  65 or has run out of inodes; or is mounted read-only;
  66 or the user has reached their quota for the file system.
  67 .IP *
  68 The directory in which the core dump file is to be created does
  69 not exist.
  70 .IP *
  71 The
  72 .B RLIMIT_CORE
  73 (core file size) or
  74 .B RLIMIT_FSIZE
  75 (file size) resource limits for the process are set to zero; see
  76 .BR getrlimit (2)
  77 and the documentation of the shell's
  78 .I ulimit
  79 command
  80 .RI ( limit
  81 in
  82 .BR csh (1)).
  83 .IP *
  84 The binary being executed by the process does not have read
  85 permission enabled.
  86 .IP *
  87 The process is executing a set-user-ID (set-group-ID) program
  88 that is owned by a user (group) other than the real user (group)
  89 ID of the process.
  90 (However, see the description of the
  91 .BR prctl (2)
  92 .B PR_SET_DUMPABLE
  93 operation, and the description of the
  94 .I /proc/sys/fs/suid_dumpable
  95 .\" FIXME . Perhaps relocate discussion of /proc/sys/fs/suid_dumpable
  96 .\" and PR_SET_DUMPABLE to this page?
  97 file in
  98 .BR proc (5).)
  99 .PP
 100 In addition,
 101 a core dump may exclude part of the address space of the process if the
 102 .BR madvise (2)
 103 .B MADV_DONTDUMP
 104 flag was employed.
 105 .SS Naming of core dump files
 106 By default, a core dump file is named
 107 .IR core ,
 108 but the
 109 .I /proc/sys/kernel/core_pattern
 110 file (since Linux 2.6 and 2.4.21)
 111 can be set to define a template that is used to name core dump files.
 112 The template can contain % specifiers which are substituted
 113 by the following values when a core file is created:
 114 .PP
 115 .RS 4
 116 .PD 0
 117 .TP 4
 118 %%
 119 a single % character
 120 .TP
 121 %p
 122 PID of dumped process
 123 .TP
 124 %u
 125 (numeric) real UID of dumped process
 126 .TP
 127 %g
 128 (numeric) real GID of dumped process
 129 .TP
 130 %s
 131 number of signal causing dump
 132 .TP
 133 %t
 134 time of dump, expressed as seconds since the
 135 Epoch, 1970-01-01 00:00:00 +0000 (UTC)
 136 .TP
 137 %h
 138 hostname (same as \fInodename\fP returned by \fBuname\fP(2))
 139 .TP
 140 %e
 141 executable filename (without path prefix)
 142 .TP
 143 %E
 144 pathname of executable,
 145 with slashes (\(aq/\(aq) replaced by exclamation marks (\(aq!\(aq).
 146 .TP
 147 %c
 148 core file size soft resource limit of crashing process (since Linux 2.6.24)
 149 .PD
 150 .RE
 151 .PP
 152 A single % at the end of the template is dropped from the
 153 core filename, as is the combination of a % followed by any
 154 character other than those listed above.
 155 All other characters in the template become a literal
 156 part of the core filename.
 157 The template may include \(aq/\(aq characters, which are interpreted
 158 as delimiters for directory names.
 159 The maximum size of the resulting core filename is 128 bytes (64 bytes
 160 in kernels before 2.6.19).
 161 The default value in this file is "core".
 162 For backward compatibility, if
 163 .I /proc/sys/kernel/core_pattern
 164 does not include "%p" and
 165 .I /proc/sys/kernel/core_uses_pid
 166 (see below)
 167 is nonzero, then .PID will be appended to the core filename.
 168
 169 Since version 2.4, Linux has also provided
 170 a more primitive method of controlling
 171 the name of the core dump file.
 172 If the
 173 .I /proc/sys/kernel/core_uses_pid
 174 file contains the value 0, then a core dump file is simply named
 175 .IR core .
 176 If this file contains a nonzero value, then the core dump file includes
 177 the process ID in a name of the form
 178 .IR core.PID .
 179 .SS Piping core dumps to a program
 180 Since kernel 2.6.19, Linux supports an alternate syntax for the
 181 .I /proc/sys/kernel/core_pattern
 182 file.
 183 If the first character of this file is a pipe symbol (\fB|\fP),
 184 then the remainder of the line is interpreted as a program to be
 185 executed.
 186 Instead of being written to a disk file, the core dump is given as
 187 standard input to the program.
 188 Note the following points:
 189 .IP * 3
 190 The program must be specified using an absolute pathname (or a
 191 pathname relative to the root directory, \fI/\fP),
 192 and must immediately follow the '|' character.
 193 .IP *
 194 The process created to run the program runs as user and group
 195 .IR root .
 196 .IP *
 197 Command-line arguments can be supplied to the
 198 program (since kernel 2.6.24),
 199 delimited by white space (up to a total line length of 128 bytes).
 200 .IP *
 201 The command-line arguments can include any of
 202 the % specifiers listed above.
 203 For example, to pass the PID of the process that is being dumped, specify
 204 .I %p
 205 in an argument.
 206 .SS Controlling which mappings are written to the core dump
 207 Since kernel 2.6.23, the Linux-specific
 208 .IR /proc/PID/coredump_filter
 209 file can be used to control which memory segments are written to the
 210 core dump file in the event that a core dump is performed for the
 211 process with the corresponding process ID.
 212
 213 The value in the file is a bit mask of memory mapping types (see
 214 .BR mmap (2)).
 215 If a bit is set in the mask, then memory mappings of the
 216 corresponding type are dumped; otherwise they are not dumped.
 217 The bits in this file have the following meanings:
 218 .PP
 219 .PD 0
 220 .RS 4
 221 .TP
 222 bit 0
 223 Dump anonymous private mappings.
 224 .TP
 225 bit 1
 226 Dump anonymous shared mappings.
 227 .TP
 228 bit 2
 229 Dump file-backed private mappings.
 230 .TP
 231 bit 3
 232 Dump file-backed shared mappings.
 233 .\" file-backed shared mappings of course also update the underlying
 234 .\" mapped file.
 235 .TP
 236 bit 4 (since Linux 2.6.24)
 237 Dump ELF headers.
 238 .TP
 239 bit 5 (since Linux 2.6.28)
 240 Dump private huge pages.
 241 .TP
 242 bit 6 (since Linux 2.6.28)
 243 Dump shared huge pages.
 244 .RE
 245 .PD
 246 .PP
 247 By default, the following bits are set: 0, 1, 4 (if the
 248 .B CONFIG_CORE_DUMP_DEFAULT_ELF_HEADERS
 249 kernel configuration option is enabled), and 5.
 250 The value of this file is displayed in hexadecimal.
 251 (The default value is thus displayed as 33.)
 252
 253 Memory-mapped I/O pages such as frame buffer are never dumped, and
 254 virtual DSO pages are always dumped, regardless of the
 255 .I coredump_filter
 256 value.
 257
 258 A child process created via
 259 .BR fork (2)
 260 inherits its parent's
 261 .I coredump_filter
 262 value;
 263 the
 264 .I coredump_filter
 265 value is preserved across an
 266 .BR execve (2).
 267
 268 It can be useful to set
 269 .I coredump_filter
 270 in the parent shell before running a program, for example:
 271
 272 .in +4n
 273 .nf
 274 .RB "$" " echo 0x7 > /proc/self/coredump_filter"
 275 .RB "$" " ./some_program"
 276 .fi
 277 .in
 278 .PP
 279 This file is only provided if the kernel was built with the
 280 .B CONFIG_ELF_CORE
 281 configuration option.
 282 .SH NOTES
 283 The
 284 .BR gdb (1)
 285 .I gcore
 286 command can be used to obtain a core dump of a running process.
 287
 288 If a multithreaded process (or, more precisely, a process that
 289 shares its memory with another process by being created with the
 290 .B CLONE_VM
 291 flag of
 292 .BR clone (2))
 293 dumps core, then the process ID is always appended to the core filename,
 294 unless the process ID was already included elsewhere in the
 295 filename via a %p specification in
 296 .IR /proc/sys/kernel/core_pattern .
 297 (This is primarily useful when employing the LinuxThreads implementation,
 298 where each thread of a process has a different PID.)
 299 .\" Always including the PID in the name of the core file made
 300 .\" sense for LinuxThreads, where each thread had a unique PID,
 301 .\" but doesn't seem to serve any purpose with NPTL, where all the
 302 .\" threads in a process share the same PID (as POSIX.1 requires).
 303 .\" Probably the behavior is maintained so that applications using
 304 .\" LinuxThreads continue appending the PID (the kernel has no easy
 305 .\" way of telling which threading implementation the userspace
 306 .\" application is using). -- mtk, April 2006
 307 .SH EXAMPLE
 308 The program below can be used to demonstrate the use of the
 309 pipe syntax in the
 310 .I /proc/sys/kernel/core_pattern
 311 file.
 312 The following shell session demonstrates the use of this program
 313 (compiled to create an executable named
 314 .IR core_pattern_pipe_test ):
 315 .PP
 316 .in +4n
 317 .nf
 318 .RB "$" " cc \-o core_pattern_pipe_test core_pattern_pipe_test.c"
 319 .RB "$" " su"
 320 Password:
 321 .RB "#" " echo \(dq|$PWD/core_pattern_pipe_test %p \
 322 UID=%u GID=%g sig=%s\(dq > \\"
 323 .B "    /proc/sys/kernel/core_pattern"
 324 .RB "#" " exit"
 325 .RB "$" " sleep 100"
 326 .BR "^\\" "                     # type control-backslash"
 327 Quit (core dumped)
 328 .RB "$" " cat core.info"
 329 argc=5
 330 argc[0]=</home/mtk/core_pattern_pipe_test>
 331 argc[1]=<20575>
 332 argc[2]=<UID=1000>
 333 argc[3]=<GID=100>
 334 argc[4]=<sig=3>
 335 Total bytes in core dump: 282624
 336 .fi
 337 .in
 338 .SS Program source
 339 \&
 340 .nf
 341 /* core_pattern_pipe_test.c */
 342
 343 #define _GNU_SOURCE
 344 #include <sys/stat.h>
 345 #include <fcntl.h>
 346 #include <limits.h>
 347 #include <stdio.h>
 348 #include <stdlib.h>
 349 #include <unistd.h>
 350
 351 #define BUF_SIZE 1024
 352
 353 int
 354 main(int argc, char *argv[])
 355 {
 356     int tot, j;
 357     ssize_t nread;
 358     char buf[BUF_SIZE];
 359     FILE *fp;
 360     char cwd[PATH_MAX];
 361
 362     /* Change our current working directory to that of the
 363        crashing process */
 364
 365     snprintf(cwd, PATH_MAX, "/proc/%s/cwd", argv[1]);
 366     chdir(cwd);
 367
 368     /* Write output to file "core.info" in that directory */
 369
 370     fp = fopen("core.info", "w+");
 371     if (fp == NULL)
 372         exit(EXIT_FAILURE);
 373
 374     /* Display command\-line arguments given to core_pattern
 375        pipe program */
 376
 377     fprintf(fp, "argc=%d\\n", argc);
 378     for (j = 0; j < argc; j++)
 379         fprintf(fp, "argc[%d]=<%s>\\n", j, argv[j]);
 380
 381     /* Count bytes in standard input (the core dump) */
 382
 383     tot = 0;
 384     while ((nread = read(STDIN_FILENO, buf, BUF_SIZE)) > 0)
 385         tot += nread;
 386     fprintf(fp, "Total bytes in core dump: %d\\n", tot);
 387
 388     exit(EXIT_SUCCESS);
 389 }
 390 .fi
 391 .SH SEE ALSO
 392 .BR bash (1),
 393 .BR gdb (1),
 394 .BR getrlimit (2),
 395 .BR mmap (2),
 396 .BR prctl (2),
 397 .BR sigaction (2),
 398 .BR elf (5),
 399 .BR proc (5),
 400 .BR pthreads (7),
 401 .BR signal (7)