1core(5) File Formats Manual core(5)
2
6 core - core dump file
7
9 The default action of certain signals is to cause a process to termi‐
10 nate and produce a core dump file, a file containing an image of the
11 process's memory at the time of termination. This image can be used in
12 a debugger (e.g., gdb(1)) to inspect the state of the program at the
13 time that it terminated. A list of the signals which cause a process
14 to dump core can be found in signal(7).
15 A process can set its soft RLIMIT_CORE resource limit to place an upper
17 limit on the size of the core dump file that will be produced if it re‐
18 ceives a "core dump" signal; see getrlimit(2) for details.
19 There are various circumstances in which a core dump file is not pro‐
21 duced:
22 • The process does not have permission to write the core file. (By
24 default, the core file is called core or core.pid, where pid is the
25 ID of the process that dumped core, and is created in the current
26 working directory. See below for details on naming.) Writing the
27 core file fails if the directory in which it is to be created is not
28 writable, or if a file with the same name exists and is not writable
29 or is not a regular file (e.g., it is a directory or a symbolic
30 link).
31 • A (writable, regular) file with the same name as would be used for
33 the core dump already exists, but there is more than one hard link
34 to that file.
35 • The filesystem where the core dump file would be created is full; or
37 has run out of inodes; or is mounted read-only; or the user has
38 reached their quota for the filesystem.
39 • The directory in which the core dump file is to be created does not
41 exist.
42 • The RLIMIT_CORE (core file size) or RLIMIT_FSIZE (file size) re‐
44 source limits for the process are set to zero; see getrlimit(2) and
45 the documentation of the shell's ulimit command (limit in csh(1)).
46 However, RLIMIT_CORE will be ignored if the system is configured to
47 pipe core dumps to a program.
48 • The binary being executed by the process does not have read permis‐
50 sion enabled. (This is a security measure to ensure that an exe‐
51 cutable whose contents are not readable does not produce a—possibly
52 readable—core dump containing an image of the executable.)
53 • The process is executing a set-user-ID (set-group-ID) program that
55 is owned by a user (group) other than the real user (group) ID of
56 the process, or the process is executing a program that has file ca‐
57 pabilities (see capabilities(7)). (However, see the description of
58 the prctl(2) PR_SET_DUMPABLE operation, and the description of the
59 /proc/sys/fs/suid_dumpable file in proc(5).)
60 • /proc/sys/kernel/core_pattern is empty and /proc/sys/ker‐
62 nel/core_uses_pid contains the value 0. (These files are described
63 below.) Note that if /proc/sys/kernel/core_pattern is empty and
64 /proc/sys/kernel/core_uses_pid contains the value 1, core dump files
65 will have names of the form .pid, and such files are hidden unless
66 one uses the ls(1) -a option.
67 • (Since Linux 3.7) The kernel was configured without the CONFIG_CORE‐
69 DUMP option.
70 In addition, a core dump may exclude part of the address space of the
72 process if the madvise(2) MADV_DONTDUMP flag was employed.
73 On systems that employ systemd(1) as the init framework, core dumps may
75 instead be placed in a location determined by systemd(1). See below
76 for further details.
77 Naming of core dump files
79 By default, a core dump file is named core, but the /proc/sys/ker‐
80 nel/core_pattern file (since Linux 2.6 and 2.4.21) can be set to define
81 a template that is used to name core dump files. The template can con‐
82 tain % specifiers which are substituted by the following values when a
83 core file is created:
84 %% A single % character.
86 %c Core file size soft resource limit of crashing process (since
87 Linux 2.6.24).
88 %d Dump mode—same as value returned by prctl(2) PR_GET_DUMPABLE
89 (since Linux 3.7).
90 %e The process or thread's comm value, which typically is the same
91 as the executable filename (without path prefix, and truncated
92 to a maximum of 15 characters), but may have been modified to
93 be something different; see the discussion of /proc/pid/comm
94 and /proc/pid/task/tid/comm in proc(5).
95 %E Pathname of executable, with slashes ('/') replaced by exclama‐
96 tion marks ('!') (since Linux 3.0).
97 %g Numeric real GID of dumped process.
98 %h Hostname (same as nodename returned by uname(2)).
99 %i TID of thread that triggered core dump, as seen in the PID
100 namespace in which the thread resides (since Linux 3.18).
101 %I TID of thread that triggered core dump, as seen in the initial
102 PID namespace (since Linux 3.18).
103 %p PID of dumped process, as seen in the PID namespace in which
104 the process resides.
105 %P PID of dumped process, as seen in the initial PID namespace
106 (since Linux 3.12).
107 %s Number of signal causing dump.
108 %t Time of dump, expressed as seconds since the Epoch, 1970-01-01
109 00:00:00 +0000 (UTC).
110 %u Numeric real UID of dumped process.
111 A single % at the end of the template is dropped from the core file‐
113 name, as is the combination of a % followed by any character other than
114 those listed above. All other characters in the template become a lit‐
115 eral part of the core filename. The template may include '/' charac‐
116 ters, which are interpreted as delimiters for directory names. The
117 maximum size of the resulting core filename is 128 bytes (64 bytes be‐
118 fore Linux 2.6.19). The default value in this file is "core". For
119 backward compatibility, if /proc/sys/kernel/core_pattern does not in‐
120 clude %p and /proc/sys/kernel/core_uses_pid (see below) is nonzero,
121 then .PID will be appended to the core filename.
122 Paths are interpreted according to the settings that are active for the
124 crashing process. That means the crashing process's mount namespace
125 (see mount_namespaces(7)), its current working directory (found via
126 getcwd(2)), and its root directory (see chroot(2)).
127 Since Linux 2.4, Linux has also provided a more primitive method of
129 controlling the name of the core dump file. If the /proc/sys/ker‐
130 nel/core_uses_pid file contains the value 0, then a core dump file is
131 simply named core. If this file contains a nonzero value, then the
132 core dump file includes the process ID in a name of the form core.PID.
133 Since Linux 3.6, if /proc/sys/fs/suid_dumpable is set to 2 ("suid‐
135 safe"), the pattern must be either an absolute pathname (starting with
136 a leading '/' character) or a pipe, as defined below.
137 Piping core dumps to a program
139 Since Linux 2.6.19, Linux supports an alternate syntax for the
140 /proc/sys/kernel/core_pattern file. If the first character of this
141 file is a pipe symbol (|), then the remainder of the line is inter‐
142 preted as the command-line for a user-space program (or script) that is
143 to be executed.
144 Since Linux 5.3.0, the pipe template is split on spaces into an argu‐
146 ment list before the template parameters are expanded. In earlier ker‐
147 nels, the template parameters are expanded first and the resulting
148 string is split on spaces into an argument list. This means that in
149 earlier kernels executable names added by the %e and %E template param‐
150 eters could get split into multiple arguments. So the core dump han‐
151 dler needs to put the executable names as the last argument and ensure
152 it joins all parts of the executable name using spaces. Executable
153 names with multiple spaces in them are not correctly represented in
154 earlier kernels, meaning that the core dump handler needs to use mecha‐
155 nisms to find the executable name.
156 Instead of being written to a file, the core dump is given as standard
158 input to the program. Note the following points:
159 • The program must be specified using an absolute pathname (or a path‐
161 name relative to the root directory, /), and must immediately follow
162 the '|' character.
163 • The command-line arguments can include any of the % specifiers
165 listed above. For example, to pass the PID of the process that is
166 being dumped, specify %p in an argument.
167 • The process created to run the program runs as user and group root.
169 • Running as root does not confer any exceptional security bypasses.
171 Namely, LSMs (e.g., SELinux) are still active and may prevent the
172 handler from accessing details about the crashed process via
173 /proc/pid.
174 • The program pathname is interpreted with respect to the initial
176 mount namespace as it is always executed there. It is not affected
177 by the settings (e.g., root directory, mount namespace, current
178 working directory) of the crashing process.
179 • The process runs in the initial namespaces (PID, mount, user, and so
181 on) and not in the namespaces of the crashing process. One can uti‐
182 lize specifiers such as %P to find the right /proc/pid directory and
183 probe/enter the crashing process's namespaces if needed.
184 • The process starts with its current working directory as the root
186 directory. If desired, it is possible change to the working direc‐
187 tory of the dumping process by employing the value provided by the
188 %P specifier to change to the location of the dumping process via
189 /proc/pid/cwd.
190 • Command-line arguments can be supplied to the program (since Linux
192 2.6.24), delimited by white space (up to a total line length of 128
193 bytes).
194 • The RLIMIT_CORE limit is not enforced for core dumps that are piped
196 to a program via this mechanism.
197 /proc/sys/kernel/core_pipe_limit
199 When collecting core dumps via a pipe to a user-space program, it can
200 be useful for the collecting program to gather data about the crashing
201 process from that process's /proc/pid directory. In order to do this
202 safely, the kernel must wait for the program collecting the core dump
203 to exit, so as not to remove the crashing process's /proc/pid files
204 prematurely. This in turn creates the possibility that a misbehaving
205 collecting program can block the reaping of a crashed process by simply
206 never exiting.
207 Since Linux 2.6.32, the /proc/sys/kernel/core_pipe_limit can be used to
209 defend against this possibility. The value in this file defines how
210 many concurrent crashing processes may be piped to user-space programs
211 in parallel. If this value is exceeded, then those crashing processes
212 above this value are noted in the kernel log and their core dumps are
213 skipped.
214 A value of 0 in this file is special. It indicates that unlimited pro‐
216 cesses may be captured in parallel, but that no waiting will take place
217 (i.e., the collecting program is not guaranteed access to /proc/<crash‐
218 ing-PID>). The default value for this file is 0.
219 Controlling which mappings are written to the core dump
221 Since Linux 2.6.23, the Linux-specific /proc/pid/coredump_filter file
222 can be used to control which memory segments are written to the core
223 dump file in the event that a core dump is performed for the process
224 with the corresponding process ID.
225 The value in the file is a bit mask of memory mapping types (see
227 mmap(2)). If a bit is set in the mask, then memory mappings of the
228 corresponding type are dumped; otherwise they are not dumped. The bits
229 in this file have the following meanings:
230 bit 0 Dump anonymous private mappings.
232 bit 1 Dump anonymous shared mappings.
233 bit 2 Dump file-backed private mappings.
234 bit 3 Dump file-backed shared mappings.
235 bit 4 (since Linux 2.6.24)
236 Dump ELF headers.
237 bit 5 (since Linux 2.6.28)
238 Dump private huge pages.
239 bit 6 (since Linux 2.6.28)
240 Dump shared huge pages.
241 bit 7 (since Linux 4.4)
242 Dump private DAX pages.
243 bit 8 (since Linux 4.4)
244 Dump shared DAX pages.
245 By default, the following bits are set: 0, 1, 4 (if the CON‐
247 FIG_CORE_DUMP_DEFAULT_ELF_HEADERS kernel configuration option is en‐
248 abled), and 5. This default can be modified at boot time using the
249 coredump_filter boot option.
250 The value of this file is displayed in hexadecimal. (The default value
252 is thus displayed as 33.)
253 Memory-mapped I/O pages such as frame buffer are never dumped, and vir‐
255 tual DSO (vdso(7)) pages are always dumped, regardless of the core‐
256 dump_filter value.
257 A child process created via fork(2) inherits its parent's coredump_fil‐
259 ter value; the coredump_filter value is preserved across an execve(2).
260 It can be useful to set coredump_filter in the parent shell before run‐
262 ning a program, for example:
263 $ echo 0x7 > /proc/self/coredump_filter
265 $ ./some_program
266 This file is provided only if the kernel was built with the CON‐
268 FIG_ELF_CORE configuration option.
269 Core dumps and systemd
271 On systems using the systemd(1) init framework, core dumps may be
272 placed in a location determined by systemd(1). To do this, systemd(1)
273 employs the core_pattern feature that allows piping core dumps to a
274 program. One can verify this by checking whether core dumps are being
275 piped to the systemd-coredump(8) program:
276 $ cat /proc/sys/kernel/core_pattern
278 |/usr/lib/systemd/systemd-coredump %P %u %g %s %t %c %e
279 In this case, core dumps will be placed in the location configured for
281 systemd-coredump(8), typically as lz4(1) compressed files in the direc‐
282 tory /var/lib/systemd/coredump/. One can list the core dumps that have
283 been recorded by systemd-coredump(8) using coredumpctl(1):
284 $ coredumpctl list | tail -5
286 Wed 2017-10-11 22:25:30 CEST 2748 1000 1000 3 present /usr/bin/sleep
287 Thu 2017-10-12 06:29:10 CEST 2716 1000 1000 3 present /usr/bin/sleep
288 Thu 2017-10-12 06:30:50 CEST 2767 1000 1000 3 present /usr/bin/sleep
289 Thu 2017-10-12 06:37:40 CEST 2918 1000 1000 3 present /usr/bin/cat
290 Thu 2017-10-12 08:13:07 CEST 2955 1000 1000 3 present /usr/bin/cat
291 The information shown for each core dump includes the date and time of
293 the dump, the PID, UID, and GID of the dumping process, the signal
294 number that caused the core dump, and the pathname of the executable
295 that was being run by the dumped process. Various options to core‐
296 dumpctl(1) allow a specified coredump file to be pulled from the sys‐
297 temd(1) location into a specified file. For example, to extract the
298 core dump for PID 2955 shown above to a file named core in the current
299 directory, one could use:
300 $ coredumpctl dump 2955 -o core
302 For more extensive details, see the coredumpctl(1) manual page.
304 To (persistently) disable the systemd(1) mechanism that archives core
306 dumps, restoring to something more like traditional Linux behavior, one
307 can set an override for the systemd(1) mechanism, using something like:
308 # echo "kernel.core_pattern=core.%p" > \
310 /etc/sysctl.d/50-coredump.conf
311 # /lib/systemd/systemd-sysctl
312 It is also possible to temporarily (i.e., until the next reboot) change
314 the core_pattern setting using a command such as the following (which
315 causes the names of core dump files to include the executable name as
316 well as the number of the signal which triggered the core dump):
317 # sysctl -w kernel.core_pattern="%e-%s.core"
319
321 The gdb(1) gcore command can be used to obtain a core dump of a running
322 process.
323 In Linux versions up to and including 2.6.27, if a multithreaded
325 process (or, more precisely, a process that shares its memory with an‐
326 other process by being created with the CLONE_VM flag of clone(2))
327 dumps core, then the process ID is always appended to the core file‐
328 name, unless the process ID was already included elsewhere in the file‐
329 name via a %p specification in /proc/sys/kernel/core_pattern. (This is
330 primarily useful when employing the obsolete LinuxThreads implementa‐
331 tion, where each thread of a process has a different PID.)
332
334 The program below can be used to demonstrate the use of the pipe syntax
335 in the /proc/sys/kernel/core_pattern file. The following shell session
336 demonstrates the use of this program (compiled to create an executable
337 named core_pattern_pipe_test):
338 $ cc -o core_pattern_pipe_test core_pattern_pipe_test.c
340 $ su
341 Password:
342 # echo "|$PWD/core_pattern_pipe_test %p UID=%u GID=%g sig=%s" > \
343 /proc/sys/kernel/core_pattern
344 # exit
345 $ sleep 100
346 ^\ # type control-backslash
347 Quit (core dumped)
348 $ cat core.info
349 argc=5
350 argc[0]=</home/mtk/core_pattern_pipe_test>
351 argc[1]=<20575>
352 argc[2]=<UID=1000>
353 argc[3]=<GID=100>
354 argc[4]=<sig=3>
355 Total bytes in core dump: 282624
356 Program source
358 /* core_pattern_pipe_test.c */
360 #define _GNU_SOURCE
362 #include <sys/stat.h>
363 #include <fcntl.h>
364 #include <limits.h>
365 #include <stdio.h>
366 #include <stdlib.h>
367 #include <unistd.h>
368 #define BUF_SIZE 1024
370 int
372 main(int argc, char *argv[])
373 {
374 ssize_t nread, tot;
375 char buf[BUF_SIZE];
376 FILE *fp;
377 char cwd[PATH_MAX];
378 /* Change our current working directory to that of the
380 crashing process. */
381 snprintf(cwd, PATH_MAX, "/proc/%s/cwd", argv[1]);
383 chdir(cwd);
384 /* Write output to file "core.info" in that directory. */
386 fp = fopen("core.info", "w+");
388 if (fp == NULL)
389 exit(EXIT_FAILURE);
390 /* Display command-line arguments given to core_pattern
392 pipe program. */
393 fprintf(fp, "argc=%d\n", argc);
395 for (size_t j = 0; j < argc; j++)
396 fprintf(fp, "argc[%zu]=<%s>\n", j, argv[j]);
397 /* Count bytes in standard input (the core dump). */
399 tot = 0;
401 while ((nread = read(STDIN_FILENO, buf, BUF_SIZE)) > 0)
402 tot += nread;
403 fprintf(fp, "Total bytes in core dump: %zd\n", tot);
404 fclose(fp);
406 exit(EXIT_SUCCESS);
407 }
408
410 bash(1), coredumpctl(1), gdb(1), getrlimit(2), mmap(2), prctl(2),
411 sigaction(2), elf(5), proc(5), pthreads(7), signal(7), systemd-core‐
412 dump(8)
413Linux man-pages 6.04 2023-02-05 core(5)