man5/core.5

   1 .\" Copyright (c) 2006 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  13 .\" manual page may be incorrect or out-of-date.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.
  16 .\"
  17 .\" Formatted or processed versions of this manual, if unaccompanied by
  18 .\" the source, must acknowledge the copyright and authors of this work.
  19 .\"
  20 .TH CORE 5 2006-04-03 "Linux" "Linux Programmer's Manual"
  21 .SH NAME
  22 core \- core dump file
  23 .SH DESCRIPTION
  24 The default action of certain signals is to cause a process to terminate
  25 and produce a
  26 .IR "core dump file" ,
  27 a disk file containing an image of the process's memory at
  28 the time of termination.
  29 A list of the signals which cause a process to dump core can be found in
  30 .BR signal (7).
  31
  32 A process can set its soft
  33 .BR RLIMIT_CORE
  34 resource limit to place an upper limit on the size of the core dump file
  35 that will be produced if it receives a "core dump" signal; see
  36 .BR getrlimit (2)
  37 for details.
  38
  39 There are various circumstances in which a core dump file is
  40 not produced:
  41 .IP *
  42 The process does not have permission to write the core file.
  43 (By default the core file is called
  44 .IR core ,
  45 and is created in the current working directory.
  46 See below for details on naming.)
  47 Writing the core file will fail if the directory in which
  48 it is to be created is non-writable,
  49 or if a file with the same name exists and
  50 is not writable
  51 or is not a regular file
  52 (e.g., it is a directory or a symbolic link).
  53 .IP *
  54 A (writable, regular) file with the same name as would be used for the
  55 core dump already exists, but there is more than one hard link to that
  56 file.
  57 .IP *
  58 The file system where the core dump file would be created is full;
  59 or has run out of inodes; or is mounted read only;
  60 or the user has reached their quota for the file system.
  61 .IP *
  62 The directory in which the core dump file is to be created does
  63 not exist.
  64 .IP *
  65 .B RLIMIT_CORE
  66 or
  67 .B RLIMIT_FSIZE
  68 resource limits for a process are set to zero (see
  69 .BR getrlimit (2)).
  70 .IP *
  71 The binary being executed by the process does not have read
  72 permission enabled.
  73 .IP *
  74 The process is executing a set-user-ID (set-group-ID) program
  75 that is owned by a user (group) other than the real user (group)
  76 ID of the process.
  77 (However, see the description of the
  78 .BR prctl (2)
  79 .B PR_SET_DUMPABLE
  80 operation, and the description of the
  81 .I /proc/sys/fs/suid_dumpable
  82 file in
  83 .BR proc (5).)
  84 .SS Naming of core dump files
  85 By default, a core dump file is named
  86 .IR core ,
  87 but the
  88 .I /proc/sys/kernel/core_pattern
  89 file
  90 (since Linux 2.6 and 2.4.21)
  91 can be set to define a template that is used to name core dump files.
  92 The template can contain % specifiers which are substituted
  93 by the following values when a core file is created:
  94 .nf
  95
  96   %%  A single % character
  97   %p  PID of dumped process
  98   %u  real UID of dumped process
  99   %g  real GID of dumped process
 100   %s  number of signal causing dump
 101   %t  time of dump (seconds since 0:00h, 1 Jan 1970)
 102   %h  hostname (same as 'nodename' returned by \fBuname\fP(2))
 103   %e  executable filename
 104
 105 .fi
 106 A single % at the end of the template is dropped from the
 107 core filename, as is the combination of a % followed by any
 108 character other than those listed above.
 109 All other characters in the template become a literal
 110 part of the core filename.
 111 The template may include '/' characters, which are interpreted
 112 as delimiters for directory names.
 113 The maximum size of the resulting core filename is 64 bytes.
 114 The default value in this file is "core".
 115 For backward compatibility, if
 116 .I /proc/sys/kernel/core_pattern
 117 does not include "%p" and
 118 .I /proc/sys/kernel/core_uses_pid
 119 (see below)
 120 is non-zero, then .PID will be appended to the core filename.
 121
 122 Since version 2.4, Linux has also provided
 123 a more primitive method of controlling
 124 the name of the core dump file.
 125 If the
 126 .I /proc/sys/kernel/core_uses_pid
 127 file contains the value 0, then a core dump file is simply named
 128 .IR core .
 129 If this file contains a non-zero value, then the core dump file includes
 130 the process ID in a name of the form
 131 .IR core.PID .
 132 .SH NOTES
 133 The
 134 .BR gdb (1)
 135 .I gcore
 136 command can be used to obtain a core dump of a running process.
 137
 138 If a multithreaded process (or, more precisely, a process that
 139 shares its memory with another process by being created with the
 140 .B CLONE_VM
 141 flag of
 142 .BR clone (2))
 143 dumps core, then the process ID is always appended to the core filename,
 144 unless the process ID was already included elsewhere in the
 145 filename via a %p specification in
 146 .IR /proc/sys/kernel/core_pattern .
 147 (This is primarily useful when employing the LinuxThreads implementation,
 148 where each thread of a process has a different PID.)
 149 .\" Always including the PID in the name of the core file made
 150 .\" sense for LinuxThreads, where each thread had a unique PID,
 151 .\" but doesn't seem to serve any purpose with NPTL, where all the
 152 .\" threads in a process share the same PID (as POSIX.1 requires).
 153 .\" Probably the behavior is maintained so that applications using
 154 .\" LinuxThreads continue appending the PID (the kernel has no easy
 155 .\" way of telling which threading implementation the userspace
 156 .\" application is using). -- mtk, April 2006
 157 .SH SEE ALSO
 158 .BR gdb (1),
 159 .BR getrlimit (2),
 160 .BR prctl (2),
 161 .BR sigaction (2),
 162 .BR elf (5),
 163 .BR proc (5),
 164 .BR pthreads (7),
 165 .BR signal (7)