man2/shmop.2

   1 .\" Copyright 1993 Giorgio Ciucci (giorgio@crcc.it)
   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  13 .\" manual page may be incorrect or out-of-date.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 .\" Modified Sun Nov 28 17:06:19 1993, Rik Faith (faith@cs.unc.edu)
  24 .\"          with material from Luigi P. Bai (lpb@softint.com)
  25 .\" Portions Copyright 1993 Luigi P. Bai
  26 .\" Modified Tue Oct 22 22:04:23 1996 by Eric S. Raymond <esr@thyrsus.com>
  27 .\" Modified, 5 Jan 2002, Michael Kerrisk <mtk.manpages@gmail.com>
  28 .\" Modified, 19 Sep 2002, Michael Kerrisk <mtk.manpages@gmail.com>
  29 .\"     Added SHM_REMAP flag description
  30 .\" Modified, 27 May 2004, Michael Kerrisk <mtk.manpages@gmail.com>
  31 .\"     Added notes on capability requirements
  32 .\" Modified, 11 Nov 2004, Michael Kerrisk <mtk.manpages@gmail.com>
  33 .\"     Language and formatting clean-ups
  34 .\"     Changed wording and placement of sentence regarding attachment
  35 .\"             of segments marked for destruction
  36 .\"
  37 .\" FIXME . Add an example program to this page.
  38 .\" FIXME Linux 2.6.9 added SHM_EXEC, which should be documented
  39 .TH SHMOP 2 2013-02-12 "Linux" "Linux Programmer's Manual"
  40 .SH NAME
  41 shmat, shmdt \- shared memory operations
  42 .SH SYNOPSIS
  43 .nf
  44 .B #include <sys/types.h>
  45 .B #include <sys/shm.h>
  46
  47 .BI "void *shmat(int " shmid ", const void *" shmaddr ", int " shmflg );
  48
  49 .BI "int shmdt(const void *" shmaddr );
  50 .fi
  51 .SH DESCRIPTION
  52 .BR shmat ()
  53 attaches the shared memory segment identified by
  54 .I shmid
  55 to the address space of the calling process.
  56 The attaching address is specified by
  57 .I shmaddr
  58 with one of the following criteria:
  59 .LP
  60 If
  61 .I shmaddr
  62 is NULL,
  63 the system chooses a suitable (unused) address at which to attach
  64 the segment.
  65 .LP
  66 If
  67 .I shmaddr
  68 isn't NULL
  69 and
  70 .B SHM_RND
  71 is specified in
  72 .IR shmflg ,
  73 the attach occurs at the address equal to
  74 .I shmaddr
  75 rounded down to the nearest multiple of
  76 .BR SHMLBA .
  77 Otherwise
  78 .I shmaddr
  79 must be a page-aligned address at which the attach occurs.
  80 .PP
  81 If
  82 .B SHM_RDONLY
  83 is specified in
  84 .IR shmflg ,
  85 the segment is attached for reading and the process must have
  86 read permission for the segment.
  87 Otherwise the segment is attached for read and write
  88 and the process must have read and write permission for the segment.
  89 There is no notion of a write-only shared memory segment.
  90 .PP
  91 The (Linux-specific)
  92 .B SHM_REMAP
  93 flag may be specified in
  94 .I shmflg
  95 to indicate that the mapping of the segment should replace
  96 any existing mapping in the range starting at
  97 .I shmaddr
  98 and continuing for the size of the segment.
  99 (Normally an
 100 .B EINVAL
 101 error would result if a mapping already exists in this address range.)
 102 In this case,
 103 .I shmaddr
 104 must not be NULL.
 105 .PP
 106 The
 107 .BR brk (2)
 108 value of the calling process is not altered by the attach.
 109 The segment will automatically be detached at process exit.
 110 The same segment may be attached as a read and as a read-write
 111 one, and more than once, in the process's address space.
 112 .PP
 113 A successful
 114 .BR shmat ()
 115 call updates the members of the
 116 .I shmid_ds
 117 structure (see
 118 .BR shmctl (2))
 119 associated with the shared memory segment as follows:
 120 .IP
 121 .I shm_atime
 122 is set to the current time.
 123 .IP
 124 .I shm_lpid
 125 is set to the process-ID of the calling process.
 126 .IP
 127 .I shm_nattch
 128 is incremented by one.
 129 .PP
 130 .BR shmdt ()
 131 detaches the shared memory segment located at the address specified by
 132 .I shmaddr
 133 from the address space of the calling process.
 134 The to-be-detached segment must be currently
 135 attached with
 136 .I shmaddr
 137 equal to the value returned by the attaching
 138 .BR shmat ()
 139 call.
 140 .PP
 141 On a successful
 142 .BR shmdt ()
 143 call the system updates the members of the
 144 .I shmid_ds
 145 structure associated with the shared memory segment as follows:
 146 .IP
 147 .I shm_dtime
 148 is set to the current time.
 149 .IP
 150 .I shm_lpid
 151 is set to the process-ID of the calling process.
 152 .IP
 153 .I shm_nattch
 154 is decremented by one.
 155 If it becomes 0 and the segment is marked for deletion,
 156 the segment is deleted.
 157 .PP
 158 After a
 159 .BR fork (2)
 160 the child inherits the attached shared memory segments.
 161
 162 After an
 163 .BR execve (2)
 164 all attached shared memory segments are detached from the process.
 165
 166 Upon
 167 .BR _exit (2)
 168 all attached shared memory segments are detached from the process.
 169 .SH RETURN VALUE
 170 On success
 171 .BR shmat ()
 172 returns the address of the attached shared memory segment; on error
 173 .I (void\ *)\ \-1
 174 is returned, and
 175 .I errno
 176 is set to indicate the cause of the error.
 177
 178 On success
 179 .BR shmdt ()
 180 returns 0; on error \-1 is returned, and
 181 .I errno
 182 is set to indicate the cause of the error.
 183 .SH ERRORS
 184 When
 185 .BR shmat ()
 186 fails,
 187 .I errno
 188 is set to one of the following:
 189 .TP
 190 .B EACCES
 191 The calling process does not have the required permissions for
 192 the requested attach type, and does not have the
 193 .B CAP_IPC_OWNER
 194 capability.
 195 .TP
 196 .B EIDRM
 197 \fIshmid\fP points to a removed identifier.
 198 .TP
 199 .B EINVAL
 200 Invalid
 201 .I shmid
 202 value, unaligned (i.e., not page-aligned and \fBSHM_RND\fP was not
 203 specified) or invalid
 204 .I shmaddr
 205 value, or can't attach segment at
 206 .IR shmaddr ,
 207 or
 208 .B SHM_REMAP
 209 was specified and
 210 .I shmaddr
 211 was NULL.
 212 .TP
 213 .B ENOMEM
 214 Could not allocate memory for the descriptor or for the page tables.
 215 .PP
 216 When
 217 .BR shmdt ()
 218 fails,
 219 .I errno
 220 is set as follows:
 221 .TP
 222 .B EINVAL
 223 There is no shared memory segment attached at
 224 .IR shmaddr ;
 225 or,
 226 .\" The following since 2.6.17-rc1:
 227 .I shmaddr
 228 is not aligned on a page boundary.
 229 .SH CONFORMING TO
 230 SVr4, POSIX.1-2001.
 231 .\" SVr4 documents an additional error condition EMFILE.
 232
 233 In SVID 3 (or perhaps earlier)
 234 the type of the \fIshmaddr\fP argument was changed from
 235 .I "char *"
 236 into
 237 .IR "const void *" ,
 238 and the returned type of
 239 .BR shmat ()
 240 from
 241 .I "char *"
 242 into
 243 .IR "void *" .
 244 (Linux libc4 and libc5 have the
 245 .I "char *"
 246 prototypes; glibc2 has
 247 .IR "void *" .)
 248 .SH NOTES
 249 Using
 250 .BR shmat ()
 251 with
 252 .I shmaddr
 253 equal to NULL
 254 is the preferred, portable way of attaching a shared memory segment.
 255 Be aware that the shared memory segment attached in this way
 256 may be attached at different addresses in different processes.
 257 Therefore, any pointers maintained within the shared memory must be
 258 made relative (typically to the starting address of the segment),
 259 rather than absolute.
 260 .PP
 261 On Linux, it is possible to attach a shared memory segment even if it
 262 is already marked to be deleted.
 263 However, POSIX.1-2001 does not specify this behavior and
 264 many other implementations do not support it.
 265 .LP
 266 The following system parameter affects
 267 .BR shmat ():
 268 .TP
 269 .\" FIXME A good explanation of the rationale for the existence
 270 .\" of SHMLBA would be useful here
 271 .B SHMLBA
 272 Segment low boundary address multiple.
 273 Must be page aligned.
 274 For the current implementation the
 275 .B SHMLBA
 276 value is
 277 .BR PAGE_SIZE .
 278 .\" FIXME That last sentence isn't true for all Linux
 279 .\" architectures (i.e., SHMLBA != PAGE_SIZE for some architectures)
 280 .\" -- MTK, Nov 04
 281 .PP
 282 The implementation places no intrinsic limit on the per-process maximum
 283 number of shared memory segments
 284 .RB ( SHMSEG ).
 285 .SH SEE ALSO
 286 .BR brk (2),
 287 .BR mmap (2),
 288 .BR shmctl (2),
 289 .BR shmget (2),
 290 .BR capabilities (7),
 291 .BR shm_overview (7),
 292 .BR svipc (7)