1SHM_OPEN(P) POSIX Programmer's Manual SHM_OPEN(P)
2
6 shm_open - open a shared memory object (REALTIME)
7
9 #include <sys/mman.h>
10 int shm_open(const char *name, int oflag, mode_t mode);
12
15 The shm_open() function shall establish a connection between a shared
16 memory object and a file descriptor. It shall create an open file
17 description that refers to the shared memory object and a file descrip‐
18 tor that refers to that open file description. The file descriptor is
19 used by other functions to refer to that shared memory object. The name
20 argument points to a string naming a shared memory object. It is
21 unspecified whether the name appears in the file system and is visible
22 to other functions that take pathnames as arguments. The name argument
23 conforms to the construction rules for a pathname. If name begins with
24 the slash character, then processes calling shm_open() with the same
25 value of name refer to the same shared memory object, as long as that
26 name has not been removed. If name does not begin with the slash char‐
27 acter, the effect is implementation-defined. The interpretation of
28 slash characters other than the leading slash character in name is
29 implementation-defined.
30 If successful, shm_open() shall return a file descriptor for the shared
32 memory object that is the lowest numbered file descriptor not currently
33 open for that process. The open file description is new, and therefore
34 the file descriptor does not share it with any other processes. It is
35 unspecified whether the file offset is set. The FD_CLOEXEC file
36 descriptor flag associated with the new file descriptor is set.
37 The file status flags and file access modes of the open file descrip‐
39 tion are according to the value of oflag. The oflag argument is the
40 bitwise-inclusive OR of the following flags defined in the <fcntl.h>
41 header. Applications specify exactly one of the first two values
42 (access modes) below in the value of oflag:
43 O_RDONLY
45 Open for read access only.
46 O_RDWR Open for read or write access.
48 Any combination of the remaining flags may be specified in the value of
51 oflag:
52 O_CREAT
54 If the shared memory object exists, this flag has no effect,
55 except as noted under O_EXCL below. Otherwise, the shared memory
56 object is created; the user ID of the shared memory object shall
57 be set to the effective user ID of the process; the group ID of
58 the shared memory object is set to a system default group ID or
59 to the effective group ID of the process. The permission bits of
60 the shared memory object shall be set to the value of the mode
61 argument except those set in the file mode creation mask of the
62 process. When bits in mode other than the file permission bits
63 are set, the effect is unspecified. The mode argument does not
64 affect whether the shared memory object is opened for reading,
65 for writing, or for both. The shared memory object has a size of
66 zero.
67 O_EXCL If O_EXCL and O_CREAT are set, shm_open() fails if the shared
69 memory object exists. The check for the existence of the shared
70 memory object and the creation of the object if it does not
71 exist is atomic with respect to other processes executing
72 shm_open() naming the same shared memory object with O_EXCL and
73 O_CREAT set. If O_EXCL is set and O_CREAT is not set, the result
74 is undefined.
75 O_TRUNC
77 If the shared memory object exists, and it is successfully
78 opened O_RDWR, the object shall be truncated to zero length and
79 the mode and owner shall be unchanged by this function call. The
80 result of using O_TRUNC with O_RDONLY is undefined.
81 When a shared memory object is created, the state of the shared memory
84 object, including all data associated with the shared memory object,
85 persists until the shared memory object is unlinked and all other ref‐
86 erences are gone. It is unspecified whether the name and shared memory
87 object state remain valid after a system reboot.
88
90 Upon successful completion, the shm_open() function shall return a non-
91 negative integer representing the lowest numbered unused file descrip‐
92 tor. Otherwise, it shall return -1 and set errno to indicate the error.
93
95 The shm_open() function shall fail if:
96 EACCES The shared memory object exists and the permissions specified by
98 oflag are denied, or the shared memory object does not exist and
99 permission to create the shared memory object is denied, or
100 O_TRUNC is specified and write permission is denied.
101 EEXIST O_CREAT and O_EXCL are set and the named shared memory object
103 already exists.
104 EINTR The shm_open() operation was interrupted by a signal.
106 EINVAL The shm_open() operation is not supported for the given name.
108 EMFILE Too many file descriptors are currently in use by this process.
110 ENAMETOOLONG
112 The length of the name argument exceeds {PATH_MAX} or a pathname
113 component is longer than {NAME_MAX}.
114 ENFILE Too many shared memory objects are currently open in the system.
116 ENOENT O_CREAT is not set and the named shared memory object does not
118 exist.
119 ENOSPC There is insufficient space for the creation of the new shared
121 memory object.
122 The following sections are informative.
125
127 None.
128
130 None.
131
133 When the Memory Mapped Files option is supported, the normal open()
134 call is used to obtain a descriptor to a file to be mapped according to
135 existing practice with mmap(). When the Shared Memory Objects option
136 is supported, the shm_open() function shall obtain a descriptor to the
137 shared memory object to be mapped.
138 There is ample precedent for having a file descriptor represent several
140 types of objects. In the POSIX.1-1990 standard, a file descriptor can
141 represent a file, a pipe, a FIFO, a tty, or a directory. Many imple‐
142 mentations simply have an operations vector, which is indexed by the
143 file descriptor type and does very different operations. Note that in
144 some cases the file descriptor passed to generic operations on file
145 descriptors is returned by open() or creat() and in some cases returned
146 by alternate functions, such as pipe(). The latter technique is used by
147 shm_open().
148 Note that such shared memory objects can actually be implemented as
150 mapped files. In both cases, the size can be set after the open using
151 ftruncate(). The shm_open() function itself does not create a shared
152 object of a specified size because this would duplicate an extant func‐
153 tion that set the size of an object referenced by a file descriptor.
154 On implementations where memory objects are implemented using the
156 existing file system, the shm_open() function may be implemented using
157 a macro that invokes open(), and the shm_unlink() function may be
158 implemented using a macro that invokes unlink().
159 For implementations without a permanent file system, the definition of
161 the name of the memory objects is allowed not to survive a system
162 reboot. Note that this allows systems with a permanent file system to
163 implement memory objects as data structures internal to the implementa‐
164 tion as well.
165 On implementations that choose to implement memory objects using memory
167 directly, a shm_open() followed by an ftruncate() and close() can be
168 used to preallocate a shared memory area and to set the size of that
169 preallocation. This may be necessary for systems without virtual mem‐
170 ory hardware support in order to ensure that the memory is contiguous.
171 The set of valid open flags to shm_open() was restricted to O_RDONLY,
173 O_RDWR, O_CREAT, and O_TRUNC because these could be easily implemented
174 on most memory mapping systems. This volume of IEEE Std 1003.1-2001 is
175 silent on the results if the implementation cannot supply the requested
176 file access because of implementation-defined reasons, including hard‐
177 ware ones.
178 The error conditions [EACCES] and [ENOTSUP] are provided to inform the
180 application that the implementation cannot complete a request.
181 [EACCES] indicates for implementation-defined reasons, probably hard‐
183 ware-related, that the implementation cannot comply with a requested
184 mode because it conflicts with another requested mode. An example might
185 be that an application desires to open a memory object two times, map‐
186 ping different areas with different access modes. If the implementa‐
187 tion cannot map a single area into a process space in two places, which
188 would be required if different access modes were required for the two
189 areas, then the implementation may inform the application at the time
190 of the second open.
191 [ENOTSUP] indicates for implementation-defined reasons, probably hard‐
193 ware-related, that the implementation cannot comply with a requested
194 mode at all. An example would be that the hardware of the implementa‐
195 tion cannot support write-only shared memory areas.
196 On all implementations, it may be desirable to restrict the location of
198 the memory objects to specific file systems for performance (such as a
199 RAM disk) or implementation-defined reasons (shared memory supported
200 directly only on certain file systems). The shm_open() function may be
201 used to enforce these restrictions. There are a number of methods
202 available to the application to determine an appropriate name of the
203 file or the location of an appropriate directory. One way is from the
204 environment via getenv(). Another would be from a configuration file.
205 This volume of IEEE Std 1003.1-2001 specifies that memory objects have
207 initial contents of zero when created. This is consistent with current
208 behavior for both files and newly allocated memory. For those implemen‐
209 tations that use physical memory, it would be possible that such imple‐
210 mentations could simply use available memory and give it to the process
211 uninitialized. This, however, is not consistent with standard behavior
212 for the uninitialized data area, the stack, and of course, files.
213 Finally, it is highly desirable to set the allocated memory to zero for
214 security reasons. Thus, initializing memory objects to zero is
215 required.
216
218 None.
219
221 close() , dup() , exec() , fcntl() , mmap() , shmat() , shmctl() ,
222 shmdt() , shm_unlink() , umask() , the Base Definitions volume of
223 IEEE Std 1003.1-2001, <fcntl.h>, <sys/mman.h>
224
226 Portions of this text are reprinted and reproduced in electronic form
227 from IEEE Std 1003.1, 2003 Edition, Standard for Information Technology
228 -- Portable Operating System Interface (POSIX), The Open Group Base
229 Specifications Issue 6, Copyright (C) 2001-2003 by the Institute of
230 Electrical and Electronics Engineers, Inc and The Open Group. In the
231 event of any discrepancy between this version and the original IEEE and
232 The Open Group Standard, the original IEEE and The Open Group Standard
233 is the referee document. The original Standard can be obtained online
234 at http://www.opengroup.org/unix/online.html .
235IEEE/The Open Group 2003 SHM_OPEN(P)