This section
describes all of the system calls.
This introduction describes all error return codes,
gives an overview of privileges,
threads and lightweight processes,
and defines common terms.
The constants
ARG_MAX,
SYS_OPEN,
and so on,
are implementation-specific constants defined in
limits.h.
See
limits(F).
Return values
The system call manual pages describe system call return values.
Most of these calls have one or more error returns.
An error condition is indicated by an otherwise
impossible return value.
This is almost always -1 or the
NULL pointer;
the individual descriptions specify the details.
Errors
Besides returning to the caller a value that indicates
that an error has occurred,
a system call also sets the external variable
errno to a value that identifies the reason for the error.
errno
is not cleared on successful calls, so it should be tested only
after an error has been indicated.
Each system call description attempts to list
all possible error numbers.
The following is a complete list of the error numbers and
their names as defined in errno.h.
1 EPERM Not privileged
Typically this error indicates
an attempt to modify a file in some way forbidden
by the privilege mechanism, or restricted to the owner of the file.
It is also returned when an attempt is made to open a device already
open by another process.
See the
``Privileges''
section and
``Access permissions''
in the
``Definitions''
section below.
2 ENOENT No such file or directory
A file name is specified and the file should exist but doesn't,
or one of the directories in a pathname does not exist.
3 ESRCH No such process
No process can be found corresponding to the process identifier
specified.
4 EINTR Interrupted system call
An asynchronous signal (such as interrupt or quit),
which the user has elected to catch,
occurred during a system service routine.
If execution is resumed
after processing the signal,
it will appear as if the interrupted routine call
returned this error condition.
5 EIO I/O error
Some physical I/O error has occurred.
This error may in some cases occur
on a call following the one to which it actually applies.
6 ENXIO No such device or address
I/O on a special file refers to a subdevice which does not
exist, or exists beyond the limit of the device.
It may also occur when, for example, a tape drive
is not on-line or no disk pack is loaded on a drive.
7 E2BIG Arg list too long
An argument list longer than
ARG_MAX
bytes is presented to a member of the
exec family of routines.
The argument list limit is the sum of the size of the argument
list plus the size of the environment's exported shell variables.
8 ENOEXEC Exec format error
A request is made to execute a file
which, although it has the appropriate permissions,
does not start with a valid format (see
a.out(F)).
9 EBADF Bad file number
Either a file descriptor refers to no open file, or a
read (respectively, write) request is made to
a file that is open only for writing (respectively, reading).
10 ECHILD No child processes
A wait routine
was executed by a process that had no existing or unwaited-for child processes.
11 EAGAIN Resource temporarily unavailable
For example, the fork routine failed because the system's process table
is full or the user is not allowed to create any more processes,
or a system call failed because of insufficient memory or swap space.
12 ENOMEM Insufficient or invalid memory
During execution of an exec, brk, or sbrk routine,
a program asks for more space than the system is able to supply.
This is not a temporary condition; the maximum size
is a system parameter.
The error may also occur if the arrangement
of text, data, and stack segments
requires too many segmentation registers, or
if there is not enough swap space during the
fork routine.
If this error occurs on a resource associated with
Remote File Sharing (RFS),
it indicates a memory depletion
which may be temporary, dependent on system activity at the time the
call was invoked.
13 EACCES Permission denied
An attempt was made to access a file in a way forbidden
by the access control mechanism.
The access control mechanism grants or denies a process
permission to access an object based on a comparison of the access
attributes of the calling process, the access attributes of
the object being accessed, and the type of access requested.
The access control attributes of the process are its real and
effective uids, and the privileges assigned to the process
when it was created.
The access control attributes of the object are the permission mode
bits, and possibly the Access Control List (ACL). assigned
to the object; these mechanisms are known as Discretionary Access
Control (DAC).
Failure of the process to meet the access control checks causes denial
of the requested access and the return of EACCES.
See
``Access permissions''
in the
``Definitions''
section below for more information.
14 EFAULT Bad address
The system encountered a hardware fault in attempting to
use an argument of a routine.
For example, errno potentially may be set to EFAULT
any time a routine that takes a pointer argument is passed an invalid address,
if the system can detect the condition.
Because systems will differ in their ability to reliably detect a bad
address, on some implementations passing a bad address to a routine
will result in undefined behavior.
15 ENOTBLK Block device required
A non-block file was mentioned where a block device was required
(for example, in a call to the mount routine).
16 EBUSY Device busy
The device or resource is currently unavailable.
An attempt was made to do one of the following: mount a device that
was already mounted;
unmount a device on which there is an active file
(open file, current directory, mounted-on file, active text segment);
enable accounting when it is already enabled;
or, open a device that is in the process of closing.
17 EEXIST File exists
An existing file was mentioned in an inappropriate context
(for example, call to the link routine).
18 EXDEV Cross-device link
A link to a file on another device was attempted.
19 ENODEV No such device
An attempt was made to apply an inappropriate operation to a device
(for example, read a write-only device, open a device not yet allocated).
20 ENOTDIR Not a directory
A non-directory was specified where a directory is required
(for example, in a path prefix or as an argument to the
chdir routine).
21 EISDIR Is a directory
An attempt was made to perform an operation not appropriate for a directory,
such as
write(S).
22 EINVAL Invalid argument
An invalid argument was specified
(for example, unmounting a non-mounted device
or specifying an undefined signal in a call to
sigaction(S)
or
kill(S)).
23 ENFILE File table overflow
The system file table is full (that is,
SYS_OPEN files are open, and temporarily no more files can be opened).
24 EMFILE Too many open files
The number of open file descriptors has reached the current soft limit.
The maximum number of open file descriptors is a per-process resource
determined by the tuneable parameter SFNOLIM, the soft or current limit,
specifying the max number of open files.
HFNOLIM is the hard limit or the maximum value of SFNOLIM.
The defaults for these tuneables are 64 (SFNOLIM) and 2048 (HFNOLIM).
See
ulimit(C)
and
setrlimit(S)
for information on getting and setting these values.
Also see
``Process limit parameters'' in Monitoring and tuning the system.
25 ENOTTY Not a typewriter
A call was made to the ioctl routine specifying a file that
is not a special character device.
26 ETXTBSY Text file busy
An attempt was made to execute a pure-procedure
program that is currently open for writing.
Also an attempt to open for writing or to remove a pure-procedure
program that is being executed.
27 EFBIG File too large
The size of a file exceeded the maximum file size,
FCHR_MAX
(see getrlimit).
28 ENOSPC No space left on device
While writing an ordinary file or creating a directory entry,
there is no free space left on the device.
In the fcntl routine, the setting or removing of record locks on a file
cannot be accomplished because there are no more record entries
left on the system.
29 ESPIPE Illegal seek
A call to the lseek routine was issued to a pipe.
30 EROFS Read-only file system
An attempt to modify a file or directory was made
on a file system which was mounted read-only.
31 EMLINK Too many links
An attempt to make more than the maximum number of links,
LINK_MAX, to a file.
32 EPIPE Broken pipe
A write on a pipe for which there is no process to read the data.
This condition normally generates a signal;
the error is returned if the signal is ignored.
33 EDOM Math argument out of domain of func
The argument of a function in the math package (3M)
is out of the domain of the function.
34 ERANGE Math result not representable
The value of a function in the math package (3M)
is not representable within machine precision.
35 ENOMSG No message of desired type
An attempt was made to receive a message of a type
that does not exist on the specified message queue (see
msgop(S)).
36 EIDRM Identifier removed
This error is returned to processes that resume execution due to the removal
of a message or semaphore identifier from the system (see
msgop(S),
semop(S),
msgctl(S),
and
semctl(S)).
37 ECHRNG Channel number out of range
38 EL2NSYNC Level 2 not synchronized
39 EL3HLT Level 3 halted
40 EL3RST Level 3 reset
41 ELNRNG Link number out of range
42 EUNATCH Protocol driver not attached
43 ENOCSI No CSI structure available
44 EL2HLT Level 2 halted
45 EDEADLK Deadlock condition
A deadlock situation was detected and avoided.
This error pertains to file and record locking.
46 ENOLCK No record locks available
There are no more locks available.
The system lock table is full (see
fcntl(S)).
47-49 Reserved numbers
58-59 Reserved numbers
60 ENOSTR Device not a stream
A putmsg or getmsg
system call was attempted on
a file descriptor that is not a
STREAMS
device.
61 ENODATA No data available
62 ETIME Timer expired
The timer set for a
STREAMS
ioctl call has expired.
The cause of this error is device specific and could indicate either
a hardware or software failure, or perhaps a timeout value
that is too short for the specific operation.
The status of the ioctl operation is indeterminate.
63 ENOSR Out of stream resources
During a
STREAMS
open, either no
STREAMS
queues or no
STREAMS
head data structures were available.
This is a temporary condition; one may recover from it
if other processes release resources.
64 ENONET Machine is not on the network
This error is Remote File Sharing (RFS) specific.
It occurs when users try to advertise,
unadvertise, mount, or unmount remote resources while the machine has not
done the proper startup to connect to the network.
65 ENOPKG Package not installed
This error occurs when users attempt to use
a system call from a package which has not been installed.
66 EREMOTE Object is remote
This error is
RFS
specific.
It occurs when users try to advertise
a resource which is not on the local machine, or try to
mount/unmount a device (or pathname) that is on a remote machine.
67 ENOLINK Link has been severed
This error is
RFS
specific.
It occurs when the link (virtual
circuit) connecting to a remote machine is gone.
68 EADV Advertise error
This error is
RFS
specific.
It occurs when users try to advertise
a resource which has been advertised already, or try to stop
RFS
while there are resources still advertised, or try to force
unmount a resource when it is still advertised.
69 ESRMNT Srmount error
This error is
RFS
specific.
It occurs when an attempt is made to stop
RFS
while resources are still mounted by remote machines,
or when a resource is readvertised with a client list that
does not include a remote machine that currently has the resource mounted.
70 ECOMM Communication error on send
This error is
RFS
specific.
It occurs when the current process is waiting for a
message from a remote machine, and the virtual circuit fails.
71 EPROTO Protocol error
Some protocol error occurred.
This error is device specific,
but is generally not related to a hardware failure.
74 EMULTIHOP Multihop attempted
This error is
RFS
specific.
It occurs when users try to
access remote resources which are not directly accessible.
76 EDOTDOT Error 76
This error is
RFS
specific.
A way for the server to tell the client that a process has transferred
back from mount point.
77 EBADMSG Not a data message
During a read, getmsg, or ioctlI_RECVFD system call
to a STREAMS device, something has come to the head of the queue
that can't be processed.
That something depends on the system call:
read
control information or a passed file descriptor
getmsg
passed file descriptor
ioctl
control or data information
78 ENAMETOOLONG File name too long
The length of the path argument exceeds PATH_MAX, or the
length of a path component exceeds NAME_MAX while
_POSIX_NO_TRUNC is in effect; see
limits(F).
79 EOVERFLOW Value too large for defined data type
80 ENOTUNIQ Name not unique on network
Given log name not unique.
81 EBADFD File descriptor in bad state
Either a file descriptor refers to no open file or a read request was
made to a file that is open only for writing.
82 EREMCHG Remote address changed
83 ELIBACC Cannot access a needed shared library
Trying to exec an a.out that requires a static shared library
and the static shared library doesn't exist or
the user doesn't have permission to use it.
84 ELIBBAD Accessing a corrupted shared library
Trying to exec an a.out that requires a static shared library
(to be linked in) and exec could not load the static shared library.
The static shared library is probably corrupted.
85 ELIBSCN.lib section in a.out corrupted
Trying to exec an a.out that requires a static shared library
(to be linked in) and there was erroneous data in the .lib
section of the a.out.
The .lib section tells exec what static shared libraries are needed.
The a.out is probably corrupted.
86 ELIBMAX Attempting to link in more shared libraries than system limit
Trying to exec an a.out that requires more static shared
libraries than is allowed on the current configuration of the system.
See your system administration guide.
87 ELIBEXEC Cannot exec a shared library directly
Attempting to exec a shared library directly.
88 EILSEQ Illegal byte sequence
Illegal byte sequence.
Handle multiple characters as a single character.
89 ENOSYS Operation not applicable
90 ELOOP Too many symbolic links in pathname traversal
91 ERESTART Restartable system call
Interrupted system call should be restarted.
92 ESTRPIPE Streams pipe error
Streams pipe error (not externally visible).
93 ENOTEMPTY Directory not empty
94 EUSERS Too many users
95 ENOTSOCK Socket operation on non-socket
96 EDESTADDRREQ Destination address required
A required address was omitted from an operation on a transport endpoint.
Destination address required.
97 EMSGSIZE Message too long
A message sent on a transport provider was larger than the internal message
buffer or some other network limit.
98 EPROTOTYPE Protocol wrong type for socket
A protocol was specified that does not support the semantics of
the socket type requested.
99 ENOPROTOOPT Protocol not available
A bad option or level was specified when getting or setting options
for a protocol.
120 EPROTONOSUPPORT Protocol not supported
The protocol has not been configured into the system
or no implementation for it exists.
121 ESOCKTNOSUPPORT Socket type not supported
The support for the socket type has not been configured into the system
or no implementation for it exists.
122 EOPNOTSUPP Operation not supported on transport endpoint
For example, trying to accept a connection on a datagram transport endpoint.
123 EPFNOSUPPORT Protocol family not supported
The protocol family has not been configured into the system
or no implementation for it exists.
Used for the Internet protocols.
124 EAFNOSUPPORT Address family not supported by protocol family
An address incompatible with the requested protocol was used.
125 EADDRINUSE Address already in use
User attempted to use an address already in use,
and the protocol does not allow this.
126 EADDRNOTAVAIL Cannot assign requested address
Results from an attempt to create a transport endpoint with
an address not on the current machine.
127 ENETDOWN Network is down
Operation encountered a dead network.
128 ENETUNREACH Network is unreachable
Operation was attempted to an unreachable network.
129 ENETRESET Network dropped connection because of reset
The host you were connected to crashed and rebooted.
130 ECONNABORTED Software caused connection abort
A connection abort was caused internal to your host machine.
131 ECONNRESET Connection reset by peer
A connection was forcibly closed by a peer.
This normally results from a loss of the connection on the remote
host due to a timeout or a reboot.
132 ENOBUFS No buffer space available
An operation on a transport endpoint or pipe was not performed
because the system lacked sufficient buffer space or because
a queue was full.
133 EISCONN Transport endpoint is already connected
A connect request was made on an already connected transport endpoint;
or, a sendto or sendmsg request on a connected
transport endpoint specified a destination when already connected.
134 ENOTCONN Transport endpoint is not connected
A request to send or receive data was disallowed because the transport
endpoint is not connected and (when sending a datagram) no address was supplied.
138 ENAVAIL No more maps allowed or available
When setting a map with
mapchan(ADM),
the limit of the NEMAP tuneable was exceeded and no more maps were
allowed to be set.
Or, when getting a map with
mapchan(ADM),
there is no map available for the device specified.
143 ESHUTDOWN Cannot send after transport endpoint shutdown
A request to send data was disallowed because the transport endpoint has already
been shut down.
144 ETOOMANYREFS Too many references: cannot splice
145 ETIMEDOUT Connection timed out
A connect or send request failed because the connected party
did not properly respond after a period of time.
(The timeout period is dependent on the communication protocol.)
146 ECONNREFUSED Connection refused
No connection could be made because the target machine actively refused it.
This usually results from trying to connect to a service
that is inactive on the remote host.
147 EHOSTDOWN Host is down
A transport provider operation failed because the destination host was down.
148 EHOSTUNREACH No route to host
A transport provider operation was attempted to an unreachable host.
149 EALREADY Operation already in progress
An operation was attempted on a non-blocking object that already had an
operation in progress.
150 EINPROGRESS Operation now in progress
An operation that takes a long time to complete (such as a connect)
was attempted on a non-blocking object.
151 ESTALE Stale NFS file handle
152 ENOLOAD Could not load the required loadable module
153 ERELOC Relocation error when module being loaded
154 ENOMATCH No symbol was found matching the given specification
156 EBADVER Version number mismatch
157 ECONFIG Configured kernel resource exhausted
158 ECANCELLED Asynchronous I/O canceled
160 ENOTAUTH Not authenticated
An attempt was made to access a server, but access was denied because
the user was not authenticated to the server (and the reason for this
was not a severed link).
Privileges
All of the sensitive system operations that require special privileges
have been identified and specific privileges defined for one or more of
these services.
A process may perform a sensitive service only if it has the required
privilege.
By default, the system runs with a privilege module known as
the Super User Module (SUM).
This module gives the privileged user ID,
uid ``0'' in the delivered system,
all of these privileges.
Thus the SUM retains compatibility with traditional UNIX systems,
wherein the root login is given unrestricted system access.
Individual privileges can also be assigned to IDs other than
uid 0, so that these IDs can run privileged commands.
Privileges and IDs are managed using the
adminrole(ADM),
adminuser(ADM),
and
tfadmin(ADM)
commands.
Privileges are assigned to trusted executable files using the
filepriv(ADM)
command and the
filepriv(S)
system call.
For a discussion of privileges see
``Privileges'' in Programming with system calls and libraries.
Following is the list of privileges defined in sys/privilege.h:
0 P_OWNER
Required to change the attributes of a file (that is, information kept
in the file's inode)
that is not owned by the effective uid of the calling process.
See
``Access permissions''
in the
``Definitions''
section below.
1 P_AUDIT
Required to manipulate the security audit mechanisms.
2 P_COMPAT
Overrides specific restrictions that are imposed solely for the
confinement of covert channels.
3 P_DACREAD
Overrides Discretionary Access Control (DAC) restrictions but only for
operations that do not alter objects (that is, read and execute
permissions).
See
``Access permissions''
in the
``Definitions''
section below.
4 P_DACWRITE
Overrides Discretionary Access Control restrictions but only for
operations that alter objects (that is, write permission).
See
``Access permissions''
in the
``Definitions''
section below.
5 P_DEV
Required to set or get device security attributes to change the device
level when it is in private state, and to access a device when it is in
private state.
This privilege is also used for special ioctl
for window management and to download trusted software to a terminal driver.
6 P_FILESYS
Required for privileged operations on a file system that have relatively
low sensitivity, including the creation of links to directories, setting
the effective root directory, and making special files.
7 P_MACREAD
Currently unused.
8 P_MACWRITE
Currently unused.
9 P_MOUNT
Mount or unmount a file system.
10 P_MULTIDIR
Currently unused.
11 P_SETPLEVEL
Currently unused.
12 P_SETSPRIV
Administrative privilege required to set the inheritable and fixed
privileges on files.
This privilege overrides access and ownership restrictions.
13 P_SETUID
Required in order to set the real and effective user and group
IDs of a process.
14 P_SYSOPS
Required to perform several general system operations that have only
minor security implications.
15 P_SETUPRIV
Privilege required for an otherwise unprivileged process to set the
inheritable and fixed privileges on a file.
This privilege does not override
access or ownership restrictions.
16 P_DRIVER
Provides compatibility with device drivers developed by third party vendors.
It is used when a sensitive operation needs to be limited to a privileged process.
17 P_RTIME
Required by processes that do real-time operations.
18 P_MACUPGRADE
Currently unused.
19 P_FSYSRANGE
Override file system range restrictions.
20 P_SETFLEVEL
Currently unused.
21 P_AUDITWR
Required to write miscellaneous audit records to the audit trail.
22 P_TSHAR
Required to raise the priority of a time sharing process or to set
the user priority limit to a value greater than 0.
23 P_PLOCK
Required to lock a process in memory.
24 P_CORE
Required to dump a core image of a process that is either privileged,
setuid, or setgid.
This privilege is not required to dump the core image of a process that
does not meet the above conditions.
25 P_LOADMOD
Required to perform selective operations associated with loadable modules.
26 P_BIND
Required to bind a process, an LWP,
or a set of LWPs to a specific processor.
P_ALLPRIVS
Represents all possible privileges.
Large File Support
With Large File Support (LFS), file size and related sizes
are no longer tied to the 32 bit limit formerly in effect for
C language data types.
Compiler extensions allow for a C application to conveniently
manipulate files of such size via 64 bit
integers (called ``long long'').
In prior releases, the size of a file was limited to
only as much information as could be addressed using a signed long integer.
On most implementations that is 2 gigabytes ([2 to the 31] -1).
Although LFS is designed to allow the kernel to handle
files containing up to
([2 to the 63] -1) bytes, the current implementation of UnixWare can handle
only files that contain up to ([2 to the 40] -1) bytes, or one terabyte
of disk store.
A sparse file can
exhibit length up to 2^63 bytes.
NOTE:
Large file support is intended for specialized applications which
need to deal with more than 2GB of data in a single file,
and not for general purpose text files and similar.
Prerequisites for programming with large files
The file system on which the application is creating large files must
be a vxfs file system created with the largefiles
option.
See the description of the largefiles command option on
mkfs_vxfs(ADM)
also see
``Large files'' in Understanding filesystem types.
Large file support can be added to an existing vxfs file
system using
fsadm_vxfs(ADM).
In addition, the file size limit (the RLIMIT_FSIZE resource)
must be set to RLIM_INFINITY
(see
setrlimit(S)).
Note that if your application does not support large files, it does not
need any modifications to run correctly on a large file aware file
system.
If it happens to access a file greater than 2GB in size, it will fail
gracefully, assuming normal error checking practices.
Programming considerations for large file support
Large File Support provides a mechanism to cause certain data types,
data structures, and system calls (listed in
``System calls and library routines that map to 32 and 64 bit versions'')
to switch from
``regular'' to ``large''.
Applications can be made to handle either regular or large files in
one of two ways.
First, you can define the macro
_FILE_OFFSET_BITS at compile time by using the -D
option to the cc command,
(that is, -D _FILE_OFFSET_BITS=num).
Here, num can be 32, signifying regular files,
or 64, signifying large files.
NOTE:
If you do not define _FILE_OFFSET_BITS at compile
time, no mapping occurs, and thus it defaults to the mode
that is native on your system (currently 32 bits).
These interfaces, on 64 bit systems, will have 64 bit sizes.
Refer to
sys/types.h
for details on the effects of defining
_FILE_OFFSET_BITS
to 32 in a native 64 bit environment.
Explicit 32 bit versions of the data types are defined there.
They are not documented elsewhere.
For example,
-D _FILE_OFFSET_BITS=64
maps existing calls that need to know about file size to their
64 bit (``large'') versions
automatically, resulting in more portable code than explicit calls
to the 64 bit versions.
Second, explicit calls can be made to new 64 bit versions of system calls,
and library routines that use the new data types and data structures.
These are transitional interfaces, however, and may not be available in
future releases.
If you choose to use these new interfaces, we recommend that you
pre-define their corresponding feature test macros to ensure that
the associated functions are declared.
Defining the macro _LARGEFILE_SOURCE
to be 1 before including any header, enables the
functionality defined for _LFS_LARGEFILE.
Defining the macro _LARGEFILE64_SOURCE before including
any header, enables the functionality defined for _LFS64_LARGEFILE,
LFS64_ASYNCHRONOUS_IO, and LFS64_STDIO.
These symbolic constants are discussed on
unistd(F).
System calls and library routines that map to 32 and 64 bit versions
A subset of the system-provided data types and structures are mapped to 64 bit
versions and uses of the existing data types are automatically
switched to the 64 bit versions if you have defined the
_FILE_OFFSET_BITS macro to be 64
at compile time.
This table lists those data type names and data structure names.
They are defined in several existing headers, and include the following:
Header
Existing data type names
Data type used when mapped
using _FILE_OFFSET_BITS=64
Many, but not all, common system utilities have been updated to work
with large files.
The most notable omissions are the shells, which have not been
updated to be large file aware.
What this means is that while the utilities listed below are large
file aware and can, for example, create large files -- you cannot
create large files using shell redirection, even on a large-file-aware
file system.
The following set of system utilities have been made large-file aware,
and a notice stating that has been added to each manual page.
Historically,
the operating system has used the concept of ``process'' to serve
two roles:
A process was the (schedulable) executor of user programs
with a single ``thread of control''.
A process was a requisitioner and user of system resources
(for example, address spaces, open files).
In the current version of the operating system:
There are extensive facilities for the writing of
``multithreaded'' user programs
available via the Threads Library (see
thread(THREAD)):
Several ``related'' or ``sibling'' threads (at least one)
can be associated with a single process.
Each thread has common access to most of the resources of the process.
Each thread has an independent path of computation.
The threads of a process can communicate and coordinate
with each other and, if used wisely, can solve some problems more
effectively than a single threaded process.
If the underlying hardware supports multiple processors,
there can be more than one thread running truly concurrently.
In support of the Threads Library,
the primary execution entity managed by the operating system kernel
has become the ``lightweight process'' (LWP).
A thread must be ``bound'' to an LWP before the thread can be executed.
In general, a given LWP can be switched among more than one thread.
The notion of ``process'' is largely retained
for the management of most resources.
Some of the features that had been maintained per process
are now maintained independently per thread.
These fall primarily in the areas of scheduling and signal handling.
Some features that appear to be maintained per thread
(for example, timers)
are actually maintained internally (by the kernel) per LWP.
The per-thread functionality is achieved by ``wrapper'' functions
in the Threads Library for certain system calls.
Consequently, applications that attempt to use LWP features directly
(without the Threads Library) will experience different semantics.
Direct use of LWPs should be avoided.
These are not a standard interface and
are subject to change in future releases of the operating system.
Use the threads interfaces instead.
An application that does not use the Threads Library
(and does not use LWP features directly)
will run in a process with a single thread of control.
System call semantics are compatible to those
in earlier releases of the operating system.
The manual pages often use the term
``process'' as the execution entity.
For example:
The system call failed because the calling process lacked
the required privilege.
A more complete description would be:
The system call failed because the calling thread was bound to an LWP
whose containing process lacked the required privilege.
Nevertheless, for convenience of language,
the manual pages will
not dwell on the hierarchy of execution entities.
Thus, the expression ``calling process'' will continue to be used.
Threads will be mentioned only if technically relevant.
Similarly, LWPs will be mentioned only if there is some significant
difference between the behavior of a thread and the behavior of
the underlying LWP.
These distinctions will usually be found in sections titled
``Considerations for Threads Programming'' and
``Considerations for Lightweight Processes,''
respectively.
An overview of changes in system call usage
for multithreaded applications
is given in the following table.
Category
Considerations
Access to resources
Access rights (for example, file permissions, resource limits)
are a property of the containing process and are shared by sibling threads.
Accounting/Profiling
Statistics are gathered at the process level
and represent the combined usage of all contained threads.
Address Space
Related threads share (by definition) the same address space;
modifications to the address space by one thread can be perceived
by sibling threads.
Blocked System Calls
While one thread is blocked, sibling threads might still be executing.
Security Attributes
Security attributes (for example, privileges)
are an attribute of the containing process
and are shared by sibling threads.
File Descriptors
Open file descriptors are a process resource
and available to any sibling thread;
if used concurrently, actions by one thread can interfere with
those of a sibling.
File System Control/Status
File system operations (for example, creating/removing files, mounting/umounting
file systems) are essentially the same for threads programming as for
programming with processes.
Identity
Process identification numbers
(for example, process ID, group ID) are defined at the
process level and are shared by sibling threads.
IPC
Access to IPC resources is controlled at the process level;
resources made available by one thread are available to siblings.
Signals
Signal disposition (that is, default/ignore/catch) is defined
at the process level and is shared by related threads; however,
caught signals are delivered to only one thread for handling.
Signal masks are maintained per thread.
Alternate signal stacks are not part of the Threads Library
but are supported by LWPs.
Scheduling
Scheduling class and attributes are maintained per thread.
System Control/Status
General system operations (for example, get/set time, get/set system name)
remain the same for programming with threads as for programming with
processes.
Definitions
Access permissions
Access checking is performed whenever a subject (a process)
tries to access an object (such as a file, directory, or IPC
resource).
Permission to access an object is granted or denied on the basis of
mode bits and Access Control Lists (ACLs).
Mode bits and ACLs are collectively known as
Discretionary Access Control (DAC).
ACLs are supported on each filesystem object on Secure File
System (sfs) or Veritas File System (vxfs) file system types,
and on Inter-Process Communication (IPC) objects.
See
acl(S)
and
aclipc(S).
Access checking is performed by the system in the following order:
DAC checks are performed on every component of the
pathname, including the object itself.
The standard file access permission bits and ACL (if
present) of each component are checked to determine if the process
requesting access to the object has the required permissions
(read, write, and/or execute/search).
Each access mode requested is checked separately using the following
algorithm:
If the effective user ID of the process is equal to the user
ID of the owner of the file, and the requested access mode
bit is set in the ``owner'' bits of the mode, access is granted;
otherwise access checking continues.
If the effective user ID of the process matches the user
ID in a ``user'' ACL entry, and the requested
access mode bit is set in the ``group''
bits of the mode and in the matching ACL entry, access is granted;
otherwise, or if no ACL is present, access checking continues.
If the effective group ID (or any of the supplementary
group IDs of the process) matches the
group ID specified in any ``group''
ACL entry (or the owning group of the file,
if no ACL is present),
and the requested access mode bit is set in the
``group'' bits of the
mode and in the matching ACL entry (if present),
access is granted; otherwise, access checking continues.
If the above checks fail, and the requested access mode bit is set
in the ``other'' bits of the mode, access is granted; otherwise,
access is denied (EACCES is returned).
(Note that if a process's user ID matches a ``user''
ACL entry and the group ID matches one or more ``group''
entries, only the ``user'' ACL entry is used to determine access.)
If DAC access checks fail to give the requested access
to the calling process,
the privileges of the calling process are examined to determine
if the calling process has the privilege required to grant
the access requested
(P_DACREAD for DAC read and execute/search access,
P_DACWRITE for DAC write access).
Background process group
Any process group that is not the foreground process group
of a session that has established a connection
with a controlling terminal.
Controlling process
A session leader that established a connection to a controlling terminal.
Controlling terminal
A terminal that is associated with a session.
Each session may have, at most,
one controlling terminal associated with it and a controlling terminal may
be associated with only one session.
Certain input sequences from the
controlling terminal cause signals to be sent to process groups in the session
associated with the controlling terminal; see
termio(HW).
Directory
Directories organize files into a hierarchical system where
directories are the nodes in the hierarchy.
A directory is a file that catalogues the list of files, including
directories (sub-directories), that are directly beneath it in the hierarchy.
Entries in a directory file are called links.
A link associates a file identifier with a filename.
By convention, a directory contains at least two links, . (dot)
and .. (dot-dot).
The link called dot refers to the directory itself while dot-dot
refers to its parent directory.
The root directory, which is the top-most node of the hierarchy, has itself
as its parent directory.
The pathname of the root directory is / and the
parent directory of the root directory is /.
Downstream
In a stream, the direction from stream head to driver.
Driver
In a stream, the driver provides the interface between
peripheral hardware and the stream.
A driver can also be a pseudo-driver,
such as a multiplexor or log driver (see
log(HW)),
which is not associated with a hardware device.
Effective user ID and effective group ID
An active process has an effective user
ID
and an effective group
ID
that are used to determine file access permissions (see below).
The effective user
ID
and effective group
ID
are equal to the process's real user
ID
and real group
ID
respectively, unless the process
or one of its ancestors evolved from a file that had the set-user-ID
bit or set-group
ID
bit set
(see
exec(S)).
File descriptor
A file descriptor is a small integer used to do I/O on a file.
The value of a file descriptor is from 0 to (NOFILES-1).
A process may have no more than NOFILES file descriptors
open simultaneously.
See
getrlimit(S).
A file descriptor is returned by system calls such as open,
or pipe.
The file descriptor is used as an argument by calls such as
read, write, ioctl, and close.
File name
Names consisting of 1 to NAME_MAX characters may be used to name
an ordinary file, special file or directory.
These characters may be selected from the set of all character values
excluding \0 (null) and the
ASCII
code for / (slash).
Note that it is generally unwise to use *, ?, [,
or ] as part of file names because of the special meaning attached
to these characters by the shell (see
sh(C)).
Although permitted, the use of unprintable
characters in file names should be avoided.
A file name is sometimes referred to as a pathname component.
The interpretation of a pathname component is dependent on the values of
NAME_MAX and _POSIX_NO_TRUNC associated with the path prefix of that
component.
If any pathname component is longer than NAME_MAX and
_POSIX_NO_TRUNC is in effect for the path prefix of that component
(see
fpathconf(S)
and
limits(F)),
it shall be considered an
error condition in that implementation.
Otherwise, the implementation shall use the first
NAME_MAX bytes of the pathname component.
Foreground process group
Each session that has established a connection with a controlling terminal
will distinguish one process group of the session as the foreground process group
of the controlling terminal.
This group has certain privileges when accessing
its controlling terminal that are denied to background process groups.
The schedulable execution entity that is managed by the kernel.
There will be at least one lightweight process (LWP) for each process,
possibly many.
The set of lightweight processes
(called ``related LWPs'' or ``sibling LWPs'')
associated with a given process
(sometimes called the ``containing process'')
share all process attributes except
scheduling context, signal mask, alternate signal stack, and
high resolution timers.
Message
In a stream, one or more blocks of data or information, with associated
STREAMS control structures.
Messages can be of several defined types, which identify
the message contents.
Messages are the only means of transferring data and communicating within a stream.
Message queue
In a stream, a linked list of messages awaiting processing by
a module or driver.
Message queue identifier
A message queue identifier (msqid) is a unique positive integer
created by a msgget system call.
Each msqid has a message queue and a data structure associated with it.
The data structure is referred to as msqid_ds
and contains the following members:
Here are descriptions of the fields of the msqid_ds structure:
msg_perm
is an ipc_perm structure that
specifies the message operation permission (see below).
This structure includes the following members:
uid_t cuid; /* creator user id */
gid_t cgid; /* creator group id */
uid_t uid; /* user id */
gid_t gid; /* group id */
mode_t mode; /* r/w permission */
ushort seq; /* slot usage sequence # */
key_t key; /* key */
*msg_first
is a pointer to the first message on the queue.
*msg_last
is a pointer to the last message on the queue.
msg_cbytes
is the current number of bytes on the queue.
msg_qnum
is the number of messages currently on the queue.
msg_qbytes
is the maximum number of bytes allowed on the queue.
msg_lspid
is the process
ID
of the last process that performed a
msgsnd
operation.
msg_lrpid
is the process
ID
of the last process that performed a
msgrcv
operation.
msg_stime
is the time of the last
msgsnd
operation.
msg_rtime
is the time of the last
msgrcv
operation
msg_ctime
is the time of the last
msgctl
operation that changed a member of the above structure.
Message operation permissions
In the msgop and msgctl
system call descriptions, the permission required
for an operation is given as {token}, where token is the type
of permission needed, interpreted as follows:
00400
READ by user
00200
WRITE by user
00040
READ by group
00020
WRITE by group
00004
READ by others
00002
WRITE by others
Read and write permissions on a msqid are
granted to a process if one or more of the following are true:
The calling process has the P_OWNER privilege.
The effective user
ID
of the process matches msg_perm.cuid
or msg_perm.uid in the data structure associated with
msqid and the appropriate bit of the ``user'' portion (0600) of
msg_perm.mode is set.
The effective group
ID
of the process matches msg_perm.cgid or msg_perm.gid
and the appropriate bit of the ``group'' portion (060) of
msg_perm.mode is set.
The appropriate bit of the ``other'' portion (006) of
msg_perm.mode is set.
Otherwise, the corresponding permissions are denied.
Module
A module is an entity containing processing
routines for input and output data.
It always exists in the middle of a
stream, between the stream's head and a driver.
A module is the STREAMS counterpart to the commands
in a shell pipeline except that a module contains a pair
of functions which allow independent bidirectional (downstream
and upstream) data flow and processing.
Multiplexor
A multiplexor is a driver that allows streams
associated with several user processes to be
connected to a single driver, or
several drivers to be connected to a single
user process.
STREAMS does not provide a general multiplexing
driver, but does provide the facilities for
constructing them and for connecting multiplexed
configurations of streams.
Orphaned process group
A process group in which the parent of every member in the group is
either itself a member of the group, or is not a member of the process
group's session.
Pathname
A pathname is a null-terminated character string
starting with an optional slash (/),
followed by
zero or more directory names separated by slashes, optionally followed
by a filename.
If a pathname begins with a slash, the path search begins at the
root directory.
Otherwise, the search begins from the current working directory.
A slash by itself names the root directory.
Unless specifically stated otherwise,
the null pathname is treated as if it named
a non-existent file.
Process ID
Each process in the system is uniquely identified during its lifetime by
a positive integer called a process
ID.
A process
ID
may not be reused
by the system until the process lifetime, process group lifetime and
session lifetime ends for any process ID, process group
ID
and session
ID
equal to that process ID.
Parent process ID
A new process is created by a currently active
process (see
fork(S)).
The parent process
ID
of a process is the process
ID
of its creator.
Privilege
Having appropriate privilege means having the capability to perform
sensitive system operations (see
procpriv(S))
or having the ability to override system restrictions.
Process group
Each process in the system is a member of a process group that is
identified by a process group ID.
Any process that is not a process
group leader may create a new process group and become its leader.
Any process that is not a process group leader may join an existing
process group that shares the same session as the process.
A newly created process joins the process group of its parent.
Process group leader
A process group leader is a process whose process
ID
is the same as
its process group ID.
Process group ID
Each active process is a member of a process group and is
identified by a positive integer called the process group ID.
This
ID
is the process
ID
of the group leader.
This grouping permits the signaling of related processes (see
kill(S)).
Process lifetime
A process lifetime begins when the process is forked and ends after it
exits, when its termination has been acknowledged by its parent process.
See
wait(S).
Process group lifetime
A process group lifetime begins when the process group is created by its
process group leader, and ends when the lifetime of the last process in the group
ends or when the last process in the group leaves the group.
Read queue
In a stream, the message queue in a module or driver containing messages
moving upstream.
Real user ID and real group ID
Each user allowed on the system is
identified by a positive integer (0 to UID_MAX) called a real user
ID.
Each user is also a member of a group.
The group is identified by a positive integer called the real group
ID.
An active process has a real user
ID
and real group
ID
that are set to the real user
ID
and real group
ID,
respectively, of the user responsible for the creation of the process.
Root directory and current working directory
Each process has associated with it a concept of a root directory and
a current working directory for the purpose of resolving pathname searches.
The root directory of a process need not be the root directory of the root
file system.
Saved user ID and saved group ID
The saved user
ID
and saved group
ID
are the values of the effective user
ID
and effective group
ID
prior to an exec of a file (see
exec(S)).
Semaphore identifier
A semaphore identifier (semid) is a unique positive integer created by a
semget system call.
Each semid has a set of semaphores and a data structure associated with it.
The data structure is referred to as semid_ds
and contains the following members:
struct ipc_perm sem_perm; /* operation permission struct */
struct sem *sem_base; /* ptr to first semaphore in set */
ushort sem_nsems; /* number of sems in set */
time_t sem_otime; /* last operation time */
time_t sem_ctime; /* last change time */
/* Times measured in secs since */
/* 00:00:00 GMT, Jan. 1, 1970 */
Here are descriptions of the fields of the semid_ds structure:
sem_perm
is an ipc_perm structure that
specifies the semaphore operation permission (see below).
This structure includes the following members:
uid_t uid; /* user id */
gid_t gid; /* group id */
uid_t cuid; /* creator user id */
gid_t cgid; /* creator group id */
mode_t mode; /* r/a permission */
ushort seq; /* slot usage sequence number */
key_t key; /* key */
sem_nsems
is equal to the number of semaphores in the set.
Each semaphore in the set is referenced by a nonnegative integer
referred to as a
sem_num.
sem_num
values run sequentially from 0 to the value of sem_nsems minus 1.
sem_otime
is the time of the last semop operation.
sem_ctime
is the time of the last semctl
operation that changed a member of the above structure.
A semaphore is a data structure called sem
that contains the following members:
Here are descriptions of the fields of the sem structure:
semval
is a non-negative integer that is the actual value of the semaphore.
sempid
is equal to the process
ID
of the last process that performed a semaphore operation on this semaphore.
semncnt
is a count of the number of processes that are currently suspended
awaiting this semaphore's semval to become greater than its current value.
semzcnt
is a count of the number of processes that are currently suspended
awaiting this semaphore's semval to become 0.
Semaphore operation permissions
In the semop and semctl
system call descriptions, the permission required
for an operation is given as {token}, where token is the type
of permission needed interpreted as follows:
00400
READ by user
00200
ALTER by user
00040
READ by group
00020
ALTER by group
00004
READ by others
00002
ALTER by others
Read and alter permissions on a semid are
granted to a process if one or more of the following are true:
The calling process has the P_OWNER privilege.
The effective user
ID
of the process matches sem_perm.cuid
or sem_perm.uid in the data structure associated with
semid and the appropriate bit of the
``user'' portion (0600) of
sem_perm.mode is set.
The effective group
ID
of the process matches sem_perm.cgid or
sem_perm.gid and the appropriate bit of the ``group'' portion
(060) of sem_perm.mode is set.
The appropriate bit of the ``other'' portion (06) of
sem_perm.mode is set.
Otherwise, the corresponding permissions are denied.
Session
A session is a group of processes identified by a common
ID
called a session
ID, capable of establishing a connection with a controlling terminal.
Any process that is not a process group leader may create a new session
and process group, becoming the session leader of the session and process
group leader of the process group.
A newly created process joins the session of its creator.
Session ID
Each session in the system is uniquely identified during its lifetime by
a positive integer called a session ID, the process
ID
of its session leader.
Session leader
A session leader is a process whose session
ID
is the same as its process and process group ID.
Session lifetime
A session lifetime begins when the session is created by its session
leader, and ends when the lifetime of the last process that is a member
of the session ends, or when the last process that is a member in the
session leaves the session.
Shared memory identifier
A shared memory identifier (shmid)
is a unique positive integer created by a
shmget system call.
Each shmid has a segment of memory
(referred to as a shared memory segment)
and a data structure associated with it.
(Note that these shared memory segments
must be explicitly removed by the user
after the last reference to them is removed.)
The data structure is referred to as shmid_ds
and contains the following members:
struct ipc_perm shm_perm; /* operation permission struct */
int shm_segsz; /* size of segment */
struct region *shm_amp; /* ptr to region structure */
char pad[4]; /* for swap compatibility */
pid_t shm_lpid; /* pid of last operation */
pid_t shm_cpid; /* creator pid */
ushort shm_nattch; /* number of current attaches */
ushort shm_cnattch; /* used only for shminfo */
time_t shm_atime; /* last attach time */
time_t shm_dtime; /* last detach time */
time_t shm_ctime; /* last change time */
/* Times measured in secs since */
/* 00:00:00 GMT, Jan. 1, 1970 */
Here are descriptions of the fields of the
shmid_ds structure:
shm_perm
is an ipc_perm structure that
specifies the shared memory operation permission (see below).
This structure includes the following members:
uid_t cuid; /* creator user id */
gid_t cgid; /* creator group id */
uid_t uid; /* user id */
gid_t gid; /* group id */
mode_t mode; /* r/w permission */
ushort seq; /* slot usage sequence # */
key_t key; /* key */
shm_segsz
specifies the size of the shared memory segment in bytes.
shm_cpid
is the process
ID
of the process that created the shared memory identifier.
shm_lpid
is the process
ID
of the last process that performed a
shmop operation.
shm_nattch
is the number of processes that currently have this segment attached.
shm_atime
is the time of the last shmat operation
(see
shmop(S)).
shm_dtime
is the time of the last shmdt operation
(see
shmop(S)).
shm_ctime
is the time of the last shmctl
operation that changed one of the members of the above structure.
Shared memory operation permissions
In the shmop and shmctl
system call descriptions, the permission required
for an operation is given as {token}, where token is the type
of permission needed interpreted as follows:
00400
READ by user
00200
WRITE by user
00040
READ by group
00020
WRITE by group
00004
READ by others
00002
WRITE by others
Read and write permissions on a shmid are
granted to a process if one or more of the following are true:
The calling process has the P_OWNER privilege.
The effective user
ID
of the process matches shm_perm.cuid or
shm_perm.uid in the data structure associated with
shmid and the appropriate bit of the
``user'' portion (0600) of shm_perm.mode is set.
The effective group
ID
of the process matches shm_perm.cgid or
shm_perm.gid and the appropriate bit of the ``group'' portion
(060) of shm_perm.mode is set.
The appropriate bit of the ``other'' portion (06) of
shm_perm.mode is set.
Otherwise, the corresponding permissions are denied.
Special processes
The process with
ID
0 and the process with
ID
1 are special processes referred to as
proc0 and proc1; see
kill(S).
proc0 is the process scheduler.
proc1 is the initialization process
(init); proc1 is the ancestor of every other
process in the system and is used to control the process structure.
STREAMS
A set of kernel mechanisms that support the development of
network services and data communication drivers.
It defines interface standards for character input/output
within the kernel and between the kernel and user level processes.
The STREAMS mechanism is composed of utility routines,
kernel facilities and a set of data structures.
Stream
A stream is a full-duplex data path within the kernel
between a user process and driver routines.
The primary components are a stream head,
a driver and zero or more modules
between the stream head and driver.
A stream is analogous to a shell pipeline except that
data flow and processing are bidirectional.
Stream head
In a stream, the stream head is the end of the stream
that provides the interface between the stream and a user process.
The principal functions of the stream head are processing
STREAMS-related system calls,
and passing data and information between a user
process and the stream.
Superuser and privilege
By default, the system is runs with the Super User Module (SUM) installed as
the privilege module.
This modules privilege policy allows a process to be recognized as a
superuser process that is granted all the privileges listed in the
``Privileges''
section above, if its effective user ID is ``0''.
Such a process has unrestricted access to the system.
In addition, because the system supports the discrete privileges
defined in the
``Privileges''
section,
a user can acquire a subset of the recognized privileges through the
Trusted Facilities Management (TFM) database.
For a discussion of privileges see
``Privileges'' in Programming with system calls and libraries.
Threads
A sequential series of instructions that can be logically executed concurrently
with other such sequences within a single process.
Also called ``thread-of-control''.
The Threads Library provides interfaces for the
creation, management, and removal of threads.
On some architectures,
the Threads Library can use lightweight processes (LWPs) to have
different threads of a process executed simultaneously.
Upstream
In a stream, the direction from driver to stream head.
Write queue
In a stream, the message queue in a module or
driver containing messages moving downstream.