1.0.X version of the Linux unit - for compatibility only. This document describes the LINUX unit for Free Pascal. The unit was written by Michael van Canneyt. It works only on the Linux/BSD operating systems. Pointer to record. Record containing one entry in the result of Filename matching the search criteria Next element in result of Glob call. Record describing floating point register in signal handler. Decimal part of floating point value Exponent of floating point value Pointer to record. Record describing floating point unit in signal handler. Pointer to record Record describing the context of the program when it receives a signal The above records contain information about the processor state and process state at the moment a signal is sent to your program. Low part of gs register High part of gs register Low part of FS register High part of FS register Low part of ES register High part of ES register Low part of DS register High part of DS register EDI register ESI register EBP register ESP register EBX register EDX register ECX register EAX register Interrupt number Error register (?) IP (Instruction Pointer) register Low part of CS register High part of CS register Flags register ESP register at signal. Low part of SS register High part of SS register Floating Point unit state Old signal mask ? Function prototype for call. Function prototype for the call. Pointer to type. Signal restorer function prototype Pointer to type Pointer to record. Record used in call. Funcion called when signal is triggered. Signal mask. Flags for SigAction Obsolete, don't use Record describing an inode (file) in the call. Device number Pad byte. Do not use Inode number of file File mode Number of links to file. File owner UID File owner GID Pad byte. Do not use File size Block size Number of blocks used Last access time Pad byte. Do not use Last modification time. Pad byte. Do not use Creation time Pad byte. Do not use Pad byte. Do not use Pad byte. Do not use Record describing a file system in the call. File system type Block size Total number of blocks Number of free blocks Number of available blocks Number of files ? ? Max name length for files. Pad bytes. Do not use. Pointer to record Record used in and calls File descriptor. Do not use. Location in directory listing File size Pointer to records ? ? ? Pointer to record. Record used in the function to return files in a directory. Inode number of file Offset in directory. Record length Name of file Number of control characters in record. Number of control characters in record. Terminal I/O description record (small) input mode flags output mode flags control mode flags local mode flags line discipline control characters Terminal I/O description record input mode flags output mode flags control mode flags local mode flags line discipline control characters Record used in to set file access and modificaton times. Access time Modification time Pointer to array. Array containing file descriptor bitmask for the call. Record specifying a time inteval. Number of seconds Number of milliseconds Pointer to record Time interval for the function. Number of seconds Number of nanoseconds Record used to return kernel information in function. The elements of this record are null-terminated C style strings, you cannot access them directly. System name Computer name Release number Version number Machine type Domain name Last operating system error Linuxerror is the variable in which the procedures in the linux unit report errors. Standard error output (deprecated)

StdErr Is a Text variable, corresponding to Standard Error or diagnostic output. It is connected to file descriptor 2. It can be freely used, and will be closed on exit.

This variable is deprecated. Please use the corresponding variable in the system unit for disgnostic messages.

Get/Set process priority Get/set process group priority Get/set user priority call test: read allowed call test: write allowed call test: execute allowed call test: file exists. Sigaction options: Do not receive notification when child processes stop Sigaction options: ? Sigaction options: Call the signal handler on an alternate signal stack. Sigaction options: Provide behaviour compatible with BSD signal semantics Sigaction options: ? Sigaction options: Do not prevent the signal from being received when it is handled. Sigaction options: Restore the signal action to the default state. Sigprocmask flags: Add signals to the set of blocked signals. Sigprocmask flags: Remove signals from the set set of blocked signals. Sigprocmask flags: Set of blocked signals is given. Signal handler: Default signal handler Signal handler: Ignore signal Signal handler: error Signal: HUP (Hangup) Signal: INT (Interrupt) Signal: QUIT Signal: ILL (Illegal instruction) Signal: TRAP (Trace trap) Signal: ABRT (Abort) Signal: IOT (IOT trap) Signal: BUS (bus error) Signal: FPE (Floating point error) Signal: KILL (unblockable) Signal: USR1 (User-defined signal 1) Signal: SEGV (Segmentation violation) Signal: USR2 (User-defined signal 2) Signal: PIPE (Broken pipe Signal: ALRM (Alarm clock) Signal: TERM (Terminate) Signal: STKFLT (Stack Fault) Signal: CHLD (child status changed) Signal: CONT (Continue) Signal: STOP (Stop, unblockable) Signal: TSTP (keyboard stop) Signal: TTIN (Terminal input, background) Signal: TTOU (Terminal output, background) Signal: URG (Socket urgent condition) Signal: XCPU (CPU limit exceeded) Signal: XFSZ (File size limit exceeded) Signal: VTALRM (Virtual alarm clock) Signal: PROF (Profiling alarm) Signal: WINCH (Window/Terminal size change) Signal: IO (I/O operation possible) Signal: POLL (Pollable event) Signal: PWR (power failure restart) Signal: Unused command: Get close-on-exec flag command: Set close-on-exec flag command: Get filedescriptor flags command: Set filedescriptor flags command: Get lock command: Set lock command: Test lock command: get owner of filedescriptor events command: Set owner of filedescriptor events File ( record) mode: File type bit mask File ( record) mode: Socket File ( record) mode: Link File ( record) mode: Regular file File ( record) mode: Block device File ( record) mode: Directory File ( record) mode: Character device File ( record) mode: FIFO File ( record) mode: UID bit set File ( record) mode: GID bit set File ( record) mode: Sticky bit set File ( record) mode: Other permission bits mask File ( record) mode: Other read permission File ( record) mode: Other write permission File ( record) mode: Others execute permission File ( record) mode: Group permission bits mask File ( record) mode: Group read permission File ( record) mode: Group write permission File ( record) mode: Others execute permission File ( record) mode: Owner permission bits mask File ( record) mode: Owner read permission File ( record) mode: Owner write permission File ( record) mode: Others execute permission File system type (): (ext2) Old second extended File system type (): (ext2) Second extended File system type (): (ext) Extended File system type (): ISO 9660 File system type (): Minix File system type (): Minix 3.0 File system type (): Minix V2 File system type (): MSDOS (FAT) File system type (): NFS File system type (): PROC fs File system type (): XIA Shared lock Exclusive lock unlock Non-blocking operation memory access: page can be read memory access: page can be written memory access: page can be executed memory access: page can not be accessed map type: Share changes map type: Changes are private map type: Bitmask for type of mapping map type: Interpret addr exactly map type: Don't use a file File Input/Output routines

Functions for handling file input/output.

NameDescription
Duplicate a file handle
Copy one file handle to another
General file control
Close file descriptor
Flush file descriptor
Open new file descriptor
Read from file descriptor
Position in file
Truncate file
Write to file descriptor
Get file descriptor of pascal file
Wait for input from file descriptor
Wait for input from pascal file
 General File handling routines

Functions for handling files on disk.

NameDescription
Check access rights on file
Return name part of file
Change owner of file
Change access rights on file
Return directory part of file
Split filename in parts
Return full-grown filename
Set lock on a file
Match filename to searchpattern
Search for a file in a path
Return filesystem information
Return file information
Rename file
Return information on a link
Create a link
Read contents of a symbolic link
Create a symbolic link
Set the file creation mask
Remove a file
Change file timestamps
 Pipes, FIFOs and streams

Functions for creating and managing pipes.

NameDescription
Create a pipe
Create pipes to program's input and output
Make a fifo
Close a pipe
Open a pipe for to program's input or output
 Directory handling routines

Functions for reading and searching directories.

NameDescription
Close directory handle
Return files matching a search expression
Free result of Glob
Open directory for reading
Read directory entry
Seek directory
Seek directory
 Process handling

Functions for managing processes and programs.

NameDescription
Create a thread
Execute process with command-line list
Execute process with command-line list and environment
Search in path and execute process with command list
Execute process
Execute process with environment
Search in path and execute process
Spawn child process
Get effective group id
Get environment variable
Get effective user id
Get group id
Get process id
Get parent process id
Get process priority
Get user id
Change priority of process
Change priority of process
Execute shell command
Wait for child process to terminate
 Signals

Functions for managing and responding to signals.

NameDescription
Send alarm signal to self
Send arbitrary signal to process
Wait for signal to arrive
Set signal action
Set signal action
See if signals are waiting
Set signal processing mask
Send signal to self
Sets signal mask and waits for signal
Waits for a specific amount of time
 System information

Functions for retrieving system information such as date and time.

NameDescription
Return system date
Return system date and time
Return system domain name
Return epoch time
Return system host name
Return system timezone
Return system time
Return system time
Return name of timezone file
Read timezone file contents
Return general system information
Return system information
 Terminal functions

Functions for controlling the terminal to which the process is connected.

NameDescription
Set terminal to raw mode
Set terminal reading speed
Set terminal writing speed
General IO control call
See if filedescriptor is a terminal
Wait till all output was written
Suspend transmission or receipt of data
Discard data written to terminal
Get terminal attributes
Return PID of foreground process
Send data for specific time
Set terminal attributes
Set foreground process
Name of tty file
 Utility routines

Auxiliary functions that are useful in connection with the other functions.

NameDescription
Create an array of pchars from string
Convert epoch time to local time
Clear item of select filedescriptors
Check item of select filedescriptors
Set item of select filedescriptors
Clear all items in select filedecriptors
Convert local time to epoch time
Map a file into memory
Unmap previously mapped memory file
Convert octal to digital
Check file mode for block device
Check file mode for character device
Check file mode for directory
Check file mode for FIFO
Check file mode for symboloc link
Check file mode for regular file
Check file mode for socket
Create an array of pchars from string
 Check file access

Access tests user's access rights on the specified file. Mode is a mask existing of one or more of the following:

R_OK
User has read rights.
W_OK
User has write rights.
X_OK
User has execute rights.
F_OK
File exists.

The test is done with the real user ID, instead of the effective user ID. If access is denied, or an error occurred, False is returned.

LinuxError is used to report errors:

sys_eaccess
The requested access is denied, either to the file or one of the directories in its path.
sys_einval
Mode was incorrect.
sys_enoent
A directory component in Path doesn't exist or is a dangling symbolic link.
sys_enotdir
A directory component in Path is not a directory.
sys_enomem
Insufficient kernel memory.
sys_eloop
Path has a circular symbolic link.
 Schedule an alarm signal to be delivered

Alarm schedules an alarm signal to be delivered to your process in Sec seconds. When Sec seconds have elapsed, Linux will send a SIGALRM signal to the current process. If Sec is zero, then no new alarm will be set. Whatever the value of Sec, any previous alarm is cancelled.

The function returns the number of seconds till the previously scheduled alarm was due to be delivered, or zero if there was none.

Create a set of pipe file handlers

AssignePipe creates a pipe, i.e. two file objects, one for input, one for output. What is written to Pipe_out, can be read from Pipe_in.

This call is overloaded. The in and out pipe can take three forms: an typed or untyped file, a text file or a file descriptor.

If a text file is passed then reading and writing from/to the pipe can be done through the usual Readln(Pipe_in,...) and Writeln(Pipe_out,...) procedures.

The function returns True if everything went succesfully, False otherwise.

In case the function fails and returns False, LinuxError is used to report errors:

sys_emfile
Too many file descriptors for this process.
sys_enfile
The system file table is full.
 Assign stream for in and output to a program

AssignStream creates a 2 or 3 pipes, i.e. two (or three) file objects, one for input, one for output,(and one for standard error) the other ends of these pipes are connected to standard input and output (and standard error) of Prog. Prog is the name of a program (including path) with options, which will be executed.

What is written to StreamOut, will go to the standard input of Prog. Whatever is written by Prog to it's standard output can be read from StreamIn. Whatever is written by Prog to it's standard error read from StreamErr, if present.

Reading and writing happens through the usual Readln(StreamIn,...) and Writeln (StreamOut,...) procedures.

You should not use Reset or Rewrite on a file opened with POpen. This will close the file before re-opening it again, thereby closing the connection with the program.

The function returns the process ID of the spawned process, or -1 in case of error.

In case of error (return value -1) LinuxError is used to report errors:

sys_emfile
Too many file descriptors for this process.
sys_enfile
The system file table is full.

Other errors include the ones by the fork and exec programs

Return basename of a file

Returns the filename part of Path, stripping off Suf if it exists. The filename part is the whole name if Path contains no slash, or the part of Path after the last slash. The last character of the result is not a slash, unless the directory is the root directory.

None. Sets flags in record.

CFMakeRaw sets the flags in the Termios structure Tios to a state so that the terminal will function in Raw Mode.

For an example, see .

None. Set input baud rate in record CFSetISpeed Sets the input baudrate in the TermIOS structure Tios to Speed. None. Set output baud rate in record CFSetOSpeed Sets the output baudrate in the Termios structure Tios to Speed. None. Change owner of file Chown sets the User ID and Group ID of the file in Path to NewUid, NewGid The function returns True if the call was succesfull, False if the call failed.

Errors are returned in LinuxError.

sys_eperm
The effective UID doesn't match the ownership of the file, and is not zero. Owner or group were not specified correctly.
sys_eaccess
One of the directories in Path has no search (=execute) permission.
sys_enoent
A directory entry in Path does not exist or is a symbolic link pointing to a non-existent directory.
sys_enomem
Insufficient kernel memory.
sys_erofs
The file is on a read-only filesystem.
sys_eloop
Path has a reference to a circular symbolic link, i.e. a symbolic link, whose expansion points to itself.
 Change file permission bits

Chmod Sets the Mode bits of the file in Path to NewMode. Newmode can be specified by 'or'-ing the following:

S_ISUID
Set user ID on execution.
S_ISGID
Set Group ID on execution.
S_ISVTX
Set sticky bit.
S_IRUSR
Read by owner.
S_IWUSR
Write by owner.
S_IXUSR
Execute by owner.
S_IRGRP
Read by group.
S_IWGRP
Write by group.
S_IXGRP
Execute by group.
S_IROTH
Read by others.
S_IWOTH
Write by others.
S_IXOTH
Execute by others.
S_IRWXO
Read, write, execute by others.
S_IRWXG
Read, write, execute by groups.
S_IRWXU
Read, write, execute by user.

Errors are returned in LinuxError.

sys_eperm
The effective UID doesn't match the ownership of the file, and is not zero. Owner or group were not specified correctly.
sys_eaccess
One of the directories in Path has no search (=execute) permission.
sys_enoent
A directory entry in Path does not exist or is a symbolic link pointing to a non-existent directory.
sys_enomem
Insufficient kernel memory.
sys_erofs
The file is on a read-only filesystem.
sys_eloop
Path has a reference to a circular symbolic link, i.e. a symbolic link, whose expansion points to itself.
 Clone current process (create new thread)

Clone creates a child process which is a copy of the parent process, just like does. In difference with Fork, however, the child process shares some parts of it's execution context with its parent, so it is suitable for the implementation of threads: many instances of a program that share the same memory.

When the child process is created, it starts executing the function Func, and passes it Args. The return value of Func is either the explicit return value of the function, or the exit code of the child process.

The sp pointer points to the memory reserved as stack space for the child process. This address should be the top of the memory block to be used as stack.

The Flags determine the behaviour of the Clone call. The low byte of the Flags contains the number of the signal that will be sent to the parent when the child dies. This may be bitwise OR'ed with the following constants:

CLONE_VM
Parent and child share the same memory space, including memory (un)mapped with subsequent mmap calls.
CLONE_FS
Parent and child have the same view of the filesystem; the chroot, chdir and umask calls affect both processes.
CLONE_FILES
the file descriptor table of parent and child is shared.
CLONE_SIGHAND
the parent and child share the same table of signal handlers. The signal masks are different, though.
CLONE_PID
PArent and child have the same process ID.

Clone returns the process ID in the parent process, and -1 if an error occurred.

On error, -1 is returned to the parent, and no child is created.

sys_eagain
Too many processes are running.
sys_enomem
Not enough memory to create child process.
 Close directory file descriptor

CloseDir closes the directory pointed to by p. It returns zero if the directory was closed succesfully, -1 otherwise.

For an example, see .

Errors are returned in LinuxError. Create an array of null-terminated strings

CreateShellArgV creates an array of 3 PChar pointers that can be used as arguments to ExecVE the first elements in the array will contain /bin/sh, the second will contain -c, and the third will contain prog.

The function returns a pointer to this array, of type PPChar.

None. Extract directory part from filename

Returns the directory part of Path. The directory is the part of Path before the last slash, or empty if there is no slash. The last character of the result is not a slash, unless the directory is the root directory.

None. Duplicate a file handle

Makes NewFile an exact copy of OldFile, after having flushed the buffer of OldFile in case it is a Text file or untyped file. Due to the buffering mechanism of Pascal, this has not the same functionality as the dup call in C. The internal Pascal buffers are not the same after this call, but when the buffers are flushed (e.g. after output), the output is sent to the same file. Doing an lseek will, however, work as in C, i.e. doing a lseek will change the fileposition in both files.

The function returns False in case of an error, True if successful.

In case of errors, Linuxerror is used to report errors.

sys_ebadf
OldFile hasn't been assigned.
sys_emfile
Maximum number of open files for the process is reached.
 Duplicate one filehandle to another

Makes NewFile an exact copy of OldFile, after having flushed the buffer of OldFile in the case of text or untyped files.

NewFile can be an assigned file. If newfile was open, it is closed first. Due to the buffering mechanism of Pascal, this has not the same functionality as the dup2 call in C. The internal Pascal buffers are not the same after this call, but when the buffers are flushed (e.g. after output), the output is sent to the same file. Doing an lseek will, however, work as in C, i.e. doing a lseek will change the fileposition in both files.

The function returns True if succesful, false otherwise.

In case of error, Linuxerror is used to report errors.

sys_ebadf
OldFile hasn't been assigned.
sys_emfile
Maximum number of open files for the process is reached.
 Convert epoch time to local time

Converts the epoch time (=Number of seconds since 00:00:00 , January 1, 1970, corrected for your time zone ) to local date and time.

This function takes into account the timzeone settings of your system.

None Execute process (using argument list) Replaces the currently running program with the program, specified in path. Path is split into a command and it's options. The executable in path is NOT searched in the path. The current environment is passed to the program. On success, execl does not return.

Errors are reported in LinuxError:

sys_eacces
File is not a regular file, or has no execute permission. A compononent of the path has no search permission.
sys_eperm
The file system is mounted noexec.
sys_e2big
Argument list too big.
sys_enoexec
The magic number in the file is incorrect.
sys_enoent
The file does not exist.
sys_enomem
Not enough memory for kernel, or to split command line.
sys_enotdir
A component of the path is not a directory.
sys_eloop
The path contains a circular reference (via symlinks).
 Execute process (using argument list, environment)

Replaces the currently running program with the program, specified in path. Path is split into a command and it's options. The executable in path is searched in the path, if it isn't an absolute filename. The environment in ep is passed to the program. On success, execle does not return.

Errors are reported in LinuxError:

sys_eacces
File is not a regular file, or has no execute permission. A compononent of the path has no search permission.
sys_eperm
The file system is mounted noexec.
sys_e2big
Argument list too big.
sys_enoexec
The magic number in the file is incorrect.
sys_enoent
The file does not exist.
sys_enomem
Not enough memory for kernel, or to split command line.
sys_enotdir
A component of the path is not a directory.
sys_eloop
The path contains a circular reference (via symlinks).
 Execute process (using argument list, environment; search path) Replaces the currently running program with the program, specified in path. Path is split into a command and it's options. The executable in path is searched in the path, if it isn't an absolute filename. The current environment is passed to the program. On success, execlp does not return.

Errors are reported in LinuxError:

sys_eacces
File is not a regular file, or has no execute permission. A compononent of the path has no search permission.
sys_eperm
The file system is mounted noexec.
sys_e2big
Argument list too big.
sys_enoexec
The magic number in the file is incorrect.
sys_enoent
The file does not exist.
sys_enomem
Not enough memory for kernel, or to split command line.
sys_enotdir
A component of the path is not a directory.
sys_eloop
The path contains a circular reference (via symlinks).
 Execute process Replaces the currently running program with the program, specified in path. It gives the program the options in args. This is a pointer to an array of pointers to null-terminated strings. The last pointer in this array should be nil. The current environment is passed to the program. On success, execv does not return.

Errors are reported in LinuxError:

sys_eacces
File is not a regular file, or has no execute permission. A compononent of the path has no search permission.
sys_eperm
The file system is mounted noexec.
sys_e2big
Argument list too big.
sys_enoexec
The magic number in the file is incorrect.
sys_enoent
The file does not exist.
sys_enomem
Not enough memory for kernel.
sys_enotdir
A component of the path is not a directory.
sys_eloop
The path contains a circular reference (via symlinks).
 Execute process using environment Replaces the currently running program with the program, specified in path. It gives the program the options in args, and the environment in ep. They are pointers to an array of pointers to null-terminated strings. The last pointer in this array should be nil. On success, execve does not return.

Errors are reported in LinuxError:

sys_eacces
File is not a regular file, or has no execute permission. A compononent of the path has no search permission.
sys_eperm
The file system is mounted noexec.
sys_e2big
Argument list too big.
sys_enoexec
The magic number in the file is incorrect.
sys_enoent
The file does not exist.
sys_enomem
Not enough memory for kernel.
sys_enotdir
A component of the path is not a directory.
sys_eloop
The path contains a circular reference (via symlinks).
 Execute process, search path Replaces the currently running program with the program, specified in path. The executable in path is searched in the path, if it isn't an absolute filename. It gives the program the options in args. This is a pointer to an array of pointers to null-terminated strings. The last pointer in this array should be nil. The current environment is passed to the program. On success, execvp does not return.

Errors are reported in LinuxError:

sys_eacces
File is not a regular file, or has no execute permission. A compononent of the path has no search permission.
sys_eperm
The file system is mounted noexec.
sys_e2big
Argument list too big.
sys_enoexec
The magic number in the file is incorrect.
sys_enoent
The file does not exist.
sys_enomem
Not enough memory for kernel.
sys_enotdir
A component of the path is not a directory.
sys_eloop
The path contains a circular reference (via symlinks).
 Clear all file descriptors in set

FD_ZERO clears all the filedescriptors in the file descriptor set fds.

For an example, see .

None. Clears a filedescriptor in a set

FD_Clr clears file descriptor fd in filedescriptor set fds.

For an example, see .

None. , Check whether a filedescriptor is set

FD_Set Checks whether file descriptor fd in filedescriptor set fds is set.

For an example, see .

None. Set a filedescriptor in a set

FD_Set sets file descriptor fd in filedescriptor set fds.

For an example, see .

None. Close file descriptor

fdClose closes a file with file descriptor Fd. The function returns True if the file was closed successfully, False otherwise.

For an example, see .

Errors are returned in LinuxError. Flush kernel file buffer

fdflush flushes the Linux kernel file buffer, so the file is actually written to disk. This is NOT the same as the internal buffer, maintained by Free Pascal. The function returns True if the call was successful, false if an error occurred.

For an example, see .

Errors are returned in LinuxError. Open file and return file descriptor

fdOpen opens a file in PathName with flags flags One of the following:

Open_RdOnly
File is opened Read-only
Open_WrOnly
File is opened Write-only
Open_RdWr
File is opened Read-Write

The flags may beOR-ed with one of the following constants:

Open_Creat
File is created if it doesn't exist.
Open_Excl
If the file is opened with Open_Creat and it already exists, the call wil fail.
Open_NoCtty
If the file is a terminal device, it will NOT become the process' controlling terminal.
Open_Trunc
If the file exists, it will be truncated.
Open_Append
the file is opened in append mode. Before each write, the file pointer is positioned at the end of the file.
Open_NonBlock
The file is opened in non-blocking mode. No operation on the file descriptor will cause the calling process to wait till.
Open_NDelay
Idem as Open_NonBlock
Open_Sync
The file is opened for synchronous IO. Any write operation on the file will not return untill the data is physically written to disk.
Open_NoFollow
if the file is a symbolic link, the open fails. (linux 2.1.126 and higher only)
Open_Directory
if the file is not a directory, the open fails. (linux 2.1.126 and higher only)

PathName can be of type PChar or String. The optional mode argument specifies the permissions to set when opening the file. This is modified by the umask setting. The real permissions are Mode and not umask. The return value of the function is the filedescriptor, or a negative value if there was an error.

Errors are returned in LinuxError. Read data from file descriptor

fdRead reads at most size bytes from the file descriptor fd, and stores them in buf. The function returns the number of bytes actually read, or -1 if an error occurred. No checking on the length of buf is done.

Errors are returned in LinuxError. Set file pointer position.

fdSeek sets the current fileposition of file fd to Pos, starting from SeekType, which can be one of the following:

Seek_Set
Pos is the absolute position in the file.
Seek_Cur
Pos is relative to the current position.
Seek_end
Pos is relative to the end of the file.

The function returns the new fileposition, or -1 of an error occurred.

For an example, see .

Errors are returned in LinuxError. Truncate file on certain size. fdTruncate sets the length of a file in fd on size bytes, where size must be less than or equal to the current length of the file in fd. The function returns True if the call was successful, false if an error occurred. Errors are returned in LinuxError. Write data to file descriptor

fdWrite writes at most size bytes from buf to file descriptor fd. The function returns the number of bytes actually written, or -1 if an error occurred.

Errors are returned in LinuxError. Expand filename to fully qualified path FExpand expands Path to a full path, starting from root, eliminating directory references such as . and .. from the result. None Lock a file (advisory lock)

FLock implements file locking. it sets or removes a lock on the file F. F can be of type Text or File, or it can be a linux filedescriptor (a longint) Mode can be one of the following constants :

LOCK_SH
sets a shared lock.
LOCK_EX
sets an exclusive lock.
LOCK_UN
unlocks the file.
LOCK_NB
This can be OR-ed together with the other. If this is done the application doesn't block when locking.

The function returns True if successful, False otherwise.

If an error occurs, it is reported in LinuxError. Check whether filename matches wildcard specification

FNMatch returns True if the filename in Name matches the wildcard pattern in Pattern, False otherwise.

Pattern can contain the wildcards * (match zero or more arbitrary characters) or ? (match a single character).

None. Search for file in search path. FSearch searches in DirList, a colon separated list of directories, for a file named Path. It then returns a path to the found file. An empty string if no such file was found. Split filename into path, name and extension FSplit splits a full file name into 3 parts : A Path, a Name and an extension (in ext). The extension is taken to be all letters after the last dot (.). None. Retrieve filesystem information. FSStat returns in Info information about the filesystem on which the file Path resides, or on which the file with file descriptor fd resides. Info is of type statfs. The function returns True if the call was succesfull, False if the call failed.

LinuxError is used to report errors.

sys_enotdir
A component of Path is not a directory.
sys_einval
Invalid character in Path.
sys_enoent
Path does not exist.
sys_eaccess
Search permission is denied for component inPath.
sys_eloop
A circular symbolic link was encountered in Path.
sys_eio
An error occurred while reading from the filesystem.
 Retrieve information about a file

FStat gets information about the file specified in one of the following:

Path
a file on the filesystem.
Fd
a valid file descriptor.
F
an opened text file or untyped file.

and stores it in Info, which is of type stat. The function returns True if the call was succesfull, False if the call failed.

LinuxError is used to report errors.

sys_enoent
Path does not exist.
 File control operations.

Read a file's attributes. Fd is an assigned file, or a valid file descriptor. Cmd speciefies what to do, and is one of the following:

F_GetFd
Read the close_on_exec flag. If the low-order bit is 0, then the file will remain open across execve calls.
F_GetFl
Read the descriptor's flags.
F_GetOwn
Get the Process ID of the owner of a socket.
F_SetFd
Set the close_on_exec flag of Fd. (only the least siginificant bit is used).
F_GetLk
Return the flock record that prevents this process from obtaining the lock, or set the l_type field of the lock of there is no obstruction. Arg is a pointer to a flock record.
F_SetLk
Set the lock or clear it (depending on l_type in the flock structure). if the lock is held by another process, an error occurs.
F_GetLkw
Same as for F_Setlk, but wait until the lock is released.
F_SetOwn
Set the Process or process group that owns a socket.

LinuxError is used to report errors.

sys_ebadf
Fd has a bad file descriptor.
sys_eagain or sys_eaccess
For F_SetLk, if the lock is held by another process.
 Create child process Fork creates a child process which is a copy of the parent process. Fork returns the process ID in the parent process, and zero in the child's process. (you can get the parent's PID with ).

On error, -1 is returned to the parent, and no child is created.

sys_eagain
Not enough memory to create child process.
Rename file

FRename renames the file OldName to NewName. NewName can be in a different directory than OldName, but it cannot be on another partition (device). Any existing file on the new location will be replaced.

If the operation fails, then the OldName file will be preserved.

The function returns True on succes, False on failure.

On error, errors are reported in LinuxError. Possible errors include:

sys_eisdir
NewName exists and is a directory, but OldName is not a directory.
sys_exdev
NewName and OldName are on different devices.
sys_enotempty or sys_eexist
NewName is an existing, non-empty directory.
sys_ebusy
OldName or NewName is a directory and is in use by another process.
sys_einval
NewName is part of OldName.
sys_emlink
OldPath or NewPath already have tha maximum amount of links pointing to them.
sys_enotdir
part of OldName or NewName is not directory.
sys_efault
For the pchar case: One of the pointers points to an invalid address.
sys_eaccess
access is denied when attempting to move the file.
sys_enametoolong
Either OldName or NewName is too long.
sys_enoent
a directory component in OldName or NewName didn't exist.
sys_enomem
not enough kernel memory.
sys_erofs
NewName or OldName is on a read-only file system.
sys_eloop
too many symbolic links were encountered trying to expand OldName or NewName
sys_enospc
the filesystem has no room for the new directory entry.
 Return the system date Returns the current date. None Return system date and time Returns the current date and time. The time is corrected for the local time zone. This procedure is equivalent to the and GetTime calls. None Return current domain name Get the domain name of the machine on which the process is running. An empty string is returned if the domain is not set. None. Return effective group ID Get the effective group ID of the currently running process. None. Return effective user ID Get the effective user ID of the currently running process. None. Return value of environment variable.

GetEnv returns the value of the environment variable in P. If the variable is not defined, nil is returned. The value of the environment variable may be the empty string. A PChar is returned to accomodate for strings longer than 255 bytes, TERMCAP and LS_COLORS, for instance.

None. Return the current unix time

returns the number of seconds since 00:00:00 gmt, january 1, 1970. it is adjusted to the local time zone, but not to DST.

no errors Return file selector GetFS returns the file selector that the kernel provided for your file. In principle you don' need this file selector. Only for some calls it is needed, such as the call or so. In case the file was not opened, then -1 is returned. Return real group ID Get the real group ID of the currently running process. None. Return host name Get the hostname of the machine on which the process is running. An empty string is returned if hostname is not set. None. Return local timzeone information

GetLocalTimeZone returns the local timezone information. It also initializes the TZSeconds variable, which is used to correct the epoch time to local time.

There should never be any need to call this function directly. It is called by the initialization routines of the Linux unit.

Return current process ID Get the Process ID of the currently running process. None. Return parent process ID Get the Process ID of the parent process. None. Return process priority

GetPriority returns the priority with which a process is running. Which process(es) is determined by the Which and Who variables. Which can be one of the pre-defined Prio_Process, Prio_PGrp, Prio_User, in which case Who is the process ID, Process group ID or User ID, respectively.

For an example, see .

Error checking must be done on LinuxError, since a priority can be negative.

sys_esrch
No process found using which and who.
sys_einval
Which was not one of Prio_Process, Prio_Grp or Prio_User.
 Return current system time

Returns the current time of the day, adjusted to local time. Upon return, the parameters are filled with

hour
Hours since 00:00 today.
min
minutes in current hour.
sec
seconds in current minute.
sec100
hundreds of seconds in current second.
msec
milliseconds in current second.
usec
microseconds in current second.
 None Return kernel time of day in GMT

GetTimeOfDay returns the number of seconds since 00:00, January 1 1970, GMT in a timeval record. This time NOT corrected any way, not taking into account timezones, daylight savings time and so on.

It is simply a wrapper to the kernel system call. To get the local time, .

None. Return name of timezone information file

GetTimezoneFile returns the location of the current timezone file. The location of file is determined as follows:

  1. If /etc/timezone exists, it is read, and the contents of this file is returned. This should work on Debian systems.
  2. If /usr/lib/zoneinfo/localtime exists, then it is returned. (this file is a symlink to the timezone file on SuSE systems)
  3. If /etc/localtime exists, then it is returned. (this file is a symlink to the timezone file on RedHat systems)
 If no file was found, an empty string is returned. Return current user ID Get the real user ID of the currently running process. None. Find filenames matching a wildcard pattern

Glob returns a pointer to a glob structure which contains all filenames which exist and match the pattern in Path. The pattern can contain wildcard characters, which have their usual meaning.

Returns nil on error, and LinuxError is set.

sys_enomem
No memory on heap for glob structure.
others
As returned by the opendir call, and sys_readdir.
 Free result of call

Releases the memory, occupied by a pglob structure. P is set to nil.

For an example, see .

None General kernel IOCTL call. This is a general interface to the Unix/ linux ioctl call. It performs various operations on the filedescriptor Handle. Ndx describes the operation to perform. Data points to data needed for the Ndx function. The structure of this data is function-dependent, so we don't elaborate on this here. For more information on this, see various manual pages under linux. Errors are reported in LinuxError. They are very dependent on the used function, that's why we don't list them here Set permission on IO ports

IOperm sets permissions on Num ports starting with port From to Value. The function returns True if the call was successfull, False otherwise.

Note:

  • This works ONLY as root.
  • Only the first 0x03ff ports can be set.
  • When doing a , the permissions are reset. When doing a they are kept.
 Errors are returned in LinuxError Check if filehandle is a TTY (terminal)

Check if the filehandle described by f is a terminal. f can be of type

  1. longint for file handles;
  2. Text for text variables such as input etc.

Returns True if f is a terminal, False otherwise.

No errors are reported Is file a block device S_ISBLK checks the file mode m to see whether the file is a block device file. If so it returns True. Is file a character device S_ISCHR checks the file mode m to see whether the file is a character device file. If so it returns True. Is file a directory S_ISDIR checks the file mode m to see whether the file is a directory. If so it returns True Is file a FIFO S_ISFIFO checks the file mode m to see whether the file is a fifo (a named pipe). If so it returns True. Is file a symbolic link S_ISLNK checks the file mode m to see whether the file is a symbolic link. If so it returns True Is file a regular file S_ISREG checks the file mode m to see whether the file is a regular file. If so it returns True Is file a unix socket S_ISSOCK checks the file mode m to see whether the file is a socket. If so it returns True. Send a signal to a process Send a signal Sig to a process or process group. If Pid>0 then the signal is sent to Pid, if it equals -1, then the signal is sent to all processes except process 1. If Pid<-1 then the signal is sent to process group -Pid. The return value is zero, except in case three, where the return value is the number of processes to which the signal was sent.

LinuxError is used to report errors:

sys_einval
An invalid signal is sent.
sys_esrch
The Pid or process group don't exist.
sys_eperm
The effective userid of the current process doesn't math the one of process Pid.
 Return information about symbolic link. Do not follow the link LStat gets information about the link specified in Path, and stores it in Info, which is of type stat. Contrary to FStat, it stores information about the link, not about the file the link points to. The function returns True if the call was succesfull, False if the call failed.

LinuxError is used to report errors.

sys_enoent
Path does not exist.
 Create a hard link to a file Link makes NewPath point to the same file als OldPath. The two files then have the same inode number. This is known as a 'hard' link. The function returns True if the call was succesfull, False if the call failed.

Errors are returned in LinuxError.

sys_exdev
OldPath and NewPath are not on the same filesystem.
sys_eperm
The filesystem containing oldpath and newpath doesn't support linking files.
sys_eaccess
Write access for the directory containing Newpath is disallowed, or one of the directories in OldPath or {NewPath} has no search (=execute) permission.
sys_enoent
A directory entry in OldPath or NewPath does not exist or is a symbolic link pointing to a non-existent directory.
sys_enotdir
A directory entry in OldPath or NewPath is nor a directory.
sys_enomem
Insufficient kernel memory.
sys_erofs
The files are on a read-only filesystem.
sys_eexist
NewPath already exists.
sys_emlink
OldPath has reached maximal link count.
sys_eloop
OldPath or NewPath has a reference to a circular symbolic link, i.e. a symbolic link, whose expansion points to itself.
sys_enospc
The device containing NewPath has no room for another entry.
sys_eperm
OldPath points to . or .. of a directory.
 Convert local time to epoch (unix) time Converts the Local time to epoch time (=Number of seconds since 00:00:00 , January 1, 1970 ). None Create FIFO (named pipe) in file system MkFifo creates named a named pipe in the filesystem, with name PathName and mode Mode.

LinuxError is used to report errors:

sys_emfile
Too many file descriptors for this process.
sys_enfile
The system file table is full.
 Create memory map of a file

MMap maps or unmaps files or devices into memory. The different fields of the argument m determine what and how the mmap maps this:

address
Address where to mmap the device. This address is a hint, and may not be followed.
size
Size (in bytes) of area to be mapped.
prot

Protection of mapped memory. This is a OR-ed combination of the following constants:

PROT_EXEC
The memory can be executed.
PROT_READ
The memory can be read.
PROT_WRITE
The memory can be written.
PROT_NONE
The memory can not be accessed.
flags

Contains some options for the mmap call. It is an OR-ed combination of the following constants:

MAP_FIXED
Do not map at another address than the given address. If the address cannot be used, MMap will fail.
MAP_SHARED
Share this map with other processes that map this object.
MAP_PRIVATE
Create a private map with copy-on-write semantics.
MAP_ANONYMOUS
fd does not have to be a file descriptor.

One of the options MAP_SHARED and MAP_PRIVATE must be present, but not both at the same time.

fd
File descriptor from which to map.
offset
Offset to be used in file descriptor fd.

The function returns a pointer to the mapped memory, or a -1 in case of en error.

On error, -1 is returned and LinuxError is set to the error code:

Sys_EBADF
fd is not a valid file descriptor and MAP_ANONYMOUS was not specified.
Sys_EACCES
MAP_PRIVATE was specified, but fd is not open for reading. Or MAP_SHARED was asked and PROT_WRITE is set, fd is not open for writing
Sys_EINVAL
One of the record fields Start, length or offset is invalid.
Sys_ETXTBUSY
MAP_DENYWRITE was set but the object specified by fd is open for writing.
Sys_EAGAIN
fd is locked, or too much memory is locked.
Sys_ENOMEM
Not enough memory for this operation.
 Unmap previously mapped memory block

MUnMap unmaps the memory block of size Size, pointed to by P, which was previously allocated with .

The function returns True if successful, False otherwise.

For an example, see .

In case of error the function returns False and LinuxError is set to an error value. See for possible error values. Suspend process for a short time

NanoSleep suspends the process till a time period as specified in req has passed. Then the function returns. If the call was interrupted (e.g. by some signal) then the function may return earlier, and rem will contain the remaining time till the end of the intended period. In this case the return value will be -1, and LinuxError will be set to EINTR

If the function returns without error, the return value is zero.

If the call was interrupted, -1 is returned, and LinuxError is set to EINTR. If invalid time values were specified, then -1 is returned and LinuxError is set to EINVAL. Set process priority

Nice adds -N to the priority of the running process. The lower the priority numerically, the less the process is favored. Only the superuser can specify a negative N, i.e. increase the rate at which the process is run.

Errors are returned in LinuxError

sys_eperm
A non-superuser tried to specify a negative N, i.e. do a priority increase.
 Convert octal to decimal value

Octal will convert a number specified as an octal number to it's decimal value.

This is useful for the call, where permissions are specified as octal numbers.

No checking is performed whether the given number is a correct Octal number. e.g. specifying 998 is possible; the result will be wrong in that case. Open directory for reading OpenDir opens the directory f, and returns a pdir pointer to a Dir record, which can be used to read the directory structure. If the directory cannot be opened, nil is returned. Errors are returned in LinuxError. Wait for a signal

Pause puts the process to sleep and waits until the application receives a signal. If a signal handler is installed for the received sigal, the handler will be called and after that pause will return control to the process.

For an example, see .

Close file opened with

PClose closes a file opened with . It waits for the command to complete, and then returns the exit status of the command.

For an example, see

LinuxError is used to report errors. If it is different from zero, the exit status is not valid. Pipe file to standard input/output of program POpen runs the command specified in Cmd, and redirects the standard in or output of the command to the other end of the pipe F. The parameter rw indicates the direction of the pipe. If it is set to 'W', then F can be used to write data, which will then be read by the command from stdinput. If it is set to 'R', then the standard output of the command can be read from F. F should be reset or rewritten prior to using it. F can be of type Text or File. A file opened with POpen can be closed with Close, but also with . The result is the same, but PClose returns the exit status of the command Cmd. Errors are reported in LinuxError and are essentially those of the Execve, Dup and AssignPipe commands. Read entry from directory

ReadDir reads the next entry in the directory pointed to by p. It returns a pdirent pointer to a structure describing the entry. If the next entry can't be read, Nil is returned.

For an example, see .

Errors are returned in LinuxError. , Read destination of symbolic link

ReadLink returns the file the symbolic link name is pointing to. The first form of this function accepts a buffer linkname of length maxlen where the filename will be stored. It returns the actual number of characters stored in the buffer.

The second form of the function returns simply the name of the file.

On error, the first form of the function returns -1; the second one returns an empty string. LinuxError is set to report errors:

SYS_ENOTDIR
A part of the path in Name is not a directory.
SYS_EINVAL
maxlen is not positive, or the file is not a symbolic link.
SYS_ENAMETOOLONG
A pathname, or a component of a pathname, was too long.
SYS_ENOENT
the link name does not exist.
SYS_EACCES
No permission to search a directory in the path
SYS_ELOOP
Too many symbolic links were encountered in translating the pathname.
SYS_EIO
An I/O error occurred while reading from the file system.
SYS_EFAULT
The buffer is not part of the the process's memory space.
SYS_ENOMEM
Not enough kernel memory was available.
 Read the timezone file and initialize time routines

ReadTimeZoneFile reads the timezone file fn and initializes the local time routines based on the information found there.

There should be no need to call this function. The initialization routines of the linux unit call this routine at unit startup.

None. Seek to position in directory

SeekDir sets the directory pointer to the off-th entry in the directory structure pointed to by p.

For an example, see .

Errors are returned in LinuxError. Wait for events on file descriptors

Select checks one of the file descriptors in the FDSets to see if its status changed.

readfds, writefds and exceptfds are pointers to arrays of 256 bits. If you want a file descriptor to be checked, you set the corresponding element in the array to 1. The other elements in the array must be set to zero. Three arrays are passed : The entries in readfds are checked to see if characters become available for reading. The entries in writefds are checked to see if it is OK to write to them, while entries in exceptfds are cheked to see if an exception occorred on them.

You can use the functions , or to manipulate the individual elements of a set.

The pointers can be Nil.

N is the largest index of a nonzero entry plus 1. (= the largest file-descriptor + 1).

TimeOut can be used to set a time limit. If TimeOut can be two types :

  1. TimeOut is of type PTime and contains a zero time, the call returns immediately. If TimeOut is Nil, the kernel will wait forever, or until a status changed.
  2. TimeOut is of type Longint. If it is -1, this has the same effect as a Timeout of type PTime which is Nil. Otherwise, TimeOut contains a time in milliseconds.

When the TimeOut is reached, or one of the file descriptors has changed, the Select call returns. On return, it will have modified the entries in the array which have actually changed, and it returns the number of entries that have been changed. If the timout was reached, and no decsriptor changed, zero is returned; The arrays of indexes are undefined after that. On error, -1 is returned.

On error, the function returns -1, and Errors are reported in LinuxError :

SYS_EBADF
An invalid descriptor was specified in one of the sets.
SYS_EINTR
A non blocked signal was caught.
SYS_EINVAL
N is negative or too big.
SYS_ENOMEM
Select was unable to allocate memory for its internal tables.
 Wait for event on typed ontyped file. SelectText executes the call on a file of type Text. You can specify a timeout in TimeOut. The SelectText call determines itself whether it should check for read or write, depending on how the file was opened : With Reset it is checked for reading, with Rewrite and Append it is checked for writing. See . SYS_EBADF can also mean that the file wasn't opened. Set process priority

SetPriority sets the priority with which a process is running. Which process(es) is determined by the Which and Who variables. Which can be one of the pre-defined constants:

Prio_Process
Who is interpreted as process ID
Prio_PGrp
Who is interpreted as process group ID
Prio_User
Who is interpreted as user ID

Prio is a value in the range -20 to 20.

For an example, see .

Error checking must be done on LinuxError, since a priority can be negative.

sys_esrch
No process found using which and who.
sys_einval
Which was not one of Prio_Process, Prio_Grp or Prio_User.
sys_eperm
A process was found, but neither its effective or real user ID match the effective user ID of the caller.
sys_eacces
A non-superuser tried to a priority increase.
 Execute and feed command to system shell Shell invokes the bash shell (/bin/sh), and feeds it the command Command (using the -c option). The function then waits for the command to complete, and then returns the exit status of the command, or 127 if it could not complete the or calls. Errors are reported in LinuxError. Install signal handler

Changes the action to take upon receipt of a signal. Act and Oldact are pointers to a SigActionRec record. SigNum specifies the signal, and can be any signal except SIGKILL or SIGSTOP.

If Act is non-nil, then the new action for signal SigNum is taken from it. If OldAct is non-nil, the old action is stored there. Sa_Handler may be SIG_DFL for the default action or SIG_IGN to ignore the signal. Sa_Mask Specifies which signals should be ignord during the execution of the signal handler. Sa_Flags Speciefies a series of flags which modify the behaviour of the signal handler. You can 'or' none or more of the following :

SA_NOCLDSTOP
If signum is SIGCHLD do not receive notification when child processes stop.
SA_ONESHOT or SA_RESETHAND
Restore the signal action to the default state once the signal handler has been called.
SA_RESTART
For compatibility with BSD signals.
SA_NOMASK or SA_NODEFER
Do not prevent the signal from being received from within its own signal handler.

LinuxError is used to report errors.

sys_einval
an invalid signal was specified, or it was SIGKILL or SIGSTOP.
sys_efault
Act,OldAct point outside this process address space
sys_eintr
System call was interrupted.
 Return set of currently pending signals Sigpending allows the examination of pending signals (which have been raised while blocked.) The signal mask of pending signals is returned. None Set list of blocked signals

Changes the list of currently blocked signals. The behaviour of the call depends on How :

SIG_BLOCK
The set of blocked signals is the union of the current set and the SSet argument.
SIG_UNBLOCK
The signals in SSet are removed from the set of currently blocked signals.
SIG_SETMASK
The list of blocked signals is set so SSet.

If OldSSet is non-nil, then the old set is stored in it.

LinuxError is used to report errors.

sys_efault
SSet or OldSSet point to an adress outside the range of the process.
sys_eintr
System call was interrupted.
 Raise a signal (send to current process) SigRaise sends a Sig signal to the current process. None. Set signal mask and suspend process till signal is received SigSuspend temporarily replaces the signal mask for the process with the one given in Mask, and then suspends the process until a signal is received. None Install signal handler (deprecated)

Signal installs a new signal handler for signal SigNum. This call has the same functionality as the SigAction call. The return value for Signal is the old signal handler, or nil on error.

LinuxError is used to report errors :

SIG_ERR
An error occurred.
 Split string in list of null-terminated strings

StringToPPChar splits the string S in words, replacing any whitespace with zero characters. It returns a pointer to an array of pchars that point to the first letters of the words in S. This array is terminated by a Nil pointer.

The function does not add a zero character to the end of the string unless it ends on whitespace.

The function reserves memory on the heap to store the array of PChar; The caller is responsible for freeing this memory.

This function can be called to create arguments for the various Exec calls.

None. Create a symbolic link

SymLink makes Newpath point to the file in OldPath, which doesn't necessarily exist. The two files DO NOT have the same inode number. This is known as a 'soft' link.

The permissions of the link are irrelevant, as they are not used when following the link. Ownership of the file is only checked in case of removal or renaming of the link.

The function returns True if the call was succesfull, False if the call failed.

Errors are returned in LinuxError.

sys_eperm
The filesystem containing oldpath and newpath does not support linking files.
sys_eaccess
Write access for the directory containing Newpath is disallowed, or one of the directories in OldPath or NewPath has no search (=execute) permission.
sys_enoent
A directory entry in OldPath or NewPath does not exist or is a symbolic link pointing to a non-existent directory.
sys_enotdir
A directory entry in OldPath or NewPath is nor a directory.
sys_enomem
Insufficient kernel memory.
sys_erofs
The files are on a read-only filesystem.
sys_eexist
NewPath already exists.
sys_eloop
OldPath or NewPath has a reference to a circular symbolic link, i.e. a symbolic link, whose expansion points to itself.
sys_enospc
The device containing NewPath has no room for another entry.
 Return kernel system information

SysInfo returns system information in Info. Returned information in Info includes:

uptime
Number of seconds since boot.
loads
1, 5 and 15 minute load averages.
totalram
total amount of main memory.
freeram
amount of free memory.
sharedram
amount of shared memory.
bufferram
amount of memory used by buffers.
totalswap
total amount of swapspace.
freeswap
amount of free swapspace.
procs
number of current processes.
 None. Terminal control: Wait till all data was transmitted

TCDrain waits until all data to file descriptor Fd is transmitted.

The function returns True if the call was succesfull, False otherwise.

Errors are reported in LinuxError Terminal control: Suspend transmission of data

TCFlow suspends/resumes transmission or reception of data to or from the file descriptor Fd, depending on the action Act.

This can be one of the following pre-defined values:

TCOOFF
suspend reception/transmission
TCOON
resume reception/transmission
TCIOFF
transmit a stop character to stop input from the terminal
TCION
transmit start to resume input from the terminal.

The function returns True if the call was succesfull, False otherwise.

Errors are reported in LinuxError. Terminal control: Discard data buffer

TCFlush discards all data sent or received to/from file descriptor fd. QSel indicates which queue should be discard. It can be one of the following pre-defined values :

TCIFLUSH
input buffer
TCOFLUSH
output buffer
TCIOFLUSH
both input and output buffers

The function returns True if the call was succesfull, False otherwise.

Errors are reported in LinuxError. Terminal Control: Get terminal attributes TCGetAttr gets the terminal parameters from the terminal referred to by the file descriptor fd and returns them in a TermIOS structure tios. The function returns True if the call was succesfull, False otherwise. Errors are reported in LinuxError Terminal control: Get process group TCGetPGrp returns the process group ID of a foreground process group in Id The function returns True if the call was succesfull, False otherwise. Errors are reported in LinuxError Terminal control: Send break TCSendBreak Sends zero-valued bits on an asynchrone serial connection decsribed by file-descriptor Fd, for duration Duration. The function returns True if the action was performed successfully, False otherwise. Errors are reported in LinuxError. Terminal control: Set attributes

TCSetAttr sets the terminal parameters you specify in a TermIOS structureTios for the terminal referred to by the file descriptor Fd.

OptAct specifies an optional action when the set need to be done, this could be one of the following pre-defined values:

TCSANOW
set immediately.
TCSADRAIN
wait for output.
TCSAFLUSH
wait for output and discard all input not yet read.

The function Returns True if the call was succesfull, False otherwise.

For an example, see .

Errors are reported in LinuxError. Terminal control: Set process group

TCSetPGrp Sets the Process Group Id to Id. The function returns True if the call was successful, False otherwise.

For an example, see .

Errors are returned in LinuxError. Terminal control: Get terminal name

TTYNameReturns the name of the terminal pointed to by f. f must be a terminal. f can be of type:

  1. longint for file handles;
  2. Text for text variables such as input etc.
 Returns an empty string in case of an error. Linuxerror may be set to indicate what error occurred, but this is uncertain. Return current location in a directory

TellDir returns the current location in the directory structure pointed to by p. It returns -1 on failure.

For an example, see .

Errors are returned in LinuxError. Set file creation mask. Change the file creation mask for the current user to Mask. The current mask is returned. Return system name. Uname gets the name and configuration of the current linux kernel, and returns it in unamerec. LinuxError is used to report errors. Unlink (i.e. remove) a file.

UnLink decreases the link count on file Path. Path can be of type PathStr or PChar. If the link count is zero, the file is removed from the disk. The function returns True if the call was succesfull, False if the call failed.

For an example, see .

Errors are returned in LinuxError.

sys_eaccess
You have no write access right in the directory containing Path, or you have no search permission in one of the directory components of Path.
sys_eperm
The directory containing pathname has the sticky-bit set and the process's effective uid is neither the uid of the file to be deleted nor that of the directory containing it.
sys_enoent
A component of the path doesn't exist.
sys_enotdir
A directory component of the path is not a directory.
sys_eisdir
Path refers to a directory.
sys_enomem
Insufficient kernel memory.
sys_erofs
Path is on a read-only filesystem.
 Set access and modification times of a file (touch). Utime sets the access and modification times of a file. the utimbuf record contains 2 fields, actime, and modtime, both of type Longint. They should be filled with an epoch-like time, specifying, respectively, the last access time, and the last modification time. For some filesystem (most notably, FAT), these times are the same.

Errors are returned in LinuxError.

sys_eaccess
One of the directories in Path has no search (=execute) permission.
sys_enoent
A directory entry in Path does not exist or is a symbolic link pointing to a non-existent directory.

Other errors may occur, but aren't documented.

Wait for a process to terminate

WaitPid waits for a child process with process ID Pid to exit. The value of Pid can be one of the following:

Pid < -1
Causes WaitPid to wait for any child process whose process group ID equals the absolute value of pid.
Pid = -1
Causes WaitPid to wait for any child process.
Pid = 0
Causes WaitPid to wait for any child process whose process group ID equals the one of the calling process.
Pid > 0
Causes WaitPid to wait for the child whose process ID equals the value of Pid.

The Options parameter can be used to specify further how WaitPid behaves:

WNOHANG
Causes Waitpid to return immediately if no child hasexited.
WUNTRACED
Causes WaitPid to return also for children which are stopped, but whose status has not yet been reported.
__WCLONE
Causes WaitPid also to wait for threads created by the call.

Upon return, it returns the exit status of the process, or -1 in case of failure.

For an example, see .

Errors are returned in LinuxError. Seek option: Set absolute position. Seek option: Set position relative to current position. Seek option: Set position relative to end of file. Bitmask to determine access mode in open flags. File open mode: Read only File open mode: Write only File open mode: Read/Write File open mode: Create if file does not yet exist. File open mode: Open exclusively File open mode: No TTY control. File open mode: Truncate file to length 0 File open mode: Append to file File open mode: Open in non-blocking mode File open mode: Alias for File open mode: Write to disc at once File open mode: Minimize caching effects File open mode: Open for 64-bit I/O File open mode: File must be directory. File open mode: Fail if file is symbolic link. : Do not wait : Also report stopped but untraced processes : Wait on any process : Wait processes from current process group : Wait on clone processes only. IOCTL call number: get Terminal Control settings Register describing system calls. Register 1 Register 2 Register 3 Register 4 Register 5 Register 6 Pointer to record. Alias for record Alias for record Device descriptor type Block size Pointer to record. Alias for record. Pointer to record. Alias for type. Alias for type. Alias for record. Record describing a timezone Minutes west of GMT Daylight savings time Pointer to record. Alias for record. Pointer to record. Alias for record. Signal set type Pointer to signal set. Socket option Signal information record pad bytes size. Do not use. Size type Socket options Socket options Signal Stack size error Alternate stack registers record Stack pointer Flags Stack size Alias for type Pointer to record Pointer to record Error number of last operation. option: Signal mask to be sent at exit option: VM shared between processes option: fs info shared between processes option: open files shared between processes option: signal handlers shared between processes option: PID shared between processes Clone function prototype. option: Do not wait for processes to terminate. option: Also report children wich were stopped but not yet reported Waitpid option: Wait for clone children only Input file descriptor of pipe pair. Output file descriptor of pipe pair. Array describing a pipe pair of filedescriptors. Command-line string type. Filename path part string type. Filename directory part string type. Filename name part string type. Filename extension part string type. Record describing terminal window size. Number of rows Number of columns Number of pixels in horizontal direction Number of pixels in vertical direction Alias for record. Alias for record Input speed Output speed Alias for record. Alias for record. Alias for record. Pointer to record Record with system information, used by the call. Number of seconds since boot. 1, 5 and 15 minute load averages. total amount of main memory. amount of free memory. amount of shared memory. amount of memory used by buffers. total amount of swapspace. amount of free swapspace. number of current processes. ? Pointer to record. Execute system call. SysCall can be used to execute a direct system call. The call parameters must be encoded in regs and the call number must be specified by callnr. The call result is returned, and any modified registers are in regs None. Indicates whether daylight savings time is active. Seconds west of GMT Timezone name. Set the current system time.

SetTime sets the system time to hour, min, Sec. This is the kernel time, so it is in GMT. The date is not touched. The function returns True if the call was executed corretly, False otherwise.

You must be root to execute this call. Errors are returned in Set the current system date.

SetDate sets the system date to year, month, day. This is the kernel date, so it is in GMT. The time is not touched. The function returns True if the call was executed corretly, False otherwise.

You must be root to execute this call. Errors are returned in Set the current system date and time

SetDate sets the system date and time to year, month, day, hour, min, Sec. This is the kernel date/time, so it is in GMT. The time is not touched. The function returns True if the call was executed corretly, False otherwise.

You must be root to execute this call. Errors are returned in Exit the current process

ExitProcess exits the currently running process, and report Val as the exit status.

If this call is executed, the normal unit finalization code will not be executed. This may lead to unexpected errors and stray files on your system. It is therefore recommended to use the Halt call instead. None. Wait for process to terminate.

WaitProcess waits for process PID to exit. WaitProcess is equivalent to the call:

WaitPid(PID,@result,0)

Handles of Signal interrupts (errno=EINTR), and returns the Exitcode of Process PID (>=0) or -Status if it was terminated

None. Extract the exit status from the result. WEXITSTATUS can be used to extract the exit status from Status, the result of the call. Return the signal that caused a process to exit. WTERMSIG extracts from Status the signal number which caused the process to exit. Return the exit code from the process. WSTOPSIG is an alias for . Check whether the process exited normally WIFEXITED checks Status and returns True if the status indicates that the process terminated normally, i.e. was not stopped by a signal. Check whether the process is currently stopped. WIFSTOPPED checks Status and returns true if the process is currently stopped. This is only possible if WUNTRACED was specified in the options of . Check whether the process was exited by a signal. WIFSIGNALED returns True if Status indicates that the process exited because it received a signal. Construct an exit status based on an return code and signal. W_EXITCODE combines ReturnCode and Signal to a status code fit for WaitPid. Construct an exit status based on a signal. W_STOPCODE constructs an exit status based on Signal, which will cause to return True option: Memory grows downward (like a stack) option: Ignored. option: Ignored. option: lock the pages in memory. option: Do not reserve swap pages for this memory. Record containing mmap args. Address pointer Size of region. Memory protection flags Flags File descriptor Offset: Multiple of page size Set I/O privilege level IoPL sets the I/O privilige level. It is intended for completeness only, one should normally not use it.