NAME
sticky
- Description of the `sticky' (S_ISVTX) bit functionality
DESCRIPTION
A special file mode, called the
sticky bit
(mode
S_ISVTX
),
is used to indicate special treatment for directories.
See
chmod(2)
or the file
/usr/include/sys/stat.h
STICKY FILES
For regular files, the use of mode
S_ISVTX
is reserved and can be set only by the super-user.
NetBSD
does not currently treat regular files that have the sticky bit set
specially, but this behavior might change in the future.
STICKY DIRECTORIES
A directory whose
``sticky bit''
is set becomes a
directory in which the deletion of files is restricted.
A file in a sticky directory may only be removed or renamed
by a user if the user has write permission for the directory and
the user is the owner of the file, the owner of the directory,
or the super-user.
This feature is usefully applied to directories such as
/tmp
which must be publicly writable but should deny users the license
to arbitrarily delete or rename each others' files.
Any user may create a sticky directory.
See
chmod(1)
for details about modifying file modes.
HISTORY
The sticky bit first appeared in V7, and this manual page appeared
in section 8.
Its initial use was to mark sharable executables
that were frequently used so that they would stay in swap after
the process exited.
Sharable executables were compiled in a special way so their text
and read-only data could be shared amongst processes.
vi(1)
and
sh(1)
were such executables.
This is where the term
``sticky''
comes from - the program would stick around in swap, and it would
not have to be fetched again from the file system.
Of course as long as there was a copy in the swap area, the file
was marked busy so it could not be overwritten.
On V7 this meant that the file could not be removed either, because
busy executables could not be removed, but this restriction was
lifted in BSD releases.
To replace such executables was a cumbersome process.
One had first to remove the sticky bit, then execute the binary so
that the copy from swap was flushed, overwrite the executable, and
finally reset the sticky bit.
Later, on SunOS 4, the sticky bit got an additional meaning for
files that had the bit set and were not executable: read and write
operations from and to those files would go directly to the disk
and bypass the buffer cache.
This was typically used on swap files for NFS clients on an NFS
server, so that swap I/O generated by the clients on the servers
would not evict useful data from the server's buffer cache.
BUGS
Neither
open(2)
nor
mkdir(2)
will create a file with the sticky bit set.