RBFS(4) RBFS(4)
NAME
rbfs - ring buffer memory file system
SYNOPSIS
rbfs [ -Dipsu ] [ -m mountpoint ] [ -S srvname ] [ -l
linelen ] [ -n nlines ]
DESCRIPTION
Rbfs starts a process that mounts itself (see bind(2)) on
mountpoint (default /n/rb). The rbfs process implements a
file tree rooted at dir, keeping all files in memory. Ini-
tially the file tree is empty. Each file is modeled as a
ring buffer of writes. The default number of writes is
1024, and the default maximum write size is 256. The -n and
-l flags alter these defaults.
The -D option enables a trace of general debugging messages.
The -i flag tells rbfs to use file descriptors 0 and 1 for
its communication channel rather than create a pipe. This
makes it possible to use rbfs as a file server on a remote
machine: the file descriptors 0 and 1 will be the network
channel from rbfs to the client machine.
The -p flag causes rbfs to make its memory `private' (see
proc(3)) so that its files are not accessible through the
debugging interface.
The -s (-S) flag causes rbfs to post its channel on
/srv/rbfs (/srv/srvname) rather than mounting it on
mountpoint, enabling multiple clients to access its files.
However, it does not authenticate its clients and its imple-
mentation of groups is simplistic, so it should not be used
for precious data.
The -u option permits rbfs to consume as much memory as
needed; without it, rbfs will limit its consumption to some
arbitrary amount, currently 768MB (enough to hold a CD
image).
This program is useful mainly as an example of how to write
a user-level file server. It can also be used to provide
high-performance temporary files.
SOURCE
/sys/src/cmd/rbfs
SEE ALSO
bind(2), ramfs(4),
Page 1 Plan 9 (printed 12/21/24)
RBFS(4) RBFS(4)
BUGS
Should be integrated with ramfs. Should provide per-file
limits.
Page 2 Plan 9 (printed 12/21/24)