OOPS(8) OOPS(8)
NAME
oops - process status
SYNOPSIS
oops [ acegklstuvwx# ] [ core ] [ swap ] [ namelist ]
DESCRIPTION
Oops prints information about processes. Normally, only the
invoker's processes are candidates to be printed by oops.
Specifying a causes other user's processes to be candidates
to be printed; specifying x includes processes without con-
trol terminals in the candidate pool.
All output formats include, for each process, the process id
PID, control terminal of the process TT, cpu time used by
the process TIME (this includes both user and system time),
the state STAT of the process, and an indication of the com-
mand which is running. The state is given by a sequence of
four letters, e.g. ``RWNA''. The first letter indicates the
runnability of the process: R for a runnable process, T for
a stopped process, P for a process in page wait, D for one
in disk (or other short term) wait, S for one sleeping for
less than about 20 seconds, and I for an idle (sleeping
longer than about 20 seconds) process. The second letter is
W if the process is swapped out, blank if it is resident in
memory. The third letter is a blank if the process is run-
ning with normal scheduling priority, N if the priority is
reduced (nice), < if the priority is raised. The final let-
ter is an A if special paging behaviour has been requested.
Here are the options:
a asks for information about all processes with terminals
(ordinarily only one's own processes are displayed).
c causes the system's internal idea of the command name
to be used, rather than the command arguments in the
process' address space. This is more reliable, if less
informative, since the process is free to destroy the
latter information.
e asks for the environment (environ(7)) to be printed.
g includes process group leaders (usually shells) which
are normally excluded.
k causes the file /vmcore is used in place of /dev/kmem
and /dev/mem. This is used for postmortem system
debugging.
OOPS(8) OOPS(8)
l asks for a long listing, with fields PPID, CP, PRI, NI,
ADDR, SIZE, RSS and WCHAN as described below.
s adds the size SSIZ of the kernel stack of each process
(for use by system maintainers) to the basic output
format.
tx restricts output to processes whose controlling tty is
x (which should be specified as printed by oops, e.g.
t3 for tty3, tco for console, td0 for ttyd0, t? for
processes with no tty, etc). This option must be the
last one given.
u allegedly user oriented output is produced. This
includes fields USER, %CPU, NICE, SIZE, and RSS as
described below.
v virtual memory statistics; includes fields RE, SL,
PAGEIN, SIZE, RSS, SRS, TSIZ, TRS, %CPU and %MEM,
described below.
w use a wide output format (132 columns rather than 80);
if repeated, use arbitrarily wide output. This infor-
mation is used to decide how much of long argument
lists to print.
x asks about even processes with no terminal.
# A process number may be given, in which case the output
is restricted to that process. This option must also
be last.
A second argument tells oops where to look for core if the k
option is given, instead of /vmcore. A third argument is the
name of a swap file to use instead of the default /dev/drum.
If a fourth argument is given, it is taken to be the file
containing the system's namelist. Otherwise, /unix is used.
Fields which are not common to all output formats:
USER the owner of the process
%CPU cpu utilization of the process; this is a decaying
average over up to a minute of previous (real)
time. Since the time base over which this is com-
puted varies (since processes may be very young)
it is possible for the sum of all %CPU fields to
exceed 100%.
NICE (or NI) process scheduling increment (see nice(2))
SIZE virtual size of the process (in 1024 byte units)
RSS real memory (resident set) size of the process (in
1024 byte units)
SRS number of real memory pages (RSS) if and when
swapped
OOPS(8) OOPS(8)
TSIZ size of text (shared program) image
TRS size of resident (real memory) set of text
%MEM percentage of real memory used by this process.
RE residency time of the process (seconds in core)
SL sleep time of the process (seconds blocked)
PAGEIN number of disk i/o's resulting from references by
the process to pages not loaded in core.
UID numerical user-id of process owner
PPID numerical id of parent of process
CP short-term cpu utilization factor (used in
scheduling)
PRI process priority (non-positive when in non-
interruptible wait)
ADDR swap address of the process
WCHAN event on which process is waiting (an address in
the system), with the initial part of the address
trimmed off; e.g. 0x80004000 prints as 4000.
F flags associated with process as in
/usr/include/sys/proc.h:
SLOAD 000001 in core
SSYS 000002 swapper or pager process
SLOCK 000004 process being swapped out
SSWAP 000008 save area flag
STRC 000010 process is being traced
SWTED 000020 another tracing flag
SULOCK 000040 user settable lock in core
SPAGE 000080 process in page wait state
SKEEP 000100 another flag to prevent swap out
SDLYU 000200 delayed unlock of pages
SWEXIT 000400 working on exiting
SPHYSIO 000800 doing physical i/o (bio.c)
SPAGI 008000 init data space on demand, from inode
SSEQL 010000 user warned of sequential vm behavior
SUANOM 020000 user warned of anomalous vm behavior
STIMO 040000 timing out during sleep
SDETACH 080000 detached inherited by init
SNUSIG 100000 using new signal mechanism
A process that has exited and has a parent, but has not yet
been waited for by the parent is marked <defunct>; a process
which is blocked trying to exit is marked <exiting>. Oops
makes an educated guess as to the file name and arguments
given when the process was created by examining memory or
the swap area. The method is inherently somewhat unreliable
and in any event a process is entitled to destroy this
information, so the names cannot be counted on too much.
FILES
/unix system namelist
/dev/kmem kernel memory
/dev/drum swap device
OOPS(8) OOPS(8)
/vmcore core file
/dev searched to find swap device and tty names
SEE ALSO
kill(1), ps(1), pstat(8)
BUGS
Things can change while oops is running; the picture it
gives is only a close approximation to reality.
Oops duplicates ps(1); it is useful mainly for examining
kernel crash dumps, or when /proc is not mounted.