Linux and UNIX Man Pages

Linux & Unix Commands - Search Man Pages

prctl(1) [suse man page]

PRCTL(1)						      General Commands Manual							  PRCTL(1)

NAME
prctl - Process operations SYNOPSIS
prctl [-v] [-h|--help] [--version] <-q|<options....>> [command] where <options> are: --unaligned=[silent|signal|default] --fpemu=[silent|signal|default] DESCRIPTION
prctl allows you to query or control certain process behavior. Supported options are handling of unaligned memory accesses and floating point assistance faults by a process. When a process performs an unaligned memory access, by default the kernel would emulate the unaligned access correctly and log the unaligned access in syslog. This behavior can be changed so the kernel could either emulate the unaligned access correctly without logging an error or send SIGBUS to the process. Similarly when a process encounters a floating point assist fault, kernel would invoke floating point emulator and log the floating point assist fault. This behavior can be changed so the ker- nel could either emulate floating point operation without logging an error or send SIGFPE to the offending process. prctl can optionally be followed by a command. If a command is specified, prctl will exec the command with the processor behavior set to as specified by the options. If a command is not specified, prctl will fork a new shell unless the command only queried the current settings. The shell started by prctl will be as defined by the environment variable SHELL. If environment variable SHELL is not defined, shell in the password entry for the user will be started. If a shell is not found in the password entry, bash will be started. OPTIONS -v Verbose mode. In verbose mode, any new settings are confirmed with a message on stdout. --help Print usage information and exit. --version Print software version and exit. -q Query the current settings for the process options controllable by prctl. --unaligned=[silent|signal|default]] Set unaligned memory access behavior to not log the access (silent), send SIGBUS to the process (signal), or do the default (default). If a value is not specified after "=", current setting is returned. --fpemu=[silent|signal|default]] Set floating point assist fault behavior to not log the faulti (silent), send SIGFPE to the process (signal), or do the default (default). If a value is not specified after "=", current setting is returned. If an option is specified multiple times, the last one takes effect. EXAMPLES
prctl --unaligned=signal starts up a shell (as defined by the environment variable SHELL) and sets up any process running under this shell to be sent SIGBUS upon an unaligned memory access. prctl --unaligned=signal gdb tst starts up a gdb session for the program "tst" with the process set to receive SIGBUS upon unligned memory access. prctl --unaligned= displays the current setting for unaligned memory access handling. prctl --unaligned= gdb tst displays the current setting for unaligned memory access handling and starts up a gdb session for the program "tst". NOTES
prctl works on 2.4.0 and above kernels only. --fpemu= option is specific to IPF (Itanium Processor Family, previously known as IA-64) and is supported on kernels 2.4.11 and above. Other architectures and kernels may return "Invalid argument" error. prctl(2) AUTHOR
Khalid Aziz <khalid_aziz@hp.com> LICENSE
This software is made available under the GNU General Public License (GPL) Version 2. This software comes with NO WARRANTY. Process operations PRCTL(1)

Check Out this Related Man Page

prctl(1)                                                                                                                                  prctl(1)

NAME
prctl - get or set the resource controls of running processes, tasks, and projects SYNOPSIS
prctl [-P] [-t [basic | privileged | system] ] [ -n name [-srx] [-v value] [-e | -d action] [-p pid]] [-i idtype] id... The prctl utility allows the examination and modification of the resource controls associated with an active process, task, or project on the system. It allows access to the basic and privileged limits on the specified entity. See resource_controls(5) for a description of the resource controls supported in the current release of the Solaris operating system. If none of the -s, -r, -x, -v, -d, or -e options are specified, the invocation is considered a get operation. Otherwise, it is considered a modify operation. The following options are supported: -d | -e action Disables (-d) or enables (-e) the specified action on the resource control value specified by -v, -t, and -p. If any of the -v, -t, or -p options are unspecified, they match any value, privilege, or recipient pid. For example, specifying only -v modifies the first resource control with matching value, matching any privilege and recipient pid. If no matching resource control value is found, a new value is added as if -s were specified. Actions: all This action is only available with -d. It disables all actions. This fails on resource control values that have the deny global flag. deny Indicates that the resource control attempts to deny granting the resource to the process, task, project, or zone on a request for resources in excess of the resource control value. deny actions can not be enabled if the resource control has the no-deny global flag. deny actions can not be disabled if the resource control has the deny global flag. signal This action is only available with -d. It deactivates the signal action. signal=signum In the signal=signum action, signum is a signal number (or string representation of a signal). Setting a signal action on a resource control with the no-local-action global flag fails. A limited set of signals can be sent. See NOTES for additional details. -i idtype Specifies the type of the id operands. Valid idtypes are process, task, project, or zone. Also allowed are pid, taskid, projid, and zoneid. The default id type, if the -i option is omitted, is process. For a modify operation, the entity to which id operands are members is the target entity. For instance, setting a project resource con- trol on an -i process sets the resource control on the project to which each given process argument is a member. For a get operation, the resource controls are listed for all entities to which the id operands are members. For example, -i task taskid lists the task, project, and zone resource controls for the task, and for the project and zone to which that task is a member. -n name Specifies the name of the resource control to get or set. If the name is unspecified, all resource controls are retrieved. -p pid When manipulating (using -s, -r, -x, -d, or -e) a basic task project, or zone resource control values, a recipient pid can be specified using -p. When setting a new basic resource control or controls on a task, project, or zone, the -p option is required if the -i idtype option argument is not process. -P Display resource control values in semi-colon delimited format. -r Replaces the first resource control value (matching with the -t privilege) with the new value specified through the -v option. -s Set a new resource control value. This option requires the -v option. If you do not specify the -t option, basic privilege is used. If you want to set a basic task, process, or zone rctl, -p is required. If -e or -d are also specified, the action on the new rctl is set as well. For compatibility with prior releases, this option is implied if -v is specified, without any of -e, -d, -r, or -x. See resource_controls(5) for a description of unit modifiers and scaling factors you can use to express large values when setting a resource control value. -t [ basic | privileged | system ] Specifies which resource control type to set. Unless the "lowerable" flag is set for a resource control, only invocations by users (or setuid programs) who have privileges equivalent to those of root can modify privileged resource controls. See rctlblk_set_value(3C) for a description of the RCTL_GLOBAL_LOWERABLE flag. If the type is not specified, basic is assumed. For a get operation, the values of all resource control types, including system, are displayed if no type is specified. -v value Specifies the value for the resource control for a set operation. If no value is specified, then the modification (deletion, action enabling or disabling) is carried out on the lowest-valued resource control with the given type. -x Deletes the specified resource control value. If the delete option is not provided, the default operation of prctl is to modify a resource control value of matching value and privilege, or insert a new value with the given privilege. The matching criteria are dis- cussed more fully in setrctl(2). If none of the -d, -e, -v, or -x options is specified, the invocation is considered a get operation. The following operand is supported: id The ID of the entity (process, task, project, or zone) to interrogate. If the invoking user's credentials are unprivileged and the entity being interrogated possesses different credentials, the operation fails. If no id is specified, an error message is returned. Example 1: Displaying Current Resource Control Settings The following example displays current resource control settings for a task to which the current shell belongs: example$ ps -o taskid -p $$ TASKID 8 example$ prctl -i task 8 136150: /bin/ksh NAME PRIVILEGE VALUE FLAG ACTION RECIPIENT task.max-cpu-time system 18.4Es inf none - task.max-lwps system 2.15G max deny - project.max-contracts privileged 10.0K - deny - project.max-device-locked-memory privileged 127MB - deny - project.max-port-ids privileged 8.19K - deny - project.max-shm-memory privileged 508MB - deny - project.max-shm-ids privileged 128 - deny - project.max-msg-ids privileged 128 - deny - project.max-sem-ids privileged 128 - deny - project.max-crypto-memory privileged 508MB - deny - project.max-tasks system 2.15G max deny - project.max-lwps system 2.15G max deny - project.cpu-shares privileged 1 - none - zone.max-lwps system 2.15G max deny - zone.cpu-shares privileged 1 - none - Example 2: Displaying, Replacing, and Verifying the Value of a Specific Control The following examples displays, replaces, and verifies the value of a specific control on an existing project: example# prctl -n project.cpu-shares -i project group.staff project: 10: group.staff NAME PRIVILEGE VALUE FLAG ACTION RECIPIENT project.cpu-shares privileged 1 - none - system 65.5K max none - example# prctl -n project.cpu-shares -v 10 -r -i project group.staff example# prctl -n project.cpu-shares -i project group.staff project: 10: group.staff NAME PRIVILEGE VALUE FLAG ACTION RECIPIENT project.cpu-shares privileged 10 - none - system 65.5K max none - Example 3: Adjusting Resources This example uses the project.max-device-locked-memory resource. First, use id -p to find out the project id: /home/garfield> id -p uid=77880(garfield) gid=10(staff) projid=10(group.staff) Identify the resource limit value before the change. Find a process id of the processes in that project id: /home/garfield> ps -eaf UID PID PPID C STIME TTY TIME CMD ..... garfield 530 528 0 11:24:18 pts/1 0:00 -ksh /home/garfield> prctl 530 530: -ksh project.max-device-locked-memory [ no-basic deny ] 261969408 privileged deny 18446744073709551615 system deny [ max ] .... current limit is 261969408 bytes. Next, adjust the project.max-device-locked-memory limit to 300000000 for project id 10: # prctl -n project.max-device-locked-memory -v 300000000 -r -i project 10 The resource limit value after the change shows a new value of 300000000 bytes: /home/garfield> prctl 530 530: -ksh project.max-device-locked-memory [ no-basic deny ] 3 00000000 privileged deny 18446744073709551615 system deny [ max ] The following exit values are returned: 0 Success. 1 Fatal error encountered. 2 Invalid command line options were specified. /proc/pid/* Process information and control files See attributes(5) for descriptions of the following attributes: +-----------------------------+-----------------------------+ | ATTRIBUTE TYPE | ATTRIBUTE VALUE | +-----------------------------+-----------------------------+ |Availability |SUNWesu | +-----------------------------+-----------------------------+ |Interface Stability |See below. | +-----------------------------+-----------------------------+ The command-line syntax is Evolving. The human-readable output is Unstable. The parseable output is Evolving. rctladm(1M), setrctl(2), rctlblk_get_local_action(3C), attributes(5), resource_controls(5) The valid signals that can be set on a resource control block allowing local actions are SIGABRT, SIGXRES, SIGHUP, SIGSTOP, SIGTERM, and SIGKILL. Additionally, CPU time related controls can issue the SIGXCPU signal, and file size related controls can send the SIGXFSZ signal. 18 Aug 2005 prctl(1)
Man Page