pset_bind man page on NetBSD

Man page or keyword search:  
man Server   9087 pages
apropos Keyword Search (all sections)
Output format
NetBSD logo
[printable version]

PSET(3)			 BSD Library Functions Manual		       PSET(3)

NAME
     pset_create, pset_assign, pset_bind, pset_destroy — processor sets

SYNOPSIS
     #include <sys/pset.h>

     int
     pset_create(psetid_t *psid);

     int
     pset_assign(psetid_t psid, cpuid_t cpuid, psetid_t *opsid);

     int
     pset_bind(psetid_t psid, idtype_t type, id_t id, psetid_t *opsid);

     int
     pset_destroy(psetid_t psid);

DESCRIPTION
     The processor sets API provides the possibility to exclusively dedicate
     specific processors or groups of processors to processes or threads.
     After processes or threads are bound to a group of processors by the API,
     the group henceforth runs only those processes or threads.	 This section
     describes the functions used to control processor sets.

FUNCTIONS
     pset_create(psid)
	      Creates a processor set, and returns its ID into psid.

     pset_assign(psid, cpu, opsid)
	      Assigns the processor specified by cpuid to the processor set
	      specified by psid.  Stores the current processor set ID of the
	      processor or PS_NONE into opsid, if the pointer is not NULL.

	      The following actions can be specified:

		1.   If psid is set to PS_QUERY, then the current processor
		     set ID will be returned into psid, and no assignment will
		     be performed.

		2.   If psid is set to PS_MYID, then the processor set ID of
		     the calling process will be used, and psid will be
		     ignored.

		3.   If psid is set to PS_NONE, any assignment to the proces‐
		     sor will be cleared.

     pset_bind(psid, type, id, opsid)
	      Dedicates the processor set specified by psid to the target
	      specified by id.	The current processor set ID to which the tar‐
	      get is bound or PS_NONE will be returned in opsid, if the
	      pointer is not NULL.  NetBSD supports the following types of
	      targets specified by type:

		P_PID	 Process identified by the PID.

		P_LWPID	 Thread of the calling process indentified by the LID.

	      The following actions can be specified:

		1.   If psid is set to PS_QUERY, then the current processor
		     set ID to which the target is bound or PS_NONE will be
		     returned in opsid, and no binding will be performed.

		2.   If psid is set to PS_MYID, then the processor set ID of
		     the calling process will be used.

		3.   If psid is set to PS_NONE, the specified target will be
		     unbound from the processor set.

     pset_destroy(psid)
	      Destroys the processor set specified by psid.  Before destroying
	      the processor set, all related assignments of the processors
	      will be cleared, and all bound threads will be unbound.

	      If psid is PS_MYID, the processor set ID of the caller thread
	      will be used.

IMPLEMENTATION NOTES
     The pset_bind() function can return the current processor set ID to which
     the target is bound, or PS_NONE.  However, for example, the process may
     have many threads, which could be bound to different processor sets.  In
     such a case it is unspecified which thread will be used to return the
     information.

     There is an alternative thread affinity interface, see affinity(3).  How‐
     ever, processor sets and thread affinity are mutually exclusive, hence
     mixing of these interfaces is prohibited.

RETURN VALUES
     Upon successful completion these functions return 0.  Otherwise, -1 is
     returned and errno is set to indicate the error.

EXAMPLES
     An example of code fragment, which assigns the CPU whose ID is 0, for
     current process:

	     psetid_t psid;
	     cpuid_t ci = 0;

	     if (pset_create(&psid) < 0)
		     err(EXIT_FAILURE, "pset_create");

	     /* Assign CPU 0 to the processor-set */
	     if (pset_assign(psid, ci, NULL) < 0)
		     err(EXIT_FAILURE, "pset_assign");

	     /* Bind the current process to the processor-set */
	     if (pset_bind(psid, P_PID, P_MYID, NULL) < 0)
		     err(EXIT_FAILURE, "pset_bind");

	     /*
	      * At this point, CPU 0 runs only the current process.
	      */
	     perform_work();

	     if (pset_destroy(psid) < 0)
		     err(EXIT_FAILURE, "pset_destroy");

ERRORS
     The pset_create() function fails if:

     [ENOMEM]		No memory is available for creation of the processor
			set, or limit of the allowed count of the processor
			sets was reached.

     [EPERM]		The calling process is not the super-user.

     The pset_assign() function fails if:

     [EBUSY]		Another operation is performing on the processor set.

     [EINVAL]		psid or cpuid are invalid.

     [EPERM]		The calling process is not the super-user, and psid is
			not PS_QUERY.

     The pset_bind() function fails if:

     [EBUSY]		Another operation is performing on the processor set.

     [EINVAL]		psid or type are invalid.

     [EPERM]		The calling process is not the super-user, and psid is
			not PS_QUERY.

     [ESRCH]		The specified target was not found.

     The pset_destroy() function fails if:

     [EBUSY]		Another operation is performing on the processor set.

     [EPERM]		The calling process is not the super-user.

SEE ALSO
     affinity(3), cpuset(3), sched(3), schedctl(8)

STANDARDS
     This API is expected to be compatible with the APIs found in Solaris and
     HP-UX operating systems.

HISTORY
     The processor sets appeared in NetBSD 5.0.

BSD				  May 6, 2010				   BSD
[top]
                             _         _         _ 
                            | |       | |       | |     
                            | |       | |       | |     
                         __ | | __ __ | | __ __ | | __  
                         \ \| |/ / \ \| |/ / \ \| |/ /  
                          \ \ / /   \ \ / /   \ \ / /   
                           \   /     \   /     \   /    
                            \_/       \_/       \_/ 
More information is available in HTML format for server NetBSD

List of man pages available for NetBSD

Copyright (c) for man pages and the logo by the respective OS vendor.

For those who want to learn more, the polarhome community provides shell access and support.

[legal] [privacy] [GNU] [policy] [cookies] [netiquette] [sponsors] [FAQ]
Tweet
Polarhome, production since 1999.
Member of Polarhome portal.
Based on Fawad Halim's script.
....................................................................
Vote for polarhome
Free Shell Accounts :: the biggest list on the net