freebsd-dev/lib/libc/sys/mlockall.2
Mark Johnston 54a3a11421 Provide separate accounting for user-wired pages.
Historically we have not distinguished between kernel wirings and user
wirings for accounting purposes.  User wirings (via mlock(2)) were
subject to a global limit on the number of wired pages, so if large
swaths of physical memory were wired by the kernel, as happens with
the ZFS ARC among other things, the limit could be exceeded, causing
user wirings to fail.

The change adds a new counter, v_user_wire_count, which counts the
number of virtual pages wired by user processes via mlock(2) and
mlockall(2).  Only user-wired pages are subject to the system-wide
limit which helps provide some safety against deadlocks.  In
particular, while sources of kernel wirings typically support some
backpressure mechanism, there is no way to reclaim user-wired pages
shorting of killing the wiring process.  The limit is exported as
vm.max_user_wired, renamed from vm.max_wired, and changed from u_int
to u_long.

The choice to count virtual user-wired pages rather than physical
pages was done for simplicity.  There are mechanisms that can cause
user-wired mappings to be destroyed while maintaining a wiring of
the backing physical page; these make it difficult to accurately
track user wirings at the physical page layer.

The change also closes some holes which allowed user wirings to succeed
even when they would cause the system limit to be exceeded.  For
instance, mmap() may now fail with ENOMEM in a process that has called
mlockall(MCL_FUTURE) if the new mapping would cause the user wiring
limit to be exceeded.

Note that bhyve -S is subject to the user wiring limit, which defaults
to 1/3 of physical RAM.  Users that wish to exceed the limit must tune
vm.max_user_wired.

Reviewed by:	kib, ngie (mlock() test changes)
Tested by:	pho (earlier version)
MFC after:	45 days
Sponsored by:	Netflix
Differential Revision:	https://reviews.freebsd.org/D19908
2019-05-13 16:38:48 +00:00

147 lines
4.6 KiB
Groff

.\" $NetBSD: mlockall.2,v 1.11 2003/04/16 13:34:54 wiz Exp $
.\"
.\" Copyright (c) 1999 The NetBSD Foundation, Inc.
.\" All rights reserved.
.\"
.\" This code is derived from software contributed to The NetBSD Foundation
.\" by Jason R. Thorpe of the Numerical Aerospace Simulation Facility,
.\" NASA Ames Research Center.
.\"
.\" Redistribution and use in source and binary forms, with or without
.\" modification, are permitted provided that the following conditions
.\" are met:
.\" 1. Redistributions of source code must retain the above copyright
.\" notice, this list of conditions and the following disclaimer.
.\" 2. Redistributions in binary form must reproduce the above copyright
.\" notice, this list of conditions and the following disclaimer in the
.\" documentation and/or other materials provided with the distribution.
.\"
.\" THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS
.\" ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
.\" TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
.\" PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS
.\" BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
.\" CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
.\" SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
.\" INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
.\" CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
.\" ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
.\" POSSIBILITY OF SUCH DAMAGE.
.\"
.\" $FreeBSD$
.\"
.Dd May 13, 2019
.Dt MLOCKALL 2
.Os
.Sh NAME
.Nm mlockall ,
.Nm munlockall
.Nd lock (unlock) the address space of a process
.Sh LIBRARY
.Lb libc
.Sh SYNOPSIS
.In sys/mman.h
.Ft int
.Fn mlockall "int flags"
.Ft int
.Fn munlockall "void"
.Sh DESCRIPTION
The
.Fn mlockall
system call locks into memory the physical pages associated with the
address space of a process until the address space is unlocked, the
process exits, or execs another program image.
.Pp
The following flags affect the behavior of
.Fn mlockall :
.Bl -tag -width ".Dv MCL_CURRENT"
.It Dv MCL_CURRENT
Lock all pages currently mapped into the process's address space.
.It Dv MCL_FUTURE
Lock all pages mapped into the process's address space in the future,
at the time the mapping is established.
Note that this may cause future mappings to fail if those mappings
cause resource limits to be exceeded.
.El
.Pp
Since physical memory is a potentially scarce resource, processes are
limited in how much they can lock down.
A single process can lock the minimum of a system-wide
.Dq wired pages
limit
.Va vm.max_user_wired
and the per-process
.Dv RLIMIT_MEMLOCK
resource limit.
.Pp
If
.Va security.bsd.unprivileged_mlock
is set to 0 these calls are only available to the super-user.
If
.Va vm.old_mlock
is set to 1 the per-process
.Dv RLIMIT_MEMLOCK
resource limit will not be applied for
.Fn mlockall
calls.
.Pp
The
.Fn munlockall
call unlocks any locked memory regions in the process address space.
Any regions mapped after an
.Fn munlockall
call will not be locked.
.Sh RETURN VALUES
A return value of 0 indicates that the call
succeeded and all pages in the range have either been locked or unlocked.
A return value of \-1 indicates an error occurred and the locked
status of all pages in the range remains unchanged.
In this case, the global location
.Va errno
is set to indicate the error.
.Sh ERRORS
.Fn mlockall
will fail if:
.Bl -tag -width Er
.It Bq Er EINVAL
The
.Fa flags
argument is zero, or includes unimplemented flags.
.It Bq Er ENOMEM
Locking the indicated range would exceed either the system or per-process
limit for locked memory.
.It Bq Er EAGAIN
Some or all of the memory mapped into the process's address space
could not be locked when the call was made.
.It Bq Er EPERM
The calling process does not have the appropriate privilege to perform
the requested operation.
.El
.Sh SEE ALSO
.Xr mincore 2 ,
.Xr mlock 2 ,
.Xr mmap 2 ,
.Xr munmap 2 ,
.Xr setrlimit 2
.Sh STANDARDS
The
.Fn mlockall
and
.Fn munlockall
functions are believed to conform to
.St -p1003.1-2001 .
.Sh HISTORY
The
.Fn mlockall
and
.Fn munlockall
functions first appeared in
.Fx 5.1 .
.Sh BUGS
The per-process and system-wide resource limits of locked memory apply
to the amount of virtual memory locked, not the amount of locked physical
pages.
Hence two distinct locked mappings of the same physical page counts as
2 pages aginst the system limit, and also against the per-process limit
if both mappings belong to the same physical map.