1.\" 2.\" Copyright (c) 2014 Markus Pfeiffer 3.\" All rights reserved. 4.\" 5.\" Redistribution and use in source and binary forms, with or without 6.\" modification, are permitted provided that the following conditions 7.\" are met: 8.\" 1. Redistributions of source code must retain the above copyright 9.\" notice, this list of conditions and the following disclaimer. 10.\" 2. Redistributions in binary form must reproduce the above copyright 11.\" notice, this list of conditions and the following disclaimer in the 12.\" documentation and/or other materials provided with the distribution. 13.\" 14.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 15.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 18.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24.\" SUCH DAMAGE. 25.\" 26.\" 27.Dd June 5, 2014 28.Dt LOCKING 9 29.Os 30.Sh NAME 31.Nm locking 32.Nd introduction to kernel locking primitives 33.Sh DESCRIPTION 34The 35.Dx 36kernel provides several locking and synchronisation primitives, each with 37different characteristics and purposes. 38This manpage aims at giving an 39overview of the available locking primitives and their use cases, as well 40as pointers towards further information. 41.Ss Condition Variables 42Condition variables are used to wait for conditions to occur. 43In 44.Dx 45condition variables use a 46.Xr spinlock 9 47internally. 48Threads that wait on a condition variable are called waiters. 49Either just one or all waiters can be notified of changes to a 50condition variable. 51A condition variable can 52.Xr tsleep_interlock 9 53when given a 54.Xr lockmgr 9 55lock to avoid missing changes to it, or use regular 56.Xr tsleep 9 . 57.Pp 58See 59.Xr condvar 9 . 60.Ss Critical Sections 61A critical section changes the priority of the current thread to 62.Dv TDPRIT_CRIT , 63effectively avoiding preemption of the thread. 64Critical sections are a per-cpu primitive, and there is no synchronisation 65or locking between CPUs. 66.Pp 67See 68.Xr crit_enter 9 . 69.Ss Lockmgr Locks 70.Xr Lockmgr 9 71locks are the kitchen sink locking primitive for the 72.Dx 73kernel, and the most heavyweight locking mechanism. 74.Xr lockmgr 9 75locks can be shared/exclusive and recursive. 76Lockmgr locks should be used for 77.Fx 78compatibility when porting drivers that use 79.Fx Ap s 80mutexes. 81.Pp 82See 83.Xr lockmgr 9 . 84.Ss LWKT Messages 85LWKT messages can be used to pass messages between light weight kernel 86threads in the 87.Dx 88kernel. 89LWKT mesages are sent to message ports. Every light weight kernel thread 90possesses a message port, but more can be created if necessary. 91.Pp 92See 93.Xr msgport 9 . 94.Ss LWKT Serializers 95LWKT serializers provide a fast locked-bus-cycle-based serialization 96facility. 97They are used to serialize access to hardware and other subsystems. 98Serializers were designed to provide low level exclusive locks. 99.Pp 100See 101.Xr serializer 9 . 102.Ss LWKT Tokens 103LWKT tokens use 104.Xr atomic_cmpset 9 105internally and are integrated with the LWKT scheduler. 106The scheduler takes care of acquiring a token before 107rescheduling, so a thread will not be run unless all tokens for it can be 108acquired. 109Tokens are not owned by a thread, but by the CPU, and threads are only given 110references to tokens. 111.Pp 112See 113.Xr token 9 . 114.Ss MPLOCK 115The mplock is an API wrapper for the MP token. The use of this should be 116avoided at all cost, because there is only one MP token for the whole system. 117.Ss MTX Mutexes 118Mtx mutexes are a locking primitive that is based around 119.Xr atomic_cmpset_int 9 120instead of spinlocks. 121They are much faster and use less memory than 122.Xr lockmgr 9 123locks. 124Mtx mutexes can always be recursive, shared/exclusive and can be held 125across blocking calls and sleeps. 126They are also capable of passing ownership directly to a new owner 127without wakeup. 128.Pp 129See 130.Xr mutex 9 . 131.Ss Spinlocks 132Spinlocks employ a busy wait loop to acquire a lock. 133This means that this type of lock is very lightweight, 134but should only be held for a very short time, since all contenders 135will be spinning and not sleeping. 136No wakeup is necessary, because a waiter will be spinning already. 137If a thread tries to sleep while holding a spinlock, the kernel will panic. 138Spinlocks cannot recurse. 139.Pp 140They are mainly used to protect kernel structures, and to 141implement higher level locking primitives. 142.Pp 143See 144.Xr spinlock 9 . 145.Sh SEE ALSO 146.Xr atomic 9 , 147.Xr condvar 9 , 148.Xr crit_enter 9 , 149.Xr lockmgr 9 , 150.Xr mutex 9 , 151.Xr serializer 9 , 152.Xr spinlock 9 , 153.Xr tsleep 9 154.Sh AUTHORS 155.An -nosplit 156This manual page was written by 157.An Markus Pfeiffer Aq Mt markus.pfeiffer@morphism.de , 158based on comments by various 159.Dx 160authors. 161