1.\" 2.\" Copyright (c) 1996 Joerg Wunsch 3.\" 4.\" All rights reserved. 5.\" 6.\" Redistribution and use in source and binary forms, with or without 7.\" modification, are permitted provided that the following conditions 8.\" are met: 9.\" 1. Redistributions of source code must retain the above copyright 10.\" notice, this list of conditions and the following disclaimer. 11.\" 2. Redistributions in binary form must reproduce the above copyright 12.\" notice, this list of conditions and the following disclaimer in the 13.\" documentation and/or other materials provided with the distribution. 14.\" 15.\" THIS SOFTWARE IS PROVIDED BY THE DEVELOPERS ``AS IS'' AND ANY EXPRESS OR 16.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES 17.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. 18.\" IN NO EVENT SHALL THE DEVELOPERS BE LIABLE FOR ANY DIRECT, INDIRECT, 19.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT 20.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, 21.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY 22.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT 23.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF 24.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. 25.\" 26.\" $FreeBSD: src/share/man/man9/sleep.9,v 1.18.2.5 2001/12/17 11:30:19 ru Exp $ 27.\" $DragonFly: src/share/man/man9/sleep.9,v 1.2 2003/06/17 04:37:01 dillon Exp $ 28.\" " 29.Dd December 17, 1998 30.Os 31.Dt SLEEP 9 32.Sh NAME 33.Nm sleep , 34.Nm tsleep , 35.Nm asleep , 36.Nm await , 37.Nm wakeup 38.Nd wait for events 39.Sh SYNOPSIS 40.In sys/param.h 41.In sys/systm.h 42.In sys/proc.h 43.Ft int 44.Fn tsleep "void *ident" "int priority" "const char *wmesg" "int timo" 45.Ft int 46.Fn asleep "void *ident" "int priority" "const char *wmesg" "int timo" 47.Ft int 48.Fn await "int priority" "int timo" 49.Ft void 50.Fn wakeup "void *ident" 51.Ft void 52.Fn wakeup_one "void *ident" 53.Sh DESCRIPTION 54The functions 55.Fn tsleep 56and 57.Fn wakeup 58handle event-based process blocking. If a process must wait for an 59external event, it is put on sleep by 60.Nm tsleep . 61The parameter 62.Ar ident 63is an arbitrary address that uniquely identifies the event on which 64the process is being asleep. All processes sleeping on a single 65.Ar ident 66are woken up later by 67.Nm wakeup , 68often called from inside an interrupt routine, to indicate that the 69resource the process was blocking on is available now. 70.Pp 71The parameter 72.Ar wmesg 73is a string describing the sleep condition for tools like 74.Xr ps 1 . 75Due to the limited space of those programs to display arbitrary strings, 76this message should not be longer than 6 characters. 77.Pp 78The 79.Fn wakeup_one 80function is used to make the first process in the queue that is 81sleeping on the parameter 82.Fa ident 83runnable. This can prevent the system from becoming saturated 84when a large number of processes are sleeping on the same address, 85but only one of them can actually do any useful work when made 86runnable. 87.Pp 88.Nm Tsleep 89is the general sleep call. Suspends the current process until a wakeup is 90performed on the specified identifier. The process will then be made 91runnable with the specified 92.Ar priority . 93Sleeps at most 94.Ar timo 95\&/ hz seconds (0 means no timeout). If 96.Ar pri 97includes the 98.Dv PCATCH 99flag, signals are checked before and after sleeping, else signals are 100not checked. Returns 0 if awakened, 101.Er EWOULDBLOCK 102if the timeout expires. If 103.Dv PCATCH 104is set and a signal needs to be delivered, 105.Er ERESTART 106is returned if the current system call should be restarted if 107possible, and 108.Er EINTR 109is returned if the system call should be interrupted by the signal 110(return 111.Er EINTR ) . 112.Pp 113.Nm Asleep 114implements the new asynchronous sleep function. It takes the same arguments 115as 116.Fn tsleep 117and places the process on the appropriate wait queue, but 118.Fn asleep 119leaves the process runnable and returns immediately. The caller is then 120expected to, at some point in the future, call 121.Fn await 122to actually wait for the previously queued wait condition. 123If 124.Fn asleep 125is called several times, only the most recent call is effective. 126.Fn asleep 127may be called with an 128.Ar ident 129value of NULL 130to remove any previously queued condition. 131.Pp 132.Nm Await 133implements the new asynchronous wait function. When 134.Fn asleep 135is called on an identifier it associates the process with that 136identifier but does not block. 137.Fn await 138will actually block the process until 139.Fn wakeup 140is called on that identifier any time after the 141.Fn asleep . 142If 143.Fn wakeup 144is called after you 145.Fn asleep 146but before you 147.Fn await 148then the 149.Fn await 150call is effectively a NOP. 151If 152.Fn await 153is called multiple times without an intervening 154.Fn asleep , 155the 156.Fn await 157is effectively a NOP but will also call 158.Fn mswitch 159for safety. The 160.Fn await 161function allows you to override the priority and timeout values to be used. 162If the value -1 is specified for an argument, the value is taken from the 163previous 164.Fn asleep 165call. If -1 is passed for the priority you must be prepared to catch signal 166conditions if the prior call to 167.Fn asleep 168specified it in its priority. If -1 is passed for the timeout you must be 169prepared to catch a timeout condition if the prior call to 170.Fn asleep 171specified a timeout. When you use -1, it is usually a good idea to not make 172assumptions as to the arguments used by the prior 173.Fn asleep 174call. 175.Pp 176The 177.Fn asleep 178and 179.Fn await 180functions are mainly used by the kernel to shift the burden of blocking 181away from extremely low level routines and to push it onto their callers. 182This in turn allows more complex interlocking code to 183.Em backout 184of a temporary resource failure 185(such as lack of memory) in order to release major locks prior to actually 186blocking, and to then retry the operation on wakeup. This key feature is 187expected to be heavily used in SMP situations in order to allow code to make 188better use of spinlocks. A spinlock, by its very nature, cannot be used 189around code that might block. It is hoped that these capabilities will 190make it easier to migrate the SMP master locks deeper into the kernel. 191.Pp 192These routines may also be used to avoid nasty spl*() calls to get around 193race conditions with simple conditional test/wait interlocks. You simply 194call 195.Fn asleep 196prior to your test, then conditionally 197.Fn await 198only if the test fails. It is usually a good idea to cancel an 199.Fn asleep 200if you wind up never calling the related 201.Fn await , 202but it is not required. If you do not want to waste cpu calling 203.Fn asleep 204unnecessarily, you can surround the whole thing with a second test. The 205race condition is still handled by the inside 206.Fn asleep 207call. 208.Sh RETURN VALUES 209See above. 210.Sh SEE ALSO 211.Xr ps 1 , 212.Xr malloc 9 213.Sh HISTORY 214The sleep/wakeup process synchronization mechanism is very old. It 215appeared in a very early version of Unix. 216.Pp 217.Nm Tsleep 218appeared in 219.Bx 4.4 . 220.Pp 221.Nm Asleep Ns / Ns Nm await 222first appeared in 223.Fx 3.0 224and is designed to shift the burden of blocking 225away from extremely low level routines and push it up to their callers. 226.Pp 227.Nm Sleep 228used to be the traditional form. It doesn't let you specify a timeout or a 229.Ar wmesg , 230hence it has been discontinued. 231.Sh AUTHORS 232.An -nosplit 233This man page was written by 234.An J\(:org Wunsch . 235.Nm Asleep 236and 237.Nm await 238were designed and written by 239.An Matthew Dillon . 240