2 .\" Copyright (C) 2001 Jason Evans <jasone@FreeBSD.org>. All rights reserved.
4 .\" Redistribution and use in source and binary forms, with or without
5 .\" modification, are permitted provided that the following conditions
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice(s), this list of conditions and the following disclaimer as
9 .\" the first lines of this file unmodified other than the possible
10 .\" addition of one or more copyright notices.
11 .\" 2. Redistributions in binary form must reproduce the above copyright
12 .\" notice(s), this list of conditions and the following disclaimer in the
13 .\" documentation and/or other materials provided with the distribution.
15 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY
16 .\" EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
17 .\" WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
18 .\" DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY
19 .\" DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
20 .\" (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
21 .\" SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
22 .\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
23 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
24 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
54 .Nd kernel shared/exclusive lock
60 .Fn sx_init "struct sx *sx" "const char *description"
62 .Fn sx_init_flags "struct sx *sx" "const char *description" "int opts"
64 .Fn sx_destroy "struct sx *sx"
66 .Fn sx_slock "struct sx *sx"
68 .Fn sx_xlock "struct sx *sx"
70 .Fn sx_slock_sig "struct sx *sx"
72 .Fn sx_xlock_sig "struct sx *sx"
74 .Fn sx_try_slock "struct sx *sx"
76 .Fn sx_try_xlock "struct sx *sx"
78 .Fn sx_sunlock "struct sx *sx"
80 .Fn sx_xunlock "struct sx *sx"
82 .Fn sx_unlock "struct sx *sx"
84 .Fn sx_try_upgrade "struct sx *sx"
86 .Fn sx_downgrade "struct sx *sx"
88 .Fn sx_sleep "void *chan" "struct sx *sx" "int priority" "const char *wmesg" "int timo"
90 .Fn sx_xholder "struct sx *sx"
92 .Fn sx_xlocked "const struct sx *sx"
94 .Cd "options INVARIANTS"
95 .Cd "options INVARIANT_SUPPORT"
97 .Fn sx_assert "const struct sx *sx" "int what"
99 .Fn SX_SYSINIT "name" "struct sx *sx" "const char *desc"
100 .Fn SX_SYSINIT_FLAGS "name" "struct sx *sx" "const char *desc" "int flags"
102 Shared/exclusive locks are used to protect data that are read far more often
103 than they are written.
104 Shared/exclusive locks do not implement priority propagation like mutexes and
105 reader/writer locks to prevent priority inversions, so
106 shared/exclusive locks should be used prudently.
108 Shared/exclusive locks are created with either
114 is a pointer to space for a
118 is a pointer to a null-terminated character string that describes the
119 shared/exclusive lock.
124 specifies a set of optional flags to alter the behavior of
126 It contains one or more of the following flags:
127 .Bl -tag -width SX_NOADAPTIVE
129 Disable adaptive spinning, rather than sleeping, for lock operations
130 while an exclusive lock holder is executing on another CPU.
131 Adaptive spinning is the default unless the kernel is compiled with
132 .Cd "options NO_ADAPTIVE_SX" .
134 Witness should not log messages about duplicate locks being acquired.
140 Do not profile this lock.
142 Allow threads to recursively acquire exclusive locks for
145 Do not log any operations for this lock via
148 If the kernel has been compiled with
149 .Cd "options INVARIANTS" ,
153 has not been initialized multiple times without intervening calls to
155 unless this option is specified.
158 Shared/exclusive locks are destroyed with
162 must not be locked by any thread when it is destroyed.
164 Threads acquire and release a shared lock by calling
173 Threads acquire and release an exclusive lock by calling
182 A thread can attempt to upgrade a currently held shared lock to an exclusive
185 A thread that has an exclusive lock can downgrade it to a shared lock by
192 will return 0 if the shared/exclusive lock cannot be acquired immediately;
193 otherwise the shared/exclusive lock will be acquired and a non-zero value will
197 will return 0 if the shared lock cannot be upgraded to an exclusive lock
198 immediately; otherwise the exclusive lock will be acquired and a non-zero value
204 do the same as their normal versions but performing an interruptible sleep.
205 They return a non-zero value if the sleep has been interrupted by a signal
206 or an interrupt, otherwise 0.
208 A thread can atomically release a shared/exclusive lock while waiting for an
211 For more details on the parameters to this function,
216 .Cd "options INVARIANTS"
218 .Cd "options INVARIANT_SUPPORT" ,
223 for the assertions specified in
225 and panics if they are not met.
226 One of the following assertions must be specified:
227 .Bl -tag -width ".Dv SA_UNLOCKED"
229 Assert that the current thread has either a shared or an exclusive lock on the
231 lock pointed to by the first argument.
233 Assert that the current thread has a shared lock on the
238 Assert that the current thread has an exclusive lock on the
241 by the first argument.
243 Assert that the current thread has no lock on the
246 by the first argument.
249 In addition, one of the following optional assertions may be included with
256 .Bl -tag -width ".Dv SA_NOTRECURSED"
258 Assert that the current thread has a recursed lock on
260 .It Dv SA_NOTRECURSED
261 Assert that the current thread does not have a recursed lock on
266 will return a pointer to the thread which currently holds an exclusive lock on
268 If no thread holds an exclusive lock on
275 will return non-zero if the current thread holds the exclusive lock;
276 otherwise, it will return zero.
278 For ease of programming,
280 is provided as a macro frontend to the respective functions,
284 Algorithms that are aware of what state the lock is in should use either
285 of the two specific functions for a minor performance benefit.
289 macro is used to generate a call to the
291 routine at system startup in order to initialize a given
294 The parameters are the same as
296 but with an additional argument,
298 that is used in generating unique variable names for the related
299 structures associated with the lock and the sysinit routine.
302 macro can similarly be used to initialize a given
307 A thread may not hold both a shared lock and an exclusive lock on the same
309 attempting to do so will result in deadlock.
311 A thread may hold a shared or exclusive lock on an
316 lock may not be acquired while holding a mutex.
317 Otherwise, if one thread slept while holding an
319 lock while another thread blocked on the same
321 lock after acquiring a mutex, then the second thread would effectively
322 end up sleeping while holding a mutex, which is not allowed.
333 cannot assert whether the current thread does or does not hold a shared lock.
339 thread holds a shared lock.
340 They cannot ensure that the current thread holds a shared lock.
343 can only assert that the current thread does not hold an exclusive lock.