1 Userspace RCU Implementation
2 ============================
4 by Mathieu Desnoyers and Paul E. McKenney
10 ./bootstrap # skip if using tarball
18 - Forcing 32-bit build:
20 CFLAGS="-m32 -g -O2" ./configure
22 - Forcing 64-bit build:
24 CFLAGS="-m64 -g -O2" ./configure
26 - Forcing a 32-bit build with 386 backward compatibility:
28 CFLAGS="-m32 -g -O2" ./configure --host=i386-pc-linux-gnu
30 - Forcing a 32-bit build for Sparcv9 (typical for Sparc v9)
32 CFLAGS="-m32 -Wa,-Av9a -g -O2" ./configure
35 Architectures supported
36 -----------------------
38 Currently, the following architectures are supported:
40 - x86 (i386, i486, i586, i686)
57 - Linux all architectures
58 - FreeBSD 8.2/8.3/9.0/9.1/10.0 i386/amd64
70 (more testing needed before claiming support for these OS).
72 Linux ARM depends on running a Linux kernel 2.6.15 or better, GCC 4.4 or
75 The GCC compiler versions 3.3, 3.4, 4.0, 4.1, 4.2, 4.3, 4.4 and 4.5 are
76 supported, with the following exceptions:
78 - GCC 3.3 and 3.4 have a bug that prevents them from generating volatile
79 accesses to offsets in a TLS structure on 32-bit x86. These versions are
80 therefore not compatible with `liburcu` on x86 32-bit
81 (i386, i486, i586, i686).
82 The problem has been reported to the GCC community:
83 http://www.mail-archive.com/gcc-bugs@gcc.gnu.org/msg281255.html
84 - GCC 3.3 cannot match the "xchg" instruction on 32-bit x86 build.
85 See http://kerneltrap.org/node/7507
86 - Alpha, ia64 and ARM architectures depend on GCC 4.x with atomic builtins
87 support. For ARM this was introduced with GCC 4.4:
88 http://gcc.gnu.org/gcc-4.4/changes.html.
89 - Linux aarch64 depends on GCC 5.1 or better because prior versions
90 perform unsafe access to deallocated stack.
92 Clang version 3.0 (based on LLVM 3.0) is supported.
94 Building on MacOS X (Darwin) requires a work-around for processor
99 ./configure --build=i686-apple-darwin11
103 ./configure --build=x86_64-apple-darwin11
105 For developers using the Git tree:
107 This source tree is based on the autotools suite from GNU to simplify
108 portability. Here are some things you should have on your system in order to
109 compile the git repository tree :
111 - GNU autotools (automake >=1.12, autoconf >=2.69)
112 (make sure your system wide `automake` points to a recent version!)
114 (for more information, go to http://www.gnu.org/software/autoconf/)
116 If you get the tree from the repository, you will need to use the `bootstrap`
117 script in the root of the tree. It calls all the GNU tools needed to prepare
118 the tree configuration.
120 Test scripts provided in the `tests/` directory of the source tree depend
121 on `bash` and the `seq` program.
127 See the relevant API documentation files in `doc/`. The APIs provided by
128 Userspace RCU are, by prefix:
130 - `rcu_`: Read-Copy Update (see [`doc/rcu-api.md`](doc/rcu-api.md))
131 - `cmm_`: Concurrent Memory Model
132 - `caa_`: Concurrent Architecture Abstraction
133 - `cds_`: Concurrent Data Structures
134 (see [`doc/cds-api.md`](doc/cds-api.md))
135 - `uatomic_`: Userspace Atomic
136 (see [`doc/uatomic-api.md`](doc/uatomic-api.md))
142 ### Usage of all urcu libraries:
144 - Define `_LGPL_SOURCE` (only) if your code is LGPL or GPL compatible
145 before including the `urcu.h` or `urcu-qsbr.h` header. If your application
146 is distributed under another license, function calls will be generated
147 instead of inlines, so your application can link with the library.
148 - Linking with one of the libraries below is always necessary even for
149 LGPL and GPL applications.
150 - Define `URCU_INLINE_SMALL_FUNCTIONS` before including Userspace RCU
151 headers if you want Userspace RCU to inline small functions (10
152 lines or less) into the application. It can be used by applications
153 distributed under any kind of license, and does *not* make the
154 application a derived work of Userspace RCU.
156 Those small inlined functions are guaranteed to match the library
157 content as long as the library major version is unchanged.
158 Therefore, the application *must* be compiled with headers matching
159 the library major version number. Applications using
160 `URCU_INLINE_SMALL_FUNCTIONS` may be unable to use debugging
161 features of Userspace RCU without being recompiled.
163 There are multiple flavors of liburcu available:
171 The API members start with the prefix "urcu_<flavor>_", where
172 <flavor> is the chosen flavor name.
175 ### Usage of `liburcu-memb`
177 1. `#include <urcu/urcu-memb.h>`
178 2. Link the application with `-lurcu-memb`
180 This is the preferred version of the library, in terms of
181 grace-period detection speed, read-side speed and flexibility.
182 Dynamically detects kernel support for `sys_membarrier()`. Falls back
183 on `urcu-mb` scheme if support is not present, which has slower
184 read-side. Use the --disable-sys-membarrier-fallback configure option
185 to disable the fall back, thus requiring `sys_membarrier()` to be
186 available. This gives a small speedup when `sys_membarrier()` is
187 supported by the kernel, and aborts in the library constructor if not
191 ### Usage of `liburcu-qsbr`
193 1. `#include <urcu/urcu-qsbr.h>`
194 2. Link with `-lurcu-qsbr`
196 The QSBR flavor of RCU needs to have each reader thread executing
197 `rcu_quiescent_state()` periodically to progress. `rcu_thread_online()`
198 and `rcu_thread_offline()` can be used to mark long periods for which
199 the threads are not active. It provides the fastest read-side at the
200 expense of more intrusiveness in the application code.
203 ### Usage of `liburcu-mb`
205 1. `#include <urcu/urcu-mb.h>`
206 2. Link with `-lurcu-mb`
208 This version of the urcu library uses memory barriers on the writer
209 and reader sides. This results in faster grace-period detection, but
210 results in slower reads.
213 ### Usage of `liburcu-signal`
215 1. `#include <urcu/urcu-signal.h>`
216 2. Link the application with `-lurcu-signal`
218 Version of the library that requires a signal, typically `SIGUSR1`. Can
219 be overridden with `-DSIGRCU` by modifying `Makefile.build.inc`.
222 ### Usage of `liburcu-bp`
224 1. `#include <urcu/urcu-bp.h>`
225 2. Link with `-lurcu-bp`
227 The BP library flavor stands for "bulletproof". It is specifically
228 designed to help tracing library to hook on applications without
229 requiring to modify these applications. `urcu_bp_init()`, and
230 `urcu_bp_unregister_thread()` all become nops, whereas calling
231 `urcu_bp_register_thread()` becomes optional. The state is dealt with by
232 the library internally at the expense of read-side and write-side
238 Each thread that has reader critical sections (that uses
239 `urcu_<flavor>_read_lock()`/`urcu_<flavor>_read_unlock()` must first
240 register to the URCU library. This is done by calling
241 `urcu_<flavor>_register_thread()`. Unregistration must be performed
242 before exiting the thread by using `urcu_<flavor>_unregister_thread()`.
247 Reader critical sections must be protected by locating them between
248 calls to `urcu_<flavor>_read_lock()` and `urcu_<flavor>_read_unlock()`.
249 Inside that lock, `rcu_dereference()` may be called to read an RCU
255 `rcu_assign_pointer()` and `rcu_xchg_pointer()` may be called anywhere.
256 After, `urcu_<flavor>_synchronize_rcu()` must be called. When it
257 returns, the old values are not in usage anymore.
260 ### Usage of `liburcu-defer`
262 - Follow instructions for either `liburcu-memb`, `liburcu-qsbr`,
263 `liburcu-mb`, `liburcu-signal`, or `liburcu-bp` above.
264 The `liburcu-defer` functionality is pulled into each of
265 those library modules.
266 - Provides `urcu_<flavor>_defer_rcu()` primitive to enqueue delayed
267 callbacks. Queued callbacks are executed in batch periodically after
268 a grace period. Do _not_ use `urcu_<flavor>_defer_rcu()` within a
269 read-side critical section, because it may call
270 `urcu_<flavor>_synchronize_rcu()` if the thread queue is full. This
271 can lead to deadlock or worse.
272 - Requires that `urcu_<flavor>_defer_barrier()` must be called in
273 library destructor if a library queues callbacks and is expected to
274 be unloaded with `dlclose()`.
276 Its API is currently experimental. It may change in future library releases.
279 ### Usage of `urcu-call-rcu`
281 - Follow instructions for either `liburcu-memb`, `liburcu-qsbr`,
282 `liburcu-mb`, `liburcu-signal`, or `liburcu-bp` above.
283 The `urcu-call-rcu` functionality is pulled into each of
284 those library modules.
285 - Provides the `urcu_<flavor>_call_rcu()` primitive to enqueue delayed
286 callbacks in a manner similar to `urcu_<flavor>_defer_rcu()`, but
287 without ever delaying for a grace period. On the other hand,
288 `urcu_<flavor>_call_rcu()`'s best-case overhead is not quite as good
289 as that of `urcu_<flavor>_defer_rcu()`.
290 - Provides `urcu_<flavor>_call_rcu()` to allow asynchronous handling
291 of RCU grace periods. A number of additional functions are provided
292 to manage the helper threads used by `urcu_<flavor>_call_rcu()`, but
293 reasonable defaults are used if these additional functions are not
294 invoked. See [`doc/rcu-api.md`](doc/rcu-api.md) in userspace-rcu
295 documentation for more details.
298 ### Being careful with signals
300 The `liburcu-signal` library uses signals internally. The signal handler is
301 registered with the `SA_RESTART` flag. However, these signals may cause
302 some non-restartable system calls to fail with `errno = EINTR`. Care
303 should be taken to restart system calls manually if they fail with this
304 error. A list of non-restartable system calls may be found in
307 Read-side critical sections are allowed in a signal handler,
308 except those setup with `sigaltstack(2)`, with `liburcu-memb` and
309 `liburcu-mb`. Be careful, however, to disable these signals
310 between thread creation and calls to `urcu_<flavor>_register_thread()`,
311 because a signal handler nesting on an unregistered thread would not be
312 allowed to call `urcu_<flavor>_read_lock()`.
314 Read-side critical sections are _not_ allowed in a signal handler with
315 `liburcu-qsbr`, unless signals are disabled explicitly around each
316 `urcu_qsbr_quiescent_state()` calls, when threads are put offline and around
317 calls to `urcu_qsbr_synchronize_rcu()`. Even then, we do not recommend it.
320 ### Interaction with mutexes
322 One must be careful to do not cause deadlocks due to interaction of
323 `urcu_<flavor>_synchronize_rcu()` and RCU read-side with mutexes. If
324 `urcu_<flavor>_synchronize_rcu()` is called with a mutex held, this
325 mutex (or any mutex which has this mutex in its dependency chain) should
326 not be acquired from within a RCU read-side critical section.
328 This is especially important to understand in the context of the
329 QSBR flavor: a registered reader thread being "online" by
330 default should be considered as within a RCU read-side critical
331 section unless explicitly put "offline". Therefore, if
332 `urcu_qsbr_synchronize_rcu()` is called with a mutex held, this mutex,
333 as well as any mutex which has this mutex in its dependency chain should
334 only be taken when the RCU reader thread is "offline" (this can be
335 performed by calling `urcu_qsbr_thread_offline()`).
338 ### Interaction with `fork()`
340 Special care must be taken for applications performing `fork()` without
341 any following `exec()`. This is caused by the fact that Linux only clones
342 the thread calling `fork()`, and thus never replicates any of the other
343 parent thread into the child process. Most `liburcu` implementations
344 require that all registrations (as reader, `defer_rcu` and `call_rcu`
345 threads) should be released before a `fork()` is performed, except for the
346 rather common scenario where `fork()` is immediately followed by `exec()` in
347 the child process. The only implementation not subject to that rule is
348 `liburcu-bp`, which is designed to handle `fork()` by calling
349 `urcu_bp_before_fork`, `urcu_bp_after_fork_parent` and
350 `urcu_bp_after_fork_child`.
352 Applications that use `urcu_<flavor>_call_rcu()` and that `fork()`
353 without doing an immediate `exec()` must take special action. The
354 parent must invoke `urcu_<flavor>_call_rcu_before_fork()` before the
355 `fork()` and `urcu_<flavor>_call_rcu_after_fork_parent()` after the
356 `fork()`. The child process must invoke
357 `urcu_<flavor>_call_rcu_after_fork_child()`. Even though these three
358 APIs are suitable for passing to `pthread_atfork()`, use of
359 `pthread_atfork()` is **STRONGLY DISCOURAGED** for programs calling the
360 glibc memory allocator (`malloc()`, `calloc()`, `free()`, ...) within
361 `urcu_<flavor>_call_rcu` callbacks. This is due to limitations in the
362 way glibc memory allocator handles calls to the memory allocator from
363 concurrent threads while the `pthread_atfork()` handlers are executing.
367 - call to `free()` from callbacks executed within
368 `urcu_<flavor>_call_rcu` worker threads,
369 - executing `urcu_<flavor>_call_rcu` atfork handlers within the glibc
370 pthread atfork mechanism,
372 will sometimes trigger interesting process hangs. This usually
373 hangs on a memory allocator lock within glibc.
376 ### Thread Local Storage (TLS)
378 Userspace RCU can fall back on `pthread_getspecific()` to emulate
379 TLS variables on systems where it is not available. This behavior
380 can be forced by specifying `--disable-compiler-tls` as configure
384 ### Usage of `DEBUG_RCU` & `--enable-rcu-debug`
386 By default the library is configured with internal debugging
387 self-checks disabled.
389 For always-on debugging self-checks:
390 ./configure --enable-rcu-debug
392 For fine grained enabling of debugging self-checks, build
393 userspace-rcu with DEBUG_RCU defined and compile dependent
394 applications with DEBUG_RCU defined when necessary.
396 Warning: Enabling this feature result in a performance penalty.
399 ### Usage of `DEBUG_YIELD`
401 `DEBUG_YIELD` is used to add random delays in the code for testing
407 By default the library is configured to use synchronization primitives
408 adequate for SMP systems. On uniprocessor systems, support for SMP
409 systems can be disabled with:
411 ./configure --disable-smp-support
413 theoretically yielding slightly better performance.
416 ### Usage of `--enable-cds-lfht-iter-debug`
418 By default the library is configured with extra debugging checks for
419 lock-free hash table iterator traversal disabled.
421 Building liburcu with --enable-cds-lfht-iter-debug and rebuilding
422 application to match the ABI change allows finding cases where the hash
423 table iterator is re-purposed to be used on a different hash table while
424 still being used to iterate on a hash table.
426 This option alters the rculfhash ABI. Make sure to compile both library
427 and application with matching configuration.
433 In addition to the usual `make check` target, Userspace RCU features
434 `make regtest` and `make bench` targets:
436 - `make check`: short tests, meant to be run when rebuilding or
437 porting Userspace RCU.
438 - `make regtest`: long (many hours) test, meant to be run when
439 modifying Userspace RCU or porting it to a new architecture or
441 - `make bench`: long (many hours) benchmarks.
447 You can contact the maintainers on the following mailing list:
448 `lttng-dev@lists.lttng.org`.