diff options
author | Linus Torvalds <torvalds@linux-foundation.org> | 2014-06-12 18:48:15 -0700 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2014-06-12 18:48:15 -0700 |
commit | c29deef32e3699e40da3e9e82267610de04e6b54 (patch) | |
tree | 820ab21fe399225f7341499e461ee793a180d414 /Documentation | |
parent | f9da455b93f6ba076935b4ef4589f61e529ae046 (diff) | |
parent | bd01ec1a13f9a327950c8e3080096446c7804753 (diff) | |
download | linux-c29deef32e3699e40da3e9e82267610de04e6b54.tar.gz linux-c29deef32e3699e40da3e9e82267610de04e6b54.tar.bz2 linux-c29deef32e3699e40da3e9e82267610de04e6b54.zip |
Merge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull more locking changes from Ingo Molnar:
"This is the second round of locking tree updates for v3.16, offering
large system scalability improvements:
- optimistic spinning for rwsems, from Davidlohr Bueso.
- 'qrwlocks' core code and x86 enablement, from Waiman Long and PeterZ"
* 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86, locking/rwlocks: Enable qrwlocks on x86
locking/rwlocks: Introduce 'qrwlocks' - fair, queued rwlocks
locking/mutexes: Documentation update/rewrite
locking/rwsem: Fix checkpatch.pl warnings
locking/rwsem: Fix warnings for CONFIG_RWSEM_GENERIC_SPINLOCK
locking/rwsem: Support optimistic spinning
Diffstat (limited to 'Documentation')
-rw-r--r-- | Documentation/mutex-design.txt | 252 |
1 files changed, 135 insertions, 117 deletions
diff --git a/Documentation/mutex-design.txt b/Documentation/mutex-design.txt index 1dfe62c3641d..ee231ed09ec6 100644 --- a/Documentation/mutex-design.txt +++ b/Documentation/mutex-design.txt @@ -1,139 +1,157 @@ Generic Mutex Subsystem started by Ingo Molnar <mingo@redhat.com> +updated by Davidlohr Bueso <davidlohr@hp.com> - "Why on earth do we need a new mutex subsystem, and what's wrong - with semaphores?" +What are mutexes? +----------------- -firstly, there's nothing wrong with semaphores. But if the simpler -mutex semantics are sufficient for your code, then there are a couple -of advantages of mutexes: +In the Linux kernel, mutexes refer to a particular locking primitive +that enforces serialization on shared memory systems, and not only to +the generic term referring to 'mutual exclusion' found in academia +or similar theoretical text books. Mutexes are sleeping locks which +behave similarly to binary semaphores, and were introduced in 2006[1] +as an alternative to these. This new data structure provided a number +of advantages, including simpler interfaces, and at that time smaller +code (see Disadvantages). - - 'struct mutex' is smaller on most architectures: E.g. on x86, - 'struct semaphore' is 20 bytes, 'struct mutex' is 16 bytes. - A smaller structure size means less RAM footprint, and better - CPU-cache utilization. +[1] http://lwn.net/Articles/164802/ - - tighter code. On x86 i get the following .text sizes when - switching all mutex-alike semaphores in the kernel to the mutex - subsystem: +Implementation +-------------- - text data bss dec hex filename - 3280380 868188 396860 4545428 455b94 vmlinux-semaphore - 3255329 865296 396732 4517357 44eded vmlinux-mutex +Mutexes are represented by 'struct mutex', defined in include/linux/mutex.h +and implemented in kernel/locking/mutex.c. These locks use a three +state atomic counter (->count) to represent the different possible +transitions that can occur during the lifetime of a lock: - that's 25051 bytes of code saved, or a 0.76% win - off the hottest - codepaths of the kernel. (The .data savings are 2892 bytes, or 0.33%) - Smaller code means better icache footprint, which is one of the - major optimization goals in the Linux kernel currently. + 1: unlocked + 0: locked, no waiters + negative: locked, with potential waiters - - the mutex subsystem is slightly faster and has better scalability for - contended workloads. On an 8-way x86 system, running a mutex-based - kernel and testing creat+unlink+close (of separate, per-task files) - in /tmp with 16 parallel tasks, the average number of ops/sec is: +In its most basic form it also includes a wait-queue and a spinlock +that serializes access to it. CONFIG_SMP systems can also include +a pointer to the lock task owner (->owner) as well as a spinner MCS +lock (->osq), both described below in (ii). - Semaphores: Mutexes: +When acquiring a mutex, there are three possible paths that can be +taken, depending on the state of the lock: - $ ./test-mutex V 16 10 $ ./test-mutex V 16 10 - 8 CPUs, running 16 tasks. 8 CPUs, running 16 tasks. - checking VFS performance. checking VFS performance. - avg loops/sec: 34713 avg loops/sec: 84153 - CPU utilization: 63% CPU utilization: 22% +(i) fastpath: tries to atomically acquire the lock by decrementing the + counter. If it was already taken by another task it goes to the next + possible path. This logic is architecture specific. On x86-64, the + locking fastpath is 2 instructions: - i.e. in this workload, the mutex based kernel was 2.4 times faster - than the semaphore based kernel, _and_ it also had 2.8 times less CPU - utilization. (In terms of 'ops per CPU cycle', the semaphore kernel - performed 551 ops/sec per 1% of CPU time used, while the mutex kernel - performed 3825 ops/sec per 1% of CPU time used - it was 6.9 times - more efficient.) - - the scalability difference is visible even on a 2-way P4 HT box: - - Semaphores: Mutexes: - - $ ./test-mutex V 16 10 $ ./test-mutex V 16 10 - 4 CPUs, running 16 tasks. 8 CPUs, running 16 tasks. - checking VFS performance. checking VFS performance. - avg loops/sec: 127659 avg loops/sec: 181082 - CPU utilization: 100% CPU utilization: 34% - - (the straight performance advantage of mutexes is 41%, the per-cycle - efficiency of mutexes is 4.1 times better.) - - - there are no fastpath tradeoffs, the mutex fastpath is just as tight - as the semaphore fastpath. On x86, the locking fastpath is 2 - instructions: - - c0377ccb <mutex_lock>: - c0377ccb: f0 ff 08 lock decl (%eax) - c0377cce: 78 0e js c0377cde <.text..lock.mutex> - c0377cd0: c3 ret + 0000000000000e10 <mutex_lock>: + e21: f0 ff 0b lock decl (%rbx) + e24: 79 08 jns e2e <mutex_lock+0x1e> the unlocking fastpath is equally tight: - c0377cd1 <mutex_unlock>: - c0377cd1: f0 ff 00 lock incl (%eax) - c0377cd4: 7e 0f jle c0377ce5 <.text..lock.mutex+0x7> - c0377cd6: c3 ret - - - 'struct mutex' semantics are well-defined and are enforced if - CONFIG_DEBUG_MUTEXES is turned on. Semaphores on the other hand have - virtually no debugging code or instrumentation. The mutex subsystem - checks and enforces the following rules: - - * - only one task can hold the mutex at a time - * - only the owner can unlock the mutex - * - multiple unlocks are not permitted - * - recursive locking is not permitted - * - a mutex object must be initialized via the API - * - a mutex object must not be initialized via memset or copying - * - task may not exit with mutex held - * - memory areas where held locks reside must not be freed - * - held mutexes must not be reinitialized - * - mutexes may not be used in hardware or software interrupt - * contexts such as tasklets and timers - - furthermore, there are also convenience features in the debugging - code: - - * - uses symbolic names of mutexes, whenever they are printed in debug output - * - point-of-acquire tracking, symbolic lookup of function names - * - list of all locks held in the system, printout of them - * - owner tracking - * - detects self-recursing locks and prints out all relevant info - * - detects multi-task circular deadlocks and prints out all affected - * locks and tasks (and only those tasks) + 0000000000000bc0 <mutex_unlock>: + bc8: f0 ff 07 lock incl (%rdi) + bcb: 7f 0a jg bd7 <mutex_unlock+0x17> + + +(ii) midpath: aka optimistic spinning, tries to spin for acquisition + while the lock owner is running and there are no other tasks ready + to run that have higher priority (need_resched). The rationale is + that if the lock owner is running, it is likely to release the lock + soon. The mutex spinners are queued up using MCS lock so that only + one spinner can compete for the mutex. + + The MCS lock (proposed by Mellor-Crummey and Scott) is a simple spinlock + with the desirable properties of being fair and with each cpu trying + to acquire the lock spinning on a local variable. It avoids expensive + cacheline bouncing that common test-and-set spinlock implementations + incur. An MCS-like lock is specially tailored for optimistic spinning + for sleeping lock implementation. An important feature of the customized + MCS lock is that it has the extra property that spinners are able to exit + the MCS spinlock queue when they need to reschedule. This further helps + avoid situations where MCS spinners that need to reschedule would continue + waiting to spin on mutex owner, only to go directly to slowpath upon + obtaining the MCS lock. + + +(iii) slowpath: last resort, if the lock is still unable to be acquired, + the task is added to the wait-queue and sleeps until woken up by the + unlock path. Under normal circumstances it blocks as TASK_UNINTERRUPTIBLE. + +While formally kernel mutexes are sleepable locks, it is path (ii) that +makes them more practically a hybrid type. By simply not interrupting a +task and busy-waiting for a few cycles instead of immediately sleeping, +the performance of this lock has been seen to significantly improve a +number of workloads. Note that this technique is also used for rw-semaphores. + +Semantics +--------- + +The mutex subsystem checks and enforces the following rules: + + - Only one task can hold the mutex at a time. + - Only the owner can unlock the mutex. + - Multiple unlocks are not permitted. + - Recursive locking/unlocking is not permitted. + - A mutex must only be initialized via the API (see below). + - A task may not exit with a mutex held. + - Memory areas where held locks reside must not be freed. + - Held mutexes must not be reinitialized. + - Mutexes may not be used in hardware or software interrupt + contexts such as tasklets and timers. + +These semantics are fully enforced when CONFIG DEBUG_MUTEXES is enabled. +In addition, the mutex debugging code also implements a number of other +features that make lock debugging easier and faster: + + - Uses symbolic names of mutexes, whenever they are printed + in debug output. + - Point-of-acquire tracking, symbolic lookup of function names, + list of all locks held in the system, printout of them. + - Owner tracking. + - Detects self-recursing locks and prints out all relevant info. + - Detects multi-task circular deadlocks and prints out all affected + locks and tasks (and only those tasks). + + +Interfaces +---------- +Statically define the mutex: + DEFINE_MUTEX(name); + +Dynamically initialize the mutex: + mutex_init(mutex); + +Acquire the mutex, uninterruptible: + void mutex_lock(struct mutex *lock); + void mutex_lock_nested(struct mutex *lock, unsigned int subclass); + int mutex_trylock(struct mutex *lock); + +Acquire the mutex, interruptible: + int mutex_lock_interruptible_nested(struct mutex *lock, + unsigned int subclass); + int mutex_lock_interruptible(struct mutex *lock); + +Acquire the mutex, interruptible, if dec to 0: + int atomic_dec_and_mutex_lock(atomic_t *cnt, struct mutex *lock); + +Unlock the mutex: + void mutex_unlock(struct mutex *lock); + +Test if the mutex is taken: + int mutex_is_locked(struct mutex *lock); Disadvantages ------------- -The stricter mutex API means you cannot use mutexes the same way you -can use semaphores: e.g. they cannot be used from an interrupt context, -nor can they be unlocked from a different context that which acquired -it. [ I'm not aware of any other (e.g. performance) disadvantages from -using mutexes at the moment, please let me know if you find any. ] - -Implementation of mutexes -------------------------- - -'struct mutex' is the new mutex type, defined in include/linux/mutex.h and -implemented in kernel/locking/mutex.c. It is a counter-based mutex with a -spinlock and a wait-list. The counter has 3 states: 1 for "unlocked", 0 for -"locked" and negative numbers (usually -1) for "locked, potential waiters -queued". - -the APIs of 'struct mutex' have been streamlined: - - DEFINE_MUTEX(name); +Unlike its original design and purpose, 'struct mutex' is larger than +most locks in the kernel. E.g: on x86-64 it is 40 bytes, almost twice +as large as 'struct semaphore' (24 bytes) and 8 bytes shy of the +'struct rw_semaphore' variant. Larger structure sizes mean more CPU +cache and memory footprint. - mutex_init(mutex); +When to use mutexes +------------------- - void mutex_lock(struct mutex *lock); - int mutex_lock_interruptible(struct mutex *lock); - int mutex_trylock(struct mutex *lock); - void mutex_unlock(struct mutex *lock); - int mutex_is_locked(struct mutex *lock); - void mutex_lock_nested(struct mutex *lock, unsigned int subclass); - int mutex_lock_interruptible_nested(struct mutex *lock, - unsigned int subclass); - int atomic_dec_and_mutex_lock(atomic_t *cnt, struct mutex *lock); +Unless the strict semantics of mutexes are unsuitable and/or the critical +region prevents the lock from being shared, always prefer them to any other +locking primitive. |