1 2Concurrency Managed Workqueue (cmwq) 3 4September, 2010 Tejun Heo <tj@kernel.org> 5 Florian Mickler <florian@mickler.org> 6 7CONTENTS 8 91. Introduction 102. Why cmwq? 113. The Design 124. Application Programming Interface (API) 135. Example Execution Scenarios 146. Guidelines 157. Debugging 16 17 181. Introduction 19 20There are many cases where an asynchronous process execution context 21is needed and the workqueue (wq) API is the most commonly used 22mechanism for such cases. 23 24When such an asynchronous execution context is needed, a work item 25describing which function to execute is put on a queue. An 26independent thread serves as the asynchronous execution context. The 27queue is called workqueue and the thread is called worker. 28 29While there are work items on the workqueue the worker executes the 30functions associated with the work items one after the other. When 31there is no work item left on the workqueue the worker becomes idle. 32When a new work item gets queued, the worker begins executing again. 33 34 352. Why cmwq? 36 37In the original wq implementation, a multi threaded (MT) wq had one 38worker thread per CPU and a single threaded (ST) wq had one worker 39thread system-wide. A single MT wq needed to keep around the same 40number of workers as the number of CPUs. The kernel grew a lot of MT 41wq users over the years and with the number of CPU cores continuously 42rising, some systems saturated the default 32k PID space just booting 43up. 44 45Although MT wq wasted a lot of resource, the level of concurrency 46provided was unsatisfactory. The limitation was common to both ST and 47MT wq albeit less severe on MT. Each wq maintained its own separate 48worker pool. A MT wq could provide only one execution context per CPU 49while a ST wq one for the whole system. Work items had to compete for 50those very limited execution contexts leading to various problems 51including proneness to deadlocks around the single execution context. 52 53The tension between the provided level of concurrency and resource 54usage also forced its users to make unnecessary tradeoffs like libata 55choosing to use ST wq for polling PIOs and accepting an unnecessary 56limitation that no two polling PIOs can progress at the same time. As 57MT wq don't provide much better concurrency, users which require 58higher level of concurrency, like async or fscache, had to implement 59their own thread pool. 60 61Concurrency Managed Workqueue (cmwq) is a reimplementation of wq with 62focus on the following goals. 63 64* Maintain compatibility with the original workqueue API. 65 66* Use per-CPU unified worker pools shared by all wq to provide 67 flexible level of concurrency on demand without wasting a lot of 68 resource. 69 70* Automatically regulate worker pool and level of concurrency so that 71 the API users don't need to worry about such details. 72 73 743. The Design 75 76In order to ease the asynchronous execution of functions a new 77abstraction, the work item, is introduced. 78 79A work item is a simple struct that holds a pointer to the function 80that is to be executed asynchronously. Whenever a driver or subsystem 81wants a function to be executed asynchronously it has to set up a work 82item pointing to that function and queue that work item on a 83workqueue. 84 85Special purpose threads, called worker threads, execute the functions 86off of the queue, one after the other. If no work is queued, the 87worker threads become idle. These worker threads are managed in so 88called thread-pools. 89 90The cmwq design differentiates between the user-facing workqueues that 91subsystems and drivers queue work items on and the backend mechanism 92which manages thread-pools and processes the queued work items. 93 94The backend is called gcwq. There is one gcwq for each possible CPU 95and one gcwq to serve work items queued on unbound workqueues. Each 96gcwq has two thread-pools - one for normal work items and the other 97for high priority ones. 98 99Subsystems and drivers can create and queue work items through special 100workqueue API functions as they see fit. They can influence some 101aspects of the way the work items are executed by setting flags on the 102workqueue they are putting the work item on. These flags include 103things like CPU locality, reentrancy, concurrency limits, priority and 104more. To get a detailed overview refer to the API description of 105alloc_workqueue() below. 106 107When a work item is queued to a workqueue, the target gcwq and 108thread-pool is determined according to the queue parameters and 109workqueue attributes and appended on the shared worklist of the 110thread-pool. For example, unless specifically overridden, a work item 111of a bound workqueue will be queued on the worklist of either normal 112or highpri thread-pool of the gcwq that is associated to the CPU the 113issuer is running on. 114 115For any worker pool implementation, managing the concurrency level 116(how many execution contexts are active) is an important issue. cmwq 117tries to keep the concurrency at a minimal but sufficient level. 118Minimal to save resources and sufficient in that the system is used at 119its full capacity. 120 121Each thread-pool bound to an actual CPU implements concurrency 122management by hooking into the scheduler. The thread-pool is notified 123whenever an active worker wakes up or sleeps and keeps track of the 124number of the currently runnable workers. Generally, work items are 125not expected to hog a CPU and consume many cycles. That means 126maintaining just enough concurrency to prevent work processing from 127stalling should be optimal. As long as there are one or more runnable 128workers on the CPU, the thread-pool doesn't start execution of a new 129work, but, when the last running worker goes to sleep, it immediately 130schedules a new worker so that the CPU doesn't sit idle while there 131are pending work items. This allows using a minimal number of workers 132without losing execution bandwidth. 133 134Keeping idle workers around doesn't cost other than the memory space 135for kthreads, so cmwq holds onto idle ones for a while before killing 136them. 137 138For an unbound wq, the above concurrency management doesn't apply and 139the thread-pools for the pseudo unbound CPU try to start executing all 140work items as soon as possible. The responsibility of regulating 141concurrency level is on the users. There is also a flag to mark a 142bound wq to ignore the concurrency management. Please refer to the 143API section for details. 144 145Forward progress guarantee relies on that workers can be created when 146more execution contexts are necessary, which in turn is guaranteed 147through the use of rescue workers. All work items which might be used 148on code paths that handle memory reclaim are required to be queued on 149wq's that have a rescue-worker reserved for execution under memory 150pressure. Else it is possible that the thread-pool deadlocks waiting 151for execution contexts to free up. 152 153 1544. Application Programming Interface (API) 155 156alloc_workqueue() allocates a wq. The original create_*workqueue() 157functions are deprecated and scheduled for removal. alloc_workqueue() 158takes three arguments - @name, @flags and @max_active. @name is the 159name of the wq and also used as the name of the rescuer thread if 160there is one. 161 162A wq no longer manages execution resources but serves as a domain for 163forward progress guarantee, flush and work item attributes. @flags 164and @max_active control how work items are assigned execution 165resources, scheduled and executed. 166 167@flags: 168 169 WQ_NON_REENTRANT 170 171 By default, a wq guarantees non-reentrance only on the same 172 CPU. A work item may not be executed concurrently on the same 173 CPU by multiple workers but is allowed to be executed 174 concurrently on multiple CPUs. This flag makes sure 175 non-reentrance is enforced across all CPUs. Work items queued 176 to a non-reentrant wq are guaranteed to be executed by at most 177 one worker system-wide at any given time. 178 179 WQ_UNBOUND 180 181 Work items queued to an unbound wq are served by a special 182 gcwq which hosts workers which are not bound to any specific 183 CPU. This makes the wq behave as a simple execution context 184 provider without concurrency management. The unbound gcwq 185 tries to start execution of work items as soon as possible. 186 Unbound wq sacrifices locality but is useful for the following 187 cases. 188 189 * Wide fluctuation in the concurrency level requirement is 190 expected and using bound wq may end up creating large number 191 of mostly unused workers across different CPUs as the issuer 192 hops through different CPUs. 193 194 * Long running CPU intensive workloads which can be better 195 managed by the system scheduler. 196 197 WQ_FREEZABLE 198 199 A freezable wq participates in the freeze phase of the system 200 suspend operations. Work items on the wq are drained and no 201 new work item starts execution until thawed. 202 203 WQ_MEM_RECLAIM 204 205 All wq which might be used in the memory reclaim paths _MUST_ 206 have this flag set. The wq is guaranteed to have at least one 207 execution context regardless of memory pressure. 208 209 WQ_HIGHPRI 210 211 Work items of a highpri wq are queued to the highpri 212 thread-pool of the target gcwq. Highpri thread-pools are 213 served by worker threads with elevated nice level. 214 215 Note that normal and highpri thread-pools don't interact with 216 each other. Each maintain its separate pool of workers and 217 implements concurrency management among its workers. 218 219 WQ_CPU_INTENSIVE 220 221 Work items of a CPU intensive wq do not contribute to the 222 concurrency level. In other words, runnable CPU intensive 223 work items will not prevent other work items in the same 224 thread-pool from starting execution. This is useful for bound 225 work items which are expected to hog CPU cycles so that their 226 execution is regulated by the system scheduler. 227 228 Although CPU intensive work items don't contribute to the 229 concurrency level, start of their executions is still 230 regulated by the concurrency management and runnable 231 non-CPU-intensive work items can delay execution of CPU 232 intensive work items. 233 234 This flag is meaningless for unbound wq. 235 236@max_active: 237 238@max_active determines the maximum number of execution contexts per 239CPU which can be assigned to the work items of a wq. For example, 240with @max_active of 16, at most 16 work items of the wq can be 241executing at the same time per CPU. 242 243Currently, for a bound wq, the maximum limit for @max_active is 512 244and the default value used when 0 is specified is 256. For an unbound 245wq, the limit is higher of 512 and 4 * num_possible_cpus(). These 246values are chosen sufficiently high such that they are not the 247limiting factor while providing protection in runaway cases. 248 249The number of active work items of a wq is usually regulated by the 250users of the wq, more specifically, by how many work items the users 251may queue at the same time. Unless there is a specific need for 252throttling the number of active work items, specifying '0' is 253recommended. 254 255Some users depend on the strict execution ordering of ST wq. The 256combination of @max_active of 1 and WQ_UNBOUND is used to achieve this 257behavior. Work items on such wq are always queued to the unbound gcwq 258and only one work item can be active at any given time thus achieving 259the same ordering property as ST wq. 260 261 2625. Example Execution Scenarios 263 264The following example execution scenarios try to illustrate how cmwq 265behave under different configurations. 266 267 Work items w0, w1, w2 are queued to a bound wq q0 on the same CPU. 268 w0 burns CPU for 5ms then sleeps for 10ms then burns CPU for 5ms 269 again before finishing. w1 and w2 burn CPU for 5ms then sleep for 270 10ms. 271 272Ignoring all other tasks, works and processing overhead, and assuming 273simple FIFO scheduling, the following is one highly simplified version 274of possible sequences of events with the original wq. 275 276 TIME IN MSECS EVENT 277 0 w0 starts and burns CPU 278 5 w0 sleeps 279 15 w0 wakes up and burns CPU 280 20 w0 finishes 281 20 w1 starts and burns CPU 282 25 w1 sleeps 283 35 w1 wakes up and finishes 284 35 w2 starts and burns CPU 285 40 w2 sleeps 286 50 w2 wakes up and finishes 287 288And with cmwq with @max_active >= 3, 289 290 TIME IN MSECS EVENT 291 0 w0 starts and burns CPU 292 5 w0 sleeps 293 5 w1 starts and burns CPU 294 10 w1 sleeps 295 10 w2 starts and burns CPU 296 15 w2 sleeps 297 15 w0 wakes up and burns CPU 298 20 w0 finishes 299 20 w1 wakes up and finishes 300 25 w2 wakes up and finishes 301 302If @max_active == 2, 303 304 TIME IN MSECS EVENT 305 0 w0 starts and burns CPU 306 5 w0 sleeps 307 5 w1 starts and burns CPU 308 10 w1 sleeps 309 15 w0 wakes up and burns CPU 310 20 w0 finishes 311 20 w1 wakes up and finishes 312 20 w2 starts and burns CPU 313 25 w2 sleeps 314 35 w2 wakes up and finishes 315 316Now, let's assume w1 and w2 are queued to a different wq q1 which has 317WQ_CPU_INTENSIVE set, 318 319 TIME IN MSECS EVENT 320 0 w0 starts and burns CPU 321 5 w0 sleeps 322 5 w1 and w2 start and burn CPU 323 10 w1 sleeps 324 15 w2 sleeps 325 15 w0 wakes up and burns CPU 326 20 w0 finishes 327 20 w1 wakes up and finishes 328 25 w2 wakes up and finishes 329 330 3316. Guidelines 332 333* Do not forget to use WQ_MEM_RECLAIM if a wq may process work items 334 which are used during memory reclaim. Each wq with WQ_MEM_RECLAIM 335 set has an execution context reserved for it. If there is 336 dependency among multiple work items used during memory reclaim, 337 they should be queued to separate wq each with WQ_MEM_RECLAIM. 338 339* Unless strict ordering is required, there is no need to use ST wq. 340 341* Unless there is a specific need, using 0 for @max_active is 342 recommended. In most use cases, concurrency level usually stays 343 well under the default limit. 344 345* A wq serves as a domain for forward progress guarantee 346 (WQ_MEM_RECLAIM, flush and work item attributes. Work items which 347 are not involved in memory reclaim and don't need to be flushed as a 348 part of a group of work items, and don't require any special 349 attribute, can use one of the system wq. There is no difference in 350 execution characteristics between using a dedicated wq and a system 351 wq. 352 353* Unless work items are expected to consume a huge amount of CPU 354 cycles, using a bound wq is usually beneficial due to the increased 355 level of locality in wq operations and work item execution. 356 357 3587. Debugging 359 360Because the work functions are executed by generic worker threads 361there are a few tricks needed to shed some light on misbehaving 362workqueue users. 363 364Worker threads show up in the process list as: 365 366root 5671 0.0 0.0 0 0 ? S 12:07 0:00 [kworker/0:1] 367root 5672 0.0 0.0 0 0 ? S 12:07 0:00 [kworker/1:2] 368root 5673 0.0 0.0 0 0 ? S 12:12 0:00 [kworker/0:0] 369root 5674 0.0 0.0 0 0 ? S 12:13 0:00 [kworker/1:0] 370 371If kworkers are going crazy (using too much cpu), there are two types 372of possible problems: 373 374 1. Something beeing scheduled in rapid succession 375 2. A single work item that consumes lots of cpu cycles 376 377The first one can be tracked using tracing: 378 379 $ echo workqueue:workqueue_queue_work > /sys/kernel/debug/tracing/set_event 380 $ cat /sys/kernel/debug/tracing/trace_pipe > out.txt 381 (wait a few secs) 382 ^C 383 384If something is busy looping on work queueing, it would be dominating 385the output and the offender can be determined with the work item 386function. 387 388For the second type of problems it should be possible to just check 389the stack trace of the offending worker thread. 390 391 $ cat /proc/THE_OFFENDING_KWORKER/stack 392 393The work item's function should be trivially visible in the stack 394trace. 395