OpenE2K
/
linux
6
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Documentation/workqueue.txt: convert to ReST markup 

... and move to Documentation/core-api folder.

Signed-off-by: Silvio Fricke <silvio.fricke@gmail.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
This commit is contained in:
Silvio Fricke 2016-10-28 10:14:11 +02:00 committed by Jonathan Corbet
parent 24755a55b0
commit e7f08ffb18
3 changed files with 122 additions and 114 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
Documentation/core-api/index.rst
View File

@ -7,6 +7,8 @@ Kernel and driver related documentation.
.. toctree:: .. toctree::
:maxdepth: 1 :maxdepth: 1
workqueue
.. only:: subproject .. only:: subproject
Indices Indices

162
Documentation/workqueue.txt → Documentation/core-api/workqueue.rst
View File

@ -1,21 +1,14 @@
====================================
Concurrency Managed Workqueue (cmwq) Concurrency Managed Workqueue (cmwq)
====================================
September, 2010 Tejun Heo <tj@kernel.org> :Date: September, 2010
Florian Mickler <florian@mickler.org> :Author: Tejun Heo <tj@kernel.org>
:Author: Florian Mickler <florian@mickler.org>
CONTENTS
1. Introduction
2. Why cmwq?
3. The Design
4. Application Programming Interface (API)
5. Example Execution Scenarios
6. Guidelines
7. Debugging
1. Introduction Introduction
============
There are many cases where an asynchronous process execution context There are many cases where an asynchronous process execution context
is needed and the workqueue (wq) API is the most commonly used is needed and the workqueue (wq) API is the most commonly used
@ -32,7 +25,8 @@ there is no work item left on the workqueue the worker becomes idle.
When a new work item gets queued, the worker begins executing again. When a new work item gets queued, the worker begins executing again.
2. Why cmwq? Why cmwq?
=========
In the original wq implementation, a multi threaded (MT) wq had one In the original wq implementation, a multi threaded (MT) wq had one
worker thread per CPU and a single threaded (ST) wq had one worker worker thread per CPU and a single threaded (ST) wq had one worker
@ -71,7 +65,8 @@ focus on the following goals.
the API users don't need to worry about such details. the API users don't need to worry about such details.
3. The Design The Design
==========
In order to ease the asynchronous execution of functions a new In order to ease the asynchronous execution of functions a new
abstraction, the work item, is introduced. abstraction, the work item, is introduced.
@ -102,7 +97,7 @@ aspects of the way the work items are executed by setting flags on the
workqueue they are putting the work item on. These flags include workqueue they are putting the work item on. These flags include
things like CPU locality, concurrency limits, priority and more. To things like CPU locality, concurrency limits, priority and more. To
get a detailed overview refer to the API description of get a detailed overview refer to the API description of
alloc_workqueue() below. ``alloc_workqueue()`` below.
When a work item is queued to a workqueue, the target worker-pool is When a work item is queued to a workqueue, the target worker-pool is
determined according to the queue parameters and workqueue attributes determined according to the queue parameters and workqueue attributes
@ -136,7 +131,7 @@ them.
For unbound workqueues, the number of backing pools is dynamic. For unbound workqueues, the number of backing pools is dynamic.
Unbound workqueue can be assigned custom attributes using Unbound workqueue can be assigned custom attributes using
apply_workqueue_attrs() and workqueue will automatically create ``apply_workqueue_attrs()`` and workqueue will automatically create
backing worker pools matching the attributes. The responsibility of backing worker pools matching the attributes. The responsibility of
regulating concurrency level is on the users. There is also a flag to regulating concurrency level is on the users. There is also a flag to
mark a bound wq to ignore the concurrency management. Please refer to mark a bound wq to ignore the concurrency management. Please refer to
@ -151,23 +146,25 @@ pressure. Else it is possible that the worker-pool deadlocks waiting
for execution contexts to free up. for execution contexts to free up.
4. Application Programming Interface (API) Application Programming Interface (API)
=======================================
alloc_workqueue() allocates a wq. The original create_*workqueue() ``alloc_workqueue()`` allocates a wq. The original
functions are deprecated and scheduled for removal. alloc_workqueue() ``create_*workqueue()`` functions are deprecated and scheduled for
takes three arguments - @name, @flags and @max_active. @name is the removal. ``alloc_workqueue()`` takes three arguments - @``name``,
name of the wq and also used as the name of the rescuer thread if ``@flags`` and ``@max_active``. ``@name`` is the name of the wq and
there is one. also used as the name of the rescuer thread if there is one.
A wq no longer manages execution resources but serves as a domain for A wq no longer manages execution resources but serves as a domain for
forward progress guarantee, flush and work item attributes. @flags forward progress guarantee, flush and work item attributes. ``@flags``
and @max_active control how work items are assigned execution and ``@max_active`` control how work items are assigned execution
resources, scheduled and executed. resources, scheduled and executed.
@flags:
WQ_UNBOUND ``flags``
---------
``WQ_UNBOUND``
Work items queued to an unbound wq are served by the special Work items queued to an unbound wq are served by the special
worker-pools which host workers which are not bound to any worker-pools which host workers which are not bound to any
specific CPU. This makes the wq behave as a simple execution specific CPU. This makes the wq behave as a simple execution
@ -184,20 +181,17 @@ resources, scheduled and executed.
* Long running CPU intensive workloads which can be better * Long running CPU intensive workloads which can be better
managed by the system scheduler. managed by the system scheduler.
WQ_FREEZABLE ``WQ_FREEZABLE``
A freezable wq participates in the freeze phase of the system A freezable wq participates in the freeze phase of the system
suspend operations. Work items on the wq are drained and no suspend operations. Work items on the wq are drained and no
new work item starts execution until thawed. new work item starts execution until thawed.
WQ_MEM_RECLAIM ``WQ_MEM_RECLAIM``
All wq which might be used in the memory reclaim paths **MUST**
All wq which might be used in the memory reclaim paths _MUST_
have this flag set. The wq is guaranteed to have at least one have this flag set. The wq is guaranteed to have at least one
execution context regardless of memory pressure. execution context regardless of memory pressure.
WQ_HIGHPRI ``WQ_HIGHPRI``
Work items of a highpri wq are queued to the highpri Work items of a highpri wq are queued to the highpri
worker-pool of the target cpu. Highpri worker-pools are worker-pool of the target cpu. Highpri worker-pools are
served by worker threads with elevated nice level. served by worker threads with elevated nice level.
@ -206,8 +200,7 @@ resources, scheduled and executed.
each other. Each maintain its separate pool of workers and each other. Each maintain its separate pool of workers and
implements concurrency management among its workers. implements concurrency management among its workers.
WQ_CPU_INTENSIVE ``WQ_CPU_INTENSIVE``
Work items of a CPU intensive wq do not contribute to the Work items of a CPU intensive wq do not contribute to the
concurrency level. In other words, runnable CPU intensive concurrency level. In other words, runnable CPU intensive
work items will not prevent other work items in the same work items will not prevent other work items in the same
@ -223,22 +216,25 @@ resources, scheduled and executed.
This flag is meaningless for unbound wq. This flag is meaningless for unbound wq.
Note that the flag WQ_NON_REENTRANT no longer exists as all workqueues Note that the flag ``WQ_NON_REENTRANT`` no longer exists as all
are now non-reentrant - any work item is guaranteed to be executed by workqueues are now non-reentrant - any work item is guaranteed to be
at most one worker system-wide at any given time. executed by at most one worker system-wide at any given time.
@max_active:
@max_active determines the maximum number of execution contexts per ``max_active``
CPU which can be assigned to the work items of a wq. For example, --------------
with @max_active of 16, at most 16 work items of the wq can be
``@max_active`` determines the maximum number of execution contexts
per CPU which can be assigned to the work items of a wq. For example,
with ``@max_active`` of 16, at most 16 work items of the wq can be
executing at the same time per CPU. executing at the same time per CPU.
Currently, for a bound wq, the maximum limit for @max_active is 512 Currently, for a bound wq, the maximum limit for ``@max_active`` is
and the default value used when 0 is specified is 256. For an unbound 512 and the default value used when 0 is specified is 256. For an
wq, the limit is higher of 512 and 4 * num_possible_cpus(). These unbound wq, the limit is higher of 512 and 4 *
values are chosen sufficiently high such that they are not the ``num_possible_cpus()``. These values are chosen sufficiently high
limiting factor while providing protection in runaway cases. such that they are not the limiting factor while providing protection
in runaway cases.
The number of active work items of a wq is usually regulated by the The number of active work items of a wq is usually regulated by the
users of the wq, more specifically, by how many work items the users users of the wq, more specifically, by how many work items the users
@ -247,13 +243,14 @@ throttling the number of active work items, specifying '0' is
recommended. recommended.
Some users depend on the strict execution ordering of ST wq. The Some users depend on the strict execution ordering of ST wq. The
combination of @max_active of 1 and WQ_UNBOUND is used to achieve this combination of ``@max_active`` of 1 and ``WQ_UNBOUND`` is used to
behavior. Work items on such wq are always queued to the unbound achieve this behavior. Work items on such wq are always queued to the
worker-pools and only one work item can be active at any given time thus unbound worker-pools and only one work item can be active at any given
achieving the same ordering property as ST wq. time thus achieving the same ordering property as ST wq.
5. Example Execution Scenarios Example Execution Scenarios
===========================
The following example execution scenarios try to illustrate how cmwq The following example execution scenarios try to illustrate how cmwq
behave under different configurations. behave under different configurations.
@ -265,7 +262,7 @@ behave under different configurations.
Ignoring all other tasks, works and processing overhead, and assuming Ignoring all other tasks, works and processing overhead, and assuming
simple FIFO scheduling, the following is one highly simplified version simple FIFO scheduling, the following is one highly simplified version
of possible sequences of events with the original wq. of possible sequences of events with the original wq. ::
TIME IN MSECS EVENT TIME IN MSECS EVENT
0 w0 starts and burns CPU 0 w0 starts and burns CPU
@ -279,7 +276,7 @@ of possible sequences of events with the original wq.
40 w2 sleeps 40 w2 sleeps
50 w2 wakes up and finishes 50 w2 wakes up and finishes
And with cmwq with @max_active >= 3, And with cmwq with ``@max_active`` >= 3, ::
TIME IN MSECS EVENT TIME IN MSECS EVENT
0 w0 starts and burns CPU 0 w0 starts and burns CPU
@ -293,7 +290,7 @@ And with cmwq with @max_active >= 3,
20 w1 wakes up and finishes 20 w1 wakes up and finishes
25 w2 wakes up and finishes 25 w2 wakes up and finishes
If @max_active == 2, If ``@max_active`` == 2, ::
TIME IN MSECS EVENT TIME IN MSECS EVENT
0 w0 starts and burns CPU 0 w0 starts and burns CPU
@ -308,7 +305,7 @@ If @max_active == 2,
35 w2 wakes up and finishes 35 w2 wakes up and finishes
Now, let's assume w1 and w2 are queued to a different wq q1 which has Now, let's assume w1 and w2 are queued to a different wq q1 which has
WQ_CPU_INTENSIVE set, ``WQ_CPU_INTENSIVE`` set, ::
TIME IN MSECS EVENT TIME IN MSECS EVENT
0 w0 starts and burns CPU 0 w0 starts and burns CPU
@ -322,13 +319,15 @@ WQ_CPU_INTENSIVE set,
25 w2 wakes up and finishes 25 w2 wakes up and finishes
6. Guidelines Guidelines
==========
* Do not forget to use WQ_MEM_RECLAIM if a wq may process work items * Do not forget to use ``WQ_MEM_RECLAIM`` if a wq may process work
which are used during memory reclaim. Each wq with WQ_MEM_RECLAIM items which are used during memory reclaim. Each wq with
set has an execution context reserved for it. If there is ``WQ_MEM_RECLAIM`` set has an execution context reserved for it. If
dependency among multiple work items used during memory reclaim, there is dependency among multiple work items used during memory
they should be queued to separate wq each with WQ_MEM_RECLAIM. reclaim, they should be queued to separate wq each with
``WQ_MEM_RECLAIM``.
* Unless strict ordering is required, there is no need to use ST wq. * Unless strict ordering is required, there is no need to use ST wq.
@ -337,30 +336,31 @@ WQ_CPU_INTENSIVE set,
well under the default limit. well under the default limit.
* A wq serves as a domain for forward progress guarantee * A wq serves as a domain for forward progress guarantee
(WQ_MEM_RECLAIM, flush and work item attributes. Work items which (``WQ_MEM_RECLAIM``, flush and work item attributes. Work items
are not involved in memory reclaim and don't need to be flushed as a which are not involved in memory reclaim and don't need to be
part of a group of work items, and don't require any special flushed as a part of a group of work items, and don't require any
attribute, can use one of the system wq. There is no difference in special attribute, can use one of the system wq. There is no
execution characteristics between using a dedicated wq and a system difference in execution characteristics between using a dedicated wq
wq. and a system wq.
* Unless work items are expected to consume a huge amount of CPU * Unless work items are expected to consume a huge amount of CPU
cycles, using a bound wq is usually beneficial due to the increased cycles, using a bound wq is usually beneficial due to the increased
level of locality in wq operations and work item execution. level of locality in wq operations and work item execution.
7. Debugging Debugging
=========
Because the work functions are executed by generic worker threads Because the work functions are executed by generic worker threads
there are a few tricks needed to shed some light on misbehaving there are a few tricks needed to shed some light on misbehaving
workqueue users. workqueue users.
Worker threads show up in the process list as: Worker threads show up in the process list as: ::
root 5671 0.0 0.0 0 0 ? S 12:07 0:00 [kworker/0:1] root 5671 0.0 0.0 0 0 ? S 12:07 0:00 [kworker/0:1]
root 5672 0.0 0.0 0 0 ? S 12:07 0:00 [kworker/1:2] root 5672 0.0 0.0 0 0 ? S 12:07 0:00 [kworker/1:2]
root 5673 0.0 0.0 0 0 ? S 12:12 0:00 [kworker/0:0] root 5673 0.0 0.0 0 0 ? S 12:12 0:00 [kworker/0:0]
root 5674 0.0 0.0 0 0 ? S 12:13 0:00 [kworker/1:0] root 5674 0.0 0.0 0 0 ? S 12:13 0:00 [kworker/1:0]
If kworkers are going crazy (using too much cpu), there are two types If kworkers are going crazy (using too much cpu), there are two types
of possible problems: of possible problems:
@ -368,7 +368,7 @@ of possible problems:
1. Something being scheduled in rapid succession 1. Something being scheduled in rapid succession
2. A single work item that consumes lots of cpu cycles 2. A single work item that consumes lots of cpu cycles
The first one can be tracked using tracing: The first one can be tracked using tracing: ::
$ echo workqueue:workqueue_queue_work > /sys/kernel/debug/tracing/set_event $ echo workqueue:workqueue_queue_work > /sys/kernel/debug/tracing/set_event
$ cat /sys/kernel/debug/tracing/trace_pipe > out.txt $ cat /sys/kernel/debug/tracing/trace_pipe > out.txt
@ -380,9 +380,15 @@ the output and the offender can be determined with the work item
function. function.
For the second type of problems it should be possible to just check For the second type of problems it should be possible to just check
the stack trace of the offending worker thread. the stack trace of the offending worker thread. ::
$ cat /proc/THE_OFFENDING_KWORKER/stack $ cat /proc/THE_OFFENDING_KWORKER/stack
The work item's function should be trivially visible in the stack The work item's function should be trivially visible in the stack
trace. trace.
Kernel Inline Documentations Reference
======================================
.. kernel-doc:: include/linux/workqueue.h

2
MAINTAINERS
View File

@ -13101,7 +13101,7 @@ T: git git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git
S: Maintained S: Maintained
F: include/linux/workqueue.h F: include/linux/workqueue.h
F: kernel/workqueue.c F: kernel/workqueue.c
F: Documentation/workqueue.txt F: Documentation/core-api/workqueue.rst
X-POWERS MULTIFUNCTION PMIC DEVICE DRIVERS X-POWERS MULTIFUNCTION PMIC DEVICE DRIVERS
M: Chen-Yu Tsai <wens@csie.org> M: Chen-Yu Tsai <wens@csie.org>