linux-toradex.git/kernel/workqueue.c, branch v5.19-rc8

workqueue: Wrap flush_workqueue() using a macro

2022-06-07T17:07:14+00:00

Since flush operation synchronously waits for completion, flushing
system-wide WQs (e.g. system_wq) might introduce possibility of deadlock
due to unexpected locking dependency. Tejun Heo commented at [1] that it
makes no sense at all to call flush_workqueue() on the shared WQs as the
caller has no idea what it's gonna end up waiting for.

Although there is flush_scheduled_work() which flushes system_wq WQ with
"Think twice before calling this function! It's very easy to get into
trouble if you don't take great care." warning message, syzbot found a
circular locking dependency caused by flushing system_wq WQ [2].

Therefore, let's change the direction to that developers had better use
their local WQs if flush_scheduled_work()/flush_workqueue(system_*_wq) is
inevitable.

Steps for converting system-wide WQs into local WQs are explained at [3],
and a conversion to stop flushing system-wide WQs is in progress. Now we
want some mechanism for preventing developers who are not aware of this
conversion from again start flushing system-wide WQs.

Since I found that WARN_ON() is complete but awkward approach for teaching
developers about this problem, let's use __compiletime_warning() for
incomplete but handy approach. For completeness, we will also insert
WARN_ON() into __flush_workqueue() after all in-tree users stopped calling
flush_scheduled_work().

Link: https://lore.kernel.org/all/YgnQGZWT%2Fn3VAITX@slm.duckdns.org/ [1]
Link: https://syzkaller.appspot.com/bug?extid=bde0f89deacca7c765b8 [2]
Link: https://lkml.kernel.org/r/49925af7-78a8-a3dd-bce6-cfc02e1a9236@I-love.SAKURA.ne.jp [3]
Signed-off-by: Tetsuo Handa 
Signed-off-by: Tejun Heo

workqueue: Restrict kworker in the offline CPU pool running on housekeeping CPUs

2022-04-21T22:31:04+00:00

When a CPU is going offline, all workers on the CPU's pool will have their
cpus_allowed cleared to cpu_possible_mask and can run on any CPUs including
the isolated ones. Instead, set cpus_allowed to wq_unbound_cpumask so that
the can avoid isolated CPUs.

Signed-off-by: Zqiang 
Signed-off-by: Tejun Heo

Merge branch 'for-5.18' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq

2022-03-23T19:40:51+00:00

Pull workqueue updates from Tejun Heo:
 "Nothing major. Just follow-up cleanups from Lai after the earlier
  synchronization simplification"

* 'for-5.18' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: Convert the type of pool->nr_running to int
  workqueue: Use wake_up_worker() in wq_worker_sleeping() instead of open code
  workqueue: Change the comments of the synchronization about the idle_list
  workqueue: Remove the mb() pair between wq_worker_sleeping() and insert_work()

sched/isolation: Use single feature type while referring to housekeeping cpumask

2022-02-16T14:57:55+00:00

Refer to housekeeping APIs using single feature types instead of flags.
This prevents from passing multiple isolation features at once to
housekeeping interfaces, which soon won't be possible anymore as each
isolation features will have their own cpumask.

Signed-off-by: Frederic Weisbecker 
Signed-off-by: Peter Zijlstra (Intel) 
Reviewed-by: Juri Lelli 
Reviewed-by: Phil Auld 
Link: https://lore.kernel.org/r/20220207155910.527133-5-frederic@kernel.org

workqueue: Decouple HK_FLAG_WQ and HK_FLAG_DOMAIN cpumask fetch

2022-02-16T14:57:54+00:00

To prepare for supporting each feature of the housekeeping cpumask
toward cpuset, prepare each of the HK_FLAG_* entries to move to their
own cpumask with enforcing to fetch them individually. The new
constraint is that multiple HK_FLAG_* entries can't be mixed together
anymore in a single call to housekeeping cpumask().

This will later allow, for example, to runtime modify the cpulist passed
through "isolcpus=", "nohz_full=" and "rcu_nocbs=" kernel boot
parameters.

Signed-off-by: Frederic Weisbecker 
Signed-off-by: Peter Zijlstra (Intel) 
Reviewed-by: Juri Lelli 
Reviewed-by: Phil Auld 
Acked-by: Tejun Heo 
Link: https://lore.kernel.org/r/20220207155910.527133-3-frederic@kernel.org

workqueue: Convert the type of pool->nr_running to int

2022-01-12T17:46:36+00:00

It is only modified in associated CPU, so it doesn't need to be atomic.

tj: Comment updated.

Signed-off-by: Lai Jiangshan 
Signed-off-by: Tejun Heo

workqueue: Use wake_up_worker() in wq_worker_sleeping() instead of open code

2022-01-12T17:40:32+00:00

The wakeup code in wq_worker_sleeping() is the same as wake_up_worker().

Signed-off-by: Lai Jiangshan 
Signed-off-by: Tejun Heo

workqueue: Change the comments of the synchronization about the idle_list

2022-01-12T17:39:15+00:00

The access to idle_list in wq_worker_sleeping() is changed to be
protected by pool->lock, so the comments above idle_list can be changed
to "L:" which is the meaning of "access with pool->lock held".

And the outdated comments in wq_worker_sleeping() is removed since
the function is not called with rq lock held any more, idle_list is
dereferenced with pool lock now.

Signed-off-by: Lai Jiangshan 
Signed-off-by: Tejun Heo

workqueue: Remove the mb() pair between wq_worker_sleeping() and insert_work()

2022-01-12T17:37:28+00:00

In wq_worker_sleeping(), the access to worklist is protected by the
pool->lock, so the memory barrier is unneeded.

Signed-off-by: Lai Jiangshan 
Signed-off-by: Tejun Heo

Merge branch 'workqueue/for-5.16-fixes' into workqueue/for-5.17

2022-01-10T17:54:04+00:00

for-5.16-fixes contains two subtle race conditions which were introduced by
scheduler side code cleanups. The branch didn't get pushed out, so merge
into for-5.17.