diff mbox series

[PATCH-wq,v2,3/5] workqueue: Thaw frozen pwq in workqueue_apply_unbound_cpumask()

Message ID 20240203154334.791910-4-longman@redhat.com
State New
Headers
Series workqueue: Enable unbound cpumask update on ordered workqueues |

Commit Message

Waiman Long Feb. 3, 2024, 3:43 p.m. UTC 
  workqueue_apply_unbound_cpumask() cannot proceed with an ordered
workqueue if its dfl_pwq is still frozen. Just do a sleep wait for
it to be thawed may not work in some cases if pwq_release_workfn() is
somehow prevented from being called due to resources (e.g. wq_pool_mutex)
that are held by its caller.

To break the logjam, we have to actively check if the frozen dfl_pwq
is ready to be thawed and call thaw_pwq() directly if so.

Signed-off-by: Waiman Long <longman@redhat.com>
---
 kernel/workqueue.c | 30 ++++++++++++++++++++++++++++--
 1 file changed, 28 insertions(+), 2 deletions(-)

Comments

kernel test robot Feb. 4, 2024, 10:15 a.m. UTC | #1 
Hi Waiman,

kernel test robot noticed the following build warnings:

[auto build test WARNING on tj-wq/for-next]
[also build test WARNING on next-20240202]
[cannot apply to linus/master v6.8-rc2]
[If your patch is applied to the wrong git tree, kindly drop us a note.
And when submitting patch, we suggest to use '--base' as documented in
https://git-scm.com/docs/git-format-patch#_base_tree_information]

url:    https://github.com/intel-lab-lkp/linux/commits/Waiman-Long/workqueue-Skip-__WQ_DESTROYING-workqueues-when-updating-global-unbound-cpumask/20240203-234626
base:   https://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git for-next
patch link:    https://lore.kernel.org/r/20240203154334.791910-4-longman%40redhat.com
patch subject: [PATCH-wq v2 3/5] workqueue: Thaw frozen pwq in workqueue_apply_unbound_cpumask()
config: x86_64-randconfig-122-20240204 (https://download.01.org/0day-ci/archive/20240204/202402041854.YeHAF3wV-lkp@intel.com/config)
compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240204/202402041854.YeHAF3wV-lkp@intel.com/reproduce)

If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp@intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202402041854.YeHAF3wV-lkp@intel.com/

sparse warnings: (new ones prefixed by >>)
   kernel/workqueue.c:361:40: sparse: sparse: duplicate [noderef]
   kernel/workqueue.c:361:40: sparse: sparse: multiple address spaces given: __percpu & __rcu
>> kernel/workqueue.c:6373:25: sparse: sparse: incompatible types in comparison expression (different address spaces):
   kernel/workqueue.c:6373:25: sparse:    struct pool_workqueue *
   kernel/workqueue.c:6373:25: sparse:    struct pool_workqueue [noderef] __rcu *

vim +6373 kernel/workqueue.c

  6356	
  6357	/*
  6358	 * Check the given ordered workqueue to see if its non-default pwq's have
  6359	 * zero reference count and if so thaw the frozen default pwq.
  6360	 *
  6361	 * Return:
  6362	 * %true if dfl_pwq has been thawed or %false otherwise.
  6363	 */
  6364	static bool ordered_workqueue_ref_check(struct workqueue_struct *wq)
  6365	{
  6366		int refs = 0;
  6367		struct pool_workqueue *pwq;
  6368	
  6369		if (!READ_ONCE(wq->dfl_pwq->frozen))
  6370			return true;
  6371		mutex_lock(&wq->mutex);
  6372		for_each_pwq(pwq, wq) {
> 6373			if (pwq == wq->dfl_pwq)
  6374				continue;
  6375			refs += pwq->refcnt;
  6376		}
  6377		if (!refs)
  6378			thaw_pwq(wq->dfl_pwq);
  6379		mutex_unlock(&wq->mutex);
  6380		return !refs;
  6381	}
  6382
Waiman Long Feb. 4, 2024, 4:07 p.m. UTC | #2 
On 2/4/24 05:15, kernel test robot wrote:
> Hi Waiman,
>
> kernel test robot noticed the following build warnings:
>
> [auto build test WARNING on tj-wq/for-next]
> [also build test WARNING on next-20240202]
> [cannot apply to linus/master v6.8-rc2]
> [If your patch is applied to the wrong git tree, kindly drop us a note.
> And when submitting patch, we suggest to use '--base' as documented in
> https://git-scm.com/docs/git-format-patch#_base_tree_information]
>
> url:    https://github.com/intel-lab-lkp/linux/commits/Waiman-Long/workqueue-Skip-__WQ_DESTROYING-workqueues-when-updating-global-unbound-cpumask/20240203-234626
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/tj/wq.git for-next
> patch link:    https://lore.kernel.org/r/20240203154334.791910-4-longman%40redhat.com
> patch subject: [PATCH-wq v2 3/5] workqueue: Thaw frozen pwq in workqueue_apply_unbound_cpumask()
> config: x86_64-randconfig-122-20240204 (https://download.01.org/0day-ci/archive/20240204/202402041854.YeHAF3wV-lkp@intel.com/config)
> compiler: gcc-9 (Debian 9.3.0-22) 9.3.0
> reproduce (this is a W=1 build): (https://download.01.org/0day-ci/archive/20240204/202402041854.YeHAF3wV-lkp@intel.com/reproduce)
>
> If you fix the issue in a separate patch/commit (i.e. not just a new version of
> the same patch/commit), kindly add following tags
> | Reported-by: kernel test robot <lkp@intel.com>
> | Closes: https://lore.kernel.org/oe-kbuild-all/202402041854.YeHAF3wV-lkp@intel.com/
>
> sparse warnings: (new ones prefixed by >>)
>     kernel/workqueue.c:361:40: sparse: sparse: duplicate [noderef]
>     kernel/workqueue.c:361:40: sparse: sparse: multiple address spaces given: __percpu & __rcu
>>> kernel/workqueue.c:6373:25: sparse: sparse: incompatible types in comparison expression (different address spaces):
>     kernel/workqueue.c:6373:25: sparse:    struct pool_workqueue *
>     kernel/workqueue.c:6373:25: sparse:    struct pool_workqueue [noderef] __rcu *

OK, I didn't realize that a __rcu tag is added to the dfl_pwq in 6.9. 
Will change the patch series to use the appropriate helpers to avoid 
this kind of warnings.

Cheers,
Longman
diff mbox series

Patch

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index f089e532758a..ee934c2c6ea8 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -6354,6 +6354,32 @@  void thaw_workqueues(void)
 }
 #endif /* CONFIG_FREEZER */
 
+/*
+ * Check the given ordered workqueue to see if its non-default pwq's have
+ * zero reference count and if so thaw the frozen default pwq.
+ *
+ * Return:
+ * %true if dfl_pwq has been thawed or %false otherwise.
+ */
+static bool ordered_workqueue_ref_check(struct workqueue_struct *wq)
+{
+	int refs = 0;
+	struct pool_workqueue *pwq;
+
+	if (!READ_ONCE(wq->dfl_pwq->frozen))
+		return true;
+	mutex_lock(&wq->mutex);
+	for_each_pwq(pwq, wq) {
+		if (pwq == wq->dfl_pwq)
+			continue;
+		refs += pwq->refcnt;
+	}
+	if (!refs)
+		thaw_pwq(wq->dfl_pwq);
+	mutex_unlock(&wq->mutex);
+	return !refs;
+}
+
 static int workqueue_apply_unbound_cpumask(const cpumask_var_t unbound_cpumask)
 {
 	LIST_HEAD(ctxs);
@@ -6378,12 +6404,12 @@  static int workqueue_apply_unbound_cpumask(const cpumask_var_t unbound_cpumask)
 
 			if (!(wq->flags & __WQ_ORDERED_EXPLICIT)) {
 				wq->flags &= ~__WQ_ORDERED;
-			} else if (pwq && pwq->frozen) {
+			} else if (pwq && !ordered_workqueue_ref_check(wq)) {
 				int i;
 
 				for (i = 0; i < 10; i++) {
 					msleep(10);
-					if (!pwq->frozen)
+					if (ordered_workqueue_ref_check(wq))
 						break;
 				}
 				if (WARN_ON_ONCE(pwq->frozen))