Merge tag 'wq-for-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq

Pull workqueue updates from Tejun Heo:

 - Work items can now be disabled and enabled, and cancel_work_sync()
   and disable_work() can be called form atomic contexts for BH work
   items.

   This closes feature gap with tasklet and should allow converting all
   existing tasklet users to BH workqueues.

 - Improve pool sharing for unbound workqueues with strict affinity.

 - Misc changes including doc updates, improved debug annotations and
   cleanups.

* tag 'wq-for-6.10' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: Use "@..." in function comment to describe variable length argument
  workqueue: Add destroy_work_on_stack() in workqueue_softirq_dead()
  workqueue: remove unnecessary import and function in wq_monitor.py
  workqueue: Introduce enable_and_queue_work() convenience function
  workqueue: add function in event of workqueue_activate_work
  workqueue: Cleanup subsys attribute registration
  workqueue: Use list_last_entry() to get the last idle worker
  workqueue: Move attrs->cpumask out of worker_pool's properties when attrs->affn_strict
  workqueue: Use INIT_WORK_ONSTACK in workqueue_softirq_dead()
  workqueue: Allow cancel_work_sync() and disable_work() from atomic contexts on BH work items
  workqueue: Remember whether a work item was on a BH workqueue
  workqueue: Remove WORK_OFFQ_CANCELING
  workqueue: Implement disable/enable for (delayed) work items
  workqueue: Preserve OFFQ bits in cancel[_sync] paths

include/linux/workqueue.h

@@ -51,20 +51,23 @@ enum work_bits {
 	 * data contains off-queue information when !WORK_STRUCT_PWQ.
 	 *
 	 * MSB
-	 * [ pool ID ] [ OFFQ flags ] [ STRUCT flags ]
-	 *                 1 bit        4 or 5 bits
+	 * [ pool ID ] [ disable depth ] [ OFFQ flags ] [ STRUCT flags ]
+	 *                  16 bits          1 bit        4 or 5 bits
 	 */
 	WORK_OFFQ_FLAG_SHIFT	= WORK_STRUCT_FLAG_BITS,
-	WORK_OFFQ_CANCELING_BIT = WORK_OFFQ_FLAG_SHIFT,
+	WORK_OFFQ_BH_BIT	= WORK_OFFQ_FLAG_SHIFT,
 	WORK_OFFQ_FLAG_END,
 	WORK_OFFQ_FLAG_BITS	= WORK_OFFQ_FLAG_END - WORK_OFFQ_FLAG_SHIFT,
 
+	WORK_OFFQ_DISABLE_SHIFT	= WORK_OFFQ_FLAG_SHIFT + WORK_OFFQ_FLAG_BITS,
+	WORK_OFFQ_DISABLE_BITS	= 16,
+
 	/*
 	 * When a work item is off queue, the high bits encode off-queue flags
 	 * and the last pool it was on. Cap pool ID to 31 bits and use the
 	 * highest number to indicate that no pool is associated.
 	 */
-	WORK_OFFQ_POOL_SHIFT	= WORK_OFFQ_FLAG_SHIFT + WORK_OFFQ_FLAG_BITS,
+	WORK_OFFQ_POOL_SHIFT	= WORK_OFFQ_DISABLE_SHIFT + WORK_OFFQ_DISABLE_BITS,
 	WORK_OFFQ_LEFT		= BITS_PER_LONG - WORK_OFFQ_POOL_SHIFT,
 	WORK_OFFQ_POOL_BITS	= WORK_OFFQ_LEFT <= 31 ? WORK_OFFQ_LEFT : 31,
 };
@@ -96,7 +99,9 @@ enum wq_misc_consts {
 };
 
 /* Convenience constants - of type 'unsigned long', not 'enum'! */
-#define WORK_OFFQ_CANCELING	(1ul << WORK_OFFQ_CANCELING_BIT)
+#define WORK_OFFQ_BH		(1ul << WORK_OFFQ_BH_BIT)
+#define WORK_OFFQ_FLAG_MASK	(((1ul << WORK_OFFQ_FLAG_BITS) - 1) << WORK_OFFQ_FLAG_SHIFT)
+#define WORK_OFFQ_DISABLE_MASK	(((1ul << WORK_OFFQ_DISABLE_BITS) - 1) << WORK_OFFQ_DISABLE_SHIFT)
 #define WORK_OFFQ_POOL_NONE	((1ul << WORK_OFFQ_POOL_BITS) - 1)
 #define WORK_STRUCT_NO_POOL	(WORK_OFFQ_POOL_NONE << WORK_OFFQ_POOL_SHIFT)
 #define WORK_STRUCT_PWQ_MASK	(~((1ul << WORK_STRUCT_PWQ_SHIFT) - 1))
@@ -180,6 +185,9 @@ struct workqueue_attrs {
 	 * Below fields aren't properties of a worker_pool. They only modify how
 	 * :c:func:`apply_workqueue_attrs` select pools and thus don't
 	 * participate in pool hash calculations or equality comparisons.
+	 *
+	 * If @affn_strict is set, @cpumask isn't a property of a worker_pool
+	 * either.
 	 */
 
 	/**
@@ -465,7 +473,7 @@ void workqueue_softirq_dead(unsigned int cpu);
  * @fmt: printf format for the name of the workqueue
  * @flags: WQ_* flags
  * @max_active: max in-flight work items, 0 for default
- * remaining args: args for @fmt
+ * @...: args for @fmt
  *
  * For a per-cpu workqueue, @max_active limits the number of in-flight work
  * items for each CPU. e.g. @max_active of 1 indicates that each CPU can be
@@ -559,6 +567,14 @@ extern bool flush_delayed_work(struct delayed_work *dwork);
 extern bool cancel_delayed_work(struct delayed_work *dwork);
 extern bool cancel_delayed_work_sync(struct delayed_work *dwork);
 
+extern bool disable_work(struct work_struct *work);
+extern bool disable_work_sync(struct work_struct *work);
+extern bool enable_work(struct work_struct *work);
+
+extern bool disable_delayed_work(struct delayed_work *dwork);
+extern bool disable_delayed_work_sync(struct delayed_work *dwork);
+extern bool enable_delayed_work(struct delayed_work *dwork);
+
 extern bool flush_rcu_work(struct rcu_work *rwork);
 
 extern void workqueue_set_max_active(struct workqueue_struct *wq,
@@ -666,6 +682,32 @@ static inline bool schedule_work(struct work_struct *work)
 	return queue_work(system_wq, work);
 }
 
+/**
+ * enable_and_queue_work - Enable and queue a work item on a specific workqueue
+ * @wq: The target workqueue
+ * @work: The work item to be enabled and queued
+ *
+ * This function combines the operations of enable_work() and queue_work(),
+ * providing a convenient way to enable and queue a work item in a single call.
+ * It invokes enable_work() on @work and then queues it if the disable depth
+ * reached 0. Returns %true if the disable depth reached 0 and @work is queued,
+ * and %false otherwise.
+ *
+ * Note that @work is always queued when disable depth reaches zero. If the
+ * desired behavior is queueing only if certain events took place while @work is
+ * disabled, the user should implement the necessary state tracking and perform
+ * explicit conditional queueing after enable_work().
+ */
+static inline bool enable_and_queue_work(struct workqueue_struct *wq,
+					 struct work_struct *work)
+{
+	if (enable_work(work)) {
+		queue_work(wq, work);
+		return true;
+	}
+	return false;
+}
+
 /*
  * Detect attempt to flush system-wide workqueues at compile time when possible.
  * Warn attempt to flush system-wide workqueues at runtime.

include/trace/events/workqueue.h

@@ -64,13 +64,15 @@ TRACE_EVENT(workqueue_activate_work,
 
 	TP_STRUCT__entry(
 		__field( void *,	work	)
+		__field( void *,	function)
 	),
 
 	TP_fast_assign(
 		__entry->work		= work;
+		__entry->function	= work->func;
 	),
 
-	TP_printk("work struct %p", __entry->work)
+	TP_printk("work struct %p function=%ps ", __entry->work, __entry->function)
 );
 
 /**

tools/workqueue/wq_monitor.py

@@ -32,16 +32,13 @@ https://github.com/osandov/drgn.
   rescued  The number of work items executed by the rescuer.
 """
 
-import sys
 import signal
-import os
 import re
 import time
 import json
 
 import drgn
-from drgn.helpers.linux.list import list_for_each_entry,list_empty
-from drgn.helpers.linux.cpumask import for_each_possible_cpu
+from drgn.helpers.linux.list import list_for_each_entry
 
 import argparse
 parser = argparse.ArgumentParser(description=desc,
@@ -54,10 +51,6 @@ parser.add_argument('-j', '--json', action='store_true',
                     help='Output in json')
 args = parser.parse_args()
 
-def err(s):
-    print(s, file=sys.stderr, flush=True)
-    sys.exit(1)
-
 workqueues              = prog['workqueues']
 
 WQ_UNBOUND              = prog['WQ_UNBOUND']