Re: kdump regression compared to v2.6.35

[Tejun Heo]

Can you please try the following patch?

Thanks.

diff --git a/kernel/workqueue.c b/kernel/workqueue.c
index a2dccfc..f57cd6e 100644
--- a/kernel/workqueue.c
+++ b/kernel/workqueue.c
@@ -1224,6 +1224,8 @@ __acquires(&gcwq->lock)
 {
 	struct global_cwq *gcwq = worker->gcwq;
 	struct task_struct *task = worker->task;
+	static unsigned int cnt;
+	int rc;

 	while (true) {
 		/*
@@ -1232,8 +1234,11 @@ __acquires(&gcwq->lock)
 		 * it races with cpu hotunplug operation.  Verify
 		 * against GCWQ_DISASSOCIATED.
 		 */
-		if (!(gcwq->flags & GCWQ_DISASSOCIATED))
-			set_cpus_allowed_ptr(task, get_cpu_mask(gcwq->cpu));
+		if (!(gcwq->flags & GCWQ_DISASSOCIATED)) {
+			rc = set_cpus_allowed_ptr(task, get_cpu_mask(gcwq->cpu));
+			if (rc && ++cnt < 10)
+				printk("XXX set_cpus_allowed_ptr() failed w/ %d\n", rc);
+		}

 		spin_lock_irq(&gcwq->lock);
 		if (gcwq->flags & GCWQ_DISASSOCIATED)
@@ -1985,13 +1990,16 @@ repeat:
 		struct cpu_workqueue_struct *cwq = get_cwq(tcpu, wq);
 		struct global_cwq *gcwq = cwq->gcwq;
 		struct work_struct *work, *n;
+		bool bound;

 		__set_current_state(TASK_RUNNING);
 		mayday_clear_cpu(cpu, wq->mayday_mask);

 		/* migrate to the target cpu if possible */
 		rescuer->gcwq = gcwq;
-		worker_maybe_bind_and_lock(rescuer);
+		printk("XXX %s: rescuer dispatching to cpu%u\n", wq->name, gcwq->cpu);
+		bound = worker_maybe_bind_and_lock(rescuer);
+		printk("XXX %s: rescuer done binding, bound=%d\n", wq->name, bound);

 		/*
 		 * Slurp in all works issued via this workqueue and
@@ -3558,8 +3566,7 @@ static int __init init_workqueues(void)
 		spin_lock_init(&gcwq->lock);
 		INIT_LIST_HEAD(&gcwq->worklist);
 		gcwq->cpu = cpu;
-		if (cpu == WORK_CPU_UNBOUND)
-			gcwq->flags |= GCWQ_DISASSOCIATED;
+		gcwq->flags |= GCWQ_DISASSOCIATED;

 		INIT_LIST_HEAD(&gcwq->idle_list);
 		for (i = 0; i < BUSY_WORKER_HASH_SIZE; i++)
@@ -3583,6 +3590,7 @@ static int __init init_workqueues(void)
 		struct global_cwq *gcwq = get_gcwq(cpu);
 		struct worker *worker;

+		gcwq->flags &= ~GCWQ_DISASSOCIATED;
 		worker = create_worker(gcwq, true);
 		BUG_ON(!worker);
 		spin_lock_irq(&gcwq->lock);
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/