Currently it calls __blk_mq_run_hw_queue(), which depends on the
CPU placement being correct. This means it's not possible to call
blk_mq_start_hw_queues(q) from a context that is correct for all
queues, leading to triggering the
WARN_ON(!cpumask_test_cpu(raw_smp_processor_id(), hctx->cpumask));
in __blk_mq_run_hw_queue().
Reported-by: Ming Lei <tom.leiming@gmail.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
        clear_bit(BLK_MQ_S_STOPPED, &hctx->state);
 
        preempt_disable();
-       __blk_mq_run_hw_queue(hctx);
+       blk_mq_run_hw_queue(hctx, false);
        preempt_enable();
 }
 EXPORT_SYMBOL(blk_mq_start_hw_queue);