On 1/2/19 5:11 AM, Ming Lei wrote: >> diff --git a/drivers/block/sunvdc.c b/drivers/block/sunvdc.c >> index 9c0553dd13e7..6eddd17a864a 100644 >> --- a/drivers/block/sunvdc.c >> +++ b/drivers/block/sunvdc.c >> @@ -181,7 +181,7 @@ static void vdc_blk_queue_start(struct vdc_port *port) >> * allocated a disk. >> */ >> if (port->disk && vdc_tx_dring_avail(dr) * 100 / VDC_TX_RING_SIZE >= 50) >> - blk_mq_start_hw_queues(port->disk->queue); >> + blk_mq_start_stopped_hw_queue(port->disk->queue, true); >> } >> >> static void vdc_finish(struct vio_driver_state *vio, int err, int waiting_for) > > oops, we should have called blk_mq_start_stopped_hw_queues(): > > diff --git a/drivers/block/sunvdc.c b/drivers/block/sunvdc.c > index 9c0553dd13e7..ab7f7ab4c472 100644 > --- a/drivers/block/sunvdc.c > +++ b/drivers/block/sunvdc.c > @@ -181,7 +181,7 @@ static void vdc_blk_queue_start(struct vdc_port *port) > * allocated a disk. > */ > if (port->disk && vdc_tx_dring_avail(dr) * 100 / VDC_TX_RING_SIZE >= 50) > - blk_mq_start_hw_queues(port->disk->queue); > + blk_mq_start_stopped_hw_queues(port->disk->queue, true); > } > > static void vdc_finish(struct vio_driver_state *vio, int err, int waiting_for) Indeed, that was a silly oversight on my part. Ming, will you send a "proper" patch so I can queue it up? -- Jens Axboe
