Re: [PATCH 05/10] vfio: ccw: Suppress unused event parameter

[Pierre Morel]

On 26/04/2018 09:36, Dong Jia Shi wrote:
> * Pierre Morel <pmorel@xxxxxxxxxxxxxxxxxx> [2018-04-19 16:48:08 +0200]:
>
> > The fsm_func_t function type definition does not need the event
> > parameter since all functions are in a state/event table.
> Not sure if this will still be right after you addressed all of the
> comments on this version. But no hurt to point out some nits that I
> noticed below.
>
> > Signed-off-by: Pierre Morel <pmorel@xxxxxxxxxxxxxxxxxx>
> > ---
> >   drivers/s390/cio/vfio_ccw_fsm.c     | 34 +++++++++++-----------------------
> >   drivers/s390/cio/vfio_ccw_private.h |  5 ++---
> >   2 files changed, 13 insertions(+), 26 deletions(-)
> >
> > diff --git a/drivers/s390/cio/vfio_ccw_fsm.c b/drivers/s390/cio/vfio_ccw_fsm.c
> > index f9855cd..f8ded70 100644
> > --- a/drivers/s390/cio/vfio_ccw_fsm.c
> > +++ b/drivers/s390/cio/vfio_ccw_fsm.c
> > @@ -55,8 +55,7 @@ static int fsm_io_helper(struct vfio_ccw_private *private)
> >   	}
> >   }
> >
> > -static int fsm_notoper(struct vfio_ccw_private *private,
> > -			enum vfio_ccw_event event)
> > +static int fsm_notoper(struct vfio_ccw_private *private)
> >   {
> >   	struct subchannel *sch = private->sch;
> >
> > @@ -65,36 +64,31 @@ static int fsm_notoper(struct vfio_ccw_private *private,
> >   	 * Probably we should send the machine check to the guest.
> >   	 */
> >   	css_sched_sch_todo(sch, SCH_TODO_UNREG);
> > -	private->state = VFIO_CCW_STATE_NOT_OPER;
> > -	return private->state;
> > +	return VFIO_CCW_STATE_NOT_OPER;
> This change should belong to a former patch.
>
> [...]
>
> > @@ -109,17 +103,14 @@ static int fsm_disabled_irq(struct vfio_ccw_private *private,
> >   /*
> >    * Deal with the ccw command request from the userspace.
> >    */
> > -static int fsm_io_request(struct vfio_ccw_private *private,
> > -			   enum vfio_ccw_event event)
> > +static int fsm_io_request(struct vfio_ccw_private *private)
> >   {
> > -	union orb *orb;
> >   	struct ccw_io_region *io_region = &private->io_region;
> > +	union orb *orb = (union orb *)io_region->orb_area;
> >   	struct mdev_device *mdev = private->mdev;
> >
> >   	private->state = VFIO_CCW_STATE_BOXED;
> >
> > -	orb = (union orb *)io_region->orb_area;
> > -
> Irrelevant change for this patch.
>
> >   	io_region->ret_code = cp_init(&private->cp, mdev_dev(mdev), orb);
> >   	if (io_region->ret_code)
> >   		goto err_out;
> > @@ -139,15 +130,13 @@ static int fsm_io_request(struct vfio_ccw_private *private,
> >   	return VFIO_CCW_STATE_BUSY;
> >
> >   err_out:
> > -	private->state = VFIO_CCW_STATE_IDLE;
> > -	return private->state;
> > +	return VFIO_CCW_STATE_IDLE;
> Should move to other patch.
>
> [...]
OK, I will move these changes in patch 4/10.

--
Pierre Morel
Linux/KVM/QEMU in BÃblingen - Germany