On Wed, 12 Dec 2012 01:33:46 +0100
Alexander Graf <[email protected]> wrote:

> 
> On 11.12.2012, at 13:34, Cornelia Huck wrote:
> 
> > On Tue, 11 Dec 2012 11:09:55 +0100
> > Alexander Graf <[email protected]> wrote:
> > 
> >> 
> >> On 10.12.2012, at 10:03, Cornelia Huck wrote:
> >> 
> >>> On Sun, 9 Dec 2012 13:12:37 +0100
> >>> Alexander Graf <[email protected]> wrote:
> >>> 
> >>>> 
> >>>> On 07.12.2012, at 13:29, Cornelia Huck wrote:
> >>>> 
> >>>>> This will be needed by the new virtio-ccw transport.
> >>>>> 
> >>>>> Signed-off-by: Cornelia Huck <[email protected]>
> >>>>> ---
> >>>>> arch/s390/include/asm/ccwdev.h |  5 +++++
> >>>>> drivers/s390/cio/device_ops.c  | 12 ++++++++++++
> >>>>> 2 files changed, 17 insertions(+)
> >>>>> 
> >>>>> diff --git a/arch/s390/include/asm/ccwdev.h 
> >>>>> b/arch/s390/include/asm/ccwdev.h
> >>>>> index 1cb4bb3..9ad79f7 100644
> >>>>> --- a/arch/s390/include/asm/ccwdev.h
> >>>>> +++ b/arch/s390/include/asm/ccwdev.h
> >>>>> @@ -18,6 +18,9 @@ struct irb;
> >>>>> struct ccw1;
> >>>>> struct ccw_dev_id;
> >>>>> 
> >>>>> +/* from asm/schid.h */
> >>>>> +struct subchannel_id;
> >>>>> +
> >>>>> /* simplified initializers for struct ccw_device:
> >>>>> * CCW_DEVICE and CCW_DEVICE_DEVTYPE initialize one
> >>>>> * entry in your MODULE_DEVICE_TABLE and set the match_flag correctly */
> >>>>> @@ -226,5 +229,7 @@ int ccw_device_siosl(struct ccw_device *);
> >>>>> // FIXME: these have to go
> >>>>> extern int _ccw_device_get_subchannel_number(struct ccw_device *);
> >>>>> 
> >>>>> +extern void ccw_device_get_schid(struct ccw_device *, struct 
> >>>>> subchannel_id *);
> >>>>> +
> >>>>> extern void *ccw_device_get_chp_desc(struct ccw_device *, int);
> >>>>> #endif /* _S390_CCWDEV_H_ */
> >>>>> diff --git a/drivers/s390/cio/device_ops.c 
> >>>>> b/drivers/s390/cio/device_ops.c
> >>>>> index ec7fb6d..2ad832f 100644
> >>>>> --- a/drivers/s390/cio/device_ops.c
> >>>>> +++ b/drivers/s390/cio/device_ops.c
> >>>>> @@ -763,6 +763,18 @@ _ccw_device_get_subchannel_number(struct 
> >>>>> ccw_device *cdev)
> >>>>>         return cdev->private->schid.sch_no;
> >>>>> }
> >>>>> 
> >>>>> +/**
> >>>>> + * ccw_device_get_schid - obtain a subchannel id
> >>>>> + * @cdev: device to obtain the id for
> >>>>> + * @schid: where to fill in the values
> >>>>> + */
> >>>>> +void ccw_device_get_schid(struct ccw_device *cdev, struct 
> >>>>> subchannel_id *schid)
> >>>> 
> >>>> If you make this
> >>>> 
> >>>> u32 ccw_device_get_schid(struct ccw_device *cdev)
> >>>> 
> >>>> and you return the u32 (architected) value of the schid, not the 
> >>>> internal struct, then you can save yourself the struct export (patch 
> >>>> 2/5) and the ugly cast in patch 4/5 to get the u32 value.
> >>> 
> >>> I really prefer using the structure instead.
> >>> 
> >>> Moreover, there's a patch based on this that switches non-kvm users to
> >>> this new interface (getting rid of an old, broken interface) already
> >>> queued in the linux-s390 tree AFAIK.
> >> 
> >> Well, then base on top of that patch and add another interface that allows 
> >> you to receive a schid as integer value. Randomly casting structs to u32 
> >> in random code across the tree is just pure ugly.
> > 
> > We seem to have different ideas of 'ugly' :)
> > 
> > I really prefer to have a function called ccw_device_get_schid()
> > provide a struct subchannel_id and not a u32 - a subchannel id is,
> > after all, what the caller asks for. The fact that a subchannel id may
> > fit into a u32 is an implementation detail, as well as the fact that an
> > instruction takes a 32 bit value containing a subchannel id. Passing
> > around a 32 bit value instead of a subchannel id basically discards the
> > information that we're dealing with a subchannel id and not with a
> > random 32 bit value.
> > 
> >> 
> >> An alternative would be to create a union around the struct and to return 
> >> that one, so that you can access the u32 value or the struct value 
> >> depending on the user's needs. That way the u32 cast is part of the API 
> >> and not implicit, as it would be today.
> > 
> > What's the API here? I think we have two parts:
> > 
> > - The API to get the current subchannel id for a ccw device. For the
> > reasons above, I think this should use struct subchannel_id.
> 
> Fine with me.
> 
> But as you are mentioning above the fact that a subchannel id is a u32 
> internally is an implementation detail. Thus casting a struct schid * to a 
> u32 * is violating that exact rule.
> 
> > - The hardware/hypervisor API for addressing a subchannel, which
> > basically requires that a register contains a subchannel id. The easiest
> > way to do this is to cast to an integer. I don't see any need to play
> > games with unions here, after all we're using C and not Pascal ;)
> 
> Then just create a small inline function in the header file that converts a 
> struct schid * to u32. I want to make sure we keep implementation details in 
> files that deal with the implementation details, not in files that deal with 
> using the interfaces.
> 

I checked what the common I/O layer does for the I/O instructions in
drivers/s390/cio/ioasm.h and think that's really the best solution:
pass the subchannel id directly in the register. No information is
lost, no confusing casts, and I think the code looks nicer.

Updated guest patch set coming now.

--
To unsubscribe from this list: send the line "unsubscribe kvm" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to