Re: [qemu-s390x] [PATCH 10/15] s390-bios: Support for running format-0/1 channel programs

[Cornelia Huck]

On Mon, 14 Jan 2019 13:44:12 -0500
"Jason J. Herne" <address@hidden> wrote:

> On 1/7/19 2:02 PM, Jason J. Herne wrote:
> >>> @@ -190,6 +247,9 @@ struct ciw {
> >>>       __u16 count;
> >>>   };
> >>> +#define CU_TYPE_VIRTIO          0x3832
> >>> +#define CU_TYPE_DASD            0x3990  
> >>
> >> No other dasd types we want to support? :) (Not sure if others are out
> >> in the wild. Maybe FBA?)
> >>  
> > 
> > I have no idea. I assumed 3390 was the only thing we supported. Perhaps 
> > 3380? I'd need to 
> > find a test device, which I could probably do ... I'll look more into this. 
> >  
> 
> After a few discussions with folks in the lab we've decided that we don't see 
> a ton of 
> value in supporting anything other than 3990 at the moment. Anything else 
> would be older 
> (3380) and/or rare to see in the wild (and very difficult to test). As for 
> emulated setups 
> like z/VM, a user can just use 3390 instead of FBA. So I recommend we move 
> forward with 
> 3390/3990 support for now. We can always add in others types if/when we need 
> them.

Sounds reasonable.

What about calling the #define above CU_TYPE_DASD_3990 instead? Just to
make clear that there are other dasd types out there, but we only
support that particular one (at least at the moment).