Re: [PATCH for-5.2 3/6] pc-bios/s390-ccw: Move the inner logic of find_subch() to a separate function

[Cornelia Huck]

On Wed, 29 Jul 2020 13:13:17 +0200
Thomas Huth <thuth@redhat.com> wrote:

> On 29/07/2020 10.54, Cornelia Huck wrote:
> > On Tue, 28 Jul 2020 20:37:31 +0200
> > Thomas Huth <thuth@redhat.com> wrote:
> >   
> >> Move the code to a separate function to be able to re-use it from a
> >> different spot later.
> >>
> >> Signed-off-by: Thomas Huth <thuth@redhat.com>
> >> ---
> >>  pc-bios/s390-ccw/main.c | 99 ++++++++++++++++++++++++-----------------
> >>  1 file changed, 57 insertions(+), 42 deletions(-)
> >>
> >> diff --git a/pc-bios/s390-ccw/main.c b/pc-bios/s390-ccw/main.c
> >> index 9b64eb0c24..9477313188 100644
> >> --- a/pc-bios/s390-ccw/main.c
> >> +++ b/pc-bios/s390-ccw/main.c
> >> @@ -51,6 +51,60 @@ unsigned int get_loadparm_index(void)
> >>      return atoui(loadparm_str);
> >>  }
> >>  
> >> +static int check_sch_no(int dev_no, int sch_no)  
> > 
> > check_subchannel()? You verify dev_no as well, if supplied.  
> 
> Ok.
> 
> >> +{
> >> +    bool is_virtio;
> >> +    Schib schib;
> >> +    int r;
> >> +
> >> +    blk_schid.sch_no = sch_no;
> >> +    r = stsch_err(blk_schid, &schib);
> >> +    if (r == 3 || r == -EIO) {
> >> +        return -EIO;  
> > 
> > -ENODEV? It means that you either have no devices, or an invalid
> > subchannel set.  
> 
> We don't have ENODEV in the s390-ccw bios... but I could introduce it, I
> guess :-)

I always forget that we have only a subset of error codes here :)

ENODEV looks like a reasonable value to have, though.