Re: [Qemu-devel] [PATCH v5 3/9] s390x/pci: add stubs

[Cornelia Huck]

On Thu, 24 Aug 2017 11:50:54 +0200
Halil Pasic <address@hidden> wrote:

> On 08/24/2017 11:09 AM, Cornelia Huck wrote:
> > On Thu, 24 Aug 2017 09:43:48 +0200
> > Christian Borntraeger <address@hidden> wrote:
> >   
> >> On 08/24/2017 09:38 AM, Christian Borntraeger wrote:  
> >>>
> >>>
> >>> On 08/23/2017 05:54 PM, Cornelia Huck wrote:    
> >>>> Some non-pci code calls into zpci code. Provide some stubs for builds
> >>>> without pci.
> >>>>
> >>>> Reviewed-by: Thomas Huth <address@hidden>
> >>>> Signed-off-by: Cornelia Huck <address@hidden>
> >>>> ---
> >>>>  hw/s390x/Makefile.objs   |  3 +-
> >>>>  hw/s390x/s390-pci-stub.c | 74 
> >>>> ++++++++++++++++++++++++++++++++++++++++++++++++
> >>>>  2 files changed, 76 insertions(+), 1 deletion(-)
> >>>>  create mode 100644 hw/s390x/s390-pci-stub.c  
> >   
> >>>> +/* hw/s390x/sclp.c */
> >>>> +void s390_pci_sclp_configure(SCCB *sccb)
> >>>> +{
> >>>> +}
> >>>> +
> >>>> +void s390_pci_sclp_deconfigure(SCCB *sccb)
> >>>> +{
> >>>> +}    
> >>>
> >>> shouldnt these function set an error code in the sccb, e.g.
> >>> something like
> >>>
> >>>    sccb->h.response_code = cpu_to_be16(SCLP_RC_INVALID_SCLP_COMMAND);
> >>>
> >>>
> >>>     
> >>
> >> Oh you have something like that in patch 7. Maybe move?  
> > 
> > Does not really change anything in practice, but I can move it.
> >   
> 
> You mean these stubs are not supposed to be reachable and are just
> for making the linker happy, or? If that's the case I would prefer
> having that expressed by something like assert(false) or even 
> #define NOT_REACHABLE assert(false).

Without the last patch that makes virtio-pci depend on pci for s390x,
you can't compile without pci anyway ;)

> 
> Otherwise patch looks good, but I did not a full review on it,
> so let's try this:
> Acked-by: Halil Pasic <address@hidden>
> 

Thanks!