RE: [PATCH v7 00/27] vfio: Adopt iommufd

[Duan, Zhenzhong]

>-----Original Message-----
>From: Cédric Le Goater <clg@redhat.com>
>Sent: Wednesday, November 22, 2023 4:07 PM
>Subject: Re: [PATCH v7 00/27] vfio: Adopt iommufd
>
>
>>> The series is pushed on top of vfio-next in the vfio-8.2 tree :
>>>
>>>    https://github.com/legoater/qemu/commits/vfio-8.2
>>>
>>> with a little extra to deal with a PPC build failure.
>>
>> Thanks Cédric. That's strange I didn't see this failure on my env
>> which has CONFIG_VFIO_PCI enabled by default for PPC.
>
>
>The compile issue shows with --without-default-devices.
>
>VFIO is always selected (see ppc/Kconfig) but VFIO_PCI is not when
>--without-default-devices is used. Hence the compile failure because
>of the common vfio-pci routines exported in pci.c.

Clear, thanks

>
>Ideally, we should use an 'imply VFIO' in ppc/Kconfig because VFIO
>is not a required subsystem for the pseries machine. If that was
>the case, we wouldn't compile the VFIO EEH hooks defined in
>hw/ppc/spapr_pci_vfio.c :
>
>   bool spapr_phb_eeh_available(SpaprPhbState *sphb);
>   int spapr_phb_vfio_eeh_set_option(SpaprPhbState *sphb,
>                                     unsigned int addr, int option);
>   int spapr_phb_vfio_eeh_get_state(SpaprPhbState *sphb, int *state);
>   int spapr_phb_vfio_eeh_reset(SpaprPhbState *sphb, int option);
>   int spapr_phb_vfio_eeh_configure(SpaprPhbState *sphb);
>   void spapr_phb_vfio_reset(DeviceState *qdev);

Indeed, I reproduced same after some try out.
Maybe need a stub file for those functions.

>
>It is not that simple to fix. The simpler approach is to force compile
>of VFIO PCI in ppc/Kconfig with a 'select VFIO_PCI'. we should improve
>that.

Yes, not a blocking issue after your fix. I can have a try when I'm idle.

Thanks
Zhenzhong