Re: [pci PATCH v5 1/4] pci: Add pci_sriov_configure_simple for PFs that don't manage VF resources

[Keith Busch <keith.busch@intel.com>]

On Mon, Mar 12, 2018 at 11:09:34AM -0700, Alexander Duyck wrote:
> On Mon, Mar 12, 2018 at 10:40 AM, Keith Busch <keith.busch@intel.com> wrote:
> > On Mon, Mar 12, 2018 at 10:21:29AM -0700, Alexander Duyck wrote:
> >> diff --git a/include/linux/pci.h b/include/linux/pci.h
> >> index 024a1beda008..9cab9d0d51dc 100644
> >> --- a/include/linux/pci.h
> >> +++ b/include/linux/pci.h
> >> @@ -1953,6 +1953,7 @@ static inline void pci_mmcfg_late_init(void) { }
> >>  int pci_vfs_assigned(struct pci_dev *dev);
> >>  int pci_sriov_set_totalvfs(struct pci_dev *dev, u16 numvfs);
> >>  int pci_sriov_get_totalvfs(struct pci_dev *dev);
> >> +int pci_sriov_configure_simple(struct pci_dev *dev, int nr_virtfn);
> >>  resource_size_t pci_iov_resource_size(struct pci_dev *dev, int resno);
> >>  void pci_vf_drivers_autoprobe(struct pci_dev *dev, bool probe);
> >>  #else
> >
> > I recommend stubbing 'pci_sriov_configure_simple' or defining it to
> > NULL in the '#else' section here so you don't need to repeat the "#ifdef
> > CONFIG_PCI_IOV" in each driver wishing to use this function. Otherwise
> > looks fine to me.
> 
> My concern with defining it as NULL is that somebody may end up
> calling it in the future directly and that may end up causing issues.
> One thought I have been debating is moving it to a different file. I
> am just not sure where the best place to put something like this would
> be. I could move this function to drivers/pci/pci.c if everyone is
> okay with it and then I could just strip the contents out by wrapping
> them in a #ifdef instead.

Okay, instead of NULL, a stub implementation in the header file may
suffice when CONFIG_PCI_IOV is not defined:

static inline int pci_sriov_configure_simple(struct pci_dev *dev, int nr_virtfn)
{
	return -ENOSYS;
}

See pci_iov_virtfn_bus, pci_iov_virtfn_devfn, pci_iov_add_virtfn, or
pci_enable_sriov for other examples.