Re: [PATCH v11 04/11] PCI: dwc: Move devm_pci_alloc_host_bridge() to the beginning of dw_pcie_host_init()

[Bjorn Helgaas]

On Thu, Mar 13, 2025 at 04:45:26PM -0400, Frank Li wrote:
> On Thu, Mar 13, 2025 at 02:22:54PM -0500, Bjorn Helgaas wrote:
> > On Thu, Mar 13, 2025 at 11:38:40AM -0400, Frank Li wrote:
> > > Move devm_pci_alloc_host_bridge() to the beginning of dw_pcie_host_init().
> > > Since devm_pci_alloc_host_bridge() is common code that doesn't depend on
> > > any DWC resource, moving it earlier improves code logic and readability.
> > >
> > > Signed-off-by: Frank Li <Frank.Li@xxxxxxx>
> > > ---
> > >  drivers/pci/controller/dwc/pcie-designware-host.c | 12 ++++++------
> > >  1 file changed, 6 insertions(+), 6 deletions(-)
> > >
> > > diff --git a/drivers/pci/controller/dwc/pcie-designware-host.c b/drivers/pci/controller/dwc/pcie-designware-host.c
> > > index c57831902686e..52a441662cabe 100644
> > > --- a/drivers/pci/controller/dwc/pcie-designware-host.c
> > > +++ b/drivers/pci/controller/dwc/pcie-designware-host.c
> > > @@ -452,6 +452,12 @@ int dw_pcie_host_init(struct dw_pcie_rp *pp)
> > >
> > >  	raw_spin_lock_init(&pp->lock);
> > >
> > > +	bridge = devm_pci_alloc_host_bridge(dev, 0);
> > > +	if (!bridge)
> > > +		return bridge;
> >
> > This returns NULL (0) where it previously returned -ENOMEM.  Callers
> > interpret zero as "success", so I think it should stil return -ENOMEM.
> 
> It should be -ENOMEM. Sorry for that. Strange, not sure what happen when
> I copy/past code.
> 
> Do you need respin it or you can fix it?

I fixed it locally.  But you should fix it, too, in case we do another
spin for other reasons.

> > I tentatively changed it back to -ENOMEM locally, let me know if
> > that's wrong.
> >
> > > +	pp->bridge = bridge;
> > > +