Re: [PATCH v2] PCI: s32g: Fix ports parsing

[Frank Li <Frank.li@nxp.com>]

On Tue, Feb 03, 2026 at 07:56:23AM +0100, Vincent Guittot wrote:
> On Mon, 2 Feb 2026 at 20:46, Frank Li <Frank.li@nxp.com> wrote:
> >
> > On Mon, Feb 02, 2026 at 08:23:30PM +0100, Vincent Guittot wrote:
> > > On Mon, 2 Feb 2026 at 18:42, Frank Li <Frank.li@nxp.com> wrote:
> > > >
> > > > On Mon, Feb 02, 2026 at 04:10:50PM +0100, Vincent Guittot wrote:
> > > > > No error return is missing after the loop resulting in removing the
> > > > > ports from the list.
> > > > >
> > > > > Fixes: 5cbc7d3e316e ("PCI: s32g: Add NXP S32G PCIe controller driver (RC)")
> > > > > Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
> > > > > ---
> > > > >
> > > > > Change since v1:
> > > > > - handle correctly the case when there is no child port
> > > > >
> > > > >  drivers/pci/controller/dwc/pcie-nxp-s32g.c | 8 ++++----
> > > > >  1 file changed, 4 insertions(+), 4 deletions(-)
> > > > >
> > > > > diff --git a/drivers/pci/controller/dwc/pcie-nxp-s32g.c b/drivers/pci/controller/dwc/pcie-nxp-s32g.c
> > > > > index 47745749f75c..b3ec38099fa3 100644
> > > > > --- a/drivers/pci/controller/dwc/pcie-nxp-s32g.c
> > > > > +++ b/drivers/pci/controller/dwc/pcie-nxp-s32g.c
> > > > > @@ -282,12 +282,12 @@ static int s32g_pcie_parse_ports(struct device *dev, struct s32g_pcie *s32g_pp)
> > > > >
> > > > >               ret = s32g_pcie_parse_port(s32g_pp, of_port);
> > > > >               if (ret)
> > > > > -                     goto err_port;
> > > > > +                     break;
> > > > >       }
> > > > >
> > > > > -err_port:
> > > > > -     list_for_each_entry_safe(port, tmp, &s32g_pp->ports, list)
> > > > > -             list_del(&port->list);
> > > >
> > > > break and goto is logical equal here. So only need below if (ret).
> > >
> > > Yes it's similar that why I removed it, the goto becomes useless and
> > > make code less readable
> >
> > But don't suggest change in this fixes patch, which should focus on the
> > key part.
>
> It's not a suggested change, it's the fix. err_port is now wrong
> because it's the path for all cases.

But "goto" is the same as orginal "break". Really fix is below "if (ret)".

Frank

>
> >
> > Frank
> >
> > >
> > > >
> > > > Frank
> > > >
> > > > > +     if (ret)
> > > > > +             list_for_each_entry_safe(port, tmp, &s32g_pp->ports, list)
> > > > > +                     list_del(&port->list);
> > > > >
> > > > >       return ret;
> > > > >  }
> > > > > --
> > > > > 2.43.0
> > > > >