On Fri, Aug 14, 2015 at 04:08:10PM -0500, Rob Herring wrote:
> On Fri, Aug 14, 2015 at 11:19 AM, Marc Zyngier <marc.zyngier@xxxxxxx> wrote:
> > Both pci-host-generic and Pseries parse the "linux,pci-probe-only"
> > to engage the PCI_PROBE_ONLY mode, and both have a subtle bug that
> > can be triggered if the property has no parameter.
>
> Humm, I bet we could break a lot of machines if we fixed the core code
> to properly make pp->value NULL when there is no value.
>
> > Provide a generic implementation that can be used by both.
> >
> > Signed-off-by: Marc Zyngier <marc.zyngier@xxxxxxx>
> > ---
> > drivers/of/of_pci.c | 30 ++++++++++++++++++++++++++++++
> > include/linux/of_pci.h | 3 +++
> > 2 files changed, 33 insertions(+)
> >
> > diff --git a/drivers/of/of_pci.c b/drivers/of/of_pci.c
> > index 5751dc5..a4e29ff 100644
> > --- a/drivers/of/of_pci.c
> > +++ b/drivers/of/of_pci.c
> > @@ -118,6 +118,36 @@ int of_get_pci_domain_nr(struct device_node *node)
> > EXPORT_SYMBOL_GPL(of_get_pci_domain_nr);
> >
> > /**
> > + * of_pci_check_probe_only - Setup probe only mode if linux,pci-probe-only
> > + * is present and valid
> > + *
> > + * @node: device tree node that may contain the property (usually "chosen")
> > + *
> > + */
> > +void of_pci_check_probe_only(struct device_node *node)
> > +{
> > + const int *prop;
> > + int len;
> > +
> > + if (!node)
> > + return;
> > +
> > + prop = of_get_property(node, "linux,pci-probe-only", &len);
>
> It is preferred to use of_property_read_u32 to avoid just these types
> of problems.
I don't know enough OF to really understand this, but I infer that
this is a suggestion for improving the patch. Should I be waiting for
a v3 series?
Bjorn
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
