[PATCH v4 2/6] dt-bindings: PCI: Add bindings for Brcmstb endpoint device voltage regulators

Rob Herring robh at kernel.org
Thu Apr 8 18:55:38 BST 2021


On Thu, Apr 08, 2021 at 12:58:05PM -0400, Jim Quinlan wrote:
> On Thu, Apr 8, 2021 at 12:20 PM Rob Herring <robh at kernel.org> wrote:
> >
> > On Tue, Apr 06, 2021 at 02:25:49PM -0400, Jim Quinlan wrote:
> > > On Tue, Apr 6, 2021 at 1:32 PM Mark Brown <broonie at kernel.org> wrote:
> > > >
> > > > On Tue, Apr 06, 2021 at 01:26:51PM -0400, Jim Quinlan wrote:
> > > > > On Tue, Apr 6, 2021 at 12:47 PM Mark Brown <broonie at kernel.org> wrote:
> > > >
> > > > > > No great problem with having these in the controller node (assming it
> > > > > > accurately describes the hardware) but I do think we ought to also be
> > > > > > able to describe these per slot.
> >
> > PCIe is effectively point to point, so there's only 1 slot unless
> > there's a PCIe switch in the middle. If that's the case, then it's all
> > more complicated.
> >
> > > > > Can you explain what you think that would look like in the DT?
> > > >
> > > > I *think* that's just some properties on the nodes for the endpoints,
> > > > note that the driver could just ignore them for now.  Not sure where or
> > > > if we document any extensions but child nodes are in section 4 of the
> > > > v2.1 PCI bus binding.
> > >
> > > Hi Mark,
> > >
> > > I'm a little confused -- here is how I remember the chronology of the
> > > "DT bindings" commit reviews, please correct me if I'm wrong:
> > >
> > > o JimQ submitted a pullreq for using voltage regulators in the same
> > > style as the existing "rockport" PCIe driver.
> > > o After some deliberation, RobH preferred that the voltage regulators
> > > should go into the PCIe subnode device's DT node.
> >
> > IIRC, that's because you said there isn't a standard slot.
> Admittedly, I'm not exactly sure what you mean by a "standard slot".
> Our PCIIe HW does not support  hotplug or have a presence bit or
> anything like that.  Our root complex has one port; it can only
> directly connect to a single switch or endpoint. This connection shows
> up as slot0.  The voltage regulator(s) involved depend on a GPIO that
> turns the power  on/off for the connected device/chip.  The gpio pin
> can vary from board to board but this is easily handled in our DT.
> Some boards have regulators that are always on and not associated with
> a GPIO pin -- these have no representation in our DT.

By standard slot, I mean you have standard voltage rails 12V and 3.3V 
(or 1.5 and 3.3 for mini PCIe) and PERST# signal, no other extra 
things to make a device discoverable, and the timing for 
those rails and PERST# follow what the spec defines.

There's also CLKREQ, WAKE, and hotplug detect signals, but I think those 
are all optional and could be tied off. I think most PCI h/w is not 
hotplug capable.

Rob



More information about the linux-arm-kernel mailing list