Hi, + Kishon On Fri, Mar 04, 2022 at 09:23:02PM -0800, Lizhi Hou wrote: > Create device tree binding document for PCIe endpoint bus. > I'm currently working on a PCI endpoint function driver for MHI bus [1] and hence interested in this topic. Comments below. [1] https://lore.kernel.org/lkml/20220502060611.58987-9-manivannan.sadhasivam@xxxxxxxxxx/ > Signed-off-by: Sonal Santan <sonal.santan@xxxxxxxxxx> > Signed-off-by: Max Zhen <max.zhen@xxxxxxxxxx> > Signed-off-by: Lizhi Hou <lizhi.hou@xxxxxxxxxx> > --- > .../devicetree/bindings/bus/pci-ep-bus.yaml | 72 +++++++++++++++++++ > 1 file changed, 72 insertions(+) > create mode 100644 Documentation/devicetree/bindings/bus/pci-ep-bus.yaml > > diff --git a/Documentation/devicetree/bindings/bus/pci-ep-bus.yaml b/Documentation/devicetree/bindings/bus/pci-ep-bus.yaml > new file mode 100644 > index 000000000000..0ca96298db6f > --- /dev/null > +++ b/Documentation/devicetree/bindings/bus/pci-ep-bus.yaml > @@ -0,0 +1,72 @@ > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause) > +%YAML 1.2 > +--- > +$id: http://devicetree.org/schemas/bus/pci-ep-bus.yaml# > +$schema: http://devicetree.org/meta-schemas/core.yaml# > + > +title: PCIe Endpoint Bus binding > + > +description: | > + PCIe device may use flattened device tree to describe apertures in its > + PCIe BARs. The Bus PCIe endpoint node is created and attached under the > + device tree root node for this kind of device. Then the flatten device > + tree overlay for this device is attached under the endpoint node. > + > + The aperture address which is under the endpoint node consists of BAR > + index and offset. It uses the following encoding: > + On top of Rob's reply: Currently, the BAR memory for the PCI endpoint device is either allocated dynamically using pci_epf_alloc_space() or we need to pass the address in "phys_addr" field of "struct pci_epf_bar". In most of the PCI endpoint devices, we need to use a fixed memory region as the BAR. Since there is no devicetree integration for PCI endpoint subsystem, I've been using the 2nd approach of obtaining the BAR address from PCI endpoint controller devicetree node and passing it to "phys_addr" of "struct pci_epf_bar". Ideally, the BAR information should come from the devicetree. But we cannot use just "pci-ep-bus" node. I've been thinking about the below structure: pcie_ep: pcie-ep@40000000 { compatible = "pcie-ep"; .... pci_epf_0: pci-epf@100000 { reg = <0x100000 0x1000>; # BAR0 }; pci_epf_1: pci-epf@200000 { reg = <0x200000 0x1000>; # BAR1 }; }; Where, "pci-epf@" represents each of the PCI endpoint functions implemented by this device (note that there can be more than one function per endpoint device) and "reg" has the BAR address/size for that function. Rob, what do you think? Thanks, Mani > + 0xIooooooo 0xoooooooo > + > + Where: > + > + I = BAR index > + oooooo oooooooo = BAR offset > + > + The endpoint is compatible with 'simple-bus' and contains 'ranges' > + property for translating aperture address to CPU address. > + > +allOf: > + - $ref: /schemas/simple-bus.yaml# > + > +maintainers: > + - Lizhi Hou <lizhi.hou@xxxxxxxxxx> > + > +properties: > + compatible: > + contains: > + const: pci-ep-bus > + > + "#address-cells": > + const: 2 > + > + "#size-cells": > + const: 2 > + > + ranges: true > + > +patternProperties: > + "^.*@[0-9a-f]+$": > + description: hardware apertures belong to this device. > + type: object > + > +required: > + - compatible > + - "#address-cells" > + - "#size-cells" > + - ranges > + > +additionalProperties: false > + > +examples: > + - | > + bus { > + #address-cells = <2>; > + #size-cells = <2>; > + pci-ep-bus@e0000000 { > + compatible = "pci-ep-bus", "simple-bus"; > + #address-cells = <2>; > + #size-cells = <2>; > + ranges = <0x0 0x0 0x0 0xe0000000 0x0 0x2000000 > + 0x20000000 0x0 0x0 0xe4200000 0x0 0x40000>; > + }; > + }; > -- > 2.27.0 > -- மணிவண்ணன் சதாசிவம்
