While working on upstream support for the Toshiba TC9564 SoC I
learned that its PCIe endpoint function nodes included the
device_type property, indicating each is a "pci" device.  Herve
Codina pointed out that this is not correct; only a PCI device
"that implements a PCI bus" should define that property.

I looked at other upstream devicetree files and I found four other
places with this same mistake.  (All are for Qualcomm platforms,
and all are for platforms using a TC956x SoC.)

This series removes the device_type property where it is defined
for a PCI endpoint function.

It also removes the #address-cells, #size-cells, and ranges
properties in these cases, with a plan to restore them only when
they are known to be needed (to implement pci-ep-bus sub-nodes).

					-Alex

Alex Elder (5):
  arm64: dts: qcom: qcs6490-rb3gen2: clean up PCI function nodes
  arm64: dts: qcom: qcs6490-rb3gen2-industrial-mezzanine: clean up PCI
    function nodes
  arm64: dts: qcom: lemans-evk-ifp-mezzanine: clean up PCI function
    nodes
  arm64: dts: qcom: monaco-evk: clean up PCI function nodes
  arm64: dts: qcom: qcs6490-thundercomm-minipc-g1iot: clean up PCI
    function nodes

 .../boot/dts/qcom/lemans-evk-ifp-mezzanine.dtso  |  8 --------
 .../boot/dts/qcom/monaco-evk-ifp-mezzanine.dtso  |  8 --------
 .../qcs6490-rb3gen2-industrial-mezzanine.dtso    | 16 ----------------
 arch/arm64/boot/dts/qcom/qcs6490-rb3gen2.dts     |  8 --------
 .../qcom/qcs6490-thundercomm-minipc-g1iot.dts    |  8 --------
 5 files changed, 48 deletions(-)


base-commit: 064935ef6be39d086f4c40766d2b76181b70b3e1
-- 
2.53.0

On 8/7/26 9:58 PM, Alex Elder wrote:
> While working on upstream support for the Toshiba TC9564 SoC I
> learned that its PCIe endpoint function nodes included the
> device_type property, indicating each is a "pci" device.  Herve
> Codina pointed out that this is not correct; only a PCI device
> "that implements a PCI bus" should define that property.
> 
> I looked at other upstream devicetree files and I found four other
> places with this same mistake.  (All are for Qualcomm platforms,
> and all are for platforms using a TC956x SoC.)
> 
> This series removes the device_type property where it is defined
> for a PCI endpoint function.
> 
> It also removes the #address-cells, #size-cells, and ranges
> properties in these cases, with a plan to restore them only when
> they are known to be needed (to implement pci-ep-bus sub-nodes).
> 
> 					-Alex

For the series:

Reviewed-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Konrad

On Fri, Aug 07, 2026 at 02:58:39PM -0500, Alex Elder wrote:
> While working on upstream support for the Toshiba TC9564 SoC I
> learned that its PCIe endpoint function nodes included the
> device_type property, indicating each is a "pci" device.  Herve
> Codina pointed out that this is not correct; only a PCI device
> "that implements a PCI bus" should define that property.
> 
> I looked at other upstream devicetree files and I found four other
> places with this same mistake.  (All are for Qualcomm platforms,
> and all are for platforms using a TC956x SoC.)
> 
> This series removes the device_type property where it is defined
> for a PCI endpoint function.
> 
> It also removes the #address-cells, #size-cells, and ranges
> properties in these cases, with a plan to restore them only when
> they are known to be needed (to implement pci-ep-bus sub-nodes).

Thanks for cleaning the mess!

Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@oss.qualcomm.com>

- Mani

-- 
மணிவண்ணன் சதாசிவம்