On 06/09/2022 10:29, Abel Vesa wrote: > On 22-09-06 08:45:22, Krzysztof Kozlowski wrote: >> On 05/09/2022 16:45, Abel Vesa wrote: >>> Convert Qualcomm FastRPC bindings to yaml format, so that we could validate >>> dt-entries correctly and any future additions can go into yaml format. >>> >>> Use compute-cb@ subnodes instead of just cb@. >>> >>> Also add qcom,non-secure-domain, qcom,glink-channels and >>> qcom,smd-channels missing properties to make sure dtbs_check doesn't >>> fail right off the bat. >> >> qcom,non-secure-domain is in original binding, so I don't understand why >> it is being "added". >> > > Yeah, my bad, I should've added this line to the changes since v4. > >>> >>> Co-developed-by: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx> >>> Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx> >>> Co-developed-by: David Heidelberg <david@xxxxxxx> >>> Signed-off-by: David Heidelberg <david@xxxxxxx> >>> Signed-off-by: Abel Vesa <abel.vesa@xxxxxxxxxx> >>> --- >>> >>> Changes since v6: >>> * renamed the parent node name in the example from smd-edge to glink-edge >>> >>> .../devicetree/bindings/misc/qcom,fastrpc.txt | 88 ------------- >>> .../bindings/misc/qcom,fastrpc.yaml | 118 ++++++++++++++++++ >> >> As you can see in Rob's bot report - the patchset introduces errors and >> is not bisectable. > > Please note that Rob's bot report is for v6. I see report as a reply to this patch, so for v7. Why do you think it is v6? > > v7 fixes the errors reported, by using glink-edge instead of smd-edge. > > Looking at all QCOM SoCs that have fastrpc node in devicetree, they all > seem to be using glink-edge. I was not talking about these errors (they were separate issue). I am talking about wrong path error. > >> >> You also need to fix qcom,glink-edge.yaml >> > > I don't see why, with the changes I made in v7, there are no errors > anymore. There are, but not from tooling. The error is wrong path. It should be converted to proper schema $ref. Best regards, Krzysztof
