Hi,
On Thu, Mar 22, 2018 at 11:11 PM, Manu Gautam [off-list ref] wrote:
quoted hunk ↗ jump to hunk
@@ -1414,6 +1556,12 @@ int qcom_qmp_phy_create(struct device *dev, struct device_node *np, int id)
}, {
.compatible = "qcom,qmp-v3-usb3-phy",
.data = &qmp_v3_usb3phy_cfg,
+ }, {
+ .compatible = "qcom,sdm845-qmp-usb3-phy",
+ .data = &qmp_v3_usb3phy_cfg,
+ }, {
+ .compatible = "qcom,sdm845-qmp-usb3-uni-phy",
+ .data = &qmp_v3_usb3_uniphy_cfg,
As per my comments on the bindings patch, having two compatible
strings that both map to "qmp_v3_usb3phy_cfg" smells a little wrong.
* If the agreement in the bindings patch is that we somehow need to
keep "qcom,qmp-v3-usb3-phy" around then the sdm845 device tree file
should list:
compatible = "qcom,sdm845-qmp-usb3-phy", "qcom,qmp-v3-usb3-phy";
...and then you can get rid of the "qcom,sdm845-qmp-usb3-phy" in this
table (it will use the secondary compatible string to pick the right
entry).
* If the agreement is to get rid of "qcom,qmp-v3-usb3-phy" then it
should go away from the table.
-Doug