On 29-07-20, 21:00, Sivaprakash Murugesan wrote:
> Commit cc1e06f033af ("phy: qcom: qmp: Use power_on/off ops for PCIe")
> changed phy ops from init/deinit to power on/off, due to this phy enable
> is getting called after PCIe init.
>
> On some platforms like ipq8074 phy should be inited before accessing the
> PCIe register space, otherwise the system would hang.
Have you verified that this causes no regression on sdm845 and other
platforms?
> So move phy_power_on API before PCIe init.
>
> Fixes: commit cc1e06f033af ("phy: qcom: qmp: Use power_on/off ops for PCIe")
Is this a fix..? You are updating sequencing for a new platform
> Co-developed-by: Selvam Sathappan Periakaruppan <speri...@codeaurora.org>
> Signed-off-by: Selvam Sathappan Periakaruppan <speri...@codeaurora.org>
> Signed-off-by: Sivaprakash Murugesan <sivap...@codeaurora.org>
> ---
> drivers/pci/controller/dwc/pcie-qcom.c | 12 ++++++------
> 1 file changed, 6 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/pci/controller/dwc/pcie-qcom.c
> b/drivers/pci/controller/dwc/pcie-qcom.c
> index 3aac77a..e1b5651 100644
> --- a/drivers/pci/controller/dwc/pcie-qcom.c
> +++ b/drivers/pci/controller/dwc/pcie-qcom.c
> @@ -1265,18 +1265,18 @@ static int qcom_pcie_host_init(struct pcie_port *pp)
>
> qcom_ep_reset_assert(pcie);
>
> - ret = pcie->ops->init(pcie);
> + ret = phy_power_on(pcie->phy);
> if (ret)
> return ret;
>
> - ret = phy_power_on(pcie->phy);
> + ret = pcie->ops->init(pcie);
> if (ret)
> - goto err_deinit;
> + goto err_disable_phy;
>
> if (pcie->ops->post_init) {
> ret = pcie->ops->post_init(pcie);
> if (ret)
> - goto err_disable_phy;
> + goto err_deinit;
> }
>
> dw_pcie_setup_rc(pp);
> @@ -1295,10 +1295,10 @@ static int qcom_pcie_host_init(struct pcie_port *pp)
> qcom_ep_reset_assert(pcie);
> if (pcie->ops->post_deinit)
> pcie->ops->post_deinit(pcie);
> -err_disable_phy:
> - phy_power_off(pcie->phy);
> err_deinit:
> pcie->ops->deinit(pcie);
> +err_disable_phy:
> + phy_power_off(pcie->phy);
>
> return ret;
> }
> --
> 2.7.4
--
~Vinod
