On Wed, Oct 19, 2016 at 4:24 PM, Vivek Gautam
<vivek.gau...@codeaurora.org> wrote:
> CC: Srinivas Kandagatla
>
>
> On 10/19/2016 04:13 PM, Vivek Gautam wrote:
>>
>> Qualcomm SOCs have QMP phy controller that provides support
>> to a number of controller, viz. PCIe, UFS, and USB.
>> Add a new driver, based on generic phy framework, for this
>> phy controller.
>>
>> USB3-phy changes: Based on phy-msm-ssusb-qmp driver available on
>> msm-4.4 kernel @codeaurora[1].
>> PCIe-phy changes: Based on msm8996-pcie-phy driver posted by
>> Srinivas [2].
>>
>> [1]
>> https://source.codeaurora.org/quic/la/kernel/msm-4.4/log/?h=caf/3.18/msm-3.18
>> [2] https://patchwork.kernel.org/patch/9318947/
>>
>> Signed-off-by: Vivek Gautam <vivek.gau...@codeaurora.org>
>> Cc: Kishon Vijay Abraham I <kis...@ti.com>
>> ---
[snip]
>> +static int qcom_qmp_phy_com_init(struct qcom_qmp_phy *qphy)
>> +{
>> + const struct qmp_phy_init_cfg *cfg = qphy->cfg;
>> + void __iomem *serdes = qphy->serdes;
>> + int ret;
>> +
>> + mutex_lock(&qphy->phy_mutex);
>> + if (qphy->init_count++) {
>> + mutex_unlock(&qphy->phy_mutex);
>> + return 0;
>> + }
>> +
>> + ret = reset_control_deassert(qphy->phy_rst);
>> + if (ret) {
>> + dev_err(qphy->dev, "phy reset deassert failed\n");
>> + return ret;
Pointed out by Kbuild-test: drivers/phy/phy-qcom-qmp.c:677:2-8:
preceding lock on line 668
The mutex has to be unlocked for all error cases.
Will fix this in the next patch version.
[snip]
Thanks
Vivek
--
Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum,
a Linux Foundation Collaborative Project
