| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20220201183912.GA583363@bhelgaas>
Date: Tue, 1 Feb 2022 12:39:12 -0600
From: Bjorn Helgaas <helgaas@...nel.org>
To: Prasad Malisetty <quic_pmaliset@...cinc.com>
Cc: agross@...nel.org, bjorn.andersson@...aro.org,
lorenzo.pieralisi@....com, robh@...nel.org, kw@...ux.com,
bhelgaas@...gle.com, linux-pci@...r.kernel.org,
linux-arm-msm@...r.kernel.org, linux-kernel@...r.kernel.org,
quic_vbadigan@...cinc.com, quic_ramkri@...cinc.com,
manivannan.sadhasivam@...aro.org, swboyd@...omium.org
Subject: Re: [PATCH v2] PCI: qcom: Add system PM support
On Tue, Feb 01, 2022 at 11:37:56PM +0530, Prasad Malisetty wrote:
> Add suspend_noirq and resume_noirq callbacks to handle
> System suspend and resume in dwc pcie controller driver.
s/System/system/
s/pcie/PCIe/ as you did below.
> When system suspends, send PME turnoff message to enter
> link into L2 state. Along with powerdown the PHY, disable
> pipe clock, switch gcc_pcie_1_pipe_clk_src to XO if mux is
> supported and disable the pcie clocks, regulators.
>
> When system resumes, PCIe link will be re-established and
> setup rc settings.
>
> Signed-off-by: Prasad Malisetty <quic_pmaliset@...cinc.com>
> Reported-by: kernel test robot <lkp@...el.com>
The kernel test robot reported the lack of system power management?
I'm impressed ;) I doubt this "Reported-by" is useful. If it *is*,
please include a link to the report.
> ---
> Changes since v1:
> - Removed unnecessary logs and modified log level suggested by Manivannan.
> - Removed platform specific callbacks as PM support is generic.
> ---
> drivers/pci/controller/dwc/pcie-qcom.c | 97 ++++++++++++++++++++++++++++++++++
> 1 file changed, 97 insertions(+)
>
> diff --git a/drivers/pci/controller/dwc/pcie-qcom.c b/drivers/pci/controller/dwc/pcie-qcom.c
> index c19cd506..d1dd6c7 100644
> --- a/drivers/pci/controller/dwc/pcie-qcom.c
> +++ b/drivers/pci/controller/dwc/pcie-qcom.c
> @@ -73,6 +73,8 @@
>
> #define PCIE20_PARF_Q2A_FLUSH 0x1AC
>
> +#define PCIE20_PARF_PM_STTS 0x24
Indent with tabs (not spaces) as the surrounding code does.
> #define PCIE20_MISC_CONTROL_1_REG 0x8BC
> #define DBI_RO_WR_EN 1
>
> @@ -1616,6 +1618,100 @@ static int qcom_pcie_probe(struct platform_device *pdev)
> return ret;
> }
>
> +static int qcom_pcie_send_pme_turnoff_msg(struct qcom_pcie *pcie)
> +{
> + int ret = 0;
Drop unnecessary init.
> + u32 val = 0, poll_val = 0;
Drop unnecessary "val" init.
> + u64 l23_rdy_poll_timeout = 100000;
Add "microseconds" comment.
> + struct dw_pcie *pci = pcie->pci;
> + struct device *dev = pci->dev;
> +
> + val = readl(pcie->elbi + PCIE20_ELBI_SYS_CTRL);
> + val |= BIT(4);
> + writel(val, pcie->elbi + PCIE20_ELBI_SYS_CTRL);
> +
> + ret = readl_poll_timeout((pcie->parf + PCIE20_PARF_PM_STTS), poll_val,
> + (poll_val & BIT(5)), 10000, l23_rdy_poll_timeout);
> + if (!ret)
> + dev_info(dev, "PM_Enter_L23 is received\n");
> + else
> + dev_err(dev, "PM_Enter_L23 is NOT received.PARF_PM_STTS 0x%x\n",
> + readl_relaxed(pcie->parf + PCIE20_PARF_PM_STTS));
> +
> + return ret;
> +}
> +
> +static void qcom_pcie_host_disable(struct qcom_pcie *pcie)
> +{
> + struct qcom_pcie_resources_2_7_0 *res = &pcie->res.v2_7_0;
qcom_pcie_host_disable() is called from qcom_pcie_pm_suspend_noirq(),
which is used for all qcom devices. But it looks like not all qcom
devices have 2_7_0 resources. Is this supposed to be used only on
certain revisions?
> +
> + /* Assert the reset of endpoint */
Superfluous comment, since the function name says the same thing.
> + qcom_ep_reset_assert(pcie);
> +
> + /* Put PHY into POWER DOWN state */
> + phy_power_off(pcie->phy);
> +
> + writel(1, pcie->parf + PCIE20_PARF_PHY_CTRL);
> +
> + pcie->ops->post_deinit(pcie);
> +
> + /* Disable PCIe clocks and regulators */
> + pcie->ops->deinit(pcie);
> +}
> +
> +static int __maybe_unused qcom_pcie_pm_suspend_noirq(struct device *dev)
> +{
> + int ret = 0;
> + struct qcom_pcie *pcie = dev_get_drvdata(dev);
> + struct dw_pcie *pci = pcie->pci;
> +
> + if (!dw_pcie_link_up(pci)) {
> + dev_dbg(dev, "Power has been turned off already\n");
> + return ret;
"return 0" here and drop the init above.
> + }
> +
> + /* Send PME turnoff msg */
Superfluous comment.
> + ret = qcom_pcie_send_pme_turnoff_msg(pcie);
> + if (ret)
> + return ret;
> +
> + /* Power down the PHY, disable clock and regulators */
> + qcom_pcie_host_disable(pcie);
> +
> + return ret;
"return 0" here.
> +}
> +
> +/* Resume the PCIe link */
> +static int __maybe_unused qcom_pcie_pm_resume_noirq(struct device *dev)
> +{
> + int ret = 0;
Drop unnecessary init.
> + struct qcom_pcie *pcie = dev_get_drvdata(dev);
> + struct dw_pcie *pci = pcie->pci;
> + struct pcie_port *pp = &pci->pp;
> +
> + /* Initialize PCIe host */
Superfluous comment.
> + ret = qcom_pcie_host_init(pp);
> + if (ret) {
> + dev_err(dev, "cannot initialize host\n");
> + return ret;
> + }
> +
> + dw_pcie_setup_rc(pp);
> +
> + /* Start the PCIe link */
Superfluous comment.
> + qcom_pcie_start_link(pci);
> +
> + ret = dw_pcie_wait_for_link(pci);
> + if (ret)
> + dev_err(dev, "Link never came up, Resume failed\n");
> +
> + return ret;
> +}
> +
> +static const struct dev_pm_ops qcom_pcie_pm_ops = {
> + SET_NOIRQ_SYSTEM_SLEEP_PM_OPS(qcom_pcie_pm_suspend_noirq, qcom_pcie_pm_resume_noirq)
qcom_pcie_pm_suspend_noirq() and qcom_pcie_pm_resume_noirq() look
nothing like their counterparts in dra7xx, exynos, imx6, intel-gw.
Any chance you could make them more similar?
> +};
> +
> static const struct of_device_id qcom_pcie_match[] = {
> { .compatible = "qcom,pcie-apq8084", .data = &apq8084_cfg },
> { .compatible = "qcom,pcie-ipq8064", .data = &ipq8064_cfg },
> @@ -1648,6 +1744,7 @@ static struct platform_driver qcom_pcie_driver = {
> .probe = qcom_pcie_probe,
> .driver = {
> .name = "qcom-pcie",
> + .pm = &qcom_pcie_pm_ops,
> .suppress_bind_attrs = true,
> .of_match_table = qcom_pcie_match,
> },
> --
> QUALCOMM INDIA, on behalf of Qualcomm Innovation Center, Inc. is a member
> of Code Aurora Forum, hosted by The Linux Foundation
>
Powered by blists - more mailing lists