From: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org>
To: Krishna chaitanya chundru <quic_krichai@quicinc.com>
Cc: helgaas@kernel.org, linux-pci@vger.kernel.org,
linux-arm-msm@vger.kernel.org, linux-kernel@vger.kernel.org,
quic_vbadigan@quicinc.com, quic_nitegupt@quicinc.com,
quic_skananth@quicinc.com, quic_ramkri@quicinc.com,
quic_parass@quicinc.com, krzysztof.kozlowski@linaro.org,
"Andy Gross" <agross@kernel.org>,
"Bjorn Andersson" <andersson@kernel.org>,
"Konrad Dybcio" <konrad.dybcio@linaro.org>,
"Manivannan Sadhasivam" <mani@kernel.org>,
"Lorenzo Pieralisi" <lpieralisi@kernel.org>,
"Krzysztof Wilczyński" <kw@linux.com>,
"Rob Herring" <robh@kernel.org>,
"Bjorn Helgaas" <bhelgaas@google.com>
Subject: Re: [PATCH v1] PCI: qcom: Add sysfs entry to change link speed dynamically
Date: Wed, 23 Aug 2023 13:43:39 +0530 [thread overview]
Message-ID: <20230823081339.GK3737@thinkpad> (raw)
In-Reply-To: <1692239684-12697-1-git-send-email-quic_krichai@quicinc.com>
On Thu, Aug 17, 2023 at 08:04:43AM +0530, Krishna chaitanya chundru wrote:
> PCIe can operate on lower GEN speed if client decided based upon
> the bandwidth & latency requirements. To support dynamic GEN speed
> switch adding this sysfs support.
>
> To change the GEN speed the link should be in L0, so first disable
> L0s & L1.
>
> L0s needs to be disabled at both RC & EP because L0s entry is
> independent. For enabling L0s both ends of the link needs to support
> it, so first check if L0s is supported on both ends and then enable
> L0s.
>
> This patch is dependent on "PCI: qcom: Add support for OPP"
> https://lore.kernel.org/linux-arm-msm/1692192264-18515-1-git-send-email-quic_krichai@quicinc.com/T/#t
>
There is also an ongoing proposal to add PCIe bandwidth controller driver [1]
which more or less does the same thing. Check if you can make use of it.
- Mani
[1] https://lore.kernel.org/linux-pci/20230817121708.53213-1-ilpo.jarvinen@linux.intel.com/
> Signed-off-by: Krishna chaitanya chundru <quic_krichai@quicinc.com>
> ---
> drivers/pci/controller/dwc/pcie-qcom.c | 141 +++++++++++++++++++++++++++++++++
> 1 file changed, 141 insertions(+)
>
> diff --git a/drivers/pci/controller/dwc/pcie-qcom.c b/drivers/pci/controller/dwc/pcie-qcom.c
> index 831d158..ad67d17 100644
> --- a/drivers/pci/controller/dwc/pcie-qcom.c
> +++ b/drivers/pci/controller/dwc/pcie-qcom.c
> @@ -241,10 +241,150 @@ struct qcom_pcie {
> const struct qcom_pcie_cfg *cfg;
> struct dentry *debugfs;
> bool suspended;
> + bool l0s_supported;
> };
>
> #define to_qcom_pcie(x) dev_get_drvdata((x)->dev)
>
> +static void qcom_pcie_icc_update(struct qcom_pcie *pcie);
> +static void qcom_pcie_opp_update(struct qcom_pcie *pcie);
> +
> +static int qcom_pcie_disable_l0s(struct pci_dev *pdev, void *userdata)
> +{
> + int lnkctl;
> +
> + pci_read_config_dword(pdev, pci_pcie_cap(pdev) + PCI_EXP_LNKCTL, &lnkctl);
> + lnkctl &= ~(PCI_EXP_LNKCTL_ASPM_L0S);
> + pci_write_config_word(pdev, pci_pcie_cap(pdev) + PCI_EXP_LNKCTL, lnkctl);
> +
> + return 0;
> +}
> +
> +static int qcom_pcie_check_l0s_support(struct pci_dev *pdev, void *userdata)
> +{
> + struct pci_dev *parent = pdev->bus->self;
> + struct qcom_pcie *pcie = userdata;
> + struct dw_pcie *pci = pcie->pci;
> + int lnkcap;
> +
> + /* check parent supports L0s */
> + if (parent) {
> + dev_err(pci->dev, "parent\n");
> + pci_read_config_dword(parent, pci_pcie_cap(parent) + PCI_EXP_LNKCAP,
> + &lnkcap);
> + if (!(lnkcap & PCI_EXP_LNKCAP_ASPM_L0S)) {
> + dev_info(pci->dev, "Parent does not support L0s\n");
> + pcie->l0s_supported = false;
> + return 0;
> + }
> + }
> +
> + pci_read_config_dword(pdev, pci_pcie_cap(pdev) + PCI_EXP_LNKCAP,
> + &lnkcap);
> + dev_err(pci->dev, "child %x\n", lnkcap);
> + if (!(lnkcap & PCI_EXP_LNKCAP_ASPM_L0S)) {
> + dev_info(pci->dev, "Device does not support L0s\n");
> + pcie->l0s_supported = false;
> + return 0;
> + }
> +
> + return 0;
> +}
> +
> +static int qcom_pcie_enable_l0s(struct pci_dev *pdev, void *userdata)
> +{
> + int lnkctl;
> +
> + pci_read_config_dword(pdev, pci_pcie_cap(pdev) + PCI_EXP_LNKCTL, &lnkctl);
> + lnkctl |= (PCI_EXP_LNKCTL_ASPM_L0S);
> + pci_write_config_word(pdev, pci_pcie_cap(pdev) + PCI_EXP_LNKCTL, lnkctl);
> +
> + return 0;
> +}
> +
> +static ssize_t qcom_pcie_speed_change_store(struct device *dev,
> + struct device_attribute *attr,
> + const char *buf,
> + size_t count)
> +{
> + unsigned int current_speed, target_speed, max_speed;
> + struct qcom_pcie *pcie = dev_get_drvdata(dev);
> + struct pci_bus *child, *root_bus = NULL;
> + struct dw_pcie_rp *pp = &pcie->pci->pp;
> + struct dw_pcie *pci = pcie->pci;
> + struct pci_dev *pdev;
> + u16 offset;
> + u32 val;
> + int ret;
> +
> + list_for_each_entry(child, &pp->bridge->bus->children, node) {
> + if (child->parent == pp->bridge->bus) {
> + root_bus = child;
> + break;
> + }
> + }
> +
> + pdev = root_bus->self;
> +
> + offset = dw_pcie_find_capability(pci, PCI_CAP_ID_EXP);
> +
> + val = readl(pci->dbi_base + offset + PCI_EXP_LNKCAP);
> + max_speed = FIELD_GET(PCI_EXP_LNKCAP_SLS, val);
> +
> + val = readw(pci->dbi_base + offset + PCI_EXP_LNKSTA);
> + current_speed = FIELD_GET(PCI_EXP_LNKSTA_CLS, val);
> +
> + ret = kstrtouint(buf, 10, &target_speed);
> + if (ret)
> + return ret;
> +
> + if (target_speed > max_speed)
> + return -EINVAL;
> +
> + if (current_speed == target_speed)
> + return count;
> +
> + pci_walk_bus(pp->bridge->bus, qcom_pcie_disable_l0s, pcie);
> +
> + /* Disable L1 */
> + val = dw_pcie_readl_dbi(pci, offset + PCI_EXP_LNKCTL);
> + val &= ~(PCI_EXP_LNKCTL_ASPM_L1);
> + dw_pcie_writel_dbi(pci, offset + PCI_EXP_LNKCTL, val);
> +
> + /* Set target GEN speed */
> + val = dw_pcie_readl_dbi(pci, offset + PCI_EXP_LNKCTL2);
> + val &= ~PCI_EXP_LNKCTL2_TLS;
> + dw_pcie_writel_dbi(pci, offset + PCI_EXP_LNKCTL2, val | target_speed);
> +
> + ret = pcie_retrain_link(pdev, true);
> + if (ret)
> + dev_err(dev, "Link retrain failed %d\n", ret);
> +
> + /* Enable L1 */
> + val = dw_pcie_readl_dbi(pci, offset + PCI_EXP_LNKCTL);
> + val |= (PCI_EXP_LNKCTL_ASPM_L1);
> + dw_pcie_writel_dbi(pci, offset + PCI_EXP_LNKCTL, val);
> +
> + pcie->l0s_supported = true;
> + pci_walk_bus(pp->bridge->bus, qcom_pcie_check_l0s_support, pcie);
> +
> + if (pcie->l0s_supported)
> + pci_walk_bus(pp->bridge->bus, qcom_pcie_enable_l0s, pcie);
> +
> + qcom_pcie_icc_update(pcie);
> +
> + qcom_pcie_opp_update(pcie);
> +
> + return count;
> +}
> +static DEVICE_ATTR_WO(qcom_pcie_speed_change);
> +
> +static struct attribute *qcom_pcie_attrs[] = {
> + &dev_attr_qcom_pcie_speed_change.attr,
> + NULL,
> +};
> +ATTRIBUTE_GROUPS(qcom_pcie);
> +
> static void qcom_ep_reset_assert(struct qcom_pcie *pcie)
> {
> gpiod_set_value_cansleep(pcie->reset, 1);
> @@ -1716,6 +1856,7 @@ static struct platform_driver qcom_pcie_driver = {
> .of_match_table = qcom_pcie_match,
> .pm = &qcom_pcie_pm_ops,
> .probe_type = PROBE_PREFER_ASYNCHRONOUS,
> + .dev_groups = qcom_pcie_groups,
> },
> };
> builtin_platform_driver(qcom_pcie_driver);
> --
> 2.7.4
>
--
மணிவண்ணன் சதாசிவம்
prev parent reply other threads:[~2023-08-23 10:14 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-17 2:34 [PATCH v1] PCI: qcom: Add sysfs entry to change link speed dynamically Krishna chaitanya chundru
2023-08-17 3:50 ` kernel test robot
2023-08-17 17:20 ` Bjorn Helgaas
2023-08-18 2:26 ` Krishna Chaitanya Chundru
2023-08-23 7:56 ` Manivannan Sadhasivam
2023-08-22 15:54 ` kernel test robot
2023-08-23 8:13 ` Manivannan Sadhasivam [this message]
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20230823081339.GK3737@thinkpad \
--to=manivannan.sadhasivam@linaro.org \
--cc=agross@kernel.org \
--cc=andersson@kernel.org \
--cc=bhelgaas@google.com \
--cc=helgaas@kernel.org \
--cc=konrad.dybcio@linaro.org \
--cc=krzysztof.kozlowski@linaro.org \
--cc=kw@linux.com \
--cc=linux-arm-msm@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pci@vger.kernel.org \
--cc=lpieralisi@kernel.org \
--cc=mani@kernel.org \
--cc=quic_krichai@quicinc.com \
--cc=quic_nitegupt@quicinc.com \
--cc=quic_parass@quicinc.com \
--cc=quic_ramkri@quicinc.com \
--cc=quic_skananth@quicinc.com \
--cc=quic_vbadigan@quicinc.com \
--cc=robh@kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.