| 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
| ||
|
Date: Mon, 8 Apr 2024 14:23:06 +0530
From: Krishna Chaitanya Chundru <quic_krichai@...cinc.com>
To: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
CC: Bjorn Andersson <andersson@...nel.org>,
Konrad Dybcio
<konrad.dybcio@...aro.org>,
Rob Herring <robh@...nel.org>,
"Krzysztof
Kozlowski" <krzysztof.kozlowski+dt@...aro.org>,
Conor Dooley
<conor+dt@...nel.org>,
Lorenzo Pieralisi <lpieralisi@...nel.org>,
Krzysztof Wilczyński <kw@...ux.com>,
Bjorn Helgaas
<bhelgaas@...gle.com>, <johan+linaro@...nel.org>,
<bmasney@...hat.com>, <djakov@...nel.org>,
<linux-arm-msm@...r.kernel.org>, <devicetree@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, <linux-pci@...r.kernel.org>,
<vireshk@...nel.org>, <quic_vbadigan@...cinc.com>,
<quic_skananth@...cinc.com>, <quic_nitegupt@...cinc.com>,
<quic_parass@...cinc.com>, <krzysztof.kozlowski@...aro.org>,
Bryan O'Donoghue
<bryan.odonoghue@...aro.org>
Subject: Re: [PATCH v9 2/6] PCI: qcom: Add ICC bandwidth vote for CPU to PCIe
path
On 4/7/2024 8:09 PM, Manivannan Sadhasivam wrote:
> On Sun, Apr 07, 2024 at 10:07:35AM +0530, Krishna chaitanya chundru wrote:
>> To access PCIe registers, PCIe BAR space, config space the CPU-PCIe
>
> Please specify whether you are referencing PCIe host controller or endpoint
> device or both.
>
>> ICC (interconnect consumers) path should be voted otherwise it may
>
> ICC is just 'Interconnect' unless I misunderstood.
>
>> lead to NoC (Network on chip) timeout. We are surviving because of
>> other driver vote for this path.
>>
>
> s/vote/voting
>
>> As there is less access on this path compared to PCIe to mem path
>> add minimum vote i.e 1KBps bandwidth always which is recommended
>> by HW team.
>>
>
> 'which is sufficient enough to keep the path active.'
>
>> When suspending, disable this path after register space access
>> is done.
>>
>> Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@...aro.org>
>> Signed-off-by: Krishna chaitanya chundru <quic_krichai@...cinc.com>
>> ---
>> drivers/pci/controller/dwc/pcie-qcom.c | 38 ++++++++++++++++++++++++++++++----
>> 1 file changed, 34 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/pci/controller/dwc/pcie-qcom.c b/drivers/pci/controller/dwc/pcie-qcom.c
>> index 14772edcf0d3..b4893214b2d3 100644
>> --- a/drivers/pci/controller/dwc/pcie-qcom.c
>> +++ b/drivers/pci/controller/dwc/pcie-qcom.c
>> @@ -245,6 +245,7 @@ struct qcom_pcie {
>> struct phy *phy;
>> struct gpio_desc *reset;
>> struct icc_path *icc_mem;
>> + struct icc_path *icc_cpu;
>> const struct qcom_pcie_cfg *cfg;
>> struct dentry *debugfs;
>> bool suspended;
>> @@ -1409,6 +1410,9 @@ static int qcom_pcie_icc_init(struct qcom_pcie *pcie)
>> if (IS_ERR(pcie->icc_mem))
>> return PTR_ERR(pcie->icc_mem);
>>
>> + pcie->icc_cpu = devm_of_icc_get(pci->dev, "cpu-pcie");
>> + if (IS_ERR(pcie->icc_cpu))
>> + return PTR_ERR(pcie->icc_cpu);
>> /*
>> * Some Qualcomm platforms require interconnect bandwidth constraints
>> * to be set before enabling interconnect clocks.
>> @@ -1418,7 +1422,19 @@ static int qcom_pcie_icc_init(struct qcom_pcie *pcie)
>> */
>> ret = icc_set_bw(pcie->icc_mem, 0, QCOM_PCIE_LINK_SPEED_TO_BW(1));
>> if (ret) {
>> - dev_err(pci->dev, "failed to set interconnect bandwidth: %d\n",
>> + dev_err(pci->dev, "failed to set interconnect bandwidth for PCIe-MEM: %d\n",
>> + ret);
>> + return ret;
>> + }
>> +
>> + /*
>> + * Since the CPU-PCIe path is only used for activities like register
>
> Again, differentiate PCIe controller and endpoint device access.
>
Ack to all comments. I will modify in next patch.
>> + * access, Config/BAR space access, HW team has recommended to use a
>> + * minimal bandwidth of 1KBps just to keep the link active.
>> + */
>> + ret = icc_set_bw(pcie->icc_cpu, 0, kBps_to_icc(1));
>> + if (ret) {
>> + dev_err(pci->dev, "failed to set interconnect bandwidth for CPU-PCIe: %d\n",
>> ret);
>> return ret;
>> }
>> @@ -1448,7 +1464,7 @@ static void qcom_pcie_icc_update(struct qcom_pcie *pcie)
>>
>> ret = icc_set_bw(pcie->icc_mem, 0, width * QCOM_PCIE_LINK_SPEED_TO_BW(speed));
>> if (ret) {
>> - dev_err(pci->dev, "failed to set interconnect bandwidth: %d\n",
>> + dev_err(pci->dev, "failed to set interconnect bandwidth for PCIe-MEM: %d\n",
>> ret);
>> }
>> }
>> @@ -1610,7 +1626,7 @@ static int qcom_pcie_suspend_noirq(struct device *dev)
>> */
>> ret = icc_set_bw(pcie->icc_mem, 0, kBps_to_icc(1));
>> if (ret) {
>> - dev_err(dev, "Failed to set interconnect bandwidth: %d\n", ret);
>> + dev_err(dev, "Failed to set interconnect bandwidth for PCIe-MEM: %d\n", ret);
>> return ret;
>> }
>>
>> @@ -1634,7 +1650,15 @@ static int qcom_pcie_suspend_noirq(struct device *dev)
>> pcie->suspended = true;
>> }
>>
>> - return 0;
>> + /*
>> + * Remove the vote for CPU-PCIe path now, since at this point onwards,
>> + * no register access will be done.
>> + */
>
> Are you sure? Didn't we see late access to DBI registers on sc7280?
>
yeah you are correct I will add a check to disable icc only in suspend
to idle case. only in suspend to ram case we see the DBI access in sc7280
-Krishna Chaitanya
>> + ret = icc_disable(pcie->icc_cpu);
>> + if (ret)
>> + dev_err(dev, "failed to disable icc path of CPU-PCIe: %d\n", ret);
>
> s/failed to disable icc path/Failed to disable Interconnect path between CPU-PCIe
>
>> +
>> + return ret;
>> }
>>
>> static int qcom_pcie_resume_noirq(struct device *dev)
>> @@ -1642,6 +1666,12 @@ static int qcom_pcie_resume_noirq(struct device *dev)
>> struct qcom_pcie *pcie = dev_get_drvdata(dev);
>> int ret;
>>
>> + ret = icc_enable(pcie->icc_cpu);
>> + if (ret) {
>> + dev_err(dev, "failed to enable icc path of CPU-PCIe: %d\n", ret);
>
> Same as above.
>
> - Mani
>
Powered by blists - more mailing lists