[PATCH] PCI: keystone: deinitialize endpoint on remove

[PATCH] PCI: keystone: deinitialize endpoint on remove

[Guangshuo Li]

ks_pcie_probe() initializes the DesignWare PCIe endpoint with
dw_pcie_ep_init(). The probe failure path calls dw_pcie_ep_deinit()
when endpoint register initialization fails, but the remove path does
not perform the corresponding endpoint teardown after a successful
probe.

The successful endpoint initialization also calls pci_epc_init_notify().
Without the matching teardown on removal, the EPC initialization state
and resources allocated by the DesignWare endpoint core are left
active after the driver is removed.

Notify the endpoint framework about deinitialization and call
dw_pcie_ep_deinit() before disabling runtime PM and the PHYs.

This issue was found by manual code inspection.

Fixes: 23284ad677a9 ("PCI: keystone: Add support for PCIe EP in AM654x Platforms")
Cc: stable@vger.kernel.org
Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
---
 drivers/pci/controller/dwc/pci-keystone.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/drivers/pci/controller/dwc/pci-keystone.c b/drivers/pci/controller/dwc/pci-keystone.c
index 602516239a57..35787e727c5c 100644
--- a/drivers/pci/controller/dwc/pci-keystone.c
+++ b/drivers/pci/controller/dwc/pci-keystone.c
@@ -1349,9 +1349,17 @@ static void ks_pcie_remove(struct platform_device *pdev)
 {
 	struct keystone_pcie *ks_pcie = platform_get_drvdata(pdev);
 	struct device_link **link = ks_pcie->link;
+	struct dw_pcie *pci = ks_pcie->pci;
+	const struct ks_pcie_of_data *data;
 	int num_lanes = ks_pcie->num_lanes;
 	struct device *dev = &pdev->dev;
 
+	data = of_device_get_match_data(dev);
+	if (data->mode == DW_PCIE_EP_TYPE) {
+		pci_epc_deinit_notify(pci->ep.epc);
+		dw_pcie_ep_deinit(&pci->ep);
+	}
+
 	pm_runtime_put(dev);
 	pm_runtime_disable(dev);
 	ks_pcie_disable_phy(ks_pcie);
-- 
2.43.0

Re: [PATCH] PCI: keystone: deinitialize endpoint on remove

[sashiko-bot]

> ks_pcie_probe() initializes the DesignWare PCIe endpoint with
> dw_pcie_ep_init(). The probe failure path calls dw_pcie_ep_deinit()
> when endpoint register initialization fails, but the remove path does
> not perform the corresponding endpoint teardown after a successful
> probe.
> 
> The successful endpoint initialization also calls pci_epc_init_notify().
> Without the matching teardown on removal, the EPC initialization state
> and resources allocated by the DesignWare endpoint core are left
> active after the driver is removed.
> 
> Notify the endpoint framework about deinitialization and call
> dw_pcie_ep_deinit() before disabling runtime PM and the PHYs.
> 
> This issue was found by manual code inspection.
> 
> Fixes: 23284ad677a9 ("PCI: keystone: Add support for PCIe EP in AM654x Platforms")
> Cc: stable@vger.kernel.org
> Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>

Sashiko has reviewed this patch and found no issues. It looks great!

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260916080030.2960608-1-lgs201920130244@gmail.com?part=1

Re: [PATCH] PCI: keystone: deinitialize endpoint on remove

[Marek Vasut]

On 9/16/26 10:00 AM, Guangshuo Li wrote:
> ks_pcie_probe() initializes the DesignWare PCIe endpoint with
> dw_pcie_ep_init(). The probe failure path calls dw_pcie_ep_deinit()
> when endpoint register initialization fails, but the remove path does
> not perform the corresponding endpoint teardown after a successful
> probe.
> 
> The successful endpoint initialization also calls pci_epc_init_notify().
> Without the matching teardown on removal, the EPC initialization state
> and resources allocated by the DesignWare endpoint core are left
> active after the driver is removed.
> 
> Notify the endpoint framework about deinitialization and call
> dw_pcie_ep_deinit() before disabling runtime PM and the PHYs.
> 
> This issue was found by manual code inspection.
> 
> Fixes: 23284ad677a9 ("PCI: keystone: Add support for PCIe EP in AM654x Platforms")
> Cc: stable@vger.kernel.org
> Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
> ---
>   drivers/pci/controller/dwc/pci-keystone.c | 8 ++++++++
>   1 file changed, 8 insertions(+)
> 
> diff --git a/drivers/pci/controller/dwc/pci-keystone.c b/drivers/pci/controller/dwc/pci-keystone.c
> index 602516239a57..35787e727c5c 100644
> --- a/drivers/pci/controller/dwc/pci-keystone.c
> +++ b/drivers/pci/controller/dwc/pci-keystone.c
> @@ -1349,9 +1349,17 @@ static void ks_pcie_remove(struct platform_device *pdev)
>   {
>   	struct keystone_pcie *ks_pcie = platform_get_drvdata(pdev);
>   	struct device_link **link = ks_pcie->link;
> +	struct dw_pcie *pci = ks_pcie->pci;
> +	const struct ks_pcie_of_data *data;
>   	int num_lanes = ks_pcie->num_lanes;
>   	struct device *dev = &pdev->dev;
>   
> +	data = of_device_get_match_data(dev);

Could you maybe cache the mode in struct keystone_pcie {} instead ?

> +	if (data->mode == DW_PCIE_EP_TYPE) {
> +		pci_epc_deinit_notify(pci->ep.epc);
> +		dw_pcie_ep_deinit(&pci->ep);

Would it make sense to make dw_pcie_ep_deinit() call 
pci_epc_deinit_notify() , to avoid duplication in controller drivers ?

> +	}
> +
>   	pm_runtime_put(dev);
>   	pm_runtime_disable(dev);
>   	ks_pcie_disable_phy(ks_pcie);

-- 
Best regards,
Marek Vasut

Re: [PATCH] PCI: keystone: deinitialize endpoint on remove

[Guangshuo Li]

Hi Marek,

Thanks for the review.

On Thu, 17 Sept 2026 at 06:56, Marek Vasut <marek.vasut@mailbox.org> wrote:
>
> On 9/16/26 10:00 AM, Guangshuo Li wrote:
> > ks_pcie_probe() initializes the DesignWare PCIe endpoint with
> > dw_pcie_ep_init(). The probe failure path calls dw_pcie_ep_deinit()
> > when endpoint register initialization fails, but the remove path does
> > not perform the corresponding endpoint teardown after a successful
> > probe.
> >
> > The successful endpoint initialization also calls pci_epc_init_notify().
> > Without the matching teardown on removal, the EPC initialization state
> > and resources allocated by the DesignWare endpoint core are left
> > active after the driver is removed.
> >
> > Notify the endpoint framework about deinitialization and call
> > dw_pcie_ep_deinit() before disabling runtime PM and the PHYs.
> >
> > This issue was found by manual code inspection.
> >
> > Fixes: 23284ad677a9 ("PCI: keystone: Add support for PCIe EP in AM654x Platforms")
> > Cc: stable@vger.kernel.org
> > Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
> > ---
> >   drivers/pci/controller/dwc/pci-keystone.c | 8 ++++++++
> >   1 file changed, 8 insertions(+)
> >
> > diff --git a/drivers/pci/controller/dwc/pci-keystone.c b/drivers/pci/controller/dwc/pci-keystone.c
> > index 602516239a57..35787e727c5c 100644
> > --- a/drivers/pci/controller/dwc/pci-keystone.c
> > +++ b/drivers/pci/controller/dwc/pci-keystone.c
> > @@ -1349,9 +1349,17 @@ static void ks_pcie_remove(struct platform_device *pdev)
> >   {
> >       struct keystone_pcie *ks_pcie = platform_get_drvdata(pdev);
> >       struct device_link **link = ks_pcie->link;
> > +     struct dw_pcie *pci = ks_pcie->pci;
> > +     const struct ks_pcie_of_data *data;
> >       int num_lanes = ks_pcie->num_lanes;
> >       struct device *dev = &pdev->dev;
> >
> > +     data = of_device_get_match_data(dev);
>
> Could you maybe cache the mode in struct keystone_pcie {} instead ?
>
> > +     if (data->mode == DW_PCIE_EP_TYPE) {
> > +             pci_epc_deinit_notify(pci->ep.epc);
> > +             dw_pcie_ep_deinit(&pci->ep);
>
> Would it make sense to make dw_pcie_ep_deinit() call
> pci_epc_deinit_notify() , to avoid duplication in controller drivers ?
>
> > +     }
> > +
> >       pm_runtime_put(dev);
> >       pm_runtime_disable(dev);
> >       ks_pcie_disable_phy(ks_pcie);
>
> --
> Best regards,
> Marek Vasut

Agreed. I'll cache the mode in struct keystone_pcie and move
pci_epc_deinit_notify() into dw_pcie_ep_deinit() so controller drivers
do not need to duplicate it. I'll send a v2 with those changes.

Thanks,
Guangshuo

Re: [PATCH] PCI: keystone: deinitialize endpoint on remove

[Marek Vasut]

On 9/21/26 4:15 PM, Guangshuo Li wrote:
> Hi Marek,
> 
> Thanks for the review.
> 
> On Thu, 17 Sept 2026 at 06:56, Marek Vasut <marek.vasut@mailbox.org> wrote:
>>
>> On 9/16/26 10:00 AM, Guangshuo Li wrote:
>>> ks_pcie_probe() initializes the DesignWare PCIe endpoint with
>>> dw_pcie_ep_init(). The probe failure path calls dw_pcie_ep_deinit()
>>> when endpoint register initialization fails, but the remove path does
>>> not perform the corresponding endpoint teardown after a successful
>>> probe.
>>>
>>> The successful endpoint initialization also calls pci_epc_init_notify().
>>> Without the matching teardown on removal, the EPC initialization state
>>> and resources allocated by the DesignWare endpoint core are left
>>> active after the driver is removed.
>>>
>>> Notify the endpoint framework about deinitialization and call
>>> dw_pcie_ep_deinit() before disabling runtime PM and the PHYs.
>>>
>>> This issue was found by manual code inspection.
>>>
>>> Fixes: 23284ad677a9 ("PCI: keystone: Add support for PCIe EP in AM654x Platforms")
>>> Cc: stable@vger.kernel.org
>>> Signed-off-by: Guangshuo Li <lgs201920130244@gmail.com>
>>> ---
>>>    drivers/pci/controller/dwc/pci-keystone.c | 8 ++++++++
>>>    1 file changed, 8 insertions(+)
>>>
>>> diff --git a/drivers/pci/controller/dwc/pci-keystone.c b/drivers/pci/controller/dwc/pci-keystone.c
>>> index 602516239a57..35787e727c5c 100644
>>> --- a/drivers/pci/controller/dwc/pci-keystone.c
>>> +++ b/drivers/pci/controller/dwc/pci-keystone.c
>>> @@ -1349,9 +1349,17 @@ static void ks_pcie_remove(struct platform_device *pdev)
>>>    {
>>>        struct keystone_pcie *ks_pcie = platform_get_drvdata(pdev);
>>>        struct device_link **link = ks_pcie->link;
>>> +     struct dw_pcie *pci = ks_pcie->pci;
>>> +     const struct ks_pcie_of_data *data;
>>>        int num_lanes = ks_pcie->num_lanes;
>>>        struct device *dev = &pdev->dev;
>>>
>>> +     data = of_device_get_match_data(dev);
>>
>> Could you maybe cache the mode in struct keystone_pcie {} instead ?
>>
>>> +     if (data->mode == DW_PCIE_EP_TYPE) {
>>> +             pci_epc_deinit_notify(pci->ep.epc);
>>> +             dw_pcie_ep_deinit(&pci->ep);
>>
>> Would it make sense to make dw_pcie_ep_deinit() call
>> pci_epc_deinit_notify() , to avoid duplication in controller drivers ?
>>
>>> +     }
>>> +
>>>        pm_runtime_put(dev);
>>>        pm_runtime_disable(dev);
>>>        ks_pcie_disable_phy(ks_pcie);
>>
>> --
>> Best regards,
>> Marek Vasut
> 
> Agreed. I'll cache the mode in struct keystone_pcie and move
> pci_epc_deinit_notify() into dw_pcie_ep_deinit() so controller drivers
> do not need to duplicate it. I'll send a v2 with those changes.
Thank you