[PATCH] usb: hcd-pci: avoid double disable on surprise removal

[PATCH] usb: hcd-pci: avoid double disable on surprise removal

[Konrad Dybcio]

A PCI HCD can be surprise-removed after suspend_common() disables it
and before resume_common() re-enables it. Avoid a second call to
pci_disable_device() in the remove path when no enable reference
remains.

Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
---
 drivers/usb/core/hcd-pci.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/core/hcd-pci.c b/drivers/usb/core/hcd-pci.c
index cd223475917ef199511311eda5c6aaef073b0ad4..34c4d352b40c40c54cf17c30fc264833ba05bcd6 100644
--- a/drivers/usb/core/hcd-pci.c
+++ b/drivers/usb/core/hcd-pci.c
@@ -348,7 +348,9 @@ void usb_hcd_pci_remove(struct pci_dev *dev)
 	usb_put_hcd(hcd);
 	if ((hcd_driver_flags & HCD_MASK) < HCD_USB3)
 		pci_free_irq_vectors(dev);
-	pci_disable_device(dev);
+
+	if (pci_is_enabled(dev))
+		pci_disable_device(dev);
 }
 EXPORT_SYMBOL_GPL(usb_hcd_pci_remove);
 

---
base-commit: f5f84daefcd92d7a630066635ecea1433ed5eac7
change-id: 20261020-topic-hub_pci-903b8c18d17b

Best regards,
-- 
Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>

Re: [PATCH] usb: hcd-pci: avoid double disable on surprise removal

[Alan Stern]

On Tue, Oct 20, 2026 at 05:05:39PM +0200, Konrad Dybcio wrote:
> A PCI HCD can be surprise-removed after suspend_common() disables it
> and before resume_common() re-enables it. Avoid a second call to
> pci_disable_device() in the remove path when no enable reference
> remains.

Does the double disable actually cause any problems?

Alan Stern

> Signed-off-by: Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> ---
>  drivers/usb/core/hcd-pci.c | 4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/core/hcd-pci.c b/drivers/usb/core/hcd-pci.c
> index cd223475917ef199511311eda5c6aaef073b0ad4..34c4d352b40c40c54cf17c30fc264833ba05bcd6 100644
> --- a/drivers/usb/core/hcd-pci.c
> +++ b/drivers/usb/core/hcd-pci.c
> @@ -348,7 +348,9 @@ void usb_hcd_pci_remove(struct pci_dev *dev)
>  	usb_put_hcd(hcd);
>  	if ((hcd_driver_flags & HCD_MASK) < HCD_USB3)
>  		pci_free_irq_vectors(dev);
> -	pci_disable_device(dev);
> +
> +	if (pci_is_enabled(dev))
> +		pci_disable_device(dev);
>  }
>  EXPORT_SYMBOL_GPL(usb_hcd_pci_remove);
>  
> 
> ---
> base-commit: f5f84daefcd92d7a630066635ecea1433ed5eac7
> change-id: 20261020-topic-hub_pci-903b8c18d17b
> 
> Best regards,
> -- 
> Konrad Dybcio <konrad.dybcio@oss.qualcomm.com>
> 
> 

Re: [PATCH] usb: hcd-pci: avoid double disable on surprise removal

[Alan Stern]

On Tue, Oct 20, 2026 at 03:52:28PM -0400, Alan Stern wrote:
> On Tue, Oct 20, 2026 at 05:05:39PM +0200, Konrad Dybcio wrote:
> > A PCI HCD can be surprise-removed after suspend_common() disables it
> > and before resume_common() re-enables it. Avoid a second call to
> > pci_disable_device() in the remove path when no enable reference
> > remains.
> 
> Does the double disable actually cause any problems?

Also, is your patch any different from this one from two days ago?

https://lore.kernel.org/linux-usb/20261018164134.134106-1-cole@unwrap.rs/

Alan Stern