On Tue, Mar 12, 2019 at 03:05:48PM +0300, Sergey Miroshnichenko wrote:
> During the safe removal procedure, a Data Link Layer State Changed event
> may occur after pciehp_power_off_slot(), and it is handled when the slot is
> already set to OFF_STATE. This results in re-enabling the device and makes
> it impossible to actually safely remove it.
>
> Clear out the Presence Detect Changed and Data Link Layer State Changed
> events when the disabled slot has settled down.
>
> It is still possible to re-enable the device if it remains in the slot
> after pressing the Attention Button by pressing it again.
>
> Signed-off-by: Sergey Miroshnichenko <s.miroshnichenko@xxxxxxxxx>
Reviewed-by: Lukas Wunner <lukas@xxxxxxxxx>
Did this work correctly before v4.19 or is it a regression caused by
d331710ea78f ("PCI: pciehp: Become resilient to missed events")?
Either way, it seems reasonable to backport it to stable and
v4.19 is the earliest that it should apply cleanly to, so:
Cc: stable@xxxxxxxxxxxxxxx # v4.19+
Thanks a lot for catching this, I wasn't able to test the code paths
which are only exercised if a power controller is present (because
Thunderbolt doesn't have one).
Lukas
> ---
> drivers/pci/hotplug/pciehp_ctrl.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/drivers/pci/hotplug/pciehp_ctrl.c b/drivers/pci/hotplug/pciehp_ctrl.c
> index 3f3df4c29f6e..905282a8ddaa 100644
> --- a/drivers/pci/hotplug/pciehp_ctrl.c
> +++ b/drivers/pci/hotplug/pciehp_ctrl.c
> @@ -115,6 +115,10 @@ static void remove_board(struct controller *ctrl, bool safe_removal)
> * removed from the slot/adapter.
> */
> msleep(1000);
> +
> + /* Ignore link or presence changes caused by power off */
> + atomic_and(~(PCI_EXP_SLTSTA_DLLSC | PCI_EXP_SLTSTA_PDC),
> + &ctrl->pending_events);
> }
>
> /* turn off Green LED */
> --
> 2.20.1
