Re: [RFT PATCH 2/2] usb: port: Don't block port power-off on false peer failures

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 16.2.2024 17.20, Alan Stern wrote:
On Fri, Feb 16, 2024 at 04:12:30PM +0200, Mathias Nyman wrote:
Several unused ports may share the same bogus location data in ACPI
PLD tables. This causes port peering failures as these several unused
USB2 and USB3 ports suddenly match based on their location.

Don't print the "usb: port power management may be unreliable" warning,
or block port power-off in case peering failed for two ports with
connect type set to USB_PORT_NOT_USED.

What if one port is marked USB_PORT_NOT_USED but its bogus location data
causes the system to think it should peer with a port that _is_ in use
(and maybe is already peered with a different port)?


This could happen, but based on the DELL case it seems unlikely.

The bogus location entries are filled with zeroes, while the usable ports
start their location group position with a "1".

If neither is peered and one is marked USB_PORT_NOT_USED, and their location match,
then they will peer successfully. No change due to this patch.

If one port is already peered to some other port then peering will fail,
warning message is printed and usb_port_block_power_off = 1; is set.
No change here either.

Also two USB_PORT_NOT_USED ports with bogus matching location will
peer successfully as long as neither one is peered with some other port.
No change here either.

Could make sense to skip both peering and matching with ports that are
not used.

Thanks
Mathias



[Index of Archives]     [Linux Media]     [Linux Input]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Old Linux USB Devel Archive]

  Powered by Linux