On Thu, Feb 22, 2024 at 04:46:16PM +0200, Mathias Nyman wrote: > On 22.2.2024 16.06, Greg KH wrote: > > On Thu, Feb 22, 2024 at 03:38:18PM +0200, Mathias Nyman wrote: > > > Ports with _UPC (USB Port Capability) ACPI objects stating they are > > > "not connectable" are not wired to any connector or internal device. > > > They only exist inside the host controller. > > > > > > These ports may not have an ACPI _PLD (Physical Location of Device) > > > object. > > > > > > Rework the code so that _UPC is read even if _PLD does not exist, and > > > make sure the port->connect_type is set to "USB_PORT_NOT_USED" instead > > > of "USB_PORT_CONNECT_TYPE_UNKNOWN". > > > > > > Signed-off-by: Mathias Nyman <mathias.nyman@xxxxxxxxxxxxxxx> > > > > Does patch 2/2 need this? If so, why isn't it marked for stable? > > 2/2 alone fixes the real world port peering problem seen. > > This is something I stumbled upon while debugging that issue. > This patch just makes sure we don't skip marking some unused ports as > unused due to how we parse ACPI tables. Ok, so should patch 1/2 go to usb-next and patch 2/2 go to usb-linus? confused, greg k-h
