Re: Panther Lake UCSI: PPM forces DP alt mode pin assignment D (2 lanes) for C/E-only partner, no 4K60 (was: [BUG] Type-C Subsystem xHCI (8086:e331, Panther Lake) — repeated err -71 / port reset failures under sustained DP+HID load)

[Heikki Krogerus <heikki.krogerus@linux.intel.com>]

On Fri, Oct 23, 2026 at 10:27:32PM +0200, Rodolfo Semprevivo wrote:
> Hi,
> 
> follow-up to my report below, with a corrected analysis. The xHCI was
> the wrong suspect: the -71 errors come from the monitor's internal USB
> 2.0 hub, which sits on the standalone xHCI (8086:e37d, 00:14.0), not on
> 8086:e331. They only appear because the monitor resets when the DP link
> fails.
> 
> Root cause: on the left TB4 port the PPM/EC configures DP alt mode with
> pin assignment D (2 lanes), although the monitor only advertises C and E:
> 
>   port0-partner.0: svid ff01, active yes, vdo 0x00140045
>   displayport/pin_assignment: C E   (nothing selected)
>   displayport/configuration: [USB] sink
>   xe: Port D/TC#1: TC port mode reset (disconnected -> dp-alt) pin
> assignment: D max lanes: 2
>   xe: [LTTPR 1] Link Training passed at link rate = 540000, lane count = 2
> 
> 2x HBR2 is not enough for 3840x2160@60 (594 MHz), so the picture drops
> regardless of bpc/HDR. Lower resolutions at 60 Hz work. This also
> explains why the dGPU route via the MUX failed identically: the lane
> split is decided by the firmware, not by the GPU driver.
> 
> For comparison, an Alder Lake ASUS GU603ZW on the same kernel, with the
> same monitor and cable:
> 
>   i915: Port D/TC#1: ... pin assignment: C max lanes: 4
>   i915: Link Training passed at link rate = 540000, lane count = 4
> 
> and 4K60 is stable.
> 
> Linux cannot correct this. Writes to pin_assignment/configuration fail
> with EOPNOTSUPP:
> 
>   typec_displayport port0-partner.0: firmware doesn't support alternate mode
> overriding
>   ucsi_acpi USBC000:00: UCSI_GET_PDOS failed (-95)
> 
> So this looks like a firmware issue (reported to ASUS as well).
> Questions:
> 1. Is the PPM choosing pin assignment D for a partner that does not
>    advertise D a known issue on Panther Lake?

I have not heard of this particular problem before.

> 2. Should Linux warn about it (partner VDO lacks D but FIA reports D),
>    or avoid the resulting 2-lane DSC mode?

You tell me. If you think that this would be useful, prepare the patch
for it.

> System: ASUS ROG Zephyrus G16 GU606AX (Panther Lake, Core Ultra 9 386H),
> BIOS GU606AX.304, kernel 7.3.0-4 (CachyOS), monitor ASUS ProArt
> PA169CDV. Bug report: https://paste.cachyos.org/p/83e1bcf.log
> Full drm.debug log available on request.

Thanks,

-- 
heikki