Intel-GFX Archive on lore.kernel.org
 help / color / mirror / Atom feed
* [i915] HP Victus 15-fa1xxx: eDP-1 link training failure after s2idle resume, PPS VDD wakeref warning
@ 2026-09-03 15:12 Varun Arora
  2026-09-03 16:59 ` Jani Nikula
  0 siblings, 1 reply; 2+ messages in thread
From: Varun Arora @ 2026-09-03 15:12 UTC (permalink / raw)
  To: intel-gfx

[-- Attachment #1: Type: text/plain, Size: 3724 bytes --]

Hi,

I'm seeing a reproducible i915 eDP suspend/resume failure on an HP Victus
15-fa1xxx with an Intel Core i5-12450H and NVIDIA RTX 2050.

HARDWARE
--------

Laptop:
HP Victus 15-fa1xxx

CPU:
12th Gen Intel(R) Core(TM) i5-12450H

iGPU:
Intel Alder Lake-P GT1 [UHD Graphics]
PCI ID: 8086:46a3
Driver: i915

dGPU:
NVIDIA GeForce RTX 2050
Driver: nvidia / nvidia-drm

Internal panel:
1920x1080 @ 144Hz
eDP-1
The internal panel is physically connected to the Intel iGPU.

BIOS:
F.23 (03/30/2026)


SOFTWARE
--------

OS:
EndeavourOS / Arch Linux

Desktop:
XFCE

Session:
Wayland

Suspend:
s2idle only

Kernels tested:
7.1.11-arch1-1
6.18.48-1-lts

NVIDIA driver:
610.57.04


PROBLEM
-------

After s2idle suspend/resume, the internal display frequently takes
20-30 seconds to return.

Sometimes it takes several minutes, and sometimes the display does not
recover and continues blinking/going black until I reboot.

The kernel repeatedly reports:

    [drm] *ERROR* [CONNECTOR:...:eDP-1][ENCODER:...:DDI A/PHY A][DPRX]
    Failed to enable link training

The failure can repeat approximately every 12 seconds while the display
is trying to recover.


BOOT-TIME ERRORS
----------------

There is also an eDP-related failure during BOOT, before the desktop:

    [drm] [ENCODER:...:DDI B/PHY B]
    failed to retrieve link info, disabling eDP

Immediately around this I also get:

    drm_WARN_ON(intel_dp->pps.vdd_wakeref)

with a warning in:

    intel_pps_vdd_on_unlocked()

The call path includes:

    intel_pps_vdd_on_unlocked()
    __intel_display_power_get_domain()
    intel_pps_vdd_on()
    intel_dp_detect()
    intel_display_driver_register()
    i915_driver_probe()

After some suspend/resume failures I have also seen:

    i915 raw-wakerefs=1 wakelocks=1 on cleanup

and:

    intel_runtime_pm_driver_release


TESTS ALREADY PERFORMED
-----------------------

1. i915.enable_psr=0

   No fix.

2. i915.enable_dc=0

   No fix for the underlying problem.
   The eDP link-training failure still occurs.

3. 144Hz -> 60Hz

   Did not solve the problem.

4. Tested both:

   7.1.11-arch1-1
   6.18.48-1-lts

   Both reproduce the problem.

5. NVIDIA runtime power management

   The NVIDIA GPU successfully entered runtime suspended state,
   but the internal Intel-connected display still failed to resume
   correctly.

6. Intel Xe driver

   I experimentally tested:

       i915.force_probe=!46a3 xe.force_probe=46a3

   This did not solve the problem.

   Instead it produced PSR AUX errors, severe display corruption,
   and another very long recovery loop.

   I therefore returned to i915.


SUSPEND
-------

The system only exposes s2idle:

    cat /sys/power/mem_sleep

returns:

    [s2idle]

Attempting to use deep suspend gives:

    Invalid argument


IMPORTANT
---------

The internal panel is physically connected to the Intel iGPU.

DRM devices show:

    card0 -> NVIDIA
    card1 -> i915

and the internal connector is:

    card1-eDP-1


QUESTION
--------

Could the combination of:

    failed to retrieve link info, disabling eDP
    drm_WARN_ON(intel_dp->pps.vdd_wakeref)
    intel_pps_vdd_on_unlocked()
    Failed to enable link training

indicate an i915 PPS / panel power sequencing / BIOS handoff problem?

Has this been seen on Alder Lake HP laptops, particularly HP Victus
systems?

Is there a known i915 patch, quirk, or debugging procedure for this
PPS/VDD wakeref + eDP link-training failure?

I would especially appreciate advice from anyone familiar with i915
PPS, eDP AUX/link training, and suspend/resume on Alder Lake.

I am trying to determine whether this is primarily an i915 issue,
an HP firmware issue, or an interaction with the NVIDIA hybrid graphics
configuration.

Thanks.

[-- Attachment #2: Type: text/html, Size: 4465 bytes --]

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2026-09-03 16:59 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2026-09-03 15:12 [i915] HP Victus 15-fa1xxx: eDP-1 link training failure after s2idle resume, PPS VDD wakeref warning Varun Arora
2026-09-03 16:59 ` Jani Nikula

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox