[BUG] driver: [AMD ACP] System Deadlock & Playback Freeze due to snd_hda_intel and acp63 conflict on ThinkPad P14s Gen 5

[UsQuake]

Describe the bug
On the Lenovo ThinkPad P14s Gen 5 AMD, the audio subsystem causes a complete system deadlock, freezing all video/audio playback and permanently illuminating the hardware mic-mute LED.

Looking at the ALSA info, it appears to be a dual-binding/initialization conflict. The legacy snd_hda_intel driver successfully binds to the AMD Ryzen HD Audio Controller (Card 1: HD-Audio Generic for the ALC257 codec). However, the AMD ACP driver (snd_soc_ps_mach) also loads simultaneously, creating Card 2 (acp63) for the same hardware subsystem. This turf war over the DSP control causes PipeWire/PulseAudio to hang indefinitely.

Workaround tried: The only way to restore system stability and allow video playback to flow normally is to completely blacklist all related audio modules (snd_sof, snd_hda_intel, snd_soc_ps_mach, snd_pci_ps, etc.).

To Reproduce
Steps to reproduce the behavior:

1. Boot Fedora 43 with Kernel 6.18.13 on a Lenovo ThinkPad P14s Gen 5 AMD without any modprobe blacklists.

2. The kernel loads both snd_hda_intel and acp_ps_mach modules.

3. Open a web browser and attempt to play a YouTube video, or run speaker-test / pavucontrol in the terminal.

4. The media player freezes at 0:00, the system I/O stalls, and the hardware mic-mute LED turns on solid.

Reproduction Rate
100% (All the time upon booting without the blacklist workaround).

Expected behavior
The kernel should properly quirk this specific Lenovo model to use either the legacy HDA driver OR the newer SOF/ACP driver exclusively. The ALC257 codec should be initialized without causing a deadlock.

Impact
Showstopper. I am forced to completely blacklist the internal audio drivers to maintain basic system stability.

Environment
As I am using a packaged Linux distribution (Fedora), here are the package versions instead of git hashes:

1. Branch name and commit hash of the 3 repositories:

   • linux (kernel driver): 6.18.13-200.fc43.x86_64

   • sof (firmware): alsa-sof-firmware-2025.12.2-1.fc43.noarch

   • soft (tools & topology): alsa-utils default

2. Name of the topology file

   • Topology: Default topology loaded for AMD ACP6.3 / Realtek ALC257

3. Name of the platform(s) on which the bug is observed.

   • Platform: LENOVO 21MES00400 (ThinkPad P14s Gen 5 AMD), BIOS: R2LET36W (1.17)

Screenshots or console output
Here is the relevant dmesg snippet showing both snd_hda_intel and snd_pci_ps attempting to enable the device concurrently:

[    6.114006] snd_hda_intel 0000:c4:00.1: enabling device (0000 -> 0002)
[    6.115008] snd_hda_intel 0000:c4:00.1: Handle vga_switcheroo audio client
[    6.115569] snd_hda_intel 0000:c4:00.6: enabling device (0000 -> 0002)
...
[    6.146014] input: ThinkPad Extra Buttons as /devices/platform/thinkpad_acpi/input/input13
[    6.153309] snd_pci_ps 0000:c4:00.5: enabling device (0000 -> 0002)
[    6.164690] snd_hda_codec_alc269 hdaudioC1D0: ALC257: picked fixup  for PCI SSID 17aa:0000
[    6.165009] snd_hda_codec_alc269 hdaudioC1D0: autoconfig for ALC257: line_outs=1 (0x14/0x0/0x0/0x0/0x0) type:speaker

Snippet from alsa-info.sh showing the dual-card split:

!!Soundcards recognised by ALSA
!!-----------------------------
 0 [Generic        ]: HDA-Intel - HD-Audio Generic
                      HD-Audio Generic at 0x905c8000 irq 157
 1 [Generic_1      ]: HDA-Intel - HD-Audio Generic
                      HD-Audio Generic at 0x905c0000 irq 158
 2 [acp63          ]: acp63 - acp63
                      LENOVO-21MES00400-ThinkPadP14sGen5AMD

alsa_info.txt
audio_crash_log.txt