Re: [RFC PATCH v4] platform/x86: panasonic-laptop: Add platform_profile support

[Wu Zhipeng]

Hi Alex,

Thanks for v4 and for settling the mapping — cool/quiet/performance it is.

I applied this v4 to the v7.3 driver source and built it out-of-tree against
Debian's 7.2.9 headers (the unit currently runs 7.2.9, not 7.3; the driver
code is v4 verbatim). The 8-line CF-SV1-2 DMI entry is applied on top;
without it the profile device does not bind on this model. BIOS V2.00L24,
i5-1145G7.

  $ cat /sys/firmware/acpi/platform_profile_choices
  cool quiet performance

  target        CEFM   EPLE   read back
  cool          0x0    0x1    cool
  quiet         0x1    0x1    quiet
  performance   0x0    0x0    performance

The passive+unlocked combination (SEFM 1 + SEPL 0, written straight to the
EC) reads back as custom; the ABI itself rejects custom (EINVAL) and
max-power (EOPNOTSUPP). The values are the same through the legacy
/sys/firmware/acpi/... interface and the class device, and writing through
either frontend works.

Suspend/resume: this unit only offers s2idle (no deep S3). quiet and cool
both survived an s2idle cycle with CEFM/EPLE unchanged, and a profile write
after resume still reached the EC. No errors in dmesg on load or on profile
changes.

Evidence (choices, round-trip, custom, resume logs):
https://github.com/numuly/cf-sv1-2-platform-profile-evidence/tree/v4-test-20261026/driver

Tested-by: Wu Zhipeng <zepwu@xxxxxxxxxxx>

The CF-SV1-2 DMI entry follows as a separate patch, as you suggested.

Thanks,
Wu