https://bugzilla.kernel.org/show_bug.cgi?id=220796

--- Comment #6 from Javier Folcini ([email protected]) ---
Follow-up to comment #4 (AMD 21T0CTO1WW), with corrections:

1. BIOS fix: Lenovo's AMD BIOS 1.12 (R2ZET29W) changelog lists "Fixed an issue
that
   linux system shows 'thinkpad_acpi: ThinkPad ACPI EC access misbehaving,
disabling
   thermal sensors access'". Confirmed on 1.27 (R2ZET38W / EC R2ZHT38W): no
AE_TYPE,
   /proc/acpi/ibm/{fan,thermal} present, fan speed and level control work with
a
   stock kernel (7.0.0-38). Intel 21SX owners should check for the equivalent
R30ET
   update (it is not on LVFS, see fwupd/firmware-lenovo#603).

2. Correction to #4: I wrote that the EC's autonomous fan control "works" and
that the
   impact was only loss of monitoring. That was wrong. With EC auto (level
auto), in
   the Performance platform profile the fan stayed at <=400 RPM even at 96 C
Tctl.
   In Balanced (where thinkpad_acpi enables DYTC AMT) EC auto did ramp
(3200-3900 RPM).
   This may be related to #3 (fan control breaking while auto-cpufreq runs).
   Note: my unit also shows the fan being cut out after 16-50 s at speed even
in EC
   auto, with bogus tach values on restart; I suspect a hardware fault on my
unit
   (warranty pending), so treat that part with caution.

3. New, with the fixed BIOS: thinkpad_acpi logs "secondary fan control detected
&
   enabled" on this single-fan model. The fan2 probe (EC 0x31 bit 0 select)
returns
   the same tach as fan1, so fan2_input mirrors fan1_input, and every speed
read now
   does read-modify-write cycles on EC 0x31. A quirk to disable second-fan
detection
   for 21T0/21T1 (and probably 21SX/21SY) seems appropriate.
Point 2 is the sensitive one. If the warranty technician finds the fan was
faulty, the Balanced-vs-Performance difference might partly reflect your fan
rather than the firmware. The draft already says so. If you'd rather keep it
strictly factual, you could leave out point 2 and post it after the repair.

-- 
You may reply to this email to add a comment.

You are receiving this mail because:
You are watching the assignee of the bug.

_______________________________________________
acpi-bugzilla mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/acpi-bugzilla

Reply via email to