https://bugzilla.kernel.org/show_bug.cgi?id=220796
Javier Folcini ([email protected]) changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #4 from Javier Folcini ([email protected]) --- Adding data for the AMD variant of this model family — same ECRD failure, same root cause. Machine: ThinkPad E14 Gen 7 (AMD), model 21T0CTO1WW BIOS: R2ZET27W (1.10), EC: R2ZHT27W OS: Ubuntu 24.04.5, kernel 7.0.0-38-generic thinkpad_acpi: v0.26 Boot log (identical signature to the Intel reports): thinkpad_acpi: ThinkPad ACPI Extras v0.26 thinkpad_acpi: ThinkPad BIOS R2ZET27W (1.10 ), EC R2ZHT27W thinkpad_acpi: Lenovo ThinkPad E14 Gen 7, model 21T0CTO1WW thinkpad_acpi: acpi_evalf((null), dd, ...) failed: AE_TYPE thinkpad_acpi: ThinkPad ACPI EC unable to access EC version thinkpad_acpi: acpi_evalf((null), dd, ...) failed: AE_TYPE thinkpad_acpi: ThinkPad ACPI EC access misbehaving, disabling thermal sensors access thinkpad_acpi: acpi_evalf((null), dd, ...) failed: AE_TYPE thinkpad_acpi: ThinkPad ACPI EC access misbehaving, fan status and control unavailable Findings from analyzing the DSDT on this unit: 1. EC0.ECRD takes a Reference argument and dereferences it (DerefOf(Arg0)); thinkpad_acpi's acpi_evalf passes a plain integer offset, so the call aborts with AE_TYPE. This kills all EC access through ACPI (EC version, thermal sensors, fan). 2. Despite the broken ACPI path, the EC register map appears to be the classic ThinkPad layout: a local workaround module reads the tachometer via the kernel's ec_read() at offsets 0x84 (LSB) / 0x85 (MSB) and gets correct, stable RPM values on this unit. The EC itself behaves normally (autonomous fan control works); only the ACPI method call convention is broken. 3. The ACPI fan device (\_SB.FAN0, PNP0C0B) is not usable as a fallback: its _FST returns a firmware variable (CRPM) that only ever holds an OS-commanded speed, so it reads 0 RPM under normal EC-managed operation. acpi_fan binds with "Using Microsoft fan extensions" and exposes only a 2-state (0/1) cooling device. Symptom difference vs. the Intel reports: on this AMD unit the EC keeps running the fan autonomously (spins/stops normally per its own curve), so the impact is loss of monitoring and control from Linux, not a silent fan. Happy to provide an acpidump or test patches on this unit. -- 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
