On 11/21, Mihai Dragan wrote: > On 11/21, Mihai Dragan wrote: > > Thank you, > > > > systat / vmstat shows an average of 450 interrupts per second constantly > > coming from 'acpi0' after cold boot. > > I started disabling IO devices from BIOS and it turns out that the > > smartcard reader was a red herring: > > it's the Thunderbolt 3 controller that seems to be causing this. It's the > > only device that stops the issue from reproducing when disabled. > > > > I tried toggling the available TB3 options available in BIOS, but nothing > > makes a difference. > > It's interesting that the spurious interrupts are generated regardless if > > there's a device plugged into the thunderbolt port or not. > > Thunderbolt functionality seems perfectly fine, in both cold and warm boot. > > > > > > On 11/21, Todd C. Miller wrote: > > > On Sun, 21 Nov 2021 14:30:09 +0000, Mihai Dragan wrote: > > > > > > > Can I check from user space the interrupt count a device generates ? > > > > Any advice or hints on how to debug this further would be greatly appre > > > > ciated. > > > > > > Compare the output of "vmstat -i" after both cold and warm boot. > > > > > > - todd > > > > -- > > > I'm not the first one to hit this: > https://www.kevinthomas.dev/posts/openbsd-thinkpadt480.html > > In my case, enabling Thunderbolt BIOS Assist doesn't prevent the spurious > interrupts. It does however cause the kernel to crash on shutdown / reboot. > I'll keep looking into. > -- >
ok, so apparently this is an old issue with the ACPI tables on the t480 series. As it happens the user who detailed it on the Lenovo forums was using OpenBSD too: https://forums.lenovo.com/t5/Other-Linux-Discussions/FYI-Linux-May-Not-Support-Thunderbolt-Native-Mode/m-p/4057604?page=1#4064963 Not an OpenBSD issue at all. Thanks, Mihai --
