Acpi_task pegging CPU

Hi All, I have beta 5 installed and now my acpi_task thread in the kernel is pegging a core in the CPU at 100%. This, of course, causes the fan to spin wildly. I just updated the bits to the latest, and the same issue is occuring.

Dell G7 laptop, lots of cores.

Thanks!

1 Like

Hello, @MikeC

I can’t offer much in terms of help on an actual fix for your issue, but maybe you could try (just as a temporal workaround, mind you) disabling ACPI from the boot-loader? (if it helps, you can make that setting “permanent” by setting acpi false in the ~/config/settings/kernel/drivers/kernel file).

In any case, opening a bug report with your hardware info sounds like a good idea, no?

Please consult the Reporting bugs guide.

1 Like

This happens when the the lock/signaling between hardware controller and OS don’t work. It never waits when it should, instead loops forever.
Most likely due to your hardware having some other mechanism for syncing, but could also be a bug. Your best option is to report a bug with your exact hardware so we can try to find the bug or implement support for that kind of hardware…

If you look for the SD card reader in that Dell G7 in Devices, is the driver using ACPI? The syncing mechanism would be through GPIO interrupts, but Haiku didn’t support them when I added ACPI device support to the SD driver, so it falls back to polling. I don’t think support was added since.

1 Like

Gosh, there are so many ACPI devices in the list. Nothing stands out as a SD device however.

Never mind… the polling in the ACPI SD card driver is on a separate thread (it would show up as “SD bus poller”, so not “acpi_task”), so it’s most likely something else.