Comment 34 for bug 1774950

Here's a more succinct update of how things seem to be progressing:

- There's a bugzilla.kernel.org bug report that describes this same behaviour: https://bugzilla.kernel.org/show_bug.cgi?id=199025

- A kernel bisection between v4.14 and v4.15 gives this result:
git bisect bad
a192aa923b66a435aae56983c4912ee150bc9b32 is the first bad commit
commit a192aa923b66a435aae56983c4912ee150bc9b32
Author: Rafael J. Wysocki <email address hidden>
Date: Mon Oct 16 03:29:55 2017 +0200

    ACPI / LPSS: Consolidate runtime PM and system sleep handling

    Move the LPSS-specific code from acpi_lpss_runtime_suspend()
    and acpi_lpss_runtime_resume() into separate functions,
    acpi_lpss_suspend() and acpi_lpss_resume(), respectively, and
    make acpi_lpss_suspend_late() and acpi_lpss_resume_early() use
    them too in order to unify the runtime PM and system sleep
    handling in the LPSS driver.

    Signed-off-by: Rafael J. Wysocki <email address hidden>
    Acked-by: Greg Kroah-Hartman <email address hidden>
    Reviewed-by: Ulf Hansson <email address hidden>

:040000 040000 b4ef369c698a10cec8d6224fdc3b8f287eb853b1 20327192ca09ecc0cf0b24b62fdb915ea303b988 M drivers

- from an email list I saw that @kaihengfeng has contacted a kernel.org developer who has written a patch to test.

- I don't know if I'm supposed to test this patch (i might just be added to the email list because I opened this bug report) or whether Kai-Heng Feng will post here to let others try a kernel with the patch applied and see if it works.