[Bug 280028] S3 suspend of Thinkpad x270 stopped working after freebsd-update to 14.1-RELEASE-p1

From: <bugzilla-noreply_at_freebsd.org>
Date: Sat, 29 Jun 2024 04:52:45 UTC
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=280028

--- Comment #17 from cos <freebsd.org@fydagy.netizen.se> ---
I am quite confident the old 14.1 works. This laptop is my daily driver, so
there were a couple of weeks of trouble-free suspend-resume. I am also still
able to consistently reproduce the correct behaviour with 14.1 while
consistently hitting the bug with 14.1-p1.

On the affected host, everything is installed using freebsd-update. I'm able to
build a custom kernel if that helps.

One thing I find unexpected, which while likely is nothing, is maybe worth
mentioning. Both kernels in 14.1-p1 are '14.1-RELEASE' according to `uname -r`.
I.e. no '-p1' suffix on the newer one. That is also the case on another of my
machines. My guess it that it just means no actual new kernel was included in
14.1-p1. Besides, the bug is triggered by switching the boot environment while
keeping the 14.0-p6 kernel.

The output of dmesg filtered for obviously uninteresting differences show:

    -Timecounter "TSC-low" frequency 1247997666 Hz quality 1000
    +Timecounter "TSC-low" frequency 1248000369 Hz quality 1000
    -drmn0: [drm] [ENCODER:102:DDI B/PHY B] is disabled/in DSI mode with an
ungated DDI clock, gate it
    -drmn0: [drm] [ENCODER:116:DDI C/PHY C] is disabled/in DSI mode with an
ungated DDI clock, gate it

Where I suspect those lines about ungated graphics stuff seems likely to be
related, having read that graphics is "always" the cause for suspend-resume
trouble.

The output from kldstat appears to only differ in load order:

    --- kldstat,14.0-RELEASE-p6,be_20240603 2024-06-29 05:36:31.201634000 +0200
    +++ kldstat,14.0-RELEASE-p6,be_20240620 2024-06-29 05:52:09.217792000 +0200
    @@ -1,8 +1,8 @@
     Id Refs Address                Size Name
      1  144 0xffffffff80200000  1d345b0 kernel
    - 2    1 0xffffffff81f35000   5d5958 zfs.ko
    - 3    1 0xffffffff8250b000    1c3a8 geom_eli.ko
    - 4    1 0xffffffff82529000     7718 cryptodev.ko
    + 2    1 0xffffffff81f36000    1c3a8 geom_eli.ko
    + 3    1 0xffffffff81f53000     7718 cryptodev.ko
    + 4    1 0xffffffff81f5b000   5d5958 zfs.ko
      5    1 0xffffffff835f9000     3558 fdescfs.ko
      6    1 0xffffffff835fd000   1b3920 i915kms.ko
      7    2 0xffffffff837b1000    7c050 drm.ko

Analysis of procstat output is slightly outside of my current comfort zone.
Please find the files attached.

-- 
You are receiving this mail because:
You are the assignee for the bug.