Re: Issues with suspend/resume

On Fri, 26 Sep 2025, Ben Hutton wrote:

Hi Ben,

> Since upgrading from 1500061 I've been having issues with suspend. Firstly 
> often the machine will restart when closing the lid (triggering suspend) but 
> mostly USB-A issues. This appears to occur after you resume then plug in 
> mains power. I get the following in dmesg.
>
> ugen0.2: <Logitech USB Optical Mouse> at usbus0 (disconnected)
> usbhid0: at uhub1, port 1, addr 1 (disconnected)
> hms0: detached
> hidbus0: detached
> usbhid0: detached
>
>
> Unplugging and plugging the device (mostly a mouse) does not work. The only 
> resolutions come from rebooting or going again into suspend and plugging the 
> mouse of power in before it starts up. I haven't found another way as yet to 
> bring the USB-A ports back online. Note that the USB-C ports continue to 
> work. Also note that if I get USB-A working again but suspending/resuming the 
> laptop crashes and reboots the next time I unplug devices. Though I'm not 
> sure it's the USB-A that's causing it, since I'm also unplugging USB-C 
> devices at the same time.
>
> This started on FreeBSD 15.0-ALPHA1 and now on ALPHA3. I have filed a bug 
> report (id: 289396) however so far no activity so I'm looking for ways to 
> debug this. Also DRM is compiled from the latest ports as of the 21st (Last 
> Sunday).
>
> Note the laptop is a Thinkpad P1 Gen3 which has up until recently worked very 
> well with suspend/resume. I've also run a full hardware test. Note that this 
> has hybrid graphics mode with an Nivida Quadro T2000 with the following drm 
> drivers.
>
> drm-66-kmod-6.6.25.1500064_5
> libdrm-2.4.123,1
> linux-rl9-libdrm-2.4.123
> nvidia-drm-66-kmod-580.82.07.1500064_3
>
> I'm more than happy to assist in debugging this issue but I am unsure where 
> to begin. I'm assuming compiling the debug kernel could help? How do I then 
> capture relevant debugging information?
>
> Note I have the same version (FreeBSD 15.0-ALPHA3) running on another laptop 
> (Thinkpad T530) which appears to working well with suspend/resume. The other 
> laptop doesn't have an Nvidia GPU. I'm suspecting this may be an issue with 
> the Nvidia drm driver?

But NVIDIA has nothing to do with XHCI and USB not working, would it?

nvidia-drm should be fixed; but ALPHA1..3 did not have the LinuxKPI
yet causing trouble.  Whatever it is now given releng/15.0 was not branched (ALPHA5?)
should have it all updated but you'll need to recomplie nvidia etc.

> Hopefully this all makes sense. Happy to provide any other 
> information/logging etc.

I wonder more if this is USBDBc related but only the client part should have
gone in.  Then there are also the suspend/resume changes that went in.

hah, looking at the PR:

acpi_pwr_switch_consumer: new power state D3cold is not the one requested D3hot
acpi_pwr_switch_consumer: new power state D3cold is not the one requested D3hot
acpi_pwr_switch_consumer: new power state D3cold is not the one requested D3hot
acpi_pwr_switch_consumer: new power state D3cold is not the one requested D3hot
acpi_pwr_switch_consumer: new power state D3cold is not the one requested D3hot

You may want to talk to obiwac@ (I'll add the Cc:)

/bz

-- 
Bjoern A. Zeeb                                                     r15:7