Huh. So I checked out Linus' master (currently up to 6.11-rc5) and it
seems to be doing the right thing now; I left a USB-C ALT monitor setup
plugged in when I suspended, then came back to my 4K monitor setup via
TB and it came up in the right resolution and everything.
Excellent news, so I'll keep using Linus' master until 6.11 is released.
Oh, and to answer your question, no, my system doesn't (perceptively)
wake up when suspended if I connect/disconnect USB/TB cables.
... and unfortunately all I have now is (power-hungry) s0ix sleep.
-Kenny
On 8/25/24 20:06, Lukas Wunner wrote:
[cc += Mika, linux-usb]
On Wed, Aug 21, 2024 at 03:05:59PM -0700, Kenneth Crudup wrote:
Subject says it all, but to recap my laptop doesn't detect Thunderbolt
topology changes when resuming or coming out of hibernate; i.e., the only
time a TB topology change happens is if a TB cable is disconnected while
suspended or hibernated, but if one is connected, or a different TB setup
altogether is connected when the system resumes it doesn't notice the
topology change until I disconnect and reconnect.
I'm currently running 6.10.6, but this has been going on for a while.
[ 0.000000] DMI: Dell Inc. XPS 9320/0KNXGD, BIOS 2.12.0 04/11/2024
...
[ 0.136807] smpboot: CPU0: 12th Gen Intel(R) Core(TM) i7-1280P (family:
0x6, model: 0x9a, stepping: 0x3)
This commit went into v6.11-rc1 and will at least detect replacement
of PCI devices (to a certain extent):
https://git.kernel.org/linus/9d573d19547b
However PCI is layered on top of (tunneled through) the Thunderbolt
switching fabric and that's where the real problem likely is here.
Maybe you can open a bug at bugzilla.kernel.org and attach full dmesg
and lspci -vvv output in the working case (device attachment at runtime)
and the non-working case (device attachment during system sleep).
Does the machine wake up if you attach devices during system sleep?
Are you suspending to ACPI S0ix, S3 or S4?
Thanks,
Lukas
--
Kenneth R. Crudup / Sr. SW Engineer, Scott County Consulting, Orange
County CA
