Conversation

Jarkko Sakkinen

Hmm... X1 Thinkpad does not have Wifi after waking up from suspend in Fedora 41. #fedora
4
1
0
i need to debug this when have time, it's just so frustrating that this same issue always comes up with any new laptop ;-)
1
0
1

@jarkko
Did you buy the X1 Carbon with certified Fedora installation? At least Gen 12 is currently available with Fedora - that's awesome since even if one would like to do own installation, there'd be a reference one with possibly custom packages to refer to in case of problems.

1
0
1
@timojyrinki i got it from company i formatted the hard drive without looking what is in ;-)
0
0
1

@jarkko

I had an HP laptop that would go to airplane mode. Never found a way to disable it. From but reports it was something built into the machine and couldn't be overridden by the os when resuming. I had to manually tell gnome to turn off airplane mode each time.

Oddly enough rotating the screen caused the airplane mode to turn on as well.

1
0
1
@shertson i'll root cause the bug and send bug report to fedora bugzilla :-)
1
0
0
@shertson If I had to do blind guess I'd guess ACPI tables but could be some other thing
0
0
0

@jarkko hmmm, I believe mine is gen9 and it works. Need to check the version tomorrow.

1
0
0
@ikkeT debugging wifi issues requires motivation because each test requires reboot ;-)
1
0
1

Jarkko Sakkinen

Edited 1 month ago
@ikkeT done https://bugzilla.redhat.com/show_bug.cgi?id=2323992

I'd probably fix this myself if I had extra bandwidth given how good oops transcripts this gave (despite somewhat alien subsystem). Happy to test the patch tho if/when it is available.

Unfortunately I'm busy compiling huge Rust project that is melting my laptop ATM ;-) Someone else gets the fun part...
1
1
1
@ikkeT @kernellogger BTW now is my ark kernel moment, or to try out the cool scripts you made once the patch is available.

Exactly kind of bug where you feel more confident to give tested-by when you see how the fix flies in distribution (alike) kernel.
2
0
2

@jarkko @kernellogger I'd expect this will get some eyes soon, as Lenovo is widely used internally at RH.

1
0
1
@ikkeT @kernellogger This is why I moved from Ubuntu to Fedora basically. Most of kernel developers I know use it, and most tools I use are also used by other kernel developers :-) For instance I recently reported an mbsync bug with a quick response time because it is quite popular among kernel developers.
1
0
1

@jarkko @kernellogger Someone mentioned it had broken around 6.11

0
0
0

@jarkko I had the same with my T14s Gen4 AMD. It has some non-intel wifi and I had to chase around for firmwares to get it to behave, but still refuses to work until I remove the driver module and insert it again.

2
0
1
@harald As a kernel developer I recommend instead disabling suspend and reporting the bug ;-) Usually kernel oops is a potential memory corruption, which can by definition have undefined collateral effects.

E.g. do "rtcwake -s 15 -m mem" and dump dmesg contents to the bug report.

When it comes to disabling suspend, it seems that making "lid close" not do anything is particularly hard these days so I made this short screencast of it :-)

I tried all combinations, and for the reasons I have no idea of, both upower and logind configuration need to be edited.
1
4
0

@harald

FWIW: I have a T14s G4 AMD since a few days and wifi and suspend-to-idle work fine out of the box for me afaics (01:00.0 Network controller [0280]: Qualcomm Technologies, Inc QCNFA765 Wireless Network Adapter [17cb:1103] (rev 01))

CC: @jarkko

3
0
1

@kernellogger @jarkko that's the one. I had issues like really lousy speeds after returning from suspend and such. Ubuntu 24.10.

1
0
0

@jarkko @kernellogger You are possibly blessed with and Intel Wifi-card then, which is nice.

0
0
0

@triskelion @harald @jarkko

FWIW, that is a T14, not a T14s (there iirc is not T14s G5 AMD)

1
0
1

@harald @jarkko

hmm, interesting; I've seen "lousy [wifi] speeds" two or three times, but reconnecting solved that.

0
0
0

@jarkko disabling suspend on a laptop is not ok for me.

1
0
1

@jarkko @ikkeT

+1

FWIW, changed quite a few things again today; fixes some bugs along the way and likely introduced a few new ones; if

curl --silent 'https://gitlab.com/knurd42/linux/-/raw/arkify-arkify/arkify' | bash

does not work at any time let me know.

1
0
0

@jarkko

BTW, is your bug maybe related to the one currently being fixed in stable: https://lore.kernel.org/all/20241106120301.338302382@linuxfoundation.org/

Only looked briefly, so I might be totally off there.

CC: @ikkeT

1
0
1
@kernellogger @ikkeT I'll look into this later on :-) I worked around the bug by disabling suspend on lid close for now. Just wanted to get dmesg dumped to a bug report fast...
1
0
0
@harald ah, but you could still take the log ;-) rtcwake spell that i put just put the machine to suspend for 15 seconds and wakes it up automatically.
0
0
0
@kernellogger @ikkeT i'll try this thanks. does it work with secure boot?
1
0
0

@jarkko @ikkeT

Not really, copr afaik does not allow signing those kernels in a way that would not expose the signing key to the world at the same time. :-/

It is signed with some "test" key iirc, but that one is untrusted.

0
0
0