Open ydzhou opened 1 year ago
dmesg for suspend/resume
[ 338.400900] xhci_hcd 0000:08:00.0: Controller not ready at resume -19
[ 338.400963] xhci_hcd 0000:08:00.0: PCI post-resume error -19!
[ 338.400978] xhci_hcd 0000:08:00.0: HC died; cleaning up
[ 338.400989] xhci_hcd 0000:08:00.0: PM: dpm_run_callback(): pci_pm_resume+0x0/0xf0 returns -19
[ 338.400994] xhci_hcd 0000:08:00.0: PM: failed to resume: error -19
Cc @Redecorating
what mac model is this? has it ever worked in the past (if so, do you know when or what kernel versions)? can you reproduce the issue on the official, not-t2, ubuntu kernel?
also, is normal, not s2idle suspend fine? or maybe if you don't load the xhci_hcd or don't plug in any usb devices, see if it works then?
The mac is 2019 Macbook Air. This is my first time to try t2linux, so it never work in the past. Yes, I can reproduce in non t2 ubuntu kernel. Yes, I also tried to not plugin any usb thing, still running into the this issue. xhci_hcd is a built-in module. I cannot unload.... Yeah s2idle does not work, but deep mode suspend works.
BTW, are you able to reproduce s2idle issue on your side?
My kernel is 6.2 (latest t2linux ubuntu kernel)
The mac is 2019 Macbook Air.
This is my first time to try t2linux, so it never work in the past.
Yes, I can reproduce in non t2 ubuntu kernel.
you might be able to file an upstream bug with ubuntu then.
Yes, I also tried to not plugin any usb thing, still running into the this issue.
xhci_hcd is a built-in module. I cannot unload....
ah right
Yeah s2idle does not work, but deep mode suspend works.
BTW, are you able to reproduce s2idle issue on your side?
I haven't tested that, I'm not sure when I'll get time.
Can you do me a favor and try a simple test? I wanna verify if this is my device issue or this is upstream kernel bug...
echo 's2idle' > /sys/power/mem_sleep
systemctl suspend
Then wake up
dmesg | grep xhci
Hi there! I have 2020 MBA updated to sonoma. Installed Ubuntu 6.9.8-3-t2-jammy and experienced issue with not working suspend/hibernate. I know about same issue on other distros like debian/Arch.
Recently find out it appeared new workaround based on service which shut down brcmfmac/apple-bce modules before suspend and start them after resume. I tried that with no luck (see https://github.com/t2linux/T2-Debian-and-Ubuntu-Kernel/issues/53#issuecomment-2221615015)
Is there exists some workaround for ubuntu or maybe the issue will be addressed in future? AdityaGarg8 said that he is not working for ubuntu.
ps. here is my last attempt to run deepSleep (https://github.com/t2linux/T2-Debian-and-Ubuntu-Kernel/issues/53#issuecomment-2170545969)
Hi there! I have 2020 MBA updated to sonoma. Installed Ubuntu 6.9.8-3-t2-jammy and experienced issue with not working suspend/hibernate. I know about same issue on other distros like debian/Arch.
Recently find out it appeared new workaround based on service which shut down brcmfmac/apple-bce modules before suspend and start them after resume. I tried that with no luck (see #53 (comment))
Is there exists some workaround for ubuntu or maybe the issue will be addressed in future? AdityaGarg8 said that he is not working for ubuntu.
ps. here is my last attempt to run deepSleep (#53 (comment))
I'm test compiling the Xanmod kernel. Let's see if it works there.
Good news. Suspend worked on Xanmod kernels with the workaround!
Good news. Suspend worked on Xanmod kernels with the workaround!
AWESOME! Thanx! So excited to update)
Good news. Suspend worked on Xanmod kernels with the workaround!
I've released the Xanmod kernels, but suspend is still not that stable as it should be. You can now try them
Good news. Suspend worked on Xanmod kernels with the workaround!
I've released the Xanmod kernels, but suspend is still not that stable as it should be. You can now try them
AdityaGarg8 thank you!
Updated my ubuntu to 6.9.8-4
but got same behavior - system looks like going to sleep but unable to wakeup. I'll try to check my sleepmode later. Also I'm doubt about current suspend-fix-t2.service because with original ExecStart/ExecStop I got error after hit cmd
sudo systemctl enable suspend-fix-t2.service
Can you please share your workaround fix service?
Good news. Suspend worked on Xanmod kernels with the workaround!
I've released the Xanmod kernels, but suspend is still not that stable as it should be. You can now try them
AdityaGarg8 thank you! Updated my ubuntu to 6.9.8-4 but got same behavior - system looks like going to sleep but unable to wakeup. I'll try to check my sleepmode later. Also I'm doubt about current suspend-fix-t2.service because with original ExecStart/ExecStop I got error after hit cmd
sudo systemctl enable suspend-fix-t2.service
Can you please share your workaround fix service?
I've been following this:
https://wiki.t2linux.org/guides/postinstall/#suspend-workaround
And yes, I am also often encountering the same issue you told. I'm trying tweaking the kernel config a bit to make it more stable.
Also make sure you install xanmod
sudo apt update
sudo apt install linux-t2-xanmod
The service won't work on normal kernels.
got it, thanx I thought you swaped kernel in t2 tag
Also make sure you install xanmod
sudo apt update sudo apt install linux-t2-xanmod
Installed xanmod and checked suspending. System now suspending and resuming correctly with the workaround but I experienced with not worked touchpad and keyboard. Even after enabling fix-service keyboard and touchpad stop to work, noticed that backlit opposite start to work)
Hi @AdityaGarg8 Few days ago updated my ubuntu with xanmod and did some researches.
Here is uname of my MBA2020:
Linux mba 6.10.2-x64v3-t2-jammy-xanmod1 #xanmod1 SMP PREEMPT_DYNAMIC Mon Jul 29 02:07:55 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
Keyboard backlit always stop to work after reloading apple-bce. Here is journalctl after run
sudo service start suspend-fix-t2.service
backlit.off.log
From time to time the workaround is working, but sometimes is not. I noticed that it's always fail on resume if suspend-fix-t2.service just enabled. But if I run service with cmd from previous (from 1 in above) before suspend then KBD/WiFi will work after resume. Here is logs. suspend.fail.log suspend.ok.log suspend.ok.with.lid.log
All test was with uncommented rows with brcmfmac_wcc and brcmfmac in /etc/systemd/system/suspend-fix-t2.service
Should I create new issue in github.com/t2linux/T2-Ubuntu ?
Hi @AdityaGarg8 Few days ago updated my ubuntu with xanmod and did some researches.
- Here is uname of my MBA2020:
Linux mba 6.10.2-x64v3-t2-jammy-xanmod1 #xanmod1 SMP PREEMPT_DYNAMIC Mon Jul 29 02:07:55 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
- Keyboard backlit always stop to work after reloading apple-bce. Here is journalctl after run
sudo service start suspend-fix-t2.service
backlit.off.log- From time to time the workaround is working, but sometimes is not. I noticed that it's always fail on resume if suspend-fix-t2.service just enabled. But if I run service with cmd from previous Update build.sh #1 before suspend then KBD/WiFi will work after resume. Here is logs. suspend.fail.log suspend.ok.log suspend.ok.with.lid.log
All test was with uncommented rows with brcmfmac_wcc and brcmfmac in /etc/systemd/system/suspend-fix-t2.service
Should I create new issue in github.com/t2linux/T2-Ubuntu ?
Someone got suspend working by adding acpi_osi=!Darwin
to the kernel parameters and removed this script.
Hi @AdityaGarg8 Few days ago updated my ubuntu with xanmod and did some researches.
- Here is uname of my MBA2020:
Linux mba 6.10.2-x64v3-t2-jammy-xanmod1 #xanmod1 SMP PREEMPT_DYNAMIC Mon Jul 29 02:07:55 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux
- Keyboard backlit always stop to work after reloading apple-bce. Here is journalctl after run
sudo service start suspend-fix-t2.service
backlit.off.log- From time to time the workaround is working, but sometimes is not. I noticed that it's always fail on resume if suspend-fix-t2.service just enabled. But if I run service with cmd from previous Update build.sh #1 before suspend then KBD/WiFi will work after resume. Here is logs. suspend.fail.log suspend.ok.log suspend.ok.with.lid.log
All test was with uncommented rows with brcmfmac_wcc and brcmfmac in /etc/systemd/system/suspend-fix-t2.service Should I create new issue in github.com/t2linux/T2-Ubuntu ?
Someone got suspend working by adding
acpi_osi=!Darwin
to the kernel parameters and removed this script.
acpi_osi=!Darwin in grub?
Yes grub
Yes grub
I just checked it.
GRUB_CMDLINE_LINUX_DEFAULT="acpi_osi=!Darwin quiet splash"
sudo update-grub
reboot
also disabled suspend-fix-t2.service
aaaaand got system resume after few seconds noticed same error in log which I saw while testing the fail attempts with the workaround:
Jul 31 20:40:14 mba gnome-shell[1312]: JS ERROR: Failed to initialize fprintd service: Gio.IOErrorEnum:
GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available
asyncCallback@resource:///org/gnome/gjs/modules/core/overrides/Gio.js:114:23
Jul 31 20:40:14 mba systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully.
here is log: try.to.sleep.log
then changed kernel to Linux mba 6.6.43-1-t2-jammy
and got same result
log:
try.to.sleep.t2.lts.log
I had some success when testing the patch and now i believe it can be made to work somehow really hate mocos
Yes grub
I just checked it.
GRUB_CMDLINE_LINUX_DEFAULT="acpi_osi=!Darwin quiet splash"
sudo update-grub
reboot
also disabled suspend-fix-t2.service
aaaaand got system resume after few seconds noticed same error in log which I saw while testing the fail attempts with the workaround:
Jul 31 20:40:14 mba gnome-shell[1312]: JS ERROR: Failed to initialize fprintd service: Gio.IOErrorEnum: GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available asyncCallback@resource:///org/gnome/gjs/modules/core/overrides/Gio.js:114:23 Jul 31 20:40:14 mba systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully.
here is log: try.to.sleep.log
then changed kernel to
Linux mba 6.6.43-1-t2-jammy
and got same result log: try.to.sleep.t2.lts.logI had some success when testing the patch and now i believe it can be made to work somehow really hate mocos
Nice to know you got it working. I'll fiddle with this when I get time.
Yes grub
I just checked it.
GRUB_CMDLINE_LINUX_DEFAULT="acpi_osi=!Darwin quiet splash"
sudo update-grub
reboot
also disabled suspend-fix-t2.service aaaaand got system resume after few seconds noticed same error in log which I saw while testing the fail attempts with the workaround:Jul 31 20:40:14 mba gnome-shell[1312]: JS ERROR: Failed to initialize fprintd service: Gio.IOErrorEnum: GDBus.Error:net.reactivated.Fprint.Error.NoSuchDevice: No devices available asyncCallback@resource:///org/gnome/gjs/modules/core/overrides/Gio.js:114:23 Jul 31 20:40:14 mba systemd[1]: NetworkManager-dispatcher.service: Deactivated successfully.
here is log: try.to.sleep.log then changed kernel to
Linux mba 6.6.43-1-t2-jammy
and got same result log: try.to.sleep.t2.lts.log I had some success when testing the patch and now i believe it can be made to work somehow really hate mocosNice to know you got it working. I'll fiddle with this when I get time.
yep, just for clarity, it's barely working - only with xanmod + workaround, without backlit and resume is not stabil but it gave me hope thanks
Hi
When I use s2idle sleep mode, I run into an issue with my usb controller and I think those errors prevent my computer to completely shutdown or to sleep again
And here is the message during shutdown, then my computer stuck there.
lsusb output
0000:08:00.0
is managing Bus 003 and 004