~mil/sxmo-tickets#442: 
Crust suspend is broken

Since upgrading to SXMO 1.6 (PMOS 21.12, Wayland), I have been able to enter Crust mode. This is both after entering the screen lock through the power button, or by manually invoking sxmo_screenlock.sh crust. Invoking the latter from terminal result in error message rtcwake: write error. Additionally, the script prints out an additional error crontab: can't open 'user': No such file or directory, but that's predictably fixed with a simple touch /etc/crontab/user (but this file should probably exist in the default installation).

Status
RESOLVED FIXED
Submitter
Lorenzo Martinico
Assigned to
No-one
Submitted
5 months ago
Updated
2 months ago
Labels
No labels applied.

Lorenzo Martinico 5 months ago · edit

For additional context, here is the output of dmesg when the lockscreen script is invoked:

[ 3560.243945] PM: suspend entry (deep)
[ 3560.244097] Filesystems sync: 0.000 seconds
[ 3560.244687] Freezing user space processes ... (elapsed 0.002 seconds) done.
[ 3560.247042] OOM killer disabled.
[ 3560.247050] Freezing remaining freezable tasks ... (elapsed 0.001 seconds) done.
[ 3560.248483] printk: Suspending console(s) (use no_console_suspend to debug)
[ 3560.250033] musb-sunxi 1c19000.usb: Error unknown readb offset 112
[ 3560.399220] sunxi-mmc 1c0f000.mmc: data error, sending stop command
[ 3560.399246] sunxi-mmc 1c0f000.mmc: send stop command failed
[ 3560.399263] mmc0: error -110 reading status reg of PM func
[ 3560.399277] PM: dpm_run_callback(): mmc_bus_suspend+0x0/0x80 returns -110
[ 3560.399300] mmcblk mmc0:aaaa: PM: failed to suspend async: error -110
[ 3560.416101] PM: Some devices failed to suspend, or early wake event detected
[ 3560.470897] musb-sunxi 1c19000.usb: Error unknown writeb offset 112
[ 3560.551379] OOM killer enabled.
[ 3560.551393] Restarting tasks ... done.
[ 3560.555295] PM: suspend exit
[ 3560.581838] sunxi-mmc 1c0f000.mmc: data error, sending stop command

On 01/04/2022 5:47 PM Lorenzo Martinico outgoing@sr.ht wrote:

Since upgrading to SXMO 1.6 (PMOS 21.12, Wayland), I have been able to enter Crust mode. This is both after entering the screen lock through the power button, or by manually invoking sxmo_screenlock.sh crust. Invoking the latter from terminal result in error message rtcwake: write error. Additionally, the script prints out an additional error crontab: can't open 'user': No such file or directory, but that's predictably fixed with a simple touch /etc/crontab/user (but this file should probably exist in the default installation).

-- View on the web: https://todo.sr.ht/~mil/sxmo-tickets/442

~earboxer referenced this from #478 3 months ago

~earboxer 3 months ago

I've also been getting rtcwake: write error recently when I try to run rtcwake -m mem -s 60. (PinePhone 1.2 on linux 5.17.0-rc8). It's sets the alarm, but fails to suspend. -m disk and -m freeze similarly fail.

If I use -m off, then it actually works, except for the part about waking itself up.

Note: I've never had a pleasant crust experience on this device (lengthy reboot required, never have had it "resume" from suspend using anything but the power button), so I'm not really sure how it's supposed to work.

~alpabrz 2 months ago*

I had the same with 5.17 kernel, but not sure whether my issue is related. My dmesg spammed something along the lines of musb_bus_suspend [some-number]: trying to suspend as a_wait_bcon while active. There is a comparable issue from mobian and also a similar problem described in their wiki, but for the PineTab. Seemingly something going wrong with USB OTG in the kernel. Recompiled kernel with USB_CONFIG_GADGET=m (it is builtin in pmOS kernel) and then removing the module usb_f_rndis from the initrd. USB networking seems to be gone now but at least I can suspend again^^ Don't know anything about this stuff though and how it works, maybe I'm on the wrong track here.

BTW, suspend for me was also always rather finicky. My cron jobs often fail to wake up the device :( Therefore I stopped using them for wakeup purposes.

~alpabrz 2 months ago

hm just found your upstream issue ~earboxer https://gitlab.com/postmarketOS/pmaports/-/issues/1478 maybe wise to link it here for reference.

~arjunaithal 2 months ago

The crust seems to be working fine for me. Running 5.17 kernel on sxmo head branch.

~phartman REPORTED FIXED 2 months ago

Register here or Log in to comment, or comment via email.