Arch Linux

glenb77

Member

Registered: 2014-10-02

Posts: 68

[Solved] resume problem with asus tp412ua with nvme drive

just got above computer. having strange problem with resuming from suspend. the first time it suspends it resumes fine. the second time it suspends, it does come up and appears normal but nothing will run and i eventually get an uptime reading in excess of 30. any command entered into terminal just does nothing. if i switch screens using ctl-alt-f(x) and try to login, it hangs but i will see several messages reporting hung tasks.

i added ram and swapped the ssd drive that came with the computer for a larger m.2 nvme made by mushkin. so far i have tried acpiphp.disable=1 and nvme_core.default_ps_max_latency_us=0 as kernel params and nothing changed. i don't know to determine for certain that the nvme is the problem.

i use startx (xinit) as opposed to a gdm. i tried enabling the gnome gdm with same result. also same issue whether or not in X or wayland, happens from terminal the same way as in gnome.

i have a dual boot with win10 and that suspends and resumes fine.

computer is intel i3 with 620 graphics. does anyone have any ideas as to what to try next ? . thanks

edit: i believe the issue is the nvme drive; in sys/class/nvme/nvme0 there is reset_controller, state, rescan_controller. after first sucessful suspend/resume. state shows active. after second attempt after resume, state shows 'connecting' and issuing 'echo 1 > reset_controller returns device busy.

i can live with the issue, but would like to know any suggestions or if there is a way to exclude the drive from the sleep function and just keep it active during sleep.

edit: i opened a ticket with mushkin support on this and the rep sent me a link of the exact problem in a ms surface project for linux. for anyone else who runs into this the workaround i am using is:

in:    /sys/bus/pci/devices/xxxx.xxxx.xx.x (device for nvme controller). d3cold_allowed needs to be set to 0. i don't know if i used the right approach. but i made a simple one line script to set the parameter to 0 (echo 0 > /sys/bus/pci/devices/0000:02:00.0/d3cold_allowed) and run it from a service unit to set at bootup. it is working. i am not sure if there is a power consequence or not. but it now sleeps/suspends reliably every time.

Last edited by glenb77 (2019-06-05 00:40:57)

osxi

Member

Registered: 2014-08-30

Posts: 1

Re: [Solved] resume problem with asus tp412ua with nvme drive

Good to know! I’m about to pick one of these up today myself. Is that a PCIe 3.0 x4 NVMe drive by any chance?

Last edited by osxi (2019-06-09 16:30:36)

lolop

Member

Registered: 2015-02-19

Posts: 37

Re: [Solved] resume problem with asus tp412ua with nvme drive

I do not understand changed the setting on the periphery (d3cold_allowed) does not work at home. I have an intel ssd 760p 1Toon PCIe port 3.0 x4 (cm Asus 320M-k).
Do you have any other ideas?
Thank you