03:05 <leezu> One hypothesis is that the hardware does not properly support mem_sleep=deep and Linux should default to mem_sleep=s2idle on this platform (which works). Can the mem_sleep default be specified via device tree files? Or how are platform specific defaults implemented?

04:21 <steev> leezu: i don't believe it can be specified in dts files, but that's just a matter of editing the /etc/systemd/system/sleep.conf (or somewhere thereabouts) file typically

04:21 <steev> wait, maybe

04:22 <steev> ah, nope, i was misreading my grep of the dts files :(

04:24 <steev> i haven't actually looked into the pm_test stuff

04:30 <steev> bamse: looks like i'm re-installing on the flex 5g.... grub update screwed up somehow and now i can't even load grub modules

12:14 <leezu> steev: systemd doesn't support configuring mem_sleep https://github.com/systemd/systemd/issues/13451 but systemd.tmpfiles can be used to write s2idle to /sys/power/mem_sleep upon every boot. In the coming days I'll take a look at what ChromeOS does on this platform. If they also use s2idle, it would be great to somehow make that the platform default..

15:17 <steev> iirc, cros uses upstart, not systemd

16:04 <steev> leezu: fwiw, my c630 already has s2idle in /sys/power/mem_sleep here

16:22 <macc24> is it possible to brick a sdm845 machine if i accidentally wipe its emmc?

16:40 <leezu> steev: I've installed Debian on the device https://github.com/hexdump0815/imagebuilder/issues/47 For your c630, is it with cros or custom linux? I wonder if the s2idle default on c630 is from the kernel or set by userspace

16:41 <steev> it's linux (debian based) installed via https://github.com/aarch64-laptops/debian-cdimage (i think is the repo with the releases)

16:42 <steev> according to the kernel documentation, s2idle is always present, shallow and deep only show up if supported

16:50 <leezu> So the sc7180-trogdor-lazor-r3 may incorrectly claim deep sleep support, or there's a bug in recent (?) kernels that broke the deep sleep support? I'll take a look what uses when running on sc7180-trogdor-lazor-r3

16:51 <leezu> macc24: is it a chromebook?

16:51 <leezu> * I'll take a look what CrOS uses when running on sc7180-trogdor-lazor-r3

16:54 <steev> leezu: let me rebuild 5.15 for its latest and double check there, i only have 5.18 and 5.19 rc2 installed currently

17:00 <steev> bamse: also, the ETA of my x13s is Tuesday, 21/6 so.... get those patches sent ;)

17:02 <robclark> leezu: sc7180-trogdor-* are chromebooks.. AFAIU we just use s2idle but that is something sboyd (who doesn't seem to be in this channel) or dianders know more about

17:02 <steev> sboyd is in linux-msm (i just pinged him a little bit ago with a question)

17:09 <macc24> leezu: nope

17:09 <macc24> it's a handheld gaming console running android

17:37 <steev> the answer to that is, maybe, but unlikely

17:43 <steev> leezu: 5.15 also shows only s2idle on the c630

17:59 <leezu> steev: Both 5.15 and 5.18 show deep and default to deep on sc7180-trogdor-*. "AFAIU we just use s2idle" Do you mean aarch64-laptops sets s2idle in userspace for all chromebooks?

18:00 <steev> leezu: no, afaik, the aarch64-laptops project doesn't do any configuration of sleep modes in the packages

18:04 <steev> also the name is kinda a misgnomer-ish, it was more about running linux on WoA devices, which just happened to be aarch64 laptops

20:43 <steev> bamse: so, turns out, the way you get the recovery media is that you have to go to the lenovo site ON the flex 5g... luckily, i'm still able to boot windows by selecting it from the uefi menu, so i'm downloading it now (they also don't let you download it to removable media because.... reasons? - not even sure what that's about considering you can just copy the files across once they're downloaded...) but sheesh

20:49 <qzed> steev: is that some Gualcomm-related shenanigans? for pretty much any Surface device you can go to the MS website and download a full driver pack (usually fairly recent drivers)... only for the Pro X you have to enter the serial number and you only get a recovery image with older drivers, not a pack with the latest drivers

20:49 <steev> qzed: i think it's lenovo specific - it's also an image of windows (Win10Prox64ARM for 82AK) with the lenovo drivers added

20:50 <steev> the problem i was having was, the latest w11 update broke grub, then grub updated and broke it further

20:50 <qzed> ouch

20:51 <steev> it used to be i could "set prefix=(hd5,gpt5)/boot/grub; insmod normal; normal" and then boot just fine, but now the insmod fails because of missing symbols

20:51 <qzed> sounds fun

20:51 <steev> so, i think the first thing i'll do when the x13s arrives is... download the recovery media

20:52 <steev> i was definitely lucky that i left windows around "just in case"

20:52 <qzed> heh yeah, that sounds like a good idea

20:52 <steev> because you can't download the recovery media on another pc, apparently

20:53 <qzed> kinda sounds like no one at lenovo ever though of "what is a recovery media supposed to do"

20:53 <qzed> you could probably spoof that somehow? but yeah... not sure what they're thinking

20:56 <steev> actually, now that i've gotten the recovery media "ordered" it does show up on this computer if i run their recovery software

21:28 <hexdump0815> macc24: didn't you have deep sleep somehow working on your lazor at some point?

21:28 <hexdump0815> leezu: i right now have no access to a trogdor device and thus cannot check

21:41 <hexdump0815> macc24: i mean this one - https://oftc.irclog.whitequark.org/aarch64-laptops/2022-03-25#30771427

23:29 <steev> qzed: i wonder if it's foxconn? i'm doing the restore now, and one of the "minimized" windows is the arm driver install script, and the "author" is listed as Foxconn

23:29 <steev> maybe because foxconn did all the actual work, they don't know how to do it "properly"

23:32 <qzed> maybe, not sure if MS relies on them though... haven't seen anything foxconn related on the spx yet

23:34 <steev> it'd be part of the restore process when it installs the drivers - at least, that's where i saw it, though you'd have to un-minimize the cmd windows as they pop up, to see it

23:35 <qzed> I think MS has that pretty streamlined and hidden away behind their windows update screens. never saw any minimized windows

23:36 <steev> ah, MS must have a different restore process, this seems to just be WinRE and WIM files

23:37 <qzed> I think MS did most (of not all) of the SPX work themselves... as far as I can tell they outsourced the Surface Go line and that does show...

23:38 <steev> alas, the few 7z files i can see are all password protected

23:39 <steev> the rest are CRI/IMZ files

23:42 <qzed> yeah that sounds quite a bit different than what I can see in the SPX recovery image, that more or less looks like someone copied a windows install and zipped it up in swm files with some config files for partitioning (and other stuff I don't know) and then zipped that all up again into a .zip

23:42 <qzed> there's a boot.wim but that just looks like a recovery disk or something like that

23:46 <steev> probably the usual of, someone did it once, and it works, so don't do anything else

23:54 <steev> dang, i had to tell it no to windows 11 like 3 times during setup