chomwitt | hi. looking for some desktop neutral power management interfaces i found uswsusp . But it's not included in Chimaera. | 11:26 |
---|---|---|
chomwitt | any similar alternatives ? | 11:26 |
chomwitt | i've tried also pm-utils but looks unmaintained for many years | 11:27 |
chomwitt | xmm it seems that uswsusp is also stale | 11:40 |
fsmithred | chomwitt, I think there's been no activity on pm-utils because it works. | 12:39 |
chomwitt | ok fsmithred . i just had the impression that standby and hibernation are operations that with new hardware should need updates | 13:53 |
fsmithred | I'm not on new hardware, but I haven't heard any complaints about that. It's easy enough to try it and see if it works. | 13:55 |
chomwitt | ok. i ll try pm-utils | 14:08 |
chomwitt | i wonder though why it's not listed in related arch page both pm-utils and uswsusp | 14:14 |
chomwitt | https://wiki.archlinux.org/title/Power_management | 14:14 |
chomwitt | i found a reference to uswspc in another arch wiki page as a lower power management interface | 14:17 |
chomwitt | and interestingly in a debian handbook power management is presented as problematic : https://debian-handbook.info/browse/squeeze/sect.power-management.html | 14:21 |
chomwitt | i install pm-utils in my old laptop for starters and it seem to work | 14:22 |
chomwitt | s/install/installed | 14:23 |
chomwitt | in my desktop pm-suspend worked ok. pm-hibernate no. | 15:31 |
chomwitt | maybe in a desktop pm-hibernate is not needed but anyway it doest work. | 15:33 |
chomwitt | s/doest/does not | 15:33 |
xrogaan | my mail doesn't seem to get through the mailing list. | 15:48 |
golinux | xrogaan: pm about the ml issue | 18:35 |
rwp | chomwitt, For hibernate to work one needs compatible hardware+drivers which are most often the problem. Needs enough swap to hold all memory. Needs that set in /etc/initramfs-tools/conf.d/resume eg RESUME=/dev/sda6 | 21:15 |
rwp | If all of those things converge okay then hibernate to disk and resume from it will work. | 21:15 |
rwp | Oh and make sure if the resume location is modified that the initramfs is rebuilt for it. | 21:16 |
gnarface | also if you have one of those bioses that has a setting for "memory hole at 15k" or whatever, it needs to be in some specific state (ON, i think, last i checked) | 21:17 |
gnarface | (maybe it's OFF actually, i am not sure) | 21:17 |
chomwitt | thanks rwp and gnarface | 21:25 |
chomwitt | i dont understand how acpid and pm-tools are related. There is not stated conflict so i guess acpid is user for automatic event triggered suspend and pm-tools for manual ? | 21:28 |
chomwitt | s/user/used | 21:29 |
rwp | acpid and pm-tools are not related. acpid handles ACPI events. pm-tools interacts with Pavel Machek's power management system. | 21:29 |
rwp | Though an acpi rule might trigger a suspend and then cross over into using pm-tools. | 21:30 |
chomwitt | rwp, i have 64 GB RAM so i would need 64GB swap for hibernation to succeed? | 21:30 |
rwp | When the power button is pressed it generates an ACPI event. That event is handled by the acpid. | 21:30 |
rwp | Regarding 64GB of RAM, WOOT! But yes I think that means you need at least 64GB of swap. | 21:31 |
rwp | There are those who will advocate for the compressed ram driver but I think that is playing with fire since it is data dependent if it can compress it enough to fit a smaller size or not. | 21:31 |
chomwitt | so acpi can trigger power management command but is a more general facility | 21:31 |
rwp | Also I don't think it can be a file system swap file, but not sure, that's a Don't Know for sure from me. But I think it would be riskier than regular swap. | 21:32 |
rwp | I think if you are short of swap, but wanted to test that everything else was working, that one could boot with mem=2g or whatever the syntax is and limit memory of that boot and then try it with the limited memory being active small enough to fix the current swap. I think. | 21:33 |
rwp | Most often a kernel driver is reported as being buggy and unable to resume from suspend. Problems fighting kernel driver bugs and hibernate are legion. | 21:34 |
gnarface | i do also remember a long ago discussion about defragging system ram after purging all caches to make the required hiberation/swap partition something less than the entirety of physical ram but i don't know if that ever actually worked and do recall it required a static and well understood list of running programs | 21:36 |
gnarface | theoretically it might be possible, but the failing case is indistinguishable from all the others | 21:36 |
gnarface | kindof a problem with hibernation in general | 21:36 |
gnarface | but if you could get your entire ram footprint down to for example, 2gb of that 64gb, and make sure it was all at the FRONT of the addressable space, you could in theory get away with only 2gb of swap for hibernate | 21:37 |
gnarface | i feel pity for you though if it really came to having to try that | 21:38 |
gnarface | if i had a machine with 64gb of physical ram i'd probably just leave it on | 21:38 |
gnarface | oh yea, and on the drivers front - if you're using a nvidia card i can save you some time: just give up on this now | 22:10 |
gnarface | they don't even come out of sleep right, let alone hibernate | 22:11 |
gnarface | but they'll sorta appear to be mostly fine, superficially, which well end up being somehow worse | 22:11 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!