Bjornn- | mission accomplished | 00:41 |
---|---|---|
gnarface | anyone seen or heard of a fix for cpu core "soft-lockups" on suspend? might be relevant to AMD motherboards and NVidia video cards, particularly more recent ones like the 1060. | 06:17 |
gnarface | happening on the ascii backport kernel | 06:18 |
gnarface | (and backport nvidia drivers) | 06:18 |
gnarface | can not reliably be reproduced | 06:18 |
gnarface | but always happens on suspend (maybe 1/5 times or so by current statistics?) | 06:18 |
furrywolf | my ascii backport kernel box hangs on suspend about 1/3rd of the time, on an intel laptop with amd video. lol | 06:23 |
gnarface | hmmm. thanks for the info, furrywolf. it looks like beowulf is using about the same kernel version so i doubt upgrading is gonna fix that... | 06:42 |
gnarface | hmm, although maybe i could just try building something later than 5.x for it... | 06:43 |
furrywolf | suspend has always been one of the least-reliable things, yet also one of the hardest to troubleshoot... | 06:44 |
gnarface | i've been avoiding dealing with this for my own machines because i just leave them on but now i'm trying to help a windows refugee get full hardware support back | 06:44 |
gnarface | suspend seems to be a sticking point | 06:44 |
gnarface | furrywolf: have you looked into any bios-level mitigations, like disabling "c1e" power saving or stuff like that? | 06:46 |
gnarface | intel might call it something else | 06:46 |
furrywolf | not on this box. I'm not sure it even provides any configurability of such things. | 06:46 |
gnarface | there is a corresponding kernel config param for intel boards, i couldn't find the amd equivalent... | 06:47 |
gnarface | i think i saw someone mention setting intel_idle.max_cstate=7 or something like that (in conjunction with a bug that was supposed to have been fixed since 4.15 or 4.17) | 06:49 |
furrywolf | bbl, wolfy bedtime | 06:51 |
gnarface | google really thinks i am interested in ryzen-specific results about this, but it seems to be happening to everyone | 07:54 |
gnarface | (and this board is actually much older) | 07:54 |
gnarface | thanks for the beowulf mini live beta iso. very timely, very useful. | 13:14 |
msiism | fsmithred: Question on Refracta releases: Is there any particular reason the checksum file for the images has the ".txt" extension? I mean, this is not necessary, is it? | 13:39 |
msiism | I'm asking because I'm working on another project and was wondering how checksum files and signatures are best named and organized. | 13:40 |
fsmithred | msiism, that's so you can read it online in web browser | 13:43 |
msiism | Oh, good to know. | 13:44 |
msiism | fsmithred: Thanks. | 13:45 |
fsmithred | gnarface, maybe intel.pstate? | 13:45 |
fsmithred | I had to mess with that to get rid of some lockups | 13:46 |
fsmithred | msiism, the preferred method is to put the checksums in one file and then sign that file with a detached sig | 13:48 |
msiism | Yeah, I figured. | 13:50 |
msiism | With a reasonably modern 'sha256sum', you can then use --ignore-missing to make life easy for yourself. | 13:52 |
onefang | Security isn't meant to be convenient. | 13:52 |
fsmithred | what are you ignoring? | 13:52 |
msiism | fsmithred: Missing files. | 13:53 |
msiism | fsmithred: So you don't get a barrage of "this and that is missing", when you verify the checksums. | 13:54 |
fsmithred | oh, I always read them to compare | 13:54 |
msiism | I tend to use 'sha256sum -c --ignore-missing <checksum file>' | 13:54 |
fsmithred | yeah, that's easier. | 13:55 |
msiism | onefang: That's debatable. The question is, I think: Convenient for which party? Sure, security should make attacks inconvenient. | 13:56 |
onefang | I was joking around with the usual phrase "trade off between security and convenience". | 13:57 |
msiism | Ok, I missed the joke then… :) | 13:58 |
onefang | Knew I should have put a smiley after it. lol | 13:59 |
gnarface | fsmithred: yea that's what google told me too, but it's AMD | 13:59 |
msiism | onefang: I was doing some Lisp execises in parallel. So, my focus was kind of limited. | 14:00 |
* yeti looks twice... yay msiism! | 14:03 | |
yeti | _o/ | 14:03 |
msiism | yeti: Hello. :) | 14:03 |
msiism | yeti: I've had a look at your website recently. Turns out it's pretty stable. ;) | 14:05 |
yeti | I have a website? | 14:05 |
msiism | I think so, yes. | 14:05 |
yeti | nah... just some notes... boring stuff for nearly everyone | 14:06 |
onefang | The most stable web site is the one that doesn't exist. | 14:06 |
msiism | yeti: The colors make up for any such shortcomings. | 14:06 |
yeti | aaah that one | 14:07 |
yeti | that will fade away | 14:07 |
yeti | that was from my life before orgmode | 14:07 |
yeti | :-Þ | 14:07 |
msiism | I see. | 14:07 |
msiism | But it should be archived. | 14:07 |
yeti | even worse... | 14:08 |
yeti | https://web.archive.org/web/20160312183735/http://yeti.freeshell.org/ | 14:09 |
yeti | the beginings hat a fake CET font | 14:09 |
yeti | that was later thrown to gitlab | 14:12 |
msiism | I see. The gitlab version is an improvement. | 14:15 |
yeti | but it's the dark side too | 14:16 |
yeti | msgithub, googlegitlab | 14:16 |
yeti | :-( | 14:17 |
msiism | Yeah, luckly I know somone who runs a server with his own gitlab instance that I can use. | 14:18 |
msiism | Also (a bit off-topic, but still): The FSF is going to launch a code hosting platform, IIRC. | 14:18 |
yeti | I have gitea @home but still on my main notebook | 14:19 |
yeti | tat needs a permanent home on some armish helper | 14:19 |
yeti | someday | 14:19 |
yeti | then it at least will get a public onion address | 14:19 |
yeti | or more | 14:20 |
yeti | and the old pages will probably merge in an org-brain when it gets a stable exporter | 14:20 |
yeti | http://yeti.freeshell.org/orgy/fastspin/fastspin.html | 14:21 |
yeti | boring but useful... | 14:22 |
yeti | https://forums.parallax.com/discussion/comment/1479243/ | 14:22 |
yeti | motivational | 14:22 |
yeti | LP & RR are the way to go | 14:22 |
* msiism bookmarks that | 14:25 | |
yeti | That's only a short bait... | 14:25 |
yeti | didn't help un that microcontrollerfirum | 14:25 |
yeti | they want to stay mousepushers | 14:25 |
yeti | and ditch perfectionism | 14:26 |
yeti | glue your code together and if time is a problem, add doc later | 14:26 |
yeti | ifen if fairli imperfect, it is a seed that'll grow later | 14:27 |
yeti | http://yeti.freeshell.org/orgy/fastspin/fastspin.html some git fairly lon menawhile | 14:27 |
yeti | and yield feedback | 14:28 |
yeti | more than the old colourful stuff | 14:28 |
yeti | which even repelled some because of the colour | 14:28 |
* yeti soon gets bitten by gl... | 14:29 | |
yeti | määän... looking at all the typos... need a caffeine soup, | 14:31 |
yeti | https://yeti.tilde.institute/boc.html#org374e3e1 <<< spicing org stuff up with css is not prob | 14:32 |
yeti | https://github.com/fniessen/org-html-themes <<< a readthedoc like style | 14:33 |
yeti | org in git + that style might be a substitute for doc in a wiki | 14:34 |
yeti | I think life is easier with orgmode and while I prefer boring layouts now, it can do the opposite too | 14:35 |
* msiism has yet to try Emacs | 14:37 | |
yeti | org/babel is the interesting part and dired and tramp | 14:40 |
yeti | http://www.howardism.org/Technical/Emacs/literate-devops.html | 14:40 |
yeti | it keeps scrips, results and doc in one place... | 14:41 |
yeti | without it these parts tend to diverge | 14:41 |
yeti | lots of uglyness | 14:42 |
msiism | "howardism" – that's my kind of humor, obviously. | 14:42 |
yeti | someday emacs will be my desktop | 14:45 |
ballball | can anyone tell me what this means? https://defuse.ca/b/y3s53xif | 15:05 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!