gchristensen changed the topic of #nixos-chat to: NixOS but much less topical || https://logs.nix.samueldr.com/nixos-chat
drakonis has quit [Read error: Connection reset by peer]
drakonis has joined #nixos-chat
lassulus_ has joined #nixos-chat
lassulus has quit [Ping timeout: 252 seconds]
lassulus_ is now known as lassulus
tertl3 has joined #nixos-chat
jasongrossman has quit [Ping timeout: 252 seconds]
Lisanna has quit [Quit: Lisanna]
lopsided98 has quit [Quit: Disconnected]
emily has joined #nixos-chat
lopsided98 has joined #nixos-chat
<drakonis1> the acme-missile function makes me want to play sc2
jasongrossman has joined #nixos-chat
jasongrossman has quit [Ping timeout: 245 seconds]
thefloweringash has joined #nixos-chat
{`-`}_ has joined #nixos-chat
{`-`} has quit [Remote host closed the connection]
drakonis1 has quit [Quit: WeeChat 2.2]
jasongrossman has joined #nixos-chat
drakonis has quit [Remote host closed the connection]
<sphalerite> setuid programs vs daemons proxying access to a resource: opinions?
<andi-> I do not like setuid proccesses but I am not convinced the alternatives are better.. The tradeoff is also: some runtime issues for a short time vs system uptime. Services allow you to corrupt more memory, trigger issues.. etc... I guess it depends?!
<andi-> If you do not have knowledge on how to write secure setuid binaries: write a service instead.
sir_guy_carleton has quit [Quit: WeeChat 2.2]
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
<sphalerite> andi-: I'm not sure how services allow you to corrupt more memory?
<andi-> well they have a lifetime that is multiple "invocations"
<andi-> so they are a nice target for say spraying stuff around the heap and then later on executing an exploit that jumps to/executes some of the code you put there earlier..
<andi-> with setuid you always start from a "clean" memory view. No state from previous runs.
<sphalerite> aaah right
<andi-> (you can also fork() befor handling the request in your service...)
<andi-> it really depends on what kind of thing you are trying to protect
jasongrossman has joined #nixos-chat
sir_guy_carleton has joined #nixos-chat
__monty__ has joined #nixos-chat
jasongrossman has quit [Quit: ERC (IRC client for Emacs 26.1)]
sir_guy_carleton has quit [Quit: WeeChat 2.2]
sir_guy_carleton has joined #nixos-chat
jasongrossman has joined #nixos-chat
obadz has quit [Ping timeout: 268 seconds]
<srhb> Hms.. More new laptop questions. I see my frequency capping at a very low level despite not maxing temperature limit of 97C. Is that power throttling?
<srhb> (I'm new to this area)
obadz has joined #nixos-chat
emily has quit [Ping timeout: 252 seconds]
emily has joined #nixos-chat
<gchristensen> srhb: do you see anything in dmesg about throttling?
<gchristensen> srhb: your /etc/nixos/hardware-configuration.nix's cpuFreqGovernor might be set to powersave or something
<srhb> gchristensen: I don't see anything in dmesg. It is set to powersave, but I thought that it would still turbo boost whenever possible.
<srhb> Weird..
<srhb> Yeah, changing to performance eg doesn't change anything. Really weird.
<srhb> I think it must be some sort of power throttle. Q_Q
<srhb> Or maybe there's some other temperature limit I'm hitting somewhere.
<sphalerite> srhb: how low is very low?
<sphalerite> srhb: my laptop can go to 3.8GHz but will never sustain more than 2.8 or something
<sphalerite> I believe it does run at 3.8 sometimes, but never long enough for sampling to see it
<sphalerite> basically, if you want full speed get a desktop :D
<sphalerite> (or server)
lopsided98 has quit [Ping timeout: 276 seconds]
lopsided98 has joined #nixos-chat
<srhb> sphalerite: Right, of course. I expect throttling, but I expect throttling to happen for a reason. Right now I can't _see_ why it throttles. For instance, I can get a single core to 3.9 GHz (max boost) but if I stress everything, I get 3.5 across all cores. But temperature is fine, so I'm simply unsure what's causing the throttling to happen.
<gchristensen> oh I don't think most CPUs can hit maxboost on all cores
<gchristensen> what model is your CPU?
<srhb> I don't think it can either :)
<srhb> It is the max single core boost after all..
<srhb> i7-7820HQ
<srhb> I'm just trying to understand the 3.5 GHz cap.
sir_guy_carleton has quit [Quit: WeeChat 2.2]
drakonis has joined #nixos-chat
<emily> srhb: is this on a laptop?
<emily> oh
* emily suspects not, but just in case links https://github.com/erpalma/lenovo-throttling-fix in case it's that problem
<srhb> emily: It is, thanks. :)
<srhb> emily: Sounds very related, I am on a p51.
<emily> hmm, I'd have kind of hoped the thermals on the bigger laptops would be sufficiently better for it to not be as much of a problem :<
<emily> FWIW, though, when experiencing this problem I get dmesg entries about throttling
<srhb> emily: Thermals are excellent. It seems likely that somehow the trip point is synthetically (and erroneously) lowered
<emily> also, you're never going to get the Turbo Boost speed on all cores
<emily> not sure if that's what you're saying, but
<emily> re "I can get a single core to 3.9 GHz (max boost) but if I stress everything, I get 3.5 across all cores."
<srhb> Right. :) I'm trying to understand "why 3.5 specifically"
<srhb> Ie. what limit am I hitting.
* emily nods
<srhb> That I can't reach full 3.9 is just fine. And 3.5 might be just fine as well. I'm really curious as to _what_ limit I'm hitting though, because I can't find it :D
<emily> https://en.wikipedia.org/wiki/Intel_Turbo_Boost suggests it's based on Complicated Factors
<srhb> Like, full 8 core mprime load maxes me out at +65C, so temperature sounds unlikely
<srhb> Thanks!
<emily> and probably happens at too low a level for you to observe that much about it
<srhb> Yeah, it might be a fool's errand.. :-P
<emily> doing the undervolting thing might help you get better results
<emily> it's too scary for me to have tried it yet
<emily> I've just made my peace with laptops being kind of slow to some degree
<srhb> Already undervolted. Doesn't seem to make much of a difference.
<srhb> (I mean, aside from even better thermals...)
<srhb> This thing is fantastically fast compared to what I came from, so I'm very happy in general.
<Ralith> srhb: intel's windows-based tuning tools let you set max freq per number of active cores
<srhb> Ralith: Really! So the information is there, somewhere...
<Ralith> power throttling is possible too, ofc
lassulus has quit [Ping timeout: 252 seconds]
lassulus has joined #nixos-chat
<andi-> I almost forgot how slow rotating disks are... Migrating a service from one cluster to another...
ekleog has quit [Quit: brb]
ekleog has joined #nixos-chat
<sphalerite> infinisil: did you come along to the hackerspace at nixcon last year?
<infinisil> sphalerite: I only stayed a bit on the evening of the last conference day, not the hackday
<sphalerite> I can't remember which day it was that we went there
<infinisil> Why you ask?
<sphalerite> have you heard of 34c3?
<infinisil> Nope
<sphalerite> because it's the Munich CCC, uh…
<sphalerite> idk what to call it
<infinisil> Yeah no idea about that
<sphalerite> *c3 is the Chaos Communications Congress (iirc), a big conference of the (mostly German-language I think?) hacker scene
<sphalerite> anyway, it's good stuff :)
<infinisil> I see
<emily> C3 is well-known in my decidedly non-German circles, at least
<joepie91> yeah, def not just german
<joepie91> definitely come visit if you can get a ticket :P
* emily might be at C3!
<emily> (but is also shy)
<andi-> I am not going this year... after many years.. A bit burned out on that front..
<joepie91> emily: possibly C3 needs an introvert assembly again :P
<joepie91> hijacked some other absent group's table a few years ago to produce an ad-hoc introvert assembly, and contrary to what you might expect it actually worked really well!
<emily> haha
<andi-> I also think for an intovert it is a good atmosphere.. I usually hate many people and there I feel pretty at home.
<emily> I mean, mostly only considering going because of a bunch of people I know from other channels going (or getting tickets wouldn't feel worth it anyway)
<joepie91> it more or less worked out to "a place where you can sit and interact with others, and if all of a sudden you unannouncedly drop out of the conversation and start doing your own thing then that's fine too"
<andi-> if you run into issues with tickets I might be able to help out...
<joepie91> IRCesque I guess, 'lurking is fine'
<joepie91> anyhow, are there plans for a NixOS assembly at C3, or?
__monty__ has quit [Quit: Off to bed.]
<jasongrossman> joepie91: What a good idea.
<jasongrossman> joepie91: I mean the introvert table.