looks like it is going to take another 3-4 hours before all the files are backed-up

i am going to let this continue while i'll go get somethings done. Se ya all later. Bye RT

o/

see ya daemon

What I've read many people in the FreeBSD community are against pulseaudio... (and some major ports have pulseaudio deactivated by default - chromium for example) How comes? What would be the native FreeBSD way of handling audio streams with multiple input-/output-devices?

And what UI application would I use for the native method? (I am currently using pavucontrol to manage it)

I think chromium uses stdio

I remember trying to it working with xrdp audio and found that out.

sndio?

sndio.

but it could also support pulseaudio... which would make some web-apps (like MS-Teams :puke ) be able to detect different mics, etc.

I don't think pulseaudio is actually needed for that

I mean, I have two mics, and with the default chromium build (pkg) I only see "default" as selectable

with pulseaudio I would be able to see both mics and would be able to select which to use

(well, I don't know if chromium can do it)

chromium's audio latency always seemed bad enough to me without adding pulseaudio on top

I don't have the link on hand, but I've read a blog where they described that chromium with pulseaudio would support that use case

:D

I wouldn't use chromium either if it wasn't for my customer who uses M$ Teams

There is a pulseaudio bridge/plugin for sndio too.

and M$ teams tells me that firefox on FreeBSD is not a supported platform

It didn't work for my case though... I think it's one a one way thing and don't know much about mic support with it.

But maybe look at that.

Nope.

don't know for Safari - I've got a mac-machine on hand, but there I'd only use firefox to work with (where surprisingly M$ teams works)

so, okay... sndio is native for FreeBSD... how would I manage my different audio streams that come from different applications? (in pulseaudio you can tell which application outputs to which output-device, etc.)

the only major difference between firefox on freebsd vs. on any other platform that I know of is the lack of some rights management stuff

sndio isn't native for freebsd

the reason I am asking is: I've started a ungoogled-chromium build with pulseaudio support yesterday and my machine is still working on it :D

*an

last chromium build I tried took 9 hours at -j4, iirc

on a 3.2GHz skylake

(4 real cores, no htt)

I have an i5-8350U CPU @ 1.70GHz - so I guess it will take days :D

so I am highly doubting that approach, because I would need to build it every other day when updates arrive

anyway, freebsd's actual native audio is what you see in /dev/dsp*, /dev/mixer*, /dev/sndstat etc.

RhodiumToad: thx for that info

which is basically OSS-like

would I be able to add filters, like noise-reduction, etc. on top of the native audio? how would I handle that natively in FreeBSD?

the thing is, I wouldn't use pulseaudio if it's possible to cover these use cases natively, I guess

there's an optional bass/treble control, but no arbitrary filters in the driver, that sort of thing has to be done in userspace

man sound for the details

it doesn't have the flexibility of pulseaudio, certainly

Been using freebsd for about 5 days it's pretty good

it has its good and bad points, like most things

I know hardware support is sometimes a problem I have it on my raspberry pi but it doesn't support the wifi

yeah. thing is that most manufacturers don't provide drivers, and often they don't even provide specs

been trying to do something probably impractical, getting x86 game servers running in a linux jail on arm using qemu, but it won't work on 32bit because arm freebsd doesnt have a 32bit compatibility module

hm? I thought it did, I don't have any arm64 hardware myself but I've seen people mentioning running 32-bit arm stuff on 64-bit

I could try a 32bit Linux jail

it might not be compiled in to the GENERIC kernel by default

oh, looks like it is

options COMPAT_FREEBSD32 # Compatible with FreeBSD/arm

it might not have userland libraries for 32-bit, though (hard to tell from the makefiles)

it'd still be possible to run a 32-bit freebsd jail under 64-bit, but you'd have to build or install an armv7 system image in the jail

I'll try a 32bit Linux jail and see if it works

alright... after 14 hours the ungoogled-chromium port with pulseaudio build finished :D damn...

looking forward to the next update builds XD

RhDoc what's your system config/setup?

sozuba, How are you doing with FreeBSD & it heating up your computer?

the OS is running on a Thinkpad T480 with an i5-8350U CPU @ 1.70GHz

RhDoc, How much RAM & swap is there?

parv: I went insane and installed 64GB RAM (and likewise for SWAP - tho I am not sure if this is needed)

RhDoc, Nice!😎

Ah okay. My E450 has an i5-4300U @ 2.900 GHz max with 8 GB of RAM. I guess mine might take double the time

(the RAM is more or less required for work, because of SW development and running VMs and containers)

does your T480 heat-up at all? in general i mean, not when compiling.

like when booting up

RhDoc cool

nope, not that much

and during runtime I am setting dev.hwpstate_intel.N.epp, depending on needs or AC plugged in or not, etc.

most builds don't need huge amounts of RAM, though there are exceptions (cough ghc cough, cough gcc cough)

no idea what that it, will make a note. I am new to this, still learning the basics

(somehow I didn't get powerd to work)

With "hwpstate_intel", "powerd" does not work IIRC. If "hwpstate_intel" disabled then that may

with "sysctl dev.hwpstate_intel.[0-7].epp=[0-100]" you can influence the CPU performance (where 0 ist max performance and 100 is max economic - for battery usage)

1 hr 30 mins before the last bits of my data are backed up, still 47 GB to go

hwpstate_intell and powerd, will remeber that

I remember RhodiumToad talked about his T80, yesterday, having issues with powerd

T480* I GUESS

that wasn't me

RhDoc, forum.opnsense.org/index.php?topic=26590.0

Title: [solved] With 22.1, is PowerD needed on modern (Coffee Lake+) Intel CPUs?

RhodiumToad oh sorry, i guess i could have mixed up with the flood of messages

it happens

:)

sozuba: well, I have mentioned it here somedays in the past

Could be you as well, cause both you and RT start with Rh :D

parv: yeah, I've read that one - but I couldn't see my CPU speed "shifting" - so I created scripts setting dev.hwpstate_intel.N.epp myself as needed :D

i mean the handl

I mean, I started using FreeBSD last week... I still need to dig through all the handbooks to get more familiar

RhDoc, ;-) You could try switching "power_profile" (0x0 or 0x1) "service" & set the "C" state for "economy" & "performance" profiles. In any case, I do also have a script to twiddle "*epp" MIB depending on if I am compiling or not

yeah 32bit linux cant run in freebsd arm64

says bad system call (segfault)

There is also a wiki page on power tuning wiki.freebsd.org/TuningPowerConsumption

Title: TuningPowerConsumption - FreeBSD Wiki

parv: thx for that input - will check it out

RhDoc last week? cool. As a daily driver?

sozuba: yeah, I am trying - so far it looks promising

Nice :)

the one thing I still need to setup is a linux VM in bhyve so I can redirect docker-cli calls to for SW development

besides that I think everything is working as needed - for now :D

RhDoc: i'm a fan of sysutils/vm-bhyve

good to know :)

makes things quite a bit easier. i host several linux and freebsd guests

To me "power_profile" service is rather obtuse. Who the fuck remembers to supply "0x0" for "economy" profile & "0x1" for "performance?! Not this one ... until I read the damn "rc" source umpteenth time.

rtprio: well, from what I've read so far I think I will stick to the native bhyve/bhyvectl - not sure what vm-bhyve will solve for me :D

but I'll have a look on that

... Oh! Should add local patch to accept "battery", "eco", "perf(ormance)?", & "ac" strings too

... along with "slow" & "fast"

parv: I've read that wiki page - I think I got the idea from there to set te .epp values for Intel Speed Shift CPUs

hyvoid: if you want to try 32-bit freebsd in a jail, you can grab the armv7 distribution files (base.txz and so on) and untar them in a jail directory, and see if that works

hyvoid: (you don't need the kernel)

RhDoc, ok

backup finally done, now about to format my laptop and install freebsd. I will follow your video as well for manual install, if that's feasible :) daemon

sozuba, Try perhaps in a VM first?

parv that's already done

after around ~6 months of tryting to read and try in a vm, i decided to install it about a month or more ago

unfortunately, my laptop got heated as soon as the booting process starts

i tried for a week and then gave up

but thought i will try to give it some more energy before settling back to linux and so, came here to ask for advice and help, people were kind enough to give me hope. Esp RT Rhodium Toad.

that's the gist :)

Alright. Good luck

thanks :)

Hello, I have an issue when mounting sshfs filesystems on FreeBSD 13.1 I get the following error:

mount_fusefs: /dev/fuse on /usr/home/diego/blog: Operation not permitted

svragv: are you running as root, if not do you have sysctl `vfs.usermount=1` and finally does the appropriate user have wx permissions to the dir?

I assume your ssh already works fine for the given user

Yes the ssh works fine, let me try the sysctl think

that worked, thanks :)

hey guys, the freebsd package mirrors contain an "/{ABI}}/latest/packagesite.pkg" and "/{ABI}}/latest/packagesite.txz" files, as well as meta.txz and meta.conf. these files are essentially the same in my experience. which ones are to be used?

because older pkg utility used txz, newer uses pkg extension

and what about meta.conf/meta.txz?

meta.txz basically contains the conf with a public key and signature

it seems that the newer pkg utility only uses the .conf, is the .txz to be phased out over time, as with the packagesite.txz?

i have booted into live dvd, is it possible for me to install package on to the live dvd?

okay i can

i just made a typo

so if i edit /var/cron/tab/$user do i need to kill the cron process to load it?

Demosthenex: why not just use crontab?

pkg: failed to extract pkg-static: Failed to create dir '/usr/local/sbin' -> can i not install packages in the live dvd? i know its temporary and it gets stored in the ram (at least in some linux distros). any idea?

I don't know how exactly the live dvd environment works, but with that error I guess it's just mounted read-only ("mount" should tell you in detail)

RhDoc: writing the file outside of crontab

i'm assuming i can do a service cron restart

Demosthenex: just out of curiosity - why outside of crontab? some kind of automated cron mgmt?

RhDoc sounds logical. Thank you

RhDoc: yes

The community is awesome :) forums.freebsd.org/threads/freebsd-…ages-on-read-only-file-system.76317

Title: FreeBSD LiveCD - installing a small set of tools from packages on read-only file system | The FreeBSD Forums

that's what i was lookingfor

13.x production ready?

yes

i would say so at this point. being weve been 13.{0-2}

I've been using 13.x since it came out, on my desktop.

for what type of production use are you looking for ?

as in ... "i plan on running webservers, shell servers, nas appliances, firewalls, mail ..."

wait, I am running 13.2 trying to run it as a daily-driver... is it not production ready yet? :D

lol

I mean, I haven't seen any bugs as of now

i can't note any downfalls at all on the network side of things as thats all solid , bgp, ospf, routing and related architecture

(and I have a MacBook next to me in a shelf, ready to drop in if anything happens

)

CmdLnKid: remember the bridge issue we talked about? turned out I had networkmgr installed (well... haven't thought of that one) which deleted the bridge during start-up :D

I thought that this thing would just give me "a nice little widget in a xfce panel" to quickly switch networks - well... kicked that package and now everything runs perfectly

hi there, I do have a few question about debugging pfsync -- is this a good place to ask questions for it, or can you recommend a more firewall focused channel?

apollo13: I'd advise that you just ask the questions in the channel; if someone knows the answer, they'll answer. If someone knows a better place to get you your answer, they'll presumably suggest it.

will do :)

Asking for people isn't really the best way to go about getting help :)

What I am wondering about pfsync: if I look into pfctl -s states I do not see a rule number on the second node. Should it have one? I am trying to figure out why as soon as a failover happens the pf filter log shows packages as blocked

according to pfctl -s states | grep "ip:22" the states are sycned and so I am somewhat at a loss where to start looking

I also see pfsync traffic on the sync interface, so "states" are most certainly synced

also "pfctl -s info" confirms that both firewalls have 10k states, since only the first one is active the second firewall shouldn't have that many states if sync were not working

so assuming pfsync per se works, what might cause issues for pf to match up those states properly?

So hopefully someone who uses pf can help. :)

rhdoc, good to hear you got it corrected and i lead you at the least to it

nothing wrong with ipfw

It's what I learned first, which I'm convinced is 95% of what dictates what firewall anyone is most comfortable with.

well even if my problem might be fixed with ipfw; I am someone who wants to understand stuff, so I'd really like to get to the bottom of this

can be said about pf and ipf as well but for the most part if you really want some qaulity i would vote ob ipfw

also firewall failovers that are causing connection hickups are annoying :D

pf user myself

apollo13: I wasn't suggesting that it'd be fixed with ipfw, just that I can't help with it :)

original ipf user for a long time

Well, for a long time, there wasn't a choice - it was ipf or nothing.

and that shit frustrated the shit out of me... but it always got the job done

debdrup, true

gues that could be said

I've had to admin ipf setups, mostly in protest.

hahaha

don't blame ya

There may or may not have been cussing involved.

that shit got long real quick

till that firewall manager software came out

dunno if its still around

but even that was a task in itself to convert between pf, ipf, ipfw

good ole days ... wipfw.sourceforge.net

Title: WIPFW

wow, that brings back memories.

not good memories, but memories.

haha sometimes its all you could do

a place that wanted me to consult with them were talking about ipfw in their interview, but when i got in to work on the first day i found out they were talking about that windows port

lololol

so that pretty much explained every single one of their performance issues

feel sorry for ya on that one

and they were having copious amounts of that one

yeet

they'd paid the bosses nephew or something like that to set up a firewall for them and they'd heard about ipfw from me because i know the boss and i kept singing its praises (well, i still do..)

kid must've had some idea of what to do, but to end up with ipfw on windows... i just don't know, man

some communication wires got more than crossed

shit that sounds like the situation im in now. sales guy is tech savy, he setup a bunch of shit just cause it works now

wife just started working there and its all way outa date

of course it is

so whos getting the contract while sales is busy ...

yours truly

ever since the mid-2000s when opensource had started taking hold, management have been attempting to forget that systems operation is an ongoing process, not a once-in-a-lifetime business invenstment

happy wife happy life

it used to be that at least some part of management understood that it was a thing you paid a another company a support contract to maintain

then opensource made everything better by not being proprietary bullshit, and management liked it because it meant that it was free and you could just hire someone to do it for you - then they started noticing that it looked like the people they hired weren't doing anything, so they fired them

and then boom ... places very much like manpower.com came into play but for SaaS

ends of eras

the ircd that underpinned an equivalent of AOL here in Denmark (called Opasia, I'm sure some of the other Danes in here remember it) was online until like 2018 because the ISP it was hosted at simply forgot all about it

leaky-cauldron.opasia.dk:6667 worked up until quite recently

i'm definitely octothorpe not sorry about the flashbacks i just gave those Danes :P

here in denmark at least, it never got to a SaaS point - instead there was just a bunch of aging infrastructure everywhere

then i got sick in 2016 and haven't been able to work since, so i've no clue what it's like now

wait a second here... did you happenn to send out some ram absolutely free to some guy in the states for a '07 DELL LATITUDE

and paid for the shipping

i may be confusing you for a guy i know through only irc but i owe this guy a ton

Nope, that ain't me.

For being only around 5 million Danes in the world (closer to 6 now, I suppose), there are a lot of us online, though.

DLange: acmesh-official/acme.sh #4659 would probably warrant an update of acme.sh to 3.0.6 (github.com/freebsd/freebsd-ports/tree/main/security/acme.sh)

Title: acme.sh runs arbitrary commands from a remote server · Issue #4659 · acmesh-official/acme.sh · GitHub

4659 – sysinstall changes the rc.conf file ? bugs.freebsd.org/bugzilla/show_bug.cgi?id=4659

dvl: ^

o/ moviuro :)

hi DLange , sorry if I pinged you for no good reason ( DLange ? d-vl ?)

no worries, moviuro

was just funny

The name resemblance was uncanny

yeah, we both start with d<tab> :)

dlang<tab> (I looked at d-vl's handle on github)

haha, yes dlangille is damn close

choosing unique nicks is (usually) hard :-)

(someone had a particularly surreal /quit message)

Hi I have " 890243072 86524960 1 ms-basic-data (41G)" line when i do gpart show ada0. Now when i try to mount it with " mount -t ntfs-3g /dev/ada0p1 /mnt", i get "mount: /dev/ada0p1: Invalid fstype: Invalid argument".

fuse is loaded "14 1 0xffffffff828ed000 11cd8 fusefs.ko"

what am i doing wrong?

are you sure it's ntfs and not exfat?

i tried exfat too, it returns the same result

If I were an OS I wouldn't mount NTFS either :D

i vaguely remember creating this partition, so no idea what it has, i want to mount and see before wipe

RhDoc haha, i haven't used ntfs in over more than a few years. but after my limbo with the heating issue i mentioned yesterday i was on tmp linux distru and did a lot of tem stuff.

temp*

ntfs-3g.probe -w /dev/ada0p1 || echo $?

this partition seems to be one of that

any chance you created the partition but never actually made a filesystem in it?

that command did return any result

chances that it could be

how do i find out if tht's the case?

also try file -s /dev/ada0p1

ah the powerful file command, thanks :)

the output is like three lines, can you suggest me a pastebin packaage in freebsd that you guys use?

bsd.to is currently offline, anything else for me to paste?

simple way, pipe the output into | nc termbin.com 9999

ah termbin, thanks again

dpaste.org is a bit nicer and can be posted to with a simple curl command, e.g. | curl -s -F 'format=url' -F 'lexer=_code' -F 'content=<-' dpaste.org/api

(I have a little script that puts the output of curl into the clipboard so I can just paste it into irc

nc termbin is most easiest.But my nature i dodn;t remeber much :(

That's cool. you have a github?

did you try mounting as ntfs-3g /dev/ada0p1 /mnt ?

I haven't put the script on github, but if you want it, dpaste.org/v6iuD

Title: dpaste/v6iuD (Plain Code)

strange, it worked. thanks

requires xclip and uses the X selection, you can tweak the options to have it use the clipboard

earlier i tried that command based on forums.freebsd.org/threads/cant-mou…ows-10-efi-ntfs-ms-basic-data.69697 and

Title: Can't mount Windows 10 EFI NTFS (ms-basic-data) | The FreeBSD Forums

it failed

thanks xclip and xselection works with X and not in wayland right?

no idea, I don't use wayland

presumably there is some clipboard service in whatever you're using, and in turn that ought to have a command-line client

There should be, Thanks again

by the way on an update, i am about to installfreebsd on my laptop. One more observation, my laptop starts heating up even with livedvd or from memstick too, as soon as it reaches the boot page

as I already said, the boot process will run one core at 100% until the kernel starts running

Is your fan working?

i backed up everything frommy system, and i was about to wipe the drive, that's when i noticed this ntfs partition, looks like i have some data in it and i am using scp to back it up

the question is, why is running one core at 100% an issue, and why isn't it cooling down reasonably well afterwards

Yes i remember, and i hope that can be figured out and sorted

i am full of hope now :)

(maybe loader needs to be improved to not need 100% of cpu, but that's something of a separate issue)

Did I screw up my system by having used the portsnap instead of using the git repository (and having installed pkgs as well)?

Tha main issue is not the heat emited by the loader, but, like you said, why doesn't it cool down after. Once this is sorted, i do plan to post my finding and may raise i bug, if its waranted?

somehow I've read that portsnap is/was deprecated, but on the other hand it's 2023 and it's still part of the system (and no mention of deprecation in the man-page), etc.

I am confused :D

warranted*

well... to cool down a working fan (and non-blocked air-ways) is needed :D

I mean, on my system, when compiling (like the chromium build Ii did) the CPU was getting hot and I could feel a lot of hot air being pushed out of the laptop

but as soon as the build finished it cooled down pretty rapidly

hence I am suspecting inproper cooling of the CPU

RhDoc sorry i miessed your message, yes fan is working and this heat up dooesn't happen with linux

missed*

interesting

may be i should change my irc username/handle to type_king?

so your system stays hot even when idling around?

typo_king*

:facepalm

hahaha

RhDoc yup, significantly. But this was more than a month ago. I am going to check if that's the case still, may be i did something wrong.

while you're still on linux, you might also want to check how hot it gets if you run some cpu-heavy process

also, if possible, check what cpu performance settings linux is using

I am no longer on linux, i have my system booted freebsd through a memory stick and i am connected with it over ssh on a family memebers windows laptop. But the good thing is i haven;t wiped my linux install yet, i can reboot and get the info

once the data i am copying from the newly identified partiiton, i will do just that

after*

sozuba: you can use "stress" to put some load on your CPU

you mean while i am on freebsd live dvd? or when in linux?

well, on both... for diagnostics - it will put your CPU cores on 100% utiliization

RhodiumToad,how long are you available? cause i don'tknow what information to look for when i am inlinux. Your suggestion was a good idea, i plan to take as much  of system and configuration info to a file so i can use that in freebsd if need be.

to see how the cpu temperatures are evolving with 100% load

I'll be around off and on for the next several hours

that's more than enough, thank you very much

I know little of linux (or of the latest intel power/performance management stuff) so I can't give specific advice for how best to collect the info in linux

RhDoc that's good, will do the stress test. But i hope that won't kill my cpu or burn my board? cause this is the only machine i have and i can't afford another for a while.

sozuba: I mean, all "modern" machines should automatically throttle down when it's getting too hot

avoid stressing it for long periods and keep an eye on the cpu core temps, assuming you have a tool to show that in linux

I can deal with the how best to collect, i would need your guidance on what to look for. I promise i won't need to be spoon spend, i will do my due deligence, read and research

I think the last burnt CPU I had, when overclocking, was back in the early 2000s

I'll keep that in mind, and yes i do have a tool to check the temps

what you're looking for, if it's a recent intel cpu, are the "Intel Speed Shift" settings, especially energy/performance preferences

RhDoc, that last machine i destoryed was a msi gt60 with a bowl of oat meal spilt on it :d

mine is an i5 4th gen

i have a thinkpad e450

what's the specific cpu model?

CPU: Intel i5-4300U (4) @ 2.494GHz

Haswell arch, so that's an older one

I think that's too old for hwpstate_intel, not sure

older, but still new enough to protect itself from burning up (throttle down or shutdown)

yup, i believe this laptop was released in 2015

RhDoc that's good to hear:)

powerd should be able to handle it, though

i'll check regarding this hwpstate_intel

that's the freebsd name, I don't know what linux would call it

i'll look it up

i have y tasks setup and noted downnow :)

my*

i am going to relogin through a more stable cygwin than browser

brb

back

okay, now backedup, i am going back to linux to get as much information as possible

afk, back in a bit

RhodiumToad: while reading about hwpstate_intel, i came across this bug -> bugs.freebsd.org/bugzilla/show_bug.cgi?id=253288 . It seems similar to what i am facing.The title is a bid unrelated to me specifically, as it says "modern Thinkpad" , mine isn't, its old if not ancient. :D what is your view?

Title: 253288 – hwpstate_intel: modern ThinkPads wedge under any kind of load or during boot

sorry just noticed your afk message:p

may be i should try this with freebsd 12.4?

while the bug i posted talks about freezingand unable to boot, it also has alot ofreportings of high temperature and hwstate_intel issues mostly affecting lenovo laptops

well, you can give it a try to install 12.4 and see if it's the same behavior

woah - considering that bugzilla bug I was lucky getting the T480 :D

RhDoc: hehehe

sozuba: that bug should be fixed in 13.2

yuripv: Mine isn'texactly that bug,but just something related. The patch provided heps avoid freeze and ubootable thinkpads, and over heating when pstate is disabled. But mine is heating even when i have the installatio image. i am currentlu going to check using 12.4 installation image and see what happens.

back

RhodiumToad: do you manage busy big postgresql installation?

running on freebsd?

welcome back

concrete_houses: not at present

RhodiumToad: There is a tool called s-tui that helps monitor and also perform stress test while reporting on load and monitor. What's the max temperature should i stop the stress test? i currently went 85+ and changed it to monitor mode and the temperatures are loering now.

how long under stress did it take to get to 85+?

I wouldn't go over that

and how quickly is it dropping after?

3-4minutes to heat up

and 5-7 minutes to cool down

there is an option to save the data toa csv file, which I amdoing

i will run it a few times and then get you the csv file

meanwhile i have alsow downlaoded 12.4 and 14.0 images to see if they have this issue

as before i booted into linux,no issues of temperature at all, the average cpu temperature ranges from 44-48 C for regular use of multipe tabs opened and terminal, etc

ok. and what temps are you seeing under freebsd?

I _think_ coretemp might not be loaded by default, in which case kldload coretemp

then sysctl dev.cpu | grep temperature should give current temps

RhodiumToad: i haven't yet checked the numbers yet, but i could literally feel the heat,almost burning. I will note down the temperature with freebsd once i amdone with linux

tests*

RhodiumToad: Thank you , will nnote that down

sysctl dev.cpu will show a bunch of cpu-related info, including current clock speed, available speeds, C-state usage, temps

ƒƒthat's cool. i also have sensor information fromlmsensors pacakge onlinux and the system hardware information, etc. I am doing this sui to get monitor and stress data for load and temp

s-tui*

RhodiumToad: looks like Freebsd 12.4 doesn't have that heating issue. Although i will perform more checks and share. For now, the temperature remains low comparedwith Freebsd 13.2, sysctl dev.cpu gives 53,53,50,50

will check for Freebsd 13.2 and Freebsd 14 as well

the deperature is not much high from linux though

huh.

does your reaction mean the results are not realistic?

no, just odd that 12.4 would be better than 13.2

RhodiumToad: i guess, its proably the introduction of intel pstate, but what do i know, i am just a noob :) my guess is as good as that of an ant

anyway these are preliminary

will keep you posted

given the age of your cpu, I don't think hwpstate_intel would even attach

but that can be checked in sysctl if you boot 13.2

will do, i amdetermined to get to the bottomof this :)

Freebsd 14,doesn't boot.none of the image versions boot,they end up with mountfilesystem> prompt

huh

what messages are before that?

freebsd 13.2 temperature starts at 44 C after boot in live dvd andreaches to 54 and stabiises therewithout doing anything just leaving the laptop idle. Will try Freebsd 12.4 again.

searchign for "some filesystem" i don't remember

i have let the laptopcool for few minutes and plan to run this temp range idle test on 12.4 again

on 13.2, is that with powerd running or not? and can you show the whole sysctl dev.cpu output?

sure, for dev.cpu output. How do i find out if powerdis running or not?

All the freebsd tests are in the live iso

service powerd onestatus

I think the iso does not enable powerd.

i'll do the 13.2 again then,share the output with you and the do the 12.4 again

oh okay,fair enough then

You are right powerd is not running

on live dvd

do you still want tosee the full dev.cpu output?

yes

ok

RhodiumToad: termbin.com/conc

huh

_is_ powerd running?

RhodiumToad: no

service powerd onestatus -> powerd is notrunning

... it's showing the cpu clock stepped down to almost the lowest speed, all cores spending most of their time in C2 state, and yet the temp is ~50C?

that makes zero sense

that's scary when things don't make sense

:(

should i post a bug report? or wait to analyse this further,make a full install, perform the same tests and then raise a ticket?

try starting powerd in the livecd: service powerd onestart

ok

run dev.cpu again?

yes

i'll give it 5 mins time to wait for changes?

good idea

okay :)

RhodiumToad: termbin.com/yll8

tempreature hasn't reduced 52,51,48,47

if you run top -CSH what does the cpu usage look like?

e.g. what percentage idle

cpu is 100% idle

it needs some whipping

strange, then why is my cpu heating up :(

daemon: i wish that was as simpleas that

:)

probably this should be a bug report

I mean ... have you checked your thermal compound and heat sync

can you do all of the same things we just did on 12.4 and show the output from that?

sure

your not overclocking it?

let me boot reboot to 12.4 live

it sounsd a hardware problem not software is the thing

no boot a linux live

prove its software

because it really does not sound like it

daemon: cpu 100% idle, clock speed stepped right down to lowest available, cores spending most of their time in C2 state or below,

and yet temps of ~50C

this is not anything like reasonable

daemon: you are lagte to the party, linux boots wellwith 43-48 C idle,mosty 44

50C is fine that is way below tjmax

and it seems to be specific to the OS

what cpu?

50C isn't high enough to be a problem, but it's pretty unreasonable under the conditions

haswell-era intel i5, -U submodel, don't recall the exact part

i5-4300U

in your range of 43-48C 50C is not that far off ... could simply be a sensor calibration

daemon: Intel Core i5-4300U

load it up in linux and bsd and reboot dead and check the hardware monitor in the bios

I am hanging onto the sensor misreading thing