hey folks, i need help configuring my microphone. my sound works and my drivers are loaded, but the mic only gets noise. it pics up my voice buts its all muffled in noise as if my playback and recording are mixed into one

i can also hear electricity when moving my mouse

are you having to turn the volume up a lot to hear anything?

yeah, but i only hear some crunchy noise as it tries to pickup my voice

so this could be an analog electrical issue, or it could be that your system expects some specific frobnication of audio pins in order to make it work

what type of sound chip?

(hdac chips have some stuff you can rewire using sysctl)

hdac Real Tek

here is my config from dmesg dpaste.com/HRPMHTGZ5

Title: dpaste: HRPMHTGZ5

only my rear mic is functional and connected

but just hises

sysctl dev.hdaa.4 should show you the topology of the onboard hdac

are there any gpo_* gpi_* gpio_* pins?

doesnt look like it

just nid

sysctl dev.hdaa.4.gpio_state

quite a lot of output

Title: dpaste: 8VJCJDDD7

dev.hdaa.4.gpio_state: 0=disabled 1=disabled

you could try twiddling those to see if they do anything

its readonly

sysctl dev.hdaa.0.gpio_config="0=set 1=keep" for example

try each of them individually with values set, clear, then back to disable

alright

sorry, .4 there not .0

(I was testing it on my system, where the onboard hdac is unit 0)

oh ok

also, you checked you plugged the mic in the correct place?

Is there a way to find out if a certain library will be included in the base OS of an upcoming release, and if so which version? (In my case, I'm wondering which version of OpenSSL will be included with 14.0)

yeah its in the same place in which it worked on previous OS'es

wait till feature freeze and then look at the tree?

also changing the configs dont seem to do anything

has it worked in freebsd before, or just in other OSes?

I guess that's all I can do at this point.

only in others, this is the first time i have freebsd installed on this system

but im sure the hardware is fine

vdamewood: or look at what the devs / release team are saying about it if anything

it looks like from the sysctl that the pink jack (mic) is disabled

gzar: no

but it says [DISABLED] in big letters

where, though?

the rear pink jack is nid24, and if you look through the various mixers, you'll see that nid11 has a non-disabled input from it

RhodiumToad: I'm looking through the main branch at the moment since 14 hasn't branched off main yet.

vdamewood: I've seen some discussion on mailing lists that suggests to me that the plan is to switch openssl versions at some point, but I don't know if that'll be before or after branching 14-stable

gzar: also nid35 is another mixer with input from nid24

nid35 has it disabled tho

however, some of the topology here seems... dubious

nid34* sorry

all of those mixers have mute=1

nid34,nid35,nid11 get input from nid24 which is unmmuted but the mixers themselves are muted

are you listening to the line-out/headphone-out or are you reading audio data from the mic pcm input?

i dont know exactly what it is you are asking

i am testing my recordings from audacity

are you running some program to receive the mic input?

yes

Hmm... main is still on 1.1.1.

ok, so you're using audacity to record from the pcm input

Well, I'll just watch and find out, then.

i can only record from the /dev/dspN devices

right

did you check that you're recording from the right device?

i tried all of them, only /dev/dsp4 seems to pick up something

ok, so that's probably this:

pcm4: <Realtek ALC883 (Rear Analog 7.1/2.0)> at nid 20,22,21,23 and 24,26 on hdaa4

yeah, seems like it

which corresponds to association 3 in the topology

so we trace it through: nid8 is the actual analog-to-digital part, which gets input from nid35, but it's shown as muted

what's the output of `mixer` ?

might need mixer -f /dev/mixer4

Title: dpaste: 4XDAPEW5Z

ok, I think the values listed in sysctl are capability values rather than current settings, which makes sense

RhodiumToad: it makes no sense to branch off with OpenSSL in about to become EOL, so currently lots of effort is put into integrating it

OTOH, 13 will need to be supported for a couple of years, so there's probably going to be a lot of OpenSSL 1.11 patches floating around and shared between different OS distributions

I reckon OpenBSD has a much easier time with this, given that they've become their own TLS library vendor

iirc, their API is different?

yes, but, to be fair, all they've done is speed up the deprecation that's now happening in OpenSSL 3

OpenSSL's API used to be absolutely atrocious. You had to model the TLS state machine in your software, and over the decades that has lessened. now you mostly get a Transport Layer abstraction

maybe if we stop calling it SSL, and start calling it TLS, the pressure to have a sensible Transport Layer API will force designers to… deliver that

I've mostly avoided having to deal with openssl's API

I recall at one point you had to do weird stuff like make sure your buffer was at the same memory address as before when continuing after a partial write

are there any tools i can use to inspect my sound config?

or guides? would configuring jackd or pulseaudio help here?

I don't think those would help

the snd_hda manpage might give some ideas

booting in verbose mode will also get you a bunch more info about what the driver is doing for the sound setup

oh man

might have to use a different computer for this

how can i modify some connections in the dev.hdaa.%d. sysctl settings?

the snd_hda manpage has examples

basically, put hint.hdac.4.nidNN.config lines in /boot/loader.conf (the manpage says in device.hints, but the two are equivalent)

i think i figured it out, my pink rear jack pin is not connected to any mixer, its disabled on all

another diagnostic is to do sysctl dev.hdac.4.pindump=1 which will spew a bunch of current status to console/syslog

cant i do it while the system is up? i want to test it before writting to loader.conf

do you have snd_hda compiled into the kernel or is it a module?

module

you could try putting the hint.* values in kenv, and unloading/reloading the module; I've never tried this

(it might refuse to unload)

so i have to reboot a bunch of times to test it?

... did I not just suggest a method?

you never tried it, so i'd rather not atm

ok, here's another method

you can modify the config with sysctl dev.hdaa.4.nidNN.config=whatever and then do sysctl dev.hdaa.4.reconfig=1 to reset the driver

(that'll remove and re-add the PCM (/dev/dsp*) devices, so you'll want to not be running any sound programs at the time)

I don't think I've tried that either, but I have high confidence that it will work

alright, thank you

i think im making some progress, managed to just remove the mic from showing up in audacity but that just means im messing with the right nids

hello everyone -

I have two questions

1) Is there any alternative to clonos --- some virtualization framework that provides graphics interface to create and delete virtual machines, clone them etc ? So far this is only solution I see in Freebsd clonos.convectix.com

Title: Free Open-Source Hosting Platform «ClonOS»

2) There was an article somewhere how to "consume" head, I want to start using FreeBSD 14 --- but I know the installations of applications and updates are different than release.... perhaps anyone can point me to a right place

3) I see BSD can is coming in May - I wonder if the conferences are broadcast ?

I any place that presents a functional use case for freebsd as host for virtual machines....

1 of 2: I use VirtualBox. 2 of 2: docs.freebsd.org/en/books/handbook/cutting-edge/#current and docs.freebsd.org/en/books/handbook/cutting-edge/#makeworld. 3 of 2: bsdcan.org/2023/faq.php#streaming.

Title: Chapter 25. Updating and Upgrading FreeBSD | FreeBSD Documentation Portal

V_PauAmma_V, thanks - I am exploring all links you gave me --- it seem there will be both a stream and lately recordings posted --- that is good news --- I wonder why there is no tutorial no speech or presentation on freebsd virtualization ....

V_PauAmma_V, thanks also about sugestion for virtual box.... it is a "GPL contaminated" piece.... I am very confortable using qemu-kvm on linux ---- but I really want to switch to BSD solution bhyve - freebsd ---- I really made a serial effort ----- cbsd and clonos are so far the closest to a functional production options... tried others... though elegant, they were insuficient and short-lived.... but I switched back to linux for a year or so --- so I am not

aware if there is any functional freebsd virtualisation option that can be used not only in basement experiments, but at least as a small scale virtualization....

booo! gist.github.com/igalic/d0ef61a03ec4c3f0edab127912e9201b

Title: gist:d0ef61a03ec4c3f0edab127912e9201b · GitHub

i need to upgrade the kernel? on my own?? what injustice is this??

busted, idk, will setup a new computer

acu, I don't know whether anyone is working on a GUI front end for bhyve, or even if that's feasible.

V_PauAmma_V, clonos has a web interface to cbsd ---- I installed before virt-manager in freebsd but I did not know how to make bhyve be used in creating virtual machines --- perhaps now is a bit more functional - since I did not try it for a while...

Have you asked on #bhyve?

s/on/in/

V_PauAmma_V, thanks, I joined bhyve - posted few questions there... I appreciate your feedback

aside from the man pages for wg(4) is there a guide specific to setting up wireguard on 13.2? The guys I have seen are all from ports which has specific init scripts etc which base does not have

s/guys/guides/g

i think i'll solve my sound issue by rebooting into a live linux cd

or a presistant usb stick

anyone using zrepl ?

whats that

I assume you don't

only me, someone else here might be

On that note, does anyone use sanoid? If you had switched from "sanoid" to "zrepl", or vice versa, why?

:))

I'm using zrepl because it offers pull mode

well, trying/starting to use zrepl

I'm seeing a limitation in zrepl, apparently you can't define different snapshot intervals for different mountpoints

as you define all filesystems, they all share the same snapshot interval which is kind of silly

i haven't used zrepl but have used sanoid for years. sanoid uses a separate script called syncoid for replication that can be setup to push or pull.

sanoid allows you to configure different snapshot intervals for datasets as templates, it's very configurable

turns out zrepl can do it too, but you need to have separate jobs for that

ThinkPad T590 w/ Intel HD620 + Nvidia MX250. nvidia_xorg_enable=YES in rc.conf w/ nvidia-hybrid* pkg installed. nvrun-vgl * yields "Illegal instruction (core dumped)," and nvidia-settings yields "Address boundary error."

Am I missing something for 13.2-RELEASE with pkg set to LATEST?

(Illegal instruction in Fish = Segfault in Bourne)

Probably running into the issue of packages for 13.2-RELEASE still being built on 13.1.

Quickest way to test is to grab ports from git and build the port manually.

debdrup: Should the Tree from 13.2 installation be correct, or will it need to be checked out for manual update?

coreystephanphd: So long as you're using a version of FreeBSD that isn't EOL, ports don't care.

It's only binary packages that get built on the oldest-supported version, where things like header includes can cause breakages like that.

It's also possible it's not a header inclusion bug, though - and to find out what the issue is then, you'll still need to build the port, just with debugging symbols as described on wiki.freebsd.org/DebuggingPorts

Title: DebuggingPorts - FreeBSD Wiki

debdrup: Thanks. Will try in next day or 2 and report back here. I remember you tending to be in IRC 24/7 -- true?

coreystephanphd: it's very much the opposite; when you report back, I'd prefer if you don't hilight me, and just tell the channel.

If I can help, I will - but I don't like it when people expect me to answer all their questions.

if you know all the answers, soon people will ask you all their question. --debdrup, attributed

debdrup: Roger. Glad I asked.

johnjaye: no.

I know that's meant to be a joke, but it's not fun to be hounded all the time.

I get it. Trust me. I have students who ask me things that could not be farther from what I am reasonably positioned to know and/or help.

That's not exactly what I was saying.

And this is supposed to be a community discussion zone, not a few gurus being hounded. I get that, too.

My point is more that I like helping, but people shouldn't come to me just because they expect that I'll help them.

It's the difference between volunteer work and employed work, if that makes sense.

And not work in the American sense, work in the Scandinavian balanced-work-life sense.

I do not expect anyone to help me. But I did my homework, and nvidia-hybrid-graphics simply are not working -- and it is an older machine. I cannot imagine that I am alone in this, esp. since ThinkPads are ubiquitous among *BSD geeks.

That is all. Turned to the community to see what others might have working here.

Having both an iGPU + dGPU is fairly rare.

I know the way iGPU+dGPU worked has changed from back in the T420 era, because my T420 that I used before my T480s, had iGPU+dGPU - and I can't recall if I ever got the dynamic switching to work properly.

Even on Linux that was a PITA

I think maybe I did, but I don't remember a single one of the steps involved.

wildeboskat: I mean, it's bound to be a PITA on any system, because it's all proprietary blobs.

Ubuntu on my MacBook Pro 2010 (iGPU+dGPU) also uses only the dGPU and burns through the battery

Even a AMD iGPU+dGPU solution still involves firmware blobs.

The one saving grace of the T420 was that it could switch between iGPU and dGPU in the firmware.

Oh? Cool

nvidia-hybrid-graphics as a designated port is only about 2 years old, 1 year for -RELEASE. For the T590 and similar with Intel+Nvidia hybrid, recent Linux (kernel) has it sorted so that it works without issue, especially on up-to-date distros like Arch and Fedora. But, anyway, if the only thing that is not working of the complex set of things that I set out to have working in this specific

Of course on a Mac you can forget that because there isn't even a bios menu

installation is Nvidia graphics (even everything hooked up to my dock -- a *lot* -- is working), I am doing alright, and I can move forward with working on my BSDCan lecture -- my whole reason for running the installation today :)

why are thinkpads so ubiquitous anyway? i've heard they're super common among free software types

The downside being that it's a Sandy Bridge or Ivy Bridge CPU with 2 cores..

For me: Thinkpads are the only laptops which still have a serviceable keyboard

Everything else is that ultra-thin crap

johnjaye: the old IBM Redbook design which had 8 rows on the keyboard (which stopped with the generation that the T420 is in) was EXCEPTIONALLY well built, incredibly maintainable, and sturdy enough to take a real beating.

johnjaye: Pre-2020, everything in ThinkPads was easily user servicable. And we get the positive chicken-and-egg effect of *BSD developers owning them, so the hardware tends to be well supported across the *BSDs.

Unfortunately even lenovo is shaving off half a millimetre travel with every generation now

T420 is one of the best examples of the power of the Redbook design, because it had a socketable CPU - so you could replace the CPU.

When's the last time you've seen that in a laptop?

P.S. Any of you guys/gals coming to BSDCan? :)

coreystephanphd: I think the hybrid graphics port had _just_ come out by the time I started looking at used T480s'

The X270, T4/580, and T4/590 are the last great ThinkPads. The new ThinkPad is the Frame.Work.

coreystephanphd: ah i hadn't thought about that. but yes that is definitely true

coreystephanphd: that's the impression I get too, but they're much harder to source in the EU in general and Denmark in particular.

well i wasn't going to bring up framework. but since you mention it... do you run any bsd on it?

A few of the FreeBSD developers have various models of the Framework laptops.

T480s is great. Last model with a real ethernet port!!

Albeit foldable - the T470s still had a solid one

johnjaye: I do not own a Frame.Work (poor Assistant Professor who buys used), but I can say that jcs has brought at least the 1st gen up to 100% out-of-box support in OpenBSD, and now debdrup is helpfully revealing that some key folks around here have them, too.

A framework didn't work for me because of the glossy display but I think they finally fixed that

The big problem with the Framework laptop, at least the models I've seen, is that they insist on using Realtek NICs.

well that's my fear yes, i didn't want to buy anything without already knowing freebsd works 100% on it.

debdrup: They work with and officially sell the Intel AX210, which is pan-BSD safe

or at least 90%

coreystephanphd: I'm talking about RJ45 modules.

johnjaye: wiki.freebsd.org/Laptops

Title: Laptops - FreeBSD Wiki

dmesgd.nycbug.org/index.cgi this site is also good

Title: NYC*BUG dmesgd

ah thank you

purely by chance I got into freebsd at the exact version (13) when they brought wireless support to my laptop card

I think I need to go lie down again.

johnjaye: debdrup: Again, I do not own one, but I know for a fact that is out of date because jcs worked really hard upstream to get the Framework team to add a legacy suspend/resume to the BIOS (called "Linux Mode" or something) for OpenBSD

So at least that should be double-checked in Free* but is probably a green check now :)

also a problem with the framework is that the NIC is thicker than the laptop and that it connects over USB

In other words, johnjaye , it should be 90%+

Another really compelling thing is precisely that upstream responds so well to *BSD and other OS input. They want everyone to be able to hack on it, and so they do things like 'sure, jcs, let's add that legacy suspend/resume.'

that fills me with confidence, thank you

If you want to donate $1500 to the "University of St. Thomas Faculty Computer Research Fund," I will gladly take a Frame.Work and let you know how it works ;)

i thought you were going to call it the CSRG at your university. but i digress. i've learned something new which has made this conversation more than worth it.

hrm… wonder where DragonFly keeps its rc code

oh, it's right here github.com/DragonFlyBSD/DragonFlyBSD/tree/master/etc/rc.d (as opposed github.com/freebsd/freebsd-src/tree/main/libexec/rc )

Title: DragonFlyBSD/etc/rc.d at master · DragonFlyBSD/DragonFlyBSD · GitHub