It was.

does anyone know how vm-byhve did nat internally, when it worked?

evening

I have onboard SATA and a Marvell PCIe card. In FBSD, my Marvell seems to initialize first and gets ada0, ada1, etc. All my other OSs set up the onboard SATA first. How can I get FreeBSD to do the same?

Does this cause problems, or do you just need to adjust for a bit?

I occasionally add a drive to the Marvell controller. When I do, it is moving my FreeBSD disk from ada4 to ada5 or whatever.

And what impact does this have? Does it break your boot scripts?

It drops me to single user mode. I tired reconjiggering with /boot/device.hints, but it had no effect.

For example, hints.ada.4.at=scbus0 did not help. I can do this at the boot prompt, however, and it works.

So you are able to load the kernel but it breaks your boot scripts, which are dependent on the number of the ada device, rather than some other property by which the relevant filesystem can be identified.

I was going to try labeling the disks, but I'd rather just flip the controllers to behave like I'm used to.

Yes. Kernel loads.

You could compile a custom kernel to remove the Marvell driver, and have the Marvell driver run from your boot scripts, but that's a hack the size of Jupiter.

I would probably label the disks that are relevant to bootup.

i would agree with Reinhilde

The general wisdom is instead of device names to use labels which are persistent for exactly the reasons you are seeing.

hi. can you help me find out why a lot of application crash on my computer?

i use intel drivers

Welp, wasn't fast enough.

NinerTenny: Can you `cat /var/run/dmesg.boot | nc termbin.com 9999`?

ok, so i'm seeing an oddity on desktop x11. since my lsat update on revision 13, every few days overnight my screen will die and no key/mouse will bring it back. i can switch to console, and explore. if i restart xdm, the monitor is in teh wrong modes so i have to reboot. the xorg log says all input devices were deleted. any ideas?

Demosthenex: <not really helpful, but> on my system with an AMD graphics card I have a similar experience right now, but switching to console then back to Xorg doesn't have the resolution issue (it works). On my nvidia box, this isn't happening.

zykotick9: this is a decade old thinkpad ;]

yeah, i only noticed it after updating, so i should check what changed.

I would love to use a multi monitor using FreeBSD

Helper: works fine

we should be able to look up man pages with man 'name(section)'

i think that's a little bit of an issue because handling parenthesis isn't exactly easy iirc

'man 7 development' works though

but if you can modify src/usr.bin/man/man.sh to support `man development(7)` that'd be cool

is there a way to make the other man pages "clickable"? or otherwise navigatable?

Do you mean in a terminal window? They are clickable in man.cgi.

V_PauAmma_V: i mean in a terminal window

I know that vim and/or emacs might be able to do that, but I meant… a lightweighter option maybe

have a good night folks

i mean good morning but go to bed :D

The XFCE terminal emulator (and maybe others, but that's the one I know and use) does that (including starting a browser or opening a new tab) for URLs. So some of the machinery for that is there already.

V_PauAmma_V: Are you asking if the default XFCE installation comes with a browser installed by default?

No, I was answering meena's question about manual pages (in the "name(section#)" form) being clickable in a terminal window.

V_PauAmma_V: I'm thinking I'll get a neovim config together for reading man pages (via SSH)… locally i might just try to get used to using emacs

meena: you'd have to have your terminal emulator recognize a manual page syntax and instruct it to open a new window

URL clickability is a subsection of syntax hilighting

debdrup: oh, yeah, URLs aren't a problem, just see also man pages

terminal does not get the syntax, only already rendered output, so it will have to make (really) wild guesses for 'name(section)'

modern terminals support an escape sequence for URL links, see e.g. gist.github.com/egmontkob/eb114294efbcd5adb1944c9f3cb5feda

Title: Hyperlinks in Terminal Emulators · GitHub

but you won't be able to link to another man page on the local system that way

meena: it might be easier to have a launcher that can interpret "!man <page>" to open a new terminal window and invoke the man command for the page.

niiice

debdrup: so far, gist.github.com/c2277e8d18b23289d5289ca260cffb5b this does pretty much all i want it to do

Title: man.vim · GitHub

only the colour schemes are fucking with me

I don't speak vim config, sorry.

debdrup: me neither, i had to join #neovim and get those two lines spoon fed

Well, if it works, that's great :3

tmux-yank could probably do something similar

debdrup: i don't trust yanks

Reasonable.

Hey #freebsd! Im looking in my dmesg and i see a lot of mpt1: IOC Generated Bus Reset port, see here: gist.github.com/KenwoodFox/acec02e21316de3ab5602816fcfad077

I gave it a google but, still cant figure out what its telling me, whats mpt1? Whats going on?

Title: gist:acec02e21316de3ab5602816fcfad077 · GitHub

thank you VimDiesel

kenwoodfox: is it some old scsi adapter or a vm?

yuripv: I have a pair of HBA cards installed yeah, that would make sense if thats whats causing it, one of them is, a little old kinda? But iirc still in the supported devices list :o

cdn.discordapp.com/attachments/5066…77355/1052631027073429594/image.png First one on the list,

cgit.freebsd.org/src/tree/sys/dev/mpt/mpt_cam.c#n2298 seems to suggest it's a problem with the adapter?

Title: mpt_cam.c « mpt « dev « sys - src - FreeBSD source tree

Also, VimDiesel is a bot. :3

Alternatively, given the commit for that line, it might be because it's a dual port adapter: cgit.freebsd.org/src/commit/sys/dev…be1a851e9fe60e21e0aff5b5bebd51d1d25

Title: src - FreeBSD source tree

I do have the dual port model! Hm, ok, adapter failure could be it :C sudden onset adapter failure? I had this machine running with a different os previously

Thanks for pulling up the git :3 I really could not find any resources online!

more likely disks misbehaving, so that the hba needs to reset them (or it could be legitimate resets)

I dont have any disks on that HBA, just a single tape drive,

maybe couldn't hurt to try the other channel? Its dual channel after all

kenwoodfox: if you've only got one disk, maybe try disabling the other driver via device.hints(5)

debdrup: disabling the other driver, witch driver?

Is device.hints a command?

freebsd.org/cgi/man.cgi?device.hints ?

Title: device.hints

Oh, sorry. Nevermind that.

debdrup: oh ok, np

I read it as mpt0 and mpt1.

Oh, gotcha, no prob. I only have the single device connected but idk if mpt is that device or the driver

Im gonna try the simplest repair first :3 just, switching it over to the other channel, that cant hurt

mpt is the hba driver

yuripv: Gotcha

Looks like the same bus resets whenever i try to access the drive :c Not sure what to do exactly

What generates that error? Whats a bus reset?

Oh wait, actually the error is different now cdn.discordapp.com/attachments/5066…77355/1052638175140520006/image.png

I swapped over from mpt1 to mpt0 it looks like

Can you try disabling mpt1 in device.hints?

Sure thing, what does that look like debdrup

See the device.hints(5) manual page.

ok, one sec

So maybe something like: `hint.mpt.1.disabled="1"`?

Yep

Lets give that a go, reboot right?

Yea

yep, and is it possible to replace the cabling to rule it out?

Well, you might also get by with devctl(8)

..but reboot is probably easier.

yuripv: I dont have any spare cables :c I just switched from ubuntu to bsd, havent changed anything besides moving the cable to port0

Ill give it a reboot :3 wont take too long

Just as long as you use 'shutdown -r now' and not 'reboot'.

Oh ok, ill make sure to do that

init 6

debdrup: Hm no, it seems to be about the same :c

debdrup: cdn.discordapp.com/attachments/5066…77355/1052643008069259334/image.png Sorry for the pictures to show things and not gists, i dont really have a good workflow setup atm

maybe its jamed? I got decommisioned HP LTP... in few last months, multiple jamed

lisu: no codes :c i just loaded the tape

and it was just working on ubuntu, i still have the old OS on disk, maybe i should switch back over, reconfirm everything is still working and come back?

i really did not change or even move the old hardware, i just pulled the disk and installed bsd on a fresh one

Anyone want to have a quick peak at my website? I styled it like a man page being viewed from a console

Nothing to really promote with the site. It's boring and most (normal) people wouldn't get it

sure!

Title: crontab.pw

Im going to add some small open source to it at a later date

awesome

maybe setup subdomains for people to use... *shrug*

colors okay? style okay?

just realized that VimDiesel is a bot :)

hmm we should add some chatgpt into VimDiesel so not only would you not realize it's a bot, but you'd carry on long deep conversations with it as well

thats not creepy at all is it?

I think thats a good idea

maybe he will eventually replace us all?

so does skynet

i just assume i'm already dead and the google AI has taken over my persona

freedom is an allusion after all

(no i didnt type that incorrectly)

well i read it incorrectly

already conditioned!

RoyalYork: <from the other day I think> I really hope BSDCan happen in person next year, I'd like to get my physical copy of Absolute FreeBSD signed. It is by far, the most humours "technical" book I've ever read.

"Bert"

zykotick9: Does Michael Lucas show up to the events?

RoyalYork: I spoke to him several times during the last one...

Is it like ComicCon where they charge my kids $75 for a photo, signature and a GTFO look?

heh... dunno, didn't try to get any signatures/photos... but I doubt it ;)

zykotick9, RoyalYork, FYI: The recent issue of freebsdfoundation.org/our-work/journal/browser-based-edition includes a new PAM Tricks and Tips article by Michael Lucas.

Title: Browser-Based Edition | FreeBSD Foundation

thanks

rwp: thanks.

zykotick9: I found a return flight from Toronto to Ottawa *and* hotel stay from May 17 to May 20 for $655 on expedia. Now to see if I can get the time off work so I can attend BSDCan

Thats $483USD for my American friends here listneing in

scratch that.... thats based on double occupancy pricing

Its $1310 total for *2* people to fly and stay **OR** its $1020 total for 1 person. How does this make any sense?

Whats crazy is I flew from Toronto Sudbury for $159 roundtrip a few months back. Airline fares make no sense.

I'm suprised you can get to a gate, for any flight, for only $159 roundtrip. That almost seems "reasonable".

I've got a dumb kind of rtfm question I'm hoping someone can point me to a simple answer to. I'm just trying to run a command in a jail. If I start the jail, attach, and run my script (nohup /whatever/myscript.sh &) it's fine. I can't figure out how to get that to start automatically. All attempts just tell me it can't find it. Is there a normal recipe for "this should always be running in a jail" kind of situation?

RoyalYork: I hadn't realized that BSDCan had been announced for this year, and it's in-person :D I tried to join the volunteer mailing list, got a bounce message as I was still subscribed.

dsal: maybe exec.poststart or similar, or setting it up with an rc script inside the jail

jmnbtslsQE: Yeah, I tried variants of that, but they all claimed they couldn't find my script. It seems to be doing something I don't understand on initialization.

are you sure you give the right path

I also kind of want it to run in the foreground, but that seems quite incompatible with the thing that launches stuff.

if you want foreground, one good solution is to have a script that launches a tmux session inside the jail that will run them, but it will require some setup

Yeah, quite positive. It's the same command I run once I login.

so you have no such file or directory?

for the error

i have to leave, sorry, but i will be back later

If I try to run it as a poststart I get that. If I try to run it out of /etc/rc, it just doesn't seem to go and I don't see any output.

Yeah, np. I've been sitting on this for a while.

so if I launch my script from /etc/rc as `nohup /path &` I don't see evidence of it running. If I launch it as `/path` then it does run, but causes the jail launch to hang indefinitely which breaks the thing that starts the jail.

I just installed daemontools. That got it done.

dsal, In the jailed script where you launch your script to debug "exec >/var/tmp/my.log 2>&1; set -x; echo $PATH;"

And then you can look to see more debug of what is happening.

But sure daemontools is a good industrial strength framework too.