is there a way to get a device name (e.g., /dev/ada0) by querying a WWN, serial, or other ID?

Any way to tell if/when a package was in place and ready for download from the FreeBSD repo?

I guess you can tell by the meta files timestamps? Don't really need to see the modtimes of the package itself...

I ported libfetch to linux and systemv in the 90s libfetch was so full of itself being full of freebsd that everything went straight to the internal buffers.

Heya FreeBSD people!

I have a question that might be related to the geom. Is there a way to configure how external storage drives are shown to a filemanager just by the filesystem label only?

what filemanager

rtyler: can you get any more info out if the crash dump? it looks very generic

do I need to install some newer radeon driver port to have all three monitors working that are plugged into the display/video/graphics card? (rx 6900 xt)

<sigh> Finally have a samba share working on with Windows 11.

given a value which is known to have exactly 3 bits set, anyone know a good way of finding the index of the middle bit?

(obvious way is to find the rightmost bit with ctz, clear it or shift it off, then do another ctz

What is ctz? (I was thinking of doing the same with ffs() but didn't mention it since I assume someone has a better idea)

ctz = count trailing zeros

__builtin_ctz in gcc and clang

aha

RhodiumToad: while i can let my mind wander in many different directions for "creative" solutions to this, i feel like the most sensible answer to this will be ctz yes

I actually need the indices of all three bits so there's already a ctz, I was wondering if there were a way to eliminate the data dependency

RhodiumToad: maybe my brain is slow because it's a bank holiday here. which data dependency are you talking about?

between finding the first bit and the second

ah

i'm not sure i can think of anything better than what you already have right now. it sounds about the way i'd do it and then forget about it unless it became a major problem for some reason

i am trying to set devd rules for the acpi_ibm driver but they are not working. The values are being read in /var/run/devd.pipe but not mapped to their correct functions in /etc/devd

what did you put in devd.conf exactly?

RhodiumToad: bsd.ac/bts79l5

you have a program called "brightness" ?

and did you restart devd? I'm not sure when it reads its files

sorted

thankyou, been messing around with acpi_video but it only allowed me to increment brightness in steps of 1

how can I enable powersaving on my wireless card?

I tried adding 'ifconfig_wlan0="WPA SYNCDHCP powersave"' to my rc.conf although i dont know if it does anything

the would would be a simpler place if there was a manpage that collected information on bootcode, the booting process and the different types of boot options into a single manpage.

There's some real gems in gpart(8) that should really be in the handbook or in my imaginary "boot" manpage.

I get an argument can be made that gpart(8) is an OK repository for crucial info, but it should not be the only place.

And lack of any debug facilities (outside of "add your own debug and build loader from source") is really a time waster.

/rant

Also, this is just bizarre:

Write out non-zfs bootblocks, cat them back from the partition, and there's zfs bootblocks.

I can even nuke ada0p1 (ie: gpart delete -i 1 ada0), add it back, and the zfs bootcode is still there. How is that even possible??

Even "dd if=/boot/gptboot of=/dev/ada0p1" doesn't seem to write gptboot (but also does not return an error). I am thoroughly confused.

hi

There are 784 executables (628 if excluding tests) in FreeBSD's source tree. Maybe I'll try porting them?

porting them to what?

Porting them to ...?

MinGW, MSVC, and maybe Cygwin or even OpenWatcom's DOS target

It's your time and effort.

I've done getopt and almost m4 github.com/SamuelMarks/freebsd-m4-portable github.com/SamuelMarks/freebsd-getopt-portable - and am starting to get the idea of what's missing and how to extract out the commonalities that are missing (like err.h)

Title: GitHub - SamuelMarks/freebsd-m4-portable: FreeBSD's m4 macro language processor ported to Windows & others

Trying to think of the best way to share with FreeBSD. I mean, do I send a bunch of CMakeLists.txt your way or would you prefer I don't and just maintain a separate tree?

Hmm actually maybe the best next target would be gitlab.freedesktop.org/libbsd/libbsd

Title: libbsd / libbsd · GitLab

Oh wait it's already open gitlab.freedesktop.org/libbsd/libbsd/-/merge_requests/1 - but maybe I'll start again as facebook isn't doing anything

Title: Support building natively for Windows (!1) · Merge requests · libbsd / libbsd · GitLab

spork_css: deleting and creating partitions doesn't erase their content

spork_css: gptzfsboot is much bigger than gptboot, so overwriting it with gptboot (which only writes the actual size of gptboot) doesn't necessarily overwrite all strings saying "zfs"

so it looks like my best time for IP-address to uint32 conversion is 65 cycles, about 20ns on my system

that's ipv4-only, but with full validation

about 2/3rds of that is validating the input characters and figuring out where the dots are, and 1/3rd for the decimal to binary conversion and checking for octets > 255

NetBSD's \libnbcompat is pretty neat

RhodiumToad: Got it - thanks. I did end up dd'ing /dev/zero in and that did give me a 100% confirmation the right loader was in place.

RhodiumToad: I also ended up dd-ing some zeros on the 4 non-boot ssds and forcing a gpart recovery just in case anything was on them that the loader didn't like

RhodiumToad: and lastly, I put a 13.2 first and second stage loader (ie: pmbr and gptboot) on, and no go. But putting /boot/loader from 13.2 on there, that lets me boot

RhodiumToad: so I still have no root cause here, even though I think I've cleared out any stuff that could be tickling a bug

RhodiumToad: before I go ahead and just update to 13.2 to "resolve" this is there any use in opening a bug, just so anyone else searching can find it?

probably yes

RhodiumToad: The box will be in production, so no opportunity for further troubleshooting later, but I'd like to dump the info somewhere

so it's definitively something in /boot/loader ?

RhodiumToad: Is it fair to conclude with all the swapping of boot loader bits that the issue lies somewhere in the 3rd stage loader (ie: /boot/loader)?

RhodiumToad: Sure seems like it, that's the only swap that made any difference.

that's the only explanation for what you just said

RhodiumToad: I just wish I could get more verbosity out of it when it's sitting there for 4 hours doing whatever it's doing

it's notable that loader is the first boot stage that has anything like a real zfs implementation (gptzfsboot only has just enough to find a root pool and /boot/loader)

RhodiumToad: I feel like I've run into about 10 red herrings on my way here. :)

you may not get all that much interest in the bug, because zfs was replaced wholesale between 12.x and 13.x (switching to the openzfs upstream)

RhodiumToad: Yeah, there's some old bugs dating back to even 8.x something still open, but I'm thinking a full brain dump might help someone else since I'm not on really obscure hardware and there might be a rush of 12.4 to 13.x upgraders at the end of the year

RhodiumToad: Anyhow, really appreciate your help, I absolutely needed someone to talk through this

Hi, since a couple of weeks, when I play audio I hear some noise, I think it was since I installed pulse audio (to be able to show the audio mixer in XFCE4 panel). I'm on a ThinkPad X220.

pulseaudio is kind of ew

RhodiumToad: yes. How can I remove or suspend temporarily

deinstall it?

RhodiumToad: yes, that was the problem.

does pulseaudio's rc file not give it an appropriate rtprio? so it won't run like a slug?

Title: Slug moves in fast motion - Invidious

RhodiumToad: Still couldn't nail down why dhclient was renewing via my internal DHCP server and not my ISP (but works fine on boot). Fixed it for now via reject in dhclient.conf, but I'm still curious. I don't have the logs available at the moment, but I watched it yesterday and saw that when it was renewing it was getting DHCPACK from both

servers.

Would like to figure it out at some point just out of curiosity.