00:32:36 <llua> depending on how fast memory is being claimed, usually yeah
00:44:18 <wipt> Hey there! Sorta a FreeBSD question, sorta a networking question. I'm looking to upgrade my network to 10g fiber, switches are incredibly cheap now. But I also need PCI cards for a FreeBSD machine or two. I understand that I need that card to be compatible with BSD, but does the transceiver need to be compatible as well? I'm seeing intel ones listed in the hardware compatabilty chart, and then I have to
00:44:24 <wipt> wonder if they'll be compatible with the switch then. Also don't know if I need single mode or multimode SFP+ cables.
00:45:58 <wipt> Happy to be linked to a resource. I've dug around a lot on this, and haven't found much for helpful resources, and maybe I'm overthinking it - but I'd also like to avoid buying hardware that's not compatible, sorta just want to get things up and going and forget about it for the next 15 years when 40g or whatever is cheap.
01:26:40 <ivy> wipt: Intel cards only officially support Intel-branded SFPs.  unofficially, i've had non-Intel modules work on Intel cards using the FreeBSD ix(4) driver.
01:27:56 <ivy> the modules on each end of the link do not need to be compatible with each other (that's implied by them both being 10GBASE-LR or whatever) but the module in the switch needs to be compatible with the switch, which varies by switch vendor
01:28:47 <ivy> for cables, you need to match the transceiver to the cable type.  tl;dr: always use 10GBASE-LR over single mode fibre, unless you have existing (legacy) multimode fibre or can't use singlmode for some reason
01:29:29 <ivy> if you use multimode fibre, you would need 10GBASE-SX instead, and that does need to match, i.e. both ends should -LR or both should be -SR, because they use different wavelengths
01:29:58 <ivy> i mean -SR, not -SX
01:35:03 <ivy> also make sure you buy fibre with the right connector, for 10G SFP+ you will almost certainly want duplex (two core) LC, and if you have a choice between APC or UPC, choose UPC
01:38:34 <kerneldove> llua ok ty
02:20:23 <wipt> ivy: Thanks!
05:09:12 <kerneldove> just installed 14.3 with zfs and i made the swap too little. there any way to increase the swap size?
05:09:27 <kerneldove> swapinfo says the device is /dev/nda0p2 fwiw
05:17:28 <deimosBSD> wipt: intel cards are well supported in freebsd. i use only singlemode os2 cables with intel cards because changing cards is easier than re-pulling fiber ever few years.
05:46:39 <rwp> kerneldove, If you just installed 14.3 then you have little invested.  Easiest way to change the partitioning is to start again and make another install.  It's just easier.  A few minutes of work.  And it will be done.
05:47:02 <kerneldove> ya i will if i have to rwp jc if it can be changed post install tho
05:52:11 <rwp> It's possible but it is easier to reinstall if you don't have anything invested.
05:52:20 <kerneldove> how?
05:52:26 <kerneldove> edit fstab and reboot or?
05:54:14 <rwp> Swap is a partition on the disk.  In order to make swap partition larger the zfs data partitions would need to be made smaller.
05:54:39 <kerneldove> ah damn fuck that ya i'll reinstall ty
07:35:51 <fernie> wipt: dac cable is also option if it does not need to be more than 7 meters
10:07:34 <Retrofan> Hi guys
10:18:33 <Retrofan> Strange problem here.. I suddenly lost connection to my internet using the ethernet, after unplug/replug the cable. The device is connected but no internet. I factory rested router.. but no luck. I think my device is blocked from the router for no reason; bec. any other device I try using the same Lan cable is working fine. Also, pinging the router
10:18:33 <Retrofan> ip "192.168.1.1" it retruns back "Host is down" error, and I can see on the router's admin panel an "Lease expired" warning for my device, but it never got any new ip (I tried to set static one on my device, but no luck). I want to mention that I am using that ISP all-in-one router for 11 years now..
10:19:58 <Retrofan> tried every way ever I know (like restarting the netif and routing on freebsd) but no luck.
10:20:07 <Retrofan> and I am mosaid
10:21:08 <Retrofan> These been like that for days can't find fix, so please help me..
10:51:22 <divlamir> You see any errors about your NIC in dmesg? Which way do you wanna go: static or DHCP?
10:51:49 <Retrofan> DHCP
10:52:07 <Retrofan> wait i will see
10:54:14 <divlamir> So what happens when you do: dhclient IFNAME
10:55:02 <Retrofan> only timed out error
10:57:51 <divlamir> A problem with your router seems more likely, than one with FreeBSD
10:59:01 <Retrofan> divlamir: reconnected and the router given me a new ip with renewal time, but no internet
10:59:24 <divlamir> So you can ping the router now?
10:59:56 <Retrofan> divlamir: I want to mention that the internet connection was lagging after I started using pulsaudio..
11:00:14 <Retrofan> before the internet connection stopped
11:00:28 <divlamir> I can't imagine any connection between the two
11:00:50 <divlamir> Check your defalt route: netstat -rn | grep default
11:01:24 <Retrofan> still the same Host is down
11:03:18 <Retrofan> divlamir: GIves me the gateway ip (192.168.1.1 my router) and every thing else looks fine
11:03:53 <Retrofan> I think the problem is my pc cannot connect to the gateway
11:04:38 <divlamir> It gave you an IP, didn't it?
11:04:43 <Retrofan> yeah
11:05:01 <Retrofan> but cannot connect to the gatway after I get the ip
11:05:06 <divlamir> So you *are* connected to the router
11:05:10 <Retrofan> yeah
11:05:14 <Retrofan> but no internet
11:05:19 <Retrofan> only on my device
11:08:43 <Retrofan> divlamir: So strange..
11:15:49 <polarian> ivy: if you aren't busy could you look at this for me: https://lists.freebsd.org/archives/freebsd-pf/2025-April/000723.html I posted it ages ago and bumped it but it never got any responses, been thinking about it recently because of EuroBSDCon and all where I am away I use VPN a lot but currently have to hotspot myself and use vpn via my phone.
11:15:54 <polarian> could just be a config issue on my end
11:18:16 <polarian> I know you are good with networking so I am asking you :P
11:18:39 <polarian> I think you also use wireguard yourself, I believe you were the one giving config recommendations on the ML
11:19:57 <polarian> if you could take a look at it for me I would be grateful, I can't really give you anything in return :/
11:22:20 <polarian> wait a second
11:22:39 <polarian> my config has slightly changed since this
11:22:43 <polarian> lemme try it real quick
11:24:45 <Retrofan> I also tried every port on my router.. I confirm that's the strangest thing I ever faced in my whole life
11:27:40 <Retrofan> Wait a second I read somewhare few monthes ago, that's a bug in Freebsd when replug ethernet cable the default route will be lost..
11:28:26 <mzar> you are spreading disinformation Retrofan
11:28:41 <Retrofan> heh sorry
11:28:51 <Retrofan> I only try to remember
11:30:26 <Retrofan> But my route is not lost, Just my pc can't connect to it after I get my IP for no reason :(
11:33:38 <Retrofan> Could be my MAC is the problem?
11:41:11 <LxGHTNxNG> unlikely
11:43:23 <Retrofan> So what?
11:53:02 <Retrofan> every other device useing the same cable works...
12:40:54 <Retrofan> look like it's just a really bad luck :(
14:04:46 <polarian> ah wow
14:05:05 <polarian> alright I dont know what it was, I tried now and it works for me
14:05:19 <polarian> ignore my ping ivy it suddenly just worked, best not to touch it :D
14:17:34 <Retrofan> polarian: can you help me..
14:18:09 <polarian> Retrofan: you lost internet connection when wired
14:18:16 <polarian> alright pastebin the output of ifconfig please
14:18:32 <Retrofan> heh I can't copy it put I will try
14:24:37 <thorongil> if i've got a jail running a previous version of FreeBSD and i run freebsd-update on the host, the jail will still work without modification, right?  it's its own little world, all unto itself?
14:34:00 <Retrofan> polarian: I couldn't copy it.. it's just the regular result, and I have got a ip form the router
14:34:26 <Retrofan> but then router blocks me for no resaon of using the network but I am connected..
14:34:39 <polarian> Retrofan: what is the status:
14:34:51 <Retrofan> that happend after replug the ethernet
14:35:04 <Retrofan> status?
14:35:04 <polarian> yes but the IP can be preserved
14:35:10 <polarian> just because the IP is there doesn't mean it is working
14:35:12 <polarian> yes
14:35:18 <polarian> ifconfig
14:35:21 <polarian> each iface has a status
14:35:27 <Retrofan> Ah
14:35:27 <polarian> look for the ethernet interface
14:35:29 <polarian> what is the status of it
14:35:33 <Retrofan> preserved
14:35:36 <polarian> hmm
14:35:41 <Retrofan> I forgot that heh
14:36:01 <polarian> that doesn't sound right
14:36:03 <polarian> do you have a firewall
14:36:11 <polarian> on the device
14:36:13 <Retrofan> On freebsd
14:36:27 <polarian> yes
14:36:28 <Retrofan> not really at any high level
14:36:38 <polarian> that doesnt answer the question
14:36:41 <polarian> do you have a firewall y/n
14:36:53 <Retrofan> Let me check
14:38:21 <polarian> Retrofan:  cat /etc/rc.conf | grep -e "pf" -e "firewall"
14:38:26 <polarian> do you have either enabled?
14:38:31 <polarian> if not then there is no firewall
14:38:40 <polarian> FreeBSD has no firewall enabled by default btw
14:38:44 <polarian> so if you did not set it up then you do not have one
14:39:21 <polarian> Retrofan: also have you tried restarting the network stack
14:39:33 <polarian> service netif restart
14:39:45 <V_PauAmma_V> thorongil, do you mean previous minor version (eg 13.5 vs 13.4) or previous major version (eg 14 vs 13)? If the latter, only if your kernel configuration has the appropriate COMPAT(number) (the GENERIC kernel does).
14:40:13 <Retrofan> It's enabled (but allows everything).. so I disabled it
14:40:23 <thorongil> V_PauAmma_V: previous minor version, though in the future possibly major versions, too.  thank you for the answer!
14:40:30 <Retrofan> polarian:many times
14:40:44 <polarian> alright dump your ifconfig to pastebin
14:40:47 <polarian> and share it here
14:40:49 <polarian> I have to go though
14:40:59 <polarian> if nobody helps by the time I get back I will try and help
14:41:08 <Retrofan> I said I can't no internet
14:41:10 <polarian> the ifconfig output is required, its impossible to help without seeing the config
14:41:18 <polarian> take a picture then
14:41:29 <polarian> and upload it to a cdn
14:41:43 <Retrofan> Ah ok
14:41:56 <polarian> be back later
14:43:05 <Retrofan> wait I am doing it
14:48:34 <Retrofan> Can't copy it to windows device it fails
14:48:57 <Retrofan> corrupted
20:53:27 <kerneldove> there any way to track which files in a dir have been opened/accessed recently? i'm trying to figure out exactly which files are used for compiling a bin
20:54:47 <zi> kerneldove: ktrace make?
20:55:17 <byakuren> find . -type f -atime -1h
20:55:38 <byakuren> (for accessed in the last hour)
20:56:23 <kerneldove> byakuren, will that work even if i'm on zfs? because i think i have atime off by default no?
20:58:12 <byakuren> ooh, that's a good question - it might not - give it go though, no harm in doing so - i don't have an atime-less zfs filesystem to test with at this very moment
20:58:26 <byakuren> if not - then you might have luck with: find . -type f -cmin -60
20:58:45 <kerneldove> know how i can verify atime is on or off for my zfs?
20:59:36 <byakuren> zfs get atime
20:59:47 <byakuren> oh actually - yeah it should work
21:00:09 <kerneldove> looks like atime is off
21:00:11 <kerneldove> what should work?
21:00:23 <byakuren> likewise - forgot what system i was on - and the command worked for me c:
21:00:55 <kerneldove> ah
21:01:08 <kerneldove> how can "find . -type f -atime -1h" work even when zfs atime is off?
21:06:49 <byakuren> looks like it works for hard file accesses, like touch and editing - but not for soft accesses like cat...etc - so may not work in your case
21:07:39 <kerneldove> think i could test it by doing a build then run find . -type f -atime -1m ?
21:07:56 <mosaid> Hi
21:08:03 <byakuren> yeah! give it a go
21:08:13 <mosaid> I got my connection back using wlan usb
21:08:20 <byakuren> no harm in finding out - if not, then tracking will be your friend - i'll whip up my command gimme a sec
21:08:37 <mosaid> So I think it's a cable problem
21:09:34 <kerneldove> byakuren, i cleaned the build dir so build was fresh, then ran cargo build in the repo dir, build ran, then i ran the command, and it only outputted .git/index
21:09:46 <byakuren> this should do ya if not find: https://termbin.com/994f c:
21:11:07 <kerneldove> do i replace "make" with "cargo build" ?
21:11:59 <byakuren> yeah
21:12:43 <byakuren> it'll trace whatever command you're doing following it `-f` follows child processes, which i guess you need to see ALL the files it might use
21:13:12 <kerneldove> i tried to run it but truss said unable to follow forks for pid xxx operation not permitted. does it require sudo?
21:16:21 <byakuren> huh, works fine for me without sudo-like commands - i've even just tried it in a cargo project
21:16:34 <kerneldove> wtf
21:17:17 <kerneldove> how can that be
21:17:29 <kerneldove> which cargo proj? it public?
21:17:33 <kerneldove> i'll test with same 1
21:17:45 <kerneldove> i can clone a repo
21:21:01 <byakuren> -rw-r--r--  1 byakuren byakuren  3.8M Sep 14 22:20 ./truss.log
21:21:38 <byakuren> i chose a repo i just happened to have locally that was fairly complex: https://github.com/maxcurzi/tplay.git
21:22:44 <kerneldove> just cloned it, same command truss -f -o truss.log cargo build, same operation not permitted error hmmm
21:22:50 <kerneldove> does it matter im still on 14.1?
21:22:55 <kerneldove> can't imagine why it would
21:23:30 <byakuren> errr, i don't know that for sure - but i wouldn't expect it to
21:24:46 <byakuren> you could err run the cargo build, then seperately call truss -f -o truss.log -p PID - with the PID of the cargo build, but that's a bit funky tbh and without sleep in the start of the cargo build process you'll miss the early files
21:26:20 <kerneldove> got it!
21:26:27 <kerneldove> needed to set security.bsd.unprivileged_proc_debug=1
21:26:45 <byakuren> ohhh of course, i'm so sorry
21:26:53 <kerneldove> no prob it's my fault
21:26:59 <byakuren> c:
21:27:56 <kerneldove> then i need to grep for open and read, why both jc?
21:30:59 <byakuren> in case you were curious what was read from, and what just opened - no need to if you don't want to
21:31:33 <byakuren> some build processes just open files to check they're present - not neccecarily reading from them, i didn't know if that was your use case - so included both for ya
21:31:39 <kerneldove> i tried grepping for the name of the repo (to include only matches for the dir the repo's in) and that seems to cut it down pretty good
21:32:07 <byakuren> neat! you got it :D - yeah you'll have all the data you need, just need to filter it down now
21:33:23 <kerneldove> lots of fsstatat, openat, etc
21:33:35 <kerneldove> but i think this'll work as a starting point, wow ty??
21:35:52 <byakuren> c: no problem, i'm just glad i could help out
21:36:10 <kerneldove> weird, there's only 1 read match for matches of the repo dir name. how can that be?
21:36:27 <kerneldove> there's a read() for the repo's .git dir
21:36:45 <kerneldove> lots of open and openat in the repo dir tho
21:37:58 <kerneldove> lots of realpathat too
21:39:32 <byakuren> grep open truss.log | grep $(basename $PWD) | cut -d '"' -f 2 | sort | uniq
21:39:45 <byakuren> that should filter out the file lines for you you're intrested in - deduped
21:43:24 <kerneldove> which syscall is actually reading the contents from the .rs files?
21:48:20 <byakuren> (gimme a sec, just regenerating the truss log from a cargo clean)
21:53:05 <byakuren> looks like a lot of openat and statat
21:53:45 <kerneldove> do those read the contents?
21:54:51 <byakuren> openat does
21:55:11 <kerneldove> ohhhh ok now it's making sense
21:56:11 <byakuren> \o/
22:01:23 <kerneldove> this is starting to look really good (!?) in the output there's 2 forms of path tho which is kinda weird. 1 is the absolute and other is relative. so like /home/kerneldove/testrepo/pkgdir/src/file.rs, other like pkgdir/src/lib.rs. first group is always absolute, 2nd group is relative
22:01:41 <kerneldove> i tried it on your tplay repo and it didn't happen so maybe it's not complex enough for that to happen?
22:04:00 <byakuren> are the syscalls for them at the start different? open vs openat - openat takes a directory (something like AT_FDCWD) while open does not
22:05:03 <kerneldove> i see pkgdir/src/lib.rs in both styles
22:06:19 <byakuren> i'm not too sure then i'm afraid, as i don't have that file or your repo locally to poke about with
22:06:32 <kerneldove> sec
22:10:51 <kerneldove> ok git clone --depth 1 https://github.com/alacritty/alacritty; cd alacritty; truss -f -o truss.log cargo build -p alacritty_terminal; cat truss.log | grep $(basename $PWD) | cut -d '"' -f 2 | sort | uniq
22:16:10 <byakuren>  done that
22:16:41 <kerneldove> see how 1st group of dirs is absolute /home/..., 2nd lower group is alacritty_terminal/... ?
22:18:00 <byakuren> oh yeah - but only a small number of dirs at the bottom thee
22:18:12 <kerneldove> ya, why is that?
22:22:27 <byakuren> ahh - interesting - so when you run a cargo build - it basically scans all the files in the repo at the start for dependency resolution and indexing - so once the resolution and build is formalised, it then actually uses rustc on the relevant files to compile them
22:22:36 <byakuren> so the relative ones at the end are the most relevant for you c:
22:26:18 <kerneldove> does that mean the absolute path files aren't used for the actual building? or maybe both resolution/indexing and building?
22:28:13 <byakuren> from what i can gather (i'm not a rust dev) - the relative files are used for building, the absolute ones are used for the dependancy resolution stuff - hence why README files and such are scanned initially absolute, but not at the end relative
22:29:12 <kerneldove> hm i wonder if absolutes could be deleted (as long as they aren't also relative) and the build still succeed
22:29:54 <byakuren> you'd prolly still need like the cargo.toml...etc files tho
22:30:00 <byakuren> so you can't just delete the set difference
22:30:51 <kerneldove> ya good call. ok well this is amazing, ty so much for helping me figure this out!!
22:31:26 <byakuren> yeah no worries - i'd be curious to know how it goes!
22:31:50 <byakuren> kerneldove: do feel free to ping me if i'm around and you need any help with stuff, i'd love to help out where i can c:
22:32:02 <kerneldove> ok i'll do that. what i'm gonna try to make is a tool to autoslim repos to just what needs to be there for the pkg i want to be built
22:32:10 <kerneldove> ty i appreciate it :D
22:32:37 <byakuren> :D that sounds great!