jaeger changed the topic of #crux to: CRUX 3.7 | Homepage: https://crux.nu/ | Ports: https://crux.nu/portdb/ https://crux.ninja/portdb/ | Logs: https://libera.irclog.whitequark.org/crux/
Guest59 has joined #crux
Guest59 has quit [Ping timeout: 245 seconds]
Guest59 has joined #crux
Guest59 has quit [Quit: Client closed]
TxLogicGuy has quit [Ping timeout: 260 seconds]
TxLogicGuy has joined #crux
TxLogicGuy has quit [Quit: Leaving]
stoffepojken has quit [Ping timeout: 255 seconds]
stoffepojken has joined #crux
stoffepojken has quit [Ping timeout: 255 seconds]
farkuhar has joined #crux
lavaball has joined #crux
GazL has joined #crux
<GazL> Morning all. :)
<GazL> Just a trivial bug report for a passing dev: The cron job shipped with man-db is missing a #! interpreter line.
<SiFuh> I don't recall cron ever needing an interpreter
<SiFuh> It usually uses the system shell, right? So if you want to change interpreters you'd add your own.
<lavaball> do i need CONFIG_SATA_AHCI_PLATFORM for my intel nuc atom emmc? kernel doesn't seem to find it when booting, so i have no root. enabled CONFIG_X86_INTEL_LPSS and all the other stuff suggested here: https://github.com/home-assistant/operating-system/issues/1112 still, doesn't find root and doesn't suggest the mmcblk device beneath it. just driver? ram .... number.
<lavaball> hm, says it's some system on chip thing. i guess not then.
<SiFuh> CONFIG_SATA_AHCI_PLATFORM is a system chip
<SiFuh> It's your SATA controller
<lavaball> so shouldi turn it on? the other guy here who gave me his intel nuc config doesn't have it though.
<SiFuh> Does your NUC us an SFF controler then no, but if so then you need it
<SiFuh> Chances are you need it
<lavaball> mmc0: SDHCI controller on PCI [0000:00:17.0] using ADMA .. should i enable intel_iop_adma for this?
<lavaball> hm, i can't select it. it's not there. never mind i guess.
<SiFuh> It probably requires something else to allow you to enable it
<lavaball> i've gotten that far. what was the key to make menuconfig show everything?
<SiFuh> Since it says using ADMA, I would think you are using a non-SFF and a DMA interface
<GazL> SiFuh: the job is launched by /usr/sbin/runjobs, not directly by crond itself. If you look in the script, it just uses 'nice' to laaunch the executables, so I think a #! is in order.
<GazL> All the other scripts in /etc/cron/daily have one.
<SiFuh> Maybe some fancy coding that the default shell can't handle?
deltahotel has joined #crux
deltahotel has left #crux [Leaving]
<GazL> Nope. Nothing fancy. We're just shipping a shellscript without a hashbang/hashpling/shebang/(whatever you want to call it) line.
<GazL> Anyway, I've pointed it out now, so I'll leave it to the devs. :)
<SiFuh> Heh shebang ;-)
<SiFuh> Cool stuff bro
<SiFuh> GazL: look on the BSD's and they don't use the 'shebang'
<GazL> I guerss it depends on whether the thing that calling them 'sources' them or exec's them.
<SiFuh> Yeah, I only have BSD's and CRUX machines here
<farkuhar> GazL: at least one other script in /etc/cron/daily does not have a shebang. On my system there's /etc/cron/daily/sshd, which issues the command /etc/rc.d/sshd restart.
<farkuhar> Now I'm curious to see whether a user logged in via SSH would be kicked off when this cron job runs.
<farkuhar> SiFuh: I downloaded the latest ParrotOS and booted into it just now. In a terminal emulator I typed "nvme list" but got the error "bash: nvme: command not found". So when you mentioned using ParrotOS to work around the CRUX iso's lack of nvme-cli, maybe you meant using an equivalent tool, not nvme-cli itself.
<GazL> farkuhar: they do seem to run anyway, but IMO it's safest to be explicit, even if it's only specifying /bin/sh.
<GazL> I thought a shebang was required, but apparently not. Maybe it depends on the invoking shell though.
<SiFuh> farkuhar: probably I installed it using apt get
SiFuh has quit [Remote host closed the connection]
stoffepojken has joined #crux
<farkuhar> GazL: according to crontab(1) "Unlike other cron daemons, this crond/crontab package doesn't try to do everything under the sun. It doesn't try to keep track of user's preferred shells ... Instead, it just runs all commands using /bin/sh."
SiFuh has joined #crux
<SiFuh> farkuhar: ParrotOS is a live CD and CRUX is not
<GazL> Yes, but the crontab only runs /usr/sbin/runjobs, which is what invokes the jobs in cron/daily, so what crond does isn't relevant.
<farkuhar> SiFuh: even without connecting to a network and running apt-get, you can still use GPartEd on the ParrotOS live CD. But I'm not sure if GPartEd has the low-level functionality for issuing NVMe secure erase commands.
<GazL> Anyway, it seems it does still work, even without the shebang, so "not a bug" as such. I'd still prefer to see one specified though.
<SiFuh> farkuhar: Gparted is mostly modular.
stoffepojken has quit [Read error: Connection reset by peer]
<farkuhar> SiFuh: I do recall seeing some low-level commands like ntfsclone in the progress monitor for a GPartEd operation, so maybe it would try to utilize nvme-cli if found, when performing a secure erase.
<SiFuh> Possibly
<farkuhar> Seems like there's nothing that would recommend ParrotOS over another live CD distro, if you end up using apt-get anyway (to obtain a tool that hasn't yet been shipped with the CRUX iso).
<SiFuh> I don't
<SiFuh> I use ParrotOS for a lot of other things like pen-testing
<SiFuh> Gparted doesn't support nvme-cli
<SiFuh> ParrotOS stays on a USB drive and I take it with me. Many times used it at internet cafes which don't seem to exist anymore.
<farkuhar> Did all the internet cafes in your region change their business model as a result of the COVID pandemic?
<SiFuh> No
<SiFuh> They shut down long before the fake pandemic
<SiFuh> I am guessing the reason is the introduction of celluar phones
<SiFuh> With browsers and stuff
<farkuhar> And now you don't even need USB tethering to get a laptop online. Some laptops these days have their own slot for a SIM card.
<SiFuh> I use other people's laptops ;-)
<SiFuh> Much easier carrying a no-laptop than an actual laptop
<SiFuh> What I really wanted but never bought was those ISO/HDD devices. I knoa a few of the guys here had used them
<SiFuh> THis thing
<SiFuh> IODD
<SiFuh> To be honest though, 100% so. I thought most heavy/daily Linux users were aware of distributions like ParrotOS
<SiFuh> I could expect people to not know Sourceror for example but something more mainstream is kind of interesting
<lavaball> i have that line: mmc0: SDHCI controller on PCI [0000:00:17.0] using ADMA . but i still get vfs sync error. block (0,0) not found. or something along those lines. then it suggests a lot numbers with driver? on top and ram at the end but no mmcblk device.
<ukky> lavaball: please post your full kernel config
<lavaball> says error -6 after the unkown block (0,0) thing.
<ukky> is your root on eMMC? NVMe? SATA?
<lavaball> emmc
<ukky> did you add 'rootwait'?
<lavaball> no? where od i add that?
<ukky> Kernel command line
<lavaball> i have to figure that out in kernel, i don't have a boot loader.
<lavaball> should be where i put in the root=PARTUUID= ... right?
<ukky> right, that the place
<lavaball> just rootwait behind it?
<lavaball> seriously, considering the kind of questions i'm aksing i'm amazed i got this far ...
<ukky> yes, separate single keyword
GazL has quit [Quit: leaving]
<lavaball> let's make again, like we did last summer!
<lavaball> ukky, oh, and thanks for course.
<ukky> lavaball: lets see if rootwait helps
<lavaball> well, Intel(R) Atom(TM) CPU E3815 @ 1.46GHz Fill By OEM CPU @ 1.4GHz so yeah. this could take a while. though sometimes it's done like really quick. no idea why. feel free to explain you know.
<lavaball> if you know.
<farkuhar> Guest59's timing could not have been worse. Dropped in to ask how long it should take to compile qt6-webengine/qutebrowser, and within 24 hours all those qt6 ports got a version bump. Even worse luck, if Guest59 were running the same hardware as lavaball.
<lavaball> well, he should strap himself in then, because nvim took about 5 hours.
<lavaball> the kernel was considerably fast compared to that.
<lavaball> ukky, you are the man! or the fbi!
<ukky> does your system boot?
<lavaball> yes.
<lavaball> first i thought i left the stick in.
<ukky> cool
<lavaball> but then, no stick!
<lavaball> ukky, remember what i said about me asking amazing questions? should have asked about why it's advantagous to set a root password before you boot the system for the first time.
<ukky> lavaball: you set initial root password so you can login upon reboot, right? :-)
<lavaball> well, actually i was proud keyboard worked. didn't on the other box.
<lavaball> anyway 2 down, 2 more to go!
<lavaball> also i'm really quick mounting the chroot stuff now.
<SiFuh> You asre stripping the kernel right?
<lavaball> stripping?
<SiFuh> Removing shit you don't need
<lavaball> it already feels quite naked as it is.
<lavaball> that's how i start.
<lavaball> what i don't know. i don't put it.
<SiFuh> I have noticed above and when I make the kernel config I also have this issue so I read up on it. Pain in the arse job though
<SiFuh> But anyway, you got a prompt now?
<lavaball> though, since you mentioned the stripping. i have a lot of modules this time around. do i put them in some module file to load or how does this work?
<lavaball> right now i'm in chroot changing the root passwd.
<SiFuh> Look, the easiest way and the reason I did it was to create a static kernel that was compact enough for everyone. However you can strip it down more.
<ukky> lavaball: udev auto-loads kernel modules
<lavaball> isn't that nice of udev. someone should give it some flowers or somethign.
<lavaball> all these little helper thingies. how appreciates them, i ask you? who?!
<SiFuh> Everything else is modular. You run it and see what modules automatically load and write them down. You do some general usage and if you need to load anymore modules, write them down. Once you are happy and have your list, then smash the crap out of the kernel and vacuum to junk you will never need out of it. lsmod is your buddy
<SiFuh> lavaball: I agree with you. "All these little helper thingies"... Just get out of my way
<lavaball> though i remember on my bananapi, i needed to modprobe tuntap. no idea why it didn't load automatically.
<SiFuh> I load some modules with modules.rc but that is because they are not neccessarily needed to bring the system up <---
<SiFuh> modules.rc maybe rc.modules. Can't remember
<lavaball> that was back on arm arch. long time ago.
<lavaball> already systemd though.
<SiFuh> I doubt a tunnel 'tuntap' not sure what that is but I doubt it is an autodetected HARDWARE thing.
<SiFuh> Yeah the tunnel. That's software so it will be probably be skipped by udev
<SiFuh> As an example lopp and iso9660 I load with the rc modules file because it is software based and not hardware
<lavaball> was there automatically everywhere else though. just on arm arch not.
<SiFuh> I hate automatic crap. I prefer write to scrips or something and do myself
<lavaball> well, back then i was glad when it worked.
<lavaball> you have to give some leeway to us newbies.
<SiFuh> I don't consider anyone f***ing with a kernel and experimenting and booting with the god aweful kernel as a newbie :-)
<SiFuh> You ain't no newbie in my books dude
<lavaball> you just wait. my conscientiousness is so low, i will not just convicne you that i'm a newbiew, i'll make you doubt i can read at all!
<SiFuh> Hehe
<SiFuh> If you are unsure of something in the kernel and you have the patience. You can search anything CONFIG_* and start to study it
<lavaball> not really. mostly it comes down to trial and error there.
<SiFuh> Same for me
<SiFuh> I have to pile through the documentation and test on bare metal. The if I have a system that doesn't have this configuration I have to pile through more documentation
<cruxbot> [opt.git/3.7]: python3-cffi: 1.15.1 -> 1.16.0
<cruxbot> [contrib.git/3.7]: liborcus: 0.18.1 -> 0.19.0
<cruxbot> [opt.git/3.7]: nodejs: 20.7.0 -> 20.8.0
exark has quit [Ping timeout: 245 seconds]
exark has joined #crux
<cruxbot> [core.git/3.7]: meson: 1.2.1 -> 1.2.2
<jaeger> I also load the loop module, but the kernel can figure out iso9660 on its own, for what that's worth
<SiFuh> jaeger: I never understood that. Mostly software never loads but sometimes it does
<SiFuh> But in any case, my father hated out 8086 XT in 1988 because it would only work when it felt like it.
<SiFuh> out/our
lavaball has quit [Remote host closed the connection]
<cruxbot> [xorg.git/3.7]: [notify] mesa: 23.1.8 -> 23.2.1; new optional dependencies
<cruxbot> [compat-32.git/3.7]: mesa-32: 23.1.8 -> 23.2.1
<cruxbot> [compat-32.git/3.7]: mpg123-32: 1.31.3 -> 1.32.2
groovy2shoes has joined #crux
lavaball has joined #crux
lavaball has quit [Remote host closed the connection]
lavaball has joined #crux
<cruxbot> [core.git/3.7]: kmod: 30 -> 31
<cruxbot> [compat-32.git/3.7]: kmod-32: 30 -> 31
<cruxbot> [xorg.git/3.7]: mesa: updated signature
<ukky> Did anybody try to use runit as init system instead of sysvinit in CRUX? Any disadvantages discovered?
<jaeger> Not I
<ukky> okay. I am looking for support of inter-service dependencies. sysvinit doesn't have those. runit should have them.
lavaball has quit [Remote host closed the connection]
Guest92 has joined #crux
Guest92 has quit [Client Quit]
tilman has quit [Ping timeout: 260 seconds]
tilman has joined #crux