klange changed the topic of #osdev to: Operating System Development || Don't ask to ask---just ask! || For 3+ LoC, use a pastebin (for example https://gist.github.com/) || Stats + Old logs: http://osdev-logs.qzx.com New Logs: https://libera.irclog.whitequark.org/osdev || Visit https://wiki.osdev.org and https://forum.osdev.org || Books: https://wiki.osdev.org/Books
<zid> m'ackerel
<zid> presented by gog's fedora
<gog> fish??
eryjus has joined #osdev
eryjus has left #osdev [#osdev]
vdamewood has quit [Quit: Life beckons]
smach has joined #osdev
Burgundy has quit [Ping timeout: 256 seconds]
spikeheron has quit [Quit: WeeChat 3.7.1]
[itchyjunk] has joined #osdev
Vercas6 has quit [Remote host closed the connection]
Vercas6 has joined #osdev
Vercas6 has quit [Ping timeout: 255 seconds]
Matt|home has quit [Quit: Leaving]
spikeheron has joined #osdev
<geist> Hah cute, a Ubuntu boot screen failure on a public display screen
<geist> First time I’ve ever seen that
<zid> NOT SYNCING - PANIC
<zid> seen a lot of those cus you know, hdds fail
<zid> and a surprising number of real devices just throw linux onto a small pc
<zid> or arm soc
<geist> Yeah that one I can buy. This is a Ubuntu ‘failure while checking /, press X or Y to do whatever’
<zid> gambling machines, atms, etc
<geist> Yah, RPI or something
<zid> smart ordering kiosks, these days
<zid> like mcdonalds
<geist> And not particularly thinking about the failure rate of some random board on a random SD card
vin has quit [Ping timeout: 252 seconds]
<heat> hard drive failure making it panic?
<zid> usually
Vercas6 has joined #osdev
<heat> the hdd failures i've had never panicked the kernel
<zid> reboot -> vsync not syncing - panic
<zid> vfs*
<heat> just a lot of verbose libATA log in dmesg
<heat> hm
<heat> really?
<geist> More like leeb s aaaat a eh?
* geist guffaws
<heat> what
<geist> Exactly!
<zid> https://i.stack.imgur.com/3U68k.jpg That's what linux does if you boot it without no root= that's valid
<geist> Heat gets it
<zid> not very user friendly, but that's what it does
<heat> oh I know about that panic
<heat> a classic
<geist> I’m on a boat!
<heat> good shit
<heat> i'm on my chair
<clever> zid: but not if you have an initrd
<geist> On a boat
<geist> Booting without initrd without modules is some hard core stuff
<zid> I never use initrd
<geist> Makes you feel powerful, like you have the power
<zid> I'm too lazy to set it up
<zid> I just build ext into the kernel, everything else am module
<heat> do you have to manually assemble the initrd in gentoo
<heat> because that sounds very gentoo
<zid> I never don't want ext4 loaded
<geist> Yah gentoo has that option. My gentoo install i just disable modules and build everything in
<heat> what if you switch to zfs though!!!!
<zid> I am rebuilding the kernel at that point anyway?
<clever> zfs/lvm/luks cant be done purely in kernel
<geist> And then i sit back, satisfied, arms crossed, hand in a Power Glove
<clever> they need some userland tooling to parse the metadata and create devices
<heat> sayw hat
<heat> late stage filesystems are something very funny
<heat> gimme mah god dang block based filesystems
<geist> That’s lame
<geist> Well FWIW btrfs works fine with just in kernel stuff
<heat> i want my unix filesystems like the kings of old
<clever> lvm kinda isnt even supported in-kernel
<geist> Though if you want to mount something other than the root fs within it you need some extra sub volume args, etc
<clever> the kernel just has device-mapper, which can be configured to map sub-sections of a virt device to sections of another device
<clever> lvm then deals with all of the metadata, and configuring DM
<geist> Well it’s more complicated with some of the higher level features of LVM
<geist> Like COW, thin provisioning, etc
<geist> But possible it still needs a user space process to do thes canning to set it up
<clever> ah, those didnt exist when i used lvm
<geist> Yah was fiddling with proxmox the other day and it leans on a bunch of those LVM features for backing VMs
<geist> Also for snapsnotting, etc
<heat> i'm starting to think filesystem development isn't real anymore
<CompanionCube> geist: iirc btrfs also needs the userspace bits with multi-device fses
<geist> Not sure. I think you just point mount at one of the volumes and it finds the other
<geist> *but* that might be in mount.btrfs
<geist> Good point, maybeyoucan’t linux kernel root off a multi device fs
<geist> I think there are issues with grub too
<CompanionCube> ah, seems there's a workaround: 'If you don't have an initrd, or your initrd doesn't perform a btrfs device scan, you can still mount a multi-volume btrfs filesystem by passing all the devices in the filesystem explicitly to the mount command. A suitable /etc/fstab entry would be: '
<geist> Aaah, so then the question is what about in root= command line?
<CompanionCube> geist: it also says that works
<geist> Okay, makes sense.
<geist> I haven’t looked how that works but i assume that each device has a duplicate piece of metadata that lists all the other devices’s UUIDs
<clever> zfs does do that
KaitoDaumoto has quit [Read error: Connection reset by peer]
wand has quit [Remote host closed the connection]
wand has joined #osdev
smach has quit [Read error: Connection reset by peer]
rwxr-xr-x has joined #osdev
<rwxr-xr-x> Yo yo!
<rwxr-xr-x> How is everyone?
<heat> how do you do fellow kids
<heat> lets listen to the hip hops
<rwxr-xr-x> we are on IRC
<rwxr-xr-x> fellow kids is as far from the truth as possible
<rwxr-xr-x> how are you doing heat?
<heat> fine
<heat> hbu
<rwxr-xr-x> im chillin
<rwxr-xr-x> podcasts and development
<rwxr-xr-x> not bad
<rwxr-xr-x> and vacation from work
nyah has quit [Quit: leaving]
lg has quit [Ping timeout: 252 seconds]
lg has joined #osdev
lg has quit [Ping timeout: 256 seconds]
[itchyjunk] has quit [Remote host closed the connection]
vdamewood has joined #osdev
<rwxr-xr-x> make
<heat> make clean
vin has joined #osdev
rwxr-xr-x has quit [Remote host closed the connection]
<mrvn> make mrproper
lg has joined #osdev
<clever> cat /etc/shadow
<heat> clever, $6$FoWygeEwGsV6fXwu$ys9YgcPkXpkFHO5tbkwLlMc3SClRv2l.n.hy7M4rkAD3MrW3ASRdvo67DO8im1kHjNXVLh5QKBoFfbwWQNLc10
<heat> bruteforce this
<heat> i'll wait
<clever> Loaded 1 password hash (sha512crypt, crypt(3) $6$ [SHA512 128/128 SSE2 2x])
<clever> nada on the built in wordlist
<clever> heat: brute forced "bonk" in about 4 mins, lol
<clever> you should have used a longer pw!
<heat> fuck
<heat> $6$PbXHRbD9vWQyI/CI$BH8kbpZjvb4Owp8JJ1.WMxNjZIo05YlBcgIHNNE7YA9CmHSjqM6ze26mwVQ3cu.f8nu8wZt7izHsMUTFRw8j5.
<heat> this one's a bit longer
<clever> started
<heat> i'll be remarkably impressed if you get this one
<clever> fail on the wordlist once more, back to brute forcing
<heat> well they're all salted
<clever> salt doesnt mean much for single hash cracking
<clever> salt is to prevent pre-hashing
<clever> i did learn something critical about sha256 a few days ago
<clever> basically, the sha256 hash, is just the internal hashing state
<clever> so you can convert a hash back into a state, and update it with more data
<clever> so, given sha256(data), i can trivially generate sha256(data+moredata)
<clever> so salting it like sha256(salt+data) is pointless
<clever> even if the salt is private
<heat> how is salting pointless?
<heat> your dictionary goes to shit
<clever> it just means you cant store a hash->word lookup table
<clever> you have to re-hash everything in the dictionary, with every salt
<clever> it doesnt break the dictionary, it just makes the cracking more computationally expensive
<clever> and you cant pre-calc things
<heat> if you look at crypt implementations salting is considerably more involved than just salt+data or data+salt
<clever> yeah, for reasons like what i just said with sha256
<bslsk05> ​github.com: Onyx/crypt_sha512.c at master · heatd/Onyx · GitHub
<clever> yeah, that looks similar to hmac, hashing the hash
<clever> but its also got rounds, where its hashing the hash multiple times, and many complex concats
<clever> for unsalted stuff, you can just make a big database, where you just insert hash(n),n into the table
<clever> and then you can just `select pw where hash = ?`
<heat> are you still bruteforcing
<clever> yep, lol
<clever> and then for disk usage improvements, you can make a rainbow table
<bslsk05> ​www.security.org: How Secure Is My Password? | Password Strength Checker
<clever> where you store pw, and hash(f(hash(f(hash(pw2))))) basically
<clever> where you have a function that can turn a hash into a random unrelated pw
<clever> you then generate a chain of 6 hashes, by just running the input hash thru that same function
<clever> and then `select pw from tbl where hash = ?`
<clever> if you find a hit, you know its somewhere in that hash(f(hash(f(hash(pw2))))) chain
<heat> can you give the cracking thing hints to speed it up
<clever> heat: did you include numbers or special? i think john is in ascii only mode
<heat> numbers and special :P
<clever> length?
<heat> 16
<clever> dang! :D
<heat> can you give it hints or is it hopeless
<clever> i need to first find the right flags to set the length limits
<heat> i mean im sorry to break it to you but you're probably not brute forcing it even knowing the length
<clever> yeah, 16 is pretty long
<heat> if you can hint it part of the password, it starts with "unix420"
<clever> ah, that at least shaves some chars off
<heat> even then you took 4 minutes for bonk
<heat> the magic password strength thing says 2 hours for the rest
<clever> model name : Intel(R) Core(TM) i7-7700HQ CPU @ 2.80GHz
<clever> and this cpu is rather old
<heat> not bad
<heat> probably better than your fx8350
<sbalmos> how about a 36-character password? :D
<heat> although you really should be hashing with your gpu lol
<clever> 01:00.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] Bonaire XTX [Radeon R7 260X/360]
<clever> heat: i dont think this is capable of compute
<heat> it may just be faster anyway
<sbalmos> it might be able to compute your grocery bill
kof123 has joined #osdev
<zid> damn, sha512crypt is slow
<zid> I'm getting like 4333 guesses a second
<clever> yep, thats the point! :D
<zid> (with 12 threads, and avx)
<heat> is that literally sha512crypt?
heat has quit [Remote host closed the connection]
heat has joined #osdev
<zid> it's still not done, but I don't wanna give up :(
xenos1984 has quit [Read error: Connection reset by peer]
bradd has joined #osdev
xenos1984 has joined #osdev
VY2 has joined #osdev
Vercas6 has quit [Quit: Ping timeout (120 seconds)]
Vercas6 has joined #osdev
heat has quit [Ping timeout: 256 seconds]
Vercas6 has quit [Ping timeout: 255 seconds]
jjuran has quit [Read error: Connection reset by peer]
jjuran has joined #osdev
Vercas6 has joined #osdev
ThinkT510 has quit [Quit: WeeChat 3.7.1]
ThinkT510 has joined #osdev
elastic_dog has quit [Killed (silver.libera.chat (Nickname regained by services))]
elastic_dog has joined #osdev
VY2 has quit [Remote host closed the connection]
Vercas6 has quit [Quit: Ping timeout (120 seconds)]
Vercas6 has joined #osdev
wand has quit [Remote host closed the connection]
wand has joined #osdev
zaquest has quit [Remote host closed the connection]
zaquest has joined #osdev
Vercas6 has quit [Quit: Ping timeout (120 seconds)]
Vercas6 has joined #osdev
GeDaMo has joined #osdev
Burgundy has joined #osdev
wootehfoot has joined #osdev
bradd has quit [Ping timeout: 260 seconds]
wootehfoot has quit [Read error: Connection reset by peer]
isaacwoods has joined #osdev
HeTo has quit [Remote host closed the connection]
HeTo has joined #osdev
hmmmm has joined #osdev
hmmmm has quit [Ping timeout: 260 seconds]
smach has joined #osdev
isaacwoods has quit [Quit: WeeChat 3.7.1]
cyao has joined #osdev
<cyao> How do I test if there is a device on the ATAPI bus? Is it the same as ATA PIO? (Check if bus is floating)
rwxr-xr-x has joined #osdev
<rwxr-xr-x> good morning/afternoon/evening all
<rwxr-xr-x> how's life?
<kof123> hello. you need a sticky bit chmod +t rwxr-xr-x
<rwxr-xr-x> lol
<rwxr-xr-x> people like the nick is what i've noticed
<rwxr-xr-x> drwxr-xr-x
<rwxr-xr-x> i feel like the way i have it now is the catchiest one
<rwxr-xr-x> rwxr-xr-xt is less so, and rwxrwxrwx is eh
<FireFly> why the executable bit?
cyao has quit [Remote host closed the connection]
<rwxr-xr-x> my nick has almost nothing to do with anything other than looks, rwxr-xr-x is easy for me to remember, people get it and like it, and it rolls off the tongue
<rwxr-xr-x> mostly how I make nicks
<rwxr-xr-x> 755 would make more sense
<GeDaMo> I don't think you can start a nick with a number
<rwxr-xr-x> probably not
<rwxr-xr-x> i've had a couple people here just call me 755 because its easier to type out than rwxr-xr-x
<GeDaMo> rwxr-xr-x: rw<tab> is easy enough :P
<rwxr-xr-x> Oh woah
<rwxr-xr-x> I didn't know that existed LOL
<rwxr-xr-x> that's cool
Lumia has joined #osdev
<gog> helo
<GeDaMo> EHLO
<j`ey> EL0
<gog> 200 OK
<j`ey> it was an aarch64 joke, not a http one!
<gog> yes i know :P
<j`ey> ok good, youre just normally an x86 fiend
<gog> yeh i know very little about ARM
<gog> or any architecutre not x86
<gog> i've thought about buying an rpi to play with tho
<zid> I am in this picture :(
<gog> i had a q6600
<gog> good chip
<zid> I had a q6600 with a 212 on it..
<gog> nice
<zid> I.. am using the 212.. for my desktop, rn :(
<gog> i had an aftermarket cooler for mine too
<gog> i can't remember what it was tho
<gog> something thermaltake
Vercas6 has quit [Quit: Ping timeout (120 seconds)]
<gog> mostly because i hated the pushpin stock heatsink because i installed it incorrectly more than once and overheated my cpu
<gog> so i got something with a bracket and screws
<zid> yea the stock 775 cooler sucked shit
<zid> not even as good as nvidia's new power cable
<gog> new poer cable?
<gog> they not using standard pcie power cables?
<zid> oh did you not follow along?
<gog> no i haven't had a desktop in a few years
<gog> the last nvidia desktop card i had was a standard power connector
<zid> the 4090 is using 12vhpwr connectors not EPS12V, because EPS isn't rated for much power (can deliver it fine though)
<zid> but there were lots of 'failures' on 12vhpwr by which I mean fires and melted gpus/cables
<gog> oops
<zid> turns out it gives like no feedback about whether it's fully inserted or not
<bslsk05> ​www.theregister.com: Lawsuit filed over Nvidia's melting GeForce RTX 4090 cables • The Register
<zid> and if it isn't, the pins are long enough that it connects in the wrong places if you tweak it sideways a little
<gog> so it draws too much current and fwoosh
<zid> nah, less current, but through a tiny sliver of pin
<gog> ah
<zid> instead of the full length, cus it's diagonal
<zid> someone managed to believe this cable was 'inserted'
<zid> you can see the damn line for how inserted it wasn't
<gog> eep
Lumia has quit [Ping timeout: 248 seconds]
<sortie> <heat> sortie in shambles
<sortie> Hey I'm actually happy about getting rid of leap seconds
* gog leaps
<sortie> Worst case we drift one timezone over the next 3000 years
<sortie> Meanwhile: DST
<sortie> If we screw up people a few thousand years from now, that's a pretty good compared to how well humanity is currently doing
* sortie is legit pretty prepared for y10k
<zid> are you prepared for 2038
<sortie> 2038 easy peasy
<sortie> Been prepared since I invented time_t
<sortie> The NTP rollover a bit more tricky although technically my ntpd port isn't merged and they improved it
<sortie> I look forward to the GPS rollover tho
Lumia has joined #osdev
Vercas6 has joined #osdev
<FireFly> 2038 is.. distressingly not very far away at this point
<sortie> I mean I've prepared for 2038 since idk 2011
<FireFly> lol I hadn't seen that variant
chovelyoukai has joined #osdev
Lumia has quit [Ping timeout: 256 seconds]
<sortie> “A speed-up since 2020 has also made the issue more pressing, because for the first time, a leap second might need to be removed, rather than added.“ ← Oh jeez that's going to be a nasty XK-Class Time Loop Scenario
<mrvn> Have you thrown some GPUs at it?
<mrvn> zid: ^^
<mrvn> sortie: Hey, we could build some big ass rockets and slow down the earth a bit.
* sortie drops a slightly larger ice cube in the ocean solving it once and for all
<mrvn> Why is the earth speeding up? I thought we had global warming. Heat expands and a rotating thing aslows down when the mass moves further from the center.
chovelyoukai has quit [Ping timeout: 260 seconds]
<sortie> probs a weird moon thing
<GeDaMo> Melting ice reduces weight on the crust something something
<zid> can we just wait until we have to add one next instead and have them cancel
<zid> that seems easier
dude12312414 has joined #osdev
<mrvn> GeDaMo: that would make the mantel expand so the earth slows down. again, why does it speed up?
<mrvn> I know, the moon is a mega construct and is misbehaving.
epony has quit [Ping timeout: 268 seconds]
<mrvn> time we send someone up there again to fix it.
chovelyoukai has joined #osdev
<gog> we like the moon
<sortie> https://www.bipm.org/documents/20126/64811223/Resolutions-2022.pdf/281f3160-fc56-3e63-dbf7-77b76500990f “Resolution 4“ “On the use and future development of UTC“ “decides that the maximum value for the difference (UT1-UTC) will be increased in, or before, 2035,“ ← Boom. The 2030's gonna be wild.
<mrvn> What's the current maximum value?
<gog> ±0.9
<zid> but not as much as a spoon
<zid> time to eat my baton without taking all the skin off the roof of my mouth
<zid> I hope
<gog> baton
<zid> yes, eating a french magic wand
<gog> oh my
<gog> sacre bleu
<gog> mon dieu
<gog> hon hon c'est mon baguette magique hon hon baton pain
<zid> french is a silly language
<mjg> would not it be funny if gt results were tailored, just like search results
<gog> we don't know that they aren't
<gog> yes
<gog> i want one
<gog> an important question tho
<gog> is this a technical
<zid> It is a technical frog
<zid> it is not however, technically a frog
<mrvn> is it a pipe?
dude12312414 has quit [Quit: THE RAM IS TOO DAMN HIGH]
<gog> yes
<gog> what kind of pizza should i get
<mrvn> everything but anchovies
<gog> hmmm
<zid> jalapeno and mushroom, add a meat if you want
<zid> ham maybe
<gog> that sounds pretty good
<mrvn> pineapple if you are one of those
<\Test_User> pizza-topped pizza
<zid> ask them to make sure the cheese has infinite area but finite volume
<mrvn> toppings in the crust
<mrvn> zid: that gets far to thin
<gog> idk if i want ham or spicy pepperoni
<mrvn> half'n'half?
<gog> i'm splitting it with my friend so it might be a half and half anyway
<gog> or we'll get two mediums?
<gog> idk yet
<gog> if she wants vegan then we'll do separate
<zid> are mushrooms vegan
<gog> yes
<zid> idk, they always scream when I cut them up
<gog> those aren't mushrooms
<zid> are onions vegan?
<gog> or they're characters from super mario
<gog> yes onions are vegan
<zid> why do I cry when I dismember them then
<gog> because you need a sharper knife
<zid> I have a good strat for not crying anyway
<mrvn> wear glasses or keep a mouthfull of water
<gog> onion goggles
<zid> nah, it's, don't be such a little bitch
<zid> and stop getting emotionally attached to onions
<gog> zid <3 onion
<zid> I just ate a shallot
<zid> It's like a onion, but smaller and angy
<gog> shallot good
<zid> it was a shallot + cheddar + coleslaw magic wand
<gog> ooh
<Ermine> now it's time for garlic.
<gog> garlic
<gog> i love garlic
<mrvn> you are strange. I eat garlic.
<zid> the pizza places here all give garlic dip
<gog> same
<zid> I thought it would never happen, but I infact found a dipping sauce better than garlic
<zid> I'm not sre you can gt things from iceland in iceland though
<gog> we can
<gog> there's one on suðurlandsbraut
<gog> i don't know if they have prepared foods tho
<gog> it's more of a convenience store than a supermarket
<zid> what's a braut
<gog> highway
<zid> we have lots of sutherlands
<gog> or boulevard
<zid> weirdly the main sutherland is in the very north of scotland
<gog> :thinking:+
heat has joined #osdev
xenos1984 has quit [Ping timeout: 260 seconds]
xenos1984 has joined #osdev
zaquest has quit [Remote host closed the connection]
<mjg> who on this channel loved the wire
<mjg> kazinsal: ? :)
<sortie> mjg, yes let's make second thirty second idle non-pertinent chatter
<gog> me
<heat> i love wired pages hehehehehehehehehehehe virtual memory operating systme development
<mjg> SOLARIS
<heat> 😂😂😂😂😂😂🤣🤣🤣🤣🤣
<gog> 😂
<bslsk05> ​imgflip.com: They don't know - Imgflip
<mjg> sortie: i'm gonna bench sortix vs onyx
<sortie> mjg, look forward to the results of how fast a Sortix installation is compared to an Onyx installation :)
<mjg> does sortix do smp?
<sortie> Does Onyx install? :D
<mjg> also what to do to get a compiler
<mjg> > no SMP support,
<mjg> dude
<mjg> between this and having an installer
<mjg> i think you chose wrong
<sortie> Compiler provided right there out of the box
<j`ey> no bad choices in hobby OS
<gog> that's right
<bslsk05> ​pub.sortix.org: Index of /sortix/release/1.0/builds/
<mjg> so is that dead?
<sortie> mjg, muhahaha one of us has an OS that's actually 1) installable 2) self-hosting by running make in /src out of the box 3) hosts its own website; and the one has the wonders of SMP and ACPI
<heat> <sortie> Does Onyx install? :D
<heat> why am i out here catching strays
<mjg> guilt by association
<sortie> mjg, releases are rare, it's more rolling on the master branch these days, use the nightly releases
<mjg> where.at
<bslsk05> ​pub.sortix.org: Index of /sortix/release/nightly/
<sortie> (Or volatile: If you want the experimental networking and GUI)
nyah has joined #osdev
<heat> i'll probably beat sortix everywhere
<bslsk05> ​pub.sortix.org: installation(7)
<heat> except the features i don't have
<mjg> i hope so!
<gog> sortix is a professional operating system
<sortie> heat, I will move the goal posts so you run very fast but are disqualified
<heat> i've spent way too much time on my mm to just take L's like that
<mjg> sortie: how does one plop extra files into sortix?
<bslsk05> ​pub.sortix.org: release-iso-modification(7)
<sortie> mjg, or use the volatile release instead of nightly, for the experimental networking (comes with ssh and other goodies)
<mjg> godo grief man
<heat> sortix is a fork of openbsd in spirit
<sortie> (curl too)
<heat> manpages for everything
<mjg> i'll try fresh nightly
<mjg> but if anything goes wrong i declare sortix takes a huge L
<mjg> no takebacks
* sortie laughs heat in semantic markup
<heat> mandoc is just a horrible format man
<sortie> One of us has beautiful manual pages
* sortie loves this jovial competition with heat
<heat> i'll let you know my operating system is ✨self-documenting✨
<gog> agile os
<gog> never write a comment
<heat> and by the grace of god all my utilties have a --help
<heat> none of that bsd "what's --help" shit
gildasio has quit [Quit: WeeChat 3.6]
<gog> all of my utilities have /?
<heat> /votekick gog
<zid> I use --?
<gog> i dserve it
<sortie> -? true option
<heat> use either -help or --h
<zid> --help=-?
<heat> --what
<gog> ---help
<zid> --nani-the-fuck
<heat> --dashdashhelp
<sortie> -long-option sends you to prison, --long_option sends you to the bad place
<heat> --lc_corp_moment
<sortie> heat, obey
<zid> --pause -> --yamettekudastop
<mjg> yo, sorts, why you participating in 3rd non-pertinent idle chhatter
<heat> sorts?
<mrvn> food
<heat> it's mr. sorts for you
* gog mews at food
<mrvn> MacHeat
<heat> MacHeat were the old macbooks
<heat> not anymore due to super good very fast arm64 apple silicon
<gog> m'heat
* gog tips fedora
<heat> h'eat
<sortie> mjg, you a cop?
<heat> worse, freebsd committer
<mjg> emacs.tix.tar.gz
<mjg> wtf sortix
<mjg> do you even boot
<sortie> Please it's xz
<mjg> gcc 5
<mjg> old school
<heat> very haiku
<sortie> It's a fine vintage
<mjg> how to do any networking
<heat> >x86_64-onyx-gcc --version
<mjg> what nic do you support? virtnet ? ne2000?
<heat> x86_64-onyx-gcc (GCC) 12.2.0
<heat> e1000
<mjg> no ifconfig
<sortie> mjg, the cdrom boots of nightly are slow because the bios loads the live environment off the disk, but the hard-disk installations are lighting fast (been experimentally optimzing it down to like a third of a second)
<mjg> not what i asked dawg
<mjg> howto networking
<sortie> The gcc is at 5.2.0 because honestly I have not had the time to update it, it's a bunch of work to upgrade the patches, and it's working fine enough
<heat> do the qemu thing and give it e1000
<heat> it should Just Work(tm)
<heat> ping 1.1.1.1 for a quiq test
<sortie> mjg, like I said, volatile has networking. You're using nightly which is not volatile. The networking is experimental and not merged just yet to master. That's my next task.
<mjg> huh
<mjg> where is volatile then
<heat> the gang downloads another iso
<sortie> mjg, s/nightly/volatile/g on your URL
<bslsk05> ​pub.sortix.org: Index of /sortix/release/volatile/builds/
<zid> mod_rewrite all the things
fkrauthan has quit [Quit: ZNC - https://znc.in]
fkrauthan has joined #osdev
<sortie> The volatile has all the goodies you'd want, networking, a basic GUI, ssh, curl, wget, etc.
nyah has quit [Ping timeout: 248 seconds]
<sortie> It's not fully stabilized and finished, but pretty close
fkrauthan has quit [Client Quit]
rwxr-xr-x has quit [Ping timeout: 256 seconds]
fkrauthan has joined #osdev
nyah has joined #osdev
<mjg> sortie: looks like it hung trying to create a partitin
<heat> how u running it
<sortie> What virtual machine?
<mjg> virtualbox
<sortie> VirtualBox's ATA implementation has a weird regression these days that cause corruptions
<mjg> oh?
<sortie> At least affecting Sortix. Unclear if it's my bug. It worked back when I made 1.0 and these days it tends to fail when making the initial installation early on
<heat> i've never seen issues
<sortie> qemu-img create -f qcow2 sortix.hdd 1G && qemu-system-x86_64 -enable-kvm -vga std -m 1024 -cdrom sortix.iso -boot d
<heat> probably a sortix bug
<sortie> heat, it's likely a bug on my end yeah, but a regression nonetheless in the sense it used to work and now it tends not to
<sortie> Qemu works just fine
<sortie> mjg, alternatively just use AHCI instead of ATA
<mjg> sortie: switched to ahci, same shit
<mjg> no wait, it went through this time
<sortie> AHCI is much superior anyway :)
* heat laughs in virtio-blk and nvme
<mjg> how do you assign ip by hand
<mjg> dhclient says it gets an ip offer but does not do anything with it
<sortie> You can use ifconfig(8) directly (see man ifconfig)
<mjg> wut
<mjg> a manpage
<sortie> Curious to see your /var/log/dhclient.em0.log if it's not successful
<mjg> i did try the standard invocation, the keyword 'address' is not necessary man
<sortie> ifconfig varies between operating systems
<mjg> so how do i start sshd
<sortie> If you did an installation, it asked you whether you wanted to start it on launch
<sortie> You can see with e.g. pstree or /var/log/sshd.log if it is started
<mjg> i tried the install, it failed at the end with error code 2 or somethin
<mjg> rebooted to live
<mjg> i'm not going to debug that bit
<mjg> so how do i sart it by hand
<sortie> Ah if you're in the live environment, you can pick "Start ssh server" from the advanced bootloader menu, or run /sbin/sshd manually
<mjg> nice
<sortie> Error code 2 is used for a lot of purposes, curious what message you got
<mrvn> sortie: how do you start services? inittab? sys-v init? systemd?
<mjg> did not save it, sorry
<bslsk05> ​pub.sortix.org: init(8)
<bslsk05> ​pub.sortix.org: init(5)
<sortie> mrvn, wrote my own init system, proper dependency tracking with readiness, inspired by what systemd does right, done much simpler
<mjg> sortie: i can ping it, but trying to connect to ssh gives me nothing
<mjg> ... and now it worked. looks ike some timeout
<mjg> what's the password
<mrvn> mjg: the password is secret, obviously
<sortie> The live environment root user does not have a password set
<sortie> You can set one by running passwd
<mjg> i just did
<mjg> still does not let me in
<mjg> is root login allowed?
<mjg> i don't see /etc/sshd
<mjg> grep shows /etc/default/sshd_cnofig with root login disabled
<mjg> is that what is being used?
<heat> can't wait for the new openssh version
<heat> login disabled by default
<heat> no holes in the default install in a heck of a long time
<sortie> mjg, this is standard sshd, by default password login for root is provided. To enable it, cp /etc/default/sshd_config /etc/sshd_config and change "PermitRootLogin prohibit-password" to "PermitRootLogin yes"
<mjg> enabled it and Huped the gtuy
<mjg> i'm in
<mjg> do you have tmpfs?
<heat> livecd runs on tmpfs
<sortie> The live environment is one big tmpfs
<sortie> mjg, btw you went about this the hard (manual) way, the easy way to get a Sortix VM with set up with your public key auth, is to amend the .iso with your keys: https://pub.sortix.org/sortix/release/volatile/man/man7/release-iso-modification.7.html#SSH_Into_Live_Environment
<bslsk05> ​pub.sortix.org: release-iso-modification(7)
<mjg> root@sortix ~ # cc main.c fstat1.c
<mjg> mmap: Invalid argument
<mjg> root@sortix ~ # ./a.out
<mjg> bruh
<mjg> game over
<heat> sortix doesn't have MAP_SHARED
<heat> (at least for files, can't remember what happens with anon)
<sortie> Yeah no shared memory yet. I have a partial implementation but it destabilized volatile, so removed it
<mjg> this is for anon
<heat> also no CoW, MAP_PRIVATE immediately reads in the file manually
<mjg> i'm not patching this to work around the problem
<mrvn> sortie: my linux livecd runs on dmmapper (lvm) and the installer does a pvmove to transparently install in the background.
<mjg> verdict: Onyx faster!
<heat> i win
<heat> ez
<mjg> gg heat
<sortie> You're moving the goal posts!
<mjg> i can do whatever i want dawg
<mjg> it's my contest
<heat> good thing it's the world cup rn
<mjg> fix it up and hit me up
<mjg> cc main.c openro1.c
<mjg> get this operational
<mjg> then we will talk
<heat> openro should work
<heat> oh wait yeah, the main area :/
<mjg> until then onyx > sortix
<heat> in all honestly i'm very unlikely to lose against any hobby OS
<sortie> Can I retcon no shared memory as a hardening technique due to the common misuses?
<mjg> i can make you lose against netbsd if you want
<heat> yes sounds very openbsd of you
<mjg> sortie: depends, are you a descaendant of theo?
<heat> mjg, bring it on bitch
<mjg> OH
<sortie> mjg, it is my stated goal that people casually dismiss me like they'd do with NetBSD, Haiku, and Hurd
<heat> I'M ROCK FUCKING HARD RIGHT NOW
<heat> STAY HARD
<mjg> actually i checked haiku
<mjg> it is slow of course, burt the above works!
<sortie> What does that program even test?
<mjg> does it matter?
<heat> open performance
<mjg> key is that it sets up a shared mamoery area to collect results from processes
<mjg> as for what is being benched, whatever you plug into it, and there is quote a lot
<mjg> see will-it-scale
<heat> openro1 creates a tmpfile for each thread/process and does open(path) + close on each loop
<sortie> Ah, to test contention on /tmp?
<heat> to test contention when opening in general
<mrvn> You, know, that thing that you want to be so slow that people will fix their code to not do such a silly thing.
<mjg> dirconc.c:(.text+0x78): undefined reference to `random'
<mjg> anything i can do baout this bit?
<sortie> mjg, yes, use arc4random
<mrvn> mjg: int random() { return 4; // determined by a fair dice }
<mjg> does not have to be contention, you can run singl tthreaded
<heat> mrvn, yes.. such a silly thing as filesystem access
<sortie> If you're collecting results from multiple processes at the end, might as well stash them in a file or or a pipe or just write them out
<heat> sortie, that has a lot of overhead
<sortie> heat, not if it's done at the end?
<heat> but it's not done at the end
<mrvn> heat: hammering the filesystem. Not something any program should just do.
<sortie> Each process does its thing, after a while they write the end results?
<heat> mrvn, it's not *a* program, it's the whole system that's being bench
<heat> ed
<mjg> i'm done with sortix for the time being
<heat> sortie, nope
<mjg> sortie: creates an unkillable proc
<mrvn> heat: do you test creating the file each loop and working on an existing file?
<sortie> mjg, I assume you tried with -KILL
<mjg> i tried with -9
<mjg> and no amount of ^ does the trick
<sortie> Signal numbers may vary
<mjg> ^C even
<sortie> Use -KILL
<mrvn> mjg: cltr-\?
<heat> mrvn, huh?
<mjg> sortie: no dice
<sortie> SIGKILL is the only thing that the kernel promises will definitely kill a thread
<sortie> So if it doesn't kill it, you found a bug :)
<mjg> sortie: rly, dirconc is designed to make shit bad
<heat> dirconc is great for constipation in general
<mjg> sortie: look, if you never ran dirconc on your os, it is virtually guaranteed it will cause a problem
<sortie> Not surprised dirconc locked it up :)
<mrvn> heat: openro1 creates a tmpfile for each thread/process and does open(path) + close on each loop: Is that (create + open) on each loop or create + (open on each loop) or both?
<sortie> The tmpfs implementation is a bit old and the locking may not be perfect
<mjg> sortie: messed up haiku as well :p
<heat> mrvn, create + (open + close on each loop)
<sortie> mjg, but the ext2 filesystem may work
<mjg> i did my duty
<heat> i'm fairly sure there are some other tests to test creation
<sortie> (since it's a bit more top-level atomic)
<mjg> bug reported with repro steps
<sortie> Yeah typing up a bug
<heat> ext2 will get dragged around in performance
<mrvn> heat: should favour an LRU implementation in the FS heavily then.
<heat> due to plain old filesystem access (although granted, cached) plus IPC
<mjg> sortie: so you renumbered KILL to something else?
<heat> mrvn, LRU of what?
<mrvn> heat: path walking and file opening
<heat> it heavily favours a proper cached walk, yes
<sortie> mjg, looks like I did define it to 9, but no standard requires that
<mjg> pretending to be unix does
<sortie> That wholly depends on what I can get away with, mjg
<heat> openbsd should scramble the signal numbers for the next release
<mjg> heat: select their meaning at boot time at random
<heat> LMAO
<mjg> that will show it to these pesky people who use them
<\Test_User> random for each run of *kill
<sortie> I am a de-facto Unix-spirit system, I do some things differently as it makes sense, but ultimately it's about what the compatibility pressures makes possible/infeasible
<mjg> matching signal num,ber for KILL sounds like a nobrainer dawg
<mjg> just sayin
<mjg> anywya rating the system now
<bslsk05> ​gitlab.com: Concurrent directory operations may deadlock process (#810) · Issues · sortix / Sortix · GitLab
<heat> not knowing what signal number = signal does help reduce rop gadgets
<sortie> mjg, I mean, I did that, but the set of signals also vary. Linux got a bunch of weird ones
<heat> like what?
<mjg> i'm aware of funny discrepancies
<mjg> but 1, 9 and 11 are not negotiable dawg
<mrvn> unix lack a good way for libraries to claim a signal
<sortie> Generally signal handling are bad in the first place but yeah, they require the main application to cooperate with libraries
chovelyoukai has quit [Remote host closed the connection]
chovelyoukai has joined #osdev
<mrvn> sortie: and there is only USR1 and USR2. if you need a 3rd signal you are screwed
<heat> you have a whole host of realtime signals
<heat> around 32 of them in fact
<sortie> The problem with those is that USR{1,2} are deadly by default
<mrvn> if you have rt
<sortie> So if you send them tooo early
<heat> which you do, because it's 2022
<sortie> But seriously, signals are usually bad, they have their uses, try not to use them unless you must
<heat> rt signals have been 100% required in linux for ~20 years
<sortie> mjg, thanks for trying out Sortix :)
<gog> trianglix is where it's at+'
<heat> now you try freebsd
<heat> step 1: how do I install GNU coreutils and glibc?
<sortie> trianglix is the true UI
chovelyoukai has quit [Ping timeout: 268 seconds]
<mjg> heat: it actually happens to be possible
<heat> i am aware
<heat> debian freebsd
<mjg> no
<mjg> just create a jail
<heat> cringe
<zid> It's called BSD because it's homework from berkley systems and they affix the grade letter to it
<gog> BSD+
<heat> the F in FreeBSD stands for F
<zid> No it's called freebsd because it's under the free marking system, the work will be graded when the student feels it will pass the critera only
<gog> press F to pay respects
<zid> they're not going to submit until it's worth at least a C+
<heat> a C++
gildasio has joined #osdev
GeDaMo has quit [Quit: I'm going to show these people something you don't want them to see. I'm going to show them a world without you.]
puck has quit [Excess Flood]
puck has joined #osdev
terminalpusher has joined #osdev
* geist yawns
<heat> geist, oh yeah qemu is gaining tcg avx support this next release
<heat> it's huge
<geist> ah woot. all the way up through 512?
<heat> I won't need to pretend that I care about SSE-only systems
<heat> hmm
<heat> lets see
<zid> avx1 is the best avx
<heat> no
<zid> (it's an unbiased opinion based on me only having avx1)
<heat> "TCG now supports AVX / AVX2 / F16C / FMA3 / VAES instructions. "
<gog> avx4096
<heat> that's virtual memory gog
terminalpusher has quit [Remote host closed the connection]
<geist> heat: is this 7.2 that you're referring to? it's a few RCs in
<heat> yes
<bslsk05> ​wiki.qemu.org: ChangeLog/7.2 - QEMU
<heat> also apparently really huge performance improvements on 9pfs
<zid> 9pfs finally at 9fps?
<geist> ayway all this aside i'd assume there are bugs in the first implementation of AVX software, so probably wouldn't get too excited about it
<geist> a little surprised it shows up in a point release, but dunno what in qemu world constitutes a full release vs point
<heat> i mostly just need the memcpy bits soo
smach has quit []
Burgundy has quit [Ping timeout: 260 seconds]
xenos1984 has quit [Ping timeout: 256 seconds]
ZombieChicken has joined #osdev
xenos1984 has joined #osdev
ZombieChicken has quit [Quit: WeeChat 3.6]