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
goliath has joined #osdev
goliath has quit [Client Quit]
<zid> gog`: ur backtick makes u look lopsided
<zid> Hmm should everyone get a scathing critique?
netbsduser`` has quit [Ping timeout: 258 seconds]
nickster has quit [Quit: The Lounge - https://thelounge.chat]
nickster has joined #osdev
agent314 has quit [Ping timeout: 240 seconds]
agent314 has joined #osdev
[itchyjunk] has quit [Ping timeout: 252 seconds]
[itchyjunk] has joined #osdev
nickster has quit [Quit: The Lounge - https://thelounge.chat]
agent314 has quit [Ping timeout: 255 seconds]
agent314 has joined #osdev
nickster has joined #osdev
carbonfiber has quit [Quit: Connection closed for inactivity]
nickster has quit [Quit: The Lounge - https://thelounge.chat]
nickster has joined #osdev
<bslsk05> ​'Sorry, I saw this elsewhere and had to' - ''
blockhead has quit []
heat has quit [Quit: Client closed]
[itchyjunk] has quit [Remote host closed the connection]
GeDaMo has joined #osdev
agent314 has quit [Ping timeout: 255 seconds]
agent314 has joined #osdev
MarchHare has quit [Ping timeout: 260 seconds]
zaquest has joined #osdev
netbsduser`` has joined #osdev
netbsduser`` has quit [Ping timeout: 255 seconds]
gbowne1 has quit [Quit: Leaving]
duderonomy has joined #osdev
Burgundy has joined #osdev
agent314 has quit [Ping timeout: 240 seconds]
agent314 has joined #osdev
agent314 has quit [Read error: Connection reset by peer]
agent314 has joined #osdev
agent314 has quit [Read error: Connection reset by peer]
agent314 has joined #osdev
Starfoxxes has quit [Ping timeout: 258 seconds]
agent314 has quit [Read error: Connection reset by peer]
Starfoxxes has joined #osdev
agent314 has joined #osdev
irl25519 has joined #osdev
gog has joined #osdev
zxrom has quit [Quit: Leaving]
irl25519 has quit [Client Quit]
kof123 has quit [Ping timeout: 252 seconds]
agent314 has quit [Ping timeout: 260 seconds]
agent314 has joined #osdev
ThinkT510 has quit [Ping timeout: 255 seconds]
ThinkT510 has joined #osdev
ThinkT510 has quit [Read error: Connection reset by peer]
Affliction has quit [Quit: Read error: Connection reset by beer]
Affliction has joined #osdev
ThinkT510 has joined #osdev
netbsduser`` has joined #osdev
FireFly is now known as Luci-ghoule
k0valski1889162 has joined #osdev
Left_Turn has joined #osdev
goliath has joined #osdev
zxrom has joined #osdev
Terlisimo has quit [Quit: Connection reset by beer]
Terlisimo has joined #osdev
agent314 has quit [Ping timeout: 255 seconds]
agent314 has joined #osdev
agent314 has quit [Ping timeout: 258 seconds]
Cindy has quit [Remote host closed the connection]
bnchs has joined #osdev
kof123 has joined #osdev
Turn_Left has joined #osdev
Left_Turn has quit [Ping timeout: 264 seconds]
vai has joined #osdev
[itchyjunk] has joined #osdev
heat has joined #osdev
<heat> formerly-twitter, zfs suckzzzzzzzzzz
<heat> zfsuckzzzzzzzzzz
<mcrod`> ok
<mcrod`> i admit it
<mcrod`> i love gitlab
<mcrod`> what does suck though is the really really nice features are paid
<heat> i prefer github
<mcrod`> i know what you prefer
<mcrod`> and i'm here to tell you that you are a SHIT and WRONG
<heat> yeah, email patches
<heat> git format-patch + git send-email >> shitlab
<heat> but seriously to me gitlab feels like the backrooms of opensource
<zid> gitlab is fuckugly
<mcrod`> the one thing I don't like about github (and probably gitlab) is the poor quality of accepting a merge request
<mcrod`> but, that depends on how formal your project is and how much you care
<zid> looks like the stupid t-shirt thing has made gitlab close signups, hah
<mcrod`> i dunno, at work gitlab is completely fine
zxrom has quit [Quit: Leaving]
<Luci-ghoule> it'd be nice if they could make it usable without JS
<Luci-ghoule> like, at least let me view raw files
<heat> why
<Luci-ghoule> ..because it's nice to be able to read things on websites?
<mcrod`> are you one of those people who run noscript on everything
<mcrod`> because, rest assured, you can read a file without raw capability on gitlab
<mcrod`> although i don’t disagree the feature would be nice
<Luci-ghoule> I mean by default yes, I'll generally happy to enable JS quickly enough but like
<Luci-ghoule> it's just really quite disappointing when you just get a blank page y'know?
<Luci-ghoule> graceful degradation and all that
<mcrod`> yeah i get it
<Luci-ghoule> maybe they fixed it, gitlab used to just show a blank white page without JS
<Luci-ghoule> and not that long ago
<Luci-ghoule> and like all the other forges seem happy to at least show me a file if I click a link to one
<puck> zid: they don't even do t-shirts this year, no?
<heat> i dont get noscript users
<heat> like, it's a big part of the web, clearly no one wants to roll that back, so why?
<heat> you're cherry-picking what parts of the web you do like and then hoping everything works
<heat> it reminds me of people staunchly against HTML email
<puck> honestly the gitlab thing where they lazy-load most of the file is bullshit
<heat> with the little ribbon and everything
<puck> it takes like 20 seconds for it to actually load the file because it first needs to load every single other resource on a bloated webpage
<puck> i feel complicated about JS but in a way i doubt we can get back to "normality"
<heat> i mean yeah the web page sucks but that's not on JS, github also has JS and it works fine
<puck> gitlab feels clunkier to use too
<Luci-ghoule> well as mentioned I'm typically quick to enable JS, and like, I don't have anything against it per se?
<Luci-ghoule> (quite the opposite)
<heat> so why do you have it off by default?
<Luci-ghoule> but there's also just so much unnecessary bulk from ads and such these days, so I'd rather enable on a site-by-site basis
* Luci-ghoule shrugs
<heat> oh i mean, i have ublock origin and privacybadger and those do a good job
<puck> that handles "ads" and not the "and such" :p
<Luci-ghoule> I mean it's probably enough
<Luci-ghoule> I don't feel terribly strongly about it
<zid> I wish I could enable it site by site
<zid> I used to run noscript that allowed *.domain.com
<zid> and that still broke 90% of websites, even whitelisting common node/whatever 3rd party nosts
<zid> hsots
<zid> and noscript itself was janky
<puck> i feel noscript i used recently was .. fine?
<GeDaMo> I use NoScript, I don't have any problems with it
<zid> It was buggy and had different behavior on "not installed" vs "not supposedly doing anything"
<zid> for the several years I used it
<puck> xul or webextension?
<zid> xul, at the time
bnchs is now known as Cindy
<Ermine> fkn network shares in vindovs desktop
<netbsduser``> i love how mozilla got animated with delight by the buzzwords of the late 90s/early 2000s and just implemented COM and COM'ified everything
<kof123> https://gigamonkeys.wordpress.com/2009/09/28/a-tale-of-two-rewrites/ he tells me in a recent email that he tried, in 1998, to get Netscape to open source both the 4.0 and 3.0 code bases but they wouldn’t go for it.
<bslsk05> ​gigamonkeys.wordpress.com: Duct tape context: A tale of two rewrites | A billion monkeys can't be wrong
<kof123> this is just to say, that probably started earlier :D
bauen1 has quit [Ping timeout: 255 seconds]
vai has quit [Ping timeout: 255 seconds]
xenos1984 has quit [Ping timeout: 246 seconds]
xenos1984 has joined #osdev
<zid> https://i.imgur.com/lreZ0Pc.mp4 found the new fan for the rpi5
<bslsk05> ​i.imgur.com <no title>
xenos1984 has quit [Ping timeout: 258 seconds]
<heat> yeah
<heat> i'd say it also works on the intel 14th gen processors but those will probably need to be water cooled
<heat> formerly-twitter, btw i was just going to close the zfs issue
<heat> they clearly don't care, their filesystem's design sucks and is apparently anti-sanity
<netbsduser``> what is the issue?
<bslsk05> ​github.com: Creating + writing to a file + stat() in quick succession returns bad st_blocks · Issue #13991 · openzfs/zfs · GitHub
<netbsduser``> what an interesting observation, i wonder what the cause of it is
<heat> someone here hinted that it may be the "ZIL", whatever that is
<netbsduser``> the intent log is what some call a journal
<netbsduser``> oh it's just st_blocks
<netbsduser``> i wouldn't worry about that
xenos1984 has joined #osdev
<netbsduser``> fs implementation detail which stat exposes for the curious
<heat> i would because it breaks unit tests and is clearly faulty behavior
<heat> when you fucking fsync or fdatasync and the inode doesn't fully sync
<heat> even *sync(2)* doesn't do shit
qubasa has joined #osdev
<netbsduser``> if it's recorded in the intent log then it will be replayed safely, i can only presume
<heat> i can't presume anything, fsync is fsync, sync is sync, etc
<heat> i'd be pissed if ext4 did this as well (hint: it does not)
FreeFull has joined #osdev
<formerly-twitter> heat: just wait for amotin
<formerly-twitter> i doubt he is going to flake out with some bs, but he may need time to respond
<heat> keep calm and ext4
<heat> at this point i would rather run a system on an ntfs root partition
<formerly-twitter> yo fuckwad
<formerly-twitter> speaking of filesystems
<formerly-twitter> and operating systems
<formerly-twitter> i recnetly started doing a little of how are they doing vs 20 find(1) progs each rolling over 1 million inodes
<formerly-twitter> separate trees
<heat> speaking of operating systems? you're offtopic mate
<formerly-twitter> so far netbsd and dragonflybsd ar pretty horrid
<formerly-twitter> i can't do an immediate comparison to linux becusae it does not have an explicit vnode limit
<heat> you mean like 1 million "cached" inodes?
<formerly-twitter> openbsd i/m not even going to try
<formerly-twitter> no
<formerly-twitter> literally 20 * 1 million on disk
<formerly-twitter> which you *can't* fit all in ram at the same time
<formerly-twitter> so some churn vs freeing has to take place
<heat> so what's the problem with linux?
<formerly-twitter> oh i think i misread
<formerly-twitter> so the bsds historically have an explicit limit on vnodes
<heat> NINODES moment
<formerly-twitter> and you whack existing ones once you get there
<formerly-twitter> linux does not do that
<heat> well yeah, but what's the problem?
<heat> the shrinker will shrink a buncha caches, inodes will certainly be freed
<formerly-twitter> the problem is i can't do a valid perf comparison in that ase
<formerly-twitter> without a lot of fuckery to make it even
<heat> it will never be even
<heat> even between BSDs i assume
<formerly-twitter> for example netbsd chockes itself on that vnode reclamation thing
<formerly-twitter> while freebsd does not
<formerly-twitter> so i can legitimately say the latter is faster
<heat> linux has no inode limit because the dcache is the master of the icache
<heat> as long as dentries exist they'll hold refs to inodes, hence no reclamation is possible
<formerly-twitter> it totally is possible
sortie has quit [Ping timeout: 255 seconds]
<heat> how boss
sortie has joined #osdev
sortiecat has joined #osdev
sortiecat has quit [Client Quit]
<formerly-twitter> if what you were saying was true, the kernel would die or stall indefinitely
<formerly-twitter> once you get too many dentries
<formerly-twitter> clearly it can whack them
<heat> i'm saying it has shrinkers
<formerly-twitter> there you go
<heat> but setting an explicit inode limit would be futile since dentries always hold refs to the icache
<formerly-twitter> it would not
<formerly-twitter> you would just pick what to whack
valshaped7424880 has quit [Quit: Gone]
<formerly-twitter> you do realize in the bsds namecache entries also hold a ref
valshaped7424880 has joined #osdev
<formerly-twitter> all you need is a backpointer to whack them, if particular entry makes sense
<heat> how do you do reverse inode <-> dentry matching?
<formerly-twitter> you would have a fucking list of entries
<heat> i'm not going to say anything but it sounds FUCKING PESSIM
<formerly-twitter> in the inode
<formerly-twitter> i have not checked, but i presume they already have one
<heat> linux inodes?
<heat> oh ew they have one
<formerly-twitter> there you go mofo
<formerly-twitter> presumably this is how you whack the fuckers on unmount
<heat> i dont have unmount yet but i would clear the dcache of dentries and only then clear the superblock's icache
<heat> no need for backpointers there
<mcrod`> hi
<heat> formerly-twitter, anyway do you have an ez repro for that? might be a fun stress test for onyx
<formerly-twitter> you would need to scan the entire fuckign dcache
<formerly-twitter> i do
<formerly-twitter> can it put it up later
<heat> btw did you finish up benching onyx vs other crap systems?
<formerly-twitter> no, got "distracted" by that vnode problem
<formerly-twitter> it happened to matter
<heat> fwiw the onyx i gave you does not do inode or dcache reclamation yet
<mcrod`> somehow
<mcrod`> I forgot that google domains is closing
<mcrod`> now I'm slightly fucked
<heat> why
<mcrod`> because historically, I've had nothing but problems with squarespace
<heat> then don't use squarespace
<heat> i use cloudflare
<mcrod`> i'm transferring
<mcrod`> don't worry my son
* mcrod` pet
<heat> static inline void __d_add(struct dentry *dentry, struct inode *inode)
<heat> woops
<heat> /* inode->i_lock held if inode is non-NULL */
<heat> i like these great lock annotations in the form of comments
<heat> most formally verified C code
<heat> as a typical C programmer would put it: oops, i forgor💀
<mcrod`> you're definitely a zoomer.
<heat> i am, did u forgor
zxrom has joined #osdev
<Ermine> I forgor too
<heat> 💀
kof123 has quit [Ping timeout: 252 seconds]
kof123 has joined #osdev
<formerly-twitter> heat: l mao yer suck bro
<formerly-twitter> at least not as bad as that vnode paper
<formerly-twitter> 20 minutes(!!!!) grace time
<formerly-twitter> lmao
<formerly-twitter> bbl to shit on you and with c0d3z uploaded
<heat> i found a linox bog
<heat> it raises a question tho, does no one use kmsan builds
<Ermine> kernel memory sanitizer?
<heat> yep
<Ermine> how do i check if my kernel has kmsan on ?
<heat> it does not, it's a debug feature
GeDaMo has quit [Quit: That's it, you people have stood in my way long enough! I'm going to clown college!]
<zid> That seems bad
<heat> Ermine, ugh it's just an OOM bug, bad error reporting...
<zid> oh right I ran that ryzen tool to boost the cores eh, at least I found out that it's core 8
<geist> hewwo fronds
<zid> when I reboot, remind me to knock some adjustment off that core..
<geist> how is your lovely sunday, unless it's monday in which case i'm sorry
<zid> monday is great, monday is honzuki
<zid> it's sunday instead sadly
<zid> heat did you remember to watch mf ghost ep 3
<heat> sorry i only like mf doom
<heat> mf ghost is kinda cringe ngl
<zid> It's a sequel to initial D, you have to watch it
<zid> it's part of CULTURE
<zid> It doesn't have enough eurobeat is the only problem I have with it so far
bauen1 has joined #osdev
<heat> i forgot to run checkpatchhhhh
<heat> anyway i was trying to fix an ntfs3 bug but i can't repro, yay
divine has quit [Ping timeout: 258 seconds]
kof123 has quit [Ping timeout: 252 seconds]
divine has joined #osdev
gbowne1 has joined #osdev
greaser|q is now known as GreaseMonkey
<mcrod`> it's nice when you fix a problem that you've been bashing your head against forever
<zid> I still have my problem, still playing ED every waking moment
<mcrod`> for what it's worth
<mcrod`> my bug was not rounding
<mcrod`> :(
MiningMarsh has quit [Quit: ZNC 1.8.2 - https://znc.in]
<heat> >Everything is a file in NTFS
<heat> unix people are ERECT right now
<heat> it'd be ideal if ntfs stored metadata in plain text
<heat> the driver would just be a shell script with a bespoke regex expression and a bespoke awk invocation
MiningMarsh has joined #osdev
<netbsduser``> they pilfered a lot of ideas from unix for windows nt
<netbsduser``> probably a lot of unix ideas came after they hired rick rashid, one of the creators of the Mach kernel
<heat> they should've abandoned SEH in favor of #!/bin/sh -e
FreeFull has quit []
<nortti> wasn't microsoft the biggest unix vendor at a time in the 80s, too?
<heat> i'm not sure about the biggest, but they had xenix yeah
<heat> the original i386 UNIX bringup work had microsoft copyright all over it
<mcrod`> jesus why does doxygen generate such shit output
<heat> still does in illumos
kof123 has joined #osdev
Burgundy has quit [Ping timeout: 255 seconds]
Turn_Left has quit [Read error: Connection reset by peer]
dude12312414 has joined #osdev
gog has quit [Quit: byee]
heat has quit [Read error: Connection reset by peer]
heat_ has joined #osdev