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
gog has joined #osdev
vdamewood has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
gog has quit [Ping timeout: 260 seconds]
nyah has quit [Ping timeout: 260 seconds]
thinkpol has quit [Remote host closed the connection]
thinkpol has joined #osdev
pretty_dumm_guy has quit [Quit: WeeChat 3.3]
Dreg has quit [Read error: Connection reset by peer]
Dreg has joined #osdev
srjek|home has joined #osdev
isaacwoods has quit [Quit: WeeChat 3.3]
sdfgsdfg has joined #osdev
sdfgsdfg has quit [Quit: ZzzZ]
smeso has quit [Quit: smeso]
sdfgsdfg has joined #osdev
xing_song has quit [Remote host closed the connection]
xing_song has joined #osdev
smeso has joined #osdev
<geist> MSG computers
happy-dude has joined #osdev
novasharper has joined #osdev
paulusASol has joined #osdev
Irvise_ has joined #osdev
asmwfi[m] has joined #osdev
<geist> MSX
<geist> discuss.
terrorjack has quit [Quit: The Lounge - https://thelounge.chat]
terrorjack has joined #osdev
<zid> La Mulana <3
<zid> I sneaked my advent day 9 in just in time, it's like an hour before the next one
<zid> https://gist.github.com/zid/786cc262db42a2545e79c9b31a5b4cd7 I got the 2nd part correct the first time it compiled.. surprising
<bslsk05> ​gist.github.com: 2021-day9.c · GitHub
<zid> I definitely didn't write it in a decent way but meh
zhiayang has quit [Quit: oof.]
srjek|home has quit [Ping timeout: 252 seconds]
zhiayang has joined #osdev
ElectronApps has joined #osdev
<geist> hey no pasting of solutions!
<geist> put spoiler tags
<geist> (actually my fault for using a client that inlines those links)
<klys> hurry hurry :)
<zid> 3 minutes left geist :p
<zid> less than 2!
<klange> Not like there's some limit on when you can submit...
<kazinsal> someone I know has been doing advent of code in awk and by day 4 they were regretting it
<zid> meh I might as well start on day 10 I suppose..
<moon-child> kazinsal: heh, same
ElectronApps has quit [Remote host closed the connection]
ElectronApps has joined #osdev
<klange> 137... that's my best second part placement so far >_>
<eryjus> no competition here i see
<klange> I got 101 on the first part of day 8, so you can imagine my annoyance.
<zid> hang on I am still reading :p
<moon-child> I got some place in the low hundreds on I think day 2
<moon-child> kinda lost interest and haven't done any problems since
<klange> Most of the folks who stick with it have some additional challenge they've made for themselves, like doing it in a new language or a particular unsuitable one.
<zid> Mine segfaults I win
<zid> best I ever got was 36th
<zid> not bad for C :p
xing_song has quit [Read error: Connection reset by peer]
xing_song has joined #osdev
<gorgonic-> I was doing mine in ocaml and after I spent like four hours writing support code because I refused to use the batteries included library I lost interest also
gorgonic- is now known as gorgonical
<gorgonical> why was my name mangled? anyway
<kazinsal> IRC works in mysterious ways
<gorgonical> ocaml is fun but I don't know it very well and I have arguably more productive things to be doing than spending 4+ hours on increasingly difficult problems. I was enjoying it for around 2 hours per problem
<zid> Just do them in 4 minutes smh
<gorgonical> I seriously felt that when I was doing the bingo board one. Like... this would be trivial to do in memory with ptrs, but instead I was modeling the whole bingo board like a dingus because I didn't know how else to do it in ocaml
<zid> I'm completely baffled by this code bug, whatever it is
<zid> I'm completely baffled by this code bug, whatever it is
<zid> You are not my terminal
<Mutabah> /home/zid$
<zid> It's seemingly getting inside an if that I am printfing the value for the test which it does not match
<gorgonical> Could it be coherence?
<zid> if(blah) return 1197; printf("made it\n");
<zid> Made it
<zid> line 3, score 1197
<geist> wat.
<geist> well what's after the printf?
<zid> It mirrors the value in that exact statement too
<zid> oh shit
<zid> thanks geist
* zid adds 'return 0;'
<geist> riiiight was just asbout to say if you had no return then it returned what was ever in rax/x0/etc
<geist> and it's possible i had pre-loaded it with 1197 because
<zid> How does that end up in rax I wonder
<zid> it obviously DOES
<geist> depends on the codegen really
<zid> but I wonder how.. maybe I will copy the binary
<gorgonical> compiler shenanigans
<geist> having no return shold be a big fat warning though that you shouldn't ignore
<geist> but since that is UB, the compiler can do what it wants
<Mutabah> Annoyingly, it's not on by default in far too many compilers
<zid> I have -W but got no warning :/
<gorgonical> I have learned -
<zid> or I printf spammed too much and missed it idk
<gorgonical> Wall -Wextra is my friend
<zid> I don't like wall wextra while golfing though
<gorgonical> Really I just wish c was ada actually
<geist> so the code could have been something like 'load rax, 1179; cmp blah; beq no_return; ret; no_return: printf; ret;
sdfgsdfg has quit [Quit: ZzzZ]
<geist> or... actually probably
<zid> there are 4 of those exact ifs and 1197 is the 2nd of the 4, geist
<geist> 'load rax, 1197; cmp blah; bne exit_function; printf; exit_function: ret;'
<zid> which made it double weird
<geist> sinceit's common for the compiler to branch to a common exit path like that with the return value already preloaded in a register
<zid> anyway, it worked after that
<geist> reminds me, i found a fun new undefined compiler behavior at work the other day
<geist> had a function that had an infinite loop. basically a panic 'kill the kernel right here and dont continue' function
<geist> looked something like die() { disable_ints; for(;;); }
<geist> guess what clang decided to do with that?
<Mutabah> Remove it entirely
<geist> close!
<Mutabah> Probably made the entire function `UD2`?
<klange> disable ints _after_ the loop?
<Mutabah> No loop?
<geist> any more bets?
<klange> no loop would have been my next
<geist> simon says: no loop, and just falls off the end of the function
<geist> so doubleplus bad
<geist> die: call disable_ints; ....
<Mutabah> Infinite loops are UB yo
<klange> not anymore
<geist> yep. it was an oversight *anyway*, should at the minimum stick a WFI or yield or something in there to just not burn power
<klange> c11 says if the loop expression is a constant you're fine; the older specs were based solely on the contents of the loop - loop does nothing? no loop!
<geist> ayway the compiler deciding to just fall off the end of the function is a real jewel
<klange> in a discussion about asserts over in a discord, someone posted a code sample that accidentally missed a return statement despite have a return type
<geist> so protip: no more naked infinite loops. just say no!
divine has quit [Read error: Connection reset by peer]
<geist> at least put a asm("nop") in it
<klange> like `int foo(int num) { if (!num) { some_fail_function(); } assert(num); }`
<geist> huh and no warning on that?
<zid> part 2's final step was weird, all done though
<klange> and both gcc and clang eliminated the if and forced the call to some_fail_function and __assert_failed
<geist> oh because some_fail_function() is no-return?
<klange> presumably because __assert_failed is
<geist> it's possibly because it's no return, so it figures both paths dont exit
<zid> Looks like I did p2 in 6 minutes, overtaking 1400 people compared to part 1 which took me.. forever
<geist> one i've also seen that's really annoying is something like
<klange> no return → that's UB, but __assert_failed does not return so that path must be the only correct one therefore the if must always be true therefore here's two calls and no return :D
<geist> `int foo(enum some_enum) { assert(some_enum < MAX_VALUE_OF_ENUM); ... }
<geist> and the compiler just removes the assert because it says 'enums must be within range'
<geist> and it's like yeah yeah i'm trying to assert that some dipshit didn't pass some dumb crap to me
elderK has quit [Quit: Connection closed for inactivity]
<bslsk05> ​godbolt.org: Compiler Explorer
<zid> I mean, it should not be possible for anyone to *compile* you sending dumb crap, but the C spec fucking sucks at enums
<klange> (C++, though, not C)
<zid> nobody should actually use C enums as types
<zid> they don't work
<geist> zid: yeah but they can roll some int and pass it in. but that's the real answer. C++ fixes it nicely with proper enum classes, which i think is where they get useful
<zid> all the enum values are compatible with each other for very starters
<geist> that's oen of those absolutely better things that C++ does
<zid> yea there are so many braindead things in C that would be perfect with some minor alterations since *checks watch* 1989.
<zid> Instead we got _Complex and stuff
<zid> because they were handmedowns from the C++ committee
<geist> yah C is like the little starving kid getting food from Scrooge
divine has joined #osdev
<geist> yes sir thank you sir
<kazinsal> *staring down at a bowl of sigsegvs* please sir, may I have some more
<zid> It's more like.. the crazy aunt keeps giving you christmas presents that are obviously used jumpers that are 8 sizes too big and say BLACK AND PROUD, despite you being a 9 year old white girl
<kazinsal> thinking of building a new server with a 5800X
<zid> 5800x is nice
<zid> That's the amd cpu I'd buy
<geist> ryzen you mean?
<geist> ah yesh
<kazinsal> yeah
<kazinsal> should handily kick the ass of my 2x E5 2650 machine
<geist> so interesting: i think i may have pushed my 3950x a bit too far. it was my main desktop for 2 years, slightly overclocked
[itchyjunk] has quit [Read error: Connection reset by peer]
<kazinsal> while also booting in a tiny fraction of the time
<geist> i got a 5950x to replace it and moved the 3950x into the server
<geist> and... it has crashed twice now over the last few weeks
<geist> hard lock up, no MCE. no reset. just hard stop
<kazinsal> interesting
<zid> I've still yet to kill a cpu
<kazinsal> give it a bit more voltage at stock clocks imo
<zid> My xeon's on a 50% OC
<zid> It ran out of multiplier at that point
<kazinsal> shouldn't need the full monty, just maybe 50-70 mV over stock
<zid> it's also undervolted.
<geist> yah possibly works
<zid> Or it's just.. a faulty mosfet on the mobo
<zid> who can tell
<geist> but it does make an excellent VM host chip, aside from it being crashy
<zid> I have a fun graphics card which I am like 99% sure has a faulty VRM
<zid> It works 100% flawlessly, until it enters turbo, then crashes reliably
<zid> So it's 100% stable if you.. disable the turbo, then overclock the shit out of the non-turbo performance state to be >= the turbo state
<zid> I think it enables extra power rails when it goes into turbo state, and one of them is broken
<zid> and they aren't needed except for very extreme overclocking
<geist> another nice advantage of consumer ryzen servers: ECC actually works
<geist> though you dont get very solid ECC reports
<kazinsal> yeah, no need to buy a big fancy epyc board
<kazinsal> just grab a board and a CPU and some unbuffered ECC DDR4
<geist> watching this guy dig a tunnel under his house
<geist> he's a pretty crazy guy
<gorgonical> furze is wild
<zid> colin furze :P
<geist> yep
<gorgonical> i've been wondering how he got permits for this
<zid> his accent is super similar to my home town, the one I refuse to use
<geist> exactly
<kazinsal> I was gonna say "oh, season 1 of Bosch" but that's because I'm binge rewatching it ;)
<gorgonical> I've been hemming and hawing about building a shed sauna and this guy is digging a mine tunnel in his backyard
<geist> and in UK it's very real you slam into some roman crap too i assume
<kazinsal> imagine having the ability to build something on your property
<geist> i dunno if that's so common now it's just like whatever, toss it aside
<kazinsal> this depressing reality brought to you by: living in greater vancouver
<zid> geist: you have to phone an archaeologist or whoever who comes to inspect the site
<kazinsal> it is genuinely more feasible for me to get a green card and a job in seattle and buy a house there than it is to buy a house here
<zid> if it's significant they will make you stop digging
<gorgonical> just a stone's throw away from the wondrous interior of BC though, kazinsal
<geist> zid: yah my guess is significant is probably a fairly high bar now
<gorgonical> yeah though the housing costs are wild I heat
<gorgonical> hear*
<zid> You're just less likely to find anything at all these days
<zid> because all the cities were dug up to install the buildings in the first place
<zid> and it's all in the topsoil more or less
<zid> no roman stuff is under bedrock
<geist> yah true
<zid> (bet you weren't expecting me to link his planning permission application!)
<geist> ahaha yeah you did
<zid> his postcode is PE9, mine is PE7
<zid> he's.. probably a few miles away from my house
<gorgonical> "Has the work already been started without permission? Yes"
<gorgonical> Wait wait so he completed it and then just told the council? How is that allowed??
<kazinsal> it's important to be truthful on your permit forms
<zid> ah he's ont he other side of PE, but still, more or less exactly 20 miles
<geist> that's from 2016, probably mostly about the bunker he already built
<zid> yea he's already built it
<zid> and it took 3+ years
<zid> should I go knock on his door?
<gorgonical> This is bananas honestly. So little info on this and they just said "yep sounds good, it's already done anyway"
<zid> I could just go ask him
<geist> it's funny of all the lack of safety in these vids the thing the drives me the most nuts is using those rubber tote things to hoist the rocks
<geist> and of course no hard hat
<gorgonical> I know, ordinary looking laundry basket totes
<zid> What he's doing is dangerous but it's like.. broken arm dangerous, not life dangerous
<zid> so no more dangerous than any of the stupid vehicles he builds
<kazinsal> standard youtuber dangerous
<gorgonical> I would honestly say this seems substantially safer than those vehicles. He rides those things going 25+ mph just with a tie
<geist> okay fair
<geist> surprised has any hearing at all. maybe thats why he yells all the time
<geist> the first vid of his i watched was i think the pulse jet gokart from a few years back
<geist> had to be like 140 dB
<zid> yea I think that might have been mine too
<zid> hey you even got the 'were' in his dialect right kaz :D
<kazinsal> one of the convenience of being culturally and ethnically *waves vaguely in the direction of the english isles and ireland*
<kazinsal> you can usually guess dialects of her majesty's finest language fairly well
<zid> My friends all say 'were' when they mean 'was'.. with me being so close
<zid> also 'goo' instead of 'go'
<geist> yah the proper american version of colin is Adam Savage
<geist> who is pretty much the exact opposite
<kazinsal> ha, yep
<geist> well i guess jamie is the exact opposite, but still
<kazinsal> safety conscious, usually, but pleased to blow shit up when needed
<zid> colin's opposite is some redneck
<zid> or maybe AVE
<kazinsal> ah, ol' AvE
<kazinsal> a local-ish choochsmith
<zid> 20 miles local or am I still winning? :P
<kazinsal> not quite, same province
<geist> the canadian version is Mr Carlsen
<kazinsal> pretty sure he's up in one of the k-towns
<kazinsal> either kelowna or kamloops
<geist> he's in BC too. so incredibly even keel. kinda like an ASMR of electronics
<bslsk05> ​'1947 Electronics - Will It Still Work Today? Find Out!' by Mr Carlson's Lab (00:17:58)
<geist> good to just have going in the background when doing stuff
mkopriva has joined #osdev
scoobydoo has quit [Ping timeout: 265 seconds]
scoobydoo has joined #osdev
divine has quit [Ping timeout: 250 seconds]
divine has joined #osdev
<gorgonical> Holy moly that radio still works
<gorgonical> Also that electronics work in the bottom is positively horrifying
<geist> ahaha you actually watched it
<geist> yeah it's kinda fun watching him go through old stuff, and it's so Canadian
<gorgonical> It makes me wonder how much you could change history by introducing the breadboard or even the PCB back then
<gorgonical> Instead of having to wire every component directly in rat's nest fashion
<geist> yeah
<geist> i've seen lots of this sort of electronics, my dad is always working on old tube type ham radios
<geist> if you dig around he has a few 'look around my lab' videos, and he has a ridiculous amount of equipment
<gorgonical> that intro really shows it too. Insane
mkopriva has quit []
xenos1984 has quit [Quit: Leaving.]
Coldberg has quit [Ping timeout: 260 seconds]
_whitelogger has joined #osdev
x88x88x has quit [Remote host closed the connection]
x88x88x has joined #osdev
xenos1984 has joined #osdev
x88x88x has quit [Remote host closed the connection]
x88x88x has joined #osdev
<kernelspace> Bitweasil: thanks, is SCR.NS always readable from sec/nonsec ?
xenos1984 has quit [Quit: Leaving.]
sdfgsdfg has joined #osdev
kingoffrance has quit [Ping timeout: 252 seconds]
xing_song has quit [Read error: Connection reset by peer]
xing_song has joined #osdev
xenos1984 has joined #osdev
<geist> kernelspace: the secure/insecure bit on ARM?
<geist> from quickly grepping around the ARM v8 ARM it would appear there's nothing obviously that masks it
<gorgonical> kernelspace: to echo what geist said, I checked the manual. According to D13-3395 there's no clear mask on a specific bit. Looks like the only access requirement is EL3
<geist> in armv8 it looks like SCR_EL3.NS is where it exists, and thus is only readable from EL3
<geist> so i guess that maps directly to arm32s SCR
<geist> but it being in EL3 only means most applications/oses/hypervisors cannot read it
<gorgonical> yeah, that's like optee territory
<gorgonical> opt-tee?
<gorgonical> op-tee, apparently]
<geist> yeah
<geist> i am refreshing my memory on precisely how it switches the .NS bit on exception entry/exit into EL3
<geist> is it saved as part of SPSR? since there's only one .NS bit maybe it only applies to <=EL2?
<gorgonical> I don't have much experience with in/secure states. I know there's some pretty complex routing for exceptions, but I don't know if that's architectural or expected to be enforced by the EL3 firmware
<geist> oh yeah that's what it does
<geist> SCR.NS only affects the lower levels
<geist> EL3 is always intrinsically in secure mode
<gorgonical> I see
<gorgonical> So the exception routing model is just a suggestion for compliant el3 firmware
<gorgonical> "suggestion" maybe is a better phrasing
<geist> 0b0 Indicates that EL0 and EL1 are in Secure state.
<geist> 0b1 Indicates that Exception levels lower than EL3 are in Non-secure state, so memory accesses from those Exception levels cannot access Secure memory.
<geist> well, i think it's pretty much wired up to work a particular way
<zid> 0b0e indicates woodwind mode (sorry!)
<j`ey> zid: :D
<geist> it also mentions elsewhere that if an implementation does not implement EL3 it's IMPLEMENTATION DEFINED what security mode the cpu operates under
<geist> (since it still has to set the security bits on AXI bus transactions, etc)
<geist> for some reason i was always thinking the .NS bit was mirrored all the way down, but nope
<geist> but there's a newer feature that gives secure EL2 in which case it also i think mirrors it down one more level
<geist> FEAT_SEL2
<gorgonical> I'm not really understanding D1-2316 (d.1.4.1) here
<geist> ah SCR_EL3.EEL2 is a new bit that controls secure EL2 if FEAT_SEL2 exists
<gorgonical> It says security state can only be changed by taking exception to el3.
<geist> right, or back
<gorgonical> This is that NS bit? I mean, or is this just "you're in EL3 now and that's secure"?
<geist> ie you can go from lower security to higher on an exception to EL3, and back
<gorgonical> It must be the latter, right?
<geist> it says that EL3 is *always* secure
<geist> so the SCR.NS bit controls the levels below
<geist> whether or not when you switch to EL2 or EL1 or EL0 it'll turn the secure bit back off
<gorgonical> Oh I see
<geist> so the state isn't saved anywhere on exception entry, it's part of an EL3 control register
<gorgonical> But you can have secure EL2/1/0 right? So if it's toggling NS off then how can EL3 control it?
<geist> sure in that case you simply leave SCR_EL3.NS = 0
<geist> EL3 can set it whenever it wants, it just has to be in EL3 to do it
<geist> so if you're bouncing between secure and insecure EL1s for example, you have to do it via an exception to EL3, which can then fiddle with the bit
<geist> it's part of the state the EL3 monitor would save and restore as it switches contexts
<gorgonical> Right, so this is just intel SDM-style "here is exactly what you cannot do" and you have to read in between the lines
<gorgonical> That's what my understanding of the situation was, that el3 just has to moderate the switches from in/secure states
<geist> yah, the EL2 virtualization stuff is very much that way in ARM since it's basically a bunch of tools you can build a hypervisor out of
<geist> and in general yeah, EL3 primarily exists to have a higher authority than anything else to switch secure contexts
<Belxjander> 3 layers of exceptions ???
<gorgonical> I was just thinking that it's arbitrary that we stop at el3. Like you could have a metasecure layer and tack on el4.
<gorgonical> 4 actually, there's an el0
<gorgonical> Obviously I don't know what use a metasecure layer is lol
<geist> and as we've mentioned before, apple apparently didn't decide they needed secure mode so they simply left out EL3
<geist> M1 cpus' max mode is EL2
<j`ey> Belxjander: user el0, kernel el1, hypervisor el2, secure mode el3
<gorgonical> But don't they have codesigning and tpms and stuff? That's just part of the el2 stuff?
<geist> i think their strategy is to rely more on external (or internal to the die) secondary processors
<geist> vs needing some sort of secure mode OS
<gorgonical> I see
<Belxjander> j`ey: yikes... and here I was thinking about whether to try getting something Android ARM based working on something else in a wrapper HLE layer setup... **way** too much work for the time I do have ...
<geist> makes sense if you build your own hardware: add more dedicated processors to do that stuff
<gorgonical> Belxjander: I am exaggerating a little since you can't take exceptions to el0. Though you do have an el0 timer
<geist> Belxjander: it's not too bad. actually i think it makes a nice logical sense since for the most part they nest in a fairly intuitive way
<gorgonical> Honestly one of my favorite features of ARM so far has been the system-wide counter. Holy moly does it make benchmarking virtualization and core-migrating easier
<geist> yah right?
<gorgonical> Here I am mincing again about whether SIPI sequence on x86 refreshes core tsc or not since I am doing virtualization reconfiguration on x86
<geist> FWIW riscv is also following more or less the same model: machine mode == EL3, supervisor mode = EL1, user mode EL0
<geist> and their new hypervisor extensions are kidna a pseudo EL2, though not precisely
<gorgonical> tbh the VHE part of arm confuses and disappoints me. I was hoping to be free of "but if you enable *this* bit then all this shit works differently from normal because compatibility"
<geist> yah, sadly i think the VHE part is kinda like admitting the intel model isn't half bad
<j`ey> Belxjander not sure what a HLE wrapper is..
<gorgonical> don't say it. I don't want it to be true
<geist> my initial thought a long time ago when i realized everything is just going page tables now and all the other styles of MMUs are dying out
<gorgonical> I have only limited experience with other MMUs. Once I read a paper from... Xerox? About a grid-memory MMU
<geist> and everything is little endian, etc etc. the effect that x86 has had on everything is profound
<gorgonical> That was wild
<gorgonical> part of the paper was "yeah we just fabbed this chip out to prove it would be useful" as though it were so casual. Hilarious
<geist> heh yeah
<j`ey> gorgonical: VHE just means linux didnt need a rewrite :P
<gorgonical> j`ey: cowards! the lot of them
<geist> if you consider the sum of all architectures that have ever existed, really all the modern remaining things are far closer in design than they are apart
<gorgonical> geist do you have some pointers for mmus that aren't page table centric? I have so little experience it hadn't even occurred to me you'd use something else
<geist> power/ppc uses a large single hash table, which is pretty trippy
<geist> ia64 had *two* kinds of page tables, one of which is a large virtual page table, very trippy
<gorgonical> power had to have it to go along with their crazy tlb where you can put whatever you want in any way in it, I guess
<geist> and lots of mid 80s or lower end risc machines had TLB miss exception based hardware: software provides the data structure
<gorgonical> whoa
<gorgonical> Yeah sure why not? Get a tlb miss exception and just deal with it
<geist> ultrasparc is a good example, or say SH or microblaze or nios 2 or whatnot
<geist> lots of lower end risc machines
<geist> yep. also lower end embedded ppc machines have a TLB miss mode i believe
<geist> ia64 has a tlb miss mode too, though it serves the virtual page table thing, which i think i understood at one point in my life and then forgot
<gorgonical> part of me is sad that I wasn't old enough to play with all these architectures. I sometimes feel like I got on this train late and everybody's doing a different thing now, what with the CPU trending toward an accelerator manager
<geist> yah its one of the reasons i like to keep some of them alive a bit. i graduated college right as everything collapsed. ia64 killed everything all at once
<j`ey> gorgonical: same
<geist> it was going to be the Next Big Thing and everything died at once
<gorgonical> honestly being in HPC research is a little sad. Like summit and sierra are basically just chips hosting a stack of GPUs now. I think fugaku is a little better but it's basically the same
<geist> not that a bunch of unix companies producing their own workstations with their own architectures was going to survive any longer anyway
<geist> ia64 was just the nail
<gorgonical> "how fast can we get data into and out of the GPU blackboxes?"
<gorgonical> and how much of our souls are we willing to sell to nvidia for it
<j`ey> fugaku.. if the fujitsu one, is some custom arm CPUs that's cool gorgonical !
<gorgonical> j`ey: yeah the fujitsu one. It is indeed extremely cool to deploy fast arm chips
<zid> gorgonical: if you think the cpu is an accelerator manager now, try a 486 box
<gorgonical> Other labs have the cavium thunder x2 chips, but fujitsu really took the ball and ran with it
<zid> with a discrete sound card, discrete vga card, discrete network card, discrete pci bus, blah blah blah
<gorgonical> zid: that is true. I have been lectured about the horrors of emulating PCI chips and how absolutely none of it makes sense
<gorgonical> Basically reverse-engineered code with comments like // ???
<zid> gorgonical: who doesn't like discrete L3 cache, also
Burgundy has joined #osdev
<gorgonical> but then nvidia bought arm, too! and now the ship has several compartments breached
immibis has joined #osdev
nyah has joined #osdev
zaquest has quit [Remote host closed the connection]
zaquest has joined #osdev
warlock has joined #osdev
warlock_ has joined #osdev
warlock_ is now known as doubletoker
doubletoker has quit [Client Quit]
ravan has quit [Ping timeout: 240 seconds]
dormito has quit [Ping timeout: 250 seconds]
<klange> One of my USB hubs just crashed and did not reset when unplugged... because it was retaining power from an AC adapter.
<zid> I hate haunted hw because power loss didn't cremate the ghost living in it
<zid> had that happen a few times with motherboards, no amount of power cycling would fix them until I left it turn off for 5 minutes
<zid> which led me up a bunch of garden paths trying to diagnose it
GeDaMo has joined #osdev
gog has joined #osdev
sdfgsdfg has quit [Quit: ZzzZ]
dennis95 has joined #osdev
<kernelspace> geist: i am in arm v7, wondering if NS bit is readable always, from sec or non sec workld
dormito has joined #osdev
lg has joined #osdev
xenos1984 has quit [Quit: Leaving.]
jiehuan has joined #osdev
<clever> kernelspace: i would assume it is, at least for kernel space, so you can even know which mode your in
SpikeHeron has quit [Quit: WeeChat 3.0]
the_lanetly_052 has joined #osdev
ahalaney has joined #osdev
dormito has quit [Ping timeout: 250 seconds]
gog` has joined #osdev
Kerum has joined #osdev
xing_song1 has joined #osdev
xing_song has quit [Ping timeout: 250 seconds]
xing_song1 is now known as xing_song
[itchyjunk] has joined #osdev
<kernelspace> clever: thanks
<kernelspace> and thansk all of this great helpful channel
Kerum has quit [Quit: leaving]
Kerum has joined #osdev
Coldberg has joined #osdev
dormito has joined #osdev
sahibatko has joined #osdev
ElectronApps has quit [Remote host closed the connection]
ElectronApps has joined #osdev
xenos1984 has joined #osdev
isaacwoods has joined #osdev
srjek|home has joined #osdev
heat has joined #osdev
the_lanetly_052_ has joined #osdev
the_lanetly_052 has quit [Ping timeout: 268 seconds]
gog` has quit [Read error: Connection reset by peer]
Kerum has quit [Quit: leaving]
dude12312414 has joined #osdev
friedy10 has joined #osdev
mimmy has joined #osdev
xing_song has quit [Read error: Connection reset by peer]
xing_song1 has joined #osdev
xing_song1 is now known as xing_song
mimmy_ has joined #osdev
mimmy has quit [Ping timeout: 240 seconds]
heat has quit [Ping timeout: 268 seconds]
mimmy_ has quit [Ping timeout: 250 seconds]
mimmy_ has joined #osdev
friedy10 has quit [Ping timeout: 240 seconds]
mahmutov has joined #osdev
xenos1984 has quit [Quit: Leaving.]
ElectronApps has quit [Remote host closed the connection]
kingoffrance has joined #osdev
knebulae has quit [Read error: Connection reset by peer]
k0valski18 has quit [Remote host closed the connection]
k0valski18 has joined #osdev
jiehuan has quit [Remote host closed the connection]
xing_song has quit [Read error: Connection reset by peer]
xing_song1 has joined #osdev
kingoffrance has quit [Ping timeout: 252 seconds]
isaacwoods has quit [Quit: WeeChat 3.3]
pretty_dumm_guy has joined #osdev
nvmd has joined #osdev
vdamewood has joined #osdev
Belxjander has quit [Quit: AmigaOS PPC 4.1 +E +U1 // AmIRC 68K]
xenos1984 has joined #osdev
dennis95 has quit [Quit: Leaving]
kingoffrance has joined #osdev
darkstardevx has quit [Read error: Connection reset by peer]
mimmy_ has quit [Ping timeout: 268 seconds]
mimmy_ has joined #osdev
vdamewood has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
the_lanetly_052_ has quit [Ping timeout: 260 seconds]
isaacwoods has joined #osdev
<geist> kernelspace: i'm guessing no, unless you're in EL3
<geist> though v7... maybe. but v8 seems to generally hide it, so i wouldn't be surprised if that's not consistent with v7
darkstarx has joined #osdev
darkstarx has quit [Remote host closed the connection]
darkstarx has joined #osdev
srjek|home has quit [Ping timeout: 252 seconds]
radens has joined #osdev
<kernelspace> geist: mm, well, i am in u-boot spl, from some asm code, trying to understand if i am in sec or ns
<geist> ah
<kernelspace> not very skilled on arm7 stuff, at this level. So, if i am understanding, EL3 is secure world, correct ?
mimmy_ has quit [Ping timeout: 260 seconds]
<kernelspace> so doing
<kernelspace> mrc p15, 0, r0, c1, c1, 0
<kernelspace> bit 0 seem set to 1
darkstarx has quit [Read error: Connection reset by peer]
mctpyt has quit [Ping timeout: 268 seconds]
vdamewood has joined #osdev
darkstarx has joined #osdev
darkstarx has quit [Remote host closed the connection]
darkstarx has joined #osdev
hellstabber has joined #osdev
DutchIngraham has joined #osdev
Burgundy has quit [Remote host closed the connection]
hellstabber has quit [Quit: My MacBook Air has gone to sleep. ZZZzzz…]
dormito has quit [Ping timeout: 260 seconds]
CaCode has joined #osdev
GeDaMo has quit [Remote host closed the connection]
mimmy_ has joined #osdev
Burgundy has joined #osdev
CaCode_ has joined #osdev
mimmy has joined #osdev
nexgen has joined #osdev
mimmy_ has quit [Ping timeout: 268 seconds]
CaCode has quit [Ping timeout: 268 seconds]
dormito has joined #osdev
dormito has quit [Client Quit]
<kazinsal> vacation approved, concert ticket acquired, gonna do a couple days in Seattle in March to unwind :toot:
CaCode has joined #osdev
CaCode_ has quit [Ping timeout: 256 seconds]
kingoffrance has quit [Ping timeout: 252 seconds]
<zid> kazinsal: something something sleepless hur hur
<kazinsal> that depends on a few other things going well first hur hur hur
Matt|home has quit [Quit: Leaving]
<zid> I didn't know you were into all-night games of D&D to that degree.
<kazinsal> that's more like a consolation prize
Matt|home has joined #osdev
<sortie> kazinsal, boom
<sortie> Who you seeing?
<kazinsal> Queensryche and Judas Priest
<sortie> That's awesome
<gog> nice
sortie has quit [Quit: Leaving]
sortie has joined #osdev
<kazinsal> and if I manage to muster up enough courage (which may involve this bottle of wine work sent me) I maaaaaay end up asking someone I hang out with regularly online and play a few video games with who lives in Seattle out for dinner
ahalaney has quit [Remote host closed the connection]
<j`ey> kazinsal: shoot ur shot
<sortie> go go kazinsal
<zid> kazinsal: I'm honored, which games?
Dreg has quit [Read error: Connection reset by peer]
Dreg has joined #osdev
<klange> Would anyone care to review the preliminary release notes for ToaruOS 2.0? https://gist.github.com/klange/e31be3d1aeed22ebd5efd6dc58ed19ed
<bslsk05> ​gist.github.com: release.md · GitHub
mimmy has quit [Ping timeout: 250 seconds]
<j`ey> AArch64 not Aarch64
<klange> I think that was 'aarch64' getting sentence-cased but okay
dormito has joined #osdev
<j`ey> LGTM other than that
srjek|home has joined #osdev
<bauen1> so long story short, is there a way to compile gcc with a "freestanding libstdc++-v3" without any libc for the target ?
<bauen1> that must be something that people do right ? like if you're trying to use C++ in the kernel you won't have a libc, but you wants libstdc++-v3 for the cstdint headers etc... ?
<sortie> klange, what! A release!
<sortie> How dare you
<sortie> Oh man what a beautiful wallpaper and about toaruos screenshot at the top!
<bauen1> klange: you ported your kernel to 64 bit ?
<klange> bauen1: no, I mostly wrote a new one
<klange> and then copied over the bits that didn't need any 'porting'
<klange> eg. vfs, and other high level things
CaCode_ has joined #osdev
CaCode has quit [Ping timeout: 252 seconds]
<sortie> klange: “, a loopback,” ← Maybe be explicit and say loopback interface?
<klange> ack, edited
<moon-child> klange: 'A dynamic linker (ld.so)' doesn't end in a period, unlike all the other bullets
<klange> ack, ty
mimmy has joined #osdev
<moon-child> 'Kuroko first appeared in [...]. Kuroko has [...]' feels a bit repetitive and awkward. Maybe go for 'Kuroko first appeard in [...], but has improved greatly'
<moon-child> (or maybe and instead of but, not sure which reads better)
nexgen has quit [Read error: Connection reset by peer]
kingoffrance has joined #osdev
CaCode- has joined #osdev
<zid> Great improvements..
CaCode_ has quit [Ping timeout: 252 seconds]
<sortie> klange, a bit too tired to give a proper thorough review but overall what an impressive release and the release notes really sells that
<klange> sortie: Thank you :) The plan is to post everything tomorrow, so there is plenty of time to make sure the release notes are in good order.
<eryjus> Klange, bring "release 2.0" into the opening paragraph (versus "this release")
<klange> There will be a title "ToaruOS 2.0" on the actual release notes.
nexgen has joined #osdev
<eryjus> my thought process; i think it is important to have that fact stand on its own in the first paragraph. also iirc, everything is your own code (nothing ported)... is that correct?
<eryjus> i have not gotten all the way through the notes yet, but it that called out?
<eryjus> i think it's something to be proud of and celebrate
<klange> It's in there a few times.
elderK has joined #osdev
<eryjus> ha! i was going to ask when we would see an aarch64 port.
<eryjus> nice, klange!
<eryjus> ok, bullets in "Features of Misaka" have no periods; bullets in "Changes since 1.14.1" do
<eryjus> Also recommend capitalizing "Since"
<eryjus> If you add periods to "Change ....", "New userspace utilities exposing functionality added in Misaka: top, strace, dbg, ping, cpuwidget" has no period
<zid> I recommend inverting it and changing it to Erstwhile
<klange> zid: I don't think there's a logical way to do that.
<eryjus> Consider rewording to "Many more things of which I lost track."
adachristine has joined #osdev
gog has quit [Read error: Connection reset by peer]
adachristine is now known as gog
<zid> klange: I'm surprised you even bothered to respond to that :P It's technically doable though, 'Erstwhile versions lack the following features', I hate it :D
<eryjus> ok, back to my own shitty kernel
<zid> eryjus: is that the title of your next novel? I would read it
<zid> Back to my own shitty kernel - Volume 01 [J-Novel Club][Kobo_LWNCentral].pdf
<bslsk05> ​github.com: centuryos2/journals/v0.0 at master · eryjus/centuryos2 · GitHub
<zid> what's a redmine
<eryjus> i use that for my bug tracking
<klange> < eryjus> Consider rewording to "Many more things of which I lost track." ← I've got two reasons I don't want to do this.
<eryjus> fair enough
<klange> For one, I just think the whole "don't end sentences with prepositions" is malarkey.
<kazinsal> correct
<kazinsal> natural flow > some stuffy ivy league dickhead's manual of style from 1824
<klange> And two, I'm going for a more personal spoken style, so even if it were a valid style rule ain't nobody talk that way.
<zid> Clearly it should be Things of which track I lost of.
<zid> for that added touch of pissing that random dead american off
<eryjus> add a passive voice to really piss us off
dude12312414 has quit [Quit: THE RAM IS TOO DAMN HIGH]
<klange> zid: The rule is commonly attributed to an English poet in the 1600s, John Dryden.
<kazinsal> things missing after an author-involved changelog
<klange> Paintings confirm he was a grumpy bastard.
<zid> klange: suggesting that it might not be the best way to do something != enforcing it as a rule, though
<klange> Webster has a nice article on the history and bullshittery of it: https://www.merriam-webster.com/words-at-play/prepositions-ending-a-sentence-with
<bslsk05> ​www.merriam-webster.com: Ending Sentences With Prepositions: Is it Allowed? | Merriam-Webster
<moon-child> zid: the point of grammar rules is to promote communication
<moon-child> so the question is: it is harder to understand sentences which end with prepositions?
<moon-child> I don't think so (though obviously you can argue that)
<klange> I very often find it harder to understand the twisted garbage someone produced trying to abide by "the rule".
<moon-child> yes
<zid> If you want to fix english just split 'put' into two forms thanks
<zid> I mean I guess place technically exists.. maybe we just make that the rule! new rules yay!
<moon-child> zid: if we re-place it with 'place', then where would 'put' be put?
<zid> in the past tense where it belongs, or in the present opposing placed
<zid> flip a coin
<zid> sit/sat, eat/ate, put/put
<moon-child> hmm
<moon-child> so will you replace a golf 'putter' with a 'placer'> :)
<moon-child> s/>//
<zid> That has no tense so it's fine
<klange> putt is a different word than put ;)
<zid> also yea it's a putt-er
<zid> technically
<moon-child> right
<zid> Also we teach americans about the sentence form "The man was stood still"
<zid> apparently large amounts of americans can only read the word 'stood' to mean 'was placed'
<zid> past participle? idk grammar
<zid> Quote from a prescriptivist american blog I just googled
<zid> "It is perfectly correct to say, “I was stood at the bus stop” — but only if someone picked you up physically, walked you to the bus stop, placed you down and stood you there."
<moon-child> eh I don't buy that even under the prescriptivist ideology
<moon-child> first: nothing to say you can't already have been at the bus stop when you were stood; second: nothing to say you can't have stood yourself
<zid> Yea I think it probably comes from the latter
<zid> just a unification between being forced to stand there and forced yourself to stand there
<klange> how is that any different than 'was standing'
<zid> hence: I stood
<kingoffrance> that sounds like a horrible date that went horribly wrong
<zid> It isn't, it's just how I talk
<moon-child> kingoffrance: lol
<klange> kingoffrance: that's being stood up!
<klange> aaaand 'stood' is now semantically satiated in my mind, and I'm now reading it like stewed
<zid> laid and sat also act the same way to me
Oli has joined #osdev
<kingoffrance> If it wasn't for my horse, I wouldn't have spent that year in college </lewis black>
<eryjus> zid: lay and lie. go.
Belxjander has joined #osdev
<zid> hmm?
<zid> Go where, go lay down? :p
<zid> I think this might actually be active vs passive
mimmy has quit [Ping timeout: 260 seconds]
<zid> "Where his remains lay" "where his remains were laid" "He was standing" "He was stood"
<zid> I should read a grammar book at some point
<moon-child> it's not
<moon-child> oh wait hmm maybe
<bslsk05> ​www.britannica.com: Lay, Lie, Lied, Lain: When Do We Use Which? | Britannica
<eryjus> moon-child: you beat me to bringing "lain" into this
<moon-child> :)
<zid> but you can't tell if that person has already been infected with it though
<zid> if they're a person who already uses "I stood", then they'll suggest "I lay" as normal as well!
<zid> That's my second discussion including lain in two days actually
<zid> Do you know where I can get a bear onesie?
<zid> For a 6' tall man
<moon-child> sew it your damn self!
<zid> but you don't evn understand
<zid> I thought you were an honest man
<klange> I'm going to need help breathing if I don't stop laughing...
<kazinsal> we seem to have walked into a sketch comedy show
<eryjus> klange: go lay/lie down
<zid> I am deadly serious about the onesie, please don't suggest I am joking!
<kazinsal> one of my best friends sews fursuits, I can get you a quote
<zid> Oh nevermind
<zid> If people are going to confuse it with a fursuit I'm not going within a billion miles
<zid> what about a lain daki then
<kazinsal> a full size animal onesie is just a fursuit in denial
<zid> cosplay is not furriness :(