<geist>
yah, it actually kinda makes sense. in this case the attributes are fairly sane
<heat_>
constexpr!!!
<geist>
constexpr FTW
<geist>
yeah constexpred up a few functions while i was at it
<heat_>
do you have swapping?
<geist>
no, we have a form of zram
<heat_>
aww
<geist>
no real push to do swapping
<heat_>
i was going to ask if you had done any sort of crazy linux-like PTE inversion
<heat_>
cuz of all the side channels they invert the PTE so it's literally invalid
<geist>
i did read about a new v8.3/9.3 extension that lets you redefine the permission bits in the page tables
<geist>
called something like indirect permssions
<geist>
but then it dramatically affects the way heirarhcial permissions work
<geist>
and there's another bit that lets you disable HP which then lets yo use the bits for other stuff, etc
<geist>
s/v8.3/v8.8
<heat_>
are they still developing v8?
<heat_>
what's the difference between v8 and v9 anyway?
<heat_>
except "buncha features"
<geist>
v9.0 == v8.5 + SVE
<geist>
but yeah they're apparently continuing to at least maintain the property that v9.0 features map to v8.5+ features and vice versa
<geist>
i dunno why they're doing it to be honest, except my guess is the need to declare a v9 was more of a marketing/etc decision
<geist>
and so they just sort of snapshotted what was going on with v8 and called it that, but didn't really change anything under the hood
<geist>
also note there's no v9 ARM ARM. it's just v8 with a note that says v9 == 8.5
<geist>
so now if you get the absolute latest ARM ARM it's like 50 extensions overlaied on top of each other, some of which are optional, some are mandatory, and some when uyou enable will undo the effect of previous ones. etc
<geist>
it's a huge pile of complexium
skydhash has joined #osdev
<heat_>
oh so like x86
<heat_>
nothin new
<heat_>
the optionalness is annoying (cough cough riscv) but it is what it is
skydhash has quit [Ping timeout: 268 seconds]
<nikolapdp>
i thought risc was supposed to be simpler
<zid`>
It's simple the way learning kanji is simple
<zid`>
"each little picture is a word, simples"
<nikolapdp>
heh
<kof673>
> ½, the Plan 9 Window System i heard you like 8.5, now you can run 8.5 on your 8.5
<geist>
it is a lot simpler
<geist>
it's just going to inevitably get more complicated
gog has quit [Quit: byee]
<heat_>
plan 9 is plan v8.5 + FEAT_WINDOW
srjek has quit [Ping timeout: 255 seconds]
navi has quit [Read error: Connection reset by peer]
<bslsk05>
wiki.osdev.org: OS Specific Toolchain - OSDev Wiki
<ddevault>
sortie: they have now generated it, according to binutils master
netbsduser has quit [Ping timeout: 240 seconds]
gbowne1 has quit [Remote host closed the connection]
<ddevault>
sortie: you also should drop dbxelf.h from gcc/config.gcc
<ddevault>
removed upstream in 7e0db0cdf01e9c885a29cb37415f5bc00d90c029
<ddevault>
aside, has anyone got wiki registrations working again? or is the website still stuck in limbo
<ddevault>
oh my god it works
<ddevault>
woo!
zxrom has quit [Quit: Leaving]
<ddevault>
good lord, I last used this account 10 years ago
zxrom has joined #osdev
<nikolar>
What's 10 years between friends :)
gog has joined #osdev
Terlisimo has quit [Quit: Connection reset by beer]
Terlisimo has joined #osdev
<sortie>
ddevault: Woohoo! Yeah it's been a bunch of years since I last worked on that article. I didn't originally write OS Specific Toolchain but I did do massive contributions to it back in the day. Happy that you continue to contribute to it :)
<ddevault>
yeah, but with some, as far as I can tell, nonstandard sections for things like registers
bitoff has joined #osdev
hmw[at] has quit [Quit: Leaving.]
<Mutabah>
dwarf?
<netbsduser>
how many hobby kernels actually support a proper unmount?
<dostoyevsky2>
netbsduser: is an unmount that difficult to implement?
<netbsduser>
dostoyevsky2: more so than to not do so
<netbsduser>
it just strikes me as one of those things that few people bother to do in a hobby kernel
<zid`>
If it can be resolved by rebooting
<zid`>
it's fairly low on the prio list I imagine
<zid`>
It's a very very very long process to get a hobby OS you can actually *use*
<netbsduser>
i don't have it yet because you need tracking of open files on an FS, chasing up deletion of namecache entries, chasing up vnode writeback, chasing up vnode destruction, etc
<netbsduser>
in time i will
<dostoyevsky2>
netbsduser: you could just implement some reference counting and if it drops to zero and the device is marked as to be unmounted... you unmount it
<netbsduser>
dostoyevsky2: it's not so trivial
<netbsduser>
you can do a bare minimum effort like that and be rewarded with an unmount that will *never* happen until a) everyone has stopped interacting with the FS b) all writebacks have completed at their usual hour dictated by the writeback daemons c) there has been prolonged memory and namecache pressure resulting in the gradual replacement, depth-first, of every entry in the namecache tree relevant to that FS
<dostoyevsky2>
netbsduser: I wonder if there is a small OS that could e.g. unmount and one could use as a launching pad for OS development
dude12312414 has joined #osdev
<kof673>
well...if you have "journaling" also, it may not be clean but ...give processes 30 seconds or whatever to clean up?
dude12312414 has quit [Quit: THE RAM IS TOO DAMN HIGH]
bauen1 has quit [Ping timeout: 260 seconds]
<dostoyevsky>
netbsduser: put all your kernel data structures in sqlite tables, then you can always query who is using what ;-)
<ddevault>
getcwd. hm.
bauen1 has joined #osdev
<netbsduser>
namecache to the rescue for getcwd
<ddevault>
yeah
<ddevault>
I was thinking if I need a dcache before I can do getcwd without hacks
<bslsk05>
github.com: 4.4BSD-Lite2/usr/src/lib/libc/gen/getcwd.c at master · sergev/4.4BSD-Lite2 · GitHub
<netbsduser>
but i'm of the firm opinion now that a proper namecache with pinning of parents of currently-referenced directories is a vital design element, not least to avoid this...
<ddevault>
this approach had occured to me, grab .. until you find root
<ddevault>
and the obvious flaw, having to readdir the whole bloody thing each time, also occured to me shortly thereafter
<ddevault>
will depend on how lazy I'm feeling tomorrow
PublicWiFi is now known as PUBLICWIFI
Arthuria has joined #osdev
bauen1 has quit [Ping timeout: 256 seconds]
Mach has joined #osdev
Mach has quit [Client Quit]
Mach has joined #osdev
ddevault has quit [Remote host closed the connection]
bl4ckb0ne has quit [Remote host closed the connection]
ddevault has joined #osdev
bl4ckb0ne has joined #osdev
agent314 has quit [Ping timeout: 255 seconds]
m257 has joined #osdev
agent314 has joined #osdev
m257 has quit [Ping timeout: 250 seconds]
gareppa has joined #osdev
edr has joined #osdev
zhiayang has quit [Quit: oof.]
zhiayang has joined #osdev
agent314 has quit [Ping timeout: 252 seconds]
Arthuria has quit [Killed (NickServ (GHOST command used by Guest684531))]
Arthuria has joined #osdev
Arthuria has quit [Ping timeout: 255 seconds]
bauen1 has joined #osdev
Matt|home has quit [Quit: Leaving]
dude12312414 has joined #osdev
dude12312414 has quit [Remote host closed the connection]
<zid`>
I need a 4 open socket paladin shield plzkth
tomaw has quit [Remote host closed the connection]
tomaw_ has joined #osdev
tomaw_ is now known as tomaw
<kof673>
> From Old French paladin, from Latin palladius, from Pallas, a goddess of wisdom that's the best i can do, paladinen
dostoyevsky2 has quit [Ping timeout: 240 seconds]
dostoyevsky2 has joined #osdev
xal has quit []
Mach has quit [Ping timeout: 255 seconds]
xal has joined #osdev
xal has quit [Client Quit]
xal has joined #osdev
heat has joined #osdev
<heat>
netbsduser, yep, umount is TODO here
gog has quit [Quit: Konversation terminated!]
divine has quit [Quit: Lost terminal]
MrCryo has joined #osdev
npc has quit [Remote host closed the connection]
gog has joined #osdev
PUBLICWIFI is now known as PublicWiFi
MrCryo has quit [Remote host closed the connection]
voidah has joined #osdev
voider has joined #osdev
voidah has quit [Read error: Connection reset by peer]
voider has quit [Remote host closed the connection]
* geist
yawns
voidah has joined #osdev
voidah has quit [Client Quit]
<gorgonical>
indeed
<gorgonical>
I spent all last night wrangling a shittier version of my i2c driver I wrote for benchmark comparisons
<gorgonical>
Cobbled together a rather bad sort of TLV system to transfer variable-length i2c messages between kernels in one contiguous buffer