<Mondenkind>
so probably yeah that? what's the problem then
<Cindy>
the timing i would be able to do this i guess
neo|desktop has quit [Read error: Connection reset by peer]
<Cindy>
assuming the CPU is clocked at 12MHz
lentement has joined #osdev
<Cindy>
there would only be 2000 cycles before the next interrupt
Arthuria has quit [Ping timeout: 255 seconds]
<Cindy>
whoops, i meant 75000 cycles
<Mondenkind>
so on cycle 0 you get a vsync interrupt and a timer interrupt. cycle 120000 you get a timer interrupt, 200000 vsync interrupt, 240000 timer interrupt, 360000 timer interrupt, 400000 vsync interrupt
<Cindy>
yes
<Cindy>
the timer is for the OS
<Cindy>
and vsync, you know what it is for
lentement has quit [Ping timeout: 264 seconds]
heat_ has quit [Read error: Connection reset by peer]
heat has joined #osdev
lentement has joined #osdev
lentement has quit [Ping timeout: 260 seconds]
zxrom has quit [Quit: Leaving]
lentement has joined #osdev
lentement has quit [Ping timeout: 260 seconds]
heat has quit [Read error: Connection reset by peer]
heat has joined #osdev
lentement has joined #osdev
<heat>
BUG: unable to handle page fault for address: 0000000000009c55
<heat>
either memory is fucked or any of the drivers
<heat>
wouldn't happen if i were using a microkernel
<heat>
suck it linus torvald
<heat>
i'm getting no more info and the system hanged right after :/
sbalmos has quit [Ping timeout: 260 seconds]
<Ermine>
heat: ed is the most UNBLOATED editor
sbalmos has joined #osdev
<heat>
OPTIMAL EDITOR
<heat>
mjg, does ed scale to 200 CPUs? thanks
<Ermine>
Useful in case of nuclear war I guess
lentement has quit [Ping timeout: 240 seconds]
<Ermine>
heat: do you have kdump configured
<heat>
no
<Ermine>
try, it might get you a kernel core dump
<heat>
i have no idea how it happened, i just closed a tab and the system fell over
<heat>
journalctl told me "BUG: ..." but no more info, so maybe the system capitulated right after that
<Ermine>
and iirc BUG is something what happens before panic
<heat>
BUG doesn't necessarily mean panic
<Ermine>
or is that oops...
<heat>
BUG -> oops
<heat>
unless you have panic_on_oops
<heat>
in which case, BUG -> oops -> panic
<Ermine>
May that mean that something in kernel is fucked so panic becomes imminent?
<heat>
i don't think it panic'd
<heat>
it sounded like something was spinning on something, probably a lock, which maybe the BUG just left held
<heat>
it was certainly quick enough that systemd could pick up the beginning of the BUG and write it, but not the end
<heat>
0000000000009c55 the address doesn't tell me anything. i might run a memtest tomorrow or something, just to make sure the RAM is okay
lentement has joined #osdev
lentement has quit [Ping timeout: 260 seconds]
<heat>
the kdump stuff reminds me that the linux crashdump is very funny. /proc/kcore is literally like 16TB in size if you don't interpret it
<heat>
the crash kernel's vmcore is way smaller because they don't even try to look at virtual mappings there
<heat>
then they basically do further analysis in userspace and try to compress it by skipping free pages, etc. it's a funky system
lentement has joined #osdev
* gog
dumps
lentement has quit [Ping timeout: 268 seconds]
lentement has joined #osdev
<zid>
gog good news though
<zid>
it's technically monday
lentement has quit [Ping timeout: 260 seconds]
gog has quit [Ping timeout: 252 seconds]
netbsduser has quit [Ping timeout: 252 seconds]
lentement has joined #osdev
foudfou has quit [Ping timeout: 260 seconds]
foudfou_ has joined #osdev
lentement has quit [Ping timeout: 264 seconds]
<vai>
zid: monday here : )
lentement has joined #osdev
netbsduser has joined #osdev
heat has quit [Ping timeout: 240 seconds]
heat has joined #osdev
netbsduser has quit [Ping timeout: 272 seconds]
heat has quit [Ping timeout: 256 seconds]
foudfou_ has quit [Remote host closed the connection]
[itchyjunk] has quit [Remote host closed the connection]
<clever>
the source/output, will selectively pull one line low
<clever>
the sink/dest has pullup resistors, and measures the difference
<clever>
if there is no dest, there is no pullup, and you get zero voltage on the outputs
masoudd has joined #osdev
<Ermine>
oh, so it's foolproof
lentement has quit [Ping timeout: 260 seconds]
<clever>
Ermine: i think it was meant more for emc reasons, so a loose hdmi cable isnt radiation rf energy, but it does double as a safety for the case you asked
<immibis>
welcome to new new reddit, which is hostile to users
<immibis>
stop using it
lentement has joined #osdev
k_hachig has joined #osdev
lentement has quit [Ping timeout: 268 seconds]
k_hachig has quit [Ping timeout: 268 seconds]
Matt|home has joined #osdev
k_hachig has joined #osdev
k_hachig has quit [Max SendQ exceeded]
k_hachig has joined #osdev
voidah has joined #osdev
lentement has joined #osdev
gog has quit [Quit: Konversation terminated!]
lentement has quit [Ping timeout: 260 seconds]
lentement has joined #osdev
lentement has quit [Ping timeout: 260 seconds]
<Ermine>
when I issue cpupower frequency-info, cpu frequency spikes from 400 mhz to like 1.2 ghz
gog has joined #osdev
xenos1984 has quit [Ping timeout: 268 seconds]
xenos1984 has joined #osdev
Fingel has joined #osdev
lentement has joined #osdev
lentement has quit [Ping timeout: 268 seconds]
CryptoDavid has joined #osdev
masoudd_ has joined #osdev
masoudd has quit [Ping timeout: 272 seconds]
heat has joined #osdev
masoudd_ is now known as masoudd
<heat>
Ermine, i gave you patch to dump more exception data but you never got back to me :(
<heat>
anyway ONYX STRONK ONYX GOOD ONYX FAST
<Ermine>
oh, I must have forgor it
<Ermine>
my bad
gog has quit [Ping timeout: 252 seconds]
<heat>
no problem lol
<Ermine>
I probably will open an issue then so I don't miss stuff
<heat>
yeah feel free
<heat>
it'll help track the problem
xenos1984 has quit [Ping timeout: 264 seconds]
<heat>
i have a stack of 40 patches (more incoming) that needs to be pushed, its kinda hard to diagnose other problems
<heat>
although this stack is mostly unrelated to anything booting-wise
lentement has joined #osdev
gog has joined #osdev
<Ermine>
you mean that async io stuff?
<heat>
async io stuff, ext2 truncate fixes, fixes for other commits i pushed to master, vast cleanups, time fixes, slab fixes, O_DIRECT support for block devices, fixed /sbin/login having no ctty
<heat>
it started as async io but has a bunch of other somewhat unrelated patches in between
lentement has quit [Ping timeout: 260 seconds]
<heat>
i'd say it has definitely grown to mega-branch size but the actual diff isn't /that/ large, some of the changes depend on each other
<heat>
some of the commits could definitely be cherry-picked onto master though
xenos1984 has joined #osdev
linearcannon has joined #osdev
linear_cannon has quit [Ping timeout: 264 seconds]
<geist>
heat: is this for onyx?
<geist>
i should try booting it some time. whats the oldest hardware you run on?
<heat>
yep
<heat>
should boot on anything x86_64 with sse2 and NX, i think
<heat>
i know there were some old p4s without nx, it can't run on those, not without hacks (which i can't be bothered to add, for now)
<Ermine>
geist: join Onyx cause and do your part!
<geist>
ah hmm probaby the earliest machine i have floating around that can run it is a core 2 somewhere i have
<geist>
6300 iirc
<geist>
but i have that dual nehalem xeon that might give it a workout
<Ermine>
I have a feeling that each country has such poster
<geist>
sadly i got rid of all of my 2000s era AMD machines. i mean i dont regret it but i had a decent K10 class machine in the late 2000s before switching back to intel in the early 2010s
<heat>
geist, btw, something funny i found out about: qemu arm64 is really restricted in framebuffer choices because some of the devices (e.g virtio-gpu-vga) map the fb device-nGnRE vs a cacheable mapping on the KVM side
<geist>
though honestly virtio-gpu-vga is news. what does the -vga part add?
<geist>
a plain framebuffer to go along with the object stuff?
<heat>
yeah
<geist>
oh that's nice! i'll have to look that up. plain virtio-gpu is a pita for just a screen
<heat>
there's also this "MMIO coalescing" thing that lets you not vmexit on each MMIO write, which i think QEMU uses for framebuffers
<heat>
ramfb is even simpler than virtio-gpu
<geist>
yah for the VMM in fuchsia we have multiple types of doorknobs like that, and for kickers we have an async one that doesn't stop the emulated cpu, but kicks off an ipc to the VMM
<gog>
i had a k10
<gog>
i forget which core
<gog>
i had that for a long time until building my core 2 in 20010
<GeDaMo>
Time traveller :|
lentement has joined #osdev
<geist>
yah i think it was like a Phenom II was my main desktop until i got the sandy bridge in 2012
<GeDaMo>
My main desktop is still a Phenom II :P
<geist>
there ya go. it was a fine machine, just ran pretty hot
* Ermine
gives gog a piece of cheese
* gog
is fascinated
<heat>
ches
<gog>
i almost built a phenom ii
<gog>
but the one i was looking at had tlb errata
<gog>
and i was like nah let's see what intel has
<heat>
gog rivals the entire amd engineering department
<gog>
it was a known erratum
<geist>
i remember that was the one computer i had where i had a massive cooling failure one day. the fan i had strapped on it broke free one day
<zid>
GeDaMo: I got to sqrt(3) in geometry game
<geist>
and then of course the cpu immediately went into thermal shutdown
<Ermine>
so there are options to output graphics on arm without setting up modesetting?
<heat>
yes, ramfb
<geist>
had to get a whole nother cooling situation because the plastic tab had broken off on the socket
<gog>
oops
<geist>
heat: does it show up as pci if you want it to?
<heat>
i don't think so
carbonfiber has joined #osdev
<heat>
it seems to be something you need to poke at using the fw_cfg interface, get the base address etc, and then map as cacheable
<geist>
its not even virtio? do yo uhave to detect it via device tre?
<Ermine>
heat: is this universal or qemu-specific?
<geist>
ah, hmm. okay. well i guess that's not too bad
<heat>
qemu
<heat>
basically the big difference is that both sides agree that it's cacheable
<geist>
worth writing a driver for
lentement has quit [Ping timeout: 246 seconds]
<heat>
you can also hope for armv8 FWB it seems
<heat>
(friends with benefits UwU)
<geist>
re: stacking caching parameters on KVM there's a whole pile of logic as to what happens there on real hardware, but the general idea is it'll take the strictest of the parameters
<geist>
when you hvae S1 and S2 pages with different cache bits
<heat>
isn't it super UNDEFINED?
<geist>
though i think as you pointed out before, deviece vs normal memory... that may be more complicated
<geist>
it's probably what they call somethingl ike CONSTRAINED UNDEFINED
<geist>
like it might not do what you want but it wont format your hard drive
<heat>
HCR_EL2.FWB. This allows stage 2 to override the stage 1 attribute, instead of regular attribute combining.
<heat>
this is what i was talking about
<geist>
like it may fail or not fail with one of X things, but not outside of that list
<geist>
but i dont remembe roff the top of my head, i'll look it up later
<heat>
yeah it might be the "regular attribute combining" thing, not sure how useful that is
<geist>
i can see device attribute combining in a clean way because the RGE stuff
<geist>
but device on top of normal memory or vice versa i think would be weird
<heat>
yeah worst case you fall back to nGnRnE
<geist>
it's probably like normal and device mixes in a bad way, but not a fatal way, just not really a productive way
<bslsk05>
bugzilla.redhat.com: 1679680 – RFE: please support the "ramfb" display device model
<Ermine>
Advanced Virtualization
<heat>
meanwhile x86 is the most forgiving thing ever
<heat>
have a lower 1G page with tons of mixed caching attributes? don't worry, we'll handle it for you
<heat>
x86 people don't know about REAL architectures where things are UNDEFINED and IMPLEMENTATION DEFINED
<Ermine>
sure, x86 people write for PROTECTED architecture
<heat>
x86 bluepilled arch vs arm64 red pilled arch
<Ermine>
seems like arm64 osen deal with stuff which is handled by hardware in x86 case
lentement has joined #osdev
<geist>
sort of? depends on where you look
<geist>
for the memory stuff the big difference is ARM does caching paramters 100% via page tables, instead of via physical address
<geist>
which means you get this much more complicated interaction of stuff
<heat>
not only that, you get a lot more granularity in arm64
<geist>
yah, so it makes for a much more complicated problem space
<gorgonical>
even some caching features that are totally unused
<geist>
and to make that simpler they defined a lot of bits of it as UNDEFINED or basically there be dragons there
voidah has quit [Ping timeout: 256 seconds]
<heat>
also GOD I HATE THE PXN vs XN stuff
<heat>
why would anyone
<geist>
like not all the permutations are useful and you shouldn't try to use them
<gorgonical>
personally I hold a grudge about outer shareable
<geist>
yah that i think they've pretty much declared as unused
<geist>
a little surprised they didn't remove it in v9, but really v9 is not so much a full release as a snapshot of v8
<gorgonical>
I have been told a lot of the motivation for v9 was somewhere new they could shove in the CCA stuff
<geist>
and i guess define SVE as the new baseline
<geist>
since i'm sure they're feeling some amount of heat in server space for beefier vector stuff
<heat>
sve is absolutely insane
<heat>
less insane than riscv V, but still insane
lentement has quit [Ping timeout: 252 seconds]
<heat>
RISC vendors make an easily usable vector extension challenge [IMPOSSIBLE] [GONE WILD]
<gorgonical>
I have heard complaints about risc-v's vectors
<gorgonical>
But I don't understand the beef
<geist>
oh i wouldn't call it insane, but it's clearly something your compiler should generate for you
<geist>
what is bad about the rv vector stuff is the docs are pretty hard to read
<geist>
it has the 'written by math major' problem
<geist>
i hate that
<geist>
if you ignore the whole horizontal stride stuff you can basically use it like a standard '32 128 bit registers with a bunch of fairly standard vector instructions'
<geist>
the horizontal stride stuff is where it gets weird
k_hachig has quit [Ping timeout: 256 seconds]
colona_ has quit [Ping timeout: 255 seconds]
goliath has joined #osdev
lentement has joined #osdev
Left_Turn has joined #osdev
Turn_Left has quit [Ping timeout: 256 seconds]
lentement has quit [Ping timeout: 260 seconds]
CryptoDavid has quit [Quit: Connection closed for inactivity]
zetef has joined #osdev
lentement has joined #osdev
lentement has quit [Ping timeout: 268 seconds]
colona has joined #osdev
masoudd_ has joined #osdev
masoudd has quit [Ping timeout: 256 seconds]
* zid
mashes refresh
lentement has joined #osdev
lentement has quit [Ping timeout: 255 seconds]
netbsduser has quit [Ping timeout: 264 seconds]
zetef has quit [Ping timeout: 268 seconds]
k_hachig has joined #osdev
k_hachig has quit [Client Quit]
netbsduser has joined #osdev
GeDaMo has quit [Quit: That's it, you people have stood in my way long enough! I'm going to clown college!]
Turn_Left has joined #osdev
lentement has joined #osdev
Left_Turn has quit [Ping timeout: 272 seconds]
<heat>
linix
<Ermine>
hi
lentement has quit [Ping timeout: 260 seconds]
<Cindy>
hi
<Cindy>
i find eclipse to be a better editor than vscode
<Cindy>
eclipse looks classic, like that mid 2000's UI
<Cindy>
vscode looks too modern that it's just disgusting
<Ermine>
I'm fine with vscode looks
<Cindy>
eclipse has that icon style i miss so much
<Cindy>
not flat shit
<Cindy>
but actually colorful
<heat>
what the heck
<heat>
bad opinion, sorry
<heat>
i joke about the m68k vs x86 stuff, but saying eclipse is a better editor than vscode is just wrong
<Cindy>
tell me why
<heat>
eclipse is slow, old, unstable and not an editor
<heat>
vscode is okay-ishly fast, new, stable, an editor, and has many many many more plugins than eclipse
<Cindy>
>not an editor
<heat>
eclipse is also mega-cluttered with tiny menus and dropdown bars and everything that's remotely unhelpful to finding stuff
<Cindy>
what does old mean in this context
<Cindy>
it's still being updated
<heat>
has been far surpassed by intellij IDEA for years and years
<heat>
meanwhile eclipse looks exactly the same as 10 years ago
<Cindy>
>intellij IDEA
<heat>
feels like it to
<heat>
too*
<Cindy>
bro, you got to be kidding me
<heat>
no
<heat>
you want a sane java IDE, you use IDEA
<Cindy>
IDEA takes up all your RAM
<heat>
oh and eclipse doesn't?
<Cindy>
yes
<heat>
lol
<Cindy>
when i usd intellij's IDEs, i had to close everything down
<Cindy>
like only use its IDE and nothing else
<Cindy>
with eclipse, i can have a browser, irc client, PDF reader, video player all running at the same time
<Ermine>
Onyx Studio when
<Cindy>
in a 4GB RAM environemtn
<Cindy>
it's the same shit with vscode
<Cindy>
vscode is bloated as fuck
<Ermine>
still, eclipse is java, not something you want to save ram
<heat>
congrats, eclipse 10 years ago fucking blew up 4G of ram
<heat>
maybe it changed, i'd be happy for eclipse, but i sincerely doubt it
<Cindy>
i literally have eclipse running right now
<heat>
cool, i have vscode
<Cindy>
with tor browser
<Cindy>
librewolf
<Cindy>
yes, TWO browsers running at the same time
<Cindy>
a pdf reader
<Cindy>
video player
<Cindy>
and XMPP client
<heat>
uh, i found your RAM problem?
<Cindy>
and it barely took up all my RAM
<Ermine>
heat: 10 years ago I've run Eclipse on a laptop with 1.5G RAM
<Cindy>
also eclipse's UI is better to look at, tbh
<heat>
my crummy 4G of RAM laptop with a fucking celeron passed away at the sight of eclipse
<Cindy>
^ skill issue
<Ermine>
The version of Eclipse known as ADT
[_] has joined #osdev
<heat>
oh ADT was even worse
<heat>
good luck if you wanted to have eclipse open while opening an AVD
<Ermine>
my laptop had core 2 duo tho
<Cindy>
heat's issues are skill-related
<Cindy>
how come it's running in my laptop?
<Cindy>
with all of the shit running at the same time
<Cindy>
with the swap barely used
<Ermine>
and avd 2.1 worked fine. Idr if it was at the same time
<heat>
maybe they unfucked it in the meanwhile
<heat>
how much ram is it using up atm?
<Cindy>
1GB
<heat>
my vscode instance is using 600
<Cindy>
i got all kinds of files open
<heat>
me too
<Cindy>
like 16 files
<Ermine>
last time i've looked at vscode it was using 1G
<Cindy>
one thing i like about eclipse
<Cindy>
it can actually open very large files
<Ermine>
and I usually don't have many files open
<Cindy>
without trying to read it all into RA
<Cindy>
RAM*
<Ermine>
Well gals
<Cindy>
and when i tried vscode
<heat>
you are sincerely the first person i've encountered that said "you know, eclipse is great"
<Cindy>
when i opened the same large file
<Cindy>
it took up all my RAM
<heat>
there's a first time for everything i guess
<Ermine>
let's stop discussing editors for nth time
[itchyjunk] has quit [Ping timeout: 260 seconds]
<Cindy>
and froze my system
<heat>
Ermine, IDEs
<heat>
eclipse is not an editor
<Cindy>
BECAUSE IT TRIED TO LOAD IT INTO RAM
<Ermine>
heat: nitpicking
<heat>
no, it's really important
<heat>
vim is not as feature-packed as visual studio community so comparing the two is useless
<Ermine>
with plugins you can turn any editor into an ide
<heat>
you can try
bnchs__ has joined #osdev
<Ermine>
I can
Cindy has quit [Killed (NickServ (GHOST command used by bnchs__))]
bnchs__ is now known as Cindy
<Ermine>
I don't want to
<Cindy>
heat: while at the same time
<heat>
i would say it's not exactly the same thing
<Cindy>
eclipse could load a very large file, because it was designed to
<heat>
ok
<Cindy>
it has scalability mode
<heat>
less(1) also scales
<Ermine>
Btw, vs fits into 300mb
<Cindy>
less can't edit
<heat>
nano!
<Cindy>
vscode is barely scalable
<Cindy>
it wasn't designed to open very large files
<heat>
ohno
<Cindy>
which makes it shit in comparsion to eclipse
<Ermine>
Why would you open very large files anyway
<Cindy>
Ermine: i have an auto-generated file
<heat>
one case: logs
<heat>
i don't see any other use for it
<Cindy>
that i wanna see if it's all good
<heat>
how large is your very large file
<Ermine>
large log file means you're not rotating your logs properly
<Cindy>
80MB
<heat>
vscode can totes open that
<Cindy>
depending on how much RAM you have
<heat>
Ermine, true, but i work with big ass dumps directly from serial
<Cindy>
like, any program can load a shit ton of data into RAM
<Cindy>
as long as you have enough RAM for it
<heat>
in any case ERMINE CONTRIBUTE CONTRIBUTE TO ONYX NOW