<bslsk05>
news.ycombinator.com: We are moving past legacy Freenode to a new fork | Hacker News
<klange>
Amazing, I thought for sure they'd try to hold onto the last remnants of the network they pilfered, but it seems my joking suggestion that all we lost was the name "freenode" was more prescient than humorous.
<clever>
air: yikes!!
mniip_ is now known as mniip
<clever>
2021-06-14 22:49:53 < microfracture> for those that haven't already you need to reconnect to freenode and reregister your nicks and channels. freenode has started over
<geist>
holy crap, they're not kidding. my irccloud account can't connect to freenode anymore
<clever>
air: welp, nothing is left now, it cant be called freenode
<air>
geist: they are offering refunds
smeso has quit [Quit: smeso]
superleaf1995 has joined #osdev
<superleaf1995>
qemu does not clear the RST bit on CMD after a soft reset
<superleaf1995>
the spec says "yeah dude, use this rst thing and set it to 1 bro, i will reset i promise bro" - but it never clears such bit :/
smeso has joined #osdev
srjek has quit [Ping timeout: 272 seconds]
xenos1984 has quit [Ping timeout: 244 seconds]
xenos1984 has joined #osdev
iorem has quit [Quit: Connection closed]
Vi1i has joined #osdev
Vi1i has quit [Remote host closed the connection]
mashwell has joined #osdev
mashwell has quit [Remote host closed the connection]
mashwell has joined #osdev
<johnjay>
one of those questioning your sanity moments
<johnjay>
virtualbox installs debian fine in my home directory but not on an external drive
<johnjay>
in the latter case i get a blank screen on boot up
<johnjay>
my cpu is at 100% on all 4 logical cores
<johnjay>
so something's up here
<mashwell>
external drive filesystem?
<johnjay>
yes i have a WD external drive i like to use for large files
<johnjay>
it's an IDE but still. shouldn't be this slow
<mashwell>
like, what's it is formatted in? ntfs? fat32? ext4?
<johnjay>
the weird thing is i created a debian image and said to put it in Documents and that is fine
<johnjay>
ntfs
farcas has quit [Read error: Connection reset by peer]
<johnjay>
i find this frustrating since the reason i got this drive is to hold large things like VM images...
<klange>
(* this only really works at all because slirp is in the middle giving me a reasonably reliable connection)
sortie has joined #osdev
wgrant has quit [Quit: WeeChat 2.8]
wgrant has joined #osdev
wgrant has quit [Client Quit]
wgrant has joined #osdev
piotr_ has joined #osdev
kwilczynski has quit [Changing host]
kwilczynski has joined #osdev
KidBeta has joined #osdev
Lucretia has quit [Quit: Konversation terminated!]
elastic_dog has quit [Ping timeout: 272 seconds]
Lucretia has joined #osdev
GeDaMo has joined #osdev
<GeDaMo>
I logged into Freenode yesterday to delete my account but apparently I needn't have bothered because now they've all been deleted :|
elastic_dog has joined #osdev
<klange>
I'm just waiting for Lee to rename the whole thing so he'll walk away having gained nothing of value from this, and destroying a staple of open source in the process.
<clever>
GeDaMo: they dont know how to manage the software stack, so they opted for the nuke it all solution, and spun up an entirely new network
<klange>
Of course, because everyone who actually ran the network and had any _real_ authority over it left.
<clever>
i'm surprised any of the people donating server hw still is
<klange>
It's like changing the locks on a house you forged the deed to...
<clever>
klange: i think its more like they cant even find the lock on the pet door, and now they are getting racoons in the kitchen (spammers)
<clever>
rather then find the instructions, they just knocked down the whole house
<GeDaMo>
They're trying to build a new house while still in the process of knocking down the old one :P
<bslsk05>
www.devever.net: Freenode commits suicide, is no longer a serious IRC network
<clever>
2021-06-15 06:10:31 < opal> oh they have classic.freenode.net which is the old freenode/services config
<clever>
2021-06-15 06:10:36 < opal> lmao what a migration
<clever>
GeDaMo: you can still access the old services, on that domain
<GeDaMo>
Freenode classic? Did they not learn from Coca Cola? :P
<clever>
:D
<sortie>
klange, oh jeez. A rename seems.. soo.. likely
<sortie>
I suppose in that case the only value they have left is users.. and damaged ‘competition’
klange-toaru has joined #osdev
elastic_dog has quit [Ping timeout: 244 seconds]
<klange-toaru>
Oh neat this still works.
<klange-toaru>
It's absolutely terrible, but it works.
<klange-toaru>
I wanted to write a fancy client using bim as a base, I should look into that again.
gareppa has quit [Quit: Leaving]
klange-toaru has quit [Client Quit]
mjg has joined #osdev
dennis95 has joined #osdev
<mjg>
you guys not going back to freeonde? :)
<GeDaMo>
Which one? :|
<clever>
mjg: thats not freenode anymore
<mjg>
's the joke
<klange>
What do you mean 'back'? We're here right now.
elastic_dog has joined #osdev
<sortie>
Libera is a transitional place before we transition #osdev to the federation of hobbyist operating systems ircds
<mjg>
ye is there a channel for hobby irc servers?
<Mutabah>
#freenode?
* Mutabah
vanishes into commute
<sortie>
mjg, I did set up #osdev on irc.sortix.org with that purported purpose lol
Arthuria has joined #osdev
<sortie>
But #solanum is the development channel of the ircd used by libera, that's the one I ported
<sortie>
Mutabah, I'm almost afraid to join #freenode on this new network
* Mutabah
is away (Hometime!)
dormito has quit [Ping timeout: 244 seconds]
elastic_dog has quit [Ping timeout: 272 seconds]
gog has joined #osdev
piotr_ has quit [Ping timeout: 272 seconds]
<klange>
I'm going to end up writing my own curses replacement aren't I...
<GeDaMo>
This is about writing a simple text editor but it doesn't use curses, it manually sets raw mode and outputs ANSI escape codes https://viewsourcecode.org/snaptoken/kilo/
<bslsk05>
viewsourcecode.org: Build Your Own Text Editor
<klange>
I have, of course, done all that already ;)
<GeDaMo>
:P
froggey has quit [Quit: leaving]
froggey has joined #osdev
dormito has joined #osdev
HeTo has quit [Remote host closed the connection]
HeTo has joined #osdev
gareppa has joined #osdev
elastic_dog has joined #osdev
* klange
tags a beta release
<klange>
SMP is still flakey, network stack is a horrid little PoC+PoS, but in ideal conditions the package manager works and you can install Quake, so I think we're ready for at least a preview release of the long-awaited ToaruOS 2.0.
<sortie>
Boom klange slaying
<sortie>
klange, you've been on fire lately :) It's great to see you in a productive place where you get creative osdev work done
elastic_dog has quit [Ping timeout: 272 seconds]
<sortie>
It's all really inspiring and motivating to see y'all share your cool work in here
gareppa has quit [Quit: Leaving]
gareppa has joined #osdev
<gog>
wish i had some cool work lately
<gog>
mostly it's coming up with a new experiment and then quickly abandoning it because i can't work for more than 5 minutes apparently
<sortie>
That's natural and part of the process :)
<sortie>
I do that all the time still, the secret is just to have a project of so ridiculously large scope that whatever silly idea I fancy on any given day can somehow be worked into the project
<sortie>
I just wanted to make a game mod... and so I built a Steam-like platform for updating it.. which turned into a whole cross-platform development framework... now I am making an operating system
<GeDaMo>
When do you start making your own CPU? :P
<sortie>
... don't tempt me into making a compiler
<klange>
I just wanted to stick it to The Man [the CS department heads who said I couldn't take the ECE operating systems course because it wouldn't count for credit]
<froggey>
sortie: do it
<klange>
2.0 beta release dependencies were: `fetch` works in ideal conditions, CD BIOS loader works across the Big Three (vbox, qemu, vmware), and tested ports for gcc and quake.
<sortie>
klange, keynote speaker, "To all the naysayers, I present you... PonyOS"
<klange>
So nice to see that weather widget again after the last couple of months without it :)
<sortie>
That is one sexy screenshot
HeTo has quit [Ping timeout: 264 seconds]
<klange>
This tcp stack doesn't do any out-of-order packet correction, so it's prone to stream corruption even under SLIRP, just like the good old days of very early toaru32, but if the stars align it can manage to download gcc.
<sortie>
Oh cool I see you're doing .iso builds now and not disk images
<klange>
I have been doing ISO builds for years!
<sortie>
Yeah I just recall disk images at one point
<klange>
1.0 released with a (grub) ISO, I think I've been doing CDs for as long as I've had CI.
<sortie>
The kernel uses multiboot so GRUB boot is still possible?
<klange>
aye
<sortie>
Is there ext2 support in your new version?
<klange>
j`ey: ew, what card is that configured for? looks like it failed to modeset but the compositor is unaware
<klange>
sortie: nope, on the list of things that didn't get ported and probably should be thrown out and redone
<sortie>
Ah gotcha so trying to hack up a toaru disk installation is out of the question right now
<j`ey>
klange: VMVSGA
<j`ey>
sorry VMSVGA
<sortie>
But with networking a live environment is still useful
<klange>
j`ey: ah, hm, might be missing pciids for that; try VBoxVGA or VBoxSVGA.
<klange>
vmsvga is vmware's thing, and at least current vmware player works, but *shrug*
<j`ey>
vboxvga works
<klange>
Other/Unknown (64-bit) should give at least a viable experience, though I think vbox still defaults to something other than the Intel NIC and likes to turn off audio output even when it sets up an AC'97, so YMMV on configs.
<klange>
sortie: No block device drivers yet, either, so it's a moot point; ask again in a few weeks and I should at least have gotten around to some rudimentary AHCI drivers.
<klange>
I even stripped the hacky little ATAPI driver out of the loader, it just plops the entire CD into memory before jumping to the menu.
<sortie>
As a simple “one of these decades” person your “in a few weeks” frightens and confuses me
<klange>
Hey, it did take a decade to get around to that new-fangled 64-bit thing.
<klange>
try a leading http:// the URL parser is super dumb
<j`ey>
oh ok
<j`ey>
also is there a 'top' or some kinda activity monitor?
<klange>
There's a ps but there's no CPU usage monitoring at the moment and I'm all sorts of exhausted trying to think about how to do it well with SMP... or really at all...
<j`ey>
also the version says 1.99-dirty.. so did you build the image from a non-clean git checkout?
<klange>
Did _github_ build it from a non-clean checkout? I'll have to see what it goofed...
<klange>
friggin what
<klange>
did the 1.14 build do that...
dennis95 has quit [Remote host closed the connection]
dennis95 has joined #osdev
<klange>
1.14.1 is clean, what could I build doing that's not in the .gitignore...
* klange
runs the docker build and sees what happens
<klange>
oh shoot you know what, it might be unclean because of how the workflow clones submodules
elastic_dog has quit [Read error: Connection reset by peer]
<klange>
ah, yeah, I'll have to look into that
elastic_dog has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
isaacwoods has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
dj2 has joined #osdev
nyah has joined #osdev
dj2 has quit [Quit: Client closed]
elastic_dog has joined #osdev
gareppa has quit [Quit: Leaving]
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
ahalaney has joined #osdev
HeTo has joined #osdev
skipwich has quit [Read error: Connection reset by peer]
skipwich has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
ahalaney has quit [Ping timeout: 268 seconds]
srjek has joined #osdev
KidBeta has quit [Ping timeout: 272 seconds]
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
vai has joined #osdev
rorx has quit [Ping timeout: 272 seconds]
ahalaney has joined #osdev
gareppa has joined #osdev
Arthuria has quit [Ping timeout: 265 seconds]
ahalaney has quit [Quit: Leaving]
froggey has quit [Remote host closed the connection]
froggey has joined #osdev
ahalaney has joined #osdev
nyah_ has joined #osdev
nyah has quit [Ping timeout: 265 seconds]
gareppa has quit [Ping timeout: 265 seconds]
elastic_dog has quit [Ping timeout: 265 seconds]
rorx has joined #osdev
gareppa has joined #osdev
GreaseMonkey has quit [Ping timeout: 244 seconds]
elastic_dog has joined #osdev
<johnjay>
periodically i try debian stable
<johnjay>
then immediately i get something like, gtk4 requires meson >= .54 but it has meson .4x
<johnjay>
then i remember why i never use debian-stable
srjek|home has joined #osdev
srjek_ has joined #osdev
srjek has quit [Ping timeout: 240 seconds]
srjek|home has quit [Ping timeout: 264 seconds]
f11 has joined #osdev
srjek_ has quit [Ping timeout: 240 seconds]
iorem has quit [Quit: Connection closed]
asymptotically has joined #osdev
f11 has quit [Ping timeout: 268 seconds]
f11 has joined #osdev
dennis95_ has joined #osdev
dennis95 has quit [Read error: Connection reset by peer]
dennis95_ has quit [Remote host closed the connection]
gareppa has quit [Quit: Leaving]
rorx is now known as osxbot
osxbot is now known as rorx
dennis95 has joined #osdev
flx has quit [Ping timeout: 272 seconds]
rorx is now known as osxbot
osxbot is now known as rorx
rorx is now known as osxbot
mahmutov has joined #osdev
osxbot is now known as rorx
superleaf1995 has joined #osdev
<superleaf1995>
rst bit not being clear on my rtl8139 still
<superleaf1995>
even after powern on and soft reset
<superleaf1995>
wtf is happening
superleaf1995 has quit [Quit: Client closed]
f11 has quit [Ping timeout: 240 seconds]
f11 has joined #osdev
Arthuria has joined #osdev
f11 has quit [Quit: WeeChat 3.1]
mctpyt has joined #osdev
<geist>
johnjay: oh versus unstable?
<johnjay>
no vs testing. wh enever i've tried to use stable i can never do anything
<geist>
or testing yeah
<johnjay>
not even the build system is up to date. a *build* system. o_o
<geist>
i mostly use ubuntu, which is i guess based on -testing?
<johnjay>
you know you've reached the end of the road at that point
<johnjay>
that would make more sense
<johnjay>
my guess is debian just has too rigorous standards or something
<johnjay>
even emacs documentation i have to get from the non-free repos
<geist>
from grepping the internet it looks like ubuntu is more complicated than that, but in general LTS releases are based on testing and non are based on unstable
<geist>
but strictly speaking it's not downstream from debian
<johnjay>
interesting
<johnjay>
so ubuntu is like "the usable debian" now
<mjg>
i thought using debian is for hipsters these days
<mjg>
same with arch and others
asymptotically has quit [Remote host closed the connection]
asymptotically has joined #osdev
<kazinsal>
debian is too stable to be cool
<Bitweasil>
How dare anyone want their stuff to just work.
<Bitweasil>
That's certainly not modern.
Arthuria has quit [Read error: Connection reset by peer]
<Bitweasil>
If it's not going down every two days for updates that break something, you just don't know what, is it even a real computer?
<Bitweasil>
:/
Arthuria has joined #osdev
<mjg>
i have been using ubuntu for over 4 years now, never had an issue with an update
<mjg>
(but i do know other people had)
matt|home has quit [Read error: No route to host]
<HeTo>
I have to say, getting Kubuntu installed with disk encryption was way, way harder than doing the same with Debian
<geist>
yeah same. though generally I only stick to the LTS releases for machines i care about, or mint linux for desktops
<HeTo>
and Ubuntu was supposed to be the easy one
<geist>
yah their installer i dunno
<Bitweasil>
Yeah, LTS for anything I care about, I'll run the 6 month stuff on toys, but that's about it.
<mjg>
i don't remember the details, i do have fde and i'm pretty sure it was just few clicks
<HeTo>
maybe it was because I wanted to set up the LVs as I wanted
<HeTo>
and also keep Windows
<HeTo>
so I couldn't use the one-click fde
<mjg>
i did use the entire disk, so that's that
paulbarker has quit [Quit: Updating details, brb]
paulbarker has joined #osdev
<mjg>
as in i can easily believe not going with the default runs into unfunny issues
<HeTo>
with the Debian installer, you could just choose LVM with encryption and set up the partitions and the volumes within the installer. with Kubuntu you had to use the command line on the live cd to create a cryptsetup partition and the LVM PV, VG and maybe LVs as well and then convince the installer to use those without crashing it
<mjg>
:)
<mjg>
installers suck
<kazinsal>
yeah generally speaking there's no good way to write an installer for an OS that isn't eventually going to be slightly crap to someone in some way
<kazinsal>
windows' installer is straightforward but doesn't expose enough features, openbsd's installer is straightforward but the exposure of features is "break into the shell and do it yourself", etc
<kazinsal>
and if you add too many buttons to do things now you're going to scare off normal users
brynet has quit [Quit: leaving]
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
mahmutov has quit [Ping timeout: 240 seconds]
GreaseMonkey has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
theruran has quit [Quit: Connection closed for inactivity]
Ameisen has joined #osdev
brynet has joined #osdev
abbie has quit [Quit: Bye!]
abbie has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
<geist>
oh wow, the freenode implosion is deeper than i thought
<geist>
apparently they've thrown out all the nick and chan registrations, etc
<geist>
without any real warning either
<GeDaMo>
There are now two Freenodes :P
<HeTo>
two freenodes plus probably a couple servers that are eternally netsplit unless they've shut those down
<GeDaMo>
The new one is running on different software and has none of the nick or channels
<geist>
yah was just reading the post
<geist>
oh that's a good question. do 'they' even have logins to the 'old' servers
<HeTo>
and there was no notice of it if you happened to be on one of those netsplit servers
<geist>
if they dont, then this whole forced migration may be a way to force a split
<GeDaMo>
There's now a classic.freenode.net
<HeTo>
the new network is running a new software stack. they just couldn't or didn't want to write code to do the migration (well, given that it seems rushed, they probably couldn't in the timeframe)
<HeTo>
GeDaMo: and at least some hours ago that round-robin could still lead you to one of the netsplit servers that don't have access to services
<GeDaMo>
Yeah, I think they haven't got the DNS sorted out yet
<geist>
it's also possible that some of the old servers they dont have logins to
<geist>
and they'll just remain split off like that until someone actually disconnects them
<HeTo>
yeah, could be, something like that would explain why they couldn't resolve the netsplit
<geist>
exactly
<HeTo>
since this happened a short while after a netsplit began
<Ameisen>
HeTo - well, the logs from the chats about that are... interesting.
<Ameisen>
Evidently, people outright offered to do the data migrations, and were refused because "losing all the data is part of the point"
<HeTo>
Ameisen: oh, interesting, I've heard IRC software developers generally don't want to work with the new Freenode staff, for obvious reasons
<bslsk05>
twitter: <ariadneconill> freenode finds itself in a strange situation. as far as i know, all irc developers have basically refused to help andrew. haven’t heard from syzop yet, that might be his best shot really.
<HeTo>
(they're running inspircd and anope on the new network anyway AFAIK)
carmysilna has quit [Quit: Client limit exceeded: 10000]
<Bitweasil>
Just when you thought the train wreck couldn't get any better, along comes a steel spike and punctures those propane carriers in a shower of sparks...
carmysilna has joined #osdev
<geist>
haha
<Bitweasil>
In comes Lee, "Waaaaaaaiiiiitttttttt! You forgot the thermite! YOU FORGOT THE THERMITE!"
<Bitweasil>
I cannot wrap my head around how any of this makes any sense.
<Bitweasil>
I guess I'm just not fake enough royalty?
<bslsk05>
twitter: <carlaas> Freenode's time of death 12:26:55 PST   <rasengan> Decision has been made  <rasengan> We have to drain this swamp  <rasengan> Classic tried to go on for the sake of the community, but the community has made it loud and clear-they want freenode classic to end, today!  <rasengan> #MFGA
<GeDaMo>
"last message sent on real freenode was <girst> fuck you rasengan for destroying freenode <3"
[com]buster has joined #osdev
gmacd has joined #osdev
GeDaMo has quit [Quit: Leaving.]
<HeTo>
depending on what's the real freenode there's still chat going on on capone
<HeTo>
WARNING it has no services but has impersonators on those nicks
<HeTo>
and depending on your DNS you might not get capone but the new Freenode
* dissident
scrolls up in chat. dear lord, i'd almost forgotten about jerkcity
<Bitweasil>
I don't get it.
froggey has quit [Ping timeout: 268 seconds]
srjek_ has joined #osdev
<dissident>
cartoon from back in the day. had a friend who was really into it. i didn't get it, either
<dissident>
was done with some sort of graphical chat program that ran on irc
<dissident>
guy took screenshots to make the cartoon
merry_ has joined #osdev
merry has quit [Ping timeout: 244 seconds]
merry_ is now known as merry
gmacd has quit [Remote host closed the connection]
froggey has joined #osdev
<HeTo>
hah the person squatting nickserv has +R. probably good
<HeTo>
on capone that is
<HeTo>
means if you try to identify or whatever the server tells you to identify first
<geist>
ms comic chat. funny, you can still get it and use it
<geist>
it connects to irc
<geist>
it's an old classic win32 exe so actually runs under wine pretty well too, iirc
uplime is now known as rowbot
rowbot is now known as uplime
dennis95 has quit [Quit: Leaving]
Arthuria has joined #osdev
flx has joined #osdev
ahalaney has quit [Quit: Leaving]
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
Arthuria has quit [Read error: Connection reset by peer]
dj2 has joined #osdev
Arthuria has joined #osdev
sortie has quit [Quit: Leaving]
dj2 has quit [Client Quit]
<HeTo>
probably also full of security holes
<moon-child>
I remember someone once did an experiment: they made bots with nicks that were a typo away from nickserv. Like nivkserv. They were able to snoop surprisingly many passwords that way
<moon-child>
it seems like a bit of a design flaw that authentication isn't built into the platform. (Sasl is a thing, but it's not pervasive)
<moon-child>
s/platform/protocol
Arthuria has quit [Read error: Connection reset by peer]
Arthuria has joined #osdev
<clever>
moon-child: most irc networks also have a /ns command
<clever>
so you can /ns identify password
<clever>
on some irc servers, you can also configure it so it will only ever msg the right nickserv
<clever>
this reserves a nick for a certain server, and will reject anybody attempting to claim it
<clever>
moon-child: the fact that any of these things are possible, implies that the staff have zero clue what they are doing
<clever>
i never even had more then 20 users on my network, and it had better security
<moon-child>
clever: /ns is a property of the client, not the server. And many clients also let you insert your password in some kind of settings menu and they'll PRIVMSG Nickserv :IDENTIFY ... ... for you
<danieldg>
it's also possible to set up a network to forward the "server password" to nickserv, which fixes the design flaw by actually using the protocol-defined password
<clever>
moon-child: /ns can be both client or server
<moon-child>
but my point is that reserving NickServ doesn't solve the fundamental problem. Somebody might page NivkServ. THe problem being that authentication
<clever>
via the ServAliases config on my irc server, its a server-side command
<moon-child>
clever: /anything is client
<moon-child>
til ServAliases
<clever>
moon-child: any command the client doesnt support, just gets dumped to the server naked, for most clients
<moon-child>
what now? I have never had a client that did that
<clever>
irssi is one of the few that wont do that
<danieldg>
some clients require /raw ns identify ...
<clever>
moon-child: but, is that unknown command coming from the client, or the server?
<moon-child>
client
<clever>
try /raw blabediblu
zoey has joined #osdev
<clever>
or /quote blabediblu
<moon-child>
then the server complains
<clever>
now try /quote ns help
<moon-child>
hmm, looks like hexchat will do as you say and forward unknown commands directly
<clever>
yep
Arthuria has quit [Read error: Connection reset by peer]
<gorgonical>
Boy u-boot and qemu are really getting the better of me
Arthuria has joined #osdev
<moon-child>
pidgin won't, though
<gorgonical>
Can't seem to make it boot my kernel image... Just to make sure, the load address and entry point are where u-boot should relocate the kernel to, right?
<clever>
moon-child: yeah, it varies from client to client
<clever>
2021-06-15 20:21:15 ns quote ns $0-
<clever>
moon-child: i had to add a /ns alias to irssi, to make it do /quote ns .... behind the scenes
<moon-child>
I just let znc do it for me. Never really used any clients aside from irssi and weechat, so TIL about that behaviour
<doug16k>
gorgonical, what do you mean "relocate"
<moon-child>
it seems ... not great? IRC is relatively regular, but not _that_ regular
<gorgonical>
E.g. in the u-boot environment you point it at your kernel address, the ramdisk address, and then your FDT address
<gorgonical>
But I'm not sure if this address *must* be identical to the load address/entry point that are given by the mkimage command when you make the uImage
<doug16k>
it's a binary, with a 64 byte header? or elf?
<gorgonical>
binary
<gorgonical>
err which one?
<doug16k>
your uboot thing
<gorgonical>
that's a binary, yes
<gorgonical>
that's passed with -bios to qemu-system-aarch64 to boot the u-boot environment
<doug16k>
then qemu will slap it in memory at the load address, and jump to the entry point, I never say any relocating
<doug16k>
saw
<gorgonical>
right, so then u-boot does some setup of its own
<bslsk05>
pastebin.com: => bootm 0x200000 - $fdt_addr## Booting kernel from Legacy Image at 00200000 . - Pastebin.com
<doug16k>
load address says 0
<doug16k>
entry point is 0?
<gorgonical>
No it says 0x80000
<gorgonical>
and entry point too
<doug16k>
oh my bad
<doug16k>
stupid font
<doug16k>
blind spot makes some of the zeros in your peripheral vision turn into 6
<gorgonical>
If u-boot is placed at 0x0 then it will go past 0x80000, but I figured the bootloader didn't need to stick around once it loaded the kernel etc
<gorgonical>
Maybe that's wrong?
<gorgonical>
Also, do I need to put the kernel image at 0x80000 if that's where this image says for the load address/entry point?
<doug16k>
info mtree
<doug16k>
look under cpu-memory-0
<doug16k>
what's at 0x80000
<gorgonical>
virt.flash0, but indistinctly
<gorgonical>
it's just at the low end of virt.flash0
<gorgonical>
Should this be done after loading the image in u-boot?
<doug16k>
yes
<gorgonical>
bootm loados 0x200000?
<doug16k>
why can't you load it at its natural address
<gorgonical>
there is nothing at 0x80000
<doug16k>
then don't link it there?
<gorgonical>
natural address being what exactly? the 0x80000 or 0x20000?
<doug16k>
base address
<doug16k>
load address
<doug16k>
natural being, where the linker thinks it goes
<doug16k>
"not-relocated"
<gorgonical>
This script was provided to me and so I've given it some authority in being right, but it's seeming less and less correct
<doug16k>
I don't know for sure either
<gorgonical>
The arm64 makefile I have just does mkimage and puts it at 0x80000 for both load address and entry point
<doug16k>
I am just going when I did the same thing
<gorgonical>
the System.map has the text section at 0x80000
<doug16k>
xp where you think it is
<gorgonical>
well rather 0xfff...c000080000
<doug16k>
VMA and LMA are not the same right?
<doug16k>
objdump -h I mean
<gorgonical>
??? qemu shows nothing at 0x200000 with xp
<gorgonical>
but u-boot md 0x200000 does
<doug16k>
what about just x
<gorgonical>
wait wait
<gorgonical>
after bootm loados it clears the previous addr
<gorgonical>
so before doing bootm loados qemu's xp does show the kernel at 0x200000
<gorgonical>
Hmm, vmlwk shows vma == lma with objdump -h
<gorgonical>
for .text section
Arthuria has quit [Read error: Connection reset by peer]
<doug16k>
up high ffff.... addresses?
<gorgonical>
yes
Arthuria has joined #osdev
<doug16k>
usually you make it so LMA is where it is in the ROM (true address), and VMA is after you did your startup tricks / virtual memory / relocating / whatever
<gorgonical>
So is the u-boot uImage format just like a header containing some metadata about the payload?
Arthuria has quit [Read error: Connection reset by peer]
<doug16k>
if you do that, then just telling it to convert -Obinary makes the correct image
<doug16k>
yes
Arthuria has joined #osdev
<doug16k>
the type field completely changes how the rest works
<doug16k>
basically it's binary or an elf
<gorgonical>
the load addr/entry point should point to the text section then
<gorgonical>
and that text section should be located at the address it says
Arthuria has quit [Read error: Connection reset by peer]
<gorgonical>
"ARM Linux Kernel Image (uncompressed)"
Arthuria has joined #osdev
<doug16k>
I believe that is correct
<gorgonical>
the kernel image itself is an elf, but the u-boot payload is binary
<gorgonical>
right
Arthuria has quit [Read error: Connection reset by peer]
<doug16k>
you extract a bin from the elf with objcopy -Obinary or something?
Arthuria has joined #osdev
<gorgonical>
yes
<doug16k>
size of bin file is as expected?
<doug16k>
if it is crazy size, you scrwed up linker script
<gorgonical>
Ah crap have I been using the wrong file?
<gorgonical>
wait wait
<gorgonical>
I don't think I have been using the wrong file
<gorgonical>
Yeah the vmlwk.bin and the uImage made with it are like 64 bytes different
<doug16k>
that's correct
<gorgonical>
I can't put the uImage at 0x80000 because the u-boot.bin is ~600kb though