channel logs for 2004 - 2010 are archived at http://tunes.org/~nef/logs/old/ ·· can't be searched
#osdev2 = #osdev @ Libera from 23may2021 to present
#osdev @ OPN/FreeNode from 3apr2001 to 23may2021
all other channels are on OPN/FreeNode from 2004 to present
http://bespin.org/~qz/search/?view=1&c=osdev2&y=23&m=3&d=17
00:02:00 <klange> If you want to avoid the NSA, you need to follow xkcd 378.
00:58:00 <kazinsal> the mossad/not-mossad threat model is eternal
00:59:00 <kazinsal> praise be unto mickens, may his light guide us through the d arkness and out the other side
01:15:00 <pog> the nsa can lick my taint
01:16:00 <kazinsal> sshhhhh, the NSA wizard may be into that
01:17:00 <pog> :9
01:34:00 <sakasama> Your request has reached upper management and they have informed me they do not require your permission to perform said licking.
15:10:00 <zid> where heat
15:10:00 <zid> https://preview.redd.it/25nuhqlrd9oa1.jpg?width=640&auto=webp&v=enabled&s=b630a09f4711d66c8dad28deeb1ed3ca3a3901e6
15:10:00 <zid> link that to him when he shows up
17:27:00 <heat> mjg, how's the status of virtualization on FreeBSD/NetBSD?
17:27:00 <heat> does bhyve(I think that's what freebsd has?) work more or less like KVM?
17:28:00 <mjg> it works fine for varous bsd and linxu guests, but it is turbo cumbersome to use
17:29:00 <mjg> imagine using qemu by hand
17:29:00 <heat> ... erm
17:29:00 <heat> we use qemu by hand here
17:29:00 <mjg> lol
17:29:00 <mjg> well there you have it
17:30:00 <heat> is it like qemu-system-x86_64 -machine pc,accel=bhyve or something?
17:31:00 <heat> did netbsd steal this from you or is it a FreeBSD exclusive
17:32:00 <mjg> netbsd has its own kvm-equivalent named nvmm
17:32:00 <mjg> which ultimatley works with qemu
17:32:00 <Ermine> Afaik netbsd added lua support to the kernel
17:32:00 <mjg> except it got abandoned by the author andi don' thin is being developed
17:32:00 <mjg> if anything dragonflybsd got a nvmm port which is *never* than the netbsd one
17:32:00 <Ermine> Sad story
17:33:00 <mjg> i don't know what bhyve is doing internally nor how does it compare to kvm
17:33:00 <mjg> freebsd does *not* use qemu here
17:34:00 <heat> so i've been thinking about installing one of the BSDs on my tiny rpi zero 2w, but i'd like to possibly run Onyx on it in the future, so one with decent hypervisor support is a good idea
17:35:00 <heat> wth does freebsd use if not qemu?
17:36:00 <puck> bhyve
17:36:00 <zid> [15:10] <zid> where heat
17:36:00 <zid> [15:10] <zid> https://preview.redd.it/25nuhqlrd9oa1.jpg?width=640&auto=webp&v=enabled&s=b630a09f4711d66c8dad28deeb1ed3ca3a3901e6
17:37:00 <zid> I need to increase my coke input
17:37:00 <zid> dr pepper not as acidic, I might still get some RUST
17:37:00 <mjg> heat: i don't think any of the bsds support virt on anythin but amd64
17:38:00 <mjg> heat: there is wip to get bhyve running on arm64 though
17:38:00 <heat> wow cringe
17:38:00 <mjg> brah
17:39:00 <heat> x86_64 (and *amd64* even more so) is fucking cringe
17:39:00 <mjg> the person who wrote nvmm used 'maxv' as a handle
17:40:00 <mjg> i suggest you check netbsd mailing list archives for some of the commentary :)
17:40:00 <heat> why?
17:40:00 <mjg> just do it
17:40:00 <heat> nike?
17:41:00 <puck> honestly surprised both amd and intel have their own virtualization extension, and even have that separation still
17:41:00 <heat> amd and intel have their own everything now
17:41:00 <heat> even the fucking address tagging extensions are different
17:41:00 <heat> it's rage inducing
17:42:00 <zid> well intel is dead now anyuway
17:42:00 <mjg> would not it be funny if the 64 bit stuff was all different
17:43:00 <zid> we're all amd only now
17:43:00 <mjg> as in both intel and amd having their own arch altogether
17:43:00 <heat> ITANIUUUUUUUUUUUUUUUUUUUUUUUUUUM
17:43:00 <mjg> x86-liek
17:44:00 <heat> mjg, how tf do you search in netbsd ml archives
17:44:00 <heat> i'm fairly sure there's no search functionality
17:45:00 <mjg> marc.info
17:46:00 <heat> aw
17:46:00 <heat> i hate that website
17:48:00 <mjg> the site hates you mofo
17:48:00 <heat> lore.kernel.org >>
17:48:00 <mjg> wait linukkz doing smething better?
17:49:00 <heat> wtf i can't find shit on that guy
17:51:00 <mjg> they don't worry about it and do smethong less not productive
17:51:00 <heat> ok dad
17:51:00 * mjg pats heat on the head
17:52:00 <mjg> techniaclly i am old enough to be your father, should i come from a pathological background
18:00:00 * sakasama hugs grandpa mjg.
18:20:00 <geist> ugh, discord is really starting to much more aggressively push nitro on me
18:21:00 <geist> more and more flashing bits that you have to dismiss
18:24:00 * zid fires up the nitro disco ball
18:25:00 <GeDaMo> What's nitro?
18:25:00 <zid> discord premium
18:25:00 <GeDaMo> Ah
18:26:00 <geist> they keep adding new features that they then taunt you with
18:26:00 <heat> yeah this one was annoying
18:26:00 <geist> seems to be they pushed 2 or 3 new things in the last few weeks
18:26:00 <heat> (tip: you don't need to pay to theme your discord)
18:27:00 <geist> yah their themes aren't exactly what i want anyway. pretty close, but i really want to be able to manually set the colors
18:28:00 <geist> notably i never like light text on absolute black. i always bring up the background color to some sort of light grey or whatnot
18:28:00 <geist> but seems most of their dark themes that aren't the basic 'dark' are basically black background around text
18:29:00 <geist> honestly irccloud does it perfectly: they use varying shades of dark to delineate various zones in the text, but never hard black
18:29:00 <zid> dark sidebar is still an option thank fuck
18:29:00 <GeDaMo> I hate this whole theming thing where you have to select from a list of themes or go digging around in some weird styling language
18:29:00 <zid> they removed it for a while and I had to riot
18:32:00 <geist> yah i dunno why they dont just let you individually select zone colors
18:32:00 <geist> at the end of the day it's probably just mostly CSS anyway
18:32:00 <geist> unless they have hard .png files or somewhat backing these things i guess
18:33:00 <zid> you can restyle discord with a mod
18:33:00 <zid> if you wanna rewrite their css
18:33:00 <geist> yah probably just does that right
18:34:00 <GeDaMo> I know Firefox uses CSS because I have to use it to remove annoyances :/
19:39:00 <d316d> Hey. The Linux impl of BeFS is read-only, would my patch be accepted if I added write capability for that? Or is it a total abandoware? Latest changes were in 2002
19:40:00 <heat> probably abandonware
19:40:00 <heat> but why ask us?
19:40:00 <d316d> Ask who?
21:05:00 <geist> us, as in this channel
21:05:00 <geist> oh they left
21:13:00 <mjg> i suspect there is already fuse-based befs rw
21:13:00 <mjg> and if not, the response would be to write one
21:13:00 <mjg> if anything bringing befs up might prop its removal :)
21:16:00 <geist> FWIW at least befs hasn't changed since 2002
21:17:00 <puck> ..i wonder if haiku did any upgrades
21:17:00 <geist> right, that'd be the main reason one would even want to have RW support on linux, to interop with haiku
21:18:00 <geist> otherwise RO is fine, because you use it to extract bits off your old beos disks
21:18:00 <mjg> :]
21:18:00 <mjg> if there were literaly no changes since 2002 then the thing is disconnected from the build
21:18:00 <mjg> and would require new porting effort to make it operatinal even as ro
21:18:00 <geist> haiku has a full driver which would probably be easy to wrap in a user space linux driver
21:19:00 <puck> oh actually
21:19:00 <mjg> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/log/fs/befs
21:19:00 <bslsk05> git.kernel.org: kernel/git/torvalds/linux.git - Linux kernel source tree
21:19:00 <puck> i think haiku comes with a bfs fuse driver already
21:19:00 <mjg> clearly it is getting changes, but it also has a changelog file which is not updated since 2002
21:19:00 <mjg> perhaps that's what the above person wsa referring to
21:19:00 <mjg> would be funny if the support was in fact rw now
21:20:00 <puck> i don't think so, bfs has some design things that don't mesh well; it'd be easy to check by seeing if it knows how to handle the indexes
21:20:00 <geist> yah i think i've used that linux driver from time to time, and it works well enough. befs isn't that complicated from a RO point of view. think of it as a fairly standard extent based inode tree thing with btrees for dirs and inodes located anywhere
21:20:00 <geist> for RW the indexes are the issue, yeah
21:21:00 <puck> yeah looks like it can't do indexing
21:21:00 <geist> oh and fairly generic attributes that might not mesh will with linux's notion
21:22:00 <mjg> it is ro
21:22:00 <mjg> befs_warning(sb,
21:22:00 <mjg> if (!sb_rdonly(sb)) {
21:22:00 <mjg> "No write support. Marking filesystem read-only");
21:22:00 <mjg> sb->s_flags |= SB_RDONLY;
21:22:00 <mjg> }
21:22:00 <mjg> the fuse point stands :>
21:22:00 <geist> yah and wouldn't be surprised if attributes aren't plumbed through
21:22:00 <geist> since i suspect that was before linux had suppor for generic attributes
21:22:00 <geist> wasn't i think until a more late 2000s thing that that got serious outside of mac or beos
21:23:00 <geist> (which is not entirely unrelated, lots of beos folks were mac people)
21:23:00 <puck> yeah no attr suspport
21:23:00 <puck> geist: i do feel like beos kinda took it to the next level tho :p
21:24:00 <geist> oh yeah at the time it was quite sophisticated
21:24:00 <geist> now a lot of the stuff it did is kinda normal, but back in 1998 or so you were comparing with win95 or mac classic
21:24:00 <geist> or some ridiculously priced workstation
21:25:00 <puck> idk, i don't think any other os really did the "database" thing
21:25:00 <puck> i guess WinFS is the closest equivalent
21:25:00 <geist> oh sure i'm mostly talking about beos as a whole
21:25:00 <puck> (but inherently i do believe these kinds of schemes would fail these days; everyone would just use their own extensions instead of standardizing)
21:25:00 <geist> yah dbg basically reimplemented it at apple a few years later but as a user space thing: ie, spotlight
21:26:00 <geist> not as elegant, but basically solves the same problems, just with some lag
21:26:00 <puck> i checked my copies of the hobbit bebox prototype and they seem to not mention attrs, so that wasn't a thing they had from the start. i wonder when befs became .. befs
21:26:00 <geist> when dbg started at beos and then wrote it. i think 1996+ era
21:26:00 <puck> ahh
21:26:00 <geist> i joined in 1999 and bfs was of course finished at that point
21:26:00 <puck> i should finish my hobbit emulator
21:27:00 <geist> hobbit bebox era was think of it as beos v1, and there was a fairly substantial rewrite by v2
21:27:00 <puck> yeah. i looked at the list of hobbit bebox kcalls and they .. do not make a lot of sense
21:27:00 <geist> 95/96 they had switched to PPC and were making the ppc bebox
21:27:00 <geist> yah do you have a hobbit bebox? they're rare as heck
21:27:00 <puck> i do not
21:27:00 <mjg> half-hue https://github.com/brendangregg/pmc-cloud-tools/blob/master/icache
21:27:00 <bslsk05> github.com: pmc-cloud-tools/icache at master · brendangregg/pmc-cloud-tools · GitHub
21:27:00 <puck> i do have: very spotty hyperfocus
21:27:00 <geist> i only know of one in existance
21:27:00 <puck> and: a rust-based hobbit bebox emulator
21:28:00 <geist> i have a pair of ppc beboxen, but the hobbits there were probably 10s of them at that
21:28:00 <puck> somehow three hard drive dumps from three different hobbit beboxen appeared online
21:28:00 <puck> i sent an email to someone that allegedly had one to see if they could dump the bootrom (the one part i've stubbed out, which is annoying, because it's what parses the filesystem and loads the (ELF) kernel)
21:29:00 <geist> ah. if you run into trouble i have a buddy that has one that probably has the capability of pulling the rom
21:29:00 <geist> (unless it's the same person)
21:30:00 <geist> but i do remember there was some sort of pre-bfs filesystem. iirc the boot rom on the PPC beboxes also only knew how to parse that
21:30:00 <geist> so there was some one off tool to generate boot floppies for ppc beboxes that it recognized. i think it wasn't capable of booting from a CDROM as a result
21:31:00 <puck> this guy http://www.typewritten.org/Projects/Be/9308000058.html (last update of the site is april 2022, afaict)
21:31:00 <bslsk05> www.typewritten.org: BeBox Rev. A (Hobbit) - S/N 9308000058 - Typewritten Software
21:31:00 <puck> i wonder if i could get the ppc card emulated
21:31:00 <geist> ah not the same guy. so yeah like i said i have another buddy that might be able to get it to you as well
21:32:00 <geist> also that reminds me i should consider pulling the rom for the PPC bebox, if someone hasdn't dong it before
21:32:00 <puck> i think the ppc beboxen have firmware updates available
21:32:00 <geist> oh possible
21:32:00 <puck> but yeah, it'd be appreciated; i susspect it's not too magic, and i can just do it in rust (i'm already manually loading the ELF, and it is a normal ELF)
21:32:00 <puck> (it even has symbols!)
21:33:00 <puck> the reason i dumped all the kcalls is i want to write a userspace emulator to solve the chicken-egg problem: i have a hobbit beos C compiler. for the hobbit beos
21:33:00 <puck> some day i oughta appear a bebox here. or maybe 3d-print a replica. it's a good aesthetic
21:34:00 <zid> oh bebox is a box running beos?
21:34:00 <puck> not just
21:34:00 <puck> https://upload.wikimedia.org/wikipedia/commons/a/a4/Samsung_SyncMaster_191T_and_BeBox_20081114.jpg it is a very distinct aesthetic
21:35:00 <puck> with, yes, the first instance of modern Gamer(TM) aesthetic: two LED strips that light up based on CPU usage
21:35:00 <zid> can I turn them off? :P
21:35:00 <zid> people are finally starting to add hw buttons for this
21:35:00 <puck> iirc no, but i haven't really played with one much
21:36:00 <puck> the only time was at the home computer museum in Helmond, which had one running some beta. i accidentally KDL'd it but the KDL was on the serial port so it just visually hung
21:38:00 <kof123> you could run beos ppc on some powermac systems and some other things (used to have one) ...quite happy to see this conversation, it is the one thing i joke about "real hw i used to have, no emulator" ...
21:38:00 <geist> heh yeah. you could KDL it with a ctrl-prtscren or something
21:39:00 <kof123> probably not unobtanium, but dont know what crazy prices nowadays might be
21:39:00 <geist> the major bummer about the bebox that held it back was the cache controller was basically faulty, and there was no L2, i believe
21:39:00 <puck> hmm
21:39:00 <geist> just L1 between the cpus. so there wa a lot of cases where running both cores on a task ran < 1.0
21:39:00 <puck> i'm going to have to do some fun experimenting once i have proper multithreading working for my hobbit emulator
21:40:00 <geist> but one of the fnu things you could do with beos was in the Pulse app which was just showing a bar per cpu, you could individually turn each cpu off with a toggle button
21:40:00 <puck> because .. i'm not sure AT&T documented their cache
21:40:00 <geist> so it was fairly common to just turn off a cpu
21:40:00 <puck> kof123: oh boy i should get an old world ppc mac for that some day
21:40:00 <zid> watch out for the old world macs, they're way more venemous
21:40:00 <zid> venomous*
21:41:00 <puck> ..i have too many cursed retro projects i want to do
21:41:00 <puck> i want to port AmigaOS to 68k macs (probably also old world macs by accident, because they boot into toolbox -> 68k emulation?) because i'm quite certain it's possible
21:41:00 <puck> (RTG only of course)
21:41:00 <geist> here's the old classic beos demo video: https://youtu.be/cjriSNgFHsM
21:43:00 <geist> i *do* remember though in late 99 win2k came out and though it was a bit steeper hardware requirement wise, it was starting to look pretty good
21:43:00 <puck> . o O ( 0,2BUY NOW99,99 )
21:43:00 <geist> had the win95 ui but was a solid NT machine
21:44:00 <geist> beos was starting to get eclipsed at that point as far as feeling snappy in a highly multithreaded way on SMP machines
21:45:00 <puck> i kinda want to grab a few retro machines and use them for a while as primary machines. but ii've not been very functional recently to get that up and runniing
21:46:00 <puck> (also i need to fix the debounce on my keyboard because it is Not Very Good Right Now)
21:54:00 <geist> every time i try to seriously use a retro machine to do anything i immediately hit the Web Problem. ie, if you dont have a current web browser you can't do anything
21:54:00 <geist> except irc and ssh into something else
21:56:00 <puck> to be fair, irc + some music playing + ssh would roughly cover my work needs. if i were to seriously try this for a longer time i'd probably just run haiku :p
21:56:00 <puck> (yeah not retro but it works out roughly the same. i just kinda want to get out of the monotony of my current desktop setup)
21:57:00 <geist> yah i mean with modern machines it's basically web brower + a few specialized apps
21:57:00 <geist> so to a certain extent very little of it matters in the normal case for normal users
21:58:00 <geist> i switch between mac, chromeos, windows, and linux like 16 times a day and it's all the same thing, since i just have chrome up on all of them
21:58:00 <geist> looking at the same tabs
21:58:00 <geist> kinda nice and depressing at the same time
21:58:00 <puck> yup.
21:59:00 <geist> only thing that has a kinda unique experience and is in some way 'nostalgic' is my ipad, since it's more app centric, but still close to a regular computer (especially with an attachable keyboard)
22:26:00 <heat> mjg, there's a couple of folks doing maintainership duties for old filesystems in linux
22:26:00 <heat> like the svr4 filesystem, ext2, etc
22:26:00 <heat> befs is probably included in that pack as well
22:27:00 <heat> actually there seems to be a real maintainer but that tree hasn't been touched in 5 years (https://git.kernel.org/pub/scm/linux/kernel/git/luisbg/linux-befs.git/)
22:27:00 <bslsk05> git.kernel.org: kernel/git/luisbg/linux-befs.git - Luis de Bethencourt's fork of linux.git