what did 4chan mean by running an ancient version of freebsd and getting hacked?

demido: They had servers running very old and un-patched versions of FBSD?

Not to mention the old software they were also running (which is likely the culprit as opposed to the OS itself.)

supposedly

personally i don't buy it. imo seems like an intentional leak. those vulns were present for years and woulda been used already

so i think an ancient stack was rolled out intentionally to open a door

I've heard from multiple sources recently that it was likely old PHP and/or phpMyAdmin that lead to the compromise.

ya but still, that woulda been found and used already

it's 4chan. it's prolly attacked more than gov mil and banks combined

Perhaps. If someone knew where to find it. Maybe they just found it? Who knows...

i have no inside info but i don't buy it

I'd imagine if anyone really cared, they'd look into it. I couldn't give two s**ts about it, personally.

ya i don't use the site but it's noteworthy

ek: You a dev ?

oxbar: Yes?

I'm not a FBSD dev, though... If that's what you mean. Just a regular, old Joe.

ek: j/w

jehovah's witnesses?

just wondering ?

Just wondering

will pkgbase ship in 15.0?

as far as I know that is currently the plan, but it was already planed for 14.0 and then delayed, well I didn't hear about problems with those pkgbase testrepos for -CURRENT and -STABLE

why is lib32 in base? what breaks if a world is built without it?

that you can run 32bit executables (why shouldn't it be shipped?) and why should anything break if you don't build libaries to run 32bit executables? they are completely optional (if you install a -RELEASE via the installer you get asked if you want to install the lib32 component)

pkgbase is nice

nimaje: ok, thanks

i don't know if i have any 32bit programs installed. how can i check?

i know if i without_kerebos for example, i'll need to build all my own ports with that in mind

f451: what FreeBSD do you have installed? 14.2-RELEASE ?

chances are if you don't specifically know of 32-bit installed programs and it is a recent install -- you have None

64-bit stuff has been the default for like over a decade now

SponiX: 14-stable

n270870

if it can be safely built without it then i'll do that. buildworld takes a while building lib32 and im looking to remove whats not needed

kerebos is another but ill need to remember to build all my own ports - same with hemidal

hey

i'm trying to port some linux code to freebsd

in linux i can dlsym(0,name) and get access to all the symbols in the c standard library

f451: I've never tried excluding anything from world, I just do the normal complete checkout and build

this doesn't seem to work on freebsd and i'm not sure what i'm supposed to pass to dlsym to get it to work

same for dlfunc

is this possible at all?

why do you need all the symbols of the c library?

i need one symbol

ah it looks like RTLD_DEFAULT does what i want

why is this not the same as on linux? am i misreading the docs?

because linus went off and did things his own way

looks like RTLD_DEFAULT is 0 on linux and ((void *)-2) on freebsd

that's perhaps a good reason to use the #define

i mean sure but i didn't realise passing NULL would do things differently from reading the freebsd manual

If dlsym() is called with the special handle NULL, it is interpreted as

a reference to the executable or shared object from which the call is

being made. Thus a shared object can reference its own symbols.

thought this could just access all the symbols that were currently loaded

anyway, thanks for the support

hi there, me again

scanf in freebsd doesn't support %m which is in posix 2008

is 17 years not enough?

I can't see it

izabera: there is no %m specifier in C or POSIX

>An optional assignment-allocation character 'm'.

literally in yourlink

and in the version from 9 years prior

you guys are bad at reading

>The %c, %s, and %[ conversion specifiers shall accept an optional assignment-allocation character 'm', which shall cause a memory buffer to be allocated to hold the string converted including a terminating null character. In such a case yadda yadda yadda

izabera: you said %m

that isn't a %m thought

izabera: there is no %m specifier

?

this really isn't a friendly way to speak

it's m that goes after %

izabera: there is an m flag to the %s specifier

for instance %ms

or %mc or %m[a-z]

either way, can i have that on freebsd

izabera: sure, you'll want to edit lib/libc/stdio/vfscanf.c, open a phab review when you're done and add me (des@) as reviewer

izabera: don't forget to update the manual page

can't believe I was tricked into reading posix

lol

is there a way to detect, in a src Makefile, which toolchain we're using?

[tj]: lol

desdotdev: why do you want to detect it in a Makefile? but I guess COMPILER_TYPE from share/mk/bsd.compiler.mk is the way then

Someone I know updated with pkg update and KDE broke

It seems like the missing core package is pkg install plasma6-plasma

ah, that happened to me too

is there a work around or just wait?

If you know

build from ports

maybe

they have been working on it in #freebsd-desktop

ah thanks for that irc channel

hm, from a system before the unification of kde5/6 into just kde without a version number in the package name? then install kde and remove kde5 or kde6 whichever was installed

It seems kde5 isn't fully there anymore

In packages

yes, that was renamed to just kde

tutorial for 14.1 were wrong

pkg install plasma6-plasma-desktop

@nimaje pkg install -g "plasma6-*"

I also used that and then when the login page came on the bottom left selected plasma

and its working again

I have multiple co-workers using FreeBSD and KDE as their desktop

why not pkg install kde as the handbook says? docs.freebsd.org/en/books/handbook/desktop/#kde-meta-install

I just tested that and it looks like the handbook needs to be updated

pkg install kde 'not found'

latest pkg release is missing packages, so that's why things are broken

If you had kde installed before last update, it warned you that it was going to uninstall kde.

pkg install plasma6-plasma 'not found'

mcrane: as I said, there are missing packages that were not built

This one exists pkg install plasma6-plasma-desktop

On the freebsd forum I found this pkg install --glob "plasma6-*"

and it does work

or shorted slightly to pkg install --g "plasma6-*"

or shorted slightly to pkg install -g "plasma6-*"

i thought the package builder issue was meant to be fixed now, although i can't remember where i read that

seems like additionaly to that bug in -CURRENT which made 14.2 jails crash when building packages, there is/was also a go bug where it can't handle the operating system using 5 level page tables, no idea if there where additional issues which coincidentally all happend on start of this month (afaik for the -CURRENT bug the builders were rolled back to some known good version and that shouldn't be

a problem anymore, not sure if the bug was found)

i only knew about the go issue which i believe they fixed by disabling la57 on the builders

maybe that go bug was the suspected -CURRENT bug, but as I understood it, something made the complete build crash, not just "some" packages not building

i recall there were some doubt about this because some of the missing packages didn't appear to depend on go, but a *lot* of stuff uses go and i never saw a definitive answer

webkit-gtk also had issues

I believe it was go and webkit-gtk if I remember correctly

it's fixed, but as of last report a few repos worth of packages were trying to sync at the same time

when did it get fixed?

I just checked yesterday

checking now :)

your definition of fixed and my definition of fixed are probably different

the underlying issue is fixed and the builds are finished, but they may or may not be available yet due to aforementioned package syncing

kevans: that was a couple of days ago though right? i wonder how long the sync takes

fixed == pkg has all missing packages returned.

about 22 hours ago if thunderbird isn't lying to me

oh, timekeeping is hard

"between today and tomorrow, they will be available on all mirrors." -- so presumably within the next 10-12 hours everything should be better

thanks, I'll check then and let people know once it's back working.

:)

maybe this will teach people not to play silly games with pointers

"oh, look at all the spare bits in this 64-bit pointer! i'm sure computers will never come with more memory and start using them!"

when does CHERI come to consumer hardware? then the hardware would disallow such silly stuff

probably never

AMA going on in the Discord FreeBSD right now

why?

who is "me" in that AMA?

la_mettrie: Deb Goodkind from FreeBSD Inc

Thought it was streaming video?

nimaje: codasip.com/solutions/riscv-process…i/x730-risc-v-application-processor

Last week I asked about a problem I had with 'ip <something> moved from hardware address <x> to <y>' all the time. Suggestion from someone here was to move the ip configuration from vlan interface to the bridge this vlan interface is member of. I did this, and since then those messages are gone

thanks again !

igb0 up, no ip config. vlan6 up on physical interface igb0, no ip nor ip6 configured. bridge0 has member vlan6 and the ip4 and ip6 configs

problem solved

dstolfa: I can't find prices there, so it doesn't really sound "consumer", but reads like a good step in the direction of CHERI consumer hardware

nimaje: that's because it's not out yet :P

mvanbaak: i intend to advocate that we prevent people from configuring IP address on bridge members by default, so i appreciate your example as a justification for this :-)

it showed the ip address of the vlan interface changing from mac address all the time. from the physical interface mac address to a mac address of one of my vnet jails

ivy: yeah, that would be awesome. would have saved me some hours haha

dch: ivy: helloooo. the thing with capsicumizer is that LD_PRELOAD based filesystem fakery was way too janky. the future of that was being discussed in reviews.freebsd.org/D44373

thanks valpackett for the clarification!

valpackett: thanks! iiuc this is basically implementing a chroot for capsicum, which i really don't like, i'd rather have a config file to filter paths like how The Super Capsicumizer 9000 works. maybe i should read the whole thread though...

valpackett: incidentally thanks for your work on getting 9pfs into src, this has been super useful for me

How was the AMA

hi valpackett, your post on getting freebsd working on the pixelbook is what got me into freebsd :)

oxbar: fairly interesting but it was pretty high level and mostly about stuff the foundation is doing most of which you probably already know about if you care enough to hang about here. Setesh or wolfquinn is going to publish a recording later

hm, why does that review want to remove fchdir rights? would that disallow fchdir in capsicum mode or would that allow fchdir on any directory descriptor? both seem wrong to me

desdotdev: are you around? re: traceroute