00:25imirkin: ajax: it would seem that way. i haven't personally confirmed anything.
15:21karolherbst: anybody with a dlink router? :D my advise: burn that shit :D
15:24Tomin: karolherbst: with original firmware? they are crap as usual (I've used DIR-615 and DIR-655)
15:24karolherbst: well yeah, of course with the Dlink firmware
15:24karolherbst: but the software is so broken…
15:25karolherbst: Tomin: https://pierrekim.github.io/blog/2016-09-28-dlink-dwr-932b-lte-routers-vulnerabilities.html
15:26Tom^: i think ive broken atleast 4 DIR-615
15:26karolherbst: Tom^: ohh, I am sure you made them more secure in the end
15:27Tom^: they all sort of burn up after a while if you have to much computers plugged in. :p
15:27Tom^: then i bought asus RT-N66U , havent had a issue since.
15:28Tomin: I usually just put OpenWRT (or DD-WRT) if I can and nowadays just try to avoid anything that is not supported by OpenWRT
15:29Tom^: karolherbst: also dear lord those are some serious vulnerabilities haha
15:31karolherbst: Tom^: remote "update" for the firmware
15:31karolherbst: that's the best
15:31karolherbst: so before you even put your own firmware on, the old one may be already backdoored and is backdooring the new one as well!
15:32karolherbst: that's even dumber than what lenovo did… I thought lenovo deserved place 1 for the most stupid backdoor ever, but here we go, Dlink made it
15:40Tomin: even if there weren't any security problems, the D-Link user interface is already a good reason to avoid them. It's horrible, after every change the router must be rebooted...
15:41karolherbst: yeah well, how can you trust such a router at all?
15:41karolherbst: even the firmware you uploaded could be modified
15:43Tom^: karolherbst: well now, how can we even trust gcc?
15:44karolherbst: that's a good question indeed
15:45karolherbst: because the backdoor might be hidden inside the binary and just adds the backdoor to every gcc on compile
15:46Tom^: which is why we need to compile clang with msvc and then compile gcc with clang
15:46Tom^: oh wait that will make clang to phone home. to microsoft
15:46Tom^: and yep, that is a fact. xD
15:47karolherbst: well yeah
15:52Tom^: better to just do the 5 stages of grief fast, accept it and continue on like nothing is happening. ignorance is bliss
16:34siro__: imirkin_: is DRI3 stable on nouveau and modesetting ?
16:34imirkin_: not sure what you mean by 'stable' - it won't crash or anything
16:35imirkin_: (or at least i'm not aware of any such things)
16:35imirkin_: however afaik some kde stuff misrenders greatly with it
16:36siro__: I've got lot's of system hangs with DRI3 enabled while running nine
16:37siro__: every time the process crashes (due to an assert) the kernel module dies,too
16:37imirkin_: are you doing threaded command submission?
16:38siro__: no i don't think so
16:38imirkin_: nouveau + multithreading = no
16:38imirkin_: iirc st/nine does something funky with threads
16:38imirkin_: and perhaps that blows up in our faces with DRI3
16:38siro__: it only has one pipe_contex
16:39imirkin_: could multiple threads be calling into nouveau code at the same time?
16:39imirkin_: *any* nouveau code, not just context functions
16:40siro__: no, nine locks a mutex for each d3d9 interface call
16:40imirkin_: and what is the "thread_submit" thing i've heard about?
16:41siro__: ah, that's for presenting
16:41imirkin_: which doesn't call into nouveau?
16:42siro__: I don't know
16:42imirkin_: can you disable it and see if that helps?
16:42siro__: it uses xcb and friends to present it
16:42imirkin_: (or is it off by default?)
16:43siro__: atm it's not possible to disable it
16:43imirkin_: ah ok. iirc it was an option
16:43imirkin_: either that changed, or i'm misremembering
16:43siro__: I thought about adding a dummy interface to run wine tests headless
16:43imirkin_: what assert do you get btw?
16:44siro__: I don't know, my system freezes
16:44siro__: can't do anything
16:45imirkin_: how do you know you get one?
16:47siro__: good point. I guess there's one
16:47siro__: I used llvmpipe first and found a few
16:47siro__: at the same spot nouveau freezed my system, so I guess there's another one
16:48imirkin_: usually an assert will end the program gracefully
16:48siro__: that doesn't show up in llvmpipe
16:48siro__: yes, it does on llvmpipe
16:48imirkin_: a hang tends to be due to resource mismanagement