ChanServ changed the topic of #dri-devel to: <ajax> nothing involved with X should ever be unable to find a bar
<FLHerne>
Yes it's part of the kernel
<FLHerne>
Your Vega will use amdgpu by default, I don't think `radeon` even supports it
<jekstrand>
FOr ray-tracing, GCN doesn't have any hardware so you'd be hoping for an emulation. IIRC, someone at Valve is working on one but it's going to be a while yet and they may not enable it on older hardware if they run into bugs they don't have time to sort out.
<Wally>
So if I install manjaro or another on my 7950 machine vulkan will work out of the box?
<FLHerne>
jekstrand: JoshuaAshton merged it a while ago, in the MR above, unless I misunderstood
<FLHerne>
Wally: no, 7950 is GCN 1.0 so you'll need to force it to use amdgpu
pcercuei has quit [Quit: dodo]
<Wally>
FLHerne: How?
<FLHerne>
radeon.si_support=0 amdgpu.si_support=1
<FLHerne>
in the kernel parameters
<FLHerne>
(i.e. bootloader config usually)
dliviu has quit [Quit: Going away]
dliviu has joined #dri-devel
cki has quit [Quit: Leaving...]
sdutt has quit [Ping timeout: 480 seconds]
dliviu has quit []
libv has quit [Ping timeout: 480 seconds]
ppascher has quit [Ping timeout: 480 seconds]
dliviu has joined #dri-devel
columbarius has joined #dri-devel
co1umbarius has quit [Ping timeout: 480 seconds]
libv has joined #dri-devel
dliviu has quit []
dliviu has joined #dri-devel
Wally has quit [Remote host closed the connection]
rasterman has quit [Quit: Gettin' stinky!]
Peroverik has joined #dri-devel
Wally has joined #dri-devel
mripard has quit [Read error: Connection reset by peer]
bbrezillon has quit [Read error: Connection reset by peer]
<Peroverik>
If you have even anatomical virtues on planet earth aside from spamming corrupted science to my media and telecom info channels is not entirely known most of the changes to favor my success were in eu pushed by my female fans you called as cockblocking crookedface quasimodos with big cocks injected back to me my own stem cells with programmed
<Peroverik>
structure to shrink ny dick while humiliating me in my physical locations i have also older est men on my side who know neurology working around the world where your verichip attempts are blocked me calling you shithoses their anchorcell killed with latvians 1.5 century ago such people like you livland people and it was around1million of the and
<Peroverik>
only 1 among many like 50 of such so all your delusions to handle me with whatever you come against me with seemingly mistaking my ass with voluentsry shitholes from fin and est the local heroic holes for bigcranks gonna end up like in sport facilities where i abuse you so i can let you live couple of times and redirect you liquid shit human
<Peroverik>
anatomy link to understand why those whores big cranks are owned by me quicly my testosteron production and voice is manly its the stem directed back to me causing pain for my real fans and estonisn neither other beauties liquid shit carriers wont kill anymore stop talking shit to me about your help which never took place for me and supremacy of
<Peroverik>
your genetics only stupid whore can belive that not me, your liquid shit to me.
Peroverik has quit [Quit: Connection closed]
mripard has joined #dri-devel
bbrezillon has joined #dri-devel
camus has joined #dri-devel
MrCooper_ has joined #dri-devel
MrCooper has quit [Ping timeout: 480 seconds]
Peroverik has joined #dri-devel
<Wally>
Peroverik: wtf?
eukara has joined #dri-devel
<ishitatsuyuki>
Wally: irc spambot
<Wally>
ishitatsuyuki: ah
<Wally>
ishitatsuyuki: Why do people still make them?
<ishitatsuyuki>
:shrug:
<Wally>
No one uses irc anymore except for programmers
<Wally>
kinda
<Wally>
that is moving to discord
<ishitatsuyuki>
at least you get less raid than discord
<ishitatsuyuki>
but these kind of bots still exist from before so probably they are just being reused
<Wally>
oh.... are they just old bots from old viruses and such then?
heat has quit [Ping timeout: 480 seconds]
<Peroverik>
09:40:46
<Peroverik>
You can engineer any particle striim and vechicles with electrons and particles in the chamber where geometric signaller is handlling the correct distribution of forces with computer attached so there instead of touchcreen is some subdivided geometrical ionic bonded signaller calculations are easy hence. But the stem is mine the rarest for me ionic
<Peroverik>
charge of the stem is very positive and hence waves to not kill me with heart attack caused as side so stem experiments are all successful with my birth.
<Peroverik>
Its like wireless stirring of molecules.
<Peroverik>
And asming them back together.
<Peroverik>
My stem is rarest cameleon one where it becomes possible with genetically engineered nerve fibers to have an organic shield for danger waves or evil attempts with those.
<Peroverik>
Its not alien job it was maybe but it more likely was unbelivable fluke in my mothers placenta.
<Peroverik>
I robbed hard tissues from my sister there and a- is bloodgroup.
LexSfX has quit [Ping timeout: 480 seconds]
<Peroverik>
Aliens who made our planet which never destroys like some idiots say life can pause easily destroying human colonies. Those are total giants and in vastly bigger amounts we can visit them but we will loose black holes can make us as powerful and aliens are small when arriving tmo this side and with big heads cause with ohtrageous distance between
<Peroverik>
the galaxies they lost their structure brain is mist resillient so head is larger.
LexSfX has joined #dri-devel
<Peroverik>
So i can not be delusionally stupid and americans from same northern hemosphere with china canada and russia and europe either cant be so low to high level smt yices2 netlists to verilog back shows for not big experts who can not read circuits what clocking does and how real systems should work on this hw.
<Peroverik>
I am looking into plumbing mono on linux 5.11 and up to win11 ntoskrn.dll and sys like limbo does for xnu now.
<Peroverik>
Chromeos looks nice.
boistordu_ex has joined #dri-devel
<Peroverik>
I noticed the irc channels too.
<Peroverik>
There is chinease company called elan ebits making keyboards for arm tablets i figured out fheir universal ratio nonstandart 16 10 including some uni bezel immit.
<Peroverik>
So they have keyboards for basically any tablet any kind of layout too.
<Peroverik>
Allready standart lowperf gear mode perf for those is also outrageous.
<Peroverik>
Cisc is better but those are goid chios too.
<Peroverik>
They are very cheap and tw comp makes nfc chips for their microsd slots
<Peroverik>
Taiwan
<Peroverik>
Kingconn is that company
boistordu has quit [Ping timeout: 480 seconds]
<Peroverik>
Nothing related to hypervisors i like thouth limbo and pongos i do respect as approach though.
<Peroverik>
And kernel 5.11 and above userspace pid soace allocs however they called it.
<Peroverik>
Came late but that is ok work.
<Peroverik>
We flesh out mono aot pe container handling in misc filesystem almost done
LexSfX has quit [Ping timeout: 480 seconds]
<Peroverik>
And run legal licenses of windows in the chroot.
<Peroverik>
Sorry docker i do not like hypervisors.
<Peroverik>
Limbo works for osx ios and all just a little work to manipulate shared cache with upstream libs
<Peroverik>
For full compliance
<Peroverik>
And mono just needs a patch for ntoskrn stuff
<Peroverik>
Like ndiswrapper stuff but different alot
<Peroverik>
Some kernel usersoace interrupt api to calk sys files ioctls
<Peroverik>
Mono has al syscalls and github has
<Peroverik>
All the pe chroot code for misc contsiner linux
<Peroverik>
Snowdens smi firmware attacks is crap way anyhow real attacks are easier and different
<Peroverik>
They are bus vulnerabilities.
<Peroverik>
I say goodbye now since week ago sanctions at me were dropped by my homelands court
<Peroverik>
Of law decision
LexSfX has joined #dri-devel
<Peroverik>
Ipads are fine too though even replicas are expensive but they also have all form factor keyboards
Wally has quit [Remote host closed the connection]
<Peroverik>
Ios is even better than android mostly but whatever.
<Peroverik>
Tizenos and some other scandinavian ones did not look much into
<Peroverik>
Any code could be put tobpublic domain legally anyhow
<Peroverik>
I continue of course with linux
<Peroverik>
I do not piss off anyone by such blasts or uploads i just pay for my windows blobs
<Peroverik>
For development for that platform as well as ios and osx and watchos ipados iphoneos etc.
aravind has joined #dri-devel
<Peroverik>
Though nt might be microkernel so maybe those lin userspace to windows userspace to windows kernel userspace are interrupltless
<Peroverik>
More like xnu shims
<Peroverik>
But xnu has bsd ticking for interrupt stuff
<Peroverik>
Will look into windows current kernel docs today allready.
<Peroverik>
They might had gotten rid of posix supervisor ring in userspace
<Peroverik>
So some win7 and up fibers took over.
<Peroverik>
For newcomers those make privileged all rights threads in kernel space
<Peroverik>
But those pe bin misc stuff should also work slightly confusing pid hijacking there instead
Peroverik has quit []
lemonzest has quit [Quit: WeeChat 3.3]
<imirkin>
test
<imirkin>
success.
<jekstrand>
fail
Wally has joined #dri-devel
rcf1 has quit [Ping timeout: 480 seconds]
<Wally>
What would be the best way to get amd gpu drm(I only care about gcn and vega) into a hobby kernel(and if the answer is create it from scratch, where is the documentation for the drm interfaces libdrm and the such use)?
<imirkin>
didn't someone tell you precisely what cmdline you needed?
<Wally>
different issue
<Wally>
non-linu
<imirkin>
oh, sorry
<imirkin>
basically you want to copy the drm subsystem
<imirkin>
and hook it up to the hobby kernel's interfaces
<Wally>
thats it?
<imirkin>
depending on the hobby kernel, this is either a very difficult or an insurmountable task
<Wally>
does it matter what kernel its from(ie. linux, *bsd)?
<imirkin>
other than that, that's it :)
<imirkin>
it's all the same code. it's developed on linux, freebsd & co just copy drm
<imirkin>
and apply the steps i mentioned
<Wally>
--what about the liscencing--
<imirkin>
it's MIT-licensed
<Wally>
Oh....
zackr has quit [Ping timeout: 480 seconds]
<imirkin>
there are a handful of GPL-licensed bits, but they are not important
<imirkin>
(individual drivers, etc)
<Wally>
Are there a lot of linux specific bits(ie. weird bits) or is it normal....
<Wally>
Would it take much time to do when im working on it on my spare time?
<imirkin>
depending on the kernel ... maybe a year
<imirkin>
depends how much you copy from linux, how much you hack out, etc
<Wally>
ill be hacking out any driver that is from intel and most nvidia and legacy driver proably
<imirkin>
yeah, i mean the issues are more generally hooking into the interrupt framework properly, all the locking, etc
ppascher has joined #dri-devel
Wally has quit [Quit: Page closed]
Peroverik has joined #dri-devel
rcf1 has joined #dri-devel
rcf1 has quit [Remote host closed the connection]
JohnnyonFlame has joined #dri-devel
sumits has joined #dri-devel
Peroverik has quit []
Peroverik has joined #dri-devel
frytaped has joined #dri-devel
libv_ has joined #dri-devel
libv has quit [Ping timeout: 480 seconds]
Duke`` has joined #dri-devel
lemonzest has joined #dri-devel
Company has quit [Quit: Leaving]
fxkamd has quit []
itoral has joined #dri-devel
mclasen has quit [Ping timeout: 480 seconds]
mszyprow has joined #dri-devel
Peroverik has quit [Ping timeout: 480 seconds]
Duke`` has quit [Ping timeout: 480 seconds]
Peroverik has joined #dri-devel
danvet has joined #dri-devel
tzimmermann has joined #dri-devel
alanc has quit [Remote host closed the connection]
<javierm>
I could also trade some reviews if you point me to pending patches :)
pcercuei has joined #dri-devel
<danvet>
javierm, airlied or gerd hoffmann often ack patches for 2-4 drivers
ppascher has quit [Quit: Gateway shutdown]
tursulin has joined #dri-devel
<javierm>
danvet: Ok, thanks
Hi-Angel has joined #dri-devel
ppascher has joined #dri-devel
<emersion>
danvet, now that fbdev has a new maintainer, is it still deprecated?
<emersion>
is the deprecation documented somewhere?
<danvet>
emersion, I have a draft mail asking questions about rushing in maintainer changes over the w/e
<danvet>
but airlied seems to be asleep already
<emersion>
cool
<danvet>
emersion, afaik it's not documented anywhere, but moving at least the core and fbcon maintainership away from drm makes no sense
<danvet>
because the last 5+ years anything that happened in there was done by dri-devel folks
<danvet>
for the drivers themselves I really don't care
<danvet>
as long as whomever volunteer doesn't screw up the code we still use
<javierm>
danvet: yeah, I don't see the need for a fbdev maintainer over just someone with commit access to drm-misc that takes care of reviewing / pushing fbdev related patches
<danvet>
javierm, yeah
<danvet>
but really the part that is no-go is moving maintainership for the things we still use/care about away from drm folks, rushed in over a w/e
<javierm>
danvet: agreed
<danvet>
emersion, javierm since you've both brought it up, does this sound reasonable or too snarky? https://paste.debian.net/1227358/
<danvet>
tzimmermann, ^^ maybe relevant for you too
<danvet>
imo distros still enabling fbcon are the most relevant voices here, since you carry the security horror show burden this code is
MrCooper_ is now known as MrCooper
<javierm>
danvet: I think it's OK. I would say that even for fbdev drivers it makes sense to keep it in the drm tree
<javierm>
just last week I had to push some vga16fb fixes for bugs exposed by recent changes to support simpledrm on aarch64
<danvet>
javierm, if people want the "direct pr to linus" bragging rights for something that just doesn't matter, I don't care
<javierm>
danvet: but it creates unnecessary overhead for coordination
<danvet>
but yeah maybe you can reply with that example of drm folks wanting to push fixes to drm-misc directly
<danvet>
javierm, if helge takes care of all the fallout on their end I don't care myself
<danvet>
but a good point, maybe bring it up in your own reply
<javierm>
danvet: sure, I'll reply now
<danvet>
javierm, wait a bit, since I'm adding linus
<javierm>
danvet: Ok
frieder has quit [Remote host closed the connection]
<danvet>
javierm, cc'ed you to
frieder has joined #dri-devel
<tzimmermann>
danvet, i agree with what you said. i don't main having someone maintaining fbdev. but splitting it off from drm makes no sense.
<javierm>
danvet: thanks, will answer now then
rasterman has joined #dri-devel
frieder has quit [Remote host closed the connection]
frieder has joined #dri-devel
pnowack has joined #dri-devel
<javierm>
danvet: something that I forgot to mention in my reply is that drm-misc also provides redundancy in case the fbdev maintainer is not available for some time
<danvet>
javierm, pile it on
<danvet>
tzimmermann, maybe good if you add that on list with your distro hat
<danvet>
if you feel like
<tzimmermann>
danvet, i wanted to wait for helges answer first
<tzimmermann>
but maybe i'll do
aravind has quit [Ping timeout: 480 seconds]
pjakobsson has quit []
<danvet>
tzimmermann, well helge's already busy applying reverts for code removals
<danvet>
because apparently not enough security bugs in fbdev code as-is
<danvet>
mlankhorst, tzimmermann mripard ok if I suggest to helge deller to apply for drm-misc commit rights so they can push the various fbdev fixes there?
<danvet>
not really anything of substance in drivers/gpu, but plenty under arch/parisc
<tzimmermann>
danvet, ok
<danvet>
plus I think offering that option is probably the quickest way out here without a huge mess
kts has joined #dri-devel
<tzimmermann>
danvet, we still do the PRs, so we'd see if anything 'crazy' shows up
rgallaispou has joined #dri-devel
<tzimmermann>
besides that, i welcome that we have someone looking after fbdev
rgallaispou has quit []
<danvet>
tzimmermann, oh yeah it'd be great if someone cares more about this old cruft than just complaining that no one else cares enough
<danvet>
like geert bringing up the drm/fbhelpers doesn't support low bit greyscale
<danvet>
it's really not hard to type that up
<danvet>
and if you also add support to fbdev helpers, should be as fast as native fbdev driver
Company has joined #dri-devel
gawin has joined #dri-devel
iive has joined #dri-devel
<jenatali>
kisak: I suspect you'd want someone from the fdo side of things to take a look (maybe daniels?) Seems like issues that were closed by commits in main are being referenced in the backport for whatever reason... which would be fine except a couple have been reopened since then
<daniels>
yeah, that should only actually close those issues for an MR on main
<daniels>
I suspect it's just a broken summary
kts has quit [Quit: Konversation terminated!]
aravind has quit [Ping timeout: 480 seconds]
<kisak>
the last round of staging/21.3 skipped marge-managed MRs with the CI being bogged down. Maybe the quirk behavior is from that?
mvlad has joined #dri-devel
<kisak>
I worded that wrong ... the backport MRs for 21.3.4 were handled manually because CI was having trouble. That might have lead to this quirk.
Haaninjo has joined #dri-devel
<unrelentingtech>
hey, can someone take a quick look at https://gitlab.freedesktop.org/mesa/mesa/-/issues/5737#note_1190105 ? that's a little preprocessor issue.. we might solve it on the system side actually but it's not great that a system *can* mess it up in that way
cphealy has quit [Quit: Leaving]
xlei has quit [Ping timeout: 480 seconds]
cphealy has joined #dri-devel
<daniels>
zmike: RIP your betting pool
<MrCooper>
kisak daniels: I suspect it's an artifact of that MR being created for main initially, so it considered all commits on the stable branch as part of the MR
<MrCooper>
hopefully the next rebase will clear it up
<daniels>
MrCooper: yeah exactly
anholt has quit [Quit: Leaving]
anholt has joined #dri-devel
* pinchartl
is soooo tempted to introduce __attribute__((cleanup)) in kernel code
rasterman has quit [Quit: Gettin' stinky!]
<danvet>
pinchartl, what does that do?
<daniels>
danvet: executes a destructor on a variable when it goes out of scope
<pinchartl>
it's only useful for local variables on the stack, it doesn't provide unique or shared ptr semantics for heap variables, but it's still useful
<pinchartl>
I'm trying to implement a for_each iterator macro that requires initialization and cleanup. init is easy, but there's no good way I can find to implement the cleanup automatically
<pinchartl>
and requiring drivers to call a cleanup function on an iterator variable after the loop is error-prone
<pinchartl>
with c99 or c11 support which could now be enabled in the kernel thanks to the gcc minimum version being bumped to v5.1, we could declare variables inside the for statement, which would also make iterators much nicer
<pinchartl>
list_for_each_entry_safe wouldn't require an explicit tmp variable \o/
<hakzsam>
jekstrand: congratulations for your new position! I think you mean 24th though (in the MR) :-)
<jekstrand>
Yup. Finger slip.
<daniels>
jekstrand: looking forward to more of these kinds of high-skill contributions :P
<jenatali>
jekstrand: Congrats!
Duke`` has joined #dri-devel
<jenatali>
Phoronix in 3... 2...
frieder has quit [Remote host closed the connection]
<emersion>
congrats!
<emersion>
zmike blog post in 3… 2…
<MrCooper>
congrats jekstrand, zmike didn't see that coming :)
<dschuermann>
jekstrand: congrats! :D I feel relieved, probably most of us do :P