<binary>
not sure if this is the right place here so sorry if i am false
bbrezillon has joined #dri-devel
<binary>
since i few days i observe a high memory-usage sometimes after playing a specific game, even after exiting it. the related user-space process seem to be completely gone (can't find anything with top), so i am wondering if that could be a memory leak in the kernel-part?
frankbinns2 has joined #dri-devel
aravind has joined #dri-devel
<binary>
is there a way to debug that?
bgs has quit [Remote host closed the connection]
siddh has quit [Quit: Reconnecting]
siddh has joined #dri-devel
frankbinns1 has quit [Ping timeout: 480 seconds]
<siddh>
Hello, Please take a look at these patches I sent a while ago, and let me know if any further changes are needed. All patches are reviewed once, except the last one which can be ignored. Link: https://lore.kernel.org/lkml/cover.1686075579.git.code@siddh.me/ Thanks.
<siddh>
I was thinking of following it up with separating the old macros into a new file (as suggested in earlier discussion). But I can't do that unless the fix is merged.
vliaskov has joined #dri-devel
MajorBiscuit has joined #dri-devel
lemonzest has quit [Quit: WeeChat 3.6]
senatorM has quit [Ping timeout: 480 seconds]
lemonzest has joined #dri-devel
frieder has joined #dri-devel
xroumegue has joined #dri-devel
senatorM has joined #dri-devel
swalker__ has joined #dri-devel
kj has quit [Ping timeout: 480 seconds]
rasterman has joined #dri-devel
tursulin has joined #dri-devel
tristan has joined #dri-devel
tristan is now known as Guest7468
donaldrobson has joined #dri-devel
The_ASV has joined #dri-devel
aravind has quit [Read error: Connection reset by peer]
aravind has joined #dri-devel
frankbinns2 has quit [Quit: Leaving]
frankbinns has joined #dri-devel
The_ASV has quit [Read error: Connection reset by peer]
The_ASV has joined #dri-devel
cmichael has joined #dri-devel
<senatorM>
I am not used to dealing with so big hunks to be expected to be read correctly, but likely bigwuzla does the more advanced way of same thing. Sorts of nodes with values and indexes, coeffs, and bounds.
kts has joined #dri-devel
chengcheng has joined #dri-devel
senatorM has quit [Ping timeout: 480 seconds]
donaldrobson has quit [Remote host closed the connection]
kj has joined #dri-devel
simon-perretta-img_ has joined #dri-devel
simon-perretta-img__ has joined #dri-devel
simon-perretta-img has quit [Ping timeout: 480 seconds]
simon-perretta-img_ has quit [Ping timeout: 480 seconds]
swalker__ has quit [Remote host closed the connection]
The_ASV has quit [Read error: Connection reset by peer]
The_ASV has joined #dri-devel
The_ASV has quit [Read error: Connection reset by peer]
co1umbarius has joined #dri-devel
The_ASV has joined #dri-devel
chengcheng has left #dri-devel [Leaving]
tristan__ has joined #dri-devel
Guest7468 has quit [Ping timeout: 480 seconds]
lucaceresoli has joined #dri-devel
MajorBiscuit has quit [Ping timeout: 480 seconds]
tristan__ has quit [Ping timeout: 480 seconds]
MajorBiscuit has joined #dri-devel
senatorM has joined #dri-devel
binary has quit []
alyssa has joined #dri-devel
orbea has quit [Remote host closed the connection]
orbea has joined #dri-devel
yyds has quit [Remote host closed the connection]
itoral has quit [Remote host closed the connection]
Company has joined #dri-devel
MajorBiscuit has quit [Ping timeout: 480 seconds]
Ahuj has quit [Ping timeout: 480 seconds]
cmichael has quit [Quit: Leaving]
paulk has joined #dri-devel
cmichael has joined #dri-devel
Leopold__ has quit [Remote host closed the connection]
<KungFuJesus>
MrCooper: Yes, I've read that much, like I said I realize I'm probably a decade too late. The mesa-amber project though is meant to make things continue to work with fossilized dinosaur GPUs, or the name would have me believe. I happen to have a dinousaur GPU inside a potato G3
kts has quit [Quit: Leaving]
<emersion>
is there a reason why the drivers couldn't be ported to DRI2?
bmodem has quit [Ping timeout: 480 seconds]
<kisak>
KungFuJesus: no, mesa-amber isn't a separate project with the goal of keeping fossils working. It's a legacy release branch to keep "classic" (non-gallium) drivers available as a second slotted mesa install. There's no expressed intent to add any hardware to that branch.
<KungFuJesus>
emersion: that's what I was sort of getting at, the ML post there had suggested there was a path forward for some of the DRI1 only drivers
<KungFuJesus>
The Linux kernel has already killed support after 6.3, so that is yet another thing working against me, there
<emersion>
probably because no-one was maintaining it?
<KungFuJesus>
yeah I understand the reasons I'm just lamenting the uphill battle of trying to re-enable 3d acceleration on this in 2023
<KungFuJesus>
during the transition to DRI2, was there some sort of porting guide / document out there? How similar are these interfaces?
<alyssa>
karolherbst: i would but that MR is terrifying
<alyssa>
i would say it's old enough that faith is the most qualified reviewer, she has probably forgotten enough of the patches that she can review it herself fairly xP
bgs has joined #dri-devel
<KungFuJesus>
it looks like having some out of tree module should be possible, the drivers are built as dynamic libraries
<emersion>
i don't think that's a good path forward
<KungFuJesus>
why's that?
<alyssa>
KungFuJesus: to be clear also, the amber/upstream split is more about what's maintained than the age of the part itself
<KungFuJesus>
right, this was another obsolence decision that took place at much earlier time than the gallium decision
<KungFuJesus>
emersion: were you referring to maintaining an out of tree module as not a good idea or porting the DRI1 driver to DRI2 as not a good idea?
<emersion>
out of tree
<KungFuJesus>
ah, just hard to build that way? Relies on something more than just headers?
<KungFuJesus>
or rather, it's not a very stable API
anholt_ has quit [Ping timeout: 480 seconds]
JohnnyonFlame has quit [Ping timeout: 480 seconds]
Haaninjo has joined #dri-devel
<KungFuJesus>
I suspect I'm going to find a reason r128 was not ported to DRI2 while doing it. At least one post I've found suggests it managing buffers on lower vram cards might have been part of it
<KungFuJesus>
hmm, here's the guy that added EXA support to the DDX: https://smallperturbation.com/r128-exa - looks like the goal might have been to get DRI2 support. Did his laptop die? lol
bgs has quit [Remote host closed the connection]
<DavidHeidelberg[m]>
zmike: looked at the shader-db failure, I think where is the issue, meanwhile we could stop spamming with SKIP to the stderr, since it's just at best stdout (and I would love to have option to supress these SKIP warnings): https://gitlab.freedesktop.org/mesa/shader-db/-/merge_requests/91
kts has joined #dri-devel
kzd has joined #dri-devel
Leopold has quit [Remote host closed the connection]
Leopold_ has joined #dri-devel
frieder has quit [Remote host closed the connection]
andrey-k- has joined #dri-devel
kts has quit [Remote host closed the connection]
kts has joined #dri-devel
<gfxstrand>
cwabbott, hakzsam: Any more comments on !24335?
<gfxstrand>
I'd like to get rid of my NVK blockers so that we can merge as soon as the new UAPI is ready.
andrey-k- has quit []
andrey-k- has joined #dri-devel
yyds has joined #dri-devel
<DavidHeidelberg[m]>
zmike: btw. I looked at the shader-db crash code, I guess the issue is that CI catches the crash first and start finishing.. I'll try to think how to disable it for this case
Duke`` has joined #dri-devel
anholt has joined #dri-devel
heat has joined #dri-devel
<alyssa>
gfxstrand: "I didn't bother to copy the entire
<alyssa>
VkSamplerCreateInfo into the sampler
<alyssa>
"
<alyssa>
what isn't copied? i'm confused
<alyssa>
just like wrap modes and things?
<alyssa>
and the expectation is that the driver packs the whole sampler descriptor in the driver CreateSampler and then can throw away the createinfo?
aravind has quit [Ping timeout: 480 seconds]
<gfxstrand>
alyssa: With things like `vk_image_view`, it basically contains everything in `VkImageViewCreateInfo`. With vk_sampler, I just pulled in the fields that actully need some amount of wrapping or interpretation.
<alyssa>
makes sense :+1:
andrey-k- has quit []
heat has quit [Remote host closed the connection]
heat has joined #dri-devel
benjaminl has joined #dri-devel
Leopold__ has joined #dri-devel
Leopold_ has quit [Ping timeout: 480 seconds]
andrey-k- has joined #dri-devel
cazzacarna has quit [Remote host closed the connection]
cazzacarna has joined #dri-devel
andrey-k- has quit []
junaid has joined #dri-devel
cmichael has quit [Quit: Leaving]
jkrzyszt_ has quit [Ping timeout: 480 seconds]
djbw has joined #dri-devel
haasn` has joined #dri-devel
The_ASV has quit [Read error: Connection reset by peer]
The_ASV has joined #dri-devel
Ahuj has joined #dri-devel
yyds has quit [Remote host closed the connection]
The_ASV has quit [Read error: Connection reset by peer]
<hch12907>
<gfxstrand> "rustfmt is also a lot more..." <- I remember most of its config options are locked behind `unstable`
<eric_engestrom>
I wish I had asked myself "did they break the output of clang-format in the past?" because clearly yeah they have, which means the only way it can ever be used is to freeze clang-format to one version and have every mesa dev ever install _that_ version and good luck getting everyone to agree to that
<eric_engestrom>
or nobody does any formatting ever and instead we have the CI do it (the thing I had started working on)
<eric_engestrom>
like, the CI _does_ the formatting, not _enforces_ it
tursulin has quit [Ping timeout: 480 seconds]
<eric_engestrom>
anyway, it's late, I'm out o/
<alyssa>
gfxstrand: it was fun while it lasted
<alyssa>
eric_engestrom: fwiw when I originally did clang-format CI i pinned a particular version
<alyssa>
idk if that survived to the merge
<alyssa>
it appears "yes but it's pinned to LLVM_VERSION"
The_ASV has joined #dri-devel
<eric_engestrom>
alyssa: LLVM_VERSION is what controls it
<eric_engestrom>
yup
<alyssa>
eric_engestrom: ..right, so that bumped clang-format inappropriately then
<alyssa>
I very specifically picked clang-format-14 because it's the newest available in Debian (even Debian unstable)
<alyssa>
LLVM_VERSION=15 though
<eric_engestrom>
IIRC right now we require 13+ for the inheritance feature
<eric_engestrom>
so we can pin to anything 13+
<eric_engestrom>
ok, I really need to go though xD
The_ASV has quit [Read error: Connection reset by peer]
<alyssa>
bibi
Kayden has quit [Quit: -> JF]
<alyssa>
this just straight up looks like a bug in clang-format-14 maybe
<karolherbst>
I think the idea about auto formatter is to not care if they are right or wrong, but to just let them do their thing (and maybe auto format any time that binary changes)
<alyssa>
clang-format-16 is happy with mesa main
<alyssa>
ughhhh
<alyssa>
gfxstrand: ok fine you win
<gfxstrand>
:-P
* gfxstrand
wasn't trying to win anything but also....
<alyssa>
you win... a blahaj sticker
<karolherbst>
but yeah... clang-format behaving differently based on the version sounds annoying :D
<alyssa>
blahaj_smiling.jpg
<karolherbst>
did they add some new options or something?
<karolherbst>
maybe an updated config can fix it
<gfxstrand>
Ooh! BLÅHAJ sticker! I like BLÅHAJ stickers! :D
<alyssa>
:~D
vliaskov has quit [Ping timeout: 480 seconds]
macromorgan has quit [Quit: Leaving]
The_ASV has joined #dri-devel
<DemiMarie>
<gfxstrand> "rustfmt is also a lot more..." <- Also rustfmt uses the same parser as rustc, whereas clang-format cannot due to ambiguities in the C and C++ grammars.
Ahuj has quit [Ping timeout: 480 seconds]
benjamin1 has joined #dri-devel
<senatorM>
big hunk is working on eventloop apis cli programs are just c-api things, and those would work if specific libc functionality is rewritten. but there in osx, windows, linux are so many loose apis that can not be captured by binary analysis and isolated, i suppose. starting from libforms.
Kayden has joined #dri-devel
mvlad has quit [Remote host closed the connection]
benjaminl has quit [Ping timeout: 480 seconds]
MajorBiscuit has joined #dri-devel
The_ASV has quit [Read error: Connection reset by peer]
<senatorM>
i think on linux they allow to tap into kernel with berkeley packet format and type format but after the interrupt was handled not much is there info what application did with that keypress. begs for new specification really, as i do not think there is a way but i work on it a bit first.
The_ASV has joined #dri-devel
The_ASV has quit [Read error: Connection reset by peer]
The_ASV has joined #dri-devel
Haaninjo has quit [Quit: Ex-Chat]
junaid has quit [Ping timeout: 480 seconds]
<senatorM>
i think most understood already the issue, that keyboard is toggled with edge or level triggered interrupts or gpio writes, but it needs a scheduler in the kernel where application switches those off and on what to expect from those gpios and which keys are allowed, and scheduler is for the thing, that when the app hangs, it would not make the keys unresponsive., something like that
senatorM was banned on #dri-devel by ChanServ [senatorM!*@*]
senatorM was kicked from #dri-devel by ChanServ [You are not permitted on this channel]