ChanServ changed the topic of #dri-devel to: <ajax> nothing involved with X should ever be unable to find a bar
<karolherbst> airlied: llvm-15 will be hell :(
columbarius has joined #dri-devel
co1umbarius has quit [Ping timeout: 480 seconds]
<karolherbst> maybe writing our own C parser is really not such a bad idea
pcercuei has quit [Quit: dodo]
<HdkR> karolherbst: That's the madness talking
slattann has joined #dri-devel
cphealy has joined #dri-devel
slattann has quit []
bbrezillon has quit [Read error: Connection reset by peer]
bbrezillon has joined #dri-devel
Company has quit [Quit: Leaving]
mhenning has quit [Quit: mhenning]
cef has quit [Quit: Zoom!]
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
Duke`` has joined #dri-devel
hch12907 has quit [Remote host closed the connection]
Net147 has quit [Quit: Quit]
hch12907 has joined #dri-devel
nvishwa1 has joined #dri-devel
cef has joined #dri-devel
Net147 has joined #dri-devel
cheako has quit [Quit: Connection closed for inactivity]
Net147 has quit [Quit: Quit]
Net147 has joined #dri-devel
kem has quit [Ping timeout: 480 seconds]
lemonzest has joined #dri-devel
kem has joined #dri-devel
hch12907 has quit [Ping timeout: 480 seconds]
hch12907 has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
gouchi has joined #dri-devel
Duke`` has joined #dri-devel
ppascher has quit [Ping timeout: 480 seconds]
vsyrjala has quit [Remote host closed the connection]
tobiasjakobi has joined #dri-devel
tobiasjakobi has quit [Remote host closed the connection]
Duke`` has quit [Ping timeout: 480 seconds]
Duke`` has joined #dri-devel
<karolherbst> HdkR: you won't be able to stop me
<HdkR> You're right but are you prepared for the pain of OpenCL C++?
<karolherbst> we don't have to support OpenCL C++
<HdkR> But people will want it
<karolherbst> and they get it
<karolherbst> but we don't have to support OpenCL C++ :D
<karolherbst> that's all offline compilation
<HdkR> hah
<karolherbst> yeah, we simply need to support accepting spir-vs directly
<HdkR> wait, if you only need spirv then why even do C parsing?
<karolherbst> because that's required by the spec
<karolherbst> spir-v is optional in CL 3.0 even
<HdkR> Interesting...
<HdkR> I guess basic C99 wouldn't be the worst
<karolherbst> it probably is
<karolherbst> well.. that as bad as C++, but...
<karolherbst> it's pointless to replace as long as we have llvm anyway
<karolherbst> or radeonsi still using it
<HdkR> Do the SPIRV tools thing and just use Clang's AST? :P
<karolherbst> uhhh
pcercuei has joined #dri-devel
<HdkR> Which is what the DXC HLSL->SPIRV thing does
<karolherbst> yeah.. dunno. I'd rather not have to maintain a C parser, and I think the changes in llvm-15 are probably the good ones
<karolherbst> it's just seems like translation to/from spir-v is becoming more difficult
<HdkR> What did llvm-15 change that is causing pain? I haven't looked at LLVM for a couple years now
<karolherbst> they removed the coroutine passes (llvmpipe needs it for compute) and are trying to ditch typed pointers
<HdkR> Oh right, I see talk about opaque pointers /all/ the time
<karolherbst> I think it's the right decision, it just makes it hard for everybody else
<karolherbst> although you can disable them at runtime
<HdkR> Is coroutine passes just turning in to regular passes then?
<karolherbst> dunno
<HdkR> Well, sounds like fun!
<HdkR> </s>
dv_ has quit [Ping timeout: 480 seconds]
rkanwal has joined #dri-devel
dv_ has joined #dri-devel
kts has joined #dri-devel
kts has quit []
flacks has quit [Quit: Quitter]
flacks has joined #dri-devel
adjtm has joined #dri-devel
Haaninjo has joined #dri-devel
libv_ has joined #dri-devel
hch12907 has quit [Remote host closed the connection]
hch12907 has joined #dri-devel
libv has quit [Ping timeout: 480 seconds]
nvishwa1 has quit [Ping timeout: 480 seconds]
mszyprow has joined #dri-devel
rkanwal has quit [Ping timeout: 480 seconds]
<macc24> if i have a panel, that has to be initialized via spi, but is a dsi video mode panel, how should i bind it in dts?
<macc24> downstream driver just implements entire spi-gpio by itself and i know it won't happen under mainline kernel, so i want to make my driver for it somewhat upstreamable from the beginning
<macc24> i've tried looking for panels like this but i couldn't find any like this, maybe i was using git grep wrong
<macc24> should i do it similarly to "raspberrypi,7inch-touchscreen-panel" but with spi instead of i2c?
<Mis012[m]> macc24: that i2c sounds like EDID maybe?
<Mis012[m]> so that would be a bit different
<macc24> Mis012[m]: the binding doc for it says that i2c is for "power sequencing the dsi bright and controlling backlight"
camus has quit [Ping timeout: 480 seconds]
<Mis012[m]> well, in that case it might actually not be going to the panel controller...
<Mis012[m]> but to some other IC
<Mis012[m]> because a panel controller would just take commands in-band via DSC
<macc24> dsc?
<Mis012[m]> and it got into mainline like that because how are the reviewers supposed to know...
<Mis012[m]> display stream command?
<Mis012[m]> something like that
<Mis012[m]> my memory is not the best :P
<macc24> googling points me at 'display command set'
<macc24> datasheet for panel says it's either initialized via regular dsi commands or via spi
<macc24> and the device i'm trying to get running initializes it via spi...
<macc24> the dsi vs spi init is controlled via a pin on the panel
<Mis012[m]> dcs, sounds about right :P
<Mis012[m]> oopsie
<Mis012[m]> oh, you have datasheet for panel?
<Mis012[m]> noice
<Mis012[m]> gib gib
<macc24> just google for "ams495qa04.pdf"
<Mis012[m]> sammy one?
<Mis012[m]> sounds like sammy
<macc24> yea
<macc24> don't get excited, the device is /not/ a ps vita :v
<macc24> but has the same panel somehow
<Mis012[m]> ps vita doesn't even have mmu iirc
<macc24> the datasheet for panel is confusing for me to read
<macc24> also i have no clue how to control brightness of the damn thing
* macc24 hopes it will happily accept standard dcs commands for it
<Mis012[m]> samsung OLED?
<Mis012[m]> it really won't
<macc24> oh no
<Mis012[m]> interesting
<Mis012[m]> so it's not using a samsung driver?
* macc24 shrugs
<macc24> i don't know
lumag__ has joined #dri-devel
<Mis012[m]> doesn't look like it was printed/scanned a few times
<macc24> it's also searchable too
<Mis012[m]> some alibaba hits say > Please tell us your requirements, any Drawing or Specifications or Datasheet.
<Mis012[m]> so maybe you can ask nicely for a magnachip d53e6ea8966 datasheet?
<macc24> hmmmmmmmmmmmmmmm
heat has joined #dri-devel
<macc24> oh
<macc24> samsung ld9040 panel looks similar enough to the ams495qa04
<Mis012[m]> :doubt:
<Mis012[m]> does it even use the same driver IC
<Mis012[m]> you need the driver IC datasheet
<macc24> alright, i'll go looking for it
kem has quit [Ping timeout: 480 seconds]
kem has joined #dri-devel
xroumegue has quit [Ping timeout: 480 seconds]
iive has joined #dri-devel
cheako has joined #dri-devel
mszyprow has quit [Ping timeout: 480 seconds]
Duke`` has quit [Ping timeout: 480 seconds]
ella-0_ has joined #dri-devel
Duke`` has joined #dri-devel
ella-0 has quit [Read error: Connection reset by peer]
lumag__ is now known as lumag_
nvishwa1 has joined #dri-devel
Danct12 has quit [Quit: Quitting]
lgarylmtdea^ has quit [Remote host closed the connection]
Danct12 has joined #dri-devel
gouchi has quit [Remote host closed the connection]
gouchi has joined #dri-devel
bcheng has quit [Ping timeout: 480 seconds]
OftenTimeConsuming has quit [Remote host closed the connection]
OftenTimeConsuming has joined #dri-devel
mhenning has joined #dri-devel
stuart has quit [Remote host closed the connection]
elongbug has joined #dri-devel
rkanwal has joined #dri-devel
elongbug_ has joined #dri-devel
elongbug has quit [Read error: Connection reset by peer]
alanc has quit [Remote host closed the connection]
alanc has joined #dri-devel
elongbug__ has joined #dri-devel
<airlied> karolherbst: yeah this week.task might.be llvm15
elongbug_ has quit [Ping timeout: 480 seconds]
<karolherbst> airlied: now I managed to build llvm and clang and everything and luxmark-v3.1 crashes for whatever weird reason...
<karolherbst> ehh wait... clang doesn't work
<karolherbst> ahhhhhh, I hate llvm
<karolherbst> "Couldn't create Clang invocation." ..
<karolherbst> this is all so terrible
<karolherbst> and that llvm copr doesn't work, because it already removed coroutines :(
<karolherbst> of course clang does work...
<karolherbst> (the CLI tool), but not within llvmpipe
<karolherbst> how does anybody compile llvm and doesn't throw their computers out of the window every 5 seconds
<imirkin> gentoo.
<karolherbst> yeah I mean, after you spend years maintaining your distributions llvm packages you probably know what you have to do
<imirkin> i guess the gentoo equivalent of that in fedora is "use srpm"?
<imirkin> but i don't know the complete details
<karolherbst> yeah.. I tried that, but then I need clang to use my local env and it somehow manages to run OOM
<imirkin> delightful.
<imirkin> i manage to compile everything with 6GB of ram
<imirkin> actually disk space tends to be more of an issue :)
<karolherbst> yeah... might be that parallel linking
<karolherbst> although llvm and clang really need a bunch of mem once you enable debug symbols
<imirkin> yeah, i did have to reduce parallelism, esp when building e.g. chromium
<imirkin> but llvm is also pretty stressful on the resources
<karolherbst> _but_ I do have 32 GB of ram :D and I kind of hoped that would suffice
ppascher has joined #dri-devel
apinheiro has joined #dri-devel
gouchi has quit [Remote host closed the connection]
<karolherbst> okay.. got it working I think
<karolherbst> nice.. not hitting that assert :(
<airlied> does it fix the crash?
<karolherbst> hard to tell. I still see the wrong rendering which I saw before. but I think I also saw it crashing outside of gdb
<karolherbst> I suspect that is some random unitiliazed memory going on
<karolherbst> ahh.. now it crashed
<karolherbst> maybe valgrind tells us something
<karolherbst> nothing...
<karolherbst> "Access not within mapped region at address 0x1FDD45C"
<karolherbst> seems like llvmpipe accesses some random memory, but..
vsyrjala has joined #dri-devel
<hch12907> I remember running out of memory (16GB) when I was compiling llvm, had to increase my swap space
<hch12907> and it takes a looooot of disk space too iirc
<hch12907> generally it's been a pain to build projects that depend on llvm from scratch.
mhenning has quit [Quit: mhenning]
mhenning has joined #dri-devel
mszyprow has joined #dri-devel
Duke`` has quit [Ping timeout: 480 seconds]
mszyprow has quit [Ping timeout: 480 seconds]
mszyprow has joined #dri-devel
<airlied> karolherbst: will have to printf debug it :-)
mszyprow has quit [Ping timeout: 480 seconds]
Haaninjo has quit [Quit: Ex-Chat]
<karolherbst> hch12907: if you disable parallel linking that usually helps
<karolherbst> airlied: *sigh*...
<karolherbst> airlied: can't we add a bit of debugging aid to the JIT? I noticed that dumping the asm doesn't work
<karolherbst> well.. it crashes in weird ways
<karolherbst> airlied: but I can't printf debug it, because the runtime crashes :P
<karolherbst> ehh.. well.. not always I guess
<karolherbst> ubsan is quite happy except two things:
<karolherbst> ../src/compiler/nir/nir_serialize.c:1316:40: runtime error: left shift of 1048576 by 45 places cannot be represented in type 'long int'
<karolherbst> ../src/compiler/nir/nir_serialize.c:1327:40: runtime error: left shift of negative value -855707085
<karolherbst> wait a moment...
<karolherbst> ehh no, that's the serialization path, so that's fine
<karolherbst> huh...
<karolherbst> yeah okay.. "((lc->value[0].i64 << 45) >> 45) " I guess we could also mask, but whatever
<karolherbst> the other place is something similiar
<karolherbst> uhh
cef is now known as Guest3313
apinheiro has quit [Quit: Leaving]
cef has joined #dri-devel
Guest3313 has quit [Ping timeout: 480 seconds]
pcercuei has quit [Quit: dodo]
<airlied> karolherbst: printf debug is pretty much adding llvm printfs in the backend, not using CL printf
<airlied> just have to work out what path is crashing, load global is probably a good suspect :-P
<karolherbst> airlied: ohh, right.. could add it to every load_global thing
<karolherbst> airlied: yeah.. ehhh.. now things are too slow
jewins has joined #dri-devel
heat has quit [Remote host closed the connection]
iive has quit []
<lumag_> robclark, done
morphis has quit [Ping timeout: 480 seconds]
morphis has joined #dri-devel
rkanwal has quit [Quit: rkanwal]