ChanServ changed the topic of #dri-devel to: <ajax> nothing involved with X should ever be unable to find a bar
JohnnyonFlame has joined #dri-devel
paulk-bis has joined #dri-devel
apinheiro has quit [Quit: Leaving]
paulk has quit [Ping timeout: 480 seconds]
ickle has quit [Quit: Off we go.]
ickle has joined #dri-devel
mhenning has joined #dri-devel
epoll has quit [Ping timeout: 480 seconds]
epoll has joined #dri-devel
ickle_ has joined #dri-devel
ickle has quit [Read error: Connection reset by peer]
yuq825 has joined #dri-devel
YuGiOhJCJ has joined #dri-devel
jewins has quit [Remote host closed the connection]
oneforall2 has quit [Remote host closed the connection]
oneforall2 has joined #dri-devel
paulk-ter has joined #dri-devel
paulk-bis has quit [Ping timeout: 480 seconds]
paulk has joined #dri-devel
paulk-ter has quit [Ping timeout: 480 seconds]
tutu has joined #dri-devel
columbarius has joined #dri-devel
co1umbarius has quit [Ping timeout: 480 seconds]
paulk-bis has joined #dri-devel
paulk has quit [Ping timeout: 480 seconds]
<tutu> Dylanger, hi dylanger, I found a function named XMesaDestroyContext in mesa, it has some comments "FIXME: We should destroy the screen here, but if we do so surfaces may outlive it, causing segfaults", so how could I trigger this segfaults ?
<tutu> and is there any plan to fix it ?
oneforall2 has quit [Remote host closed the connection]
oneforall2 has joined #dri-devel
aravind has joined #dri-devel
Daanct12 has joined #dri-devel
aravind has quit [Ping timeout: 480 seconds]
heat has quit [Remote host closed the connection]
heat has joined #dri-devel
heat has quit [Ping timeout: 480 seconds]
rsjw has quit [Quit: rsjw]
bmodem has joined #dri-devel
slattann has joined #dri-devel
xroumegue has quit [Ping timeout: 480 seconds]
aravind has joined #dri-devel
xroumegue has joined #dri-devel
Company has quit [Quit: Leaving]
mupuf has quit [Remote host closed the connection]
mupuf has joined #dri-devel
Compy_ has quit [Quit: Going offline, see ya! (www.adiirc.com)]
lemonzest has joined #dri-devel
oneforall2 has quit [Quit: Leaving]
oneforall2 has joined #dri-devel
Duke`` has joined #dri-devel
khfeng_ has quit [Remote host closed the connection]
khfeng_ has joined #dri-devel
bmodem1 has joined #dri-devel
<dolphin> airlied: hmm, what's the error?
<airlied> dolphin: replied to the pull request
<dolphin> we're definitely not doing build checking with clang as part of CI so not a full surprise
bmodem has quit [Ping timeout: 480 seconds]
bgs has joined #dri-devel
windleaves has joined #dri-devel
windleaves has left #dri-devel [#dri-devel]
tutu has quit [Ping timeout: 480 seconds]
mhenning has quit [Quit: mhenning]
itoral has joined #dri-devel
<dolphin> airlied: looking at the build failure, seems like a linux/bitfield problem
<dolphin> one of the "always false" problems, when an expression is used to to check that types match
dcz_ has joined #dri-devel
<airlied> dolphin: jani commented on there as well, so hopefully someone has a fix out there
Duke`` has quit [Ping timeout: 480 seconds]
alanc has quit [Remote host closed the connection]
alanc has joined #dri-devel
<dolphin> airlied: yeah, I'll poke the folks to push such a fix
danvet has joined #dri-devel
ngcortes has quit [Quit: Leaving]
Leopold has joined #dri-devel
Leopold has quit []
DemiMarieObenour[m] is now known as DemiMarie
ahajda has joined #dri-devel
jlawryno has joined #dri-devel
bgs has quit [Remote host closed the connection]
Kayden has quit [Ping timeout: 480 seconds]
jrayhawk has quit [Ping timeout: 480 seconds]
jljusten has quit [Ping timeout: 480 seconds]
jljusten has joined #dri-devel
jrayhawk has joined #dri-devel
Kayden has joined #dri-devel
<eric_engestrom> vsyrjala: remnant of https://gitlab.freedesktop.org/mesa/drm/-/merge_requests/227 apparently; reaping it
tzimmermann has joined #dri-devel
frieder has joined #dri-devel
pochu has joined #dri-devel
tursulin has joined #dri-devel
jkrzyszt has joined #dri-devel
jkrzyszt has quit [Ping timeout: 480 seconds]
FireBurn has quit [Read error: Connection reset by peer]
mszyprow has joined #dri-devel
jkrzyszt has joined #dri-devel
FireBurn has joined #dri-devel
rasterman has joined #dri-devel
swalker__ has joined #dri-devel
swalker_ has joined #dri-devel
swalker_ is now known as Guest228
lynxeye has joined #dri-devel
swalker__ has quit [Ping timeout: 480 seconds]
egalli has quit [Quit: Reconnecting]
egalli has joined #dri-devel
egalli has quit []
egalli has joined #dri-devel
vliaskov has joined #dri-devel
mvlad has joined #dri-devel
apinheiro has joined #dri-devel
camus1 has joined #dri-devel
camus has quit [Read error: Connection reset by peer]
pcercuei has joined #dri-devel
MrCooper has quit [Quit: Leaving]
MrCooper has joined #dri-devel
Company has joined #dri-devel
bbrezillon has joined #dri-devel
bbrezillon has quit [Remote host closed the connection]
bbrezillon has joined #dri-devel
vliaskov has quit []
devilhorns has joined #dri-devel
<eric_engestrom> why did we keep the meson option `dri-drivers` around? it can't be set to anything other than "nothing", I'm not sure I see the point
<eric_engestrom> motion to drop it
vliaskov has joined #dri-devel
<eric_engestrom> ah nevermind, I just saw that it is there so that trying to set it triggers this error:
<eric_engestrom> error('Mesa\'s main branch no longer has any "classic" drivers, use the "amber" branch instead.')
<eric_engestrom> retracting the motion :)
jlawryno has quit [Ping timeout: 480 seconds]
slattann has quit [Quit: Leaving.]
kts has joined #dri-devel
xantoz has quit [Read error: Connection reset by peer]
vliaskov has quit []
slattann has joined #dri-devel
slattann has quit []
JohnnyonFlame has quit [Ping timeout: 480 seconds]
kts has quit [Quit: Leaving]
jkrzyszt has quit [Remote host closed the connection]
alyssa has left #dri-devel [#dri-devel]
gawin has joined #dri-devel
Lucretia has quit [Read error: No route to host]
jkrzyszt has joined #dri-devel
Lucretia has joined #dri-devel
vliaskov has joined #dri-devel
jlawryno has joined #dri-devel
jlawryno has quit [Remote host closed the connection]
jlawryno has joined #dri-devel
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
gawin has quit [Ping timeout: 480 seconds]
itoral has quit []
itoral has joined #dri-devel
itoral has quit []
kts has joined #dri-devel
yuq825 has quit []
kts has quit [Quit: Leaving]
Akari has quit [Ping timeout: 480 seconds]
warpme____ has joined #dri-devel
fxkamd has quit []
jlawryno has quit [Ping timeout: 480 seconds]
Haaninjo has joined #dri-devel
ybogdano has joined #dri-devel
alyssa has joined #dri-devel
<alyssa> how long do I have until the branch point?
* alyssa sweats
tarceri has quit [Ping timeout: 480 seconds]
tarceri has joined #dri-devel
Shibe has quit []
gawin has joined #dri-devel
Shibe has joined #dri-devel
jlawryno has joined #dri-devel
Jeremy_Rand_Talos has quit [Remote host closed the connection]
Jeremy_Rand_Talos has joined #dri-devel
bgs has joined #dri-devel
heat has joined #dri-devel
* ccr notices the tight close-up shots of each contributor's sweaty faces and twitching trigger fingers in style of Sergio Leone spaghetti westerns.
bmodem1 has quit [Ping timeout: 480 seconds]
ajax has joined #dri-devel
ybogdano has quit [Ping timeout: 480 seconds]
<eric_engestrom> alyssa: 3 hours or so
<eric_engestrom> but if you need me to push it back I can do it tomorrow
<ajax> does anyone claim to understand all of the pipe_flush_flags, and if so, can they please write some documentation
<jenatali> ajax: You had left some comments on !19242 (CL/GL interop iface) - did you want to give that a full review?
* jenatali would like to land it
<ajax> jenatali: ooh, thanks for the reminder
* ajax trying to flush out a bunch of old MRs before the end of the year
<anholt> ajax: oh god, yeah, they're impenetrable.
jlawryno has quit [Ping timeout: 480 seconds]
Duke`` has joined #dri-devel
<jenatali> Heh I thought I knew flush flags, and then looked at our driver and we ignore all of 'em
<bnieuwenhuizen> ajax: aren't they documented in docs/gallium/context.rst ?
<ajax> _FENCE_FD isn't...
<ajax> but, yes. not in the header like every other enum in that header, but yes.
<anholt> docs separate from code is the worst.
<alyssa> eric_engestrom: woof okay
<anholt> alyssa: did you mean to re-marge the disk cache series?
<alyssa> anholt: momentarily, was waiting for someone to re-test the new version because apparently either the disk cache series or the FMA thing was regressing perf on somebody's mali platform
<anholt> ah
<anholt> hey nir folks, anyone up for a quick review of https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18218 to save the next victim some debug time?
iive has joined #dri-devel
<alyssa> eyes
<alyssa> i love reviewing nir
<alyssa> jenatali: What's a flush flag? ;-)
ybogdano has joined #dri-devel
Leopold has joined #dri-devel
mattrope has quit [Remote host closed the connection]
karolherbst has quit [Ping timeout: 480 seconds]
mattrope has joined #dri-devel
karolherbst has joined #dri-devel
devilhorns has quit []
<kisak> robclark: congratulations on freedreno reaching GL 4.5. How's it looking in the CTS side?
aravind has quit [Ping timeout: 480 seconds]
<mdnavare> agd5f: I was looking at the shared code with amdgpu dc, wanted to understand a few more data points there: 1. Have you over time seen an added advantage in terms of bugs triaging or validation due to the code sharing 2. What is the feedback you have got from your customers in terms of this approach- has it been hard or easy for them to consume this and maintain this upstream in their production trees etc?
<mdnavare> hwentlan_: If you have any insights into any of tje data points here?
<robclark> kisak: don't try to use qbo+timestamp queries, but otherwise not bad..
<alyssa> robclark: I wonder what Tegra does for that
<alyssa> Oh wait.. Tegra isn't a tiler is it
<alyssa> Hm. So there really are no GL4.x tilers then, you're a trailblazer ;-D
<robclark> nope, it is IMR
<robclark> hmm, I guess technically zink got there first.. but it does more of qbo on the cpu
<alyssa> bah, zink+turnip is just freedreno with more abstraction layers
<robclark> heheh
pjakobsson has quit [Remote host closed the connection]
MrCooper has quit [Remote host closed the connection]
MrCooper has joined #dri-devel
<alyssa> I am very curious how zink+turnip vs freedreno perf compares in the coming year
Guest228 has quit [Remote host closed the connection]
frieder has quit [Remote host closed the connection]
<agd5f> mdnavare, probably a better question for hwentlan_. My understanding from the display team is that they are able to get better testing coverage and leverage issues found on windows due to the code sharing. That said, it's not without it's pain points, but most of those are due to internal team politics. E.g., the pipeline validation code comes from the hw team and it's pretty terrible, but they won't help debug bandwidth issues if we don't use
<agd5f> their code.
<alyssa> oof
<mdnavare> agd5f: Yes the validation and testing coverage benefits makes sense, any pain points that you have seen or have got feedback from the end users/ customers/any consumers in general?
tursulin has quit [Ping timeout: 480 seconds]
tzimmermann has quit [Quit: Leaving]
<agd5f> mdnavare, nothing major comes to mind off hand.
alyssa has left #dri-devel [#dri-devel]
<eric_engestrom> jenatali: ci is already failing, because of `if (!ret != MESA_GLINTEROP_SUCCESS)` in st_interop.c
<jenatali> I saw, I cancelled
<jenatali> I wish MSVC had better warnings and/or I remembered to build with GCC before throwing at marge
jkrzyszt has quit [Remote host closed the connection]
<jenatali> Though... I guess only Clang caught it
<eric_engestrom> yeah clang has a lot more diagnostics like that, I think
<eric_engestrom> fyi the previous pipeline is still running: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19242/pipelines
<eric_engestrom> ah nvm
<eric_engestrom> not anymore
<eric_engestrom> (:
<jenatali> Took 3 tries for the "cancel all" button to actually work
<jenatali> I take it you're watching to see what merges before the branchpoint? :P
<eric_engestrom> hehe yep
<eric_engestrom> I don't want to branch off just before something important
<jenatali> 👍
<jenatali> Would be nice to get this in but the branchpoint doesn't mean too much to me, since we control our ship vehicle and tend to just follow main rather than stable branches
<jenatali> (Just quick checking if Clang catches anything else before reassigning)
srslypascal is now known as Guest262
srslypascal has joined #dri-devel
oneforall2 has quit [Quit: Leaving]
oneforall2 has joined #dri-devel
Guest262 has quit [Ping timeout: 480 seconds]
ybogdano has quit [Ping timeout: 480 seconds]
heat has quit [Remote host closed the connection]
heat has joined #dri-devel
Akari has joined #dri-devel
Leopold has quit []
<lynxeye> eric_engestrom: would be nice to give austriancoder a chance to land https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19013 before the branching.
<eric_engestrom> lynxeye, austriancoder: is it about to be merged, or does it need a while longer?
<eric_engestrom> ah, I see austriancoder just assigned it to marge :)
<austriancoder> eric_engestrom: jup..
ngcortes has joined #dri-devel
<eric_engestrom> providing the ci doesn't fail, it will be in; if it does fail, we'll see if it's a trivial fix but if it needs longer then I think I'll branch without it
apinheiro has quit [Ping timeout: 480 seconds]
<austriancoder> eric_engestrom: let's hope for the best :)
<eric_engestrom> austriancoder: that last change you did doesn't make the compiler happy
<eric_engestrom> > error: suggest explicit braces to avoid ambiguous 'else' [-Werror=dangling-else]
<eric_engestrom> I think you should just undo that last change
jkrzyszt has joined #dri-devel
<austriancoder> eric_engestrom: soo.. finally @home. commute took a little bit long. marge should be happy now
ngcortes has quit [Ping timeout: 480 seconds]
ybogdano has joined #dri-devel
<eric_engestrom> austriancoder: 🤞
<eric_engestrom> ... and merged :)
<austriancoder> eric_engestrom: it landed! Thanks for waiting :)
<eric_engestrom> austriancoder: np!
<eric_engestrom> nothing else in marge's queue, cutting the branch now :)
<jenatali> \o/
ybogdano has quit [Ping timeout: 480 seconds]
idr has quit [Remote host closed the connection]
rsjw has joined #dri-devel
mareko has quit [Read error: Connection reset by peer]
mareko has joined #dri-devel
gawin has quit [Ping timeout: 480 seconds]
gouchi has joined #dri-devel
nchery has quit [Remote host closed the connection]
jkrzyszt has quit [Remote host closed the connection]
lynxeye has quit [Quit: Leaving.]
<eric_engestrom> I should've prepared this in advance, but... what new deps are there since 22.2, other than rusticl stuff and intel-clc for rt on anv?
<eric_engestrom> for now I have this:
<eric_engestrom> A couple of notes, especially for packagers:
<eric_engestrom> - When building the Intel Vulkan driver with ray-tracing (using
<eric_engestrom> `-D intel-clc=enabled`, disabled by default), libclc is required
<eric_engestrom> (both as build and runtime dependency).
<eric_engestrom> disabled by default), introduces a bunch of new dependencies.
<eric_engestrom> - Rusticl, the OpenCL implementation (`-D gallium-rusticl=true`,
<eric_engestrom> Make sure you read docs/rusticl.rst (https://docs.mesa3d.org/rusticl)
<eric_engestrom> if you're considering enabling it.
<eric_engestrom> anything I should add?
<eric_engestrom> karolherbst: ^ for the rusticl note
<eric_engestrom> lstrano: ^ for the intel note
<karolherbst> eric_engestrom: yeah, sounds about right. Might be worth pointing out that no driver is enabled by default even when enabling it.
bgs has quit [Remote host closed the connection]
<eric_engestrom> karolherbst: will do, and how does the user change that? through the env var, or is there another setting?
<karolherbst> Yeah, through an env var. I've added documentation for those here: https://docs.mesa3d.org/envvars.html#rusticl-environment-variables
<karolherbst> (though putting there radeonsi as an example is missleading, because it's not enabled yet oops)
<eric_engestrom> got it, thanks!
<lstrano> eric_engestrom: looks good, thanks!
jlawryno has joined #dri-devel
<eric_engestrom> lstrano: thanks for confirming
<eric_engestrom> from grepping through the diff since 22.2, the only other new dep I could see was xcb-randr added in 47526556494f18cd2c02, but I don't think it's worth mentioning
<eric_engestrom> I'll leave it at that for the -rc1 email, but I'll repeat all of that in the final .0 email anyway
jkrzyszt has joined #dri-devel
jlawryno has quit [Ping timeout: 480 seconds]
yogesh_mohan has joined #dri-devel
yogesh_m1 has quit [Ping timeout: 480 seconds]
mvlad has quit [Remote host closed the connection]
dcz_ has quit [Ping timeout: 480 seconds]
ngcortes has joined #dri-devel
rasterman has quit [Quit: Gettin' stinky!]
<karolherbst> jenatali: what does "SEH exception with code 0xc0000005 thrown in the test body." mean?
<jenatali> Access violation, i.e. sigsegv
<karolherbst> ehhh
<jenatali> Context?
<karolherbst> https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18581 is hitting it and I have honestly no idea what I broke here :/
<karolherbst> or is that like a segfault in CPU code?
<jenatali> Looking
<jenatali> CPU segfault, yeah
<karolherbst> ahh
<karolherbst> yeah...
<karolherbst> I changed the gallium API :)
<jenatali> Which... it's running on a software driver so even a GPU segfault would be a CPU segfault
<karolherbst> ehh
<karolherbst> anyway, I touched launch_grid
<karolherbst> though it still makes no sense, dunno
<jenatali> I'm building it, gimme a sec
<karolherbst> I'd put my bet on this one, because that's I think the only commit touching anything related to CL on windws https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18581/diffs?commit_id=fc71b693ed5a48ce9706b8de57ba1228c73299de
<jenatali> Yep, it's that
<karolherbst> uhhh
<karolherbst> but why :(
<jenatali> Good question. var is null
<jenatali> Lemme re-run just that test with NIR debugging
<jenatali> Specifically https://gitlab.freedesktop.org/-/snippets/7156#LC402, this load is failing because the cast means the deref can't chase back to the var
<karolherbst> "vec1 8 ssa_14 = intrinsic load_deref (ssa_36) (access=0)" this one?
<karolherbst> ohh
<karolherbst> down there
<jenatali> vec1 8 ssa_339 = intrinsic load_deref (ssa_338) (access=0)
<karolherbst> ehhh
ngcortes has quit [Ping timeout: 480 seconds]
<jenatali> This is a test for loading a packed private struct kernel arg
<karolherbst> ehhhh
<karolherbst> that's very annoying as some drivers rely on proper infos there :/
apinheiro has joined #dri-devel
<jenatali> It's already a ptr_as_array when coming from vtn
<karolherbst> right.. but that's totally fine
<jenatali> Yeah just pointing out that it's not any custom special lowering I'm doing that's causing it :P
ngcortes has joined #dri-devel
<karolherbst> wonder if we should just yolo it and set the range to the entire uniform space *sigh*
<jenatali> Just some god-awful CL C
<karolherbst> but the entire range would be wrong as well..mhh
<karolherbst> though maybe shader->num_uniforms would be fine
<karolherbst> jekstrand: any thoughts on https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/18581/diffs?commit_id=fc71b693ed5a48ce9706b8de57ba1228c73299de and doing a nir_intrinsic_set_range(load, b->shader->info.num_uniforms); instead?
<karolherbst> apparently we can't always track back to the var, but maybe I also missunderstood what the intrinsic range is anyway
<karolherbst> yeah..
Duke`` has quit [Ping timeout: 480 seconds]
<jenatali> Man this generated SPIR-V is garbage :) I don't get why there's all these casts
<karolherbst> ask LLVM
<karolherbst> *insert they played us for fools meme LLVM edition*
<jekstrand> karolherbst: I don't think that's correct.
<jekstrand> karolherbst: Typically, when base/range is used, the base is added to the offset and the end address is base+range.
<jekstrand> karolherbst: You're kind-of doing half of that
<karolherbst> right...
<karolherbst> but then load_constant just sets the range to the entire size ... and I was wondering if I can just do the same
<karolherbst> I kind of thought that range is like the worst case bound of the access
<jekstrand> Yeah, that could work, probably.
<karolherbst> once I lower to ubo the info is all overwritten anyway and I just wanted to make nir_validate happy
<karolherbst> ehh well.. I use the same info
<karolherbst> yeah.. I just use num_uniforms
<karolherbst> if somebody cares they can work on limiting the range
apinheiro has quit [Ping timeout: 480 seconds]
<jenatali> Is num_uniforms a count or a size?
<jenatali> Doesn't the range need to be a size?
<karolherbst> for CL it's a size
<jenatali> Ah ok
<jekstrand> jenatali: It's always a size of some form. Sometimes a size in units of vec4s but always a size.
<jenatali> Ack
<karolherbst> I think I broke my git
<karolherbst> rip
<karolherbst> ohh wait...
<karolherbst> uhhhh
<karolherbst> I shouldn't cherry-picks ranges while rebasing
<karolherbst> nvm
nchery has joined #dri-devel
gouchi has quit [Remote host closed the connection]
ngcortes_ has joined #dri-devel
ngcortes_ has quit []
ybogdano has joined #dri-devel
danvet has quit [Ping timeout: 480 seconds]
glisse has quit [Read error: Connection reset by peer]
glisse has joined #dri-devel
Jeremy_Rand_Talos_ has joined #dri-devel
Jeremy_Rand_Talos has quit [Remote host closed the connection]
mszyprow has quit [Ping timeout: 480 seconds]
heat has quit [Remote host closed the connection]
jkrzyszt has quit [Remote host closed the connection]
heat has joined #dri-devel
Jeremy_Rand_Talos__ has joined #dri-devel
Jeremy_Rand_Talos_ has quit [Remote host closed the connection]
pcercuei has quit [Read error: Connection reset by peer]
pcercuei has joined #dri-devel
jkrzyszt has joined #dri-devel
ahajda_ has joined #dri-devel
ahajda has quit [Read error: Connection reset by peer]
ybogdano has quit [Ping timeout: 480 seconds]
pcercuei has quit [Quit: dodo]
ahajda_ has quit []
Haaninjo has quit [Quit: Ex-Chat]
jkrzyszt has quit [Ping timeout: 480 seconds]