bmodem has quit [Remote host closed the connection]
bmodem has joined #dri-devel
sgruszka has joined #dri-devel
tzimmermann_ has quit []
tzimmermann has joined #dri-devel
<tzimmermann>
danvet, airlied, the most recent drm-misc-next PR has not been merged. is that intentional?
<dolphin>
tzimmermann: when'd you send it? drm-intel-gt-next not merged either but that was on Friday
<tzimmermann>
dolphin, it was sent by mlankhorst on thursday
bmodem1 has joined #dri-devel
<dolphin>
I would assume just busyness with -fixes
bmodem has quit [Ping timeout: 480 seconds]
MajorBiscuit has joined #dri-devel
bmodem has joined #dri-devel
bmodem1 has quit [Ping timeout: 480 seconds]
<tzimmermann>
after -rc6, it might not make much sense to merge the PR now
<dolphin>
oh, no separate drm-misc-fixes? I think I keep forgetting that
bmodem has quit [Ping timeout: 480 seconds]
bmodem has joined #dri-devel
bmodem has quit [Read error: Connection reset by peer]
mvlad has joined #dri-devel
<darkbasic4>
karolherbst: I've tried darktable git master: same issue with the black images and failing kernels (but slightly different logs). Fun fact arecibo benchmark is twice as fast now, not sure if you did some wizardry in rusticl in the past week or darktable did
<karolherbst>
probably darktable? dunno
<karolherbst>
is the output fine?
<darkbasic4>
no, still black as before
<karolherbst>
ahh
<karolherbst>
sadly I have no idea if I even be able to look into more rusticl stuff this week :/ Have some work stuff and nouveau regression to take care of... though you could check if using llvmpipe is any better
<darkbasic4>
karolherbst: do I need -Dgallium-opencl="icd" for llvmpipe? Because it builds clover and I don't want darktable to bitch about multiple cl implementations
<karolherbst>
no
<karolherbst>
just need to add swrast as a gallium driver
<karolherbst>
and use RUSTICL_ENABLE=lp
<karolherbst>
darktable might complain that it's a CPU device
<karolherbst>
then use RUSTICL_DEVICE_TYPE=gpu
bmodem has joined #dri-devel
bmodem1 has joined #dri-devel
yuq825 has left #dri-devel [#dri-devel]
<darkbasic4>
karolherbst: oh wait you didn't mean to use LP_CL=1 right? What's the difference with RUSTICL_ENABLE=lp?
<karolherbst>
the old env vars won't work with rusticl anymore, because how different things are. RUSTICL_ENABLE is the only way of enabling drivers and also tidies up the mess of each driver needing their own.
bmodem has quit [Ping timeout: 480 seconds]
jfalempe has quit [Ping timeout: 480 seconds]
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
gawin has joined #dri-devel
pekkari has joined #dri-devel
<darkbasic4>
karolherbst: without RUSTICL_DEVICE_TYPE=gpu I get "discarding new device as emulated by CPU" in darktable, while adding it I get "marked as disabled" and so still no opencl acceleration. both are with RUSTICL_ENABLE=lp obviously
DPA has quit [Ping timeout: 480 seconds]
<karolherbst>
heh...
<karolherbst>
why is it marked as disabled tho?
<karolherbst>
too slow?
srslypascal is now known as Guest2113
srslypascal has joined #dri-devel
Guest2113 has quit [Ping timeout: 480 seconds]
DPA has joined #dri-devel
Ahuj has quit [Quit: Leaving]
zehortigoza has joined #dri-devel
kts has quit [Quit: Leaving]
itoral has quit [Remote host closed the connection]
<darkbasic4>
karolherbst: no idea, but if that's the reason is finds out so BEFORE starting the benchmark: https://pastebin.com/i8nV9zHi
jfalempe has quit [Remote host closed the connection]
<darkbasic4>
karolherbst: [opencl_init] due to a slow GPU the opencl flag has been set to OFF.
<darkbasic4>
WTF darktable stop overriding it :D
<karolherbst>
yeah
<karolherbst>
it does some quick check
<darkbasic4>
karolherbst: lol I've increased the cl timout and it still found a way to disable it: [opencl_init] initial status of opencl enabled flag is OFF.
<darkbasic4>
cl disabled flag is off but cl enabled flag is off as well XD
<karolherbst>
yeah dunno.. I just know I somehow made it work at some point
fxkamd has joined #dri-devel
mairacanal[m] has quit []
jfalempe has joined #dri-devel
kts has joined #dri-devel
<darkbasic4>
karolherbst: finally managed to bypass all the checks: llvmpipe works fine, no black images
dviola has quit [Ping timeout: 480 seconds]
<karolherbst>
cool
<karolherbst>
yeah.. so I suspect there is something weird happening with discrete GPUs, because the code is a little bit different for those
jkrzyszt has quit [Remote host closed the connection]
jfalempe has quit [Ping timeout: 480 seconds]
junaid has quit [Read error: No route to host]
rasterman has quit [Remote host closed the connection]
jlawryno has quit [Ping timeout: 480 seconds]
rasterman has joined #dri-devel
djbw has joined #dri-devel
camus has quit [Ping timeout: 480 seconds]
aravind has quit [Ping timeout: 480 seconds]
<clever>
javierm: abuse all the connectors :D
<javierm>
clever: haha, yeah. I read that and remembered our conversation :)
heat has joined #dri-devel
jfalempe_ has quit [Ping timeout: 480 seconds]
jfalempe has joined #dri-devel
junaid has joined #dri-devel
pH5 is now known as pza[afk]
pza[afk] has quit []
frieder has quit [Remote host closed the connection]
alyssa has left #dri-devel [#dri-devel]
fab has quit [Quit: fab]
bmodem1 has quit [Ping timeout: 480 seconds]
Duke`` has joined #dri-devel
janusz has joined #dri-devel
tzimmermann has quit [Quit: Leaving]
<janusz>
Hi, I wanted to report a bug, but when created account, I am getting this error on login: Your account is pending approval from your GitLab administrator and hence blocked. Please contact your GitLab administrator if you think this is an error. How can I proceed?
<jekstrand>
daniels: ^^
<daniels>
janusz: did you receive an email from GitLab asking you to confirm your email address?
<janusz>
daniels: yes, and I did
<janusz>
just got email that my request has been approved. thanks
Company has joined #dri-devel
Lucretia has quit [Remote host closed the connection]
sgruszka has quit [Remote host closed the connection]
Lucretia has joined #dri-devel
<daniels>
cool :) that’s the automatic mechanism in action
jfalempe_ has joined #dri-devel
Thymo_ has joined #dri-devel
tursulin has quit [Ping timeout: 480 seconds]
Thymo has quit [Ping timeout: 480 seconds]
jfalempe has quit [Ping timeout: 480 seconds]
<jekstrand>
?D
<jekstrand>
:D
fab has joined #dri-devel
MajorBiscuit has quit [Ping timeout: 480 seconds]
<karolherbst>
(we really should add a spoiler that users should use that "Sign in with" instead)
kts has quit [Quit: Leaving]
<emersion>
and have their GitLab account tied to another service's security?
<emersion>
and a proprietary service at that?
JohnnyonFlame has joined #dri-devel
<jekstrand>
I trust Google's infosec more than I trust fd.o. No offence intended to the fd.o admins, of course.
<DavidHeidelberg[m]>
I would take that as a offence. :D
<bnieuwenhuizen>
jekstrand: am actually more afraid of the "google locking me out and having no customer support" case
<karolherbst>
you can also connect github/twitter (haha lol) and gitlab
<bnieuwenhuizen>
that said, you can make an account using email and then log in with google while using the original creds as backup
<karolherbst>
that defeats the purpose of not having to wait on that activation email
<karolherbst>
(and not having to get approved by that and what not)
JohnnyonF has quit [Ping timeout: 480 seconds]
<karolherbst>
but you can sign in and the do a password reset afaik
sarahwalker has quit [Remote host closed the connection]
ybogdano has joined #dri-devel
<jekstrand>
Yeah, I'm gonna tie it to Twitter... 6 months ago, that might have sounded like a really good idea but not today. :P
<jekstrand>
I trust Marge to merge my patches the first time more than I trust Twitter to not go down at this point.
maxzor has joined #dri-devel
<karolherbst>
lol
<MrCooper>
poor Marge, getting blamed for the CI over which she has no control :)
<daniels>
jekstrand: absolutely none taken. Google are somewhat better at security than us
jlawryno has joined #dri-devel
alanc has quit [Remote host closed the connection]
<DavidHeidelberg[m]>
MrCooper: Maybe we can call the CI Bart? :D
FireBurn has joined #dri-devel
<MrCooper>
hehe
alanc has joined #dri-devel
<DavidHeidelberg[m]>
Bumping the meson in CI from the 0.61 to 0.63, anyone see obvious issue with that? I would like to use the new script bundled with it "meson env2mfile" for cross builds
orbea has quit [Remote host closed the connection]
orbea has joined #dri-devel
Lucretia has joined #dri-devel
Akari has joined #dri-devel
Lucretia has quit [Remote host closed the connection]
bgs has joined #dri-devel
<airlied>
danvet: will do next today, had a couple of days of low bandwidth :-p
bluebugs has quit [Quit: Leaving]
maxzor has quit [Ping timeout: 480 seconds]
iive has joined #dri-devel
pekkari has quit [Quit: Konversation terminated!]
jlawryno has joined #dri-devel
MajorBiscuit has joined #dri-devel
srslypascal is now known as Guest2138
srslypascal has joined #dri-devel
jlawryno has quit [Ping timeout: 480 seconds]
Guest2138 has quit [Ping timeout: 480 seconds]
jhli has quit [Remote host closed the connection]
jfalempe_ has quit [Ping timeout: 480 seconds]
lemonzest has quit [Quit: WeeChat 3.6]
Lucretia has joined #dri-devel
Haaninjo has joined #dri-devel
<DavidHeidelberg[m]>
eric_engestrom: do you have some tests you disabled due to OOM? I'm currently trying to find anything which is over the edge and could be tested if enough memory is provided (for zram MR)
mbrost has joined #dri-devel
<eric_engestrom>
DavidHeidelberg[m]: now, I haven't been able to narrow down the test list as basically any change makes it not reproduce (we think we need a very specific combo of tests that happen in parallel)
<eric_engestrom>
the OOMs do seem to happen less often when disabling some caches, so we've been exploring that
<DavidHeidelberg[m]>
eric_engestrom: btw. I ended up on your 512M, everything above so far seems to be unused, but I'll see in future with the traces
mbrost has quit [Ping timeout: 480 seconds]
<eric_engestrom>
yeah when looking at each test's peak memory footprint the largest was 241 MB; taking the 4 biggest together would never reach anything near 1 GB, much less 4 GB
<eric_engestrom>
it's been very puzzling
<eric_engestrom>
but I'm happy to run that zram patch of yours in a loop to see if it alleviates the problem
<eric_engestrom>
DavidHeidelberg[m]: do you have a branch ready to test?
<eric_engestrom>
DavidHeidelberg[m]: thanks! I'll run that overnight
mvlad has quit [Remote host closed the connection]
MajorBiscuit has quit [Quit: WeeChat 3.6]
junaid has quit [Ping timeout: 480 seconds]
<DavidHeidelberg[m]>
thank you Eric! :)
dcz_ has quit [Ping timeout: 480 seconds]
junaid has joined #dri-devel
maxzor has joined #dri-devel
maxzor has quit [Remote host closed the connection]
JohnnyonFlame has quit [Ping timeout: 480 seconds]
JohnnyonFlame has joined #dri-devel
alyssa has joined #dri-devel
* alyssa
stares at GitLab
<alyssa>
1000 open MRs
<alyssa>
review bankruptcy time? :-p
<ajax>
only!
<alyssa>
meanwhile
* alyssa
grumbles about not being able to do nir_lower_{blend, fragcolor, texcoord_replace, clip_fs} on lowered I/O
<DavidHeidelberg[m]>
alyssa: it's surprising how many MR's on last page are recently updated :D
<DavidHeidelberg[m]>
made 3 years ago, updated in last 5 days :D
<alyssa>
DavidHeidelberg[m]: heh, yeah
<alyssa>
we lost a lot of last updated information when we retargeted master->main
<ajax>
gitlab analytics notes that we're averaging 57.1 commits per day for our last 2000 commits
junaid has quit [Ping timeout: 480 seconds]
<alyssa>
ajax: of which a concerningly large 5.3 of those commits per day are mine >.>
<ajax>
633 merge requests landed in the last 35 days
<ajax>
1000 MRs ago was september 20
<alyssa>
wild
<alyssa>
I do wonder if we should be closing zombie MRs
<anholt>
20/day is about 2x faster than I expected we were pushing through. nice.
<alyssa>
stuff where the original author has stopped caring, the code is blocked on review feedback, and the change isn't interesting enough for anyone else to pick up
tarceri has quit [Ping timeout: 480 seconds]
<alyssa>
it's just tricky... feels 'rude' for anyone other than the original author to close an MR
Duke`` has quit [Ping timeout: 480 seconds]
<alyssa>
looking through the MR graveyard I see lots of stuff that probably could be closed
<alyssa>
but if I didn't author the patch and I don't work on that part of the tree, who am I to make that judgement?
<ajax>
you're 9% of the commit log, that's who you are
<alyssa>
ajax: concerning.
<alyssa>
my official title is "Software Developer" (-:
<javierm>
alyssa: maybe you can ask in the MR if the author (or reviewers) feel that it could be closed due inactivity / nobody caring anymore ?
<ajax>
i do like to go through old stuff and either update or close out while other stuff is compiling
<javierm>
alyssa: then if in some time nobody asked, well is a way for them to answer :)
<alyssa>
jekstrand: is an interesting case of the above, with a massive number of Intel MRs that jekstrand is obvious not going to finish off anytime soon but no good reason to close in itself
<javierm>
s/nobody asked/nobody responds
<ajax>
also, /reopen works. it's not like closing an MR is permanent, and even if it was you still have the commits and can make a new MR from them
<alyssa>
yeah..
danvet has quit [Ping timeout: 480 seconds]
<alyssa>
at any rate I feel like hitting 1000 open MRs, this is probably the time to figure out a policy for this
<ajax>
might be worth making a 'backburner' label for MRs that are old and unattended but still interesting enough to keep open
<alyssa>
As a reviewer at least, it's frustrating not knowing what MRs I need to look at, with so many open
<ajax>
and then like, send the 10 oldest non-backburner MRs to mesa-dev@ every week, and whatever doesn't find a champion befoe the next week gets closed
<ajax>
(would be a really good way to draw in new contributors, this, since it'd be a constant stream of maybe interesting projects)
<alyssa>
I'm listening :p
<alyssa>
for lots of these MRs it's not even clear why they were forgotten
<alyssa>
Does the change no longer apply?
<alyssa>
Do they just need a review (and a rebase)?
<alyssa>
Did the author stop caring but left the code for posterity?
<alyssa>
Some combination thereof?
<alyssa>
Without knowing... it's hard to know what to do
<ajax>
yep. all of these things, we can label, and then search for.
<alyssa>
For my own MRs, I'm trying to use the scheme where *any* MR that is not *completely ready to land conditional on reviews* gets marked Draft
<anholt>
sounds like a good goal.
<alyssa>
(If a review comment requires a nontrivial change that I don't do immediately, the MR gets marked as Draft until I make that change and am ready for a new round for review.)
<DavidHeidelberg[m]>
alyssa: time to time, when you ping author he remembers and rebase :D
<alyssa>
(And cynically for stuff in my drivers that people don't review ... the non-Draft MRs is the assign to Marge queue :-p)
* airlied
has a few MRs where I probably merged 90% of them via another MR and forgot about the original
<alyssa>
airlied: Yep. And it's hard for anybody else to tell that at a glance.
<alyssa>
And then there are the especially 'interesting' cases where an MR is blocked and the original author isn't even in the project anymore
<eric_engestrom>
gonna go to sleep now, I'll see if there's something obvious tomorrow
jfalempe has quit [Ping timeout: 480 seconds]
<eric_engestrom>
ah wait, nvm, there is something obvious: we're not using that kernel, so the config change + rebuild needs to be done on our side as well
<DavidHeidelberg[m]>
eric_engestrom: I'm also off PC, I'll check tomorrow
<eric_engestrom>
no need, it's a problem on our side :)
<eric_engestrom>
I know *what* to do, but I'll have to ask a colleague to know *where* to do it ^^
<DavidHeidelberg[m]>
good, thanks :)
fab has quit [Quit: fab]
jfalempe has joined #dri-devel
jfalempe has quit []
vliaskov has quit [Remote host closed the connection]
ahajda_ has quit [Ping timeout: 480 seconds]
gawin has quit [Ping timeout: 480 seconds]
rasterman has quit [Quit: Gettin' stinky!]
jkrzyszt has joined #dri-devel
jkrzyszt has quit [Remote host closed the connection]