<dj-death>
ah... nir_lower_io_to_vector() only works for in/out atm
Duke`` has joined #dri-devel
<haagch>
oh I think I read the rusticl compile error wrong. it's the compiler compiling rusticl that is clang 14 and failing in clang headers, and *not* that it fails because it would try to build rusticl against clang 14 headers
MrCooper has quit [Remote host closed the connection]
mrkajetanp has quit [Remote host closed the connection]
frieder has quit [Remote host closed the connection]
tobiasjakobi has joined #dri-devel
bgs has quit [Remote host closed the connection]
MrCooper has joined #dri-devel
lstrano_ has joined #dri-devel
mhenning has joined #dri-devel
cleverca22[m] has quit []
heat has joined #dri-devel
mbrost has quit [Remote host closed the connection]
mbrost has joined #dri-devel
<haagch>
I think the rusticl_opencl_bindings_rs bindgen needs the include path from the llvm/clang that the mesa build has found added to include_directories
<haagch>
manually adding include_directories('/usr/lib/llvm-mesa/lib/clang/16/include'), to it seems to completely screw the standard includes though https://hastebin.com/ohacosakaj.txt
fab has joined #dri-devel
mbrost has quit [Ping timeout: 480 seconds]
mbrost has joined #dri-devel
MajorBiscuit has quit [Ping timeout: 480 seconds]
Guest1451 has quit [Remote host closed the connection]
YuGiOhJCJ has quit [Quit: YuGiOhJCJ]
JohnnyonFlame has quit [Ping timeout: 480 seconds]
bmodem2 has quit [Ping timeout: 480 seconds]
JohnnyonFlame has joined #dri-devel
tursulin has quit [Ping timeout: 480 seconds]
alanc has joined #dri-devel
kts has quit [Quit: Leaving]
<DavidHeidelberg[m]>
eric_engestrom: Hey. Check the script for sure, invocation is very easy, you can force manual jobs, do stress testing & use regex... Running is easy: `.gitlab-ci/bin/ci_run_n_monitor.py --rev $(git rev-parse HEAD) --target "v3d-rpi4-traces" --force-manual`
ybogdano has quit [Ping timeout: 480 seconds]
ybogdano has joined #dri-devel
djbw has joined #dri-devel
jkrzyszt has quit [Remote host closed the connection]
<DavidHeidelberg[m]>
What is "Kopper"? Why there is no docs for it?
<eric_engestrom>
DavidHeidelberg[m]: thanks for pointing me to that script :)
<eric_engestrom>
from reading the code `--stress` retries the script until you ^C, right? I think that might be what I want
<DavidHeidelberg[m]>
eric_engestrom: I'll be most happy if it gets used :D
<eric_engestrom>
* retries the job
<DavidHeidelberg[m]>
eric_engestrom: exactly!
<eric_engestrom>
I'll run that overnight then :)
<DavidHeidelberg[m]>
eric_engestrom: you will see failed jobs, has clickable links to the log. We usually giving a stress test to the new jobs, like 100x times before we enable them and it's nice, since you don't have to track dependencies, click and stuff
<DavidHeidelberg[m]>
I somehow thought it's more well known... I should invest more into Ads I think :D
<eric_engestrom>
I was basically trying to do the same thing but bypassing gitlab, so that it would just be an infinite loop around `deqp-runner ...` and count the pass/fail
codingkoopa9 has quit []
<eric_engestrom>
but it's a pain to set up over serial
<eric_engestrom>
your solution is heavier/slower but so much easier :D
<DavidHeidelberg[m]>
yeah. But maybe we can add "injection" to loop into the scripts.. I was thinking that it still has huge runtime cost
<eric_engestrom>
(oh, and that way I was sure I was running it on one of the two devices where the failure has been seen, in case it's relevant; I don't think it's possible to do that when going through gitlab, right?)
<DavidHeidelberg[m]>
some variable like "REPEAT: 100" and it'll iterate... but only until LAVA won't kill it
<DavidHeidelberg[m]>
eric_engestrom: if it fails only specific device, this script cannot pick. What gitlab gives it's used.
<eric_engestrom>
yeah, I figured gitlab didn't have a way to do that
<eric_engestrom>
wait actually, since you pointed me to that "daily report" thing, I haven't looked at all of them to see if there are other devices
codingkoopa9 has joined #dri-devel
Ryback_ has joined #dri-devel
<eric_engestrom>
gonna do that once I've started `ci_run_n_monitor.py` from a server
<DavidHeidelberg[m]>
iterate the report I linked, it's like 1 month ago - 2 months ago I think and there is like 20 occurences
<DavidHeidelberg[m]>
eric_engestrom: I think you can pick the exactly "3/8" part if you specify it in the name of the job
<eric_engestrom>
yeah I was going to try that; no point using up 8x as many resources when we know the rest don't have issues :0
Leopold_ has quit []
Lucretia has quit [Read error: No route to host]
Leopold_ has joined #dri-devel
<DavidHeidelberg[m]>
eric_engestrom: if you have any questions about the script, feel free to ping me :) beware that any job name input is python regex and also script cannot "restart pipeline setup" when already run on one pipeline with specific job name(s)
<eric_engestrom>
I'm not sure I understand the second part of the last sentence?
Lucretia has joined #dri-devel
<DavidHeidelberg[m]>
eric_engestrom: if you run it once with wrong name or setting which won't invoke the job, usually second run doesn't work (you have to reset pipeline from the gitlab UI or change the code)
<DavidHeidelberg[m]>
*change the code and push again
<eric_engestrom>
"reset pipeline from the gitlab UI" what does reset mean here?
<emersion>
daniels: any plan to spin up again your DRM format modifiers docs?
<emersion>
daniels: if not, maybe i can adopt the patch?
<mupuf>
anholt: yes, I am trying to prevent unknown users from starving Marge or other devs
<anholt>
ok, then please make the jobs manual in that case.
<mupuf>
You should be able to trigger the job from the context of the MR
<mupuf>
anholt, do you know what I mean?
<anholt>
When I click play on the container jobs at the start, please don't run your job automatically if it's going to fail for reasons other than the MR breaking the job.
<mupuf>
Oh, right, I can easily detect that, good idea :)
<mupuf>
Thanks, will do!
jkrzyszt has quit [Ping timeout: 480 seconds]
<anholt>
thanks!
yogesh_m1 has quit [Ping timeout: 480 seconds]
Kayden has quit [Quit: to office & lunch?]
guru_ has quit [Remote host closed the connection]
oneforall2 has joined #dri-devel
yogesh_m1 has joined #dri-devel
oneforall2 has quit [Remote host closed the connection]
oneforall2 has joined #dri-devel
Akari has joined #dri-devel
dcz_ has quit [Ping timeout: 480 seconds]
Jeremy_Rand_Talos__ has quit [Remote host closed the connection]
srslypascal has quit [Ping timeout: 480 seconds]
Jeremy_Rand_Talos__ has joined #dri-devel
ngcortes has quit [Ping timeout: 480 seconds]
lynxeye has quit [Quit: Leaving.]
danvet has quit [Ping timeout: 480 seconds]
gawin has joined #dri-devel
Leopold_ has quit []
<gawin>
is someone familiar with converting d3d9 traces to opengl ones? I heard it's doable with wine
<DavidHeidelberg[m]>
gawin: traces like apitrace .traces ?
ybogdano has quit [Ping timeout: 480 seconds]
<gawin>
yeah
<DavidHeidelberg[m]>
never tried, but if you play d3d9 trace in wine and it converts to the gl and you `apitrace trace` wine...
mbrost has quit [Remote host closed the connection]
tobiasjakobi has quit []
mbrost has joined #dri-devel
<gawin>
this makes sense (I was suspecting there's dedicated tool)
ybogdano has joined #dri-devel
Leopold_ has joined #dri-devel
mvlad has quit [Remote host closed the connection]
rmckeever has joined #dri-devel
gouchi has joined #dri-devel
JohnnyonF has joined #dri-devel
Leopold_ has quit [Remote host closed the connection]
Leopold_ has joined #dri-devel
JohnnyonFlame has quit [Ping timeout: 480 seconds]
nchery has joined #dri-devel
<DavidHeidelberg[m]>
haha! anholt thanks for pushing me in the right direction, I just had to fake few files with drm_shim_override_file like nouveau does it
<anholt>
oh, good.
tagr_ has joined #dri-devel
tagr_ has quit [Remote host closed the connection]
<robclark>
DavidHeidelberg[m]: it would be nice to have similar for per-a6xx but would require finding registers to tell us how much is queued up in ROQ, otherwise they will start hitting false negatives
<DavidHeidelberg[m]>
robclark: how much hardcore work that would be? thinking about Jami Kettunen who reported the a540 issues recently
<DavidHeidelberg[m]>
maybe he could look into it and he has also HW available
<robclark>
you'd need to know low levels reasonably well.. in order to construct scenarios where gpu hangs at a known location, and then dump registers and try to make sense of it
<robclark>
not that this isn't really a *fix* so much.. it isn't really all that reasonable that (for normal desktop scenario) you have rendering that takes more than 500ms.. so hitting hangcheck is more likely a symptom than the root problem
<DavidHeidelberg[m]>
right, so lot of hardcore work...
<robclark>
yeah, and not really addressing the root cause
nchery has quit [Ping timeout: 480 seconds]
pzanoni has joined #dri-devel
ngcortes_ has joined #dri-devel
ahajda has quit [Ping timeout: 480 seconds]
ngcortes has quit [Ping timeout: 480 seconds]
ngcortes_ has quit []
ngcortes_ has joined #dri-devel
ngcortes_ has quit [Remote host closed the connection]
ngcortes has joined #dri-devel
pcercuei has quit [Quit: dodo]
Mangix has quit [Read error: Connection reset by peer]