02:13 gentlewonder: those millenials like me what happened to them is once in a time only, it is unique, i became very skilled and smart in the process of following LINUX advancements and all stack build ups, not all i completely agree with but all this is interesting for me enough. So i have something to fill out my days with even when being injured. The stack is good enough to put performance on steroids or do something unexpectedly neat.
07:31 MrCooper: doras: the compositor also needs to wait for fences to signal first to prevent slow clients from delaying its output frames
07:42 pq: doras, if you allowed an atomic commit to update different planes at different times, it would not be atomic anymore. Maybe you'd also then need per-plane flip completion events? Sounds really complicated to make work corerctly in both userspace and kernel, much more complicated than making userspace check the fences before committing.
09:42 tagr: karolherbst, ric96: fwiw, you might be better off building the upstream U-Boot version for Jetson Nano (p3450-0000) if you want to boot an upstream kernel
09:42 tagr: the downstream U-Boot version has a bunch of stuff that either isn't supported upstream anyway, or flat out breaks things (like that node copying stuff)
09:46 tagr: ric96: feel free to post a link to your boot log and I can take a look
10:36 ric96: tagr: its upstream uboot
12:00 tagr: ric96: upstream U-Boot shouldn't have that issue with the node copying because the node copying code doesn't exist there, as far as I know
12:18 thaytan: Lyude, I can't get the panel brightness stack to apply cleanly on Linus' tree. It looks like there are changes in drivers/gpu/drm/i915/display/intel_panel.c to use pwm_state that aren't upstream yet?
12:30 ric96: tagr: then it might be something else.
12:30 ric96: I tried grepping tegra-gpu, drm, firmware, nouveau on dmesg. I just got the expected stuff on tegra-gpu and drm. Rest was just empty.
12:33 tagr: thaytan: those changes should be in drm-misc-next
12:34 tagr: ric96: what exactly is the failure that you're seeing? unfortunately I only have half of the scrollback from your discussion with karolherbst
12:34 tagr: ric96: can you post the complete boot log somewhere?
12:37 ric96: tagr: nouveau loads but doesn't latch, no error on dmesg. This happens only on 5.9 and not on 5.8.
12:37 ric96: 5.8 works as expected.
12:37 ric96: I'll get the log in a couple of hours.
12:38 karolherbst: ric96: nouveau doesn't print a single line in dmesg, correct?
12:38 karolherbst: tagr: that kind of looks like the dts fail, but maybe with 5.9 something changed again :/
12:39 karolherbst: maybe we want to add some error prints in the platform codepath
12:39 ric96: karolherbst: correct
13:02 tagr: odd... I've been running linux-next and Nouveau works just fine for me
13:02 tagr: ric96: is this on Jetson Nano?
13:03 ric96: tagr: yes
13:08 jcristau: OAOBOAOB
13:29 thaytan: tagr, thanks
13:38 mripard: danvet_: so it looks like -rc8 is fairly incertain, I should probably just make a final PR for drm-misc-next today?
13:38 mripard: https://lore.kernel.org/lkml/CAHk-=wjSWKpiNbcMpzQ9fBUA5s11sU-STLztzQkHhCHtYvLP5A@mail.gmail.com/
13:44 thaytan: tagr, doesn't look like they are in drm-next or drm-misc-next. I'll wait for Lyude and see
13:44 danvet_: mripard, would be nice I think
13:45 danvet_: but not the end if those few patches misses
13:52 mripard: ack, I'll send it by the end of the day
15:31 Lyude: thaytan: apply against drm-tip
15:55 thaytan: Lyude, ta, that was much easier
18:28 DPA: If I have some update onscreen with X11 with dcss+etnaviv, but not with mxsfb+etnaviv, is it save to assume that it's an mxsfb issue, or could it still be something else?
18:32 DPA: s/some/something/
19:17 Lyude: btw seanpaul will try to take a look at that backlight patch that you sent out today
19:17 seanpaul: thanks Lyude!
22:33 jekstrand: airlied: How do you recommend testing vallium?
22:33 jekstrand: airlied: How close to conformance is it?