13:52 bylaws: Hi
13:53 bylaws: has anyone tested Nouveau on tegra recently
13:54 bylaws: it doesn't seem to be working on x1
13:54 bylaws: [ 564.416098] nouveau 57000000.gpu: gr: DATA_ERROR 0000009c [] ch 4 [04003e2000 xfwm4[4291]] subc 0 class b197 mthd 0d78 data 00000014
17:22 karolherbst: bylaws: what version of mesa is that with?
21:05 bylaws: karolherbst: 19.2.6 I think
21:06 karolherbst: I think I know what bug this is .. mhh
21:06 bylaws: fwiw I was also experiencing this 7months back in android
21:06 bylaws: and 5 or so months back in plasma mobile
21:06 karolherbst: bylaws: can you run kmsgears on the tty?
21:08 karolherbst: ohh, wait, the bug I am thinking about should leave you with a crash
21:08 karolherbst: bylaws: I have a jetson nano myself, so I could check if anything is up I am not aware of
21:08 karolherbst: what tegra do you have?
21:09 bylaws: sure, 10min
21:09 bylaws: x1
21:09 bylaws: device is switch
21:09 bylaws: but it's exactly the same soc
21:10 bylaws: kernel crash?
21:12 bylaws: <bylaws "sure, 10min"> I'll check tommorow
21:13 bylaws: kmsgears is in Mesa demos?
21:27 karolherbst: uhm.. no idea
21:27 karolherbst: with crash I meant a crash inside mesa
21:27 karolherbst: or.. well, rather an assert I think
21:27 karolherbst: so a release build might not hit it
21:31 karolherbst: let's see... I can check onmy nano as well
21:32 karolherbst: ahh, even my ssh key still works :)
21:32 karolherbst: bylaws: I meant kmscube, sorry
21:32 karolherbst: and start it like this: ./kmscube -D /dev/dri/card1
21:33 karolherbst: https://gitlab.freedesktop.org/mesa/kmscube.git
21:38 karolherbst: imirkin: any idea what we should do regarding the modifier stuff on the tegras?
21:39 imirkin: erm ... what's the problem?
21:39 karolherbst: we can't draw into the buffer we get from tegra drm I think? this zsbuf issue
21:40 imirkin: oh, that modifier
21:40 karolherbst: yeah
21:40 imirkin: afaik this is a self-inflicted issue
21:40 imirkin: there's no need for the depth buffer to be shared in the first place
21:40 imirkin: (at least i think)
21:40 karolherbst: ahh, makes sense
21:40 imirkin: either way, tagr_ was looking at it, no?
21:40 karolherbst: dunno if he actually was looking into ti
21:41 karolherbst: I am actually wondering why we still have the tegra drm driver in the first place anyway...
21:41 imirkin: ah ok. well if he is, he should be more than capable of resolving the problem
21:41 karolherbst: I guess jusg using kmsro might resolve the issue as well, as it does less magic
21:44 karolherbst: I think I might actually just go ahead and port it over
21:44 karolherbst: we could get rid of the entire tegra gallium driver then