03:06rhyskidd: skeggs: is there somewhere you can point me around INIT_GENERIC_CONDITION_ID_NO_PANEL_SEQ_DELAYS ?
03:06rhyskidd: skeggsb: --^
03:10skeggsb: rhyskidd: what do you mean?
03:11rhyskidd: oh, it's a new devinit script opcode
03:11rhyskidd: i'd seen 0x7 previously, and at least tried to account for this in envytools
03:11rhyskidd: but i don't think we'd previously had an idea of what exactly 0x7 did
03:12skeggsb: i don't think there's any public info to point you to about it
03:12rhyskidd: okay, no worries then
03:12skeggsb: it relates to some features we don't use (nor does the linux proprietary driver)
03:13rhyskidd: well that makes it very easy then :)
19:38hololeap: i'm trying out sway with wayland, and i'm using firefox through xwayland. sometimes, when i switch tabs, i have an issue where the screen flashes between the old tab and the new one very quickly
19:40hololeap: also, sometimes kitty, which is a terminal which can run on native wayland, will fail to draw part of it and will show the background in part
19:42hololeap: if i switch workspaces and then switch back, it always fixes it temporarily
19:42hololeap: i've also noticed this happening more under system load
19:43hololeap: i feel like it has to do with the nouveau framebuffer, but this is just a hunch. i don't know much about this stuff
19:46hololeap: there was some advice online for proprietary nvidia drivers, which said to enable Force full screen redraws on repaint
20:03imirkin: hololeap: sounds like some lack of synchronization between drawing and submitting a fb for display
20:08hololeap: where in the stack would be likely culprit?
20:09hololeap: maybe it's the interaction of firefox's hardware acceleration with xwayland... it would be nice to test this on a different gpu vendor, but this is all i've got
20:09imirkin: i don't know too much about that stack
20:09hololeap: ok, well thanks anyway :)
20:10imirkin: you might be able to get advice from the sway / wayland folk
20:10imirkin: actually re-reading your description, sounds like it's more related to glamor than anything else
20:10hololeap: yeah, nobody on their irc channels has responded to me yet, but i'll keep testing stuff
20:11imirkin: firefox is unaware of wayland -- it's just X
20:11imirkin: and it uses various X api's to draw things
20:11imirkin: and it's just inside firefox that you're getting corruption
20:12imirkin: so it stands to reason that it's somewhere in the firefox <-> X thing that it goes wrong
20:12imirkin: Xwayland uses glamor to provide acceleration
20:12hololeap: i could try a 3d game or something similar in xwayland and see if i get similar issues
20:14hololeap: imirkin: i have another problem that i should bring up too. if sway crashes on me and i'm using the framebuffer console, the framebuffer is completely unusable until i reboot. i had to disable FRAMEBUFFER_CONSOLE in the kernel config
20:15hololeap: do you know any workarounds for that kind of issue?
20:15imirkin: try flipping between vt's?
20:15imirkin: i've run into similar issues too, actually, when debugging some wayland-specific issues
20:15imirkin: i don't use wayland on a regular basis though
20:15hololeap: it doesn't work. nothing works except Alt+SysRq+REISUB
20:15imirkin: what if you ssh in
20:15imirkin: and run 'chvt 2' and then 'chvt 1' or whatever?
20:16hololeap: yeah, the system is responseive. i didn't know what to try, though. i wasn't aware of the chvt command
20:16imirkin: well, if you can't ssh in
20:16imirkin: then that's much worse
20:17hololeap: it's not a kernel panic, i know that much for sure. music playing from an online stream will keep playing
20:18imirkin: can you ssh in?
20:18hololeap: i honestly haven't tried it, but i would suspect i could. i just disabled the framebuffer terminal and called it good
20:19hololeap: actually, no, i did ssh in a couple times. i just remembered
20:20hololeap: to see the debug output from sway. basically if sway tried to start but encountered an error, it would make the local screen/keyboard unusable (except for alt+sysrq) until i rebooted
20:22hololeap: oddly, if i exited sway normally, it would bring me back to the vt just fine
20:23hololeap: and these weren't gpu-related errors from sway either. it was looking for a configuration file
23:41RSpliet: imirkin: RE setting of RON pull value. I don't think GDDR3 is relevant beyond some very rare NVA3, and a few prior cards for which memory reclocking code is a tad wonky anyway. Your resolution sounds good to me
23:41RSpliet: Also, "however I don't have the datasheet available for the H/W" gave me a little giggle. Neither do we :-P