03:55 rhyskidd: imirkin: how is the long weekend of nouveau going?
04:07 imirkin: :)
04:07 imirkin: i sent some patches
04:07 imirkin: but overall i've had more productive long weekends
04:07 imirkin: i passed out at like 9pm on sat
04:17 pabs3: any thoughts on this line? kernel: nouveau 0000:01:00.0: fifo: fault 00 [READ] at 000000ffad9a1000 engine 07 [HOST0] client 07 [HUB/HOST_CPU] reason 40 [] on channel 11 [007f534000 systemd-logind[886]]
04:18 pabs3: after I rebooted following this, one of the apps I started after reboot showed a pixels from the game that I was using just before the display freeze and that fault
04:18 pabs3: (this was 0ad and GNOME Wayland)
04:18 imirkin: you're the guy with the weird setup right?
04:19 imirkin: arm64 or something?
04:19 imirkin: with a pcie controller whose io aperture is too small
04:19 pabs3: no, just an old x86 :)
04:20 pabs3: NVIDIA Corporation GK107 [GeForce GT 740] (rev a1)
04:20 imirkin: oh ok. sorry. everyone just gets jumbled up in my head.
04:20 pabs3: np :)
04:21 imirkin: this is an address at the top of the 40-bit space...
04:21 imirkin: unfortunately i don't know how our VM is set up
04:22 imirkin: it _feels_ like it's saying that it received a read from the CPU
04:22 imirkin: and this read didn't go well
04:23 imirkin: as always, skeggsb is the one who can really explain it
04:41 rhyskidd: i've been invited onto a reverse engineering podcast to talk about nouveau
04:42 rhyskidd: anything in particular that ppl would like me to mention, beyond the typical coverage that is given at xdc each year?
04:42 rhyskidd: (obviously, i'm speaking as an individual contributor -- not speaking for the community/project etc etc)
04:43 rhyskidd: audience will be a mixture of non-gpu ppl, but likely a fairly high hardware knowledge
04:56 imirkin: rhyskidd: think about what your goal is, wrt the listeners
04:56 imirkin: educate, recruit, etc
04:56 imirkin: and try to structure your talk accordingly
05:02 pabs3:suggests the 'do not buy nVidia' line feature prominently :)
05:02 imirkin: well - presumably this is more about RE than end-user recommendations
09:38 karolherbst: rhyskidd: the question is rather is that podcast credible and more or less well known. Anyway, the biggest issues right now we have is that's not easy to have enough developers overall for a project that size. Sure the technical aspects are more or less important, but those are our smaller issues.
09:39 karolherbst: pabs3: and I wouldn't
12:16 rhyskidd: thanks. yes, i agree this is to a technical audience rather than end-users, so the recommendation line isn't relevant
12:17 rhyskidd: i've done some background listening of prior episodes, they appear credible around RE matters
14:38 karolherbst: rhyskidd: well, if there is a tiny chance people are getting interested, it is a good thing to do
14:38 karolherbst: maybe mentioning also the secboot situation
22:51 imirkin: skeggsb: that vmm_put thing definitely happens a lot. in case this makes it any easier to repro, i'm using Xorg + wmaker + that's it, no funny business
22:51 imirkin: this is on a GTX 960
22:51 imirkin: no weird software
22:52 imirkin: aterm (xterm clone from the days of yore), hexchat, chromium, emacs
22:52 imirkin: xf86-video-nouveau.
23:55 karolherbst: imirkin: from my experience it more or less happens randomly, but sometimes I don't get it at all :/
23:57 imirkin: agreed on the random part
23:59 imirkin: wow. it's amazing how much crisper the new monitor is =/