00:09 hanetzer: hrm... why xml files =_=
00:10 imirkin: why not?
00:10 hanetzer: verbose af.
00:11 imirkin: i actually think they're a pretty good fit for this
00:11 hanetzer: did those rnndb files get written by hand or are they generated somehow?
00:11 imirkin: by hand
00:11 imirkin: but they're consumed by computers
00:11 hanetzer: yeesh. Oh well, neovim will make it less terrible :)
00:16 hanetzer: honestly I wouldn't even be bothering with this if I knew of a mcu/soc that wasn't ass I could just use as a basis for a new product, even if I never bring it to fruition
00:32 rhyskidd: any takers for reviewing https://patchwork.freedesktop.org/patch/174696/
00:33 rhyskidd: i saw karolherbst sent a related temp series to the ML that contemplated the higher precision Pascal temp sensors
00:36 imirkin: well ultimately you need to capture skeggsb's attention
00:37 rhyskidd: thanks -- hopefully I've addressed the helpful comments I got on the initial RFC patch
00:45 skeggsb: rhyskidd: i'd like to see that nvkm_info() either go away, or turn into nvkm_trace()
02:32 rhyskidd: skeggsb: thanks, sent a v3 to the mailing lists with the requested change
06:51 pq: hanetzer, imirkin, karolherbst, RSpliet, ideally mmiotrace would use instruction emulation, it would make it all kinds of better, but at the time I was working on it, there was no such thing hooked up. Hence: pagefault on MMIO read, restore page, single-step, decode instruction to know which register has the interesting value, re-arm page, continue. On MMIO write similar except maybe it got the value before si
06:52 pq: ngle-stepping, I forget.
11:55 ccaione: hey rhyskidd any comment on https://bugs.freedesktop.org/show_bug.cgi?id=101220#c16 ?
15:13 Lyude: karolherbst: re your question about clocking: I don't believe so/
15:13 Lyude: *?
15:13 Lyude: I've never had any serious reclocking issues afaik
15:14 karolherbst: :(
15:21 karolherbst: Lyude: if you by any chance know _anybody_ who might be willing to give up their kepler GPU and have reclocking issues with a recent kernel, then _please_ let me know :)
15:36 Lyude: karolherbst: I will!
15:37 karolherbst: Lyude: thanks!
17:48 tobijk: airlied: it looks like you comment here came true: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=0c697fafc66830ca7d5dc19123a1d0641deaa1f6
17:48 tobijk: it breaks nouveau with 4.14: https://hastebin.com/dujadezafi.pas
18:34 airlied: tobijk: a later fix feom mlankhorst went in
18:34 airlied: though maybe that didnt fix ir
18:35 airlied: skeggsb: time to test linus :-)
18:35 tobijk: airlied: did the fix get into your fist oul lrequest?
18:35 tobijk: *pull
18:35 tobijk: or is it still pending? if have only tested linus tree for now
18:47 airlied: tobijk: in there
18:49 tobijk: airlied: oh well
19:28 tobijk: airlied: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/drivers/gpu/drm/nouveau/nouveau_fbcon.c#n227 is the faulty line, or better someting is not initialized here: drm->fbcon->helper.fbdev->flags;