00:13netz: imirkin: looks like your script still finds some data on 367.27
00:44imirkin: netz: just not the ctxsw fw
00:45netz: if you knew what it looked like, would you be able to grab it?
00:45imirkin: and not just coz those are hard-coded... the archives are no longer there
00:45imirkin: sure, but they changed the way it's uploaded and we no longer see it in the mmiotrace
00:45imirkin: although for GM107 it's not necessary, the nouveau-supplied ctxsw fw should be fine
00:45imirkin: just for GM20x+
00:46netz: major cuntery.
00:47imirkin: well, like i said, i doubt we're more than a blip on their radar
00:47imirkin: so it's probably not to make it harder for us to trace
00:56netz: I don't trust 'em. honestly I don't trust any of the major vendors anymore. too much fuckery.
01:32netz: if they do start releasing those signed blobs, can nouveau progress?
01:41imirkin: problem is that if they release those blobs like 2 years after the fact ...
08:13yoshimo: does this mean Nouveau is looking in the wrong folder to get the firmware? https://pastee.org/ykc4r
08:13yoshimo: the difference i see is the additional "gr" subfolder
08:59hakzsam: netz, imirkin yeah, tess is complicated, I looked like one or two days, did some progress but I'm working on something else currently, I will have a second look :)
09:05hakzsam: imirkin, btw, can you close the card for ms images on trello?
11:43imirkin: Yoshimo: hm, that *is* surprising
11:45Yoshimo: it is weird because the path seems to be hardcoded at least in skeggsbs github repo
11:52Yoshimo: it happens both on a kubuntu 16.04 with oibafs graphic ppa added for the current mesa stuff which i could argue is broken due to experiments and the current yakkety daily image for 16.10 which is alpha but pristine
12:08imirkin: Yoshimo: yeah, that's totally buggered
12:08imirkin: Yoshimo: change these to gr/* : https://github.com/skeggsb/nouveau/blob/master/drm/nouveau/nvkm/engine/gr/gf100.c#L1814
12:09imirkin: it happened as part of the nvkm_firmware refactor
12:09imirkin: coz we used to look in two places, now we look in just on
12:17Yoshimo: shouldn't this refactoring already be upstreamed and part of YY?
12:47imirkin: skeggsb_: --^
17:02RSpliet: disdi: don't ask to ask, just ask\
17:12karolherbst: by the way: would somebody mind going over my reclocking patches (again)? I hoped I would have found more time to spend some time on them, but as it seems, I really need one day of at the weekend and the other day I usually do other things :/