00:21 gnurou: skeggsb: airlied: would it be acceptable for you to take the GM20B (Tegra X1) clock driver disabled for the 4.6 merge window, then a very small patch that enables it during the -rc cycle, after the last device tree details are ironed out?
04:01 RSpliet: https://benchlife.info/gp104-aka-nvidia-geforce-gtx-1080-will-ship-in-may-and-no-hbm2-031112016/
04:47 mupuf: RSpliet: cool! Did not know you spoke chineese :D
04:49 RSpliet: the "no HBM2" is a bit of a bummer though
04:50 mupuf: RSpliet: yop
04:50 mupuf: I wonder when we will get the fw for this though :D
04:50 RSpliet: I'd still keep my fingers crossed they'd rather release a GTX 1060 - and a GTX1080 with HBM2 months after (rather than 1 year+ after)
04:50 mupuf: yeah, possibly
04:51 mupuf: they likely have a bad yield
04:51 RSpliet: has more to do with the availability of HBM2
04:51 mupuf: so, as soon as the yield increases, they will start selling them
04:51 mupuf: sure? where did you read this?
04:51 RSpliet: nowhere, but it's new, and AMD has dibs :-P
04:52 karolherbst: from the article I got the feeling, that the memory stuff will be different :/
04:52 karolherbst: awesome, the hw isn't released yet and my first though was, mhhh maybe memory reclocking still works
04:52 karolherbst: :D
05:38 karolherbst: is the nouveau drm pull sent out by the way?
05:39 karolherbst: or do we have still time? :D
05:42 RSpliet: what benefit would your work have to the user?
05:43 karolherbst: reclocking on maxwell?
05:44 karolherbst: well first gen maxwell
05:48 RSpliet: I'd assume that's going to be postponed for a little bit as it might require more extensive testing
05:51 karolherbst: yeah make it makes sense to test it before in a controlled way than to push it out for everybody
06:10 gnurou: skeggsb: airlied: actually please forget that last question - the clk driver is not related to the device tree issue, so we can just merge it if Ben is ok with it
06:35 karolherbst: gnurou: got my question regarding the nforce ethernet carD?
06:36 gnurou: skeggsb: ok, sent you the gm20b clk, it is less scary than it looks - mostly refactoring to reuse the GK20A code
06:36 gnurou: karolherbst: errr yes, that's the thing I forgot to look at today >_<
06:40 gnurou: karolherbst: that's some pretty old stuff, will take some effort to reach the right people for that
06:41 gnurou: I thought this chipset was fully supported under Linux?
06:41 karolherbst: gnurou: yeah I figured
06:41 karolherbst: gnurou: well it is supported
06:41 karolherbst: gnurou: but not the phy wake on lan mode
06:41 karolherbst: gnurou: I tried to RE it a bit, but it still wakes up on multicast packages
06:41 karolherbst: and it is better to implement this feature according to docs, because I have no big clue what I am doing there :D
06:42 karolherbst: gnurou: my two patches by the way: https://github.com/karolherbst/linux/commits/nforce_wol
06:43 gnurou: karolherbst: ok - I will see if I can find some docs on Monday (not at the office now)
06:47 gnurou: karolherbst: looking at your code, it seems like you already have wol running? and just need confirmation your implementation is correct?
06:47 karolherbst: gnurou: thanks a lot
06:47 karolherbst: gnurou: well phy should only wake the machine up at direct packages
06:47 karolherbst: gnurou: but the machine wakes up also on multicast
06:47 karolherbst: and stuff
06:47 gnurou: ah, I see
06:47 karolherbst: gnurou: and envytools has support for ethernet cards now too :)
06:48 bozhan: hi windows has application list for application which have to work with optimus. is there such thing for linux or does someone work in this direction... maybe on installing packages has to set somewhere that this package require 3D?
06:48 gnurou: please give me some time though, as I will probably need to obtain acks from the right people to release any information
06:48 karolherbst: gnurou: no problems
06:48 gnurou: karolherbst: I really need to have a close look at what envytools can do...
06:59 RSpliet: gnurou: the only thing it can't do is your dishes
07:49 karolherbst: gnurou: well with nvalist you can list supported devices and with nvap{eek,oke} -c $card_number you can modify the mmio stuff
07:49 karolherbst: and a lot of other iommu centric tools
07:50 karolherbst: gnurou: nvascan is good to check what can be set on the reg, but it's not perfect
07:52 karolherbst: maybe we should move all the non gpu centric stuff out into a seperated project
09:49 karolherbst: imirkin: the serious engine is open sourced, now we could look into the talos engine code and search for the issues :D
09:50 karolherbst: ohh wait, it's only an older version
09:50 karolherbst: meh, still good
12:37 pmoreau: karolherbst: Where did the pstate file went? I forgot…
12:39 karolherbst: debugfs
12:40 pmoreau: Where so in debugfs?
12:41 karolherbst: dri
12:43 pmoreau: But that's only in 4.5+, right?
12:43 karolherbst: yes
12:43 pmoreau: And I'm missing the pstate flag on 4.4… That's why I can't see it… --"
13:20 Yoshimo: do these mesa error come from the driver or is the game misbehaving? https://pastee.org/zg5rh
13:53 Maws: Should a integrated "sky lake" graphics be able to display 2560x1600 via HDMI under Linux ? (Debian) - at least the spes of MB says Supports HDMI with max. resolution 4096 x 2160 @ 24 Hz / 2560 x 1600 @ 60 Hz, but only get up 1920x1200
13:54 Maws: it's on a clean Debian testing installation
13:54 robclark: Maws, #intel-gfx
13:54 Maws: robclark, is that Linux related as well or ?
13:55 robclark: yes
13:56 Maws: ok, btw, isn't it the Nouveau driver that's being used on these intel graphics?
13:56 robclark: umm, no
13:56 Maws: hehe, ok.....:)
13:56 Maws: thnx
13:56 robclark: nouveau is for nvidia gpu's
13:57 Maws: I see, I wasn't aware that it was only nvidia (I'v been using it on another pc with nvidia graphics.......but just set up a new computer...:) )
13:59 Maws: hmmm, can't seem to be able to "send" to the intel-gfx channel for some reason - any idea ?
14:20 RSpliet: Maws: probably you should first register your nick with nickserv before you can chat there
14:21 Maws: RSpliet, Where/how do I register ? :)
14:25 RSpliet: nickserv
14:25 RSpliet: that stuff must be documented somewhere on the freenode webpages
14:26 Maws: ok, will check
14:44 karolherbst: Yoshimo: did I give you the maxwell memory reclocking stuff?
14:59 karolherbst: Yoshimo: ohh wait, you had 2gen maxwell right? ... I should keep a lit of stuff I give to people
14:59 karolherbst: jayhost`: you were the one with maxwell1 !
19:05 jayhost`: karolherbst yes i was going to say something then saw your message
19:06 jayhost`: What's up
19:09 jayhost`: I imagine my name has a grave accent for being away from keyboard
23:16 Yoshimo: Mesa: User error: GL_INVALID_OPERATION in glTexImage3D(bad target for texture), i got quite a few of these when i play games, where would i go to to get these silenced?
23:20 orbea: Yoshimo: tell the developers of the games
23:21 orbea: you can get a backtrace on them if you use "break output_if_debug" in gdb
23:24 Yoshimo: as a non-expert it could have been anything between the bare metal and the displaydriver, so i had to ask