02:35 imirkin: skeggsb: what does NOUVEAU_BO_MAP flag to nouveau_bo_new do? does it make the return bo have a ->map pointer?
02:37 skeggsb: it *used* to tell the kernel you wanted the bo allocated in mappable vram (you can't map certain areas pre-nv50)
02:37 skeggsb: we just make ttm move the buffer these days
02:37 skeggsb: (ie. it's completely ignored currently)
02:38 skeggsb: however, if you ever want to be able to map the buffer, you should still set it, in case the kernel decides to care one day
02:38 imirkin: skeggsb: ok, so it's not like libdrm gives you a pre-mapped pointer?
02:38 skeggsb: nope
02:38 imirkin: and this is very explicitly on a GART texture...
02:39 skeggsb: it's a "want a buffer that *can* be mapped at some point" flag
02:39 imirkin: gotcha
07:48 guest388459: hi
07:48 guest388459: i am using nouveau on kernel 4.6 on a gm200 chip with the latest signed firmware, so i have hw acceleration
07:49 guest388459: the problem i have is the cards fans dont spin down the way they should after booting
07:49 guest388459: they seem to be stuck at some fixed duty cycle
07:52 guest388459: and i don't seem to have any fan speeds reported by nouveau device
07:52 guest388459: so i cant regulate it manually either. any advice?
07:57 guest388459: brb
08:11 guest_993851: still have the issue :(
11:25 mvaenskae: alright, got the kernel crashed again (or rather X, acpi event shutdown still succeeded)
12:06 karolherbst_work: mvaenskae: what are you doing to cause that?
12:22 mvaenskae: karolherbst_work: using nouveau on a 320M ;)
12:25 karolherbst_work: uhh
13:17 Arthur_D2: hello, I'm having issues getting HDMI out to work on an old laptop. I am using Xubuntu 16.04, and I can see the screen and activate/deactivate in the control panel. However there's no signal to the TV, but I have gotten it to work with a previous Xubuntu version. There's a key combo to enable/disable HDMI (Fn+f4), however using it now I don't seem to get any output. Any help appreciated!
13:32 Arthur_D2: hmm, acpi_listen does not produce any output when I try the Fn+f* keys
13:33 Arthur_D2: so maybe it's not an ACPI function
13:55 Arthur_D2: hi Akien, fancy seeing you here ;)
13:58 Akien: Hi Arthur_D2 ;) Well I'm in half of the floss channels here :p
13:58 Arthur_D2: hehe
13:59 Akien: Recently dumped bumblebee+nvidia for nouveau+PRIME, so I hang out here to follow the development a bit
14:02 Arthur_D2: ah I see
14:02 Arthur_D2: I'm messing around with an old laptop, it's giving me trouble
14:03 Arthur_D2: does not give any output over HDMI even though the screen is configured
14:03 Arthur_D2: brb, reboot gonna test something
14:07 Arthur_D2: hm did not work. Dinner time tho
14:32 imirkin: mvaenskae: with only X accel, no GL?
14:56 guest_94922: so after more research, i figure nouveau doesn't support fan control/auto throttling on gm200 chips (yet) even with the new signed blobs
14:56 guest_94922: ?
14:56 guest_94922: it must be due to nvidias stupid zero rpm at idle bios. it doesnt behave like any of my previous cards where its thermals regulate its fans after gpu post
14:57 guest_94922: it needs os level firmware to run that functionality...right?
15:08 imirkin_: guest_94922: you have purchased a super-locked down gpu. nvidia needs to sign firmware in order for the fan to be controlled.
15:32 guest_94922: imirkin_: yeah but i thought they had released that firmware...thats why i bought it
15:34 imirkin_: guest_94922: ah, no - they released firmware for controlling graphics context switches
15:34 imirkin_: guest_94922: but even if they had released something, it's still anti-opensource super-locked-down hardware
15:34 guest_94922: i see. so as of today there's no 980ti out there that wont suffer from fan speed problems?
15:34 imirkin_: guest_94922: i'd highly recommend getting an AMD gpu.
15:35 imirkin_: guest_94922: no GM20x (which all GTX 950+ gpu's are)
15:35 guest_94922: well, i had one before this card and it needed a non free blob too
15:36 imirkin_: non-free blob is one thing. having a situation where (a) you have to have signed firmware and (b) the company with the key has no interest in releasing such redistributable firmware
15:36 imirkin_: is a recipe for "switch to different hardware"
15:37 guest_94922: i cant really argue with that
15:37 guest_94922: i hate nvidia
15:37 imirkin_: then stop giving them your money?
15:37 guest_94922: well, amd aren't competitive in 3d performance...
15:37 imirkin_: they're pretty competitive against nvidia on open-source drivers...
15:38 guest_94922: mh
15:38 imirkin_: and since those are the only drivers i'm willing to use ... well, you see how this thought ends.
15:38 guest_94922: yea...
15:49 guest_94922: thanks, i'll shop around
17:18 imirkin_: skeggsb: hm, my records show that glsl-vs-point-size also fails on nv4x, so it's probably not just a nv30 issue
17:22 imirkin_: does anyone happen to have a nv4x plugged in right now?
17:22 imirkin_: and would be willing to run a piglit test for me
18:16 RSpliet: imirkin_: afraid not. I'm a little low in time for the next few days and I can't just resurrect my AGP machine like that
18:17 RSpliet: ping me at the end of next week if you still need some help
18:17 imirkin_: RSpliet: no worries ... i wasn't expecting anyone to perform work to do it
18:17 imirkin_: just if you happened to be sitting at the console of one ;)