07:52Guest63615: mupuf: could you in the future add a v2 tag inside the commit message to tell us what you changed?
08:02karolherbst_tmp1: crap, forgot my irc password…
08:31karolherbst1: mupuf: and I didn't got your second mail...
08:36karolherbst1: mupuf: also, read "Chapter 20: Conditional Compilation" of the linux CodingStyle
08:38karolherbst1: mhh, allthough I am not quite sure if that applies here in fact, I also have to play around with that IS_ENABLED macro a bit
09:33t-ask: Hi, what does "Arch Linux with llvm SI scheduler + dma copy enabling patch" mean? Is this relavant for nvidia cards and if how to enable it?
12:30imirkin: t-ask: that's all for radeonsi
15:41RSpliet: that's a fat memory pipe
16:48mek42_laptop: does nouveau need to be built-in to the kernel instead of a module?
16:51mek42_laptop: ok, because I think making nouveau as built-in made my sas card not show up at boot
16:51mek42_laptop: rebuilding with nouveau modular now
16:53Tom^: will that actually make a difference?
16:53ajax: i'd be surprised, but.
16:54imirkin_: but you've been surprised before :)
16:54Tom^: i mean isnt it just staticly compiled in, and essentially does the same just that your kernel got a few kilobytes bigger. :p
16:54ajax: and that initcalls will run at a different phase of boot
16:54ajax: which _shouldn't_ affect whether any other card appears on the bus
16:55ajax: but, should is not will.
16:57mek42_laptop: well, my system booted before i rebuilt with nouveau built-in
16:57mek42_laptop: and not after
16:58mek42_laptop: and the failure problem was the same as i saw before i made the sas card drivers built-in
16:58mek42_laptop: if my rebuild boots properly i'll ask if/how i should submit a bug report
17:06mek42_laptop: ah, well, nouveau might be off the hook - same issue with rebuilt kernel where i set the nouveau bits back to modules
18:42NanoSector: anyone knows what the channel is for the Radeon kernel driver?
18:43NanoSector: oh just #radeon
18:43NanoSector: how generic
19:00karolherbst: hakzsam: I am on it
19:04karolherbst: imirkin: you should write a bot to automate your replies :p
19:05karolherbst: if somebody asks about how well nouveau supports the gpu -> auto reply: if you want proper oss support, use intel/amd
19:06karolherbst: then you don't have to write it all the time yourself
19:06karolherbst: mupuf: do you have an updated repository somewhere?
19:47karolherbst: pmoreau: I
19:47karolherbst: pmoreau: I
19:47karolherbst: the heck
19:47karolherbst: pmoreau: I've added the led patch to the repositoriy now :)
19:49pmoreau: Having some difficulties speaking, tonight? :-P
19:49karolherbst: actually I also drank some wine tonicht
19:50hakzsam: karolherbst, thanks for closing the ticket
19:50pmoreau: Do you think I could change the brightness of my 960 logo?
19:51karolherbst: try it :p
19:51Calinou: I have LEDs on my GTX 1080
19:51karolherbst: as long as you have the led pwm in the gpio table
19:51Calinou: please support them! for the year of linux desktop! rgb led 4 lyf!
19:51karolherbst: Calinou: pmoreau will prepare a image for you to try out :p
19:51pmoreau: I won’t try that tonight though, I just got back from work, I need to eat and relax a bit.
19:51karolherbst: ohhh wait
19:51Calinou: I'm on blob though
19:51karolherbst: pmoreau: is your led also a rgb one?
19:51Calinou: and blob has support for LEDs only on Windows
19:52pmoreau: karolherbst: Dunno
19:52Calinou: my case has no window though, so I can't see them anyway
19:52Yoshimo: Calinou: have you sent in the vbios for the record? ;)
19:52Calinou: Yoshimo: nope
19:52Calinou: karolherbst said they already have 1080 VBIOSes
19:52karolherbst: pmoreau: maybe it will be fine
19:52Yoshimo: i heard asus offers a special one without power target, did anyone get ahold of that one?
19:52karolherbst: Calinou: yes, and those are stupid those vbios'
19:53karolherbst: Calinou: amateur :p
19:53karolherbst: everybody knows that if you can see the LEDs your rig is 10% faster at least
19:53Yoshimo: not on nouveau though :P
19:54karolherbst: right, there it will be 20%
19:55karolherbst: pmoreau: I am pretty sure the 0x7d and 0x7e GPIOs are for the two other colours
19:55karolherbst: but we will see then
21:41karolherbst: pmoreau: I think I will add support for creating a kernel tree to that script
21:41karolherbst: so anybody can choose what they want
22:40imirkin: who ported bioshock btw?
22:40imirkin: (bioshock infinite)
22:42hakzsam: imirkin, btw, I probably fixed elemental on gm107 (this samplers/textures invalidation bs again)
22:44hakzsam: at least, it doesn't hang the gpu now
22:44imirkin: i've also pushed that iterator fix... i doubt anything other than grid was hitting it, but ... who knows
22:45hakzsam: I saw that
22:50hakzsam: imirkin, does STACK_MISMATCH dmesg fail sound familiar to you?
22:51imirkin: not really
22:51imirkin: where do you see that?
22:51hakzsam: elemental on gm107
22:52hakzsam: and a ton of ILLEGAL_INSTR_ENCODING like on fermi
22:52hakzsam: but it's definitely better :)
22:52hakzsam: I will cook a patch for tomorrow
22:58imirkin: ok, so STACK_MISMATCH is, in the gk20a docs, "API_STACK_ERROR"
22:58imirkin: which i think means that someone messed up l usage somewhere
22:58imirkin: perhasp between the header and what's written somewhere
22:59hakzsam: yeah, that makes sense
22:59NanoSector: anyone knows if karolherbst's kepler reclocking patches will be in linux 4.9?
23:00imirkin: it could also be some issue with the join stack
23:00imirkin: NanoSector: skeggsb does
23:00imirkin: since it's up to him to apply them
23:01NanoSector: it's the difference between me being excited or me being very excited about it :P
23:01NanoSector: 4.9 already has southern island amdgpu support so that's awesome