09:42PaulePanter: Hi. Just an understanding question for an article I’m writing. Are there NVIDIA cards, which Nouveau fully supports without loading proprietary Video BIOS?
09:45skeggsb: PaulePanter: none. it's impossible. we don't execute any of the x86 code in the bios, but there's a whole heap of data tables we use. you can't avoid that, those tables describe how a given vendor customised the board
09:47PaulePanter: skeggsb: Ok, so putting the VGA Option ROM into RAM and not executing it, is possible for old(?) generations?
09:48skeggsb: no, we depend on it being there for every generation
09:48skeggsb: except tegra boards.. they have some stuff hardcoded in device tree and/or code
09:49PaulePanter: skeggsb: Sorry, for not being clear.
09:49PaulePanter: By the way, what is the term NVIDIA uses? Intel uses Video BIOS Tables (VBT).
09:50PaulePanter: skeggsb: With the Video BIOS data, Nouveau can set up a simple frame buffer I guess?
09:50PaulePanter: But to user more features (acceleration, …) other firmware is needed?
10:05imirkin: PaulePanter: not a direct answer to your question, but have a look at this: https://www.phoronix.com/forums/forum/linux-graphics-x-org-drivers/open-source-nvidia-linux-nouveau/998310-nouveau-persevered-in-2017-for-open-source-nvidia-but-2018-could-be-much-better?p=998427#post998427
10:10PaulePanter: imirkin: Thank you. That answered my questions I guess. Very nice write-up.
12:40crmlt: I'm not able to start X with noveau kernel 4.15
12:40crmlt: I stuck on black screen with a trace and mouse cursor
12:41imirkin_: crmlt: which GPU?
12:41imirkin_: and pastebin trace
12:41crmlt: imirkin_: geforce 320M
12:41imirkin_: that's MCP89?
12:42imirkin_: fun. there was a VMM rewrite that landed in 4.15, and MCP89 is slightly different from others, so something could have been missed
12:42crmlt: imirkin_: definitely
12:43crmlt: i get something like nouveau 0000:02:00.0: Xwayland: failed to idle channel
12:44imirkin_: mmm... that's just it saying it's dead
12:44imirkin_: is there anythign before?
12:45crmlt: lots of things
12:45crmlt: Call trace:
12:46crmlt: could this help>?
12:46imirkin_: paste in there
12:46imirkin_: provide a link here.
12:46crmlt: i'm not able to get it
12:47crmlt: there is just blinking cursor and the tty ouput in background
12:47imirkin_: there's like 1000 of them
12:47imirkin_: that's what it should be.
12:47imirkin_: paste into it :)
12:47imirkin_: hit save.
12:47imirkin_: then you're done
12:47imirkin_: it's for when you need to paste in haste
12:48crmlt: imirkin_: i cant acces hastebin.com
12:50imirkin_: weird. loads fine for me.
12:50crmlt: well the server does work but i cant acces it on that machine
12:50imirkin_: you said you saw the blinking cursor. that's all it is paste into it, hit save (top right, or hit ^S)
12:51crmlt: well i ment I get tty ouput and blinking mouse cursor on that machine when I m trying to boot
12:51crmlt: and run X
12:52imirkin_: ah :)
12:52imirkin_: well, figure out a way to xfer the logs to a diff machine
12:52crmlt: btw what is VMM
12:57Moiman: probably Virtual Memory Manager
12:59crmlt: where I can acces log from previous boot?
13:05crmlt: I think I found it
13:08crmlt: line 129
13:10crmlt: What do you think?
13:24imirkin_: nouveau 0000:02:00.0: imem: PRAMIN exhausted
13:24imirkin_: oh goodie.
13:24imirkin_: that was in the "working" kernels
13:25imirkin_: skeggsb: BUG in nvkm_object_unmap
13:27imirkin_: pmoreau: your auto-builder's dead :( but you appear to already know that.
13:38mooch: ayyyy it worsk
13:39mooch: thanks imirkin_
13:41crmlt: imirkin_: How can avoid this issue for now?
13:42crmlt: imirkin_: I can reach GDM with 4.13 but not GNOME
13:42crmlt: but 4.13 works with ubuntu 17.10
13:47imirkin_: did you mean 4.15 on that first one?
13:47crmlt: no I cant rech GDM on 4.15
13:48imirkin_: skeggsb: you have a left-over "gp10b." file in subdev/mmu
13:48crmlt: but when I try boot with 4.13 I can reach GDM
13:49crmlt: on ubuntu 18.04
13:49imirkin_: nvaf is hooked up to g84_mmu_new instead of mcp77_mmu_new
13:50imirkin_: that definitely *sounds* odd
13:50imirkin_: but my recollection of how mcp89 works is ... incomplete.
13:51crmlt: anything else I can provide?
13:52imirkin_: a patch to fix the issue would be nice ;)