| === JanC is now known as Guest73457 | ||
| === JanC_ is now known as JanC | ||
| tjaalton | where does nvidia dkms determine if a kernel is supported or not (to build)? | 07:13 |
|---|---|---|
| ricotz | tjaalton, maybe https://www.kernel.org/ | 07:51 |
| tjaalton | huh? I mean the dkms build | 07:52 |
| ricotz | I assumed you mean how nvidia decides which kernel is still important | 07:53 |
| tjaalton | it checks the linux-headers package | 07:53 |
| ricotz | it simply tries to build it | 07:53 |
| ricotz | and conditionally applies patches based on the kernel version | 07:53 |
| tjaalton | ERROR (dkms apport): kernel package linux-headers-4.13.0-NNNN is not supported | 07:54 |
| ricotz | I haven't encountered this yet | 07:54 |
| tjaalton | tseliot is off this week, he'd have the answer | 07:54 |
| tjaalton | okay | 07:55 |
| ricotz | never ran into this using the ppa or mainline packages | 07:55 |
| tjaalton | it comes from dkms itself | 07:58 |
| tjaalton | so it refuses to build if the kernel is from a ppa. wonderful | 08:09 |
| ricotz | tjaalton, is this acually the nvidia module failing? "apport" confuses me here | 08:44 |
| tjaalton | yes | 08:44 |
| tjaalton | it's a ppa kernel based on 4.13+drm-next | 08:44 |
| tjaalton | so I'll fix nvidia to build against it and show what I got :P | 08:45 |
| ricotz | hmm, I am using the canonical-kernel-team ppa packages here | 08:45 |
| tjaalton | the apport error just confused me | 08:45 |
| ricotz | might be that they handled special | 08:45 |
| tjaalton | all it means is that it won't file a bug because it's an unofficial kernel, which is correct | 08:45 |
| tjaalton | what I didn't see was the build error | 08:45 |
| ricotz | which nvidia version? | 08:46 |
| tjaalton | the latest | 08:46 |
| ricotz | so 384.90? | 08:46 |
| tjaalton | yep | 08:46 |
| ricotz | if you are changing the drm stack it could be a legit failure though | 08:47 |
| tjaalton | yep | 08:47 |
| tjaalton | I know it is | 08:47 |
| ricotz | nvidia doesnt build on 4.14 | 08:47 |
| ricotz | ok | 08:47 |
| tjaalton | right, I'll fix that | 08:48 |
| tjaalton | actually it's drm-intel-next-queued, which has drom from 4.14 and then some | 08:48 |
| tjaalton | *Drm | 08:48 |
| tjaalton | uh | 08:48 |
| tjaalton | *drm | 08:48 |
| ricotz | I guess you have to figure this out yourself then | 08:48 |
| tjaalton | sure | 08:48 |
| ricotz | there are some hacky 4.14 patches around | 08:49 |
| ricotz | mamarley, hi, btw not a fan that you already copied 384.90 | 08:49 |
| ricotz | I was running it and didn't had issues though | 08:50 |
| tjaalton | at which point are the dkms patches applied? | 09:14 |
| tjaalton | the PATCH lines in dkms.conf are all commented out | 09:24 |
| ricotz | tjaalton, there are no patches required up to 4.13.x | 09:25 |
| tjaalton | ah | 09:25 |
| ricotz | and 4.14 is not supported | 09:25 |
| ricotz | you will have to add one | 09:25 |
| tjaalton | why keep the old crap there to confuse random people? :) | 09:25 |
| ricotz | for reference/ as how to ;) | 09:26 |
| tjaalton | ok then | 09:29 |
| mamarley | ricotz: Sorry, I thought you had already looked at it when I uploaded it without the ARM binary on Friday. | 10:02 |
| ricotz | mamarley, I see, I just confirmed that I read your message | 10:42 |
| tjaalton | wasn't too hard to make nvidia build against 4.13+dinq | 11:30 |
| tjaalton | but now I need to provide something with the kernel to check against, since it's not really 4.13 nor 4.14 | 11:31 |
| tjaalton | or, fix conftest.sh | 11:31 |
| ricotz | tjaalton, what is the goal here? | 11:35 |
| ricotz | why aren't you just using 4.14+ | 11:35 |
| tjaalton | because I canm't | 11:35 |
| tjaalton | -m | 11:36 |
| tjaalton | you'll hear about it in a few weeks | 11:36 |
| ricotz | backporting the drm stack like in the good old days... | 11:37 |
| tjaalton | alright, I have a diff to the source that adds a couple of new tests to conftest.sh, works with funny kernels and doesn't rely on version.h | 16:17 |
| tjaalton | mamarley: this seems to build at least, I have no hw to test https://pastebin.com/YFcVb8ZF | 17:10 |
| mamarley | tjaalton: I think you highlighted the wrong person there; I'm not sure what this is about. | 17:40 |
| === JanC_ is now known as JanC | ||
| tjaalton | mamarley: you don't upload new versions to the ppa? | 18:57 |
| mamarley | tjaalton: I do, but I hadn't previously been attempting to install the driver on 4.14. | 18:58 |
| tjaalton | ah | 18:58 |
| mamarley | I usually hold off until -rc5 or -rc6 or so before installing the RC kernel, since in the past damage has been done to people's filesystems by installing early RCs. | 18:59 |
| tjaalton | i'll have test feedback tomorrow | 19:01 |
| mamarley | Thanks for the patch though! | 19:02 |
| mamarley | It definitely looks more complicated than anything I would have been able to come up with myself. | 19:02 |
| tjaalton | had to do it this way so that it builds with drm-intel-next | 19:03 |
| tjaalton | and also mainlines | 19:03 |
| mamarley | I imagine if it needs that to build against drm-intel-next, it will probably also need it for 4.15 later on down the road. | 19:04 |
| tjaalton | no i mean can't do simple version comparison | 19:05 |
| tjaalton | drm-intel-next is (currently) 4.13 | 19:06 |
| mamarley | Ah, OK, sorry, I misunderstood. | 19:12 |
| ricotz | tjaalton, I don't think this will work with the real 4.14, I expected there to be GPL issues | 20:36 |
| ricotz | tjaalton, what is happening with the legacy drivers in "a few weeks"? | 20:41 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!