| === JanC_ is now known as JanC | ||
| tseliot | mamarley, ricotz: does the patch for linux 4.6 apply correctly in the 367.35 driver for you? It seems to fail here | 09:43 |
|---|---|---|
| tseliot | mamarley, ricotz: oh, it's commented out, never mind | 09:45 |
| ricotz | tseliot, make sure to make it 4.7 compatible ;) | 12:02 |
| ricotz | which will likely land soon in yakkety | 12:02 |
| tseliot | ricotz: yes, I'm going to upload 367 with support for that. I also need to fix up the legacy drivers | 12:20 |
| tseliot | the legacy drivers both build against 4.7 | 13:05 |
| ricotz | tseliot, iirc it was drm related | 13:13 |
| tseliot | ricotz: yes, it was, for 367. You never know what breaks in the kernel though ;) | 13:14 |
| tseliot | ok, 367 uploaded | 13:14 |
| mamarley | It will be interesting to see if GCC6 busts the kernel module build. | 13:25 |
| * tseliot is not looking forward to that... | 13:28 | |
| ricotz | tseliot, remember when patching the uvm module is needed it breaks on non-amd64 | 13:40 |
| ricotz | tseliot, where did you upload it? | 13:41 |
| ricotz | https://launchpad.net/ubuntu/+source/nvidia-graphics-drivers-367 | 13:41 |
| tseliot | ricotz: 367 needs to be approved by an archive admin first | 13:42 |
| ricotz | ah it is in the new-queue | 13:42 |
| tseliot | it's in yakkety-proposed NEW | 13:42 |
| tseliot | yes | 13:42 |
| ricotz | I hope you used the ppa tarball | 13:43 |
| tseliot | I did | 13:43 |
| tseliot | I always check | 13:43 |
| ricotz | good :) | 13:43 |
| ricotz | tseliot, did you understood the note above? | 13:48 |
| tseliot | ricotz: this one? "patching the uvm module is needed it breaks on non-amd64" | 13:49 |
| ricotz | yes | 13:49 |
| tseliot | I'm on the call, so, I can't really check | 13:49 |
| tseliot | *a call | 13:49 |
| ricotz | +--- a/kernel/nvidia-uvm/uvm_linux.h | 13:49 |
| ricotz | ++++ b/nvidia-uvm/uvm_linux.h | 13:49 |
| ricotz | ^ this wont be available on non-amd64 module builds and therefore make the patch-apply fail | 13:50 |
| tseliot | err... I'll think of a solution | 13:50 |
| ricotz | dkms_nvidia.conf might have a mechanism to check | 13:50 |
| tseliot | if it doesn't I'll use a macro in the code | 13:51 |
| ricotz | turning on PATCH_MATCH will at least prevent problems on <4.7 | 13:51 |
| * ricotz totally forgot about that issue | 13:52 | |
| tseliot | I think there is a check in the code | 13:53 |
| tseliot | #if LINUX_VERSION_CODE < KERNEL_VERSION(4, 7, 0) | 13:53 |
| ricotz | no, to prevent the patch to be applied | 13:54 |
| tseliot | ok, I see what you mean | 13:55 |
| tseliot | ricotz: I have a fix, and I'm going to test/upload it soon | 15:08 |
| tseliot | ok, uploading to a PPA for testing... | 15:29 |
| tjaalton | RAOF: hi, do you remember why mesa ships libGL.so link pointing to mesa/libGL.so? means that it's always there, no matter what ldconfig knows | 22:44 |
| RAOF | tjaalton: I do not recall that, no. | 22:44 |
| RAOF | tjaalton: If I were to guess, it would be to satisfy the OpenGL Linux ABI. | 22:45 |
| tjaalton | right | 22:45 |
| RAOF | Which requires that /usr/lib/libGL.so exist. | 22:45 |
| tjaalton | there's bug 1609110 | 22:45 |
| ubottu | bug 1609110 in mesa (Ubuntu) "Do we still need the mesa or mesa-egl directories in /usr/lib/<arch>?" [Undecided,New] https://launchpad.net/bugs/1609110 | 22:45 |
| tjaalton | wine is unhappy, but i'd say just cripple wine now too | 22:45 |
| RAOF | Why that's hardcoded to point at mesa rather than being an alternative dependent on the master update-alternatives state is not in my memory. | 22:45 |
| tjaalton | well it has to point somewhere :) | 22:46 |
| tjaalton | though it could be via alternates too, right | 22:46 |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!