| === himcesjf_ is now known as him-cesjf | ||
| kees | gh | 04:50 |
|---|---|---|
| kees | ignore me :) | 04:50 |
| dupondje | lets try to build drm-tip myself :) | 08:13 |
| === himcesjf_ is now known as him-cesjf | ||
| === chu is now known as yuken | ||
| === yuken is now known as Guest72076 | ||
| === Guest72076 is now known as chu | ||
| dupondje | rsync -a --exclude=dkms.conf --delete spl/ /home/dupondje/kernel/cod/tip/drm-tip/2018-02-23/debian/build/build-generic/spl/ | 10:50 |
| dupondje | rsync: change_dir "/home/dupondje/kernel/cod/tip/drm-tip/2018-02-23//spl" failed: No such file or directory (2) | 10:50 |
| dupondje | hmmm :D | 10:50 |
| dupondje | ah had to clone spl :D | 10:58 |
| dupondje | logic | 10:58 |
| apw | dupondje, or disable zfs in that build | 11:57 |
| apw | oh but you might need a fix to make that work | 11:58 |
| Snicksie | I am trying to compile a linux kernel module that works in kernel 4.4.0-112-generic, but if I compile in 4.4.0-116-generic I don't have the retpoline flag that is seemingly required from that release on (at least according to journalctl it refuses to load the module because there is a flag mismatch (retpoline vs no retpoline)). Machine is ubuntu 16.04.3 LTS. I couldn't find any hint on how to | 11:59 |
| Snicksie | compile with retpoline support. How can I do this? | 11:59 |
| dupondje | apw: well :) got pass that, but hitting another error now: | 12:01 |
| dupondje | LD acpidbg | 12:01 |
| dupondje | ld: warning: cannot find entry symbol _start; defaulting to 00000000004000f0 | 12:01 |
| dupondje | /home/dupondje/kernel/cod/tip/drm-tip/2018-02-23/debian/build/tools-perarch/tools/power/acpi/tools/acpidbg/acpidbg.o: In function `acpi_aml_set_fd': | 12:01 |
| dupondje | ah well :) | 12:01 |
| apw | are the automated drm-tip builds failing too, i assume so | 12:02 |
| apw | Snicksie, you need to have the retpoline compiler installed | 12:02 |
| apw | the latest version of the compiler which has that support | 12:03 |
| snadge | HexChat 2.12.4 Linux 4.14.18-rt15-snadge [x86_64/1.83GHz/SMP] | 12:05 |
| snadge | nobody builds a standard ubuntu kernel config with the latest hard realtime patch :( | 12:06 |
| snadge | what if you had an industrial laser you needed to control | 12:06 |
| snadge | ive tested it with 17.10, mixbus (ardour essentially) with a 128 byte audio buffer, whilst im compiling a kernel with 8 threads.. without buffer overruns :p | 12:07 |
| snadge | just in case you need to mix audio streams at low latency whilst under high system load.. i thought everyone needed to do that | 12:08 |
| snadge | lowlatency just doesn't cut the mustard guys.. you can do better ;) | 12:09 |
| dupondje | apw: http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-tip/2018-02-23/ they seem to be fine | 12:11 |
| apw | snadge, no indeed ... because they are rarely based on a version even close to what we are using | 12:12 |
| apw | dupondje, so that is confusing then | 12:12 |
| snadge | thats true.. but it at least works for me ? | 12:12 |
| snadge | saying yes and no and module to mostly the same options makes it an ubuntu kernel does it not? | 12:13 |
| snadge | aren't a lot of the important patches backported? the date on the kernel is at least new anyway | 12:14 |
| snadge | falktx didn't want to know about it either.. im guessing ubuntu-studio is the same deal? nobody wants to put their hands up for it.. but i literally just took the standard 17.10 kernel config, fed it into old config, set preempt-rt .. and pretty much hit enter for everything else | 12:16 |
| apw | snadge, we have a ton of patches on top indeed, including the bits for meltdown etc, so it all depends what your focus is | 12:16 |
| snadge | the meltdown patches are already in it from upstream | 12:17 |
| apw | well th | 12:17 |
| apw | that is something then | 12:17 |
| snadge | https://wiki.linuxfoundation.org/realtime/start shows latest "dev" and "stable" versions.. which are based on (currently) 4.14.20 (rt17) released 22nd feb.. based on 4.14.20 released 17-Feb-2018 | 12:21 |
| snadge | which is an lts branch or something, which presumably has important backports.. but i understand this is far away from what ubuntu uses for reasons which beyond my understanding | 12:21 |
| snadge | i mean why would you use an lts kernel .. thats way too logical ;) | 12:22 |
| snadge | better to use something that isn't an lts.. then backport everything yourself | 12:22 |
| snadge | im just joking.. i dont pretend to understand what the logic is behind not using the upstream lts kernel | 12:24 |
| dupondje | apw: seems like acpidbg isnt build on dailies, so could explain | 12:29 |
| frickler | apw: is there a document describing how to install retpoline gcc on xenial? I'm having similar issues compiling latest 4.13 it seems | 12:54 |
| dupondje | dpkg-deb: building package 'linux-image-4.16.0-994-generic' in '../linux-image-4.16.0-994-generic_4.16.0-994.201802222100_amd64.deb'. | 13:02 |
| dupondje | allright :D | 13:02 |
| Snicksie | apw: is there currently any compiler that has retpoline patches in ubuntu 16.04? I cannot find any, also not in the ubuntu-toolchain-r ppa | 13:40 |
| snadge | i thought it was in -proposed | 13:58 |
| === zyga_ is now known as zyga | ||
| === mamarley_ is now known as mamarley | ||
| === walac_ is now known as walac | ||
| === logan_ is now known as logan- | ||
| mozmck | Anyone here know if CONFIG_PM can be disabled in a 4.9+ kernel? I'm building a kernel with preempt-rt patches and I can disable some of the power management in menuconfig, but it doesn't let me disable CONFIG_PM. | 18:57 |
| mozmck | The help for CONFIG_PM (Device power management core functionality) says, "Selected by: PM_SLEEP [=y] && (SUSPEND [=n] || HIBERNATE_CALLBACKS [=y])" | 18:59 |
| mozmck | But I can't find PM_SLEEP or HIBERNATE_CALLBACKS | 19:00 |
| === chiluk_ is now known as chiluk | ||
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!