Re: Linux kernel config maintainablility

2016-09-11 Thread David Craven
Hi Ludo, >> What difficulties did you encounter with using allmodconfig as a basis? > I forgot the details, but maybe the list archive has some info. Changing what already works doesn't make sense. For arm and new platforms I think we can go down the defconf route, since it's really simple now..

Re: Linux kernel config maintainablility

2016-09-11 Thread Ludovic Courtès
Hi David, David Craven skribis: >> I was hoping we could come up with a way to generate a kernel config, >> either based on ‘allmodconfig’ or something like what NixOS does [0], >> but I’ve become pretty convinced that this would be difficult at best. > > What difficulties did you encounter with

Re: Linux kernel

2016-09-10 Thread Ludovic Courtès
David Craven skribis: >> Nitpick: Use ‘base32’ here, rather than in the callee (that way, base32 >> parsing happens at compile time and errors are detected early.) > > Good to know, I wondered why we always need sha256 and base32 :-) To illustrate: --8<---cut here---star

Re: Linux kernel

2016-09-09 Thread David Craven
> Seems we already discussed some of these, no? Ah yes, I missed this, sorry - No need to comment on this patch again: >> Important note: (%current-target-system) returns a GNU triplet such as >> “arm-linux-gnueabihf”, whereas (%current-system) returns a “system type” >> such as “armhf-linux”. >

Re: Linux kernel

2016-09-09 Thread Ludovic Courtès
David Craven skribis: > I reworked the patches for the linux-libre package and made some other > improvements that make customizing the linux-libre package easier. I > tested that everything builds and that there wasn't any breakage on > x86_64, but testing the arm kernel on real hardware is a wi

Re: Linux kernel

2016-09-08 Thread Leo Famulari
On Fri, Sep 09, 2016 at 12:45:13AM +0200, David Craven wrote: > I reworked the patches for the linux-libre package and made some other > improvements that make customizing the linux-libre package easier. I > tested that everything builds and that there wasn't any breakage on > x86_64, but testing t

Linux kernel

2016-09-08 Thread David Craven
I reworked the patches for the linux-libre package and made some other improvements that make customizing the linux-libre package easier. I tested that everything builds and that there wasn't any breakage on x86_64, but testing the arm kernel on real hardware is a wip. Is it easier if I push a bran

Re: Linux kernel config maintainablility

2016-09-05 Thread David Craven
Hi Ludo, > I was hoping we could come up with a way to generate a kernel config, > either based on ‘allmodconfig’ or something like what NixOS does [0], > but I’ve become pretty convinced that this would be difficult at best. What difficulties did you encounter with using allmodconfig as a basis?

Re: Linux kernel config maintainablility

2016-09-05 Thread Ludovic Courtès
.org/archive/html/guix-devel/2014-06/msg00108.html https://lists.gnu.org/archive/html/guix-devel/2014-09/msg00370.html https://lists.gnu.org/archive/html/guix-devel/2014-07/msg00044.html https://github.com/NixOS/nixpkgs/blob/master/pkgs/os-specific/linux/kernel/linux-4.7.nix https://github.c

Linux kernel config maintainablility

2016-09-02 Thread David Craven
Hi, So I'm trying to get a linux-libre-kernel for arm going. For arm there is a multi_v7_defconfig we can use. I looked at the existing kernel configs we have and can't make heads or tails of them. My question is are any kernel options needed specifically for guixsd? And can we make it more readab