patch ati-drivers for kernel 4.6

[jerith666]

Motivation for this change

fixes #18202

Things done

• Tested using sandboxing
  (nix.useSandbox on NixOS,
  or option build-use-sandbox in nix.conf
  on non-NixOS)
• Built on platform(s)
  • NixOS
  • OS X
  • Linux
• Tested compilation of all pkgs that depend on this change using nix-shell -p nox --run "nox-review wip"
• Tested execution of all binary files (usually in ./result/bin/)
• Fits CONTRIBUTING.md.

---

this uses the patch from
https://github.com/manjaro/packages-extra/commit/ddae91f2 to account
for torvalds/linux@d4edcf0d and the patch
from https://www.virtualbox.org/ticket/15298 to account for
torvalds/linux@09cbfeaf

[mention-bot]

@jerith666, thanks for your PR! By analyzing the annotation information on this pull request, we identified @edwtjo, @heydojo and @MarcWeber to be potential reviewers

[vcunat]

linux? I'm confident one should use the existing kernel parameter instead. That can be null, but that's for case when a kernel module isn't built and such patches shouldn't be needed.

[jerith666]

That's what I thought too, but when I was running my build locally, kernel was null. That seemed weird to me, although I don't see kernel being passed in explicitly where this is called (at https://github.com/jerith666/nixpkgs/blob/df9b7c1274551bc1ce0b1e71bb35e4cfbad143ae/pkgs/top-level/all-packages.nix#L11276). But I also don't understand what args callPackage is expected to fill in automatically.

At any rate, here is what I get:

$ git diff
diff --git a/pkgs/os-specific/linux/ati-drivers/default.nix b/pkgs/os-specific/linux/ati-drivers/default.nix
index d54b0ab..621d756 100644
--- a/pkgs/os-specific/linux/ati-drivers/default.nix
+++ b/pkgs/os-specific/linux/ati-drivers/default.nix
@@ -1,4 +1,4 @@
-{ stdenv, fetchurl, kernel ? null, linux, which
+{ stdenv, fetchurl, kernel ? null, which
 , xorg, makeWrapper, glibc, patchelf, unzip
 , fontconfig, freetype, mesa # for fgl_glxgears
 , # Whether to build the libraries only (i.e. not the kernel module or
@@ -72,7 +72,7 @@ stdenv.mkDerivation rec {
     ./patches/15.9-mtrr.patch
     ./patches/15.9-preempt.patch
     ./patches/15.9-sep_printf.patch ]
-  ++ optionals ( (builtins.compareVersions linux.version "4.6") >= 0 )
+  ++ optionals ( (builtins.compareVersions kernel.version "4.6") >= 0 )
                [ ./patches/kernel-4.6-get_user_pages.patch
                  ./patches/kernel-4.6-page_cache_release-put_page.patch ];

$ nixos-rebuild build -I nixpkgs=/home/matt/git/nixos/nixpkgs
building Nix...
building the system configuration...
error: value is null while a set was expected, at /home/matt/git/nixos/nixpkgs/pkgs/os-specific/linux/ati-drivers/default.nix:75:44
(use ‘--show-trace’ to show detailed location information)

[vcunat]

I expect you have a 64-bit system with driSupport32Bit = true: there you get 64-bit version build against the kernel and a 32-bit version with libsOnly == true (i.e. without kernel).

[jerith666]

The Travis CI failure was in building the OpenJDK docs, which seems spurious to me.

[jerith666]

@vcunat you're right. Have a look at 7bc91ff instead.

[domenkozar]

Did someone test this? cc @ocharles

[vcunat]

Looks reasonable and builds OK. I assume @jerith666 runs on it.

[vcunat]

It still won't build with 4.7, but that's a relatively new branch.

[jerith666]

@vcunat I do indeed run on this. I guess I should've said that explicitly in the description? I thought @domenkozar was asking for someone else to test it. Anyway, thanks for the merge!

[jerith666]

I haven't looked at 4.7, it probably won't come to the top of my list for a bit yet. :-/

[vcunat]

I see you checked the "tested execution" box; that should be enough IMHO.