Commit graph

104 commits

Author SHA1 Message Date
Matthieu Coudron b9a4e6953d kernel: fix config generation
Addresses https://github.com/NixOS/nixpkgs/issues/71803:
Kernel options are not merged as described, especially the "optional"
aspects. The error silences legitimate warnings.
2020-04-01 22:25:57 +02:00
Matthieu Coudron 2207378c9e linux: clearer origin for settings configuration
adds _file so that nix may have a chance to display what file the conflictings settings come
from.
2019-09-04 17:04:16 +09:00
volth c814d72b51 treewide: name -> pname 2019-08-17 10:54:38 +00:00
volth f3282c8d1e treewide: remove unused variables (#63177)
* treewide: remove unused variables

* making ofborg happy
2019-06-16 19:59:05 +00:00
Joachim F d7da5e2af2
Merge pull request #53826 from delroth/randstruct-custom-seed
nixos: allow customizing the kernel RANDSTRUCT seed
2019-04-16 17:49:19 +00:00
Matthieu Coudron a5a796222b linux: (re)take into account extraConfig
PR #42838 wrongly started to ignore extraConfig. This fixes that.
2019-01-31 13:34:42 +09:00
Matthieu Coudron 3bb7b3f02e linux: ability to merge structured configs
This should make the composability of kernel configurations more straigthforward.

- now distinguish freeform options from tristate ones
- will look for a structured config in kernelPatches too
one can now access the structuredConfig from a kernel via linux_test.configfile.structuredConfig
in order to reinject it into another kernel, no need to rewrite the config from scratch

The following merge strategies are used in case of conflict:
-- freeform items must be equal or they conflict (mergeEqualOption)
-- for tristate (y/m/n) entries, I use the mergeAnswer strategy which takes the best available value, "best" being defined by the user (by default "y" > "m" > "n", e.g. if one entry is both marked "y" and "n", "y" wins)
-- if one item is both marked optional/mandatory, mandatory wins (mergeFalseByDefault)
2019-01-28 09:06:33 +09:00
Joachim F f6414428ed
Merge pull request #53511 from joachifm/kernel-32bit-emu-feature-flag
linux: flag to indicate 32bit emulation support
2019-01-25 14:13:45 +00:00
Pierre Bourdon a258015d3a
kernel: extend the RANDSTRUCT seed with a user-configurable section 2019-01-24 01:42:16 +01:00
Joachim Fasting 6eea9ac868
linux: add feature flag to indicate support for 32bit emulation
Motivated by the need to warn users trying to build configurations that depend
on being able to run 32bit apps on 64bit kernels.
2019-01-06 19:52:03 +01:00
Pierre Bourdon c789f642f0
kernel/generic.nix: provide required dependencies for GCC plugins builds 2019-01-04 22:24:50 +01:00
John Ericson 0828e2d8c3 treewide: Remove usage of remaining redundant platform compatability stuff
Want to get this out of here for 18.09, so it can be deprecated
thereafter.
2018-08-30 17:20:32 -04:00
Tuomas Tynkkynen e0ce4bee57 kernel: Remove empty preUnpack 2018-07-24 23:36:05 +03:00
volth dda95bae35 [bot] treewide: remove unused 'args@' in lambdas 2018-07-20 19:54:05 +00:00
Matthieu Coudron e4fe7b31be linux: fix features system 2018-06-30 16:01:41 +03:00
Dan Peebles ff9999ad1b linux: translate config to structured config
Instead of using a string to describe kernel config, use a nix
attribute set, then converted to a string.
- allows to override the config, aka convert 'yes' into 'modules' or
vice-versa
- while for now merging different configs is still crude (last spec wins),
at least there should be only one CONFIG_XYZ value compared to the current string
config where the first defined would be used and others ignored.

[initial idea by copumpkin in 2016, a major rebase to 2018 by teto]
2018-06-30 16:01:41 +03:00
Tuomas Tynkkynen 87a68c455e kernel: Allow kernel packages to specify a custom default defconfig 2018-05-31 18:06:09 +03:00
Dominique Martinet b64f407c45 linux kernel generic: use passAsFile for kernelConfig
Otherwise get the error 'Argument list too long' when running builder
with a very long kernelConfig
2018-04-09 18:23:52 +09:00
Matthieu Coudron 3ccc65a36c buildLinux: allow to override arch/modules/builtin
- Easy override of autoModules and preferBuiltin and kernelArch parameters (currently living in `hostSystem` set).
2018-03-08 12:47:06 +09:00
Matthieu Coudron 600e977898 buildLinux: removes unnecessary parameters
and passes parameters in a single set
2018-03-08 12:45:39 +09:00
Vladimír Čunát 7d19d7ee18
Revert "linux-*: build with gcc7, but allow overriding it"
This reverts commit ae040525d8.
gcc7 is the default now.
2018-02-17 16:30:26 +01:00
Tuomas Tynkkynen 64fa0d5b97 kernel: Add deps for building 4.16 2018-02-16 01:13:30 +02:00
Vladimír Čunát ae040525d8
linux-*: build with gcc7, but allow overriding it
I expect we will revert this after general upgrade to gcc7.
See https://github.com/NixOS/nixpkgs/issues/34383
2018-02-11 18:55:18 +01:00
Tuomas Tynkkynen 4c6c919a31 Merge remote-tracking branch 'upstream/master' into HEAD
Conflicts:
	pkgs/os-specific/linux/kernel/generic.nix
2018-02-07 03:41:24 +02:00
Tuomas Tynkkynen 685ceb999e kernel: Fix missing config options from platform
common-config.nix has:

${kernelPlatform.kernelExtraConfig or ""}

and indeed kernelExtraConfig is in hostPlatform.platform, and not in
hostPlatform. (Ugh.).
2018-02-06 22:13:57 +02:00
Matthieu Coudron f620b1b693 kernel: buildLinux replaces import ./generic.nix
- defined buildLinux as generic.nix instead of manual-config.nix. This
makes kernel derivations a tad more similar to your typical derivations.
- moved $buildRoot to within the source folder, this way it doesn't have to be created before the unpackPhase
and make it easier to work on kernel source without running the unpackPhase
2018-02-07 10:07:13 +09:00
Ben Wolsieffer 7bb3a044be kernel: fix dependencies 2018-01-15 22:28:45 -05:00
Ben Wolsieffer 285181a1db kernel: remove leftover remnants of old cross compiling system 2018-01-15 12:48:51 -05:00
Ben Wolsieffer a09709bf2a kernel: don't force ignoreConfigErrors when cross compiling, only set the
default
2018-01-15 12:39:45 -05:00
Ben Gamari 7a9b6ac39a kernel: Enable cross compiling 2018-01-15 11:55:06 -05:00
Jan Malakhovski 85cd22cf7d treewide: replace addPassthru 2018-01-14 19:53:57 +00:00
Jan Malakhovski 8606dd8556 lib: change the order of arguments of addPassthru 2018-01-14 13:24:30 +00:00
Vladimír Čunát 13797ff522
linux-4.13: mark as insecure (+required generic changes)
extraMeta was being fed as passthru without being processed by stdenv,
so without those changes, adding the security attribute would be useless.
2017-12-04 09:33:39 +01:00
Tuomas Tynkkynen 0d9f2f0bb4 platforms.nix: Clean up more 'uboot' legacy
For a while now, the only thing the 'uboot' attribute does is to tell
whether to add ubootTools to kernel/initrd builds. That can be
determined with platform.kernelTarget == "uImage" just as well.
2017-11-05 17:06:59 +02:00
Jan Malakhovski 8525b78d05
linuxPackages: properly propagate features and extraConfig
`kernel.override { features = ... }` didn't work before, now it works as
expected.
2017-09-17 16:33:47 +02:00
Vladimír Čunát 7c7c83e233
buildLinux: allow overriding stdenv on each call 2017-08-20 08:24:52 +02:00
hsloan 459d07d41c kernel generic: Don't use stdenv.cross 2017-06-28 20:22:59 -04:00
Nikolay Amiantov c0e77dba0e linux: add kernelPreferBuiltin platform option
This allows to use kernelAutoModules but still compile in any options that are set so in template config.
It's helpful for ARM and maybe other platforms where defaul configurations are useful because they compile in
modules that we and udev cannot autodetect now.
2017-04-09 22:46:07 +03:00
Eelco Dolstra 72a30ae66f linux: Use $SOURCE_DATE_EPOCH as the build timestamp 2016-02-01 18:19:23 +01:00
Dan Peebles 8f9aea9ccc grsecurity: fix kernel config and uncomment grsecurity kernels 2016-01-23 16:58:44 +00:00
Vladimír Čunát 6671aff83e linux kernel determinism: unify timestamp style
Testing showed the linux build is sensitive to /usr/include/ncursesw
unless chrooted (on non-nixos).
On a single chrooted nixos machine, -A linux is binary reproducible.

CC #2281 & @alexanderkjeldaas.
2015-01-03 13:54:32 +01:00
Vladimír Čunát d8c5d95330 determinism: change some fixed timestamp to != (time_t)0
vcunat removed the unrelated glib change.
Conflicts:
	pkgs/development/libraries/glib/default.nix
	pkgs/os-specific/linux/kernel/generic.nix
	pkgs/os-specific/linux/kernel/manual-config.nix
2014-12-30 17:03:39 +01:00
Eelco Dolstra ce6b86cc68 Fix various evaluation problems
http://hydra.nixos.org/build/13616685
2014-08-22 11:57:40 +02:00
Michael Raskin efb0c56db4 Update linux_testing and enable parallel build of Linux kernel 2014-06-30 10:52:33 +04:00
Austin Seipp 80d0e31a94 kernel: allow features to be used in common-config
Signed-off-by: Austin Seipp <aseipp@pobox.com>
2014-05-18 08:49:32 -05:00
Shea Levy cb9cc87a39 kernel.passthru: Don't include meta to be consistent with stdenv.mkDerivation
Signed-off-by: Shea Levy <shea@shealevy.com>
2014-01-21 21:12:58 -05:00
Shea Levy fd999ed570 Linux: Set passthru attr to contain full passthru
This fixes #1566, thanks @wizeman

Signed-off-by: Shea Levy <shea@shealevy.com>
2014-01-21 21:06:27 -05:00
Shea Levy d18bc25b95 Rename linuxManualConfig to buildLinux
Signed-off-by: Shea Levy <shea@shealevy.com>
2014-01-21 20:05:55 -05:00
Shea Levy ac2035287f Greatly reduce kernel closure size
Based on access analysis with strace, I determined an essentially
minimal required set of files from the kernel source that was needed to
build all current kernel packages on 3.10, which ultimately resulted in
keeping 30M of source. Generalizing from that minimal set, which
required ad-hoc specifications of which headers outside of include/ and
arch/*/include and which files in the scripts/ directory should be kept,
to a policy of keeping all non-arch-specific headers that aren't part of
the drivers/ directory and the entire scripts/ directory added an
additional 17M, but there was nothing in the analysis that indicated
that that ad-hoc specification was at all complete so I think the extra
hit is worth the likely greater compatibility.

For reference, we now keep:

* All headers that are NOT in arch/${notTargetArch}/include or drivers/
* The scripts/ directory
* Makefile
* arch/${targetArch}/Makefile

IMO the most likely cause of future problems are the headers in
drivers/, but hopefully they won't actually be needed as they add 50M
Ideally kernel packages would only use include and
arch/${targetArch}/include, but alas this is observably not the case.

master:
  * $out
    * size: 234M
    * references-closure: linux-headers, glibc, attr, acl, zlib, gcc,
      coreutils, perl, bash
merge-kernel-builds:
  * $out
    * size: 152M
    * references-closure: none
  * $dev
    * size: 57M
    * references-closure: linux-headers, glibc, zlib, gcc

So even with the non-minimal set we still beat out master. Keeping the
drivers headers would make us only slightly bigger.

Signed-off-by: Shea Levy <shea@shealevy.com>
2014-01-05 06:55:47 -05:00
Shea Levy 2c38df1c5b kernel build: limit dev output footprint
This makes the disk usage footprint of building the kernel smaller in 3
ways:

1) There is no separate kernel source derivation
2) Rather than using the entire build tree, only the output of make
modules_prepare is kept in the $dev output (plus the module symbol
versioning file generated during the build)
3) Only the subset of the source tree known to be needed for external
builds is kept in $dev

Note that while 2) is supported by official kernel documentation, I
couldn't find any source describing what we need to keep for 3). I've
started with the bare minimum (the main Makefile is called by the
Makefile generated by make modules_prepare) and we can/should add more
as needed for kernelPackages.

Signed-off-by: Shea Levy <shea@shealevy.com>
2014-01-01 23:56:24 -05:00