On aarch64 we "leak" a reference to $out/lib/systemd/catalog in the lib
output. The result of that is a dependency cycle between $out and $lib.
Thus nix (rightfully) marks the build as failed. That reference
originates from an array of strings (catalog_file_dirs) in systemd
(src/src/journal/catalog.{c,h}). The only consumer (as of v242) of the
symbol is the main function of journalctl. Still libsystemd.so contains
the VALUE but not the symbol. Systemd seems to be properly using
function & data sections together with the linker flags to garbage
collect unused sections (-Wl,--gc-sections). For unknown reasons those
flags do not eliminate the unused string constants, in this case on
aarch64-linux. The hacky way is to just remove the reference after we
finished compiling. Since it can not be used (there is no symbol to
actually refer to it) there should not be any harm. It is a bit odd and
I really do not like starting these kind of hacks but there doesn't seem
to be a straight forward way at this point in time.
The reference will be replaced by the same reference the usual nukeRefs
tooling uses. The standard tooling can not / should not be uesd since
it is a bit too excessive and could potentially do us some (more) harm.
We are currently not running any tests but building them takes
signitifcant amounts of time since they account to about 40% of all the
compilation targets.
The current approach will fail when enough time has passed. We ideally
want to be reproducible even in a few years of time. So we should pick
the sources of patches wisely as otherwise we can not do that.
This reverts commit d1de23b8302d02d4699e884533906a3992f370b6.
The changes turned out to be too intrusive, so we'll patch instead.
Discussion: https://github.com/NixOS/systemd/pull/24
Fixes CVE-2018-15688 and updates latest upstream stable v239 branch.
See https://github.com/NixOS/systemd/pull/24 for details.
Co-authored-by: Andreas Rammhold <andreas@rammhold.de>
meson 0.46 no longer likes receiving both -Dmandir and --mandir. I removed the flags from the expression in favour of those in the meson setup hook. This also fixes manpages which were previously
installed to $man/lib for some reason.
Following legacy packing conventions, `isArm` was defined just for
32-bit ARM instruction set. This is confusing to non packagers though,
because Aarch64 is an ARM instruction set.
The official ARM overview for ARMv8[1] is surprisingly not confusing,
given the overall state of affairs for ARM naming conventions, and
offers us a solution. It divides the nomenclature into three levels:
```
ISA: ARMv8 {-A, -R, -M}
/ \
Mode: Aarch32 Aarch64
| / \
Encoding: A64 A32 T32
```
At the top is the overall v8 instruction set archicture. Second are the
two modes, defined by bitwidth but differing in other semantics too, and
buttom are the encodings, (hopefully?) isomorphic if they encode the
same mode.
The 32 bit encodings are mostly backwards compatible with previous
non-Thumb and Thumb encodings, and if so we can pun the mode names to
instead mean "sets of compatable or isomorphic encodings", and then
voilà we have nice names for 32-bit and 64-bit arm instruction sets
which do not use the word ARM so as to not confused either laymen or
experienced ARM packages.
[1]: https://developer.arm.com/products/architecture/a-profile
(cherry picked from commit ba52ae5048)
Following legacy packing conventions, `isArm` was defined just for
32-bit ARM instruction set. This is confusing to non packagers though,
because Aarch64 is an ARM instruction set.
The official ARM overview for ARMv8[1] is surprisingly not confusing,
given the overall state of affairs for ARM naming conventions, and
offers us a solution. It divides the nomenclature into three levels:
```
ISA: ARMv8 {-A, -R, -M}
/ \
Mode: Aarch32 Aarch64
| / \
Encoding: A64 A32 T32
```
At the top is the overall v8 instruction set archicture. Second are the
two modes, defined by bitwidth but differing in other semantics too, and
buttom are the encodings, (hopefully?) isomorphic if they encode the
same mode.
The 32 bit encodings are mostly backwards compatible with previous
non-Thumb and Thumb encodings, and if so we can pun the mode names to
instead mean "sets of compatable or isomorphic encodings", and then
voilà we have nice names for 32-bit and 64-bit arm instruction sets
which do not use the word ARM so as to not confused either laymen or
experienced ARM packages.
[1]: https://developer.arm.com/products/architecture/a-profile
The isSeccomputable flag treated Linux without seccomp as just a
normal variant, when it really should be treated as a special case
incurring complexity debt to support.
The isKexecable flag treated Linux without kexec as just a normal
variant, when it really should be treated as a special case incurring
complexity debt to support.
Updated to the latest version of the nixos-v237 branch, which fixes two
things:
* Make sure that systemd looks in /etc for configuration files.
https://github.com/NixOS/systemd/pull/15
* Fix handling of the x-initrd.mount option.
https://github.com/NixOS/systemd/pull/16
I've added NixOS VM tests for both to ensure we won't run into
regressions. The newly added systemd test only tests for that and is by
no means exhaustive, but it's a start.
Personally I only wanted to fix the former issue, because that's the one
I've been debugging. After sending in a pull request for our systemd
fork (https://github.com/NixOS/systemd/pull/17) I got a notice from
@Mic92, that he already fixed this and his fix was even better as it's
even suitable for upstream (so we hopefully can drop that patch
someday).
The reason why the second one came in was simply because it has been
merged before the former, but I thought it would be a good idea to have
tests for that as well.
In addition I've removed the sysconfdir=$out/etc entry to make sure the
default (/etc) is used. Installing is still done to $out, because those
directories that were previously into sysconfdir now get into
factoryconfdir.
Quote from commit NixOS/systemd@98067cc806ae0d2759cdd2334f230cd8548e531:
By default systemd should read all its configuration from /etc.
Therefore we rely on -Dsysconfdir=/etc in meson as default value.
Unfortunately this would also lead to installation of systemd's own
configuration files to `/etc` whereas we are limited to /nix/store. To
counter that this commit introduces two new configuration variables
`factoryconfdir` and `factorypkgconfdir` to install systemd's own
configuration into nix store again, while having executables looking
up files in /etc.
Tested this change against all of the NixOS VM tests we have in
nixos/release.nix. Between this change and its parent no new tests were
failing (although a lot of them were flaky).
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @Mic92, @tk-ecotelecom, @edolstra, @fpletzFixes: #35415Fixes: #35268
The indenting is a bit weird to follow, especially at the end of the
file (right brace without indent, but the opening brace is indented by
two spaces).
No functional change and I've verified this by building it with this
change and without and both lead to the same store paths.
Signed-off-by: aszlig <aszlig@nix.build>
Cc: @fpletz, @edolstra
volth