network-manager: install libnss DNS and mDNS plugins
Install libnss_dns.so and libnss_mdns4_minimal.so plugins for the Name
Service Switch (NSS) functionality of glibc so that name resolution
through /etc/resolv.conf and mDNS works in the initrd.
/etc/vconsole.conf must be installed always, even in generic initrds, not only host-only ones.
systemd-vconsole-setup is run at initrd stage and expects either /etc/vconsole.conf or kernel cmdline options to exist.
I have also proposed a change in systemd-vconsole-setup that makes it behave better if /etc/vconsole.conf does not exist:
https://github.com/systemd/systemd/pull/15479
But it is just a fallback. If /etc/vconsole.conf does not exist in initrd and if that patch is applied on systemd,
then the default consoel font is set despite the one being set in /etc/vconsole.conf and this setting is inherited
when new TTYs are opened. This leads to full ignorance of settings in /etc/vconsole.conf.
It is incorrect, and that is why this file must be copied to initrds always, but not only on host-only initrds.
Jonas Witschel [Sat, 18 Apr 2020 12:55:41 +0000 (14:55 +0200)]
dracut-lib.sh: quote variables in parameter expansion patterns
According to POSIX.1-2017, 2.6.2 Parameter Expansion:
${parameter%[word]} [...] The word shall be expanded to produce a
pattern.
This means if word contains variables that itself contain special
characters like asterisks or backslashes, these are treated as pattern
characters unless the variable is quoted. Try e.g. the following example
in bash, dash or (busybox) ash:
i='a\c'; j='\'; echo "${i%$j*}"
This prints "a\c" because "$j*" is expanded to "\*", escaping the
asterisk. In contrast,
i='a\c'; j='\'; echo "${i%"$j"*}"
produces the expected result "a" because the backslash is not specially
treated any more after quoting.
The quotes that this commit adds have been previously removed in commit f9c96cf56fed390841eac05c43826e62014c9188, citing issues with busybox
hush without further specifying the actual error. I tested a recent
busybox build (upstream commit 9aa751b08ab03d6396f86c3df77937a19687981b)
and couldn't find any problems. Note that the above example always
produces "a\c" in hush regardless of quoting $j, making hush unsuitable
for use with dracut, but using quotes in parameter expansions generally
works.
The unquoted variables break the "rd.luks.uuid/name" kernel command line
options in dracut 050 because
str_replace "$luksname" '\' '\\'
in modules.d/90crypt/parse-crypt.sh is not able to escape the
backslashes any more, see GH-723, GH-727: backslashes in the
systemd-cryptsetup@.service unit name stay unescaped for use in udev
(cf. commit 0f6d93eb9d63695a64002ec8b0421fbc9fc8a7a3), leading to
failures in starting the unit.
Kairui Song [Thu, 5 Mar 2020 10:44:45 +0000 (18:44 +0800)]
95znet: Add a rd.znet_ifname= option
qeth device may have a different IP for each boot, so the rd.ifname=
option will no longer work. So for znet device, introduce a
rd.znet_ifname= options, to subchannel id instead of MAC address as the
identifier and rename the interface.
Ben Howard [Fri, 3 Apr 2020 19:32:44 +0000 (13:32 -0600)]
multipath: add automatic configuration for multipath
Add support for 'rd.multipath=default' for using the default
configuration on boot. The intended purpose for this is to help support
ostree-based image boots from multipathed devices (such as Fedora and
Red Hat CoreOS).
Daniel Molkentin [Fri, 13 Mar 2020 10:39:46 +0000 (11:39 +0100)]
Makefile: really make externally defined CFLAGS work
The fix in aed52a6cfb43ebea372328fd5837c1f341c3f0df does not
work (second line will be ignored on override), so expand the implicitly
defined rule and add them there.
Dusty Mabe [Fri, 13 Mar 2020 19:40:11 +0000 (15:40 -0400)]
don't prefer $TMPDIR over --tmpdir
Ability to use `$TMPDIR` was introduced in bc1e69b but
causes the `--tmpdir` option on the command line to be
ignored. Switch to only using `$TMPDIR` if `--tmpdir`
wasn't specified.
network-manager: ensure that nm-run.sh is executed when needed
The network-manager command line hook must install a
initqueue/finished hook to ensure that nm-run.sh is executed when
there are network connections to activate.
Topi Miettinen [Sun, 1 Mar 2020 08:45:16 +0000 (10:45 +0200)]
Use TMPDIR if available
Use environment variable TMPDIR (typically /run/user/$UID) as default
temporary directory, if available. This should be more private
location than /var/tmp. Path specified with --tmpdir is takes
precedence over TMPDIR and /var/tmp is still used as last resort if
neither TMPDIR is set nor --tmpdir is used.
Martin Wilck [Wed, 29 Jan 2020 22:53:29 +0000 (23:53 +0100)]
dracut.sh: don't call fsfreeze on subvol of root file system
dracut.sh already doesn't call fsfreeze if the output file is on
the root file system. For btrfs, however, this is not sufficient.
Because fsfreeze is a superblock operation, and all btrfs subvolumes
share the same superblock, fsfreeze may freeze the entire system
if the subvolume on which the output file is written and / are
subvolumes of the same file system. Avoid this by comparing file
system UUIDs for btrfs.
Fixes: de576db3c225 ("call fsfreeze(8) on /boot to flush initramfs data & metadata to media")
Martin Wilck [Thu, 30 Jan 2020 12:48:11 +0000 (13:48 +0100)]
Add module "90nvdimm" for NVDIMM support
Detection of persistent memory devices works mostly out of the box
already. Only the "provider" modules for ndbus devices, which are responsible
to extract information of available NVDIMM devices and their configuration
from system firmware, are only indirectly linked into the module stack.
Examples for such modules are nfit.ko, nd_e820.ko, and virtio-pmem.ko.
Makefile: merge main-version and git-version earlier
With GNU Make 4.3 on both ArchLinux, and VoidLinux,
GITVERION is always empty because of bad substitution.
Change '\#' to simply '#' can fix it,
but we don't need that complation.
We can merge DRACUT_MAIN_VERSION and GITVERSION into DRACUT_FULL_VERSION.
Because, GITVERSION will be attached back to DRACUT_MAIN_VERSION in all
situation.
While we're at it, detect if we're in git worktree by:
limiting GIT_CEILING_DIRECTORIES to parent directory of
dracut's top level directory; instead of checking for .git directory,
in order to support git-worktree, in such case, .git will be a file, see
gitrepository-layout(5)
Currently there is no usb support on RPi4 in the
initrd phase as the pcie-brcmstb module is missing.
If part of the boot is handled from a USB stick
(e.g. with Ignition), the stick cannot be accessed.
Topi Miettinen [Wed, 26 Feb 2020 12:48:18 +0000 (14:48 +0200)]
Don't resolve libraries lazily if tmpdir is mounted with 'noexec'
If the temporary directory for images is mounted with 'noexec', dracut
would construct unbootable images because most dynamic libraries
aren't installed. Avoid this by not resolving library dependencies
lazily if the temporary directory is mounted with 'noexec'.
Frantisek Sumsal [Fri, 28 Feb 2020 17:07:12 +0000 (18:07 +0100)]
github: fetch git tags before testing
GitHub workflows fetch a clone of the dracut repository which doesn't
contain git tags, thus "breaking" the RPM build in certain situations
i.e.:
DRACUT_MAIN_VERSION in Makefile is defined as an output of `git describe`,
which in full git clone returns a tag with a numeric version. However,
without tags it returns SHA of the last commit, which later propagates into
`Provides:` attribute of the built RPM and can break dependency tree when
installed
Frantisek Sumsal [Mon, 24 Feb 2020 20:56:37 +0000 (21:56 +0100)]
Makefile: fix RPM build
`make rpm` usually chose `/tmp` as the `$rpmbuild` dir, which breaks the
dracut build, since it needs to execute `./configure`, but `/tmp` is
mounted with `-o noexec`, leading to:
```
/var/tmp/rpm-tmp.SwEhJO: line 46: ./configure: Permission denied
error: Bad exit status from /var/tmp/rpm-tmp.SwEhJO (%build)
RPM build errors:
Bad exit status from /var/tmp/rpm-tmp.SwEhJO (%build)
```
Harald Hoyer [Wed, 26 Feb 2020 15:53:47 +0000 (16:53 +0100)]
network-legacy/net-genrules.sh: use $name instead of $env{INTERFACE}
The original behavior of $env{INTERFACE} was undocumented and changed in
the recent udev versions, breaking the ability to bring up networking
reliably. Switching to $name directive should fix this issue.
Previously, we were doing `setup_net` from `ifup` for any setup that
wasn't DHCP, since those are already taken care of by `dhclient-script`.
The issue is that the case-statement we use to detect this doesn't catch
options like `ip=dhcp,dhcp6`.
Fix this by reworking the logic here to just check if a
`setup_net_$netif.sh` hook exists. If so, then we know that `setup_net`
will be called for this interface later.
This was causing issues in RHCOS which now ships with `ip=dhcp,dhcp6` to
support IPv6 environments[1]. The code here would make us do `setup_net`
pre-emptively which IIUC would then cause the initqueue to finish
earlier even if we had more udev netif events to process.