Detect image type using "bootctl kernel-identify $kernel",
store result in KERNEL_INSTALL_IMAGE_TYPE.
Extend layout autodetection to check the kernel image type
and pick layout=uki for UKIs.
Resolves: #25822
Read the magic first, try reading the full DOS exe header only in case
the magic check succeeds.
This avoids throwing an header read error on small dummy files as used
by test-kernel-install.
Tablets don't typically have relative coordinates (they are separated on
the kernel device layer). However, some Logitech mice report similar
supported events, so use the existence of EV_REL to determiner whether or
not the device is really a tablet.
Fixes bug introduced by 0855ce6772.
Fixes: #26600
On distros like Arch and on carbonOS, libraries end up in /usr/lib.
Thus, /lib64 should point to /usr/lib. This commit adds this
functionality as a final fallback (if neither Debian-style nor
Fedora-style multilib can be detected)
We prefer /efi as a mount point for the ESP, and use /boot as a fallback
if /efi doesn't exist. However, when root=tmpfs, neither /efi nor /boot
exist. gpt-auto should mount to /efi in this case, but it mounted to
/boot instead. This is because gpt-auto didn't check for the existence
of /boot. Here, we correct this
Starting with commit ed88bcfb7c,
udev_rules_parse_file() silently skips files it fails to open with
ENOENT error, e.g. when they are broken symlinks. As this behavior is
undocumented and it seems to be unintended, let's treat ENOENT like any
other error. This change would also simplify the implementation of the
udev rules syntax checker mentioned in #26606.
udev_rules_load(), the only user of udev_rules_parse_file(), is not
affected by this change because it essentially ignores the value
returned by the latter, the only visible difference would be a log
message issued for every udev rules file that couldn't be open because
of ENOENT.
Fixes: ed88bcfb7c ("Be more careful when checking for empty files")
IN C23, thread_local is a reserved keyword and we shall therefore
do nothing to redefine it. glibc has it defined for older standard
version with the right conditions.
v2 by Yu Watanabe:
Move the definition to missing_threads.h like the way we define e.g.
missing syscalls or missing definitions, and include it by the users.
Co-authored-by: Yu Watanabe <watanabe.yu+github@gmail.com>
This reverts commit 81cfea95e5.
The modalias seems to match a generic Logitech USB receiver even the
connected mouce is not for left hand.
Fixes#26671 and #26676.
Before this commit, if --since is used with --lines=N,
we seek to the place of --since and search afterwards
there, resulting in outputing the first N lines.
After this commit, we only do the above if --since is used without
--reverse and --lines. Otherwise we seek to the tail first and check
if the entry is within the range of --since later.
Before this commit, when --lines is specified, we jump to the tail and
search afterwards from there, thus breaking --until if used together.
After this commit:
If both --until and any of --reverse and --lines is specified, things get
a little tricky. We seek to the place of --until first. If only --reverse or
--reverse and --lines is specified, we search backwards and let the output
counter handle --lines for us. If only --lines is used, we just jump backwards
arg_lines and search afterwards from there.
Otherwise under certain conditions `va_arg()` might get garbage instead
of the expected value, i.e.:
$ sudo build-o0/systemctl disable asdfasfaf
sd_bus_message_appendv: Got uint64_t: 0
Failed to disable unit: Unit file asdfasfaf.service does not exist.
$ sudo build-o1/systemctl disable asdfasfaf
sd_bus_message_appendv: Got uint64_t: 7954875719681572864
Failed to disable unit: Invalid argument
(reproduced on an armv7hl machine)
Resolves: #26568
Follow-up to: bf1bea43f1
Related issue: https://github.com/systemd/systemd/pull/14470#discussion_r362893735
And modernize heavily while doing so.
Fixes: #21787
(Strictly speaking, this leaves a race window open: the the system is
powered off in the short interval when we linked in the prepared hwdb
file into the dir under a temporary name and are about to rename it to
the final name, then the file might be left over after all. But this
minimizes the window so much that this shouldn't be an issue in
real-life. Key after all is that with this change we'll build up the
hwdb file under O_TMPFILE, and thus are robust to power loss during the
slow operation)
Gerd Hoffmann
Yu Watanabe
Joshua Goins
Luca Boccassi
Josef Miegl
Adrian Vovk
Daan De Meyer
Lennart Poettering
Dmitry V. Levin
Cristian RodríguezandYu Watanabe
Zbigniew Jędrzejewski-Szmek
Mike Yuan
Frantisek Sumsal
наб