It's mostly wrappers for compatibility and another trivial stuff etc.
Let's keep it as public domain to make it more portable to LGPL, GPL
and BSD code.
Signed-off-by: Karel Zak <kzak@redhat.com>
CC: Sami Kerola <kerolasa@iki.fi>
CC: Ruediger Meier <ruediger.meier@ga-group.nl>
Move the helper functions that parse ISO data & sector size info into
a separate, shared header.
These will addtionally be used by libblkid's iso9660 parser.
Signed-off-by: Daniel Drake <drake@endlessm.com>
* 'kill-pidfd' of https://github.com/kerolasa/util-linux:
kill: use pidfd system calls to implement --timeout option
build-sys: add missing NR underscore to UL_CHECK_SYSCALL()
At times there is need in scripts to send multiple signals to a process.
Often these cases require some amount of waiting before follow-up signal
should be sent.
One common case is process termination, where first script tries to kill
process gracefully but if that does not work SIGKILL is sent. Functionality
like that is commonly done by periodically checking if signalled pid exist
or not, and if it does another signal is sent possibly to an unrelated
process that reused pid number. That means polling a pid is prone to a data
race. Also if the first signal immediately kills the process one polling
interval is lost in sleep.
Another example when multiple signal need to be sent is various daemon
process control situations, such as Upgrading Executable on the Fly (see
reference). This happens to be the case that inspired change author to make
sequential signaling a little bit easier.
Reference: http://nginx.org/en/docs/control.html#upgrade
Pull-request: https://github.com/karelzak/util-linux/pull/902
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
It some cases it makes sense to disable ECHO flag also when script
used in pipe. This new option allows to keep full control in user's
hands.
Signed-off-by: Karel Zak <kzak@redhat.com>
We need a proper way how to inform child (shell) that the game is
over. It seems the best is to send EOF to child rather than
immediately break PTY mainloop where we have poll(), because shell can
still produce data etc.
Signed-off-by: Karel Zak <kzak@redhat.com>
For tools like su(1) is ECHO flag unexpected for use-case like
echo 'date' | su - user
but script(1) need the echo to keep input recorded.
The patch also return execlp() use to script(1) code.
Signed-off-by: Karel Zak <kzak@redhat.com>
The parse-date.y is used only for hwclock, let's keep it together.
Note that the file (originally from gnulib) has GPLv3 license, so it's
better to make it obvious that we use it really only for hwclock (also
GPL).
Signed-off-by: Karel Zak <kzak@redhat.com>
* improve strdup_to_offset() readability
* add strdup_between_offsets() and strdup_between_structs() to have
better support for use-cases when we copy structs
Signed-off-by: Karel Zak <kzak@redhat.com>
In commit b1418ed14 (lib/path: add ul_path_stat(), fix absolute paths,
2019-10-15), a new function `ul_path_stat()` was added to "path.h". This
new function prototype causes a compiler warning on musl libc based
systems due to one of the parameters having the unknown type `struct
stat` due to the <sys/stat.h> header not being included.
Fix the warning by including the header.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
The classic way which is based on file-descriptors table size is
pretty expensive (due to table size) and forces code to do many
unnecessary close() calls. It seems better to use /proc/self/fds and
close used descriptors only.
Addresses: https://github.com/karelzak/util-linux/issues/883
Signed-off-by: Karel Zak <kzak@redhat.com>
* add ul_path_stat()
* make sure all paths for ul_path_..() functions are always
interpreted relatively to the context directory and prefix. This is
difference between ul_path_ API and standard libc "at" functions. We
do not use any exception for absolute paths. The reason is that we
need to read from prefixed paths although application assume absolute
path (/dev/sda1 means /prefix/dev/sda1 if a /prefix is defined).
Signed-off-by: Karel Zak <kzak@redhat.com>
Now the code is duplicate on many places, but all we usually need is to
remember child status. It seems good enough to have very simple
callback child_die() to inform application about a change.
The patch also add PID to all signal related callbacks.
Signed-off-by: Karel Zak <kzak@redhat.com>
This allows to control mainloop behavior from PTY applications. For
example you can write to child (shell) process independently on the
current stdin.
Signed-off-by: Karel Zak <kzak@redhat.com>
The idea is to consolidate script(1), scriptlive(1) and su(1) --pty
and use the same code everywhere.
TODO: add callbacks for stdin/out logging (necessary for script(1)).
Signed-off-by: Karel Zak <kzak@redhat.com>
Sorry detail-oriented people tend to wipe these out if they notice them.
Add in automated tools and lots of excess end-of-line spaces get wiped
out.
Addresses: https://github.com/karelzak/util-linux/pull/849
Signed-off-by: Karel Zak <kzak@redhat.com>
Add the --keep-caps option to unshare to preserve capabilities that
are granted when creating a new user namespace. This allows the child
process to retain privilege within the new user namespace without also
being UID 0.
In order to avoid closing standard streams multiple times, commit
52aa1a661 (include/closestream: avoid close more than once, 2019-06-13)
introduced code to set the standard output and error streams to `NULL`.
As musl libc defines standard streams as constant pointers, the change
causes compiler errors on systems with that libc. According to ISO C89,
being able to assign to the standard text streams is not a requirement
for any C implementation, see footnote 238 in chapter §7.19.5.6:
The primary use of the freopen function is to change the file
associated with a standard text stream (stderr, stdin, or stdout),
as those identifiers need not be modifiable lvalues to which the
value returned by the fopen function may be assigned.
This commit implements a new function `flush_standard_stream` that tries
to reliably flush standard streams without actually closing them. By not
calling fclose(3P), we can neatly avoid the issue of accessing standard
streams in an unspecified state and thus remove the infringing `NULL`
assignments.
Properly flushing standard streams without fclose(3P) proves to be more
intricate than one may expect, though, as some filesystems like NFS may
defer flushing until they see a close(3P) of the underlying descriptor.
One may call fsync(3P) to remedy that, but this may incur a heavy
performance penalty in some scenarios. To work around the issue and
still get proper errors, we duplicate the stream's file descriptor and
close that one instead, which is sufficient to cause a flush.
Note that both `close_stdout` and `close_stdout_atexit` are misnamed
after this change, as we do not actually close the streams now. In order
to avoid unnecessary code churn, we still retain their current names.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
There are two things happening in this commit. 1) Move each function
attribute to it's own line 2) move function return value after attributes so
that function inputs and outputs are on same line.
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
There is a c_isascii function that can be used.
isascii is deprecated and not available with some libcs like uClibc-ng
where it can be compile time disabled.
Turned out lsblk is passing null as argument to xstrdup(), so fix that and
add assert() to make sure promise of not returning null is kept in future.
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
The device can be inaccessible for non-root user or busy (already used
by another process). In this case it seems better to read information
from /sys.
Note that /sys does not provide struct watchdog_info.options, so we
cannot print list of supported watchdog features.
Addresses: https://github.com/karelzak/util-linux/issues/804
Signed-off-by: Karel Zak <kzak@redhat.com>
Let's use miscdev /dev/watchdog as fallback only. We need (if possible)
cdev /dev/watchdog0 as this device has entry in /sys/class/watchdog.
Signed-off-by: Karel Zak <kzak@redhat.com>
Let's keep compilers and static analyzers happy. The idea is to use
memcpy() to copy from buffer to variable and use all label as unsigned
char rather than vectorize by uint32_t.
Signed-off-by: Karel Zak <kzak@redhat.com>
The __has macros are more portable (supported by gcc as well as
clang).
The old __GNUC_PREREQ is deprecated and it should be removed in
future. (Well, the __has macros are supported since gcc 5, so we
should be patient as some old stable distros (e.g. RHEL 7) use gcc
4.x).
This patch helps clang to correctly analyze our xalloc.h stuff.
Signed-off-by: Karel Zak <kzak@redhat.com>
Recent lscpu fix caused gcc -Wnull-dereference to go off that this change
addresses.
Reference: b94acada9e
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
From the test input string ':' characters are removed:
cat x86_64-epyc_7451/sys/devices/system/cpu/vulnerabilities/spectre_v2
Mitigation: Full AMD retpoline, IBPB: conditional, STIBP: disabled, RSB filling
Signed-off-by: Sami Kerola <kerolasa@iki.fi>
It seems better to have a way to control when atexit(close_stdout()) is
used, because close stdout means that for example ASAN (or another
into binary integrated tool) is not able to print the final summary.
Signed-off-by: Karel Zak <kzak@redhat.com>