| Commit message (Collapse) | Author | Age |
... | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
RHEL explicitly disables IFLA_BRPORT_PROXYARP by renaming the enum value.
In order to support unpatched builds, we have two options:
a) redefine the enum value through missing.h and ignore the fact that it
is really unsupported, or
b) omit that enum value on rtnl_prot_info_bridge_port_types[]
As we are not actually using this netlink type anywhere, and because it
is only hooked up for the sake of completeness, this patch opts for the
former.
|
|
|
|
|
|
| |
Usually, we place the #pragma once before the copyright blurb in header files,
but in a few cases we didn't. Move those around, so that we do the same thing
everywhere.
|
|
|
|
|
| |
After all, most ETHERTYPE variables are defined in the system headers, hence
define these where we defined all other fill-ins for system headers.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This patch contains a set of little cleanups for alloc-util.h:
1. The malloc_multiply(), realloc_multiply() and memdup_multiply()
functions check allocation related parameters on overflow. Let's
move them to the separate size_multiply_overflow() function for
simplicity, code duplication prevention and possible reuse in future.
2. use SIZE_MAX from stdlib instead of ((size_t) - 1) to be more
clear.
3. The 'a'/'b' variables are renamed to 'size' and 'need' to be
more clear.'
|
|
|
|
|
|
| |
The deserialize_timestamp_value() is renamed timestamp_deserialize() to be more
consistent with dual_timestamp_deserialize()
And add the NULL check back on realtime and monotonic
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Before this patch existence of char16_t, char32_t, key_serial_t was checked
with AC_CHECK_DECLS() which doesn't actually work for types. Correct this to
use AC_CHECK_TYPES() instead.
Also, while we are at it, change the check for memfd_create() to use
AC_CHECK_DECLS() instead of AC_CHECK_FUNCS(). This is a better choice, since a
couple of syscalls are defined by glibc but not exported in the header files
(pivot_root() for example), and we hence should probably be more picky with
memfd_create() too, which glibc might decide to expose one day, but not
necessarily in the headers too.
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The time-util.c provides dual_timestamp_deserialize() function to
convert value to usec_t and set it as value of ts->monotonic and
ts->realtime.
There are some places in code which do the same but only for one
clockid_t (realtime or monotonic), when dual_timestamp_deserialize()
sets value of both.
This patch introduces the deserialize_timestamp_value() which converts
a given value to usec_t and write it to a given timestamp.
|
|
|
|
| |
let's export as little as we can
|
|
|
|
|
|
| |
We already define IFLA_PROMISCUITY and some other of these masks in
order to allow building with older headers. Define IFLA_EXT_MASK too,
which was added in the same kernel version as IFLA_PROMISCUITY (v3.10).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Empty strings were ignored in strv_join, but only if they were at the beginning
of the string. Empty strings after at least one non-empty item were treated
normally.
Previously:
{"x"} → "x"
{"x", ""} → "x"
{"x", "", ""} → "x::"
{""} → ""
{"", ""} → ""
{"", "", ""} → ""
{"", "x"} → "x"
{"", "x", ""} → "x:"
Now:
{"x"} → "x"
{"x", ""} → "x"
{"x", "", ""} → "x::"
{""} → ""
{"", ""} → ":"
{"", "", ""} → "::"
{"", "x"} → ":x"
{"", "x", ""} → ":x:"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
ISO/IEC 9899:1999 §7.21.1/2 says:
Where an argument declared as size_t n specifies the length of the array
for a function, n can have the value zero on a call to that
function. Unless explicitly stated otherwise in the description of a
particular function in this subclause, pointer arguments on such a call
shall still have valid values, as described in 7.1.4.
In base64_append_width memcpy was called as memcpy(x, NULL, 0). GCC 4.9
started making use of this and assumes This worked fine under -O0, but
does something strange under -O3.
This patch fixes a bug in base64_append_width(), fixes a possible bug in
journal_file_append_entry_internal(), and makes use of the new function
to simplify the code in other places.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Since Linux v4.4-rc1, __DEVEL__sane_behavior does not exist anymore and
is replaced by a new fstype "cgroup2".
With this patch, elogind no longer supports the old (unstable) way of
doing unified hierarchy with __DEVEL__sane_behavior and elogind now
requires Linux v4.4 for unified hierarchy.
Non-unified hierarchy is still the default and is unchanged by this
patch.
https://github.com/torvalds/linux/commit/67e9c74b8a873408c27ac9a8e4c1d1c8d72c93ff
|
| |
|
|
|
|
|
|
| |
- musl-libc has no <bits/local_lim.h>
- musl-libc has no <xlocale.h>
- added <missing.h> include to basic/mdkir.c
|
| |
|
| |
|
|
|
|
|
|
|
| |
Now that dedicated libc headers are included in their corresponding foo-util.h
headers, stdio-util.h is the only place left, where printf.h was included.
The nice thing is, that this is the only place left where we have to take care
about printf and musl-libc.
|
| |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Support for net_cls.class_id through the NetClass= configuration directive
has been added in v227 in preparation for a per-unit packet filter mechanism.
However, it turns out the kernel people have decided to deprecate the net_cls
and net_prio controllers in v2. Tejun provides a comprehensive justification
for this in his commit, which has landed during the merge window for kernel
v4.5:
https://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=bd1060a1d671
As we're aiming for full support for the v2 cgroup hierarchy, we can no
longer support this feature. Userspace tool such as nftables are moving over
to setting rules that are specific to the full cgroup path of a task, which
obsoletes these controllers anyway.
This commit removes support for tweaking details in the net_cls controller,
but keeps the NetClass= directive around for legacy compatibility reasons.
|
|
|
|
|
| |
This should be handled fine now by .dir-locals.el, so need to carry that
stuff in every file.
|
|
|
|
|
|
|
| |
The dual_timestamp_from_realtime(), dual_timestamp_from_monotonic()
and dual_timestamp_from_boottime_or_monotonic() shares the same
code for comparison given ts with delta. Let's move it to the
separate inline function to prevent code duplication.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
This reworks the coredumping logic so that the coredump handler invoked from the kernel only collects runtime data
about the crashed process, and then submits it for processing to a socket-activate coredump service, which extracts a
stacktrace and writes the coredump to disk.
This has a number of benefits: the disk IO and stack trace generation may take a substantial amount of resources, and
hence should better be managed by PID 1, so that resource management applies. This patch uses RuntimeMaxSec=, Nice=, OOMScoreAdjust=
and various sandboxing settings to ensure that the coredump handler doesn't take away unbounded resources from normally
priorized processes.
This logic is also nice since this makes sure the coredump processing and storage is delayed correctly until
/var/elogind/coredump is mounted and writable.
Fixes: #2286
|
|
|
|
|
|
| |
The time-util.c provides dual_timestamp_get() function for getting
realtime and monotonic timestamps. Let's use it instead of direct
realtime/monotonic calculation.
|
|
|
|
| |
they do
|
|
|
|
|
|
|
| |
mode_t is unsigned, so MODE_INVALID < 0 can never be true.
This fixes a possible DoS where any user could fill /run by writing to
a world-writable /run/elogind/show-status.
|
|
|
|
|
|
| |
environments
This adds some basic infrastructure in order to fix #2015.
|
|
|
|
|
|
| |
Let's make sure we read it in a way compatible with non-aligned memory.
Fixes: https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=812928
|
|
|
|
|
|
|
| |
When the user specifies "foo*" as unit name glob expression, we shouldn't turn this into "foo*.service". Hence: only
append a suffix if the specified string isn't a glob expression.
Fixes: #2397
|
| |
|
|
|
|
| |
rootfs as PID1
|
|
|
|
|
|
|
|
|
| |
core: Add flexible way to provide socket type
the socket type should be a diffrent argumet
in make_socket_fd . In this way we can set the socket
type like SOCK_STREAM SOCK_DGRAM in the address.
journal-remote: modify make_socket_fd
|
|
|
|
| |
Inspired by https://github.com/elogind/elogind/issues/2187#issuecomment-165587140
|
|
|
|
|
| |
This small addition fixes the issues #1982 and #2242.
IPv6PrivacyExtension now works as expected even when a RA is received.
|
|
|
|
|
| |
In contrast to ascii_strcasecmp_nn() it takes two character buffers with their individual length. It will then compare
the buffers up the smaller size of the two buffers, and finally the length themselves.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Not every byte sequence is valid utf8. We allow escaping of non-utf8
sequences in strings by using octal and hexadecimal escape sequences
(\123 and \0xAB) for bytes at or above 128. Users of cunescape_one
could infer whether such use occured when they received an answer
between 128 and 256 in *ret (a non-ascii one byte character). But this
is subtle and misleading: the comments were wrong, because ascii is a
subset of unicode, so c != 0 did not mean non-unicode, but rather
ascii-subset-of-unicode-or-raw-byte. This was all rather confusing, so
make the "single byte" condition explicit.
I'm not convinced that allowing non-utf8 sequences to be produced is
useful in all cases where we allow it (e.g. in config files), but that
behaviour is unchanged, just made more explicit.
This also fixes an (invalid) gcc warning about unitialized variable
(*ret_unicode) in callers of cunescape_one.
|
|
|
|
| |
Also add a coccinelle receipt to help with such transitions.
|
|
|
|
|
| |
Compare errno with zero in a way that tells gcc that
(if the condition is true) errno is positive.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
gcc is confused by the common idiom of
return errno ? -errno : -ESOMETHING
and thinks a positive value may be returned. Replace this condition
with errno > 0 to help gcc and avoid many spurious warnings. I filed
a gcc rfe a long time ago, but it hard to say if it will ever be
implemented [1].
Both conventions were used in the codebase, this change makes things
more consistent. This is a follow up to bcb161b0230f.
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=61846
|
|
|
|
|
|
|
|
|
|
|
|
| |
Rather than walking a list of valid values one-by-one, generate a
switch-case statement for the IN_SET() macro. This allows the compiler to
further optimize its code output, possibly by generating jump tables.
This effectively decreases the binary size slightly.
The implementation is based on macro overloading depending on the number of
arguments. h/t to the following post:
https://stackoverflow.com/questions/11761703/overloading-macro-on-number-of-arguments
|
|
|
|
|
| |
This also introduces a new macro siphash24_compress_byte() which is useful to add a single byte into the hash stream,
and ports one user over to it.
|
|
|
|
| |
The hash operations are not really that specific to hashmaps, hence split them into a .c module of their own.
|
| |
|
|
|
|
|
|
|
|
|
|
| |
In a couple of cases we maintain configuration settings that know an on
and off state, like a boolean, plus some additional states. We generally
parse them as booleans first, and if that fails check for specific
additional values.
This adds a generalized set of macros for parsing such settings, and
ports one use in resolved and another in networkd over to it.
|
| |
|
|
|
|
|
| |
SCTP_NODELAY is diffrent to TCP_NODELAY.
Apply proper options in case of SCTP.
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
rework C11 utf8.[ch] to use char32_t instead of uint32_t when referring
to unicode chars, to make things more expressive.
[
@zonque:
* rebased to current master
* use AC_CHECK_DECLS to detect availibility of char{16,32}_t
* make utf8_encoded_to_unichar() return int
]
|
|
|
|
| |
this is a follow-up for commit 11c3a36649e5e5e77db499c92f3
|