As workarounded by fc0cbed2db, the pair of
subsystem and sysname is not unique. For examples,
- /sys/bus/gpio and /sys/class/gpio, both have gpiochip%N. However, these point to different devpaths.
- /sys/bus/mdio_bus and /sys/class/mdio_bus,
- /sys/bus/mei and /sys/class/mei,
- /sys/bus/typec and /sys/class/typec, and so on.
Let's refuse to provide sd_device object in such cases.
To create the sd_device object of a driver, the function
sd_device_new_from_subsystem_sysname() requires "drivers" for subsystem
and e.g. "pci:iwlwifi" for sysname. Similarly, sd_device_new_from_device_id()
also requires driver subsystem. However, we have never provided a
way to get the driver subsystem ("pci" for the previous example) from
an existing sd_device object.
Let's introduce a way to get driver subsystem.
This partially reverts the commit 730b76bd2c.
Before the commit, the function returned 0 on success, but the commit
made the function always return 1, as if device->devtype is NULL, the
function returns -ENOENT in the above.
Fortunately, udev_device_get_devtype() does not propagate any
non-negative value from sd_device_get_devtype(). Hence, hopefully we can
safely revert the change.
This fixes the following issues:
- We have a journal file, which contains entries of boot A and B. Let T
be the timestamp of the _last_ entry of boot A.
If sd_journal_seek_monotonic_usec() is called for boot A with a timestamp
_after_ T, following sd_journal_next() will provide the _first_ entry of
boot A, rather than the first entry of boot B.
- We have two journal files X and Y. The file X contains entries of boot A.
Let T be the timestamp of the _last_ entry of boot A in file X. The file Y
contains entries of boot A after timestamp T.
If sd_journal_seek_monotonic_usec() is called for boot A with a
timestamp _after_ T, following sd_journal_next() will provide the
_first_ entry of boot A, whose timestamp is of course earlier than T.
To make it work without sd-event.
Prompted by recent chat:
> Hey all!
> reading man libudev, it says to use sd-device instead now. I've read that
> APIs header file and it seems it no longer has an equivalent to libudev's
> udev_monitor_get_fd, which AFAICT means I have to use sd-event to watch
> for events I'm interested in. I know I can "embed" sd-event in other event
> loops I might already have, but that seems overkill when I'm only interested
> in this one type of event and don't need sd-event for anything else.
Previously, the main process of systemd-udevd manages worker process
with their sd_device_monitor object to save the destination address.
Let's save only destination address, and drop worker's sd_device_monitor
object.
Previously, device_monitor_enable_receiving() does
- update filter,
- bind socket.
But, binding socket can be done in when the socket is opened.
Let's remove device_monitor_enable_receiving() and bind the socket in
device_monitor_new_full().
The struct itself is generic, and can be used for other ID.
Let's rename it to more generic one.
No functional change, just refactoring and preparation for later
commits.
Since Linux commit ddd1ad68826d ("net: bridge: Add netlink knobs for number
/ max learned FDB entries") [1] it is possible to limit to number of
dynamically learned fdb entries per bridge.
Add support to the systemd networkctl for the netlink bridge attributes
IFLA_BR_FDB_MAX_LEARNED and IFLA_BR_FDB_N_LEARNED.
[1] https://lore.kernel.org/all/20231016-fdb_limit-v5-0-32cddff87758@avm.de/
Signed-off-by: Gregor Herburger <gregor.herburger@ew.tq-group.com>
Since Linux commit ddd1ad68826d ("net: bridge: Add netlink knobs for number
/ max learned FDB entries") [1] it is possible to limit to number of
dynamically learned fdb entries per bridge.
Add support to the systemd netdev bridge for the new netlink attribute
IFLA_BR_FDB_MAX_LEARNED.
[1] https://lore.kernel.org/all/20231016-fdb_limit-v5-0-32cddff87758@avm.de/
Signed-off-by: Gregor Herburger <gregor.herburger@ew.tq-group.com>
This is a lot of stuff, and sometimes quite wild, let's turn this into
its own header.
All stuff color-related that just generates sequences is now in
ansi-color.h (no .c file!), and everything more complex that
probes/ineracts with terminals remains in termina-util.[ch]
This is a common case, and nothing noteworthy at all. For example, if we
establish an enumerator for listing all devices tagged by some tag, then
the per-tag dir is not going to exist if there are currently no devices
tagged that way, but that's a really common case, and doesn't really
deserve any mention, not even at debug level.
Our usual rule is that we are more lenient towards misuse for public
users of our code than for ourselves. Or in other words: when validating
parameters of our public functions (those starting with sd_…) we prefer
assert_ret() over assert().
It's time. sd-json was already done earlier in this cycle, let's now
make sd-varlink public too.
This is mostly just a search/replace job of epical proportions.
I left some functions internal (mostly IDL handling), and I turned some
static inline calls into regular calls.
Our variables for internal libraries are named 'libfoo' for the shared lib
variant, and 'libfoo_static' for the static lib variant. The only exception was
libbasic, because we didn't have a shared variant for it. But let's rename it
for consitency. This makes the build config easier to understand.
Use the new function `hashmap_ensure_replace()` where we're using
`hashmap_ensure_allocated()` and `hashmap_ensure_replace()`.
Signed-off-by: Matteo Croce <teknoraver@meta.com>
It's the same old story: 'struct sd_bus' is generally instantiated once, so
bitfields, for which we pay with more complicated code in all users of this
struct, are counterproductive. In some progs the structure may be instantiated
a few times, but it's still not worth it because we save a few bytes of memory
in one place and pay for this with many more bytes in the code.
$ size build/libsystemd.so.0.39.0{.orig,}
text data bss dec hex filename
2452757 65376 3768 2521901 267b2d build/libsystemd.so.0.39.0.orig
2451669 65376 3768 2520813 2676ed build/libsystemd.so.0.39.0
$ diff -u <(pahole build/libsystemd.so.0.39.0.orig) <(pahole build/libsystemd.so.0.39.0)
...
- /* size: 1960, cachelines: 31, members: 105 */
- /* sum members: 1944, holes: 3, sum holes: 9 */
- /* sum bitfield members: 25 bits, bit holes: 2, sum bit holes: 31 bits */
+ /* size: 1984, cachelines: 31, members: 105 */
+ /* sum members: 1971, holes: 4, sum holes: 13 */
/* member types with holes: 1, total: 1 */
i.e. 2452757 - 2451669 = 1088 extra bytes of code and slower execution, to save
24 bytes of memory per instance of the struct. (But the number of cachelines
doesn't change, so the smaller struct most likely has no effect on memory
access, and the alignment of the struct most likely means that the memory
saving is illusory too, we just end up with a few bytes of padding after the
struct.)
In the other structs, the alignment prevent the bitfield for having any effect
on memory use, but the compiler would still generate more complicated code,
i.e. we pay something for nothing.
For example:
$ diff -u <(pahole build/libsystemd.so.0.39.0.orig) <(pahole build/libsystemd.so.0.39.0)
...
struct node_callback {
struct node * node; /* 0 8 */
- _Bool is_fallback:1; /* 8: 0 1 */
+ _Bool is_fallback; /* 8 1 */
- /* XXX 7 bits hole, try to pack */
/* XXX 3 bytes hole, try to pack */
unsigned int last_iteration; /* 12 4 */
@@ -455,15 +448,13 @@
struct node_callback * callbacks_prev; /* 32 8 */
/* size: 40, cachelines: 1, members: 6 */
- /* sum members: 36, holes: 1, sum holes: 3 */
- /* sum bitfield members: 1 bits, bit holes: 1, sum bit holes: 7 bits */
+ /* sum members: 37, holes: 1, sum holes: 3 */
/* last cacheline: 40 bytes */
};
I kept the bitfield in sd_bus_slot because it prevents the struct from growing
from 112 to 120 bytes by reducing the alignment requirement for subsequent
fields, and we potentially can have this instantiated many times.
Yu Watanabe
Daan De Meyer
Luca Boccassi
David Tardon
Gregor Herburger
Lennart Poettering
Mike Yuan
Zbigniew Jędrzejewski-Szmek
Matteo Croce