Previously, only servers that statically configursd or obtained by
DHCPv4 protocol are saved in the manager state file.
NTP servers obtained by DHCPv6 could not be used by timesyncd.
Fixes#29148.
Some qdiscs (e.g. tbf) implicitly create class(es) on create.
Previously, we could not create any child qdisc under the class, as the
implicit class is tagged as foreign.
Some kind of qdisc implicitly creates a class for the qdisc, but the
created class is not notified by the kernel. So, we need to explicitly
enumerate classes after a qdisc is created.
When a node of traffic control tree is removed, all child nodes are also
removed but their removal are not notified by the kernel.
So, previously, removed TC classes or qdiscs under the removed node were
kept in the memory of networkd, and may cause failure on reconfigure.
Previously, if the input offset 'p' does not point to an entry object,
the function returns the next of the nearest entry object on
DIRECTION_DOWN, as generic_array_bisect() already returns the nearest
entry object.
If the first call of generic_array_bisect_plus_one() provides the same
offset, then it is not necessary to call the next one, as we already
know the entry object is also liked to the input data object.
Also, this make the function reuse the object returned by
generic_array_bisect_plus_one().
No functional change, just optimization.
Follow-up for ec50313d4e.
The function generic_array_bisect_plus_one() does not read any new data
objects, so the data object is still valid, and not necessary to re-read it.
Previously, if journal_file_set_offline() is called twice with 'wait = false',
the second call triggered segfaults, as the offline_state is OFFLINE_DONE,
and journal_file_set_offline_thread_join() tries to call pthread_join()
with NULL.
In sd-dhcp-client.c, we do not set the option in the DHCPDISCOVER
message when anonymized, and the specified address is ignored
anyway. So, this does not change the behavior, but suppress
misleading debugging log in dhcp4_set_request_address().
And MMapFileDescriptor always references MMapCache, hence it is not
necessary Window to have a reference to MMapCache.
This also renames the list fields 'by_fd' -> 'windows', to make them
consistent with the name of the head.
No functional change, just refactoring.
The post operations (attach the found or new window to a context, update
the keep_always flag, and calculate the address) in three steps are
equivalent, and the two searching logics are quite simple.
Let's merge them to decrease the total lines.
No functional change, just refactoring.
This reverts the following commits.
- a1640191b4
- 231741d355
These were done by my misunderstanding of the mmap cache behavior.
Also, this updates the comments added by
df04b9ed86.
In discover_next_boot(), first we find a new boot ID based on the value
stored in the entry object. Then, find the tail (or head when we are going
upwards) entry of the boot based on the _BOOT_ID= field data.
If boot IDs of an entry in the entry object and _BOOT_ID field data
are inconsistent, which may happen on corrupted journal, then previously
discover_next_boot() failed with -ENODATA.
This makes the function check if the two boot IDs in each entry are
consistent, and skip the entry if not.
Fixes the failure of `journalctl -b -1` for 'truncated' journal:
https://github.com/systemd/systemd/pull/29334#issuecomment-1736567951
When READ_FULL_FILE_UNBASE64 (or READ_FULL_FILE_UNHEX) is specified,
setting size argument by caller is difficult, as it is hard to estimate
the encoded length.
This makes when size is specified with decoding option, let's read file
more, and check decoded size later with the specified size.
On normal systems, triggering a timeout should be a bug in code or
configuration error, so I do not think we should extend the default
timeout. Also, we should not introduce a 'first class' configuration
option about that. But, making it configurable may be useful for cases
such that "an extremely highly utilized system (lots of OOM kills,
very high CPU utilization, etc)".
Closes#25441.
Then, timout_compare() becomes simplar, the timeout value becomes
consistent with what sd_netlink_get_timeout() provides.
This also drop unnecessary assignment of reply_callback.timeout after
the slot is dropped from the prioq.
Because journal_file_next_entry_for_data() provides the first entry, while
journal_file_next_entry() actually provides the next entry of the input,
this also renames it to journal_file_move_to_entry_for_data().
Also, previously, on DIRECTION_UP the function did not fall back to the
'extra' entry when all entries linked in the chained array are broken.
This also fixes the issue, and now it fall back to the extra entry.
When we reach an empty array, there are at least two possibilities:
- journal file is corrupted,
- invalid index is requested.
We cannot distinguish them here, let's simply return earlier.
If there's corruption and we are going upwards, then the 'total'
must be decreased when we go to the previous array. However,
previously, we wrongly kept or increased the number. This fixes
the behavior.
This effectively reverts d9b61db922.
In the do-while loop, we do not read any other entry array object, hence
the current object is always in the mmap cache and not necessary to re-read it.
Consider the following situation:
- There are two journal files (x and y), that contains entries for two boots (X and Y).
- The journal file x contains entries of the boot X, and y contains
entries of Y.
- Nevertheless x does not contains entries of boot Y, it contains
the _BOOT_ID= data object of boot Y. Of course, the data object is not
referenced by any entries in the journal file x.
In such situation, when the current location of sd_journal is the head
of journal y, that is, the first entry of the boot Y,
sd_journal_previous() failed without this change, and
'journalctl --boot -NUM' for boot X failed.
Fixes#29275.
Yu Watanabe
