Systemd's LLDP client also internally tracks all neighbours, and it thus
already needs a maximum already. For systemd, that is currently 128.
We don't need to implement our own rate limiting on top of that,
because if we wouldn't trust the LLDP client to get this right,
it would be DoS-able already.
Also decrease the number of maximum neighbours from 4k to 128.
Note that already previously we wouldn't ever get more than 128
entries.
nm_keyfile_read() and nm_keyfile_write() will be public API.
As such, it must be flexible and extendible for future needs.
There is already the handler callback that fully solves this
(e.g. a future handler event could request whether a certain
behavior is enabled or not).
As additional possibility for future extension, add a flags
argument. Currently no flags are implemented.
From inside a callback 4 properties are potentially interesting
to all callbacks: the currenty group, key, setting and property-name.
Refactor the code to track these properties in NMKeyfileHandlerData
and distinguish between the property name and the keyfile key.
Setting the error on the callback does not work well from bindings.
Instead, let bindings call a (future) nm_keyfile_handler_data_fail_with_error()
function on the handler_data to indicate failure.
As the keyfile handler callback will become public API, it needs to be
usable via bindings. A plain void pointer is not usable. Instead, add
a new type that can be used via introspection.
This will become public API. The enum for read and write callback
serves very similar purposes. Merge them so that we have fewer
types in the public API.
6876b73ec113 all: rework logging to append logging messages to the event queue
65ca0e4bea32 client: check length of client id for n_dhcp4_client_config_set_client_id()
6a9ca10b7ad6 client: close timer file descriptor explicitly when EPOLL_CTL_ADD fails
e5f3b27f5301 connection: don't fail dispatch in case of receive errors
68790d5ec950 connection: keep trying after a failure in send()
788b03891cd0 n-dhcp4: fix logging macro
6cd0fd5e8b3f probe: request previous address after expiration
ef11c7079ad7 connection: avoid double free of NDhcp4Outgoing
922a9a2326e7 n-dhcp4: fix initialization of the 'secs' DHCP header field
7fff54117436 outgoing: fix uninitialized variable in n_dhcp4_outgoing_append()
27547faaf258 probe: move back to INIT after lease expires
d07cb4e32884 incoming: accept options that are longer than requested
34d365e2267f client: handle invalid return codes gracefully during n_dhcp4_client_dispatch()
2e8c6017f85b n-dhcp4: use packet socket in rebinding state
9a7c82143c06 n-dhcp4: support init-reboot state
e615e3f5b063 n-dhcp4/socket: use SO_REUSEADDR on UDP socket
8dbfa1ad2549 merge branch 'bengal:log'
6e50189fba86 n-dhcp4: log outgoing packets
aea0315d3862 n-dhcp4: log incoming packets
1ca64a82393d n-dhcp4: add logging API
72c16a0028ac client-probe: fix state transitions on timer dispatch
c023ae2ad829 merge branch 'leasetime'
780f94c333ca lease: add n_dhcp4_client_lease_get_basetime()
a94378274cf4 incoming: don't handle 0xFFFFFFFF timestamps special in n_dhcp4_incoming_query_u32()
6bb277aefe0b probe: unconditionally pass ownership of message in n_dhcp4_client_probe_dispatch_io()
1cf87c90c6a1 probe: fix leaking message during client probe
55239cef2d63 lease: remove unreachable code in n_dhcp4_incoming_get_timeouts()
fa5ee7903061 lease: expose the server IP address
0cfce4a4fe07 all: merge branch 'th/build-centos76'
cae4868e5f7e all: avoid {net,linux}/if.h clashes on old distros
05859a3e4d32 all: avoid c_min() macro to work with old GCC
0be7033dd940 probe: arm timers in bound state
8735cdb0f95f ci: switch to c-util automation
a0bb7c69a11a client: fallback to CLOCK_MONOTONIC for timerfd
308cb242caf7 ci: disable valgrind run
468c93195ad3 build: update submodules
9d0f690f9ee8 ci: switch to github actions
9ba143a037bb probe: allocate memory of right size in n_dhcp4_client_probe_option_new()
2546aa2c809d util/link: suppress gcc warning
e4a01f5870a6 ci: drop broken armv7hl
2e2fbfa18f2c build: update submodules
6277dfd20eca Merge pull request #2 from NetworkManager/th/declaration-after-statement
1b660ae2d12a outgoing: avoid "-Werror=declaration-after-statement" warning with static_assert
f34a54cd9719 client/probe: fix memory leak
21e68f3bba80 client/connection: fix memory leak
git-subtree-dir: shared/n-dhcp4
git-subtree-split: 6876b73ec113328712a5cdc2ffb9497fd774c0e2
../src/nm-manager.c: In function periodic_update_active_connection_timestamps:
../src/nm-manager.c:7358:43: error: t may be used uninitialized in this function [-Werror=maybe-uninitialized]
nm_settings_connection_update_timestamp (nm_active_connection_get_settings_connection (ac),
^
Fixes: 6f3ae8a563 ('core: in periodic_update_active_connection_timestamps() use same timestamp')
We don't need the mask argument. If the caller wants to check only for certain
flags, she can do that right away with
NM_FLAGS_ANY (nm_settings_connection_autoconnect_blocked_reason_get (sett_con), flags)
When the secrets for a connection are updated, unblock autoconnection
in case it had been blocked previously due to bad or no
secrets. Otherwise we would need to manually activate the connection
or restart NM to get another try with the new secrets.
As commit ccfe5fec8d ('wifi: don't autoconnect to networks that have
never been successful') explains, Wi-Fi profiles only autoconnected
if they had no or a positive timestamp.
The problem that tried to solve is when a user accidentally clicks on a
Wi-Fi network in applet. Then the created profile may not be usable
(because of wrong credentials). To avoid indefinitely to autoconnect,
uch a profile will have a timestamp of 0, which prevents further
auto activations.
However, I find that problematic.
An important use case is pre-deploying profiles. In that case, the user
cannot set the timestamp, because the timestamp cache
/var/lib/NetworkManager/timestamps is internal, undocumented API. Also,
ifcfg-rh doesn't support the timestamp and anyway, does the timestamp
of the NMConnection does not get honored (only the one from the
timestamps file).
Maybe that could be an alternative solution here, to allow the user to
mark profiles as "I really want it to autoconnect". But that seems
unnecessary and wrong to me.
The problem really is that the user cannot do anything to ensure that
autoconnect will work tomorrow (short of editing the timestamps
database). The problem is that the property of whether a profile
every connected successfully is not in direct control of the user (it
depends on external conditions).
If the user has bogus profiles configured, those profiles should be
deleted (or autoconnect disabled) and not keep autoconnect blocked.
Also note that if you are at home and accidentally click on your
neighbour's Wi-Fi network, then you presumably still also have a working
profile to your own network. That usable profile will have a more
recent timestamp and be preferred during autoconnect already.
This reverts commit ccfe5fec8d.
https://bugzilla.redhat.com/show_bug.cgi?id=1781253
Updating the timestamp marks the keyfile database as dirty. Avoid
that, if there is no change. Of course, nm_key_file_db_set_value()
itself already checks whether the are any changes, and does nothing
if there aren't.
Simply perform the check earlier, to do nothing.
When performing a synchronous action together (like iterating
over all settings and set the current timestamp), it's nicer
to pretend that all this would happen instantaneously. That means,
ensure we use the same timestamp throughout.
On a minor point, there really is no need to call time() multiple times.
Code like
»···»···if (strcmp (tag, "net.ifnames") == 0)
»···»···»···net_ifnames = strcmp (argument, "0") != 0;
is really hard to understand (at least to me). Compare to
»···»···if (nm_streq (tag, "net.ifnames"))
»···»···»···net_ifnames = !nm_streq (argument, "0");
The autoneg/speed ethtool settings are important. If they are wrong,
the device might not get any carrier. Having no carrier means that
you may be unable to activate a profile (because depending on
configuration, carrier is required to activate a profile).
Since activating profiles are the means to configure the link settings
in NetworkManager, and activating a profile can be hampered by wrong link
settings, it's important to reset the "correct" settings, when deactivating
a profile.
"Correct" in this case means to restore the settings that were present
before NM changed the settings. Presumably, these are the right once.
Beyond that, in the future it might make sense to support configuring
the default link settings per device. So that NM will always restore a
defined, configured, working state. The problem is that per-device
settings currently are only available via NetworkManager.conf, which
is rather inflexible.
Also, when you restart NetworkManager service, it leaves the interface
up but forgets the previous setting. That possibly could be fixed by
persisting the previous link state in /run. However, it's not
implemented yet.
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/issues/356https://bugzilla.redhat.com/show_bug.cgi?id=1807171
nm_platform_ethtool_init_ring() only has one caller. It's simpler to
drop the function and implement it at the only place where it is needed.
Maybe there could be a place for a function to initialize NMEthtoolRingState,
one option after the other. However, at the moment there is only one
user, so don't implement it.
This fixes various minor issues:
- the function had a NMPlatform argument, although the argument
is not used. Thus function merely operates on a NMEthtoolRingState
instance and shouldn't have a nm_platform_*() name.
- nm_platform_ethtool_init_ring() returned a boolean, but all
code paths (except assertion failures) returned success.
- as the function returned an error status, the caller was compelled
to handle an error that could never happen.
- the option was specified by name, although we already have a more
efficient way to express the option: the NMEthtoolID. Also, the
caller already needed to resolve the name to the NMEthtoolID, so
there was no need to again lookup the ID by name.
Imagine you have a veth device. That device supports certain offload features
(like "ethtool.feature-rx-checksum") but doesn't support any ring
options. Even trying to read the current ring settings will fail.
If you try to activate that profile, NMDevice previously would always
try to fetch the ring options and log a warning and extra debugging
messages:
<trace> [1590511552.3943] ethtool[31]: ETHTOOL_GRINGPARAM, v: failed: Operation not supported
<trace> [1590511552.3944] ethtool[31]: get-ring: failure getting ring settings
<warn> [1590511552.3944] device (v): ethtool: failure getting ring settings (cannot read)
It does so, although you didn't specify any ring settings and there
was no need to fetch the ring settings to begin with.
Avoid this extra logging by only fetching the ring option when they
are actually required.
Setting the kernel token is not strictly necessary as the IPv6 address
is generated in userspace by NetworkManager. However it is convenient
for users to see that the value set in the profile is also set in the
kernel, to confirm that everything is working as expected.
The kernel allows setting a token only when 'accept_ra' is 1:
temporarily flip it if necessary. Unfortunately this will also
generate an additional Router Solicitation from kernel, but this is
not a big issue.
When a team device is assumed, we skip stage1 and imply that teamd is
already running. If this doesn't happen (for example because teamd was
manually stopped or because the interface was created in the initrd),
the team interface will continue processing traffic but will not react
to changes in the environment (e.g. carrier changes). Ensure that
teamd is running for assumed devices.
Rework qdisc synchronization. The previous implementation added all
known qdiscs and removed unneeded ones from platform; this had some
problems:
- kernel doesn't allow to add (with exclusive flag) a qdisc if one
with the same parent already exists;
- if we use the replace flag instead of add, then it becomes possible
to add a new qdisc with the same parent of an existing one. However
if the existing qdisc is of the same kind, kernel will try to to
change() it, which fails for some qdiscs (e.g. sfq).
- kernel doesn't allow to delete a qdisc with handle of zero because
that is the default qdisc and can only be replaced;
Fix that.
Usually stage1 is skipped for external or assumed devices. Add a
mechanism to call stage1 for all devices, similarly to what was
already done for stage2.
When searching an element that is lower than the first list element (for
example RTNL type "batadv"), imax will be -1 after the last iteration.
Use int instead of unsigned to make the termination condition imin > imax
work in this case. This fixes NetworkManager crashing due to an
out-of-bounds array access whenever interfaces of such types exist.
Fixes: 19ad044359 ('platform: use binary search to lookup NMLinkType for rtnl_type')
https://gitlab.freedesktop.org/NetworkManager/NetworkManager/-/merge_requests/515
We are going to expose some of this API in libnm.
The name "gendata" (for "generic data") is not very suited. Instead,
call the public API nm_setting_option_*(). This also brings no naming
conflict, because currently no API exists with such naming.
Rename the internal API, so that it matches the API that we are going
to expose next.
We already have NMEthtoolID to handle coalesce options in a way that is
convenient programmatically. That is, we can iterate over all valid
coalesce options (it's just an integer) and use that in a more generic
way.
If NMEthtoolCoalesceState names all fields explicitly, we need explicit
code that names each coalesce option. Especially since NMEthtoolCoalesceState
is an internal and intermediate data structure, this is cumbersome
and unnecessary.
Thereby it also fixes the issue that nm_platform_ethtool_init_coalesce() has a
NMPlatform argument without actually needing it.
nm_platform_ethtool_init_coalesce() does not operate on a NMPlatform
instance, and should not have the appearance of being a method of
NMPlatform.
All other users name a similar variable "coalesce" (or "coalesce_new",
"coalesce_old"). Rename the variables, to don't use different names for
similar uses.
nm-device now applies ethtool ring settings during stage 2 "device
config" of the connection activation.
ring settings will be then restored (according to what the state
was before the connection got activated on the device) when the
connection is deactivated during the device cleanup.
One thing to be noted is that unset ring settings (in the profile)
will not be touched at all by NetworkManager so that if the NIC driver
sets some default values these will be preserved unless specifically
overridden by the connection profile.
https://bugzilla.redhat.com/show_bug.cgi?id=1614700
Only in one moment we need the old and requested settings together:
during _ethtool_coalesce_set(). But for that we shouldn't track both
states in "NMEthtoolCoalesceState".
Simplify "NMEthtoolCoalesceState" to only contain one set of options.
By tracking less state, the code becomes simpler, because you don't
need to wonder where the old and requested state is used.
