Two races could prevent DHCP-learned default routes from being installed
at boot:
1. The signal from the DHCP client script could be lost leaving conf.d
updated but frr.conf stale.
2. Even when the signal was received, 'vtysh -b' could fail because the
FRR daemon chain (mgmtd→zebra→staticd) writes PID files before being
fully operational, causing netd to give up with no retry.
Fix both by refactoring netd to use libev:
- Use an inotify watch of CONF_DIR, so netd reacts directly to file
changes without depending on signal delivery.
- On backend_apply() failure, schedule a retry in 5s and call pidfile()
unconditionally so dependent services are not blocked waiting for the
Finit condition 'pid/netd' to be satisfied.
We also take this opportunity to rename /etc/netd/conf.d/ to /etc/net.d/
The extra /etc/netd/ directory level served no purpose — nothing else
lives there. Flatten to /etc/net.d/ which reads more naturally as the
drop-in directory for network configuration snippets.
Also, reduce logging a bit since each netd backend already logs success
or fail which is sufficient to know that a configuration change has been
applied or not.
Finally, with the new inotify processing in netd it's redundant to call
'initctl reload netd' from the udhcpc script, in fact it will only cause
unnecessary overhead, so we drop it.
Fixes#1438
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Add a must expression to ensure users do not set static neighbors when
the interface type is not point-to-multipoint or non-broadcast.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The MVEBU SafeXcel Crypto Engine firmware (linux-firmware minifw) was
originally added in 70c12c3 to the generic aarch64 defconfigs to silence
kernel probe failures on Marvell Armada SoCs (37xx, 7k, 8k, CN913x):
crypto-safexcel f2800000.crypto: Firmware load failed.
crypto-safexcel f2800000.crypto: HW init failed (-2)
It was then accidentally dropped in 0e2d12e (kernel upgrade to 6.18),
which rebased on a tree that predated the firmware addition.
Rather than restoring it to the generic defconfigs, move it to the two
Marvell board Config.in files where it actually belongs — consistent
with how RTL8169 firmware was handled for the Raspberry Pi CM4 IoT
Router Board Mini in 68313773.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
A legacy name length limit in firewalld triggered problems with longer
policy names. This patch to firewalld lifts that limit by checking the
backend in use, no limit for nftables.
Fixes#1389
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Increase Zebra NETLINK buffer to fix reported issues with zebra being
out of sync with kernel routes and interface changes.
Also, try '-a' option to allow other processes to delete zebra routes.
Other changes in this commit:
- Minor fixes to Finit service files for consistency
- Empty daemon.conf stub files to silence bogus mgmtd errors at startup
- Relocate all skeleton files from board/common to separate from files
that are actual Infix additions and what's package system integration
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This is a follow-up to c559b8f2 (new-frr) where netd now is responsible
for funneling static routes, as well as routes from DHCP clients, into
the kernel via Frr.
For this to work, in particular at boot, any DHCP client must wait for
netd to have started or any default routes learned will be lost.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
DNS does not allow CNAME records at an apex domain (e.g. kernelkit.org),
so GitHub Pages cannot reliably redirect kernelkit.github.io to it. A
www subdomain can carry a CNAME pointing directly to kernelkit.github.io,
giving GitHub full control over the mapping and making the redirect work.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The weird `if env | grep -q 'reset-volumes'` construct is due to Klish
not sanitizing names before constructing variables, which leads to env
variables with '-' in them, not allowed in POSIX sh.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
curl's SCP/SFTP backend (libssh2) does not support interactive password
prompting — it fails with an authentication error unless credentials are
supplied non-interactively.
Replace the curl path for SSH URIs with native scp(1), which inherits
the TTY and prompts the user for a password as expected. Both scp://
and sftp:// URIs are handled by parsing the URI into a [user@]host:path
argument. Port is extracted if present and passed via -P. If no user
is embedded in the URI, the -u flag value (remote_user) is used instead.
Host-key verification is skipped with -o StrictHostKeyChecking=no, the
same trade-off already made by the test framework.
When the destination URI ends with '/', a meaningful remote filename is
derived from the source: the datastore's on-disk filename (e.g.,
startup-config.cfg), the datastore short-name for in-memory stores, or
the source file's own basename for plain file copies.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Add input validation patterns to string-type leaves in container,
firewall, DHCP, and interface YANG models. Also use mkdtemp() for
temporary key files in keystore.
Signed-off-by: Mattias Walström <lazzer@gmail.com>
Any exception raised by fn() propagated immediately out of the retry
loop, effectively making until() a single-shot call the moment any
transient error occurred.
Fix by wrapping fn() in a try/except inside the loop and treating any
exception as a "not yet" result. The last exception is preserved and
re-raised if all attempts are exhausted, so failure output is still
meaningful. Also fix a missing newline at end of file.
Fixes#1403
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Python's urllib.request wraps OSError in URLError during the connection
phase, but NOT during response reading (getresponse()). This means a
timeout that fires after the TCP handshake completes, e.g. when a service
accepts the connection but hasn't sent a response yet, propagates as a
bare TimeoutError that bypasses the exception handler, killing the entire
until() retry loop on the first attempt.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Rename BR2_PACKAGE_NETD_FRR to BR2_PACKAGE_NETD_FRR_GRPC to match the
naming convention of the other FRR backends (VTYSH, CONF) and make the
intent self-evident.
Have the gRPC backend select BR2_PACKAGE_FRR_GRPC so FRR is built with
gRPC support when — and only when — this backend is chosen.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This is a simpler approach than both grpc and watchfrr. Since we only
really need dynamic handling of static routes (for DHCP), we can track
changes made and maintain a "diff engine" for the resulting frr.conf.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Not all systems are equally quick, let's give it while to set before
bailing out because we test too soon.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
With the latest backported fixes to Finit, we no longer need to remove
/run/netd.pid to trigger a reload (restart) of staticd.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This includes adding a netd-daemon since FRR has gone towards all
daemons should be configured from mgmtd using YANG, netd uses gRPC to
reconfigure in mgmtd
Signed-off-by: Enable grpc
This is a simple two port router board, with the same
WiFi chips as BPI-r3.
The bootloader shall be the same as the BPI-r3, the bootloader
now probes which of the two is booting by checking for the PHYS
that only exist on BPI-r3-mini.
This patch adds a background gps monitor to statd because the gpspipe
program, normally used to interface with gpsd, slows down access to the
operational datastore with *seconds*.
This background monitor is not load bearing for how chrony accesses the
gps + nmea information from gpsd, this is handled separately in SHM.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Enable gpsd and override package defaults to integrate in Infix.
- Sync with rules disabled downstream in Debian project to prevent
clashes with common vendor/device ids used for other device types
- Start gpsd instance with '-n' to ensure /dev/shm is udpated for chrony
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Add documentation noting that at least one side of a VETH pair must
remain in the host namespace. Both ends cannot be assigned to different
containers.
Updates:
- container.md: Added IMPORTANT note in Container Host Interface section
- infix-if-veth.yang: Added note in module description
- infix-if-container.yang: Added note in host identity description
Fixes: #947
Related: #941
Co-authored-by: troglobit <183517+troglobit@users.noreply.github.com>
Enable IPv6 forwarding on the left DUT's tunnel interface, matching
the IPv4 tunnel config. Without it, IPv6 packets are not forwarded
through the tunnel, causing all tunnel_basic IPv6 tests to fail.
Regression introduced when fixing issue #515 in 717c1c0.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Add Raspberry Pi 2B to the bootloader build matrix and raspberrypi-rpi2
to the SD card image builder, enabling full RPi 2B image creation.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Catch ARM 32-bit build regressions early by adding arm targets
to the familty of Infix image and release builds.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This update brings new support for editing cleartext-symmetric-keys with
the 'change' command. The cleartext-symmetric-key is of type binary and
its use differs between different key-formats.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Infix key format identities now derive from the IETF bases, so the standard
identityref accepts them without any need for deviations. This also preserves
the nacm default rules, which were inadvertently dropped before.
Also rename format identities for generality:
- wifi-preshared-key-format -> renamed: 'passphrase-key-format'
- wireguard-symmetric-key-format -> use IETF 'octet-string-key-format'
Since cleartext-symmetric-key is now type binary (base64-encoded), the
WiFi backends (station and AP) decode values before passing them to
wpa_supplicant and hostapd.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Environment variables are stored in a separate .env file from the main
container script. The config SHA256 checksum only covered the script
file, so changes to env vars went undetected -- the container was not
recreated and kept running with stale values.
Include the env file contents in the config checksum calculation so that
any change to environment variables triggers a proper container restart.
Fixes#1313
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Type a command, regret you typed so much before checking something else,
comment it out for later with Meta-# (Alt-Shift-3). The whole line is
now commented out and saved to history. Recall it later with Ctrl-P or
Up arrow, delete the leading # with Ctrl-D, and run it.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Introduced in Linux 6.17, the force_forwding flag corresponds to the
ipv4 forwarding flag, which maps perfectly to the ietf-ip.yang model.
Fixes#515
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
All build scripts use $BR2_ARCH to determine location of critical board files
and configuration. As per previous decisions, the BR2_ARCH variable controls
all naming in Infix and for consistency the defconfig files are also renamed.
This fixes the issue with missing DTB file(s) for the Raspberry Pi 2B.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- Bump toolchain to match other build targets GCC 13 -> 14
- Drop post-build.sh to match other build targets
- Add features to arm_defconfig, all except containers
- Enable chck hashes
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- Fix locations of packages to match 'make update-defconfig' order
- Fix defconfig syntax for watchdogd tempmon ...
- Enable check hashes on all defconfigs
- Add missing wireguard and wifi support to RiscV
- Rename dongle support for X86_64 build
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The RTL8169 firmware was specifically added for the Raspberry Pi Compute
Module 4 IoT Router Board Mini, it does not need to be part of defconfig.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Fixes a warning from fakeroot by adjusting the type in the permission
table. Apparently the chown + chmod operations were always done even
though the type was wrong.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
RSSI is a vendor-specific relative index (0-255), while dBm is an
absolute power measurement. Since we report dBm values from iw,
rename the leaf to signal-strength for accuracy.
Also, add a check() function with a user-friendly reminder if the forget
to include the model prefix in the URL.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
For some reason, during the last refactor possibly, the dhcp client
stopped sending the current hostname to the dhcp server. This is very
useful to check what devices are on your network by asking the home
router's dhcp server.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Also, gcc complains that ap_op may be used unset, so let's set it to the
only safe default there is.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The trick to starting the scanner mode is to have a dummy network that
cannot associate with anything.
Fixes#1193
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Was set to informational, and Zebra to debug, which is mostly useful for
developers when debugging the integration of Frr daemons.
Example of before this patch:
Nov 4 13:44:09 infix-6f-05-e1 finit[1]: staticd[4366], stopping, sending SIGTERM ...
Nov 4 13:44:09 infix-6f-05-e1 staticd[4366]: Terminating on signal
Nov 4 13:44:09 infix-6f-05-e1 zebra[4221]: [N5M5Y-J5BPG][EC 4043309121] Client 'static' (session id 0) encountered an error and is shutting down.
Nov 4 13:44:09 infix-6f-05-e1 zebra[4221]: [KQB7H-NPVW9] zebra/zebra_ptm.c:1333 failed to find process pid registration
Nov 4 13:44:09 infix-6f-05-e1 zebra[4221]: [JPSA8-5KYEA] client 26 disconnected 1 static routes removed from the rib
Nov 4 13:44:09 infix-6f-05-e1 zebra[4221]: [S929C-NZR3N] client 26 disconnected 0 static nhgs removed from the rib
Nov 4 13:44:09 infix-6f-05-e1 finit[1]: Starting staticd[5529]
Nov 4 13:44:09 infix-6f-05-e1 zebra[4221]: [V98V0-MTWPF] client 26 says hello and bids fair to announce only static routes vrf=0
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
We want the command to show the entire log file, but fast forward to and
tail at the end, handling any file rotations from syslogd.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Learned one thing, a setting not set, is a safe setting.
bss_color made a lot of iPhones failed to connect.
Remove some more options with unknown impact.
The networking guide has grown out of control for quite some time, it is
time to split it up into multiple files. This will hopefully also make
it a bit more accessible to users.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- Add missing SVG For dhcpv6 example
- Fix link to static routes in WireGuard doc
- Add WireGuard to a new VPN sub-section in nav bar
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Ensure umask is properly managed around file creation
to maintain consistent behavior across the codebase.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
If we time out waiting for a wifi dongle, or onboard chipset, instead of
causing fatal error in dagger and b0rking startup, we cheat and create a
dummy interface.
Also, fix bug in HEAD, the condition was == wifi_ap but the comment said
"if not station or scanning mode". Since wpa_supplicant is for station
mode (not AP), the cleanup should happen when != wifi_ap.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- Simplify, prefer readability
- Fix nftw() 'nopenfd' argument
- Constify
The most interesting change is the argument to nftw(), 0 -> 20, to
ensure it is allowed to open enough file descriptors to descend.
In POSIX it is undefined if nopenfd is < 1.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Also, spruce up the table a bit. Allow it to resize, let the SSID column
be flexible, and right-align column(s) with numbers.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Extend the WiFi station mode with rx-speed and tx-speed leaves,
mirroring what already exists for AP mode connected stations.
Display bitrates in 'show interface <wifi>' detailed view.
Also standardize speed units to kbps/Mbps across the YANG model.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
When fdopen() succeeds, the FILE stream takes ownership of the file
descriptor. Calling fclose() automatically closes the underlying fd, so
the subsequent close(fd) calls were double-close bugs that could cause
undefined behavior.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Simplify USB device debugging by allowing USB core to always announce
details of every new USB device to syslog.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
When the shell is blocked on sleep, signal handlers don't execute until
the command completes. This caused the script to ignore SIGTERM at
shutdown, forcing Finit to wait and eventually SIGKILL it instead.
Fix by running sleep in the background and using wait, which is
interruptible by signals.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This reverts commit bf1eca0567.
This mess up order if having more than 1 phy, have not been
reproduce the error that got me to introduce this fix from
the beginning.
This is obsolete. By remove this we expose radio as mandatory.
This since a Wi-Fi interface is not possible to create without
a radio to connect it to.
This is run per interface, not per radio so its place
in hardware.c was wrong. This also opens up for multiple
station interfaces per radio, but not for now that it locked
down in a must expression for now.
Then hostapd needs to be restarted, handle this by adding the wifi container
to the diff if a custom-phys-address has been changed on a Wi-Fi interface.
Ensure the 9pm report tool runs inside the container so that it has
permissions to write to the test/.log directory. Then mark the log
directory as 777 when done so the ghmoon 'git clean -ffdx' can clean
it up when switching between jobs.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Need to verify that client hostnames have been set before starting the
client, otherwise the test will fail due to clients getting pool lease,
which is intended.
In a real-world scenario this is not a problem. Here we've booby trapped
the server to try to trip up errors.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Used by the new NACM basic test to verify ACL rules on explicit XPaths
or modules. For this we cannot rely on put_config_dict() since it now
uses the candidate datastore for all changes.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Switch RESTCONF from PUT to PATCH for configuration updates. This gives
RESTCONF the same merge behavior as NETCONF edit-config, allowing NACM
to correctly enforce path-specific permissions instead of requiring
write access to the entire datastore.
Fixes#617
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Otherwise the file will not be updated by confd on datastore copy:
Jan 12 14:26:37 foo confd[3410]: Overwrite existing file /cfg/startup-config.cfg (y/N)? <FF>
Jan 12 14:26:37 foo confd[3410]: Error: OK, aborting.:Inappropriate ioctl for device
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- relocate src/show.py to src/bin/show/
- Refactor container() to use run_sysrepo() instead of rolling its own
- Replace sysrepocfg with copy which supports nacm
- Rename run_sysrepocfg() -> get_json()
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Since sysrepocfg does not do any NACM based on the UNIX user, we expand
the scope of the copy tool slightly to allow outputting JSON to stdout.
This allows us to replace all sysrepocfg commands in the CLI used to
show runnining/startup/factory.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Previously the sys-cli group was for interactive shell access, but with
ever changing requirements this split has become necessary.
This commit introduces the 'sysrepo' group for low-level access to all
sysrepo commands, i.e., bootstrap only. For user-level shell access a
'klish' group is added which allows users to connect to the CLI. This
is now the only group users, including the default 'admin', are members
of, effectively making the new 'copy' tool the norm.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This patch enables umask and nacm defalts in the Buildoot sysrepo
package and a patch to sysrepo to set the sysrepo group also on the
event pipes in /etc/sysrepo/, without which it would be hard to inject
any type of new event as non-root user now since the umask now prevent
all 'other' users from interacting with sysrepo.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
We could go with the old default sha512crypt, but since the default has
changed to yescrypt, as used by the 'change password' command. We use
that for consistency.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- The new SocketGroup setting allows for configurable /run/klishd.sock
- Also a bug fix for async commands, needed to fix regressions in the
klish-plugin-sysrepo commands 'change passwordk' and 'text-editor'
after the latest upgrade to sysrepo + libyang v4
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The sysrepocfg tool does not set up nacm based on $USER, and we now have
the shell 'copy' tool which does. So drop 'cfg' and while we're at it
we also drop the 'edit' alias.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The wi-fi architecture in infix require udev to run and remove the kernel automatic wlan-interface
this. Unclear reason, but it seems that udev sometimes miss the add event.
This is a decision right now, that we support multi ssid AP, but
only one station interface, this is the only thing the c-code support
right now, so only allow this in the yang validation.
Service descriptions or command arguments may contain special characters
that need to be escaped in JSON strings.
Example:
{
...
"command": "udhcpc -f -p /run/dhcp-client-wan.pid -t 3 -T 5 -A 30 -a1000 -S -R -o -O 1 -O 3 -O 6 -O 12 -O 15 -O 28 -O 42 -i wan -V "NanoPi R2S"",
...
}
Also, wrap add_services() in a try/except instead of testing for keys.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The table_width() class method was lost in b826bcb9d, so we take the
opportunit to refactor to SimpleTable.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Verified with latest Finit 4.15 that reboot after resize works.
Added a comment to the board-specific .mk file about the broken
WDT on the RK3328 SoC, or rather the NanoPi R2S board.
Fixes#1292
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Board probing cannot copy board specific files to /usr (read-only), so
let's share this script with all other aarch64 boards.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
When interfaces were moved from firewall zones to become bridge or LAG
ports (e.g., sfp1/sfp2 moved from WAN zone to lan-br), the firewall
configuration was not regenerated. This caused stale entries in
/etc/firewalld/zones/*.xml where interfaces remained listed in their
old zones despite no longer being L3 interfaces.
Root cause: firewall_change() only triggered on firewall model changes,
but interface membership changes (bridge-port/lag-port) occur in the
ietf-interfaces model. When interfaces become member ports, they
transition from L3 to L2, which affects the result of
interfaces_get_all_l3() used for default zone assignment.
Fix: Expand the diff check to also trigger firewall regeneration when
bridge-port or lag-port configuration changes, ensuring firewall zones
stay synchronized with actual L3 interface topology.
Fixes#1345
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
When users modified firewall zone services (e.g., enabling SSH on WAN),
firewalld.conf would disappear from /etc/firewalld/, causing firewalld
to fail with "Failed to load '/etc/firewalld/firewalld.conf': [Errno 2]
No such file or directory".
Root cause: After switching to the "roll in" staging directory approach
late in development, firewalld.conf generation was still conditional on
changes to "default" or "logging" settings. When other changes (zones,
services, policies) were made:
1. Code created /etc/firewalld+/ staging directory
2. Skipped generating firewalld.conf (no default/logging in diff)
3. On commit: rm -rf /etc/firewalld (deleted firewalld.conf!)
4. Renamed /etc/firewalld+ → /etc/firewalld (incomplete config)
Fix: Always generate firewalld.conf whenever firewall configuration is
being staged, not just for default/logging changes.
Fixes#1346
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Replace BusyBox udhcpc6 with OpenWrt's odhcp6c for improved DHCPv6
functionality and better integration with Router Advertisements (RA).
The primary motivation is support for a common ISP deployment scenario
where IPv6 addresses are assigned via SLAAC (from Router Advertisements)
and DHCPv6 is used in stateless/information-only mode to provide DNS
servers and other configuration options. This hybrid RA+DHCPv6 setup is
standard practice for many ISPs but was not supported by udhcpc6, which
treats RA and DHCPv6 as separate, non-integrated mechanisms.
Additional benefits of odhcp6c:
- Better IPv6 Prefix Delegation (IA-PD) support with proper handling
of delegated prefix lifetimes and renewal
- Native integration of RA-provided configuration (DNS servers, routes,
addresses) with DHCPv6-provided options
- Support for stateless DHCPv6 via information-only mode
To verify stateless DHCPv6 integration with SLAAC addresses - the ISP
scenario that motivated this migration, a new test case has been added.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Almost full support for WireGuard
admin@server:/> show interface wg0
name : wg0
type : wireguard
index : 10
mtu : 1420
operational status : up
ipv4 addresses : 10.0.0.1/24 (static)
ipv6 addresses : fd00::1/64 (static)
peers : 2
Peer 1:
public key : ROaZyvJc5DzA2XUAAeTj2YlwDsy2w0lr3t+rWj2imAk=
status : UP
endpoint : 192.168.10.2:51821
latest handshake : 2025-12-09T22:51:38+00:00
transfer tx : 1412 bytes
transfer rx : 1324 bytes
Peer 2:
public key : Om9CPLYdK3l93GauKrq5WXo/gbcD+1CeqFpobRLLkB4=
status : UP
endpoint : 2001:db8:3c4d:20::2:51822
latest handshake : 2025-12-09T22:51:38+00:00
transfer tx : 1812 bytes
transfer rx : 428 bytes
in-octets : 1752
out-octets : 3224
admin@server:/>
This is described on the internet as:
"If you are unlucky and have a device with wrong MAC address,
too bad"
For us: We set mac addresses static, and do not use hostapds, insane
MAC address allocation (radio MAC + n).
Upstream sysrepo commit 90ce1f7 modified sr_nacm_init() to explicitly switch
the session to SR_DS_RUNNING before initializing NACM. This broke our klish
workflow which started sessions on SR_DS_CANDIDATE.
This commit drops the sysrepo patch that reverts that behavior in favor of
explicitly starting sessions on SR_DS_RUNNING for NACM initialization, then
switch to SR_DS_CANDIDATE for editing operations. This works with sysrepo's
behavior: NACM reads rules from running datastore while our editing workflow
operates on candidate datastore.
Also, add prompt path shortening for deep configuration hierarchies.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
If running outside container find the netns inside the container, then
enter that netns.
lazzer@tollan ~/src/github.com/kernelkit/infix3 (add-wireguard)$ sudo ./test/nsenter d2b
(netns:d2b) #
Example usage:
./src/statd/python/yanger/yanger -p wifi0 -x "ixll -A ssh enp195s0f0 sudo" ietf-hardware | ./board/common/rootfs/usr/libexec/infix/wifi-channel-map.py
Show the congestion in the 2.4 band.
Signed-off-by: Mattias Walström <lazzer@gmail.com>
* Add support for AP (list connected stations)
* Add scan-mode (Scan without create a fully configured station)
Signed-off-by: Mattias Walström <lazzer@gmail.com>
The original 'show firewall log' just did the same as the 'show log firewall'
command. This new implementation allows showing pretty-printed firewall log
which is admittedly easier on the eyes.
Also, add 'show firewall matrix' to just show the zone matrix overview.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Leverage the operational data to present a more human-friendly view of
containers in the CLI. Replacing the podman script wrapper.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
The constaint was supposed to be 50% of one CPU core, or 500 millicores
as per Kubernetes nomenclature.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This patch adds support for flexible columns. When a column is marked
as flexible it means it can be stretched when printing the table to an
optional minimal table width. Multiple columns can be marked and when
this occurs the padding is applied equally to all columns.
Also, make sure to only add 2 char padding between columns, not always
at the end, or we will require larger terminal than necessary.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Finit is capable of setting up all basic file systems, even /run but
Infix relies on it allowing executing scripts (dagger and container)
so leave that for now.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
This is most noticeable in interface context
admin@infix-00-00-00:/config/interface/e1/> set type wifi
admin@infix-00-00-00:/config/interface/e1/> set
bind-ni-name bridge-port container-network custom-phys-address description enabled ethernet
where you previously got *all* containers for all interface-types.
It changes the datastore to running when init nacm, this
make klish misbehave:
Klish open candidate -> init nacm -> now all changes are made in running
Schema node "type" for parent "/ietf-routing:routing/control-plane-protocols/control-plane-protocol/ietf-ospf:ospf" not found;
in expr "derived-from-or-self(../../rt:type"
The cfg.ext4 disk image was created but not formatted because mkfs.ext4
wasn't found. As sbin is not in PATH for an unprivileged user in
Debian.
This caused the VM boot to fail with "No persistent storage found for /cfg".
Signed-off-by: Richard Alpe <richard@bit42.se>
Attempt to fix permissions with sudo before falling back to $HOME,
and verify directory is actually writable before proceeding.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Actually disable iitod on Styx DCP-SC-28p to work around #670, this
prohibits software control of LEDs, leaving the default HW control,
which has proven more stable on this platform.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
- Remove fixed lengths from Column class
- Add dynamic width calculation to SimpleTable
- Update service table to use new batched printing
Signed-off-by: Richard Alpe <richard@bit42.se>
Add resource-limit and resource-usage containers to YANG model. Podman,
and later conmon, enforce CPU and memory limits in a delegated cgroupsv2
hierarchy managed by Finit.
Resource usage is queried from 'podman inspect', which has more nodes
than what is currently possible to limit.
Requires Finit 4.15, or later.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
Replace manual f-string formatting with SimpleTable/Column classes.
This new "framework" handles ANSI colors and padding. Removing the
hassle of manually calculating padding. You simply specify the header
with max number of chars the data can be and if you want left/right
padding and the "framework" calculates the padding for you.
We use this new "framework" to pretty print the newly added services
statistics.
Signed-off-by: Richard Alpe <richard@bit42.se>
Add statistics container to service model with memory usage, uptime,
and restart count tracking.
Updates YANG revision to 2025-12-02.
Signed-off-by: Richard Alpe <richard@bit42.se>
This update got pretty messy, since libyang had some
breaking changes. And unfortunatly rousette, libyang-cpp
and sysrepo-cpp has not made any new release yet.
We previously run a draft version of a lot of YANG models,
netconf-server,ssh-server and mode. Now they are released
with minor changes.
infix-meta is an exception, here we had to remove the obsolete
marker of the node and value, since this is no longer allowed in
libyang.
Previously rousette did not install the yang models so we kept a copy
of them in confd. Now it is implemented in rousette, let them install
their models.
On platforms based on CN913{0,1,2}, collect all ECC related counters
so that we know if any correctable errors have been encountered since
the most recent boot.
Allow particular boards or platforms install hardware specific scripts
in /etc/support.d, via the usual /usr/share/product logic, which are
then run by /sbin/support and included in the resulting bundle.
This lets us collect hardware specific information, e.g., SoC-specific
error counters, without /sbin/support needing to know about all the
nitty gritty details.
Verify that if the system encounters a hard lockup, i.e., interrupts
are no longer being serviced, then the system's watchdog will
correctly reboot it.
Rev. A has a FLASH that Linux reconfigures to 4B-addressing, but must
be in 3B-addressing when accessed by the BootROM. Since it has no
reset input, this would not work on warm resets. Linux has a
workaround for this issue for the common case of a controlled reboot,
which this change enables on those boards (which are admittedly few
and far between, but still useful as testing vehicles).
(This won't take effect until Finit is updated to at least 4.15, which
switches over from watchdog- to reboot(2)- based reboot, by default)
By default, ensure that the target device accepts TCP connections on
the specified port before returning the device object. This aligns the
SSH implementation with {NET,REST}CONF.
This solves the issue where tests that use SSH commands early in their
execution would sometimes fail as the remote server had not started up
yet.
Store historical snapshots as compressed .json.gz files to reduce disk
usage. The operational.json file remains uncompressed for easy access.
Signed-off-by: Richard Alpe <richard@bit42.se>
Dump operational datastore to timestamped JSON snapshots every 5 minutes
(in /var/lib/statd/). The operational.json symlink always points to the
latest snapshot.
Implement hierarchical retention policy that keeps the first snapshot of
each time period (hour/day/week/month/year), providing fine-grained recent
history while preventing unbounded disk usage.
This will allow us to plot / track how the system state evolves as
well as give us somewhat fine-grained info in the case of an event,
such as a crash.
Add unit test simulating months of snapshots to verify retention
behavior using a statd stub that only runs the retention code locally
(unit test)
Signed-off-by: Richard Alpe <richard@bit42.se>
Turn on OOPS-to-panic, soft/hard lockup panic, hung-task panic, and
extra workqueue watchdog reporting. This makes latent stalls visible
instead of silently freezing, improving diagnosis of issues like the
recent resource-pressure lockup.
Signed-off-by: Joachim Wiberg <troglobit@gmail.com>
echo "> **Try Infix in GNS3!** Download the appliance from the [GNS3 Marketplace](https://gns3.com/marketplace/appliances/infix) to test Infix in a virtual network environment without hardware." >> release.md
if you want to play with low-level stuff like ethtool, you
might want to test the Intel 82545EM driver, e1000.
Note: ARM32 uses virtio-net-device (MMIO) by default.
Note: ARM32-bit uses virtio-net-device (MMIO) by default.
config QEMU_NET_BRIDGE_DEV
string "Bridge device"
Some files were not shown because too many files have changed in this diff
Show More
Reference in New Issue
Block a user
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
