Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

To avoid version mismatch.

Requires following commit from ppsi
wrs-arch: move ptpd_exports.h from wr-switch-sw repo as ppsi_exports.h

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

/etc/init.d/wr_date gets parameter "ppsi" to avoid redirect of the
output and setting the date from /update/saved_date if unable to get
the data via ntp.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Don't kill other ntpd instances.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

This fix solves the problem of jumps of a fraction of a second during
the startup. It is believed that the problem is in HDL, this commit is
just a workaround implemented in software.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Disable web interface in the default config due to security
vulnerabilities described in CVE-2023-22577.

Please note that web interface can still be enabled manually in the
config file in runtime.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Running fsm_generic_run in hal_port_state_fsm_init_all delays setting
shmemState as HAL_SHMEM_STATE_INITITALIZED, which causes the timeout
(and exit) at the startup of wrsw_rtud. wrsw_rtud is restarted by
monit. However, wrs_vlans is run only once by init and is not able to setup
correctly VLAN configurationi when wrsw_rtud is not running.
The issue is related to the suspected bug #261

.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Reduce the possible range in Kconfig to the half of the possible range.
E.g. setting job to be run every 40minutes is not trivial and */40 will
not work. Instead it will run the job at minute 0, then 40, then again
at 0 (after 40min, then after 20min).

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

And fix two typos in the script itself

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Before the fix in some rare cases, radiusvlan could report a receive
of frames from another interface. This could lead to a wrong
authentication of a device connected to a port.

The problem is caused by a behaviour that if a socket is created it
listens on all interfaces until the bind function is called. It might
happen that later calls of recvfrom function will handle frames
received from other interfaces. To prevent that, radiusvlan flushes
all frames received before socket is bound to an interface.
The lost of frames is meaningless since all frames that radiusvlan
listens for are sent periodically.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Print exact command which is or would be run to change VLAN
configuration.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

In dry run mode no changes to the VLAN configuration are applied.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Remove superfluous spaces when generating Kconfig_port_timing.in by gen_kconfig_port_timing.sh

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

fix script generating Kconfig_port_timing.in to include change from
commit:
Kconfig: allow configuration of externalPortConfigurationEnabled in all timing modes

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Disabling of SFP's TX is possible only via HAL.
Disabling it I2C directly is a risky, it can corrupt SFP's EEPROM if
HAL is running.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

The SDBfs library has been re-licensed as LGPL v2.1 in order to allow it
to be linked to proprietary software.

The permission to re-license those files has been approved by CERN in
2018 and then in 2021 (and also by its main author).

Signed-off-by: Adam Wujek <adam.wujek@creotech.pl>

Signed-off-by: Adam Wujek <adam.wujek@creotech.pl>

After reset of the CPU (ARM) GPIO pins are set as input. During
configuration they were set to 0, which had a meaning to the i2c
transfer if the last transfer before reset was not completed (e.g. by
the switch restart).

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

The I2C transfer can be interrupted by a restart of a switch. An I2C slave
is not aware about an ARM restart so it expect the continuation of the
transfer.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

To reduce the CPU load the data from one SFP is read every 300ms. In the
worst case all monitoring data is updated every 5.4s.

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

This commit reduces the size of periodic read for each SFP from 256B to 16B.
Since the I2C transfer are bit-banged, this commit reduces CPU
utilization if monitoring of SFPs is used (CONFIG_READ_SFP_DIAG_ENABLE).

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

Signed-off-by: Adam Wujek <dev_public@wujek.eu>

To be reported by SNMP as SNMPv2-MIB::sysContact.0 and SNMPv2-MIB::sysLocation.0

Signed-off-by: Adam Wujek <dev_public@wujek.eu>