Flowire FCDC1/2 - Release Notes

Version 4.1.3.13 (2019-11-01)

Release Info

• Name: Zenitel-flowire 4.1.3.13 and onwards
• Description: zenitel-flowire 4.1.3.xx can be installed on FCDC1/2 and TFIX-x and EAPFX-x.
• Version:4.1.3.13
• Date: 01.11.2019
• Status: GA
• Compatibility: With this sw the FCDC1/2 are compatible with the new FCDC3
• Upgrade: See Flowire Software Upgrade Procedure

Files included in the package:

• jffs2-spc
• jffs-spc.md5
• linux-spc
• linux-spc.md5
• nvram218.bin
• u-boot218.bin

Bug fixes and improvements

• With this sw the FCDC1/2 are compatible with the new FCDC3.

Important information

IGMP settings

As long as the system is using a managed switch, the following applies:

• “IGMP querier” should be disabled in all Flowire units
• “IGMPv2 compatability” should never be enabled in the Flowire connected to the system switch
• “IGMPv2 compatability” should always be enabled in the Flowire connected to the field equipment

More information on how to configure the Flowire’s IGMP settings

CCo settings

For optimal availability in cluster configurations, use the CCo settings in Flowire web as follows:

• Set the Flowire connected to the system switch to “CCo”
• Set the Flowires connected to the field equipment to “Not CCo”

In Exigo A/B systems, clusters must be connected to either the A or the B system. Flowire does not support redundant connections to devices in a cluster.

For 1 to 1 Flowire connections with A/B redundancy, we recommend:

• Leave the Flowires connected to the system switches on “Auto”
• Set the Flowires connected to the field equipment to “Not CCo”

More information about Flowire’s CCo settings.

For more information on configuring Flowire in Exigo systems, see our Exigo Flowire Guidelines.

Version 4.1.3.12 (2019-02-07)

Summary

Software package: zenitel-flowire-4.1.3.12.zip

Files included in the package:

• jffs2-spc
• jffs-spc.md5
• linux-spc
• linux-spc.md5
• nvram219.bin
• u-boot218.bin

Bug fixes and improvements

Flowire disable DHCP was not working.

Disabling DHCP in the web did not work. Even if the web reports DHCP as disabled, DHCP was still active. (If a static IP address was configured, Flowire could still get a DHCP address alongside the static address). This has now been fixed and disable DHCP works as intended. (MTN-1996)

Also improved IP address list:

• Listing of static netmask
• Added "DHCP" to legend text for IP address/mask
• Repaired display when no IP address/mask set on interface

Three more configuration items are now displayed to the web front page

• Default gateway
• IGMPv2 compatibility
• Configured PLC Role

The IMT scan will now give more complete information. (MTN-2215)

Important information

IGMP settings

As long as the system is using a managed switch, the following applies:

• “IGMP querier” should be disabled in all Flowire units
• “IGMPv2 compatability” should never be enabled in the Flowire connected to the system switch
• “IGMPv2 compatability” should always be enabled in the Flowire connected to the field equipment

More information on how to configure the Flowire’s IGMP settings

CCo settings

For optimal availability in cluster configurations, use the CCo settings in Flowire web as follows:

• Set the Flowire connected to the system switch to “CCo”
• Set the Flowires connected to the field equipment to “Not CCo”

In Exigo A/B systems, clusters must be connected to either the A or the B system. Flowire does not support redundant connections to devices in a cluster.

For 1 to 1 Flowire connections with A/B redundancy, we recommend:

• Leave the Flowires connected to the system switches on “Auto”
• Set the Flowires connected to the field equipment to “Not CCo”

More information about Flowire’s CCo settings.

For more information on configuring Flowire in Exigo systems, see our Exigo Flowire Guidelines.

Version 4.1.3.11 (2017-12-06)

Summary

Software package: zenitel-flowire-4.1.3.11.zip

Files included in the package:

• jffs2-spc
• jffs-spc.md5
• linux-spc
• linux-spc.md5
• nvram219.bin
• u-boot218.bin

Issues Resolved

• After "restore to factory settings" (button pressed for 15s) it was not possible to change the static IP for the unit from the web interface. Fixed. (AL-2880)
• After system wide reset, Flowire was not able to synchronize time with the NTP server running on the AMC-IP card. This mainly affected logging. Fixed. (AL-3243)

Important information

IGMP settings

As long as the system is using a managed switch, the following applies:

• “IGMP querier” should be disabled in all Flowire units
• “IGMPv2 compatability” should never be enabled in the Flowire connected to the system switch
• “IGMPv2 compatability” should always be enabled in the Flowire connected to the field equipment

More information on how to configure the Flowire’s IGMP settings

CCo settings

For optimal availability in cluster configurations, use the CCo settings in Flowire web as follows:

• Set the Flowire connected to the system switch to “CCo”
• Set the Flowires connected to the field equipment to “Not CCo”

In Exigo A/B systems, clusters must be connected to either the A or the B system. Flowire does not support redundant connections to devices in a cluster.

For 1 to 1 Flowire connections with A/B redundancy, we recommend:

• Leave the Flowires connected to the system switches on “Auto”
• Set the Flowires connected to the field equipment to “Not CCo”

More information about Flowire’s CCo settings.

For more information on configuring Flowire in Exigo systems, see our Exigo Flowire Guidelines.

Version 4.1.3.10 (2017-04-07)

Summary

Software package: zenitel-flowire-4.1.3.10.zip

Files included in the package:

• jffs2-spc
• jffs-spc.md5
• linux-spc
• linux-spc.md5
• nvram219.bin
• u-boot218.bin

Improvements

• Configure CCo
  • Possible to set “CCo” (Master) and “Not CCo” (Slave)
  • Optimizes availability in cluster configurations
• Possible to configure IP net mask
• Added ntpclient for more accurate logs
• Added a reboot button to Flowire Web
• Improved IGMP/multicast handling

Issues Resolved

• Multicast handling in clusters (MTN-1120): Stability in cluster configurations has been significantly improved

Important information

IGMP settings

As long as the system is using a managed switch, the following applies:

• “IGMP querier” should be disabled in all Flowire units
• “IGMPv2 compatability” should never be enabled in the Flowire connected to the system switch
• “IGMPv2 compatability” should always be enabled in the Flowire connected to the field equipment

More information on how to configure the Flowire’s IGMP settings

CCo settings

For optimal availability in cluster configurations, use the CCo settings in Flowire web as follows:

• Set the Flowire connected to the system switch to “CCo”
• Set the Flowires connected to the field equipment to “Not CCo”

In Exigo A/B systems, clusters must be connected to either the A or the B system. Flowire does not support redundant connections to devices in a cluster.

For 1 to 1 Flowire connections with A/B redundancy, we recommend:

• Leave the Flowires connected to the system switches on “Auto”
• Set the Flowires connected to the field equipment to “Not CCo”

More information about Flowire’s CCo settings.

For more information on configuring Flowire in Exigo systems, see our Exigo Flowire Guidelines.

Version 4.1.3.9 (17.03.2016)

Multicast handling

• The IGMP querier application will respawn in case it crashes due to an unpredictable error or assert.
• Added a configuration parameter for the netlink socket - the "buffer overrun" error is now suppressed. This is used by the IGMP snooping application.

General improvements

• It is now possible to erase the ZapCppClient configuration from the device's web interface.
• Fixed spelling errors in the web interface.

Version 4.1.3.8 (01.03.2016)

BSP

• Fixed ZapCppClient monitoring application. Note: this is only relevant for EXIGO projects.

Version 4.1.3.7 (21.01.2016)

PLC driver

• Fixed an issue where the device would reboot because it would like to inform other Flowire devices of its Ethernet bridge (list of all MAC addresses it bridges). In case of a large network the message would overflow the DSP buffer and cause the device to crash. The bridge message is limited to 240 devices.

Version 4.1.3.6 (07.01.2016)

Multicast handling

• The IGMP querier source address is correctly set even if the DHCP is turned off

BSP

• Operational ZAP
• Stripped YANG model
• Operational ZapCppClient

PLC driver

• Fixed crashes due to device rapidly changing states from connected to disconnected
• Extra debugging information

Version 4.1.3.5 (18.12.2015)

Networking

• Fixed TCP timeouts

PLC driver

• Fixed bitstream allocation error corner case and added more verbose debugging information

Version 4.1.3.4 (10.11.2015)

PLC driver

• Fix the issue where uninitialized beacon packets would crash the DSP (bitstream allocation error)

Version 4.1.3.3 (23.10.2015)

PLC driver

• Code refactored memory barrier in the FSM

Multicast handling

• Initialize receive IGMP message to zero, otherwise the message could be mangled

General improvements

• Added the possibility to turn DHCP on and off from the device web
• Added the possibility to extract trace logs from the device web
• BSP improvements
  • C++ support
  • LUA support

Version 4.1.3.2 (07.10.2015)

PLC driver

• Code prepared for limited release

General improvements

• Web interface showing the correct MAC address
• Fixed IGMP snooping daemon logging

Version 4.1.3.1 (09.09.2015)

PLC driver

• Code refactored
• Memory barrier, locking mechanism added, to stop asserts in the PLC state machine that appeared due to thread context switching

Multicast handling

• Fixed logging
• Code refactored

General improvements

• Code refactored
• Fixed a bug while setting a device static IP
• Fixed DHCP client logging

Version 4.1.3.0 (29.05.2015)

Issues with lost communication links

• An issue with broadcast streams fixed, no more long drop-outs, only some non-contiguous ping losses (can normally appear due to the topology and data throughput on the network)
• Device acts as a hub, bridge aging time is set to zero

System stability improvements

• Improved MAC stream handling to prevent loss of broadcast traffic
• Adaptation of beacon handling to prevent collisions with RTP traffic.
• Modifications in PHY level frame handling to improve system stability.
• Speeding up the tonemap calculation.
• Improvement of beacon synchronization.
• Improvement of PLC network tracking and synchronization during coupling transition period.

Issues with group calls using multicast

• Supported number of members, per one multicast group in IGMP snoopings application, is raised to 10, the software now supports maximum of 12 IGMP groups up to 10 members each
• Groups with too many members appear on Zenitel Flowire devices. Only the IGMP members from the same PLC cluster should be tracked, the solution is to accept only the IGMP reports coming from plc0 using a netlink between plcdrv and the IGMP snooping deamon
• Flowire boards used a wrong source IP address when sending IGMP queries (br0:0 instead of br0 address). The solution is to use raw packets and prepare the IP header directly in IGMP snooping daemon
• IGMP robustness variable set to 5
• Fixed IGMP membership timeout according to the IGMPv3 standard (https://tools.ietf.org/html/rfc3376)

General improvements

• Reduced device boot time
• Created a Flowire specific device WEB interface
  • Remote upgrade WEB interface
  • The device is discoverable and upgradable using the **Intercom Management Tool (IMT)
  • Monitoring information
  • Remote Network Management Key (NMK) change
  • Configurable static IP alias
• Fixed DHCP client bug

Known issues

PLC link errors cause the device to reboot. These are known errors for the Flowire DSP chipset that happen rarely in specific cases.

• Hardware freeze (no ACCESS, no RESYS) is a fault that locks the DSP in a state that cannot be exited
• Asserts in the PLC driver (RX_DATA_WACK_LAST_PB "this looks like maria") happens in a rare case when the PLC link Tonemap is recalculated too often due to a bad link in one part of the communication spectrum
• Asserts in the PLC driver that occur on a large system reboot (../../mac/ca/src/alloc.c:161: assertion failure in ca_alloc_update_beacon_periods: ctx->state == CA_STATE_IDLE || (beacon_periods[0].start_date == ctx->current_beacon_period->start_date)). The device will reset and start to function properly in less than 10 seconds. This issue was observed rarely with large system on complete power down/ups, though this is not a problem since the rest of the system won't be operational and the Flowire will recover by the time the rest of the system is operational.