Version Identification 

• Application: 2.0.1
• Stack: 3.0.0

ubisys Compact7B Stack Changes between 2.0.13 and 3.0.0

• Minor optimizations to internal flash write/erase flow
• Migrated to new 64-bit timer implementation across the entire stack
• Improved support for the future ZigBee Direct technology
• Consolidated support for All Hubs at the stack level, e.g., advanced parent link maintenance mechanism, beacon survey etc.
• Completed support for the proprietary All Hubs cluster
• Added support for dynamic cluster instantiation
• Completed support for Hybrid Zigbee Green Power technology
• Completed support for the Color Control cluster
• Added better programmatic control over attribute reports, enabled pausing and resuming attribute reports
• Enhanced the keep-alive mechanism for end devices via the end device timeout approach
• Make sure that APS security of incoming requests is mirrored to outgoing responses
• Extended maximum reporting interval (enabled by new 64-bit timer implementation)
• Enhancements and improvements of the Level Control cluster
• Enhancements and improvements of the Thermostat cluster
• Added support for the HMAC-SHA-256 encryption scheme
• Further modularized LZMA decompression implementation
• Added ZDO request queue support for a ZDO request command anticipating multiple responses
• Added ZCL request queue support for a ZCL request command anticipating multiple responses
• Added enhancements to the OTA server locating process
• Added support for Green Power GPDF frame encryption
• Optimized start-up LED indicator behavior, i.e. improve arbitration of LED usage between menu and status indications
• Added support for multiple handlers for the same GPIO
• Added generalized section support for multiple platforms
• Reject TC rejoin with MAC capabilities change to prevent security vulnerability (CSA incident no. 2021-ZP-0401)
• Added retry mechanism to the MAC association process for robustness in dense networks
• Explicitly set the destination PAN ID in case of injecting an unsolicited rejoin response due to address conflict (improve interoperability with 3^rd party stack)
• Reject a leave request during the startup process if a rejoin is already in progress
• Routing optimization, add weak provisional default route to trust center
• Indicate parent router link failure upon failed keep-alive mechanism by making use of the end device timeout approach
• Introduced a multi-pass priority scheme for network leave processing (improve modularity)
• Ensure factory new state when a Touch Link commissioning procedure is instigated (for devices which do not reboot)
• Improved the Touch Link commissioning process flow in case of router/end device joining (await confirmation for pending leave request)
• Improved the transition behavior for the Level Control cluster
• Color Control cluster: Improved the behavior when changing color modes
• Extended support for the Fan Control cluster
• Added support for the Basic client cluster
• Added framework support for processing ZCL Read Attributes response command
• Enhanced the behavior upon receipt of an OTA Image Notify command
• Consider the fact that All Hubs and Touch Link clusters are only optionally available
• Consider the fact that a remote OTA server might not support rate limiting
• Prohibited OTA downgrade by default
• Added support for the gateway attribute “current channel”
• Added support for I2C TAP
• Added support for a unified and fully configurable join/rejoin algorithm with back-offs and retries at the stack level
• Reject trust center rejoin on a distributed security network (CSA incident no. 2021-ZP-0401)
• Start TCLK exchange if an end device could not successfully finish TCLK exchange during its initial joining and has just rejoined
• Improved sleep management: Checking both transmission and reception queues for pending packets, if necessary
• Extended the advanced parent link maintenance mechanism by the new criterion „poor poll control reactivity“
• More granular startup progress notifications, including kick-off of a Touch Link commissioning procedure
• Stop ongoing join/rejoin retry attempts when Touch Link commissioning procedure is instigated
• Encapsulated the All Hubs cluster’s dependency on the Touch Link server, IAS Zone server and Time client clusters, for better modularity
• Added Enhancements to the Thermostat cluster
• Made the hard limit for the scene table capacity configurable
• Added support for the Relative Humidity Measurement cluster
• Added framework support for initiating enroll request and IAS zone status change indication via the ZCL request queue
• Added framework support for time query initiated by a Time client cluster
• Prohibit automatic CIE detection by default; application is still able to instigate it though should legacy systems require it
• Explicitly stop fast polling mode as soon as no more response frame is pending
• Resume network activity if a router device just failed a rejoin attempt
• Enhancement: PAN-ID and permit joining state changes are forwarded to application clusters
• Bug-fix: Idle processing requires thread-safety on some platforms
• Bug-fix: Improve co-existence of AODV and MTO routing algorithms
• Bug-fix: Under rare circumstances a pending outgoing attribute report got discarded inadvertently
• Bug-fix: OTA hash calculation must account for data chunks not arriving in-sequence
• Bug-fix: Make sure all code paths align the transmit options for default responses with the mating request
• Bug-fix: Cluster-level global flags and individual configuration options used to be not maintained persistently
• Bug-fix: Disentangle cluster-level policy settings for APS ACKs and default responses
• Bug-fix:Fixed Green Power OOB key encryption for emerging Green Power install code support
• Bug-fix: Nonce setup for GPD vs. GPS used to be incorrect (relates to Hybrid Green Power feature)
• Bug-fix: NWK short address might be invalidated after a failed TC rejoin when the NWK key material has not been received in time
• Bug-fix: Ensure resetting the startup parameters during a cold-start
• Bug-fix: Ensure resetting the network layer’s retransmission queue across stack restarts
• Bug-fix: On some platforms loss of timer ticks might occur due to internal timer resolution discrepancy
• Bug-fix: Need to only evaluate the router/end device capacity fields in a beacon payload when performing network discovery during initial joining
• Bug-fix: Startup parameters might be inadvertently dropped across startup retry attempts
• Bug-fix: Unexpected behavior might be encountered while resolving an ongoing PAN ID conflict (caused by typo)
• Bug-fix: Internal coordinator flag might have inadvertently been set while forming a distributed network
• Bug-fix: Client/server cluster role not always properly honored when accessing an attribute
• Bug-fix: Increment network update ID in course of network-wide channel change (improve interoperability)
• Bug-fix: Global ZCL commands used to be silently dropped if the APS security level was insufficient. Instead, create a proper response mirroring the incoming security and indicating insufficient privileges.
• Bug-fix: Under rare circumstances, a frame might linger in the transmission queue without being actually sent
• Bug-fix: Potential overflow of 64-bit timer values might occur when unsigned arithmetic is used for comparison of due times
• Bug-fix: During MAC association, if an unrelated frame is being received via data polling, MAC association might be incorrectly terminated; account for this case and retry if permissible
• Bug-fix: Periodical polling timer used to be unconditionally kept alive, even in case of data poll failure; temporarily cancel it to avoid interference with subsequent retry attempts. Once succeeded, the periodical polling timer will be reinstated for its normal usage.
• Bug-fix: When assessing MAC capability changes for a rejoining device verify the „Device Type“ and „Receiver on when Idle“ bits only
• Bug-fix: Skip route discovery table verification diagnostics in case of an end device
• Bug-fix: Limit the use of incoming APS security material to few select cases only
• Bug-fix: Periodical reporting on a cluster attribute could lead to undue delays of reports for other clusters
• Bug-fix: Startup behavior of the OnOff cluster might be incorrect due to interactions with Level Control or Color Control clusters on the same application endpoint
• Bug-fix: Brightness adjustment of the Level control cluster might be incorrect due to interactions with a Color Control cluster on the same application endpoint
• Bug-fix: Under rare circumstances, neighbor table entries might be incorrect (caused by a typo)
• Bug-fix: Under rare circumstances, out-of-sequence or truncated OTA data block could result in failed OTA download attempts