Update RELEASE_NOTES for ns-3.27
This commit is contained in:
230
RELEASE_NOTES
230
RELEASE_NOTES
@@ -9,8 +9,8 @@ http://www.nsnam.org including tutorials: http://www.nsnam.org/tutorials.html
|
||||
Consult the file CHANGES.html for more detailed information about changed
|
||||
API and behavior across ns-3 releases.
|
||||
|
||||
Release 3-dev
|
||||
=============
|
||||
Release 3.27
|
||||
============
|
||||
|
||||
Availability
|
||||
------------
|
||||
@@ -19,110 +19,168 @@ This release is not yet available.
|
||||
Supported platforms
|
||||
-------------------
|
||||
This release is intended to work on systems with the following minimal
|
||||
requirements:
|
||||
requirements (Note: not all features available on all platforms):
|
||||
- g++-4.9 or later (Note: this is an upgraded requirement from ns-3.26)
|
||||
- Apple LLVM version 7.0.2 or later
|
||||
- clang-3.3 or later
|
||||
- Python 2.7 (Python 2 series) or Python 3.4 (Python 3 series) or later
|
||||
|
||||
This release has been tested on the following platforms:
|
||||
- Ubuntu 17.04 (64 bit) with g++-6.3.0
|
||||
- Ubuntu 16.04 (64 bit) with g++-5.4.0
|
||||
- Ubuntu 14.04.5 (64 bit) with g++-4.9.4
|
||||
- Note: the default g++ version (4.8) must be upgraded
|
||||
- Fedora Core 26 (64 bit) with g++-7.1.1
|
||||
- Fedora Core 22 (64 bit) with g++-5.3.1
|
||||
- CentOS/RHEL 7.4 (64-bit) with g++-6.3.1
|
||||
- Note: the default g++ version (4.8) must be upgraded
|
||||
- CentOS/RHEL 6.8 (64-bit) with g++-4.9.2
|
||||
- Note: the default g++ version (4.4) must be upgraded
|
||||
- Note: the default Python version (2.6) must be upgraded
|
||||
- OS X Sierra 10.12.6 with Xcode 9.0 and Apple LLVM version 9.0.0
|
||||
- OS X El Capitan 10.11.6 with Xcode 8.2.1 and Apple LLVM version 8.0.0
|
||||
- OS X Yosemite 10.10.5 with Xcode 7.2.1 and Apple LLVM version 7.0.2
|
||||
- Windows 10 Linux Subsystem with g++-5.4.0
|
||||
|
||||
New user-visible features
|
||||
-------------------------
|
||||
- (wifi) 802.11ax High Efficiency (HE) physical layer modes are now supported.
|
||||
- (tcp) The SACK option and the RFC 6675 loss recovery algorithm are now supported.
|
||||
- (build) Modules can also be located in contrib/ directory (as well as src/)
|
||||
- (core) Logging timestamp precision has been improved
|
||||
- (lte) LTE carrier aggregation feature according to 3GPP Release 10 is now supported.
|
||||
- (network) CsmaNetDevice, SimpleNetDevice and WifiNetDevice support flow control.
|
||||
- (traffic-control) CsmaNetDevice, SimpleNetDevice and WifiNetDevice support flow control.
|
||||
- (traffic-control) Added support for queue discs to mark packets, and ECN support has been added to RED
|
||||
- (traffic-control) A multi-queue aware queue disc modelled after Linux mq
|
||||
- (traffic-control) Added support for detailed statistics output of queue discs
|
||||
- (traffic-control) Added support for tracing queue sojourn times
|
||||
- (tcp) The SACK option and the RFC 6675 loss recovery algorithm are now supported.
|
||||
- (tcp) A LEDBAT model has been added.
|
||||
- (wifi) 802.11ax High Efficiency (HE) physical layer modes are now supported.
|
||||
- (wifi) A new RRPAA rate control mechanism has been added.
|
||||
- (wifi) A new trace source for TXOP duration is now supported
|
||||
- (wifi) Frame capture effect can now be modelled
|
||||
- (uan) World Ocean System Simulation (WOSS) integration, and new PER models
|
||||
|
||||
Bugs fixed
|
||||
----------
|
||||
- Bug 1034 - No trace source for packet dropping from WifiMacQueue
|
||||
- Bug 1034 - wifi: No trace source for packet dropping from WifiMacQueue
|
||||
- Bug 2007 - uan: Remove deprecation on SetRxThresholdDb
|
||||
- Bug 2048 - lte: Fix scheduler resource allocation when CQI==0 (out of rante)
|
||||
- Bug 2133 - tcp: Avoid assert when receiving data in FIN_WAIT_1 or _2 states
|
||||
- Bug 2214 - tcp: Use of ScheduleNow only in selected part of the code
|
||||
- Bug 2221 - network: Remove constraint on size of ns3::Packet Tag objects
|
||||
- Bug 2248 - tcp: TCP SACK added after reviews
|
||||
- Bug 2256 - tcp: bytes in flight now updated in TcpTxBuffer
|
||||
- Bug 2263 - Support processing of multiple TCP options
|
||||
- Bug 2263 - tcp: Support processing of multiple TCP options
|
||||
- Bug 2285 - tcp: Loss of ack of SYN can cause improper connection setup
|
||||
- Bug 2367 - BlockAckManager does not remove iterators to freed items
|
||||
- Bug 2450 - LogDistancePropagationLossModel is not continuous
|
||||
- Bug 2463 - create trace source to trace the TXOP time that is actually used
|
||||
- Bug 2471 - unable to disable Block Ack agreement for 802.11n
|
||||
- Bug 2477 - DCF manager assert
|
||||
- Bug 2485 - Check for queue full should happen before checking RED thresholds
|
||||
- Bug 2491 - replace DcfManager::MY_DEBUG
|
||||
- Bug 2367 - wifi: BlockAckManager does not remove iterators to freed items
|
||||
- Bug 2450 - propagation: LogDistancePropagationLossModel is not continuous
|
||||
- Bug 2456 - lte: MemberCschedSapProvider and Member CschedSapProvider templates
|
||||
- Bug 2464 - spectrum: Store spectrum conversion matrix in CSR format
|
||||
- Bug 2467 - spectrum: Do not schedule StartRx for receivers on other models
|
||||
- Bug 2471 - wifi: unable to disable Block Ack agreement for 802.11n
|
||||
- Bug 2477 - wifi: DCF manager assert
|
||||
- Bug 2485 - traffic-control: Check queue full before checking RED thresholds
|
||||
- Bug 2491 - wifi: replace DcfManager::MY_DEBUG
|
||||
- Bug 2492 - uan: Make use of RxGain attribute in UanPhyGen class
|
||||
- Bug 2511 - HT Greenfield is not working
|
||||
- Bug 2512 - Byte mode not configured correctly in red-queue-disc-test-suite.cc
|
||||
- Bug 2513 - ParetoRandomVariable needs a "scale", not a "mean" attribute.
|
||||
- Bug 2518 - Suppress printing of list of modules for Python programs
|
||||
- Bug 2521 - Include ipv6-option.h in wscript
|
||||
- Bug 2527 - PrintRoutingTable extended to add an optional Time::Units parameter
|
||||
- Bug 2528 - 802.11n RIFS cannot be enabled
|
||||
- Bug 2529 - Missing trace when Block ACK timeout is triggered or when missing MPDUs are announced by a Block ACK response
|
||||
- Bug 2530 - Rename aodv::SetBalckListTimeout to aodv::SetBlackListTimeout
|
||||
- Bug 2532 - Inconsistencies between 802.11n MCS and NSS value reported in TXVECTOR
|
||||
- Bug 2533 - Provide a better 802.11n/ac PHY abstraction model for SIMO, MISO and MIMO
|
||||
- Bug 2535 - memory leak in bench-simulator.cc
|
||||
- Bug 2536 - fixed dead assignment and potential memory leak in wimax
|
||||
- Bug 2537 - Dead assigment on CoDel::DoDequeue
|
||||
- Bug 2538 - fixed dead assignment on tap-bridge
|
||||
- Bug 2540 - fixed dead assignment on mesh/ie-dot11s-perr
|
||||
- Bug 2541 - preamble not assigned correctly
|
||||
- Bug 2542 - dead assignment on wifi mac-low
|
||||
- Bug 2545 - Optimized build fails for flow-monitor
|
||||
- Bug 2547 - dead assignments on various tcp congestion controls
|
||||
- Bug 2551 - wifi preamble should be part of TXVECTOR
|
||||
- Bug 2552 - Remove code duplication between Yans and Spectrum
|
||||
- Bug 2559 - TCP advertised window still incorrect
|
||||
- Bug 2561 - TCP can not use Connect after a BindToNetDevice
|
||||
- Bug 2563 - Modify pybindgen version check to >= minimum version required
|
||||
- Bug 2564 - Simulation crashes when CtsTimeout is triggered for A-MPDU transmissions
|
||||
- Bug 2565 - TCP; do not wait for 2*MSL to notify socket close
|
||||
- Bug 2566 - BlockAckManager::GetNRetryNeededPackets missing some packets in the queue
|
||||
- Bug 2573 - Added TCP Ledbat
|
||||
- Bug 2577 - simulation crashes when A-MPDU and multiple TOS are used with RTS-CTS enabled
|
||||
- Bug 2578 - Assert "Internal collision but no packet in queue" unexpectedly triggered
|
||||
- Bug 2584 - MacLow triggers StartNext even if there is no TXOP
|
||||
- Bug 2587 - Avoid overflow in htcp.cc
|
||||
- Bug 2590 - Minor enhancements in red-queue-disc{.h, .cc}
|
||||
- Bug 2591 - 802.11e Block Ack mechanism cannot be enabled on HT/VHT stations
|
||||
- Bug 2594 - vht-wifi-network provides very low throughtput at MCS 6, 160 MHz, SGI
|
||||
- Bug 2601 - HT stations should use 40 MHz channel width if the configured channel width is set to 80 or 160 MHz
|
||||
- Bug 2604 - QosData frames with Block Ack policy should be separated by SIFS as long as they belong to the same TXOP
|
||||
- Bug 2605 - A HT/VHT station transmitting to a legacy access point results in a null throughput
|
||||
- Bug 2606 - Ideal rate manager does not work correctly when a HT/VHT station is transmitting to a legacy access point
|
||||
- Bug 2613 - MaxRxSequence () is sometimes too large
|
||||
- Bug 2607 - Minstrel HT manager results in an endless loop when a 802.11ac station is transmitting to a 802.11a access point
|
||||
- Bug 2614 - RIP header version should be set to 2
|
||||
- Bug 2615 - When the TXOP limit is nonzero, packets should be fragmented so that the TXOP limit is not exceeded
|
||||
- Bug 2627 - Ipv6RawSocket does not honor the bound interface when sending packets
|
||||
- Bug 2628 - Simulation crashes because of an out of range TID
|
||||
- Bug 2629 - Assert failure in MinstrelHtWifiManager::GetLowestIndex
|
||||
- Bug 2647 - ideal-wifi-manager-example crashes when NSS > 1
|
||||
- Bug 2654 - Rraa wifi manager cannot be used with 802.11b
|
||||
- Bug 2655 - A-MPDU simulation using TCP sometimes fails with message "Packet has no Traffic ID"
|
||||
- Bug 2660 - TrafficControlHelper::Default should not configure internal queue sizes
|
||||
- Bug 2665 - Ipv4RawSocket can not send packets to broadcast or multicast
|
||||
- Bug 2671 - ArpCache::Entry::SetMacAddress is misspelled
|
||||
- Bug 2673 - run-time channel switch does not update WiFi spectrum model
|
||||
- Bug 2716 - TCP socket ignoring FIN flag when in FIN_WAIT state
|
||||
- Bug 2717 - Fix mask generation for Ipv4RoutingTableEntry::CreateDefaultRoute
|
||||
- Bug 2722 - 802.11g sends DSSS spectrum signals using CreateOfdmTxPowerSpectralDensity
|
||||
- Bug 2733 - Ideal wifi manager cannot handle NSS higher than 1
|
||||
- Bug 2741 - IPv4 fragmentation fails when last fragment have to be re-fragmented.
|
||||
- Bug 2744 - 802.11n/ac with RTS/CTS is crashing for a large number of nodes
|
||||
- Bug 2751 - QueueDisc::Enqueue() order of operations
|
||||
- Bug 2757 - 802.11n/ac/ax maximum TXOP is not properly enforced
|
||||
- Bug 2758 - IPv4 sockets bound to unicast receive also subnet-directed broadcasts
|
||||
- Bug 2759 - Packets sent to broadcast address are converted to subnet-directed broadcast
|
||||
- Bug 2760 - OLSR uses unicast-bound sockets to receive broadcasts
|
||||
- Bug 2761 - Packet has no Traffic ID for CTS frames when A-MPDU is used
|
||||
- Bug 2762 - BindToNetDevice behaviour is not coherent with Linux
|
||||
- Bug 2763 - wifi management packets in the PCAPs are all marked as wrong by Wireshark
|
||||
- Bug 2769 - Set ssthresh correctly upon RTO
|
||||
- Bug 2771 - wifi: Simulation fails with IdealWifiManager if stations have different number of antennas
|
||||
- Bug 2511 - wifi: HT Greenfield is not working
|
||||
- Bug 2512 - traffic-control: Byte mode not configured correctly for RED test
|
||||
- Bug 2513 - core: ParetoRandomVariable needs a "scale", not a "mean" attribute
|
||||
- Bug 2518 - build: Suppress printing of list of modules for Python programs
|
||||
- Bug 2519 - lte: m_windowOffsetsMap.end () dereference
|
||||
- Bug 2520 - examples: TCP variant not configured in wifi-tcp.cc
|
||||
- Bug 2521 - internet: Include ipv6-option.h in wscript
|
||||
- Bug 2523 - lte: Generate the RLC PDU properly
|
||||
- Bug 2527 - internet: PrintRoutingTable adds an optional Time::Units parameter
|
||||
- Bug 2528 - wifi: 802.11n RIFS cannot be enabled
|
||||
- Bug 2529 - wifi: Missing trace for Block ACK timeout or missing MPDUs
|
||||
- Bug 2530 - aodv: Rename aodv::SetBalckListTimeout to aodv::SetBlackListTimeout
|
||||
- Bug 2531 - lte: Set HARQ feedback default value to avoid false retrasmissions
|
||||
- Bug 2532 - wifi: Inconsistencies between 802.11n MCS and NSS value in TXVECTOR
|
||||
- Bug 2533 - wifi: Provide better 802.11n/ac PHY abstraction for SIMO/MISO/MIMO
|
||||
- Bug 2534 - traffic-control: ARED and PIE examples should return 1 upon exit
|
||||
- Bug 2535 - utils: memory leak in bench-simulator.cc
|
||||
- Bug 2536 - wimax: fixed dead assignment and potential memory leak in wimax
|
||||
- Bug 2537 - traffic-control: Fix dead assigment on CoDel::DoDequeue
|
||||
- Bug 2538 - tap-bridge: fixed dead assignment on tap-bridge
|
||||
- Bug 2540 - mesh: fixed dead assignment on mesh/ie-dot11s-perr
|
||||
- Bug 2541 - wifi: preamble not assigned correctly
|
||||
- Bug 2542 - wifi: dead assignment on wifi mac-low
|
||||
- Bug 2543 - lte: Fix when EstimateUlSinr returns NO_SINR
|
||||
- Bug 2545 - flow-monitor: Optimized build fails for flow-monitor
|
||||
- Bug 2547 - tcp: dead assignments on various tcp congestion controls
|
||||
- Bug 2551 - wifi: wifi preamble should be part of TXVECTOR
|
||||
- Bug 2552 - wifi: Remove code duplication between Yans and Spectrum
|
||||
- Bug 2554 - wave: Fix static analysis warning
|
||||
- Bug 2556 - csma, point-to-point: Avoid modification of transmitted packets
|
||||
- Bug 2557 - point-to-point: Fix dead assignments in queue-discs-benchmark
|
||||
- Bug 2559 - tcp: TCP advertised window still incorrect
|
||||
- Bug 2561 - tcp: TCP can not use Connect after a BindToNetDevice
|
||||
- Bug 2563 - bindings: pybindgen version check to >= minimum version required
|
||||
- Bug 2564 - wifi: Simulation crashes when CtsTimeout triggered for A-MPDU
|
||||
- Bug 2565 - tcp: Do not wait for 2*MSL to notify socket close
|
||||
- Bug 2566 - wifi: BlockAckManager::GetNRetryNeededPackets missing some packets
|
||||
- Bug 2568 - examples: Avoid enumerating TCP variants
|
||||
- Bug 2577 - wifi: crashes when A-MPDU and multiple TOS used, upon CTS timeout
|
||||
- Bug 2578 - wifi: Unexpected assert "Internal collision but no packet in queue"
|
||||
- Bug 2584 - wifi: MacLow triggers StartNext even if there is no TXOP
|
||||
- Bug 2587 - tcp: Avoid overflow in htcp.cc
|
||||
- Bug 2590 - traffic-control: Minor enhancements in red-queue-disc{.h, .cc}
|
||||
- Bug 2591 - wifi: 802.11e Block Ack cannot be enabled on HT/VHT stations
|
||||
- Bug 2594 - wifi: vht-wifi-network very low throughtput at MCS 6, 160 MHz, SGI
|
||||
- Bug 2596 - network: EthernetTrailer::GetFcs() should be const
|
||||
- Bug 2601 - wifi: HT stations should use 40 MHz width if configured 80 or 160z
|
||||
- Bug 2604 - wifi: QosData frames separation with Block Ack enabled
|
||||
- Bug 2605 - wifi: HT/VHT station sending to legacy results in null throughput
|
||||
- Bug 2606 - wifi: Ideal rate manager for HT/VHT station to a legacy AP
|
||||
- Bug 2607 - wifi: Correct overflow in channel width loop in Minstrel HT
|
||||
- Bug 2613 - tcp: MaxRxSequence () is sometimes too large
|
||||
- Bug 2614 - internet: RIP header version should be set to 2
|
||||
- Bug 2615 - wifi: When the TXOP limit is nonzero, fragment in some cases
|
||||
- Bug 2617 - traffic-control: PfifoFastQueueDisc::Peek all bands
|
||||
- Bug 2621 - wifi: Fix mismatch between log message and action
|
||||
- Bug 2624 - lte: Inconsistent coding of type and naming
|
||||
- Bug 2627 - internet: Ipv6RawSocket does not honor the bound interface
|
||||
- Bug 2628 - wifi: Simulation crashes because of an out of range TID
|
||||
- Bug 2629 - wifi: Assert failure in MinstrelHtWifiManager::GetLowestIndex
|
||||
- Bug 2632 - lte: Build error due to uninitialized variable in rem helper
|
||||
- Bug 2646 - lte: lena-simple-eps-emu example will never work
|
||||
- Bug 2647 - wifi: ideal-wifi-manager-example crashes when NSS > 1
|
||||
- Bug 2649 - tcp: Disabling TCP SACK doesn't fall back to NewReno
|
||||
- Bug 2654 - wifi: Rraa wifi manager cannot be used with 802.11b
|
||||
- Bug 2655 - wifi: A-MPDU simulation using TCP sometimes fails
|
||||
- Bug 2658 - core: Avoid unexpected TracedValue type conversions
|
||||
- Bug 2660 - traffic-control: TrafficControlHelper::Default configuration
|
||||
- Bug 2661 - wave: Fix outdated class references in documentation
|
||||
- Bug 2665 - internet: Ipv4RawSocket can not send to broadcast or multicast
|
||||
- Bug 2666 - lte: Remove deprecated variables/config paths
|
||||
- Bug 2668 - various: Make template classes use NS_LOG_* macros
|
||||
- Bug 2671 - internet: ArpCache::Entry::SetMacAddress is misspelled
|
||||
- Bug 2673 - wifi: run-time channel switch does not update WiFi spectrum model
|
||||
- Bug 2675 - traffic-control: Use correct string value in PIE example
|
||||
- Bug 2716 - tcp: TCP socket ignoring FIN flag when in FIN_WAIT state
|
||||
- Bug 2717 - internet: Fix mask for Ipv4RoutingTableEntry::CreateDefaultRoute
|
||||
- Bug 2722 - wifi: 802.11g sends DSSS spectrum signals inappropriately
|
||||
- Bug 2731 - lte: Allow retranmission in RLC AM when PR expires, vtS overflows
|
||||
- Bug 2732 - applications: BulkSend Tx trace is fired too much
|
||||
- Bug 2733 - wifi: Ideal wifi manager cannot handle NSS higher than 1
|
||||
- Bug 2741 - internet: IPv4 fragmentation fails on last fragment fragmentation
|
||||
- Bug 2744 - wifi: 802.11n/ac with RTS/CTS crashes for a large number of nodes
|
||||
- Bug 2751 - traffic-control: QueueDisc::Enqueue() order of operations
|
||||
- Bug 2756 - internet: Ipv4AddressGenerator underestimates available addresses
|
||||
- Bug 2757 - wifi: 802.11n/ac/ax maximum TXOP is not properly enforced
|
||||
- Bug 2758 - internet: IPv4 sockets bound to unicast receiving subnet broadcast
|
||||
- Bug 2759 - internet: Packet conversion from broadcast to subnet-directed
|
||||
- Bug 2760 - internet: OLSR uses unicast-bound sockets to receive broadcasts
|
||||
- Bug 2761 - wifi: Packet has no Traffic ID for CTS frames when A-MPDU is used
|
||||
- Bug 2762 - internet: BindToNetDevice behaviour is not coherent with Linux
|
||||
- Bug 2763 - wifi: management packets in PCAPs are marked as wrong by Wireshark
|
||||
- Bug 2769 - tcp: Set ssthresh correctly upon RTO
|
||||
- Bug 2771 - wifi: IdealWifiManager problems under different number of antennas
|
||||
- Bug 2776 - wifi: Fix segmentation fault in MacLow
|
||||
- Bug 2777 - lte: Save bearer info in m_drbMap, needed for handover request
|
||||
- Bug 2781 - tcp: Do not go beyond limit in TcpTxBuffer
|
||||
- Bug 2783 - wifi: MonitorSnifferRx SNR is invariant to RxNoiseFigure
|
||||
- Bug 2786 - lte: Remove unimplemented declaration
|
||||
- Bug 2796 - tcp: NSC TCP transmission flow control not working
|
||||
|
||||
Known issues
|
||||
------------
|
||||
|
||||
Reference in New Issue
Block a user