Documentation

Release Notes

New in this release

New Features
  • Built ECI features on upstream Debian.org (buster and buster-backports) packages apt-get and dpkg-builddeps
  • Replaced all Deby target image variants with Isar compatible to LTS GNU/Debian 10 (Buster)
  • Updated RTS Hypervisor to R5.4.01.27752 (Add PCI MSI-X hypervisor emulation for Intel® Atom™ x6000 Series [Elkhart Lake] Ethernet GbE Controllers PCI ID 8086:4BA0 or 8086:4BB0, add #AC on Split Lock, a Time Coordinated Computing (TCC) feature, can be selectively turned on or off.)
  • Updated Linux Ethernet 802.1Q TSN frameworks to v0.8.9
  • Updated IEC-62541 OPC-UA runtime library
  • Updated Robot Operating System 2 to “Foxy” release
  • Updated Linux PREEMPT_RT baseline to Intel LTS (5.4/preempt-rt) v5.4.115-rt57 [id-b428b24]
  • Updated Linux Xenomai 3.1.1 w/ ipipe-core-5.4.115-x86-4 baseline to Intel LTS (5.4/yocto) v5.4.115-ipipe [id-57f2d38]
  • Updated Intel® In-band Manageability
  • Updated Intel® Secure Device Onboarding
  • Updated Trusted VM for IoT Platforms
  • Updated Intel® Time Coordinated Computing Tools (Intel® TCC Tools)
  • Updated ACRN Hypervisor to v2.5
  • Improved Edge Control Procotol Bridge OPC UA (part 4) UADP frame Plug-ins for IEEE sdt 802.1Q-2018 Enhanced Scheduled Traffic (EST)
  • Improved Edge Control Procotol Bridge OPC UA security
  • Added Intel® Edge Insights for Industrial
  • Added support for CANopen
Unsupported or Discontinued Features
N/A

Known Issues

Limitation #1
When building an ECI image, the build process fails due to
do_fetch
errors. These errors occur when the build engine is unable to reach external dependencies via the network. Most often, these errors occur due to network instability, broken mirrors, or proxy server instability.
Mitigation Actions
  • It is recommended to avoid environments with network proxies and heavy network congestion.
  • Often, repeating the build process will eventually result in a complete build.
  • Outages and offline mirrors tend to be intermittent. Attempt the build again after waiting a few hours for the issues to subside.
Limitation #2
[ECI-A] When intensive storage I/O is sustained within a virtual machine (VM) running on the ACRN hypervisor, some I/O requests may be dropped. Linux kernels will interpret this behavior as faulty storage hardware and automatically remount the affected storage partition as read-only. This situation may occur if storage benchmarks are performed within the VM.
  • Avoid sustained intensive storage I/O within the VM.
  • Assign more virtual CPUs to the VM to increase the effective I/O bandwidth.
  • Limit the number of VMs running simultaneously.
  • Avoid sharing physical CPUs between VMs.
Limitation #3
Benchmark
msi-jitter
/opt/benchmarking/msi-jitter
does not provide results on certain devices. Known affected devices are Karbon 700 and Karbon 300.
************ RESULTS (ns) ************ [79774.994363] * Max: 0 [79774.994363] * Avg: 0 [79774.994364] * Min: 0 [79774.994364] *********** *********** ***********
This behavior is currently observed only for i210 Ethernet NICs connected to cascading PCI bridges. The following is an
lspci -t
example from a Karbon 700 device:
lspci -t -[0000:00]-+-00.0 +-02.0 +-08.0 +-12.0 +-14.0 +-14.2 +-15.0 +-15.1 +-16.0 +-16.3 +-17.0 +-1b.0-[01]----00.0 +-1b.6-[02-08]----00.0-[03-08]--+-01.0-[04]----00.0 | +-02.0-[05]----00.0 | +-03.0-[06]----00.0 | +-04.0-[07]----00.0 | \-05.0-[08]-- +-1c.0-[09]----00.0 +-1c.6-[0a]----00.0 +-1c.7-[0b]----00.0 +-1e.0 +-1f.0 +-1f.3 +-1f.4 +-1f.5 \-1f.6 lscpi 00:00.0 Host bridge: Intel Corporation 8th Gen Core 8-core Desktop Processor Host Bridge/DRAM Registers [Coffee Lake S] (rev 0d) ... 00:1b.6 PCI bridge: Intel Corporation Cannon Lake PCH PCI Express Root Port #23 (rev f0) 00:1c.0 PCI bridge: Intel Corporation Cannon Lake PCH PCI Express Root Port #3 (rev f0) 00:1c.6 PCI bridge: Intel Corporation Cannon Lake PCH PCI Express Root Port #7 (rev f0) 00:1c.7 PCI bridge: Intel Corporation Cannon Lake PCH PCI Express Root Port #8 (rev f0) ... 02:00.0 PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch 03:01.0 PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch 03:02.0 PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch 03:03.0 PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch 03:04.0 PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch 03:05.0 PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch 04:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) 05:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) 06:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) 07:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) 09:00.0 Non-Volatile memory controller: Silicon Motion, Inc. Device 2263 (rev 03) 0a:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03) 0b:00.0 Ethernet controller: Intel Corporation I210 Gigabit Network Connection (rev 03)
In the above example, ethernet controller
04:00.0 , 05:00.0 , 06:00.0 , 07:00.0
would be affected as they connect to
00:1c.6 PCI bridge: Intel Corporation Cannon Lake PCH PCI Express Root Port #7 (rev f0) and PCI bridge: Pericom Semiconductor PI7C9X2G608GP PCIe2 6-Port/8-Lane Packet Switch
.
Mitigation Actions
Run
msi-jitter
benchmark on non-cascading PCI bridged ethernet controllers. For example,
./run_msijitter.sh "0000:0a:00.0" 1 10 10
, for Karbon 700. This behavior is unique to the
msi-jitter
benchmark and does not affect regular operation of the ethernet controller. The
msi-jitter
benchmark kernel driver has to be fixed with upcoming releases.
Limitation #4
Generally, tc qdisc TAPRIO gate-masks definitions are specified in terms of traffic classes (tc), so
sched-entry
S 03 5000000 means that tc 0 and 1 are open for 500us. But, when taprio s
ched-entries
are offloaded to hardware, that is, flags 0x2 , the gate-masks are specified as hardware queues, which means txq 0 and 1 are open for 500us.
To make sure that the corresponding
sched-entries
are applied, reload the igc kernel module with fully verbose debug.
rmmod igc insmod /lib/modules/5.4.59-rt36-intel-pk-preempt-rt/kernel/drivers/net/ethernet/intel/igc/igc.ko debug=16 dmesg -w [2967265.790583] igc 0000:02:00.0 eth0: Detected Tx Unit Hang Tx Queue <0> TDH <55> TDT <5a> next_to_use <5a> next_to_clean <55> buffer_info[next_to_clean] time_stamp <111ae3b16> next_to_watch <00000000737c47b3> jiffies <111af42f0> desc.status <938000> [2967267.790416] igc 0000:02:00.0 eth0: Detected Tx Unit Hang Tx Queue <0> TDH <55> TDT <5a> next_to_use <5a> next_to_clean <55> buffer_info[next_to_clean] time_stamp <111ae3b16> next_to_watch <00000000737c47b3> jiffies <111af43b8> desc.status <938000>
Limitation #5
Following is the list of known issues/limitations for Intel® Elkhart Lake Functional Safety (FuSa) Software Package (Revision 0.8.7)
Title
Description
Proof Test not supported
Proof Test is not supported in Revision 0.8.7 release
1ms/4ms PST not supported
PST is set to 100ms in Revision 0.8.7 release
False NOK
ODCC execution could result in a False NOK generation due to timer expiry
Startup STL Test Failure message
Non-fatal message; recommended to ignore the message

Product and Performance Information

1

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.