* Update submodule pointers
This is needed to address doxygen failures (Which are already fixed in
the submoduled repos).
Signed-off-by: Gaurav Aggarwal <aggarg@amazon.com>
This flag enables a user to only make local commits for various steps
and not push them to the remote.
Signed-off-by: Gaurav Aggarwal <aggarg@amazon.com>
As the manifest verification script has been moved to FreeRTOS/CI-CD-Github-Actions as a GitHub Action along with the added functionality of verifying that manifest.yml versions match the current submoduled pointers in the repository, this PR updates the CI check for manifest verification to take advantage of the new manifest-verifier action from the repository.
To validate that the FreeRTOS/CI-CD-Github-Actions/memory_statistics/paths.json configuration is correct, this PR adds a CI check mechanism to generate the memory statistics report in JSON format. Additionally, the CI check also uploads the report as an artifact (for main and release-candidate branches ONLY) for user accessibility of the generated report.
* Updated submodule libraries to latest release
* Updated pointer for FreeRTOS-Plus-TCP
* Updated README about SNTP
* Updated link for coreSNTP on freertos.org
* Updated corePKCS11 to 'update-pkcs11submodule' branch for testing
* Update corePKCS11 version in manifest.yml
* Updated corePKCS11 pointer to v3.1.0
Co-authored-by: Oscar Michael Abrina <abrinao@amazon.com>
Co-authored-by: Archit Aggarwal <architag@amazon.com>
* Replace library links with the application-protocols one
The ReadMe.txt refers to library link on FreeRTOS.org which will not be available until after the library is released in github. Instead, use top level page https://www.freertos.org/application-protocols.html which lists all the libraries in this directory which is also easy for user to navigate and could avoid this issue when releasing new library.
* Update FreeRTOS-Plus/Source/Application-Protocols/readme.txt
Co-authored-by: Archit Aggarwal <architag@amazon.com>
* Update FreeRTOS-Plus/Source/Application-Protocols/readme.txt
Co-authored-by: Paul Bartell <pbartell@amazon.com>
Co-authored-by: Archit Aggarwal <architag@amazon.com>
Co-authored-by: Paul Bartell <pbartell@amazon.com>
* Remove submodules of community and partner contribution repos, and stale directory of FreeRTOS+UDP
* Obselete NTP demo and rename of HTTP server demo folder
* Update manifest.yml
Since `mbedtls_error.c` is already part of the mbedTLS submodule, the duplicate files are removed from this repository.
Co-authored-by: Cobus van Eeden <35851496+cobusve@users.noreply.github.com>
Issue
The existing clock discipline algorithm in the coreSNTP demo yields an inaccurate wall-clock time (in UTC) that is always behind the astronomical UTC time by the chosen polling interval period of the demo. The issue with the clock discipline algorithm was that it was calculating the slew rate to make the system clock be gradually compensated for the previous time synchronization clock drift over the period of the subsequent polling time interval. Thereby, the clock slew rate correction calculated is always lower than it should be (because it does not incorporate possible clock drift that would occur over another period of time polling interval till the next time synchronization attempt).
Fix
This PR fixes the clock discipline algorithm to make the demo calculate its system time that matches the astronomical UTC time.
The fix involves simplifying the clock discipline algorithm to use a combination of step and slew correction approaches where:
1. Step correction is applied on EVERY time successful synchronization with a time server. This immediately corrects the system time to match server time.
2. Slew rate is calculated ONCE (on the second successful time synchronization with server) and applied henceforth, on every time polling interval period to compensate gradually for the system clock drift during the period.
This PR makes the following changes in the coreSNTP demo:
To suggest the most secure way of using SNTP communication, we are adding a documentation note about the Network Time Security in the SNTP demo.
Hygiene update of using the FreeRTOS/backoffAlgorithm utility library for calculating time polling period interval backoff when time requests are rejected by a time server.
Issue
There is a possible vulnerability of Denial of Service attack by keeping the UDP socket for the SNTP client task always open in the coreSNTP demo. The Denial of Service attack can occur from receiving multiple server response (duplicated or malicious) for a single SNTP time request sent by the client, and thereby, filing the socket network buffer response packets that affect future time requests.
Solution
This PR fixes this vulnerability by updating the demo to keep a UDP socket open only for the time period of waiting for server response, closing the socket on either receiving a server response or experiencing server timeout, and re-creating a UDP socket for the next polling try. This PR also adds another security functionality using a random port for UDP socket to protect against spoofing attacks from "off-network path" attackers.
* Fix code to match comment; idle prioriry task should not block to yield, as then it will hardly run at all (bug affects SMP tests)
* Add RP2040 Demo - can be used for 'main' or 'smp' branch of FreeRTOS_Kernel
* move to latest main/ of FreeRTOS-Kernel
* move RP2040 demos into Community-Supported
* Update whitespace issue
Co-authored-by: Gaurav-Aggarwal-AWS <33462878+aggarg@users.noreply.github.com>
* Add SMP to lexicon
Co-authored-by: Gaurav-Aggarwal-AWS <33462878+aggarg@users.noreply.github.com>
Mark Tuttle
tianmc1
Gaurav-Aggarwal-AWS
chinglee-iot
JD
Archit Aggarwal
tianmc1
lienng_aws
jasonpcarroll
Shivangi
alfred gedeon
Oscar Michael Abrina
Graham Sanderson
thomas-coding
Alfred Gedeon