Repository navigation
Sync csa branch with main - #1269
Merged
Merged
Conversation
…CodeDrivenDataModelProvider (#74656) * [DataModel] Notify endpoint changes on runtime AddEndpoint/RemoveEndpoint in CodeDrivenDataModelProvider * Notify endpoint changes unconditionally in AddEndpoint/RemoveEndpoint * Log cluster Startup failures in AddEndpoint without failing endpoint registration * Log endpoint and cluster ID on Startup failure and fix ASAN stack-use-after-return in test
Stop forcing chip_crypto_psa_aead_single_part for ESP-IDF >= 6, so the default multi-part PSA AEAD path is used. Single-part mode caps AES-CCM payloads at the UDP MTU: encrypt and decrypt go through a stack buffer of CHIP_CONFIG_DEFAULT_UDP_MTU_SIZE + 16 bytes and fail with CHIP_ERROR_INVALID_ARGUMENT for anything larger, which breaks large-payload (TCP) sessions that advertise MAX_TCP_MESSAGE_SIZE = 64000.
…uto-merged by platform-bot) (#74679) The ESP32 example OTA provider stores the fabric index and node id of the requester that issues QueryImage, but never checks that the peer initiating the subsequent BDX transfer is that same requester. Override BdxOtaSender::OnMessageReceived to compare the initiating session's fabric index and node id against the stored requester, rejecting a mismatch with CHIP_ERROR_INVALID_DESTINATION_NODE_ID. Ports #74130 to the ESP32 example. Fixes part of #74131
…ter (#74684)
* ResourceMonitoring: Pass cluster ID and revision together to the cluster
ResourceMonitoringCluster::ReadAttribute chose ClusterRevision with an
if-chain that fell back to ActivatedCarbonFilterMonitoring::kRevision for
any cluster it did not recognise, so a newly aliased cluster such as
WaterTankLevelMonitoring would report the wrong revision.
Follow the ModeBase design: add ResourceMonitoring::ClusterEntry
{ id, revision } with one constant per aliased cluster, take it in the
ResourceMonitoringCluster constructor and encode the stored revision.
Instance keeps its ClusterId-based signature, so existing callers are
unaffected. It maps the ID to a ClusterEntry and dies on an unknown ID.
Reported revisions for HepaFilterMonitoring and
ActivatedCarbonFilterMonitoring are unchanged.
The ResourceMonitoringCluster constructor signature changed (ClusterId
-> ClusterEntry); only Instance and the unit test use it in this tree.
Tests: add per-cluster ClusterRevision checks and pin the ClusterEntry
constants to the current revisions.
* [pre-commit.ci] auto fixes from pre-commit.com hooks
for more information, see https://pre-commit.ci
---------
Co-authored-by: pre-commit-ci[bot] <66853113+pre-commit-ci[bot]@users.noreply.github.com>
* Fixing BDX send path * Fixing MEDIAPLAYBACK 6.12 test * Fixing the MEDIAPLAYBACK 6.11 test * Fixing the CONTENTLAUNCHER 10.10 test * Fixing the Media File Management 2.2 test * Fixing the MEDIAFILEMANAGEMENT 3.1 test * Fixing the MESS 3.1 test * Fixing the MESS 3.2 test * Fixing the MESS 3.3 test * Fixing the Media File Management 2.1 test * Fixing the CONTENTLAUNCHER 10.8 test * Removing Step 5 of Content Launcher 10.8 to match the updated test plan * Matching the test script error to the updated test plan for MEDIAFILEMANAGEMENT 2.2 * Clarifying the applicability of some steps in MESS 3.2 * Ensuring we have StartTime and Duration coverage * Ensuring we cover the NULL case in MESS 3.3 * Clearing up the mandatory/optional behavior around SupportedMimeTypes * Fixing build error * clang-format errors for PR automation * autopep8 errors from PR automation * clang-format updates with the correct version * Fixing test failure reported in issue 74211 * coderabbit suggestions * coderabbit comment * Additional coderabbit comment * MESS 3.3 fix * Adopt MatterTestCommissionedDevice marker from master * autopep8 fixes * coderrabit comment
SplitArgs() doubles argList when it fills, but passed the new entry count to MemoryRealloc(), which takes a size in bytes. Once the list outgrew its initial ten entries, the "doubled" buffer held 20 bytes instead of 20 pointers, and the next argList write and the trailing NULL went out of bounds. Any ParseArgsFromString() input with nine or more tokens reached it. Multiply by sizeof(char *), and refuse to double when that would overflow size_t. Signed-off-by: Justin Wood <woody@apple.com>
test018 and test021 depend on the connectivity monitor not reporting for the peer controller. The tests removed the peer's advertisement with DNSServiceRefDeallocate alone, and dns-sd can still answer a resolve for the instance for up to about a second after that. When the monitor resolved first, test018's invoke queued behind the subscription pool slot the test holds, and test021's session came through the pool rather than the timeout. test021 also held the Matter queue from a user-initiated thread blocked on a semaphore that the main thread signals. Thread Performance Checker reports this as a priority inversion. In the runs where it did, the main thread's predicate wait returned after its 5 s timeout without seeing a check that had already run. - After setting up the session to the peer, wait until a browse sees the peer's instance removed on every interface it was seen on. - test021 starts the session request from inside a Matter queue block and waits there, so the timeout is still queued between the request and the check, and no other thread waits on the Matter queue.
* Fix for 917 provisionCrypto with provision flash * [Silabs] Fix 917 Flash provisionStorage. Add condition statement to fix slc based builds * Restyled by gn --------- Co-authored-by: Restyled.io <commits@restyled.io>
Contributor
|
The check for |
Junior Martinez (jmartinez-silabs)
approved these changes
Oct 11, 2026
sarthak shaha (Sarthak-Shaha)
approved these changes
Oct 12, 2026
Junior Martinez (jmartinez-silabs)
merged commit Oct 12, 2026
4dbaa06
into
main
112 of 115 checks passed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This PR syncs the csa branch with the main branch.
PR MUST BE MERGED WITH MERGE COMMIT - ADMIN MUST ENABLE THE OPTION