Get peer's node ID from session instead of packet header for response handling

[pan-apple]

Problem

The reports from peer node are getting dropped when received on the controller.

CHIP: [CTL] OnMessageReceived was called for unknown source node

Change overview

The unicast encrypted messages do not carry node IDs in packet headers. The controller code was using the source node ID to identify which device sent the message. This PR updates it to use session information to identify the device object.

Testing

A new test (testSendClusterTestCluster_Reporting_0002_ReportOnOff_Test) has been added to ensure that reports are received and processed on the controller.

[woody-apple]

@saurabhst @msandstedt @mspang @andy31415 @Damian-Nordic @LuDuda ?

[kghost]

I think this is part of PR #8726

Would you please help me review that PR, it fixes lots of incorrect usage of packet header like this one.

[andy31415]

For my own knowledge: how is reporting used today? I am looking through the spec and see reporting mentioned as a transaction for subscriptions, however I also am aware there is a lot of back and forth regarding how subscriptions are implemented.

How does reporting work right now, how much is based on spec and what do the tests validate?

[andy31415]

Any chance to add tests beyond end-to-end in darwin?

End to end tests are slower and darwin in particular is limited to only 5 CI runners so it is slower than any linux tests.

[andy31415]

• note: will not block PR on spec or testing - just please update the typo and we can merge.

I would like to express some concern about darwin-only tests because slow and ideally maybe some way to record what is spec and what is temporary until it better evolves.

[sagar-apple]

For my own knowledge: how is reporting used today? I am looking through the spec and see reporting mentioned as a transaction for subscriptions, however I also am aware there is a lot of back and forth regarding how subscriptions are implemented.

How does reporting work right now, how much is based on spec and what do the tests validate?

I can describe what reporting does today. Currently some clusters support configuring reports on a specific attribute. When reports are configured the server sends a report of the attributes state out to each "bound" peer.
I'm not sure if the current implementation is spec compliant but reports have been functioning as described until recently.

The added Darwin tests are just setting up these reports on the OnOff Cluster and making sure at least one report is sent out.
In sequence they are testing that the

• the controller can bind to the OnOff Cluster on accessory
• the controller can request to configure a reporting interval for the OnOff attribute
• the controller is able to receive at least 1 report
• the controller is able to disable reporting.

[github-actions]

Size increase report for "esp32-example-build" from f4c307f

File	Section	File	VM
chip-temperature-measurement-app.elf	.flash.text	-60	-60

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

Comparing ./master_artifact/chip-shell.elf and ./pull_artifact/chip-shell.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-lock-app.elf and ./pull_artifact/chip-lock-app.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-persistent-storage.elf and ./pull_artifact/chip-persistent-storage.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-bridge-app.elf and ./pull_artifact/chip-bridge-app.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-ipv6only-app.elf and ./pull_artifact/chip-ipv6only-app.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-pigweed-app.elf and ./pull_artifact/chip-pigweed-app.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-all-clusters-app.elf and ./pull_artifact/chip-all-clusters-app.elf:

sections,vmsize,filesize

Comparing ./master_artifact/chip-temperature-measurement-app.elf and ./pull_artifact/chip-temperature-measurement-app.elf:

sections,vmsize,filesize
[Unmapped],0,60
.flash.text,-60,-60

[mspang]

For my own knowledge: how is reporting used today? I am looking through the spec and see reporting mentioned as a transaction for subscriptions, however I also am aware there is a lot of back and forth regarding how subscriptions are implemented.
How does reporting work right now, how much is based on spec and what do the tests validate?

I can describe what reporting does today. Currently some clusters support configuring reports on a specific attribute. When reports are configured the server sends a report of the attributes state out to each "bound" peer.
I'm not sure if the current implementation is spec compliant but reports have been functioning as described until recently.

The added Darwin tests are just setting up these reports on the OnOff Cluster and making sure at least one report is sent out.
In sequence they are testing that the

• the controller can bind to the OnOff Cluster on accessory
• the controller can request to configure a reporting interval for the OnOff attribute
• the controller is able to receive at least 1 report
• the controller is able to disable reporting.

This is so different from spec that we didn't even think it was worth enabling it in the Android controller.

See #8757