Disable CRMP if we send/receive messages not over UDP transport

[yufengwangca]

Problem

I found CRMP is always enabled regardless of low level transport for application dispatch. This is regression from recent changes.

Currently, we use IsTransportReliable() to decide if to enable CRMP, this logic is wrong since TransportMgr is a tuple which can support TCP and UDP simultaneously.

We need to use the destination address instead to decide if use CRMP:

1. For PASE/CASE message exchange, we don't have PeerConnectionState, we can relay on the SessionEstablishmentExchangeDispatch to do this check since it has the peer destination address.

2. For Application message exchange, we can get the peer address form PeerConnectionState, and do the check.

Summary of Changes

Disable CRMP if we send/receive messages not over UDP transport

Fixes #6026

[woody-apple]

@pan-apple ?

How is this different than this? #6317

[yufengwangca]

@pan-apple ?

How is this different than this? #6317

#6317 address the CASE and PASE messages over SessionEstablishmentExchangeDispatch, but not address the application message over ApplicationExchangeDispatch

[pan-apple]

@pan-apple ?
How is this different than this? #6317

#6317 address the CASE and PASE messages over SessionEstablishmentExchangeDispatch, but not address the application message over ApplicationExchangeDispatch

Yes, this is to support Application messages over TCP (and disable CRMP for such messages).

[github-actions]

Size increase report for "esp32-example-build" from 6df63b4

File	Section	File	VM
chip-all-clusters-app.elf	.dram0.bss	0	8
chip-all-clusters-app.elf	.flash.rodata	8	8
chip-all-clusters-app.elf	.flash.text	-60	-60

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

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
.debug_info,0,76524
.debug_line,0,9554
.debug_abbrev,0,7452
.xt.prop._ZN4chip9Messaging23ExchangeMessageDispatch29IsReliableTransmissionAllowedEv,0,156
.debug_str,0,135
.strtab,0,131
.xt.prop._ZN4chip36SessionEstablishmentExchangeDispatch29IsReliableTransmissionAllowedEv,0,76
.debug_ranges,0,56
.debug_loc,0,28
.shstrtab,0,21
.symtab,0,16
.dram0.bss,8,0
.flash.rodata,8,8
.xt.prop._ZTVN4chip24LifetimePersistedCounterE,0,-1
.debug_aranges,0,-8
[Unmapped],0,-8
.debug_frame,0,-24
.flash.text,-60,-60
.xt.prop._ZN4chip36SessionEstablishmentExchangeDispatch19IsTransportReliableEv,0,-76
.xt.prop._ZN4chip9Messaging23ExchangeMessageDispatch19IsTransportReliableEv,0,-156

[github-actions]

Size increase report for "nrfconnect-example-build" from 6df63b4

File	Section	File	VM
chip-lighting.elf	text	332	332
chip-lighting.elf	rodata	184	184
chip-lighting.elf	bss	0	32
chip-lighting.elf	init_array	4	4
chip-shell.elf	text	12	12
chip-shell.elf	device_handles	4	4
chip-lock.elf	text	336	336
chip-lock.elf	rodata	184	184
chip-lock.elf	bss	0	32
chip-lock.elf	device_handles	12	12
chip-lock.elf	init_array	4	4

Full report output

BLOAT REPORT

Files found only in the build output:
    report.csv

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

sections,vmsize,filesize
.debug_info,0,407163
.debug_line,0,24599
.debug_abbrev,0,14674
.debug_loc,0,2205
.debug_str,0,1919
.strtab,0,676
.symtab,0,480
.debug_ranges,0,432
.debug_frame,0,352
text,332,332
rodata,184,184
.debug_aranges,0,168
bss,32,0
init_array,4,4

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

sections,vmsize,filesize
.debug_info,0,67
.debug_str,0,31
.debug_line,0,29
text,12,12
.strtab,0,10
device_handles,4,4
.shstrtab,0,-2
.debug_ranges,0,-8
.debug_loc,0,-87

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

sections,vmsize,filesize
.debug_info,0,199512
.debug_line,0,24120
.debug_abbrev,0,16064
.debug_loc,0,2209
.debug_str,0,1919
.strtab,0,676
.symtab,0,480
.debug_ranges,0,432
.debug_frame,0,352
text,336,336
rodata,184,184
.debug_aranges,0,168
bss,32,0
device_handles,12,12
init_array,4,4

[woody-apple]

@saurabhst @msandstedt ?