sys/net/nanocoap: fix buffer overflow in separate response handling [backport 2024.10]

[maribu]

(Partial) Backport of #21075

Contribution description

When RFC 8974 support (module nanocoap_token_ext) is in use, the request token may be longer than the buffer in the separate response context is large. This adds a check to not overflow the buffer.

Sadly, this is an API change: Preparing the separate response context can actually fail, so we need to report this with a return value.

The example application has been adapted to only proceed if the separate reply context could have been prepared, and rather directly emit a reset message if the token exceeds the static buffer.

Testing procedure

This requires a client from master, as extending the test to allow specifying the token is not backported.

Preparation

$  sudo ./dist/tools/tapsetup/tapsetup 

Running the Server

This should be done with 2024.10-branch and this PR

$ USEMODULE=nanocoap_token_ext make BOARD=native64 -C examples/nanocoap_server && examples/nanocoap_server/bin/native64/nanocoap_server.elf -w tap1
[...]
main(): This is RIOT! (Version: 2025.01-devel-261-g75828)
RIOT nanocoap example application
Waiting for address autoconfiguration...
{"IPv6 addresses": ["<IPV6_ADDR>"]}

copy the <IPV6_ADDR> part of the output

Running the Client

This should be done with master.

$ USEMODULE=nanocoap_token_ext make BOARD=native64 -C tests/net/nanocoap_cli flash term
[...]
2024-12-11 18:10:58,750 # RIOT native interrupts/signals initialized.
2024-12-11 18:10:58,751 # RIOT native board initialized.
2024-12-11 18:10:58,751 # RIOT native hardware initialization complete.
2024-12-11 18:10:58,751 # 
2024-12-11 18:10:58,752 # main(): This is RIOT! (Version: 2025.01-devel-262-g562036-sys/net/nanocoap/buffer-overflow-separate-response)
2024-12-11 18:10:58,752 # nanocoap test app
2024-12-11 18:10:58,752 # All up, running the shell now
> client_token deadbeefdeadbeefdeadbeef
2024-12-11 18:12:11,060 # client_token deadbeefdeadbeefdeadbeef
> client get fe80::24de:d0ff:fe85:bdb2 5683 /separate
2024-12-11 18:12:27,736 # client get fe80::24de:d0ff:fe85:bdb2 5683 /separate
2024-12-11 18:12:27,736 # nanocli: sending msg ID 1, 25 bytes
2024-12-11 18:13:45,417 # nanocli: msg send failed: -110

Expected result: No segfault in the server

Issues/PRs references

Partial backport of #21075

[maribu]

I did test this with native64, but without cherry-pick 2b6f65a the XFA bug triggers and the contents of the XFA resources are misaligned.

I guess I should backport 2b6f65a as well, then 😅

With that fix cherry picked, the test did succeed, so the buffer overflow is fixed as promised.

[riot-ci]

Murdock results

✔️ PASSED

2da20b3 sys/net/nanocoap: fix buffer overflow in separate response handling

Success	Failures	Total	Runtime
10215	0	10215	19m:39s

Artifacts

• Documentation preview

[maribu]

Thx :-)