netutils/mdns: raise responder stack size default to 8KiB

[ricardgb]

Summary

CONFIG_NETUTILS_MDNS_STACKSIZE defaulted to DEFAULT_TASK_STACKSIZE,
which resolves to as little as 2KiB on some targets. The bundled mjansson
mDNS responder needs a deeper stack than that — DNS record parsing with
name decompression, plus the send/receive buffers — so on those targets it
overflows its stack and hardfaults at startup, before it can answer a
single query.

This raises the default to 8KiB (the value that reliably runs the
responder) and corrects the help text, which previously claimed a "4KiB"
default that did not match DEFAULT_TASK_STACKSIZE.

Impact

• Fixes a startup hardfault when the mDNS app/daemon is enabled on targets
  with a small DEFAULT_TASK_STACKSIZE.
• No API change; configurations that already set an explicit stack size are
  unaffected.

Testing

Reproduced and fixed on a Raspberry Pi Pico 2 W (RP2350, NuttX). With the
old default (2KiB on this target) the responder overflowed its stack and
hardfaulted at startup; with the new 8KiB default it runs and answers mDNS
queries normally (verified with avahi-resolve/avahi-browse from a host).

Here are on-hardware test logs from a Raspberry Pi Pico 2 W (RP2350), NuttX 13.0.0-RC0.

The mDNS responder was built with CONFIG_NETUTILS_MDNS_STACKSIZE=8192 — the value this PR makes the default — and run over a USB CDC-NCM link (board eth0 192.168.7.1 ⇆ Linux host usb0 192.168.7.2).

Board — responder starts and answers a live query:

nsh> mdnsd_start
Local IPv4 address: 192.168.7.1
Opened 1 socket for mDNS service
Service mDNS: _http._tcp.local.:80
Hostname: nuttx
Sending announce
...
Query PTR _http._tcp.local.
  --> answer nuttx._http._tcp.local. (multicast)

Board — responder task running at the 8 KiB stack, no overflow (ps):

  TID   PID  PPID PRI POLICY   TYPE    NPX STATE    EVENT     SIGMASK          STACK COMMAND
    8     8     0 100 RR       Task      - Waiting  Semaphore 000...0000       0008080 --service _http._tcp.local --hostname nuttx --port 80

(The responder task shows an ~8 KiB stack; other tasks are ~2 KiB.)

Host — discovery / resolution / reachability:

$ avahi-browse -rt _http._tcp
=   usb0 IPv4 nuttx                     Web Site         local
   hostname = [nuttx.local]
   address = [192.168.7.1]
   port = [80]
   txt = ["other=value" "test=1"]

$ avahi-resolve -n nuttx.local
nuttx.local	192.168.7.1

$ ping nuttx.local          # name resolved via the board's responder
64 bytes from 192.168.7.1: icmp_seq=3 ttl=64 time=16.9 ms
3 packets transmitted, 3 received, 0% packet loss

The responder runs, advertises, and answers queries at the 8 KiB stack without overflowing.

---

Disclosure: this change was prepared by an AI agent (Claude Code) at the
direction of the author (@ricardgb), who reviewed it before submission.

🤖 Generated with Claude Code

[Abhishekmishra2808]

@ricardgb, please sign your commit message with git commit -s (https://nuttx.apache.org/docs/latest/contributing/guide.html#commit-requirements).
Also, rebase the branch; then the checks will pass.

[linguini1]

Please use the Assisted-by: field from the contribution guide. Also, please attach your test logs from your testing on the Pico.

[ricardgb]

@linguini1 commit trailer updated to Assisted-by: per the contribution guide (and rebased onto the latest master). Test logs from the Pico will follow in a separate comment.

[ricardgb]

@linguini1 here are on-hardware test logs from a Raspberry Pi Pico 2 W (RP2350), NuttX 13.0.0-RC0.

The mDNS responder was built with CONFIG_NETUTILS_MDNS_STACKSIZE=8192 — the value this PR makes the default — and run over a USB CDC-NCM link (board eth0 192.168.7.1 ⇆ Linux host usb0 192.168.7.2).

Board — responder starts and answers a live query:

nsh> mdnsd_start
Local IPv4 address: 192.168.7.1
Opened 1 socket for mDNS service
Service mDNS: _http._tcp.local.:80
Hostname: nuttx
Sending announce
...
Query PTR _http._tcp.local.
  --> answer nuttx._http._tcp.local. (multicast)

Board — responder task running at the 8 KiB stack, no overflow (ps):

  TID   PID  PPID PRI POLICY   TYPE    NPX STATE    EVENT     SIGMASK          STACK COMMAND
    8     8     0 100 RR       Task      - Waiting  Semaphore 000...0000       0008080 --service _http._tcp.local --hostname nuttx --port 80

(The responder task shows an ~8 KiB stack; other tasks are ~2 KiB.)

Host — discovery / resolution / reachability:

$ avahi-browse -rt _http._tcp
=   usb0 IPv4 nuttx                     Web Site         local
   hostname = [nuttx.local]
   address = [192.168.7.1]
   port = [80]
   txt = ["other=value" "test=1"]

$ avahi-resolve -n nuttx.local
nuttx.local	192.168.7.1

$ ping nuttx.local          # name resolved via the board's responder
64 bytes from 192.168.7.1: icmp_seq=3 ttl=64 time=16.9 ms
3 packets transmitted, 3 received, 0% packet loss

The responder runs, advertises, and answers queries at the 8 KiB stack without overflowing.

Disclosure: these tests were run and captured with the help of an AI agent (Claude Code, Anthropic); results are from real hardware and human-reviewed.

[cederom]

Thank you @ricardgb :-) Please remember to always attach test logs to the PR description, before the problem (segfault) and after (showing fix working).. I have copy-pasted it to the PR description from a discussion :-)

[raiden00pl]

Do we really need a description of the problem in the help message? The option description should describe the options and that's it. Please look more closely at what Claude generates.