netutils/thttpd: fix broken cgi and fixup kconfig

[TimJTi]

Summary

When trying to enable and use the thttp net utility, it was found that CGI apps didnt send any http responses of data from the CGI app back to the browser client,

Investigation revealed this was due to the net socket being opened with the O_CLOEXEC flag set.

This is the result, I believe, of a PR last year here.

See this issue.

In correcting the issue, some complier warnings were encountered and fixed, along with a Kconfig error and an omission, and the opportunity was taken to improve some help text within Kconfig.

This is why there are two commits in the PR.

Question

The cgi_child app that is spawned closes all open file descriprors except stdin, stdout, stderr and the socket's descriptor, up to a maximum number of file descriptors as set via Kconfig.

See here.

Should the Kconfig option be removed and the function use OPEN_MAX instead?

Impact

No impact on existing usage of this netutil, other than to allow it work properly now!

Testing

This has been tested as part of my own application development, using a custom SAMA5D27C-D1G processor, with the webserver running via CDC-NCM.

[cederom]

Thank you @TimJTi :-)

[xiaoxiang781216]

@TimJTi please fix nuttx side:

====================================================================================
Configuration/Tool: lincoln60/thttpd-binfs,CONFIG_ARM_TOOLCHAIN_GNU_EABI
2025-07-02 11:49:28
------------------------------------------------------------------------------------
  Cleaning...
  Configuring...
  Disabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Enabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Building NuttX...
  [1/1] Normalize lincoln60/thttpd-binfs
64d63
< CONFIG_THTTPD_CGIINBUFFERSIZ=256
Saving the new configuration file
On branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
	modified:   boards/arm/lpc17xx_40xx/lincoln60/configs/thttpd-binfs/defconfig

[TimJTi]

@TimJTi please fix nuttx side:

====================================================================================
Configuration/Tool: lincoln60/thttpd-binfs,CONFIG_ARM_TOOLCHAIN_GNU_EABI
2025-07-02 11:49:28
------------------------------------------------------------------------------------
  Cleaning...
  Configuring...
  Disabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Enabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Building NuttX...
  [1/1] Normalize lincoln60/thttpd-binfs
64d63
< CONFIG_THTTPD_CGIINBUFFERSIZ=256
Saving the new configuration file
On branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
	modified:   boards/arm/lpc17xx_40xx/lincoln60/configs/thttpd-binfs/defconfig

@xiaoxiang781216
Oops!

Can I do it as a third commit in this PR or do I need to do it via a new PR in the nuttx repo?

[cederom]

Different repo different commit/pr :-)

[xiaoxiang781216]

@TimJTi please fix nuttx side:

====================================================================================
Configuration/Tool: lincoln60/thttpd-binfs,CONFIG_ARM_TOOLCHAIN_GNU_EABI
2025-07-02 11:49:28
------------------------------------------------------------------------------------
  Cleaning...
  Configuring...
  Disabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Enabling CONFIG_ARM_TOOLCHAIN_GNU_EABI
  Building NuttX...
  [1/1] Normalize lincoln60/thttpd-binfs
64d63
< CONFIG_THTTPD_CGIINBUFFERSIZ=256
Saving the new configuration file
On branch master
Your branch is up to date with 'origin/master'.

Changes not staged for commit:
  (use "git add <file>..." to update what will be committed)
  (use "git restore <file>..." to discard changes in working directory)
	modified:   boards/arm/lpc17xx_40xx/lincoln60/configs/thttpd-binfs/defconfig

@xiaoxiang781216 Oops!

Can I do it as a third commit in this PR or do I need to do it via a new PR in the nuttx repo?

please provide pr to nuttx git, I will merge both patch together.

[TimJTi]

@xiaoxiang781216
I'm still seeing the compiler warnings that were fixed in one of the commits here. I wonder if the merge sequence was such that the warnings were "undone" by the actual real fix?

Please advise if you need me to do anything to sort this as I'm sure it was down to me LOL

[xiaoxiang781216]

@xiaoxiang781216 I'm still seeing the compiler warnings that were fixed in one of the commits here. I wonder if the merge sequence was such that the warnings were "undone" by the actual real fix?

Please advise if you need me to do anything to sort this as I'm sure it was down to me LOL

let's see what' happen when the new PR come in.

[TimJTi]

@xiaoxiang781216 I'm still seeing the compiler warnings that were fixed in one of the commits here. I wonder if the merge sequence was such that the warnings were "undone" by the actual real fix?
Please advise if you need me to do anything to sort this as I'm sure it was down to me LOL

let's see what' happen when the new PR come in.

@xiaoxiang781216 - what new PR?

[xiaoxiang781216]

If the new PR doesn't have the build break, your change is fine

[TimJTi]

If the new PR doesn't have the build break, your change is fine

Sorry - I hadn't properly rebased by branch to latest upstream so it had a mix of upstream/master and my own changes that were conflicted. All is good.