NSAPI/lwIP: Free held netbuf on close

[kjbracey]

mbed_lwip_socket_recv() takes one netbuf at a time from the netconn API,
and it holds a partially-read netbuf if necessary in order to present as
a stream for TCP.

This held netbuf was not being freed when the socket was closed.

Fixes issue #3974.

[kjbracey]

One note on this netbuf thing for future consideration - the netbuf API isn't ideal for TCP, as TCP doesn't internally hold netbufs, so netconn_recv() itself allocates a temporary netbuf just to make TCP match UDP and raw.

As we know this path is TCP-only, you could use netconn_recv_tcp_pbuf to skip that step, and just work on the pbufs.

[0xc0170]

/morph test-nightly

[mbed-bot]

Result: SUCCESS

Your command has finished executing! Here's what you wrote!

/morph test-nightly

Output

mbed Build Number: 1709

All builds and test passed!

[0xc0170]

Restarted jenkins CI, please look at the result once done

The current failures are 3 identified, one of them for instance ./mbed-client/source/m2mbase.cpp:60:23: error: 'M2MBase::lwm2m_parameters_s' has no member named 'dynamic_resource_params' , is this valid? As this patch does not touch that code

[kjbracey]

The branch was deliberately created based on an old revision, partly to test whether your CI was correctly testing the merge result with master, or just testing the tip of the branch.

It seems it might just be testing the tip. I could rebase the branch, but maybe you want to check your CI configuration?

[tommikas]

@kjbracey-arm The mbed-client part of the jenkins/pr-head build doesn't currently do the merge. There's a task to add it, but I'm not sure what the status of it is at the moment.

[tommikas]

@kjbracey-arm We'll fix it soon (TM). If you're in a hurry with this PR I'd suggest rebasing.

[kjbracey]

I'm not personally in a hurry - it's just to address the GitHub issue referenced - a slightly hard-to-cause memory leak.

So I'm happy to leave it for the moment as a testcase.

[geky]

Looks good to me 👍

[kjbracey]

@0xc0170 I believe Jenkins has been fiddled with successfully. Could you retrigger?

[tommikas]

@kjbracey-arm @0xc0170 I just restarted the job.