netsocket: several dynamic allocation results not checked

[pea-pod]

Description of defect

There are several netsocket source files that dynamically allocate without std::nothrow or otherwise checking for allocation failure. Here is the list of files, excluding tests:

• NetworkInterface.cpp: 148
• nsapi_dns.cpp: 985
• TCPSocket.cpp 269
• TLSSocketWrapper.cpp: 137, 289, 430, 462, 572

Target(s) affected by this defect ?

All with onboard networking

Toolchain(s) (name and version) displaying this defect ?

N/A

What version of Mbed-os are you using (tag or sha) ?

mbed-os-6.7.0

What version(s) of tools are you using. List all that apply (E.g. mbed-cli)

N/A

How is this defect reproduced ?

Any allocation failure should result in crashing either in allocation failure or attempting to use an unallocated resource. For the most part, the failure should be handled and returned as an error instead.

[pea-pod]

I first noticed this while trying to figure out if it would be possible inherit TCPSocket to override accept to use a MemoryPool instead of actual dynamic allocation. I wanted to allocate memory statically and not while running the server if possible.

I realized that the accept method never checks to see if the new call to the protected constructor and could fail. It is possible that I've experienced this while testing out an idea and never realized it because I did not want to get the haruspex diagram out (my hepatomancy license has lapsed).

[ciarmcom]

Thank you for raising this detailed GitHub issue. I am now notifying our internal issue triagers.
Internal Jira reference: https://jira.arm.com/browse/IOTOSM-3331