Using the cache for function bindings and not deleting shared memory resources upon request to close

[gohar94]

Description

When the host sends a message to close shared memory resources, instead of deleting the shared memory resources (i.e. files in the case of Linux) just drop the handle and let the worker delete the resources.

When the host advertises the capability for using FunctionDataCache, use that even for small objects (without imposing the minimum size constraint.)

This PR is part of the work described in Azure/azure-functions-host#7310

---

PR information

• The title of the PR is clear and informative.
• There are a small number of commits, each of which has an informative message. This means that previously merged commits do not appear in the history of the PR. For information on cleaning up the commits in your pull request, see this page.
• If applicable, the PR references the bug/issue that it fixes in the description.
• New Unit tests were added for the changes made and CI is passing.

Quality of Code and Contribution Guidelines

• I have read the contribution guidelines.

[vrdmr]

minor comments. Also, can you please add testing for cache-enabled scenarios? to see if the same shared mem is reused or not?

[gohar94]

minor comments. Also, can you please add testing for cache-enabled scenarios? to see if the same shared mem is reused or not?

That's a very good point.
The problem right now is that the caching piece will not be enabled until one of the azure storage extension PRs is checked in. That PR will not be checked in until the azure-webjobs-SDK PR package is published to nuget.org - this will happen in a few weeks as part of some deployment cycle which I don't fully understand. Once that is checked in, then we can have a subsequent PR here for the Python worker which consumes the latest host and storage extension in the tests so that we can put the tests you mentioned. Does that sounds like a fair plan?

[vrdmr]

nit changes. otherwise, LGTM. 🚢

[vrdmr]

Suggested change

	if self._function_data_cache_enabled:
	# If the cache is enabled, let the host decide when to
	# delete the resources.
	# Just drop the reference from the worker.
	to_delete = False
	else:
	# If the cache is not enabled, the worker should free
	# the resources as at this point the host has read the
	# memory maps and does not need them.
	to_delete = True
	to_delete_reference = False if self._function_data_cache_enabled else True

And please add the comment in the python docs section of this method itself.

[vrdmr]

Suggested change

	success = self._shmem_mgr.free_mem_map(map_name, to_delete)
	success = self._shmem_mgr.free_mem_map(map_name, to_delete_reference)

If above suggestion is included, please include this and also check if there is any other place to_delete is mentioned.