Call HAL free functions from C++ destructors

[kyle-cypress]

Summary of changes

Add missing calls to HAL free to the following drivers

• AnalogIn
• QSPI
• SerialBase

Impact of changes

Resources will be properly freed when using these drivers through the C++ layer.

Migration actions required

NA

Documentation

NA

---

Pull request type

[x] Patch update (Bug fix / Target update / Docs update / Test update / Refactor)
[] Feature update (New feature / Functionality change / New API)
[] Major update (Breaking change E.g. Return code change / API behaviour change)

---

Test results

[] No Tests required for this change (E.g docs only update)
[x] Covered by existing mbed-os tests (Greentea or Unittest)
[x] Tests / results supplied as part of this PR

The TIMEOUT failures are a test infrastructure issue; they do not repeat when rerun manually.
All other failures are known, pre-existing issues.

CY8CKIT_062_WIFI_BT-GCC_ARM-NET.txt
CY8CPROTO_062_4343W-GCC_ARM-NET.txt
CY8CPROTO_062S3_4343W-GCC_ARM-NET.txt
CYW9P62S1_43012EVB_01-GCC_ARM-NET.txt
GT_FT_KIT_062_BLE_GCC.txt
GT_FT_KIT_062_WIFI_BT_GCC.txt
GT_FT_KIT_062S2_43012_GCC.txt
GT_FT_P6S1_43012EVB_01_GCC.txt
GT_FT_P6S1_43438EVB_01_GCC_NET.txt
GT_FT_P6S1_43438EVB_01_GCC.txt
GT_FT_PROTO_062_4343W_GCC.txt
GT_FT_PROTO_062S3_4343W.txt
GT-FT-KIT-062S2-43012-GCC-NET.txt

---

Reviewers

---

@ARMmbed/team-cypress

[ciarmcom]

@kyle-cypress, thank you for your changes.
@ARMmbed/mbed-os-core @ARMmbed/mbed-os-maintainers please review.

[0xc0170]

I don't think this is intended to be in a patch (should be at least feature, but this could be major?).
We previously had similar attempts , have you reviewed them what challenges are there including these in destructors ? Couple of HAL were updated to with defined behavior, they should be safe to be called but the rest - I don't think so.

This makes sense to do, just need to understand the consequences here and how to make it backward compatible.

[kyle-cypress]

Do you know PR/issue #'s for the previous attempts to add destructors? I did a quick search and the closest I found was #10843 which says it was superceded by #10924, the latter of which does not modify the destructors, and it is not clear from the (lengthy) discussion whether that was a deliberate omission.

I recall discussions in the past around adding free() functions that were previously missing from the HAL layer, and I know care had to be taken there to not break existing HAL implementations that didn't define a free() function.

But I'm having trouble thinking of what issues would arise when updating the C++ layer to call free() functions that already exist in the HAL layer. The greentea tests should already be exercising those free() functions to make sure that they are safe to call.

[0xc0170]

Thanks for the references. That helps now.

@ARMmbed/mbed-os-hal Please review

[0xc0170]

I wonder how come this was not added in #10924 where we have added enable/disable output.

[0xc0170]

CI started

[mbed-ci]

Test run: SUCCESS

Summary: 6 of 6 test jobs passed
Build number : 1
Build artifacts

[adbridge]

Approved by @andypowers