NRF52: add a separate .nvictable section and allow .noinit to be used…

[drahnr]

Description

Creates a separate section for the NVIC relocated vector table.
Yields the common .noinit for application use cases.

Pull request type

[X] Fix
[ ] Refactor
[ ] New target
[ ] Feature
[X] Breaking change

This could be potentially breaking if an app does some dynamic vector table modifcation or similiar.

[pan-]

@drahnr Thanks for your submission.

Could you comment on why that would be a breaking change ? To me it seems that position and the symbol name of the vector tables remain untouched; the only change is the name of the section holding the vector table.

IAR and ARMCC linker scripts needs to be updated as well.

As a side note it would be appreciated if you backport these changes to NRF51 too.

[drahnr]

The breaking change might be applications which exploit the defined symbols based on __start_noinit which is not the NVIC table anymore, but an indepentend section.

I will look into the other linker files, but I can't promise much there since I don't have any setups for IAR nor ARMCC nor any licenses for either nor experience writing linke files for them.

[drahnr]

Also NRF51 does not have a dedicated section for the relocated table yet, so I am not sure this makes sense to tackle, especially since I can't even test any of those.

Latest commit appears to be addressing comments

[cmonr]

@pan- @marcuschangarm Mind taking another look?

[marcuschangarm]

The NRF52 looks good, thank you! I don't have enough experience with the NRF51 to review it.

[0xc0170]

/morph build

[drahnr]

@marcuschangarm there are no changes for NRF51 in this PR since I don't do any development with that platform nor are the changes directly applicable to the NRF51 without some additional preliminay work

[marcuschangarm]

You made changes to https://github.com/drahnr/mbed-os/blob/367d95e17e14aaaa41058be41be72dd445c860e2/targets/TARGET_NORDIC/TARGET_NRF5x/TARGET_NRF51/reloc_vector_table.c ?

I don't think .nvictable is defined in any of the NRF51 linker scripts.

[drahnr]

@marcuschangarm I totally messed that up, I am removing those accidentally commited changes and things should be fine.

[mbed-ci]

Build : SUCCESS

Build number : 2401
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7145/

Triggering tests

/morph test
/morph uvisor-test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2036
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7145/

[cmonr]

/morph mbed2-build

[mbed-ci]

Test : FAILURE

Build number : 2187
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7145/2187

[cmonr]

Re-enqueueing since it looks like we hit a device allocation issue.

/morph test

[marcuschangarm]

@cmonr I think we are still waiting for changes.

[cmonr]

@marcuschangarm Ah, crud. Yeah, looks like it.
Just saw the failure come up, not the preceeding comments.

[mbed-ci]

Test : FAILURE

Build number : 2192
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7145/2192

[drahnr]

Removed NRF51 related changes from the last commit.

[cmonr]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 2417
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7145/

Triggering tests

/morph test
/morph uvisor-test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2048
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7145/

[cmonr]

/morph mbed2-build

[mbed-ci]

Test : SUCCESS

Build number : 2199
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7145/2199