enable os5 for MOTE_L152RC

[dudmuck]

Description

Enable mbed-os5 for target MOTE_L152RC.
Tested using: mbed test -t GCC_ARM -m MOTE_L152RC -n mbed-os-*
Removed "default_lib": "small" for this target because c_string float printf was failing.

Pull request type

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

[cmonr]

@dudmuck Thanks for the PR. Do you think you could post the results for the test here? A text file would be fine.

Also, have you tested against the ARM and IAR compilers?

[cmonr]

@Patater Mis-clicked when assigning reviewers. Feel free to dismiss yourself or ignore this PR.

[cmonr]

@dudmuck Making sure, is this targeted for the NaMote-72 (https://os.mbed.com/platforms/NAMote-72/)? I didn't seem to find the MOTE_L152RC anywhere, and think I'm missing something.

[dudmuck]

Yes, TARGET_MOTE_L152 is NAMote-72 platform. I have tested using ARM GCC, but i dont have armcc for cortex-m3 or IAR ARM.
I tested using -n mbed-os-*
All was passing except c_string float printf because default_lib was small.

[cmonr]

@hasnainvirk Thoughts?

[0xc0170]

All was passing except c_string float printf because default_lib was small.

Should pass. As this has uARM that is not thread safe and it's very small target, I do not think enabling mbed OS 5 for this platform makes sense.

[0xc0170]

Should pass. As this has uARM that is not thread safe and it's very small target, I do not think enabling mbed OS 5 for this platform makes sense.

@ARMmbed/team-st-mcd What do you think? I checked the linker files, this target has 256k flash and 32k RAM.

[0Grit]

@0xc0170
I've been meaning to attempt a similar pull request for some time, as we would like to use OS5 with this board.

What are the general hardware requirements for OS5?

Not that it would help those who already have the board but;
I have some influence in regards to this board's design, and might be able to push for a replacement platform or changes if they make sense.

[bcostm]

What do you think? I checked the linker files, this target has 256k flash and 32k RAM.

Well, on our side we have enabled OS5 only on targets with more than 64KB Flash AND more than 16KB RAM. So, OS5 should be ok for this board...

[0Grit]

@cmonr @0xc0170 What more is needed to make this happen?

@janjongboom Any interest in OS5 support for MOTE_L152RC

[0xc0170]

All was passing except c_string float printf because default_lib was small.

This should be fixed. First, why this target has uARM if it has 256k flash and 32k RAM. If we remove this from the target, will all test pass?

@dudmuck Can you test please?

@loverdeg-ep Thanks for the feedback, this should be fine to have it enabled for Mbed OS 5. needs to pass all tests.

[janjongboom]

@loverdeg-ep Yes, very happy with this, so we can run the new LoRaWAN stack on this target too. As @0xc0170 said this target should be able to run normal default lib. xDot has same RAM and ROM and runs everything fine, with >50% space left when running LoRaWAN stack and RTOS.

[cmonr]

Will start CI when able. Still waiting on feedback from @hasnainvirk and @ashok-rao, but I think @bcostm's and @janjongboom's feedback indicate this is good to go.

[cmonr]

Also still waiting on this.

All was passing except c_string float printf because default_lib was small.

This should be fixed. First, why this target has uARM if it has 256k flash and 32k RAM. If we remove this from the target, will all test pass?

@dudmuck Can you test please?

[0xc0170]

Closing this due to inactivity. Please reopen with an update. It would be nice addition based on the feedback here

[hasnainvirk]

This platform could potentially run LoRaWAN stack on it. @dudmuck could you please investigate why you are having troubles with default lib ? It should be alright in my opinion. You may add ARM and GCC_ARM in supported toolchains.

[maclobdell]

Reopening due to new activity/fixes/test results.
@dudmuck - please add new information that you have available.

[dudmuck]

test results attached
MOTE_IAR.LOG
MOTE_GCC.LOG
MOTE_ARM.LOG

[cmonr]

@dudmuck @maclobdell @hasnainvirk thanks for getting the ball moving again with this.

However, before we can progress, the second commit needs to be removed (Merge remote-tracking branch 'upstream/master' into mote_L152RC_add_os5). Our branching model doesn't support merges from master into PRs. Instead, you need to do a rebase, using upstream/master as the base branch.

[0xc0170]

6 commits but changes are only in targets.json file (few lines), squash to one commit (also merge commit will disappear).

I reviewed the tests, nvstore failed. Is this related to this fix ? #7127 . If not , please provide details for the failure

[0xc0170]

This should actually resolve #4394

[cmonr]

@dudmuck Any progress?

[cmonr]

Closing due to no response to comments for nearly three weeks.
Feel free to reopen once an update is available.