ARMmbed / nrf51-sdk

Module to contain files provided by the nordic nRF51 SDK
Other
13 stars 17 forks source link

Update files to Nordic sdk 9.0.0 #8

Closed LiyouZhou closed 8 years ago

LiyouZhou commented 8 years ago

Tests:

clitest - all expected passes pass, all expected fails fail. build - all ble-examples build successfully pstorage - tested against https://github.com/andresag01/ble-examples-1/tree/eddystone_pstorage and works dfu - tested to work

Highlight from sdk 9 release notes:

Changes

@ARMmbed/ble-owners Please review.

rgrover commented 8 years ago

@LiyouZhou please confirm that the bootloader is still able to build and work correctly with this new SDK. I have a feeling you'll find that the bootloader doesn't work any more unless you fix the constants in hci_mem_pool_internal.h (see comments above). Jonny is currently attempting to build the bootloader for BBC. Please connect with him to align with his efforts.

Please also ensure that the memory consumption doesn't go up with this new SDK. Building any of our examples should show the heap sizes at the end. This is important.

rgrover commented 8 years ago

I would like you to discuss your approach with others members of the team. Please document the affects of switching to the new SDK; we need to ensure that we don't affect existing functionality--such as use of pstorage, bootloader, etc. Could you also please highlight the important parts of the release notes for SDKv9?

rgrover commented 8 years ago

@LiyouZhou please also be mindful that we need to hand-apply the changes from SDKv9 to the ble-nrf51822 repository on mbed.org. That repository will no longer automatically sync with the changes to the SDK.

LiyouZhou commented 8 years ago

@rgrover pstorage cmd queue size reduced to 2 in new commit. Mem usage increased by 24bytes as a result of updating the sdk. Bootloader continue to work.

rgrover commented 8 years ago

@LiyouZhou you claim: Mem usage increased by 24bytes as a result of updating the sdk. Can you account for this?

LiyouZhou commented 8 years ago

@rgrover by looking at the obj-dump the major refactorying of pstorage.c caused this increase. There is a lot of change, hence it is difficult to tell exactly what cause memory increase.

rgrover commented 8 years ago

@LiyouZhou ok. Thanks for this dive. 24bytes isn't too much to worry about.

LiyouZhou commented 8 years ago

@rgrover hold off merging, need to talk to jonny and make a small change.

ghost commented 8 years ago

sdk10 has a fix for "App_pwm occasionally gives inverted signal"

LiyouZhou commented 8 years ago

@jrobeson I'm currently in the process of upgrading to v10. Should be up shortly.