add OSHChip as an mbed target #5892

drewcassidy · 2018-01-20T00:16:23Z

Description

Add a target for the OSHChip, a miniature nRF51822 board, along with pin name definitions

Todo

get fix for ticker tests locking up (see ticker unit test contains a race condition, never runs one of the callback events #5955 and Ticker tests fix #5971)
get fix for ticker tests when using the RC clock
verify all tests on target hardware

Deploy notes

The documentation is confusing regarding the use of config parameters in target definitions. I had to read through the python code to discover the necessary key was overrides and not target_overrides. The documentation should probably be updated to include more information on this. This is being recified in ARMmbed/mbed-os-5-docs#398

mbed-ci · 2018-01-20T00:16:25Z

Automatic CI verification build not done, please verify manually.

0xc0170 · 2018-01-22T07:47:52Z

The documentation is confusing regarding the use of config parameters in target definitions. I had to read through the python code to discover the necessary key was overrides and not target_overrides. The documentation should probably be updated to include more information on this

Is this in the docs portal? Can you file the bug in https://github.com/ARMmbed/Handbook ?

confirmed to work on target hardware

How did you test ? Does this have on board debugger that could be supported by greentea to run our tests?

drewcassidy · 2018-01-22T16:11:20Z

I already made a fork of the handbook with the intent of doing a pull request but just caught a cold and havnt gotten around to it yet

I tested with just blinky, it should be identical to any other nrf51822 board besides the LF clock source and pin names

drewcassidy · 2018-01-22T18:39:19Z

I attempted to test the board using Greentea, but it wont work because it relies on the MSD programming. Theres no DAPLink version for the lpc11u35 in my programmer that directly supports the nrf51822, so resetting doesnt work, and i'm having trouble getting greentea working with pyOCD (see ARMmbed/greentea#259)

drewcassidy · 2018-01-23T05:48:08Z

After some firmware changes and connecting the UART to the programmer (which I didnt realize greentea needed) I ran all the tests

all tests pass except mbed_drivers-ticker: 2x callbacks, probably due to the decreased clock resolution when using the RC timer

0xc0170 · 2018-01-23T08:09:13Z

all tests pass except mbed_drivers-ticker: 2x callbacks, probably due to the decreased clock resolution when using the RC timer

Can you provide logs (-v verbose output) to understand the failure better? The test should be fixed or target :)

drewcassidy · 2018-01-23T08:55:57Z

the test times out, and is reported as an error

[1516696965.48][CONN][RXD] >>> Running case #11: 'Test timers: 2x callbacks'...
[1516696965.54][CONN][INF] found KV pair in stream: {{__testcase_start;Test timers: 2x callbacks}}, queued...
[1516696965.57][CONN][INF] found KV pair in stream: {{timing_drift_check_start;0}}, queued...
[1516696965.58][SERI][TXD] {{base_time;0}}
[1516697190.20][HTST][INF] test suite run finished after 240.64 sec...

0xc0170 · 2018-01-23T13:55:15Z

How to reproduce? using any nrf5x target? Tests should be green

drewcassidy · 2018-01-23T15:16:14Z

Any nRF51 target using the RC low frequency clock should probably reproduce it yeah. I’ll try testing it outside Greentea after class today

0xc0170 · 2018-01-24T10:30:51Z

Any nRF51 target using the RC low frequency clock should probably reproduce it yeah. I’ll try testing it outside Greentea after class today

Let us know your findings. All tests should be green

drewcassidy · 2018-01-25T05:15:22Z

I connected a 32KHz watch crystal to it in a breadboard, making it more or less identical to the configuration of an NRF51-DK, and the test fails with the same error. If anyone could run mbed test -n mbed-os-tests-mbed_drivers-ticker with another nRF51-based board and report back that would be helpful, but I'm starting to suspect this is an unrelated issue.

0xc0170 · 2018-01-25T08:32:49Z

I connected a 32KHz watch crystal to it in a breadboard, making it more or less identical to the configuration of an NRF51-DK, and the test fails with the same error. If anyone could run mbed test -n mbed-os-tests-mbed_drivers-ticker with another nRF51-based board and report back that would be helpful, but I'm starting to suspect this is an unrelated issue.

@mprse @maciejbocianski Can you please look at this test issue?

mprse · 2018-01-25T09:01:00Z

Hi,
I noticed that OSHChip:master branch is not up to date.

Recently synchronisation fix was provided to ticker HAL layer. Please check the following PR: #5889. This might be related to your issue (see tests results).

Please rebase OSHChip:master on the current master and try again.

drewcassidy · 2018-01-25T18:11:18Z

After rebasing, the test doesnt compile at all. It looks like aaa15bc changed the test file but I'm not sure how its getting into RTOS stuff

$ mbed test -c -v --compile -n mbed-os-tests-mbed_drivers-ticker
>>SNIP<<
        Compile [ 66.6%]: mbed_rtx_idle.cpp
        [DEBUG] Compile: arm-none-eabi-g++ -std=gnu++98 -fno-rtti -Wvla -c -Wall -Wextra -Wno-unused-parameter -Wno-missing-field-initializers -fmessage-length=0 -fno-exceptions -fno-builtin -ffunction-sections -fdata-sections -funsigned-char -MMD -fno-delete-null-pointer-checks -fomit-frame-pointer -Os -g1 -mcpu=cortex-m0 -mthumb -D__CORTEX_M0 -DNRF51 -D__MBED__=1 -DTARGET_LIKE_MBED -DTARGET_NRF51822 -DDEVICE_PORTINOUT=1 -D__MBED_CMSIS_RTOS_CM -DMBED_BUILD_TIMESTAMP=1516902332.91 -DTOOLCHAIN_object -D__CMSIS_RTOS -DTARGET_MCU_NRF51822_UNIFIED -DTOOLCHAIN_GCC -DTARGET_CORTEX_M -DTARGET_SDK11 -DARM_MATH_CM0 -DTARGET_UVISOR_UNSUPPORTED -DTARGET_NRF5 -DFEATURE_BLE=1 -DTARGET_M0 -DTARGET_MCU_NRF51 -DCMSIS_VECTAB_VIRTUAL -DTARGET_MCU_NRF51_UNIFIED -DMBED_TICKLESS -DDEVICE_INTERRUPTIN=1 -DTARGET_CORTEX -DDEVICE_I2C=1 -DDEVICE_PORTOUT=1 -DTARGET_RELEASE -DTARGET_NORDIC -DDEVICE_SERIAL=1 -DBLE_STACK_SUPPORT_REQD -DTARGET_MCU_NRF51_32K_UNIFIED -DDEVICE_PORTIN=1 -DDEVICE_SLEEP=1 -DTOOLCHAIN_GCC_ARM -DTARGET_MCU_NRF51822 -DTARGET_MCU_NORDIC_32K -DSOFTDEVICE_PRESENT -DNO_SYSTICK -DDEVICE_SPI=1 -DCMSIS_VECTAB_VIRTUAL_HEADER_FILE="cmsis_nvic.h" -DDEVICE_SPISLAVE=1 -DDEVICE_ANALOGIN=1 -DDEVICE_PWMOUT=1 -DS130 -DTARGET_OSHCHIP -DTARGET_LIKE_CORTEX_M0 -DTARGET_MCU_NRF51_32K @/Users/drewcassidy/Projects/OSHChip-mbed-tests/BUILD/tests/OSHChip/GCC_ARM/.includes_bb1986b19ee2954951aa1bc2a56ce774.txt -include /Users/drewcassidy/Projects/OSHChip-mbed-tests/BUILD/tests/OSHChip/GCC_ARM/mbed_config.h -MD -MF BUILD/tests/OSHChip/GCC_ARM/mbed-os/rtos/TARGET_CORTEX/mbed_rtx_idle.d -o BUILD/tests/OSHChip/GCC_ARM/mbed-os/rtos/TARGET_CORTEX/mbed_rtx_idle.o /Users/drewcassidy/Projects/OSHChip-mbed-tests/mbed-os/rtos/TARGET_CORTEX/mbed_rtx_idle.cpp
        [Error] mbed_rtx_idle.cpp@50,29: 'get_lp_ticker_data' was not declared in this scope
        [DEBUG] Return: 1
        [DEBUG] Output: /Users/drewcassidy/Projects/OSHChip-mbed-tests/mbed-os/rtos/TARGET_CORTEX/mbed_rtx_idle.cpp: In constructor 'RtosTimer::RtosTimer()':
        [DEBUG] Output: /Users/drewcassidy/Projects/OSHChip-mbed-tests/mbed-os/rtos/TARGET_CORTEX/mbed_rtx_idle.cpp:50:29: error: 'get_lp_ticker_data' was not declared in this scope
        [DEBUG] Output:      RtosTimer(): TimerEvent(get_lp_ticker_data()), _start_time(0), _tick(0) {
        [DEBUG] Output:                              ^~~~~~~~~~~~~~~~~~
        [DEBUG] Output: /Users/drewcassidy/Projects/OSHChip-mbed-tests/mbed-os/rtos/TARGET_CORTEX/mbed_rtx_idle.cpp:50:29: note: suggested alternative: 'get_us_ticker_data'
        [DEBUG] Output:      RtosTimer(): TimerEvent(get_lp_ticker_data()), _start_time(0), _tick(0) {
        [DEBUG] Output:                              ^~~~~~~~~~~~~~~~~~
        [DEBUG] Output:                              get_us_ticker_data

[mbed] ERROR: "/usr/bin/python" returned error code 1.
[mbed] ERROR: Command "/usr/bin/python -u /Users/drewcassidy/Projects/OSHChip-mbed-tests/mbed-os/tools/test.py -t GCC_ARM -m OSHChip --source /Users/drewcassidy/Projects/OSHChip-mbed-tests --build /Users/drewcassidy/Projects/OSHChip-mbed-tests/BUILD/tests/OSHChip/GCC_ARM --test-spec /Users/drewcassidy/Projects/OSHChip-mbed-tests/BUILD/tests/OSHChip/GCC_ARM/test_spec.json -n mbed-os-tests-mbed_drivers-ticker -v" in "/Users/drewcassidy/Projects/OSHChip-mbed-tests"
---

mprse · 2018-01-25T18:21:06Z

Looks like LOWPOWERTIMER is not available. If this feature is available for your device, then try to add LOWPOWERTIMER to target's device_has array in targets.json file. Otherwise try to disable tickless mode.

drewcassidy · 2018-01-25T18:50:56Z

Even after rebasing the issues with the 2x callbacks test remains. ~~This happens regardless of the LF clock source~~ The tests now pass when using a crystal, but time out when using the internal RC

I'm unclear on what this test is actually supposed to do?

mprse · 2018-01-26T06:54:33Z

What is the frequency of the internal clock? It looks like you still need to adjust clock frequency to your board. NRF51_DK clock frequency is defined here:

mbed-os/targets/TARGET_NORDIC/TARGET_NRF5/TARGET_SDK11/drivers_nrf/hal/nrf_rtc.h

Line 71 in 26d0c6d

#define RTC_INPUT_FREQ 32768 /**< Input frequency of the RTC instance. */

drewcassidy · 2018-01-26T06:56:42Z

What is the frequency of the internal clock?

The internal RC clock is the same fequency, 32.768 kHz

I'm unclear on what that test is supposed to be doing, but after looking at it outside greentea it looks like its suffering to a race condition? event 1 starts the ticker for event 2, but event 1 gets called again before event 2 can, resetting its clock, so event 2 never happens, if that makes sense. I dont know what the expected behavior is, but I think this is at least part of the problem

This behavior happened predictably when testing at a lower frequency, but when running with a 1ms period the behavior was erratic, even when using the crystal. I'm assuming its just trying to run too fast for the cortex-m0

the test program I used is here

drewcassidy · 2018-01-27T18:15:20Z

Once I get hold of an nRF51-DK I'll try to test this further and confirm its not specific to the OSHChip hardware. I can no longer get the tests to pass with the crystal, so I think its just very erratic

… should probably move to the unified target at some point but is outside the scope of this PR

bulislaw · 2018-02-06T09:32:46Z

@drewcassidy can you confirm this board has a working DAPLink?

drewcassidy · 2018-02-07T18:42:48Z

@drewcassidy can you confirm this board has a working DAPLink?

It comes with a DAPLink programmer to connect to it, but the board itself is too small to have one built in

cmonr · 2018-02-12T22:17:03Z

@bulislaw What is the reason for the DNM tag?

bulislaw · 2018-02-13T15:46:50Z

My question was answered. Removing the DNM tag.

cmonr · 2018-02-13T16:28:00Z

/morph build

drewcassidy · 2018-02-13T20:15:14Z

This can be merged, just note that the ticker tests still fail on the nRF51 when using the RC Low Frequency clock. I still have some debugging to do to figure out why, but its not entirely specific to this target.

0xc0170 · 2018-02-14T09:08:36Z

/morph build

mbed-ci · 2018-02-14T09:40:59Z

Build : SUCCESS

Build number : 1133
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/5892/

Triggering tests

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

mbed-ci · 2018-02-14T10:58:53Z

Test : SUCCESS

Build number : 941
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/5892/941

mbed-ci · 2018-02-14T11:22:06Z

Exporter Build : SUCCESS

Build number : 810
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/5892/

cmonr · 2018-02-14T21:15:04Z

targets/TARGET_NORDIC/TARGET_NRF5/TARGET_MCU_NRF51822_UNIFIED/TARGET_OSHCHIP/PinNames.h

+    │  SEL  D5 ╶┨ 6 └──┘11 ┠╴ D9  AREF
+    └─ CLK  D4 ╶┨ 7     10 ┠╴ A2
+           GND ╶┨ 8  ::  9 ┠╴ A3
+                ┗━━━━━━━━━━┛


That's some nice ascii art.

cmonr · 2018-02-14T21:19:17Z

Merging this in since the issue that causes the ticker problem is non-standard.
Opening up an issue to keep track of the problem (reproduced with a NRF51-DK).

0xc0170 added needs: review scope: new-target labels Jan 22, 2018

0xc0170 added needs: work and removed needs: review labels Jan 24, 2018

drewcassidy force-pushed the master branch 2 times, most recently from 81af684 to f4a5f00 Compare January 25, 2018 16:23

drewcassidy added 7 commits January 25, 2018 08:24

Add target for OSHChip

fe6b6af

add board header files

a6c6956

Clean up board definition to conform to the target spec

8cb421d

Add pinout diagram

e06d1db

fix typo 😖

5547a09

remove redundent device_has

8bf51ac

move OSHCHIP to be with other nRF51 boards

8ba0943

drewcassidy force-pushed the master branch from f4a5f00 to 8ba0943 Compare January 25, 2018 16:25

add LOWPOWERTIMER and I2C_ASYNC to nrf51822 unified

bc3d938

drewcassidy mentioned this pull request Jan 27, 2018

ticker unit test contains a race condition, never runs one of the callback events #5955

Closed

maciejbocianski mentioned this pull request Jan 30, 2018

Ticker tests fix #5971

Merged

bulislaw added the do not merge label Feb 5, 2018

drewcassidy added 2 commits February 5, 2018 21:15

remove redundent 'targets.' in overrides

b1199b0

move device_has to OSHCHIP specifically and define device name. These…

1fe0933

… should probably move to the unified target at some point but is outside the scope of this PR

bulislaw removed the do not merge label Feb 13, 2018

cmonr added needs: CI and removed needs: work labels Feb 13, 2018

cmonr added ready for merge and removed needs: CI labels Feb 14, 2018

cmonr reviewed Feb 14, 2018

View reviewed changes

cmonr merged commit ee52f90 into ARMmbed:master Feb 14, 2018

cmonr removed the ready for merge label Feb 14, 2018

cmonr mentioned this pull request Feb 14, 2018

Ticker test fails with NRF51 when LF clock is used #6098

Closed

cmonr added the release-version: 5.7.7 label Mar 5, 2018

add OSHChip as an mbed target #5892

add OSHChip as an mbed target #5892

Uh oh!

Conversation

drewcassidy commented Jan 20, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Todo

Deploy notes

Uh oh!

mbed-ci commented Jan 20, 2018

Uh oh!

0xc0170 commented Jan 22, 2018

Uh oh!

drewcassidy commented Jan 22, 2018

Uh oh!

drewcassidy commented Jan 22, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

drewcassidy commented Jan 23, 2018

Uh oh!

0xc0170 commented Jan 23, 2018

Uh oh!

drewcassidy commented Jan 23, 2018

Uh oh!

0xc0170 commented Jan 23, 2018

Uh oh!

drewcassidy commented Jan 23, 2018

Uh oh!

0xc0170 commented Jan 24, 2018

Uh oh!

drewcassidy commented Jan 25, 2018

Uh oh!

0xc0170 commented Jan 25, 2018

Uh oh!

mprse commented Jan 25, 2018

Uh oh!

drewcassidy commented Jan 25, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mprse commented Jan 25, 2018

Uh oh!

drewcassidy commented Jan 25, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mprse commented Jan 26, 2018

Uh oh!

drewcassidy commented Jan 26, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

drewcassidy commented Jan 27, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

bulislaw commented Feb 6, 2018

Uh oh!

drewcassidy commented Feb 7, 2018

Uh oh!

cmonr commented Feb 12, 2018

Uh oh!

bulislaw commented Feb 13, 2018

Uh oh!

cmonr commented Feb 13, 2018

Uh oh!

drewcassidy commented Feb 13, 2018 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

0xc0170 commented Feb 14, 2018

Uh oh!

mbed-ci commented Feb 14, 2018

Build : SUCCESS

Triggering tests

Uh oh!

mbed-ci commented Feb 14, 2018

Test : SUCCESS

Uh oh!

mbed-ci commented Feb 14, 2018

Exporter Build : SUCCESS

Uh oh!

cmonr Feb 14, 2018

Choose a reason for hiding this comment

drewcassidy commented Jan 20, 2018 •

edited

Loading

drewcassidy commented Jan 22, 2018 •

edited

Loading

drewcassidy commented Jan 25, 2018 •

edited

Loading

drewcassidy commented Jan 25, 2018 •

edited

Loading

drewcassidy commented Jan 26, 2018 •

edited

Loading

drewcassidy commented Jan 27, 2018 •

edited

Loading

drewcassidy commented Feb 13, 2018 •

edited

Loading