gpio_api.h: Clarify desired behaviour for NC #10489

kjbracey · 2019-04-26T10:01:15Z

Description

Documentation clarified, and platforms that don't conform brought into line. Simple test added.

Clarification needed as shown by #7862, #10191, #9469 and #10482.

Pull request type

[ ] Fix
[ ] Refactor
[X] Target update
[ ] Functionality change
[ ] Docs update
[ ] Test update
[ ] Breaking change

ciarmcom · 2019-04-26T11:00:30Z

@kjbracey-arm, thank you for your changes.
@ARMmbed/mbed-os-hal @ARMmbed/mbed-os-maintainers please review.

0xc0170 · 2019-04-29T07:41:27Z

Updated description looks good.

How many targets follow this? As stated, test would be good and have target code updated to this. Will this be few steps update?

This would be good to target for 5.13 (considering all targets will require update).

kjbracey · 2019-04-29T07:47:23Z

Most, but not all, implement it. As SPI now relies on it (its original constructors now construct a DigitalOut(NC) for an unused software SSEL), any that don't are going to need patching.

So I would say this and target updates and test should go for patch release.

I can look at tests and target updates later this week.

stevew817 · 2019-04-29T12:03:36Z

IMO this definitely needs to go to patch release (together with #10482) since SPI on the current release is broken for Silicon Labs targets.

See #9469 for more details as to why it broke.

kjbracey · 2019-04-29T15:18:08Z

Created a test, and waiting for its results, but by inspection, ARM SSG/FM, Atmel, RTL8195A, EFM32, TT, Wiznet and Ublox HI2110 do not handle NC as specified here.

The rest do.

(It's worth noting that gpio_irq_api does not incorporate this behaviour, and does not have a gpio_irq_is_connected call, so we are inconsistent here. I'd be inclined to bring that in line at some point.)

I've got commits for a test + changing targets (https://github.com/kjbracey-arm/mbed-os/commits/gpio_api_nc_test). Can incorporate those into this PR, or as one or more follow-ups.

0xc0170 · 2019-04-29T15:44:46Z

I've got commits for a test + changing targets (https://github.com/kjbracey-arm/mbed-os/commits/gpio_api_nc_test). Can incorporate those into this PR, or as one or more follow-ups.

👍 Lets add it here.

It would probably be worth adding tests for the ability to initialise NC pins and check `is_connected`. Some platforms are assert failing the init, and can't be 100% sure `is_connected` is working on those platforms either.

kjbracey · 2019-05-02T09:25:11Z

Added the extra commits for testing and correcting NC handling on platforms (other than EFM32 covered by #10482).

0xc0170

just SPDX identifier, LGTM otherwise

TESTS/mbed_hal/gpio/main.cpp

Check two items of defined behaviour - that you can initialise a gpio_t with NC, and you can detect that state with gpio_is_connected().

kjbracey · 2019-05-03T08:46:16Z

Should get reviews from each of the vendors though - not least for confirmation I haven't broken anything. (Not all of those are in CI, and I don't have them).

0xc0170 · 2019-05-03T09:10:00Z

@ThunderSoft123 @ARMmbed/team-wiznet @ARMmbed/team-realtek @jamesbeyond Please review gpio files for your targets changed here

0xc0170 · 2019-05-13T05:54:40Z

@ThunderSoft123 @ARMmbed/team-wiznet @ARMmbed/team-realtek @jamesbeyond Please review gpio files for your targets changed here

We would like to proceed with this pull request, please review

cc @maclobdell @MarceloSalazar

0xc0170 · 2019-05-13T06:02:24Z

I started CI job meanwhile

mbed-ci · 2019-05-13T11:28:10Z

Test run: SUCCESS

Summary: 11 of 11 test jobs passed
Build number : 1
Build artifacts

adbridge · 2019-05-17T10:02:08Z

@maclobdell @MarceloSalazar could you help get this reviewed by partners please ?

ThunderSoft123 · 2019-05-21T13:25:59Z

@0xc0170 No problem with revision.

kjbracey force-pushed the gpio_api_nc branch from d12e555 to c2db8a3 Compare April 26, 2019 10:32

ciarmcom requested review from a team April 26, 2019 11:00

ciarmcom added the needs: review label Apr 26, 2019

c1728p9 approved these changes Apr 26, 2019

View reviewed changes

0xc0170 added the release-version: 5.13.0-rc1 label Apr 29, 2019

0xc0170 mentioned this pull request Apr 29, 2019

Allow NC pins to be instantiated on Silicon Labs platforms #10482

Merged

stevew817 approved these changes May 1, 2019

View reviewed changes

0xc0170 added needs: work and removed needs: review labels May 2, 2019

kjbracey added 7 commits May 2, 2019 12:16

gpio_api.h: Clarify desired behaviour for NC

ad1acf9

It would probably be worth adding tests for the ability to initialise NC pins and check `is_connected`. Some platforms are assert failing the init, and can't be 100% sure `is_connected` is working on those platforms either.

ARM SSG/FM: Cope correctly with NC GPIO

fd68691

Atmel: Cope correctly with NC GPIO

bd69ede

RTL8195A: Cope correctly with NC GPIO

0fb64ee

TT: Cope correctly with NC GPIO

e6e3614

Wiznet: Cope correctly with NC GPIO

1eff7ec

HI2110: Cope correctly with NC GPIO

e502eb5

kjbracey force-pushed the gpio_api_nc branch from c2db8a3 to 2d07210 Compare May 2, 2019 09:23

0xc0170 requested changes May 2, 2019

View reviewed changes

TESTS/mbed_hal/gpio/main.cpp Show resolved Hide resolved

Add GPIO NC test

66f446b

Check two items of defined behaviour - that you can initialise a gpio_t with NC, and you can detect that state with gpio_is_connected().

kjbracey force-pushed the gpio_api_nc branch from 2d07210 to 66f446b Compare May 2, 2019 13:12

0xc0170 approved these changes May 2, 2019

View reviewed changes

0xc0170 removed the needs: work label May 2, 2019

0xc0170 added the needs: CI label May 2, 2019

0xc0170 added needs: review and removed needs: CI labels May 3, 2019

bjnhur approved these changes May 20, 2019

View reviewed changes

0xc0170 added ready for merge and removed needs: review labels May 21, 2019

kjbracey mentioned this pull request May 21, 2019

Fix bug Operator new[] out of memory #10616

Merged

0xc0170 merged commit f859289 into ARMmbed:master May 21, 2019

kjbracey deleted the gpio_api_nc branch May 21, 2019 13:37

0xc0170 removed the ready for merge label May 21, 2019

gpio_api.h: Clarify desired behaviour for NC #10489

gpio_api.h: Clarify desired behaviour for NC #10489

Uh oh!

Conversation

kjbracey commented Apr 26, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Pull request type

Uh oh!

ciarmcom commented Apr 26, 2019

Uh oh!

0xc0170 commented Apr 29, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

kjbracey commented Apr 29, 2019

Uh oh!

stevew817 commented Apr 29, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

kjbracey commented Apr 29, 2019

Uh oh!

0xc0170 commented Apr 29, 2019

Uh oh!

kjbracey commented May 2, 2019

Uh oh!

0xc0170 left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

kjbracey commented May 3, 2019

Uh oh!

0xc0170 commented May 3, 2019

Uh oh!

0xc0170 commented May 13, 2019

Uh oh!

0xc0170 commented May 13, 2019

Uh oh!

mbed-ci commented May 13, 2019

Test run: SUCCESS

Uh oh!

adbridge commented May 17, 2019

Uh oh!

ThunderSoft123 commented May 21, 2019

Uh oh!

Uh oh!

kjbracey commented Apr 26, 2019 •

edited

Loading

0xc0170 commented Apr 29, 2019 •

edited

Loading

stevew817 commented Apr 29, 2019 •

edited

Loading