Merge feature-hal-spec-sleep to master #6994

bulislaw · 2018-05-23T20:20:29Z

Description

New HAL Sleep API.

Pull request type

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

Sleep - within 10us Deepsleep - within 10ms Note about mbed boards with interface, moved to lpc176x, as they are target related, should be documented in the target documentation. The tests will come as separate PR, to conform to this updates to sleep API.

These boards will be re-enabled when sleep driver for them is ready. Note: This operation is done by removing "SLEEP" feature from target's "device_has" list (in targets.json config file). For NRF52_DK removing of "SLEEP" feature causes some timing issues which have influence on tests. In order to successfully disable this board we need to disable also related features like "USTICKER", "LOWPOERTIMER" and slightly modify ticker tests, so they will not be executed if usticker support is not available (by default all targets support us ticker).

…abled in deep-sleep mode. This is done because of possible limitations of lowpower ticker freq/width.

Compile this file only when DEVICE_LOWPOWERTIMER is defined.

These platforms are not consistent with the new sleep standards.

…s only when SLEEP support is enabled.

…isabled.

…bled (NRF52_DK).

I implemented the SLEEP feature for Rnesas mbed boards. The mainly changing is here. - hal_sleep() To satisfy the mbed specificationm, I implemented this function newly by using "sleep" that is one of low power mode that is incorporated in our hardware(RZ_A1). In the "sleep", peripheral and memory state are maintained, and the peripherals continue to work and can generate interrupts. - hal_deepsleep() To satisfy the mbed specificationm, I implemented this function newly by combined using "sleep" and "module standby" that is the low power mode that is incorporated in our hardware(RZ_A1). The "module standby" is peripheral module's powerdown. Also in case of our "module standby", it need to read register as dummy when access to each register.

Signed-off-by: Mahesh Mahadevan <[email protected]>

bulislaw · 2018-05-23T20:21:05Z

@cmonr First one is ready for testing.

cmonr · 2018-05-23T20:23:05Z

@bulislaw Cool. Will take a quick look at it whilst Travis CI does it thing.

cmonr · 2018-05-23T20:23:36Z

@bulislaw Is this PR just the feature branch coming into master, or are there other PRs in her as well?

bulislaw · 2018-05-23T20:32:28Z

It's feature branch, but it has some of the partners code that wasn't validated by the CI.

cmonr · 2018-05-23T20:33:43Z

Alrgiht. I saw the auto-merges close those PRs. Was hoping that this method wouldn't do that, but oh well.

From the PoV of master, the effect is the same. The changes are still gated by CI.

FYI @ARMmbed/mbed-os-maintainers

kjbracey

Looks good - particularly glad to see specification of "deep sleep".

I would like to draw to your attention one significant issue with the current deep sleep specification - the console. If using a "proper" buffered console - like UARTSerial, deep sleep is rendered non-functional, as we must lock deep sleep to get serial IRQs. This is quite unfortunate as we're encouraging and setting apps to use buffered serial (it's an absolute requirement for some), thus reducing the number of apps that can deep sleep.

The deep sleep + console combo only "works" because the legacy serial console requires you to spin to transfer data, and if you're not spinning on RX you drop input. So that terrible behaviour isn't affected by sleep at all.

I think we may need some sort of mechanism to flag "extra devices that work in deep sleep on this platform".

Issue suggesting that here: #5943

We may also want a mechanism to request "output only" on UARTSerial, so we don't claim the RX interrupt.

We are currently in something of a knot with GreenTea - we're wanting to make it use buffered serial to solve various comms problems, but just setting that on locks out deep sleep tests...

jeromecoutant

It seems that there are some dependency with #6995
I think it should be reasonable to merge first this one, and then the ticker branch (and not in parallel)

0xc0170 · 2018-05-24T08:41:38Z

I would like to draw to your attention one significant issue with the current deep sleep specification - the console. If using a "proper" buffered console - like UARTSerial, deep sleep is rendered non-functional, as we must lock deep sleep to get serial IRQs. This is quite unfortunate as we're encouraging and setting apps to use buffered serial (it's an absolute requirement for some), thus reducing the number of apps that can deep sleep.

I think we may need some sort of mechanism to flag "extra devices that work in deep sleep on this platform".

Shall this be captured in Serial specs work to capture it there ?

The flag - this comes from time to time, targets hooks to overwrite the default locking or similar approach to have this "sort of mechanism to flag"

kjbracey · 2018-05-24T09:01:23Z

Shall this be captured in Serial specs work to capture it there ?

I think we do need to tackle it from two ends, basically independently

a general HAL mechanism for letting targets declare that extra devices beyond the minimum do work in deep sleep.
something in UARTSerial+retargetting to allow write-only opening to avoid locking deep sleep for unwanted RX on platforms without the above extension.

So I think that goes to "general HAL" and "serial/stream C++/retarget layers", not necessarily anything for the "serial HAL" specifically. But perhaps a general mechanism could be done as part of the serial HAL package.

Still a bit stuck with greentea, but the latter mechanism could maybe be runtime switchable - we effectively close serial input while doing the deep sleep tests, then reopen.

0xc0170 · 2018-05-24T13:35:17Z

/morph build

0xc0170 · 2018-05-24T13:40:22Z

Will retrigger the build once emac lands (might cause some conflicts), the current build aborted. This shoud happen shortly 🤞

mbed-ci · 2018-05-24T14:53:58Z

Build : FAILURE

Build number : 2138
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/6994/

cmonr · 2018-05-24T15:16:22Z

Marking as DNM since this will be coming in a larger PR.

cmonr · 2018-05-24T20:36:50Z

Removed labeling to make 5.9 tags easier to sort through. Will be brought in with #7009.

0xc0170 and others added 18 commits May 15, 2018 13:40

Add Sleep test specification to sleep_api.h.

fb7c03a

Add test and test header to Sleep HAL API.

6f50e8e

tests-mbed_hal-sleep: add lp/us ticker overflow handling.

fd84176

tests-mbed_hal-sleep: use lp ticker data while testing deepsleep.

f6e32bb

tests-mbed_hal-sleep: remove unused variable.

126d98c

tests-mbed_hal-sleep: fix formatting.

9000866

tests-mbed_hal-sleep: decrease delay when testing if us ticker is dis…

dc389db

…abled in deep-sleep mode. This is done because of possible limitations of lowpower ticker freq/width.

ARM_CM3DS_MPS2: add protection to lp_ticker.c.

dbf5c8b

Compile this file only when DEVICE_LOWPOWERTIMER is defined.

Disable SLEEP support for MAX32630FTHR and KW41Z platforms.

f988019

These platforms are not consistent with the new sleep standards.

tests-mbedmicro-rtos-mbed-systimer: execute sleep/deepsleep test case…

6ff49b6

…s only when SLEEP support is enabled.

Disable tests-mbed_drivers-timerevent test if USTICKER support is d…

432b139

…isabled.

Disable tests-mbed_drivers-timeout test if USTICKER support is disa…

81adf90

…bled (NRF52_DK).

Remove sleep from rf52 due to failig tests

947cbe8

KW41Z: Enable SLEEP feature

371b34a

Signed-off-by: Mahesh Mahadevan <[email protected]>

Adding Sleep and usticker to MTB_ODIN_W2

3d70964

cmonr added the release-version: 5.9.0-rc1 label May 23, 2018

cmonr requested review from 0xc0170, c1728p9, a user and ithinuel and removed request for a user May 24, 2018 02:17

cmonr added the needs: review label May 24, 2018

kjbracey approved these changes May 24, 2018

View reviewed changes

jeromecoutant reviewed May 24, 2018

View reviewed changes

0xc0170 approved these changes May 24, 2018

View reviewed changes

0xc0170 added needs: CI and removed needs: review labels May 24, 2018

c1728p9 approved these changes May 24, 2018

View reviewed changes

cmonr added the do not merge label May 24, 2018

bulislaw mentioned this pull request May 24, 2018

Bring in improved HAL APIs to master #7009

Merged

cmonr removed needs: CI release-version: 5.9.0-rc1 labels May 24, 2018

ithinuel approved these changes May 25, 2018

View reviewed changes

0xc0170 closed this May 28, 2018

0xc0170 deleted the feature-hal-spec-sleep branch June 21, 2018 13:39

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Merge feature-hal-spec-sleep to master #6994

Merge feature-hal-spec-sleep to master #6994

Uh oh!

bulislaw commented May 23, 2018

Uh oh!

bulislaw commented May 23, 2018

Uh oh!

cmonr commented May 23, 2018

Uh oh!

cmonr commented May 23, 2018

Uh oh!

bulislaw commented May 23, 2018

Uh oh!

cmonr commented May 23, 2018

Uh oh!

kjbracey left a comment

Uh oh!

jeromecoutant left a comment

Uh oh!

0xc0170 commented May 24, 2018

Uh oh!

kjbracey commented May 24, 2018

Uh oh!

0xc0170 commented May 24, 2018

Uh oh!

0xc0170 commented May 24, 2018

Uh oh!

mbed-ci commented May 24, 2018

Uh oh!

cmonr commented May 24, 2018

Uh oh!

cmonr commented May 24, 2018

Uh oh!

Uh oh!

Merge feature-hal-spec-sleep to master #6994

Merge feature-hal-spec-sleep to master #6994

Uh oh!

Conversation

bulislaw commented May 23, 2018

Description

Pull request type

Uh oh!

bulislaw commented May 23, 2018

Uh oh!

cmonr commented May 23, 2018

Uh oh!

cmonr commented May 23, 2018

Uh oh!

bulislaw commented May 23, 2018

Uh oh!

cmonr commented May 23, 2018

Uh oh!

kjbracey left a comment

Choose a reason for hiding this comment

Uh oh!

jeromecoutant left a comment

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented May 24, 2018

Uh oh!

kjbracey commented May 24, 2018

Uh oh!

0xc0170 commented May 24, 2018

Uh oh!

0xc0170 commented May 24, 2018

Uh oh!

mbed-ci commented May 24, 2018

Build : FAILURE

Uh oh!

cmonr commented May 24, 2018

Uh oh!

cmonr commented May 24, 2018

Uh oh!

Uh oh!