STM32 LPTICKER : Fix tickless and LPTICKER_DELAY_TICKS #8242

jeromecoutant · 2018-09-25T12:34:59Z

Description

This should fix #7858,
and is needed to pass new test case from #8129

Test result

With that patch, all tests are OK in default mode and in TICKLESS mode for targets using LPTICKER with LPTIM.
I will push a separate PR to enable TICKLESS for them.

Restriction

In tickless mode, for targets using LPTICKER with RTC, it seems there is still some issue
with the RTC write procedure.
I will push a new PR when solution is found.

Pull request type

[x] Fix
[ ] Refactor
[ ] Target update
[ ] Functionality change
[ ] Breaking change

jeromecoutant · 2018-09-25T12:52:06Z

Note that my TICKLESS tests was including #8223

jeromecoutant · 2018-09-25T15:44:00Z

0xc0170 added the needs: preceding PR

I don't think so, you can start CI

c1728p9 · 2018-09-25T23:00:28Z

targets/targets.json

+ },
+ "lpticker_delay_ticks": {
+ "help": "https://os.mbed.com/docs/v5.9/reference/low-power-ticker.html",
+ "value": 1,


Why is only one cycle needed?

Good idea for the link. Just maybe replace "v5.9" with "latest".

In some previous patch, I have removed this delay tick, but it seems that not all the tests can pass...
Then I set back value to 1 as the default value.

c1728p9 · 2018-09-25T23:00:51Z

targets/targets.json

 }
 },
+ "overrides": {"lpticker_delay_ticks": 4},


Why are 4 needed for the L series?

Default value 1 seems not enough for low speed targets.
4 is the experimental result to pass speed test.

Is there anything in the reference manual for these devices which state the number of cycles which must elapse before writing to the match register?

c1728p9 · 2018-09-25T23:06:53Z

targets/TARGET_STM/lp_ticker.c

@@ -263,7 +258,6 @@ uint32_t lp_ticker_read(void)

 void lp_ticker_set_interrupt(timestamp_t timestamp)
 {
- lp_ticker_disable_interrupt();


Do you have any details on how this fixes the test in #8129?

yes, I should add the details you have added in #8129

c1728p9 · 2018-09-25T23:12:35Z

targets/TARGET_STM/lp_ticker.c

@@ -194,21 +191,22 @@ void lp_ticker_set_interrupt(timestamp_t timestamp)
 LptimHandle.Instance = LPTIM1;
 irq_handler = (void (*)(void))lp_ticker_irq_handler;

+ __HAL_LPTIM_CLEAR_FLAG(&LptimHandle, LPTIM_FLAG_CMPOK);
+ __HAL_LPTIM_COMPARE_SET(&LptimHandle, timestamp);
 /* CMPOK is set by hardware to inform application that the APB bus write operation to the LPTIM_CMP register has been successfully completed */
 /* Any successive write before the CMPOK flag be set, will lead to unpredictable results */
 while (__HAL_LPTIM_GET_FLAG(&LptimHandle, LPTIM_FLAG_CMPOK) == RESET) {


Will this bring back the blocking behavior?

No, it is the same behavior as you have done in #8129

The change in #8129 caused blocking behavior as mentioned by @LMESTM.

@c1728p9 my mistake - Jérôme explained to me and I actually misunderstood CMPOK (programing ok) with CMPM (comparator match) - looks ok to me now

c1728p9 · 2018-09-25T23:23:30Z

@0xc0170 why is this marked as "needs: preceding PR"? PR #8223 is pending on this.

0xc0170 · 2018-09-26T14:24:24Z

Removing, I understood it otherway around

This fixes issue with mbed_hal/lp_ticker/lp_ticker_early_match_race_test

When both tickless and LPTICKER_DELAY_TICKS are enabled some ST devices randomly get stuck sleeping forever. This is because the wake up time passed to the rtc is ignored if the previous match is about to occur. This causes the device to get stuck in sleep. This patch prevents matches from getting dropped by the rtc by deactivating the rtc wake up timer before setting a new value. Events leading up to this failure for the RTC: -1st call to lp_ticker_set_interrupt -delay until ticker interrupt is about to fire -2nd call to lp_ticker_set_interrupt -interrupt for 1st call fires and match time for 2nd call is dropped -LowPowerTickerWrapper gets ticker interrupt but treats it as a spurious interrupt and drops it since it comes in too early -device enters sleep without a wakeup source and locks up

For both implementation, RTC and LPTIM, there is some delay in the set_interrupt function due to HW constraints. Value has been set to 4 for STM32L0, because SystemClock is slower than other families.

jeromecoutant · 2018-09-27T12:49:29Z

Rebase done.

@c1728p9 maybe you can approve and start CI ?

Thx

0xc0170 · 2018-10-01T11:35:06Z

/morph build

mbed-ci · 2018-10-01T12:41:04Z

Build : SUCCESS

Build number : 3203
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/8242/

Triggering tests

/morph test
/morph export-build
/morph mbed2-build

mbed-ci · 2018-10-01T16:56:51Z

Test : FAILURE

Build number : 3003
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/8242/3003

mbed-ci · 2018-10-02T06:47:30Z

Exporter Build : ABORTED

Build number : 2795
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8242/

0xc0170 · 2018-10-02T09:38:52Z

Test needs to be restarted (device was power cycled) and exporters as well, will do once we fix exporters CI job

0xc0170 · 2018-10-04T09:25:26Z

/morph test

mbed-ci · 2018-10-04T21:56:29Z

Test : SUCCESS

Build number : 3038
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/8242/3038

0xc0170 · 2018-10-05T15:01:47Z

/morph export-build

mbed-ci · 2018-10-05T19:13:43Z

Exporter Build : SUCCESS

Build number : 2841
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/8242/

jeromecoutant mentioned this pull request Sep 25, 2018

Add LPTICKER tests #8129

Closed

0xc0170 added the needs: review label Sep 25, 2018

jeromecoutant mentioned this pull request Sep 25, 2018

Recent LP Ticker Changes Break Tickless #7858

Closed

0xc0170 added the needs: preceding PR label Sep 25, 2018

c1728p9 self-requested a review September 25, 2018 22:59

c1728p9 reviewed Sep 25, 2018

View reviewed changes

0xc0170 removed the needs: preceding PR label Sep 26, 2018

jeromecoutant added 3 commits September 27, 2018 14:30

STM32 LPTICKER with LPTIM : Fix tickless and lp wrapper

b1d23e5

This fixes issue with mbed_hal/lp_ticker/lp_ticker_early_match_race_test

STM32 LPTICKER : set default LPTICKER_DELAY_TICKS to 1

af4841d

For both implementation, RTC and LPTIM, there is some delay in the set_interrupt function due to HW constraints. Value has been set to 4 for STM32L0, because SystemClock is slower than other families.

jeromecoutant force-pushed the PR_STM32_LPTICKER branch from 4c94e30 to af4841d Compare September 27, 2018 12:48

c1728p9 approved these changes Sep 27, 2018

View reviewed changes

0xc0170 approved these changes Oct 1, 2018

View reviewed changes

0xc0170 added needs: CI and removed needs: review labels Oct 1, 2018

jeromecoutant mentioned this pull request Oct 2, 2018

All STM should have tickless mode enabled #7153

Closed

jeromecoutant mentioned this pull request Oct 4, 2018

unstable test - tests-mbedmicro-rtos-mbed-queue on NUCLEO_F746ZG #8320

Closed

0xc0170 added the release-version: 5.10.2 label Oct 5, 2018

0xc0170 added ready for merge release-version: 5.10.1 and removed needs: CI release-version: 5.10.2 labels Oct 5, 2018

0xc0170 merged commit baada0e into ARMmbed:master Oct 5, 2018

0xc0170 removed the ready for merge label Oct 5, 2018

jeromecoutant deleted the PR_STM32_LPTICKER branch October 8, 2018 08:02

mattbrown015 mentioned this pull request Oct 23, 2018

STM32L4: Adjusting RTC Synchronisation When Using MBED_TICKLESS in RTC Mode #7695

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

STM32 LPTICKER : Fix tickless and LPTICKER_DELAY_TICKS #8242

STM32 LPTICKER : Fix tickless and LPTICKER_DELAY_TICKS #8242

jeromecoutant commented Sep 25, 2018

jeromecoutant commented Sep 25, 2018

jeromecoutant commented Sep 25, 2018

c1728p9 Sep 25, 2018

bcostm Sep 26, 2018

jeromecoutant Sep 26, 2018

c1728p9 Sep 25, 2018

jeromecoutant Sep 26, 2018

c1728p9 Sep 27, 2018

c1728p9 Sep 25, 2018

jeromecoutant Sep 26, 2018

c1728p9 Sep 25, 2018

jeromecoutant Sep 26, 2018

c1728p9 Sep 27, 2018

LMESTM Sep 27, 2018

c1728p9 commented Sep 25, 2018

0xc0170 commented Sep 26, 2018

jeromecoutant commented Sep 27, 2018

0xc0170 commented Oct 1, 2018

mbed-ci commented Oct 1, 2018

mbed-ci commented Oct 1, 2018

mbed-ci commented Oct 2, 2018

0xc0170 commented Oct 2, 2018

0xc0170 commented Oct 4, 2018

mbed-ci commented Oct 4, 2018

0xc0170 commented Oct 5, 2018

mbed-ci commented Oct 5, 2018

STM32 LPTICKER : Fix tickless and LPTICKER_DELAY_TICKS #8242

STM32 LPTICKER : Fix tickless and LPTICKER_DELAY_TICKS #8242

Conversation

jeromecoutant commented Sep 25, 2018

Description

Test result

Restriction

Pull request type

jeromecoutant commented Sep 25, 2018

jeromecoutant commented Sep 25, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

c1728p9 commented Sep 25, 2018

0xc0170 commented Sep 26, 2018

jeromecoutant commented Sep 27, 2018

0xc0170 commented Oct 1, 2018

mbed-ci commented Oct 1, 2018

Triggering tests

mbed-ci commented Oct 1, 2018

mbed-ci commented Oct 2, 2018

0xc0170 commented Oct 2, 2018

0xc0170 commented Oct 4, 2018

mbed-ci commented Oct 4, 2018

0xc0170 commented Oct 5, 2018

mbed-ci commented Oct 5, 2018