ST: Watchdog: Fix timeout registers value calculation #11178

Tharazi97 · 2019-08-07T11:01:57Z

Description

Proposition of fixing: #11175
Changes only implementation of watchdog on ST devices, but problem applies mainly to the meaning of this test. I don't know if this is a right way of fixing that. @fkjagodzinski might help.

Pull request type

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

Reviewers

@maciejbocianski @mprse @jamesbeyond @fkjagodzinski

Release Notes

ciarmcom · 2019-08-07T13:00:39Z

@Tharazi97, thank you for your changes.
@mprse @maciejbocianski @jamesbeyond @fkjagodzinski @ARMmbed/mbed-os-core @ARMmbed/mbed-os-test @ARMmbed/mbed-os-maintainers please review.

jamesbeyond · 2019-08-07T14:21:54Z

TESTS/mbed_drivers/watchdog/main.cpp

@@ -18,6 +18,10 @@
 #error [NOT_SUPPORTED] Watchdog not supported for this target
 #else

+#ifndef MAX_IWDG_PR
+#define MAX_IWDG_PR 0x6


Why is the max pre-scaler defined as 6 ?

Right, I didn't look to other targets than STM and assumed that it is used like this in all boards. I'll take a peek to other drivers and make it work on all devices.

mbed-ci · 2019-08-07T14:37:24Z

Test run: FAILED

Summary: 3 of 3 test jobs failed
Build number : 1
Build artifacts

Failed test jobs:

jenkins-ci/mbed-os-ci_build-ARM
jenkins-ci/mbed-os-ci_build-GCC_ARM
jenkins-ci/mbed-os-ci_build-IAR

0xc0170 · 2019-08-08T09:33:04Z

TESTS/mbed_drivers/watchdog/main.cpp

-    // The watchdog should trigger at, or after the timeout value.
-    TEST_ASSERT(watchdog.get_timeout() >= max_timeout);
+    // The watchdog should trigger at, or after the timeout value - maximum time spend on prescaller.
+    uint32_t max_window = ceil((float)(4 << MAX_IWDG_PR) * 1000 / LSI_VALUE);


this is also one MCU family specific value. See LSI_VALUE error in the build logs.

The test need to be generic and using driver functionality (=test).

Tharazi97 · 2019-08-08T10:54:27Z

I've changed the function in stm watchdog_api.c so it rounds the value of timeout up. This makes the test pass, but it still needs to be discussed if it's the right way of doing that. @fkjagodzinski might help in clarifying this problem.

0xc0170 · 2019-08-09T12:39:21Z

TESTS/mbed_drivers/watchdog/main.cpp

@@ -225,7 +225,7 @@ void test_start_max_timeout()
    Watchdog &watchdog = Watchdog::get_instance();
    uint32_t max_timeout = watchdog.get_max_timeout();
    TEST_ASSERT_TRUE(watchdog.start(max_timeout));
-    // The watchdog should trigger at, or after the timeout value.
+    // The watchdog should trigger at, or after the timeout value;


this change is not related. the rest of the comments in this file has it as it was. please revert

0xc0170 · 2019-08-09T12:40:16Z

This currently is touching only ST targets, this should be stated (title and updated comment about the scope of this fix)

fkjagodzinski

Please clean up the commit history -- if you make an update and then revert it, this shouldn't be visible here.
Please update the commit messages -- e.g. change in stm watchdog doesn't say much. The commit message may be as long as you need it to be to explain why and what was updated. Take a look at our commit history for reference.
If the code needs an astyle update to pass the CI, please keep that in a separate commit.

fkjagodzinski · 2019-08-14T09:36:27Z

targets/TARGET_STM/watchdog_api.c


 // Convert Prescaler_divider bits (PR) of Prescaler_register (IWDG_PR) and a timeout value [ms]
 // to Watchdog_counter_reload_value bits (RL) of Reload_register (IWDG_RLR)
 #define PR_TIMEOUT_MS2RL(PR_BITS, TIMEOUT_MS) \
-	((TIMEOUT_MS) * (LSI_VALUE) / (PR2PRESCALER_DIV(PR_BITS)) / 1000UL)
+    ((uint32_t)ceil((float)(TIMEOUT_MS) * (LSI_VALUE) / (PR2PRESCALER_DIV(PR_BITS)) / 1000UL))


Avoid using floating point if possible. The code below will give same values as ceil().

Suggested change

((uint32_t)ceil((float)(TIMEOUT_MS) * (LSI_VALUE) / (PR2PRESCALER_DIV(PR_BITS)) / 1000UL))

(((TIMEOUT_MS) * (LSI_VALUE) / (PR2PRESCALER_DIV(PR_BITS)) + 999UL) / 1000UL)

fkjagodzinski · 2019-08-14T09:48:26Z

targets/TARGET_STM/watchdog_api.c

-    const uint32_t rl = PR_TIMEOUT_MS2RL(pr, config->timeout_ms);
+    uint32_t rl = PR_TIMEOUT_MS2RL(pr, config->timeout_ms);
+    if (rl > MAX_IWDG_RL) {
+        rl = MAX_IWDG_RL;


Isn't that a dead code? pr != INVALID_IWDG_PR should guarantee that rl is also valid.

fkjagodzinski · 2019-08-14T10:12:14Z

@Tharazi97 could you also update the name and the description of this PR? From what I see now, this could be as simple as ST: Watchdog: Fix timeout registers value calculation. ;)

Change the calculation method of rl so it is rounded up.

jeromecoutant

NUCLEO_L073RZ watchdog tests are now OK!
DISCO_L475VG_IOT01A watchdog tests are still OK.

Thx for the fix

0xc0170 · 2019-08-19T12:20:55Z

CI started

mbed-ci · 2019-08-20T00:09:26Z

Test run: SUCCESS

Summary: 11 of 11 test jobs passed
Build number : 2
Build artifacts

jamesbeyond · 2019-08-20T11:52:21Z

@donatieng we fixed a watchdog issue related to NUCLEO_L073RZ and someother STM target, you might need to aware of this change

ciarmcom requested review from fkjagodzinski, jamesbeyond, maciejbocianski, mprse and a team August 7, 2019 13:00

ciarmcom added the needs: review label Aug 7, 2019

jamesbeyond reviewed Aug 7, 2019

View reviewed changes

0xc0170 added needs: work and removed needs: review labels Aug 8, 2019

0xc0170 requested changes Aug 8, 2019

View reviewed changes

0xc0170 reviewed Aug 9, 2019

View reviewed changes

0xc0170 requested a review from a team August 9, 2019 12:39

Tharazi97 force-pushed the watchdog_max_timeout branch from 6e9eafd to eb558de Compare August 9, 2019 12:54

Tharazi97 changed the title ~~proposition of fix watchdog start max timeout~~ proposition of fix watchdog start max timeout (tweak ST implementation) Aug 9, 2019

Tharazi97 changed the title ~~proposition of fix watchdog start max timeout (tweak ST implementation)~~ proposition of fix watchdog start max timeout (tweak STM implementation) Aug 9, 2019

Tharazi97 force-pushed the watchdog_max_timeout branch from eb558de to 302be42 Compare August 12, 2019 07:46

Tharazi97 changed the title ~~proposition of fix watchdog start max timeout (tweak STM implementation)~~ fix watchdog start max timeout and test update config (tweak STM implementation) Aug 12, 2019

fkjagodzinski suggested changes Aug 14, 2019

View reviewed changes

Tharazi97 changed the title ~~fix watchdog start max timeout and test update config (tweak STM implementation)~~ ST: Watchdog: Fix timeout registers value calculation Aug 14, 2019

Tharazi97 force-pushed the watchdog_max_timeout branch from 302be42 to a75d43b Compare August 14, 2019 10:58

Tweak STM watchdog implementation

df43350

Change the calculation method of rl so it is rounded up.

Tharazi97 force-pushed the watchdog_max_timeout branch from a75d43b to df43350 Compare August 14, 2019 11:03

jeromecoutant approved these changes Aug 19, 2019

View reviewed changes

0xc0170 approved these changes Aug 19, 2019

View reviewed changes

fkjagodzinski approved these changes Aug 19, 2019

View reviewed changes

0xc0170 added ready for merge release-version: 5.13.4 and removed needs: work labels Aug 20, 2019

0xc0170 merged commit ce74bb5 into ARMmbed:master Aug 20, 2019

0xc0170 removed the ready for merge label Aug 20, 2019

fkjagodzinski mentioned this pull request Aug 22, 2019

Watchdog issue with timeout #11175

Closed

	((uint32_t)ceil((float)(TIMEOUT_MS) * (LSI_VALUE) / (PR2PRESCALER_DIV(PR_BITS)) / 1000UL))
	(((TIMEOUT_MS) * (LSI_VALUE) / (PR2PRESCALER_DIV(PR_BITS)) + 999UL) / 1000UL)

ST: Watchdog: Fix timeout registers value calculation #11178

ST: Watchdog: Fix timeout registers value calculation #11178

Uh oh!

Conversation

Tharazi97 commented Aug 7, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Pull request type

Reviewers

Release Notes

Uh oh!

ciarmcom commented Aug 7, 2019

Uh oh!

jamesbeyond Aug 7, 2019

Choose a reason for hiding this comment

Uh oh!

Tharazi97 Aug 7, 2019

Choose a reason for hiding this comment

Uh oh!

mbed-ci commented Aug 7, 2019

Test run: FAILED

Uh oh!

0xc0170 Aug 8, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Tharazi97 commented Aug 8, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

0xc0170 Aug 9, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Aug 9, 2019

Uh oh!

fkjagodzinski left a comment

Choose a reason for hiding this comment

Uh oh!

fkjagodzinski Aug 14, 2019

Choose a reason for hiding this comment

Uh oh!

fkjagodzinski Aug 14, 2019

Choose a reason for hiding this comment

Uh oh!

fkjagodzinski commented Aug 14, 2019

Uh oh!

jeromecoutant left a comment

Choose a reason for hiding this comment

Uh oh!

0xc0170 commented Aug 19, 2019

Uh oh!

mbed-ci commented Aug 20, 2019

Test run: SUCCESS

Uh oh!

jamesbeyond commented Aug 20, 2019

Uh oh!

Uh oh!

Tharazi97 commented Aug 7, 2019 •

edited

Loading

0xc0170 Aug 8, 2019 •

edited

Loading

Tharazi97 commented Aug 8, 2019 •

edited

Loading

0xc0170 Aug 9, 2019 •

edited

Loading