WIP: HRTIM #96

usbalbin · 2023-11-30T21:55:36Z

Work in progress. This will likely be further split up across multiple PRs.

Currently requires stm32-rs/stm32-rs#860 , thus all the CI errors

TODO:

Traitify PreloadSource, MasterPreloadSource
Update examples for new traitified changes
- hrtim/adc-trigger
Make sure the timers are started in the examples
Go through the examples once more in case of any last minute experiment leftovers

usbalbin · 2023-11-30T21:58:07Z

examples/hrtim/eev.rs

@@ -0,0 +1,113 @@
+//This example puts the timer in PWM mode using the specified pin with a frequency of 100Hz and a duty cycle of 50%.


These descriptions should be updated

Almost all of the hrtim examples are just copy pasted from examples/hrtim/hrtim.rs or examples/hrtim/simple.rs and modified without having their comments updated.

usbalbin · 2023-12-03T09:22:54Z

examples/hrtim/eev-comp.rs

+
+    out1.enable_rst_event(&cr1); // Set low on compare match with cr1
+    out1.enable_rst_event(eev_input4);
+    out1.enable_set_event(&timer); // Set high at new period


@no111u3 sort of related to #98...

I have been experimenting a bit with different ways to specify different types of event sources.

Here an event is specified by passing a reference to the thing to listen to. In this case:

reset output on compare register 1 match

reset output on external event 4 signal

set output on timer period event

The nice thing about this is that there is no way to specify the wrong event source by accident. Also there is no need to figure out what enum and variant to to use, just pass in the actual object(or reference to it).

However this requires the event source to exist before the listener which might make things more complicated?

@usbalbin thank for this great example. Yesterday I've merged some changes but it will update in some next sessions (for every timers also). So I think about correct and wrong source for event, it might be provided by internal mark field (I'm not sure that rust can do it).

However this requires the event source to exist before the listener which might make things more complicated?

Following the RM on G4 (and other) family:

Note: The clock of the slave timer or ADC must be enabled prior to receive events from the master timer, and must not be changed on-the-fly while triggers are received from the master timer.

That means consumer must be created first, I've checked it on my g474 — violation rule makes ADC crazy.

[..] Yesterday I've merged some changes [..]

Nice :)

It least when it comes to hrtim things might get quite complicated(I have not looked at a lot of other timers yet). For example I believe it is possible to use a:

DAC to output, for example, a saw tooth shape

This is routed to the comparators inverting input

The comparators output is routed to the hrtim EEV1

EEV1 is used as an rst event for an hrtim timers output.

That same hrtim timer triggers the update of the DAC's wave form, and around we go

Which object should be created first? :)

Whatever API we pick, it should preferably support doing things like that, while preventing the user from doing something that will violate the RM's rules.

examples/hrtim/adc-trigger.rs

usbalbin · 2024-01-01T21:40:16Z

examples/hrtim/adc-trigger.rs

+        .HRTIM_COMMON
+        .hr_control(&mut rcc)
+        .enable_adc_trigger1_source(Adc13Trigger::TimACmp3)
+        .enable_adc_trigger1_source(Adc13Trigger::TimACmp4)


If this enable method was instead defined on HrPwmControl then the variables cr3 and cr4 could be passed directly to that method without the need for the enum.

@no111u3 what do you think about something like this d8d2b92 to avoid enums? (Will probably update the examples at some point)

Hmm, I think it better to use one work variant instead of any workless.

examples/hrtim/adc-trigger.rs

usbalbin · 2024-01-19T13:09:10Z

I have now gotten rid of a few enums in favor of traits to force the user to pass a reference to the actual type instead of an enum variant. See examples/adc-trigger.rs for how that looks like(this is the only example updated so far, it is not yet tested).

For anyone wanting to give it a try, I am using usbalbin/stm32-rs@ab9082b until the next stm32-rs/stm32-rs release

* Rcc enable is now handled for the entire HRTIM peripheral by dp.HRTIM_COMMON.hr_control() instead of by every timer * HRTIM calibration is now mandatory, also handled by `dp.HRTIM_COMMON.hr_control()` but also by running on that result `_.wait_for_calibration(). * Add `get_state` for `HrOutput` to check what state an output is in: `Idle`, `Running` or `Fault`

* dma - update for new pac * adc - triggers * hrtim

usbalbin commented Nov 30, 2023

View reviewed changes

usbalbin commented Dec 3, 2023

View reviewed changes

usbalbin mentioned this pull request Dec 16, 2023

[DON'T MERGE] Update pac #102

Draft

usbalbin mentioned this pull request Dec 26, 2023

stm32/hrtim: refactor to deduplicate code between channels. embassy-rs/embassy#2318

Open

usbalbin commented Jan 1, 2024

View reviewed changes

examples/hrtim/adc-trigger.rs Show resolved Hide resolved

usbalbin commented Jan 1, 2024

View reviewed changes

examples/hrtim/adc-trigger.rs Outdated Show resolved Hide resolved

usbalbin commented Jan 1, 2024

View reviewed changes

examples/hrtim/adc-trigger.rs Outdated Show resolved Hide resolved

usbalbin mentioned this pull request Jan 5, 2024

Timer driven Events #98

Open

usbalbin force-pushed the hrtim_new_attempt branch 2 times, most recently from 6a4b3d0 to da8337e Compare January 19, 2024 09:21

usbalbin force-pushed the hrtim_new_attempt branch 2 times, most recently from 966c1d9 to 9ad72a4 Compare January 24, 2024 16:09

usbalbin and others added 16 commits February 16, 2024 13:15

WIP

224f0d5

Untested sort of maybe working chaos

58f44f3

Dual channel support and cargo fmt

108c7b3

Almost...

ec9ec6f

Still untested but might actually work, maybe

af66e36

Adjust doc

7e1a45c

Enable outputs in example

b56e2fd

Further cleanup

f2de47f

Tested and verified on NUCLEO-G474RE

a5e7f3a

hrtim_simple also tested and verified on NUCLEO-G474RE

2dcae09

Hrtim cleanup, add pins for all timers but HRTIM_TIMF

db94990

Start work on hrtim fault

ec58b83

Add hrtim fault example

68feb9f

TOO MANY CHANGES

ce22d9c

* dma - update for new pac * adc - triggers * hrtim

Add ADC triggers among a lot of things to hrtim

688a867

usbalbin and others added 25 commits February 16, 2024 13:15

Fix warnings for hrtim

1942ca7

HRTIM - Traitify adc_triggers

2217a97

HRTIM: Start work on capture support

4213ce0

HRTIM: Add capture.rs

0350eb0

HRTIM: Impl CompareEvent for more types

20205c7

Fmt

5a846ef

HRTIM: Traitify events

5dd9335

HRTIM: Fix bugs caused by traitification

e089ce5

HRTIM: Update adc-trigger example for traitification changes

a6ed404

HRTIM: Start timer and set trigger points in adc-trigger example

27075e1

HRTIM: Refactor adc trigger

b7bacde

HRTIM: Refactor adc trigger - example hrtim-adc-trigger

0e763b9

HRTIM: adc trigger impl into

1626954

HRTIM: adc-trigger example set duty

944985b

HRTIM: Move clear_repetition_interrupt to trait

eb075b0

HRTIM: Capture interrupts and clippy

ab4c3b5

HRTIM: Capture example

8a7e208

HRTIM: Create feature for all devices supporting hrtim

a823f21

HRTIM: Some fixes for capture example

2924811

HRTIM: Fix warning in capture example

6e1d446

HRTIM: Add forgotten traitification impls

b97dada

HRTIM: Update examples

749c064

HRTIM: Capture cleanup

47d7a43

HRTIM: Avoid unneeded type parameters on HrTimer trait

6887958

HRTIM: Move some inherent methods to traits on HrTim

5ba64f8

usbalbin force-pushed the hrtim_new_attempt branch from 9ad72a4 to 5ba64f8 Compare February 16, 2024 12:16

usbalbin added 4 commits April 19, 2024 10:36

HRTIM - Add derives for some types

066f4a3

HRTIM - Allow using capture module with DMA

8ebf645

HRTIM - Fix example

6c619a9

HRTIM - Prevent future dataraces for DMA

f8018d1

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

WIP: HRTIM #96

WIP: HRTIM #96

usbalbin commented Nov 30, 2023 •

edited

usbalbin Nov 30, 2023

usbalbin Dec 3, 2023

usbalbin Dec 3, 2023

no111u3 Dec 3, 2023

usbalbin Dec 3, 2023 •

edited

usbalbin Jan 1, 2024

usbalbin Jan 7, 2024

no111u3 Jan 7, 2024

usbalbin commented Jan 19, 2024

		@@ -0,0 +1,113 @@
		//This example puts the timer in PWM mode using the specified pin with a frequency of 100Hz and a duty cycle of 50%.

WIP: HRTIM #96

Are you sure you want to change the base?

WIP: HRTIM #96

Conversation

usbalbin commented Nov 30, 2023 • edited

usbalbin Nov 30, 2023

Choose a reason for hiding this comment

usbalbin Dec 3, 2023

Choose a reason for hiding this comment

usbalbin Dec 3, 2023

Choose a reason for hiding this comment

no111u3 Dec 3, 2023

Choose a reason for hiding this comment

usbalbin Dec 3, 2023 • edited

Choose a reason for hiding this comment

usbalbin Jan 1, 2024

Choose a reason for hiding this comment

usbalbin Jan 7, 2024

Choose a reason for hiding this comment

no111u3 Jan 7, 2024

Choose a reason for hiding this comment

usbalbin commented Jan 19, 2024

usbalbin commented Nov 30, 2023 •

edited

usbalbin Dec 3, 2023 •

edited