RFC: ADC: add support for injected channels for STM32 #99787

gautierg-st · 2025-11-21T08:44:52Z

This PR adds the support for the STM32 ADC injected channels. Starting an injected sequence means that the sequence is executed immediately. If a regular sequence is being executed, it will be paused and resumed once the injected one is done.

It makes a slight change to the ADC API by adding a boolean in the adc_sequence structure to indicate if the sequence should be injected or not. This boolean is itself hidden behind a specific Kconfig. That way, there is no difference for ADC that do not support this feature.

An ADC test is added to test the feature on an STM32U5 board. The PR is functional but the implementation stays open for discussion.

drivers/adc/adc_stm32.c

etienne-lms · 2025-11-26T09:12:27Z

drivers/adc/adc_stm32.c

+		LL_ADC_ClearFlag_JEOS(adc);
+		for (uint8_t i = 0; i < data->channel_count_injected; i++) {
+			*data->buffer_injected++ = LL_ADC_INJ_ReadConversionData32(adc,
+						   table_rank_injected[i]);


Could you add an indentation tab?

drivers/adc/adc_stm32.c

drivers/adc/Kconfig

erwango · 2025-11-26T09:34:26Z

drivers/adc/Kconfig.stm32

+config ADC_STM32_INJECTED_CHANNELS
+	bool "STM32 ADC Injected channels support"
+	select ADC_INJECTED_CHANNELS


If ADC_INJECTED_CHANNELS is part of generic ADC API, it should be possible to use in a portable application and hence I'd expect the dependency to be in the other way round:
User selects ADC_INJECTED_CHANNELS in his application and ADC_STM32_INJECTED_CHANNELS is selected in turn.
Alternatively ADC_STM32_INJECTED_CHANNELS may not be required at all.

erwango · 2025-11-26T09:37:02Z

drivers/adc/adc_stm32.c


 #if ANY_ADC_SEQUENCER_TYPE_IS(SEQUENCER_PROGRAMMABLE)
 		if (config->sequencer_type == SEQUENCER_PROGRAMMABLE) {
-#if ANY_ADC_HAS_CHANNEL_PRESELECTION


This part of the change is unclear. Should it be in a dedicated commit ?

Writing to the preselection register cannot be done if any sequence (regular or injected) is on-going. So it is necessary to set it up at the "channel setup" stage rather than at the start stage. That's also why this register is not reset when sampling is done (when injected is enabled).

That said, the code should be kept here when injected is not enabled, or else some current tests/samples will fail (because presel will be reset at the end of the sampling, and channel_setup is not necessarily called each time before starting conversions)

drivers/adc/adc_stm32.c

mathieuchopstm

Seems overall OK; some comments:

drivers/adc/Kconfig

drivers/adc/adc_stm32.c

mathieuchopstm · 2025-11-26T09:24:00Z

drivers/adc/adc_stm32.c

+#if ANY_ADC_HAS_CHANNEL_PRESELECTION
+	if (config->has_channel_preselection) {
+		/*
+		 * Each channel in the sequence must be previously enabled in PCSEL.
+		 * This register controls the analog switch integrated in the IO
+		 * level.
+		 */
+		uint32_t channel = __LL_ADC_DECIMAL_NB_TO_CHANNEL(channel_cfg->channel_id);
+
+		LL_ADC_SetChannelPreselection(adc, channel);
+	}
+#endif /* ANY_ADC_HAS_CHANNEL_PRESELECTION */


Not sure how this change relates to injected channels. Maybe it should be moved to a dedicated commit?

Writing to the preselection register cannot be done if any sequence (regular or injected) is on-going. So it is necessary to set it up at the "channel setup" stage rather than at the start stage. That's also why this register is not reset when sampling is done (when injected is enabled).

drivers/sensor/st/stm32_temp/Kconfig

drivers/sensor/st/stm32_vbat/stm32_vbat.c

mathieuchopstm · 2025-11-26T09:41:13Z

drivers/sensor/st/stm32_vbat/Kconfig

 	depends on SOC_FAMILY_STM32 && !SOC_SERIES_STM32F1X
 	select ADC
 	help
 	  Enable driver for STM32 Vbat sensor and then also ADC


Outside PR scope: inconsistency here vs. other drivers

drivers/sensor/st/stm32_vref/stm32_vref.c

dts/bindings/adc/st,stm32-adc.yaml

erwango · 2025-11-27T12:09:31Z

@anangl Unfortunately we miss and ADC maintainer. As the unique contributor to this subsystem, would you mind having a look to the api change and see if acceptable ?

tests/drivers/adc/adc_api/src/test_adc.c

etienne-lms · 2025-11-27T12:43:41Z

tests/drivers/adc/adc_api/src/test_adc.c

 	TC_PRINT("Samples read: ");
-	for (i = 0; i < BUFFER_SIZE; i++) {
-		adc_data_size_t sample_value = m_sample_buffer[i];
+	for (i = 0; i < MAX(BUFFER_SIZE, expected_count); i++) {


This will not test values were not filled above expected_count offset for inject channels' large buffer.
Should pass the buffer size as argument and use instead of BUFFER_SIZE here?

drivers/adc/adc_stm32.c

drivers/sensor/st/stm32_vbat/stm32_vbat.c

Add injected mode support in the ADC API. For hardware supporting it, this allows to inject an ADC sequence in the middle of another one. The injected sequence takes precedence over the regular sequence, interrupting it to make its own conversions. Once the injected sequence is finished, the regular sequence will start back from where it was interrupted. Signed-off-by: Guillaume Gautier <[email protected]>

gautierg-st · 2025-11-28T07:44:47Z

Last push was to fix conflicts for the release notes.

drivers/adc/Kconfig.stm32

drivers/adc/adc_stm32.c

dts/bindings/adc/st,stm32-adc.yaml

mathieuchopstm

Otherwise LGTM

mathieuchopstm · 2025-12-01T14:44:37Z

drivers/adc/Kconfig.stm32

+	bool
+	def_bool y


Suggested change

bool

def_bool y

def_bool y

(def_bool y is equivalent to bool followed by default y)

Add injected mode support for STM32 ADC driver. To avoid potential conflicts when configuring the ADC registers, some settings (like presel) are moved into the channel_setup function when injected is enabled. Also configures an ISR when DMA and injected mode are enabled. The reason is that DMA is not available for injected sequence, so an ISR is still necessary in this case. Factorizes the "start_read_xxx" functions to avoid needless duplication between sync/async and regular/injected mode. Signed-off-by: Guillaume Gautier <[email protected]>

In order to write into some ADC registers, it is necessary to have no on-going conversions (either regular, injected, or both). Until now, each time such a write was needed, the ADC was disabled unconditionally, stopping any potential on-going conversion. With the introduction of the injected mode, this could lead to problems if for example an injected is configured differently than a running regular one. In that case it would stop the regular conversion. If the application is waiting the result before starting another one, it will be stuck waiting forever. To avoid cases like this, return an error rather than stopping conversion and disabling the ADC. This reworks many functions to return an error in case it was not possible to safely disable the ADC. It stays possible to force an ADC disable even though a conversion is running, this is useful at init where we don't want this to fail. This also creates a dedicated function for the preselection register to avoid duplication. Signed-off-by: Guillaume Gautier <[email protected]>

Add injected mode support for STM32 temperature sensor. When in injected mode, this configures the sensor channel and sensor enable bit at start to prevent any conflict later: some registers can't be written while a conversion is on-going. Signed-off-by: Guillaume Gautier <[email protected]>

Add injected mode support for STM32 Vbat sensor. When in injected mode, this configures the sensor channel and sensor enable bit at start to prevent any conflict later: some registers can't be written while a conversion is on-going. Signed-off-by: Guillaume Gautier <[email protected]>

Add injected mode support for STM32 Vref sensor. When in injected mode, this configures the sensor channel and sensor enable bit at start to prevent any conflict later: some registers can't be written while a conversion is on-going. Signed-off-by: Guillaume Gautier <[email protected]>

Defines the io-injected-channels property to be able to define a seconde sequence used for injected channels. Signed-off-by: Guillaume Gautier <[email protected]>

Add a test for injected channels. The test first starts a regular asynchronous sampling of several channels, then starts an injected sampling. The test checks that both the regular and injected samplings run without issues. Signed-off-by: Guillaume Gautier <[email protected]>

The ADC has a new API to enable injected mode (Kconfig + struct member). Signed-off-by: Guillaume Gautier <[email protected]>

sonarqubecloud · 2025-12-01T16:20:39Z

Quality Gate passed

Issues
2 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud

gautierg-st force-pushed the adc_injected_channels branch from 9e9f4ab to bbd2b6c Compare November 21, 2025 14:04

gautierg-st changed the title ~~STM32 ADC: add support for injected channels~~ RFC: ADC: add support for injected channels for STM32 Nov 21, 2025

gautierg-st marked this pull request as ready for review November 21, 2025 14:25

zephyrbot added area: Boards/SoCs area: Devicetree Bindings area: Sensors Sensors platform: STM32 ST Micro STM32 area: ADC Analog-to-Digital Converter (ADC) area: Tests Issues related to a particular existing or missing test labels Nov 21, 2025

zephyrbot requested review from FRASTM, GeorgeCGV, MaureenHelm, anangl, asemjonovs, avisconti, djiatsaf-st, etienne-lms, jeppenodgaard, juickar, mathieuchopstm, teburd, tristan-google, ubieda and yperess November 21, 2025 14:26

zephyrbot assigned MaureenHelm Nov 21, 2025

etienne-lms reviewed Nov 26, 2025

View reviewed changes

erwango requested changes Nov 26, 2025

View reviewed changes

mathieuchopstm reviewed Nov 26, 2025

View reviewed changes

etienne-lms reviewed Nov 27, 2025

View reviewed changes

gautierg-st force-pushed the adc_injected_channels branch from bbd2b6c to 8e70d05 Compare November 27, 2025 15:35

zephyrbot requested review from erwango and nashif November 27, 2025 15:37

zephyrbot added the Release Notes To be mentioned in the release notes label Nov 27, 2025

zephyrbot requested review from cfriedt and jhedberg November 27, 2025 15:44

gautierg-st force-pushed the adc_injected_channels branch from 956a746 to b2be6ee Compare November 27, 2025 15:45

gautierg-st force-pushed the adc_injected_channels branch from b2be6ee to e50014e Compare November 28, 2025 07:44

gautierg-st force-pushed the adc_injected_channels branch from e50014e to f5386d9 Compare November 28, 2025 08:10

mathieuchopstm reviewed Nov 28, 2025

View reviewed changes

drivers/adc/Kconfig.stm32 Outdated Show resolved Hide resolved

drivers/adc/adc_stm32.c Outdated Show resolved Hide resolved

dts/bindings/adc/st,stm32-adc.yaml Show resolved Hide resolved

gautierg-st force-pushed the adc_injected_channels branch 3 times, most recently from 2ea924e to 0c20dfc Compare December 1, 2025 10:30

mathieuchopstm reviewed Dec 1, 2025

View reviewed changes

gautierg-st added 8 commits December 1, 2025 16:53

dts: bindings: adc: stm32: add io-injected-channels binding

a846418

Defines the io-injected-channels property to be able to define a seconde sequence used for injected channels. Signed-off-by: Guillaume Gautier <[email protected]>

doc: releases: 4.4: add the new api for adc injected mode

4064a88

The ADC has a new API to enable injected mode (Kconfig + struct member). Signed-off-by: Guillaume Gautier <[email protected]>

gautierg-st force-pushed the adc_injected_channels branch from 0c20dfc to 4064a88 Compare December 1, 2025 15:54

RFC: ADC: add support for injected channels for STM32 #99787

Are you sure you want to change the base?

RFC: ADC: add support for injected channels for STM32 #99787

Conversation

gautierg-st commented Nov 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

mathieuchopstm left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

erwango commented Nov 27, 2025

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

gautierg-st commented Nov 28, 2025

Uh oh!

Uh oh!

Uh oh!

Uh oh!

mathieuchopstm left a comment

Choose a reason for hiding this comment

Uh oh!

mathieuchopstm Dec 1, 2025 • 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!

Choose a reason for hiding this comment

Uh oh!

sonarqubecloud bot commented Dec 1, 2025

Quality Gate passed

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants

gautierg-st commented Nov 21, 2025 •

edited

Loading

mathieuchopstm Dec 1, 2025 •

edited

Loading