[topic-gpio] drivers: gpio_stm32: update to use new GPIO API #19579

erwango · 2019-10-03T13:06:35Z

Update driver code to use new configuration flags such as GPIO_ACTIVE_LOW.
~~STM32F1 series not taken into account.~~
~~Only one board converted.~~
~~No backward compatibility: gpio_configure don't configure interrupts anymore~~

Based on:

Provide arduino connector to nucleo_64 (boards: stm32: Set arduino gpio connector on nucleo 64 pins boards #19628) and nucleo_144 (boards: stm32: Set arduino gpio connector on nucleo 144 pins boards #19636)
[topic-gpio] drivers/gpio: stm32: Rework configure function exit for dual core #19646 [topic-gpio] drivers/gpio: stm32: Rework configure function exit for dual core
gpio: stm32: support disabling and reenabling interrupts on gpio pin #19256 as new API assumes possibility to disabling and re-enabling interrupts

TODO:

So far STM32F1 series are not taken into account. Extend to change to this series
Get all STM32 based boards converted to the new API
~~Done in dedicated PR [topic-gpio] boards: stm32: update to use new GPIO API #19607 in order to get useful Shippable result in current PR~~ Done in current PR
Test on all series
Identified nucleo_144 and nucleo_64 boards have different active states. ~~To be corrected in [topic-gpio] boards: stm32: update to use new GPIO API #19607~~ ~~Might be trickier than pure nucleo_64 vs nucleo_144. Need to check manuals in detail.~~ Done

EDIT:
PR description rework after push on 09-09-19

erwango · 2019-10-03T13:07:30Z

@mniestroj, FYI

erwango · 2019-10-03T13:09:38Z

Tested ok on gpio_basic_api but FATAL error when test is run w/o wire.
So this requires some more analysis.

erwango · 2019-10-03T13:16:15Z

Tested ok on gpio_basic_api but FATAL error when test is run w/o wire.
So this requires some more analysis.

Seems to be intended behavior.

galak · 2019-10-03T13:17:25Z

Tested ok on gpio_basic_api but FATAL error when test is run w/o wire.
So this requires some more analysis.

Seems to be intended behavior.

Yeah, seen the same thing myself

galak · 2019-10-03T13:19:37Z

Get all STM32 based boards converted to the new API

While maybe not the most efficient, here's an example to help convert the dts files over:

git grep -l \<st boards/arm | grep dts | xargs sed -i 's/GPIO_INT_ACTIVE_HIGH/GPIO_ACTIVE_LOW/g'

galak · 2019-10-03T13:24:46Z

Marking dnm since there are some todo items.

zephyrbot · 2019-10-03T13:58:11Z

All checks are passing now.

Tip: The bot edits this comment instead of posting a new one, so you can check the comment's history to see earlier messages.

erwango · 2019-10-03T18:24:47Z

recheck

pabigot

Observed one flag test using the zero-valued variant, which needs to be fixed. One or two other comments.

Checked on nucleo_l476rg, things passed.

drivers/gpio/gpio_stm32.c

mnkp

I believe it would be better to have all stm32 changes in one PR instead of two. It makes it difficult to do the review since both PRs modify driver code. Also by not running checks on all the boards we may miss some issues. In case of Shippable timeout it's better to add a temporary patch from @galak that fixes the issue.

drivers/gpio/gpio_stm32.c

erwango · 2019-10-07T11:12:45Z

@mnkp Thanks for review!

It makes it difficult to do the review since both PRs modify driver code.

This should not be the case PR #19067 only adds commit 717de74 which impact is limited to boards

Also by not running checks on all the boards we may miss some issues. In case of Shippable timeout it's better to add a temporary patch from @galak that fixes the issue.

Agree it's better to actually get the results on all boards. @galak how can we make this happen?

pabigot · 2019-10-07T11:32:54Z

@erwango As suggested to me by @carlescufi:

From e888dbf355ca8a63ec17ed725368ebb2ec79a200 Mon Sep 17 00:00:00 2001
From: Peter Bigot <[email protected]>
Date: Sun, 6 Oct 2019 15:05:28 -0500
Subject: [PATCH] DNM shippable.yml: increase builds to avoid timeout failures

Signed-off-by: Peter Bigot <[email protected]>
---
 .shippable.yml | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/.shippable.yml b/.shippable.yml
index 605fd41be8..b50342a5ca 100644
--- a/.shippable.yml
+++ b/.shippable.yml
@@ -6,13 +6,20 @@ env:
     global:
         - ZEPHYR_SDK_INSTALL_DIR=/opt/sdk/zephyr-sdk-0.10.3
         - ZEPHYR_TOOLCHAIN_VARIANT=zephyr
-        - MATRIX_BUILDS="5"
+        - MATRIX_BUILDS="12"
     matrix:
         - MATRIX_BUILD="1"
         - MATRIX_BUILD="2"
         - MATRIX_BUILD="3"
         - MATRIX_BUILD="4"
         - MATRIX_BUILD="5"
+        - MATRIX_BUILD="6"
+        - MATRIX_BUILD="7"
+        - MATRIX_BUILD="8"
+        - MATRIX_BUILD="9"
+        - MATRIX_BUILD="10"
+        - MATRIX_BUILD="11"
+        - MATRIX_BUILD="12"
 
 build:
     cache: false
-- 
2.20.1

Also #19562 (comment)

erwango · 2019-10-07T11:35:46Z

@pabigot

@erwango As suggested to me by @carlescufi:

Thanks, will use that.

erwango · 2019-10-11T13:29:46Z

Status following last push:

Tested on all STM32 series.

Test status:

gpio_basic_api: Passed on all available boards except stm32h747_disco_m7, cf below
gpio_api_1pin: Passed on available boards except stm32h747_disco_m7, and stm32mp157c_dk2 cf below
button: Passed on all available boards

Test failures on `stm32h747_disco_m7`:

gpio_basic_api:
Playing with other boards, I've seen several times that this specific failure can happen when pin selected is in conflict with other usage. I've not been able to identify the conflict yet, but I heavily suspsect this issue is purely a configuration issue.
gpio_api_1pin:
Can be PASSED if changing pin.

STM32H7 port is still very basic for now (only GPIO, UART) and globally a wip.
Given this status and board complexity that might explain itself tests failures, I think it is safe to enable new GPIO API on this series.

Test failures on `stm32mp157c_dk2`:

gpio_api_1pin: Failing point relates to port testing. Due to use of Linux on Cortex-A, lot of pins are configured for various purposes (ethernet, display, ...), there is no empty port available for testing matching pins available on arduino connector. I'll have a new try once [TOPIC-GPIO] gpi_api_1pin test failures #19692 is fixed.

Given board complexity that might explain tests failures, I think it is safe to enable new GPIO API on this series.

Summary

Unless reviewers comment, I consider the work done on this PR.

carlescufi · 2019-10-15T16:29:54Z

@erwango could you please rebase?

This allows compiler to inline function body and reduce overall code size. Signed-off-by: Marcin Niestroj <[email protected]>

This patch doesn't change functionality, but is only related to improved readability and reusability. Signed-off-by: Marcin Niestroj <[email protected]>

Up to now interrupts could be only configured once, with no way to disable them in runtime. Allow interrupts to be disabled in runtime and then properly reenabled on user request. This allows to ignore interrupts when software is not expecting them. The improvement over previously reverted patch [1] is that we disable interrupts only when we configure port for which interrupt line was previously selected. This for example prevents to disable interrupts line 2 in case PA2 was previously configured as interrupt source, but we are currently configuring PB2 as output. [1] 0951ce2 ("gpio: stm32: support disabling and reenabling interrupts on pin") Signed-off-by: Marcin Niestroj <[email protected]>

Update STM32 GPIO driver to support new GPIO API. Signed-off-by: Erwan Gouriou <[email protected]>

Perform few clean up in stm32 gpio driver: *Clean up uint32_t occurrences left over. *Move function gpio_stm32_flags_to_conf from const to static and remove useless parameter check *Rework error handling in functoin gpio_stm32_config *Remove gpio_stm32_int_enabled_port function and have direct call to gpio_stm32_get_exti_source Signed-off-by: Erwan Gouriou <[email protected]>

Since it is now possible to disable/re-enable interrupts and also to reconfigure an already configured interrupt, it is now required to clear non requested triggers. While it is not strictly requested, triggers are also cleared when interrupt is disabled (assuming trigger should be configured when interrupt is enabled). Signed-off-by: Erwan Gouriou <[email protected]> fixup exti

Move GPIO_ACTIVE_INT_HIGH/LOW to GPIO_ACTIVE_HIGH/LOW. Signed-off-by: Erwan Gouriou <[email protected]>

Add boards overlay to enable gpio_basic tests. Selected boards should reflect various stm32 series. Signed-off-by: Erwan Gouriou <[email protected]>

On some stm32 based boards buttons and leds configuration was wrong. Fix that. Signed-off-by: Erwan Gouriou <[email protected]>

Add arduino connection dts description and update target yaml files. Signed-off-by: Erwan Gouriou <[email protected]>

Argument 'port' in stm32_exti_set_callback function is not used, remove it. Signed-off-by: Erwan Gouriou <[email protected]>

stm32_exti_enable was returning errors on line > 32 or line pointing to non implemented line. Both conditions are hard-coded, hence there is no use to detect them dynamically in the code. Check them with assert. As a consequence, function could now be void. Additionally, enable exti irq line only if both checks are passed. Signed-off-by: Erwan Gouriou <[email protected]>

Add led and button to board description to enable basic gpio tests. Signed-off-by: Erwan Gouriou <[email protected]>

Test generates lengthy log. Signed-off-by: Erwan Gouriou <[email protected]>

Change codeowner following driver indeep rework. Signed-off-by: Erwan Gouriou <[email protected]>

erwango · 2019-10-16T06:53:06Z

@erwango could you please rebase?

@carlescufi, done.

kklimek · 2020-01-09T19:58:32Z

drivers/interrupt_controller/exti_stm32.h

 	/* trigger on rising edge */
 	STM32_EXTI_TRIG_RISING  = 0x1,
 	/* trigger on falling endge */
 	STM32_EXTI_TRIG_FALLING = 0x2,
+	/* trigger on falling endge */


erwango requested review from mnkp, pabigot and rsalveti as code owners October 3, 2019 13:06

erwango requested a review from galak October 3, 2019 13:06

erwango added the platform: STM32 ST Micro STM32 label Oct 3, 2019

galak changed the title ~~drivers: gpio_stm32: update to use new GPIO API~~ [topic-gpio] drivers: gpio_stm32: update to use new GPIO API Oct 3, 2019

galak added area: GPIO DNM This PR should not be merged (Do Not Merge) labels Oct 3, 2019

zephyrbot added area: Boards area: Tests Issues related to a particular existing or missing test labels Oct 3, 2019

erwango force-pushed the gpio_stm32 branch from fee4c3f to f0f6d22 Compare October 3, 2019 16:14

erwango force-pushed the gpio_stm32 branch 2 times, most recently from 89ea7b0 to a6958f1 Compare October 4, 2019 08:14

erwango mentioned this pull request Oct 4, 2019

[topic-gpio] boards: stm32: update to use new GPIO API #19607

Closed

erwango force-pushed the gpio_stm32 branch from a6958f1 to df1e47d Compare October 4, 2019 11:58

galak mentioned this pull request Oct 5, 2019

Convert GPIO drivers to new GPIO API #18530

Closed

pabigot requested changes Oct 5, 2019

View reviewed changes

drivers/gpio/gpio_stm32.c Show resolved Hide resolved

drivers/gpio/gpio_stm32.c Outdated Show resolved Hide resolved

drivers/gpio/gpio_stm32.c Outdated Show resolved Hide resolved

mnkp requested changes Oct 7, 2019

View reviewed changes

drivers/gpio/gpio_stm32.c Outdated Show resolved Hide resolved

drivers/gpio/gpio_stm32.c Outdated Show resolved Hide resolved

drivers/gpio/gpio_stm32.c Outdated Show resolved Hide resolved

drivers/gpio/gpio_stm32.c Outdated Show resolved Hide resolved

erwango force-pushed the gpio_stm32 branch from df1e47d to ddc3da1 Compare October 7, 2019 15:38

erwango force-pushed the gpio_stm32 branch from 2602553 to 0da014f Compare October 11, 2019 13:20

pabigot approved these changes Oct 11, 2019

View reviewed changes

erwango removed the DNM This PR should not be merged (Do Not Merge) label Oct 11, 2019

ydamigos approved these changes Oct 11, 2019

View reviewed changes

erwango mentioned this pull request Oct 15, 2019

gpio: stm32: support disabling and reenabling interrupts on gpio pin #19256

Closed

erwango force-pushed the gpio_stm32 branch from 0da014f to 260863c Compare October 15, 2019 07:35

mniestroj and others added 15 commits October 16, 2019 08:51

gpio: stm32: statify gpio_stm32_enable_int()

e1c8b17

This allows compiler to inline function body and reduce overall code size. Signed-off-by: Marcin Niestroj <[email protected]>

gpio: stm32: split helper functions from gpio_stm32_enable_int()

1883638

This patch doesn't change functionality, but is only related to improved readability and reusability. Signed-off-by: Marcin Niestroj <[email protected]>

drivers/gpio: stm32: Update to new GPIO api

fb06d5b

Update STM32 GPIO driver to support new GPIO API. Signed-off-by: Erwan Gouriou <[email protected]>

boards: stm32: Update board definition using new GPIO api

26adc25

Move GPIO_ACTIVE_INT_HIGH/LOW to GPIO_ACTIVE_HIGH/LOW. Signed-off-by: Erwan Gouriou <[email protected]>

tests: gpio_basic_api: Enable test on stm32 boards

4e159d4

Add boards overlay to enable gpio_basic tests. Selected boards should reflect various stm32 series. Signed-off-by: Erwan Gouriou <[email protected]>

boards: stm32: Fix buttons and leds configuration

c0c3f85

On some stm32 based boards buttons and leds configuration was wrong. Fix that. Signed-off-by: Erwan Gouriou <[email protected]>

boards: stm32h747i_disco: Add arduino connector header

f2e3551

Add arduino connection dts description and update target yaml files. Signed-off-by: Erwan Gouriou <[email protected]>

drivers/interrupt_controller: stm32: Clean set api from unused arg

b8afa26

Argument 'port' in stm32_exti_set_callback function is not used, remove it. Signed-off-by: Erwan Gouriou <[email protected]>

boards: stm32mp157c_dk2: Add led and button to dts

0de980e

Add led and button to board description to enable basic gpio tests. Signed-off-by: Erwan Gouriou <[email protected]>

tests: gpio_api_1pin: Tune RAM console size for stm32mp157c_dk2

54508a7

Test generates lengthy log. Signed-off-by: Erwan Gouriou <[email protected]>

CODEWONERS: Update for gpio_stm32 driver

c966c4f

Change codeowner following driver indeep rework. Signed-off-by: Erwan Gouriou <[email protected]>

erwango force-pushed the gpio_stm32 branch from 260863c to c966c4f Compare October 16, 2019 06:52

erwango mentioned this pull request Oct 16, 2019

[TOPIC-GPIO] drivers: sensor: lsm6dsl: update to new GPIO API #19839

Merged

MaureenHelm approved these changes Oct 16, 2019

View reviewed changes

MaureenHelm merged commit d170e2e into zephyrproject-rtos:topic-gpio Oct 16, 2019

kklimek reviewed Jan 9, 2020

View reviewed changes

erwango deleted the gpio_stm32 branch June 5, 2020 10:13

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[topic-gpio] drivers: gpio_stm32: update to use new GPIO API #19579

[topic-gpio] drivers: gpio_stm32: update to use new GPIO API #19579

erwango commented Oct 3, 2019 •

edited

Loading

erwango commented Oct 3, 2019

erwango commented Oct 3, 2019

erwango commented Oct 3, 2019

galak commented Oct 3, 2019

galak commented Oct 3, 2019

galak commented Oct 3, 2019

zephyrbot commented Oct 3, 2019 •

edited

Loading

erwango commented Oct 3, 2019

pabigot left a comment

mnkp left a comment

erwango commented Oct 7, 2019

pabigot commented Oct 7, 2019

erwango commented Oct 7, 2019

erwango commented Oct 11, 2019 •

edited

Loading

carlescufi commented Oct 15, 2019

erwango commented Oct 16, 2019

kklimek Jan 9, 2020

[topic-gpio] drivers: gpio_stm32: update to use new GPIO API #19579

[topic-gpio] drivers: gpio_stm32: update to use new GPIO API #19579

Conversation

erwango commented Oct 3, 2019 • edited Loading

erwango commented Oct 3, 2019

erwango commented Oct 3, 2019

erwango commented Oct 3, 2019

galak commented Oct 3, 2019

galak commented Oct 3, 2019

galak commented Oct 3, 2019

zephyrbot commented Oct 3, 2019 • edited Loading

erwango commented Oct 3, 2019

pabigot left a comment

Choose a reason for hiding this comment

mnkp left a comment

Choose a reason for hiding this comment

erwango commented Oct 7, 2019

pabigot commented Oct 7, 2019

erwango commented Oct 7, 2019

erwango commented Oct 11, 2019 • edited Loading

Test status:

Test failures on stm32h747_disco_m7:

Test failures on stm32mp157c_dk2:

Summary

carlescufi commented Oct 15, 2019

erwango commented Oct 16, 2019

kklimek Jan 9, 2020

Choose a reason for hiding this comment

erwango commented Oct 3, 2019 •

edited

Loading

zephyrbot commented Oct 3, 2019 •

edited

Loading

erwango commented Oct 11, 2019 •

edited

Loading

Test failures on `stm32h747_disco_m7`:

Test failures on `stm32mp157c_dk2`: