drivers: gpio: add interrupt support for the CH32V family #89139

nzmichaelh · 2025-04-26T18:42:49Z

The WCH External Trigger and Interrupt controller (EXTI) supports
between 8 and 22 lines where each line can trigger an interrupt on
rising edge, falling edge, or both edges. Lines are assigned to a
group, and each group has a separate interrupt. On the CH32V003/6,
there is one group of 8 lines, while on the CH32V208 there are
multiple groups with between one and six lines per group.

In the same way as the STM32 and GD32, define an EXTI driver that
configures the peripheral and an internal interface that can configure
individual lines. Bind into the GPIO driver.

Tested on the linkw and the ch32v006evt using
samples/basic/button. Note that the linkw default stack size is too small for the sample.

nzmichaelh · 2025-05-10T12:44:25Z

@miggazElquez @paulwedeck FYI

drivers/gpio/wch_gpio_ch32v00x.c

Copilot

Pull Request Overview

This PR introduces external interrupt support for the CH32V family by adding a dedicated WCH EXTI driver and integrating it with the GPIO driver, including new DTS nodes and device tree bindings.

Added new header and driver files for the WCH EXTI controller
Updated DTS files and bindings to include EXTI support for multiple CH32V variants
Enabled interrupt configuration in the GPIO driver and updated board DTS files to activate the exti nodes

Reviewed Changes

Copilot reviewed 14 out of 14 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
include/zephyr/drivers/interrupt_controller/wch_exti.h	New header defining interface and types for the WCH EXTI driver
dts/riscv/wch/ch32v208/ch32v208.dtsi, ch32v0/ch32v006.dtsi, ch32v0/ch32v003.dtsi	New DTS node for EXTI configuration on different CH32V variants
dts/bindings/interrupt-controller/wch,exti.yaml	Added binding documentation for the WCH EXTI controller
drivers/interrupt_controller/intc_wch_exti.c	New EXTI controller driver implementation
drivers/gpio/wch_gpio_ch32v00x.c	Updated GPIO driver to support interrupt configuration via EXTI
Kconfig & CMakeLists.txt files	Integrated the new driver with build and configuration systems
boards/wch//.dts	Enabled the EXTI node in board configurations

dts/bindings/interrupt-controller/wch,exti.yaml

nzmichaelh · 2025-06-02T18:40:47Z

Rebased on main.

nzmichaelh · 2025-06-09T09:35:26Z

Rebased to fix an unrelated CI error.

nzmichaelh · 2025-06-09T09:36:13Z

@SoftwareArchitekt could you check my feedback and re-review please?

SoftwareArchitekt · 2025-05-27T09:12:33Z

drivers/gpio/wch_gpio_ch32v00x.c

+	gpio_fire_callbacks(&data->callbacks, dev, BIT(line));
+}
+
+static int gpio_ch32v00x_configure_exti(const struct device *dev, gpio_pin_t pin)


this lines are new added, so could it be done?

SoftwareArchitekt · 2025-05-27T09:13:57Z

drivers/gpio/wch_gpio_ch32v00x.c

@@ -7,6 +7,7 @@
 #include <zephyr/drivers/clock_control.h>
 #include <zephyr/drivers/gpio.h>
 #include <zephyr/drivers/gpio/gpio_utils.h>
+#include <zephyr/drivers/interrupt_controller/wch_exti.h>


yes, the code in c-files are only compiled by device-tree entry. but in header are some variables which takes some memory. pls check it again

SoftwareArchitekt · 2025-05-27T09:14:20Z

drivers/gpio/wch_gpio_ch32v00x.c

+	return 0;
+}
+
+static int gpio_ch32v00x_pin_interrupt_configure(const struct device *dev, gpio_pin_t pin,


Same reply as above.

SoftwareArchitekt · 2025-05-27T09:14:31Z

drivers/gpio/wch_gpio_ch32v00x.c

+			wch_exti_set_trigger(pin, WCH_EXTI_TRIGGER_FALLING_EDGE |
+							  WCH_EXTI_TRIGGER_RISING_EDGE);
+			break;
+		default:


SoftwareArchitekt · 2025-06-10T14:40:07Z

drivers/gpio/wch_gpio_ch32v00x.c

+	const struct device *dev = user;
+	struct gpio_ch32v00x_data *data = dev->data;
+
+	gpio_fire_callbacks(&data->callbacks, dev, BIT(line));


why assigning the data here to local variable (on stack) instead of direct using in gpio_fire_callbacks?

For consistent style. Zephyr drivers typically have an initial variable block where data, config, and regs are resolved as needed. These are then used in the body.

SoftwareArchitekt · 2025-06-10T14:43:03Z

drivers/gpio/wch_gpio_ch32v00x.c

+	case DT_REG_ADDR(DT_NODELABEL(gpioa)):
+		port_id = 0;
+		break;
+#if DT_NODE_EXISTS(DT_NODELABEL(gpiob))


should this also be done on the other banks? I would never think about, that parts of device tree comes so in driver. It could help in future and the defines doesn't take memory

No, this switch is based on the hardware capabilities, not what is enabled (i.e. status = "okay") in Devicetree. All CH32V devices have GPIOA, but at least the CH32V003 does not have GPIOB.

SoftwareArchitekt · 2025-06-10T14:44:31Z

drivers/gpio/wch_gpio_ch32v00x.c

+						 enum gpio_int_mode mode,
+						 enum gpio_int_trig trigger)
+{
+	int err;


could you assign here a 0? I like to have always clear declaration

I understand the intent, but that's not Zephyr style and, with a modern compiler, lets through more bugs.

For example, if you initialise err to zero, and you intended to write:

err = do_something(); if (err < 0) { return err; }

but instead you wrote:

if (do_something() < 0) { return err; }

then the compiler will let this logic bug through. If you have the use of uninitialized variable warning on, then the compiler would warn.

SoftwareArchitekt · 2025-06-10T14:48:14Z

drivers/interrupt_controller/intc_wch_exti.c

+{
+	EXTI_TypeDef *regs = (EXTI_TypeDef *)DT_INST_REG_ADDR(0);
+
+	regs->INTENR |= BIT(line);


where I can find error checking of parameter line?

There's no real Zephyr policy on validating arguments.

petejohanson

A couple items I found from testing this on my custom v305 board. Thanks for working on this!

petejohanson · 2025-06-20T23:38:09Z

drivers/interrupt_controller/intc_wch_exti.c

+	if (callback != NULL && registration->callback != NULL) {
+		return -EALREADY;
+	}


Needs a check similar to the STM32 EXTI one, to just return 0 if a caller attempts to configure with the exact same callback and user:

Suggested change

if (callback != NULL && registration->callback != NULL) {

return -EALREADY;

}

if (registration->callback == callback && registration->user == user) {

return 0;

}

if (callback != NULL && registration->callback != NULL) {

return -EALREADY;

}

Without this, the test_gpio_config_twice_trigger test fails.

petejohanson · 2025-06-21T00:18:13Z

drivers/gpio/wch_gpio_ch32v00x.c

+	BUILD_ASSERT(ARRAY_SIZE(afio->EXTICR) == 4);
+
+	cr_id = pin / 4;
+	bit0 = (pin % 4) << 4;


From testing here, this logic looks slightly off. Need to multiply by 4, not shift:

Suggested change

bit0 = (pin % 4) << 4;

bit0 = (pin % 4) * 4;

With this in place, on top of my work in #91966 I can run the tests/drivers/gpio on my custom v305 board, with the -X gpio_loopback in place, and they all pass:

❯ west twister --device-testing -p wch32v305pad --device-serial /dev/ttyACM0 --west-runner wlink --west-flash -T tests/drivers/gpio -X gpio_loopback Renaming output directory to /home/peter/git/zmk/zephyr/twister-out.7 INFO - Using Ninja.. INFO - Zephyr version: 9996afb64654 INFO - Using 'zephyr' toolchain. INFO - Building initial testsuite list... INFO - Writing JSON report /home/peter/git/zmk/zephyr/twister-out/testplan.json Device testing on: | Platform | ID | Serial device | |-----------------------|------|-----------------| | wch32v305pad/ch32v305 | | /dev/ttyACM0 | INFO - JOBS: 8 INFO - Adding tasks to the queue... INFO - Added initial list of jobs to queue INFO - Total complete: 4/ 4 100% built (not run): 0, filtered: 17, failed: 0, error: 0 INFO - 18 test scenarios (18 configurations) selected, 17 configurations filtered (14 by static filter, 3 at runtime). INFO - 1 of 1 executed test configurations passed (100.00%), 0 built (not run), 0 failed, 0 errored, with no warnings in 42.31 seconds. INFO - 7 of 7 executed test cases passed (100.00%) on 1 out of total 1094 platforms (0.09%). INFO - 1 test configurations executed on platforms, 0 test configurations were only built. Hardware distribution summary: | Board | ID | Counter | Failures | |-----------------------|------|-----------|------------| | wch32v305pad/ch32v305 | | 1 | 0 | INFO - Saving reports... INFO - Writing JSON report /home/peter/git/zmk/zephyr/twister-out/twister.json INFO - Writing xunit report /home/peter/git/zmk/zephyr/twister-out/twister.xml... INFO - Writing xunit report /home/peter/git/zmk/zephyr/twister-out/twister_report.xml... INFO - Run completed

Nice! Done.

nzmichaelh · 2025-06-21T18:27:36Z

Rebased on main.

The WCH External Trigger and Interrupt controller (EXTI) supports between 8 and 22 lines where each line can trigger an interrupt on rising edge, falling edge, or both edges. Lines are assigned to a group, and each group has a separate interrupt. On the CH32V003/6, there is one group of 8 lines, while on the CH32V208 there are multiple groups with between one and six lines per group. In the same way as the STM32 and GD32, define an EXTI driver that configures the peripheral and an internal interface that can configure individual lines. Signed-off-by: Michael Hope <michaelh@juju.nz>

The WCH GPIO peripheral integrates with the EXTI and supports firing interrupts when a GPIO pin changes. Add optional support for firing a callback on rising edge, falling edge, or both edges. Tested on the `linkw` and the `ch32v006evt` using `samples/basic/button`. Signed-off-by: Michael Hope <michaelh@juju.nz>

Now that Zephyr has support, enable by default. Signed-off-by: Michael Hope <michaelh@juju.nz>

sonarqubecloud · 2025-06-21T19:10:42Z

Quality Gate passed

Issues
1 New issue
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

petejohanson

LGTM. Cherry picked the changes onto my v305 work and the full GPIO tests pass.

nzmichaelh · 2025-06-25T17:27:49Z

@VynDragon @nordicjm PTAL. I'd submit this before Friday's freeze if practical.

VynDragon

Took a quick look at the code and tested button sample on linkw that now works. Looks good to me.

nzmichaelh mentioned this pull request Apr 26, 2025

drivers: pinctrl: enable the AFIO clock on the CH32V003/20x/30x #89138

Merged

nzmichaelh force-pushed the ch32v-exti branch 2 times, most recently from bf5e86b to 453af71 Compare May 10, 2025 12:28

nzmichaelh marked this pull request as ready for review May 10, 2025 12:43

nzmichaelh requested review from nordicjm and henrikbrixandersen May 10, 2025 12:43

nzmichaelh self-assigned this May 10, 2025

github-actions bot added area: Interrupt Controller platform: WinChipHead area: GPIO area: RISCV RISCV Architecture (32-bit & 64-bit) labels May 10, 2025

github-actions bot requested review from carlocaione, edersondisouza, fkokosinski, katsuster, kgugala, mgielda, mnkp, npitre, tgorochowik, VynDragon and ycsin May 10, 2025 12:44

nordicjm previously approved these changes May 12, 2025

View reviewed changes

nzmichaelh dismissed nordicjm’s stale review via 7a08a75 May 12, 2025 17:47

nzmichaelh force-pushed the ch32v-exti branch from 453af71 to 7a08a75 Compare May 12, 2025 17:47

VynDragon previously approved these changes May 13, 2025

View reviewed changes

drivers/gpio/wch_gpio_ch32v00x.c Show resolved Hide resolved

kartben requested a review from Copilot May 13, 2025 16:36

Copilot AI reviewed May 13, 2025

View reviewed changes

dts/bindings/interrupt-controller/wch,exti.yaml Outdated Show resolved Hide resolved

nzmichaelh dismissed VynDragon’s stale review via 4c61048 May 13, 2025 17:40

nzmichaelh force-pushed the ch32v-exti branch from 4c61048 to 014e4a9 Compare May 26, 2025 17:20

nzmichaelh dismissed VynDragon’s stale review via 0a5ed6d May 26, 2025 17:26

nzmichaelh force-pushed the ch32v-exti branch from 014e4a9 to 0a5ed6d Compare May 26, 2025 17:26

nzmichaelh requested a review from SoftwareArchitekt May 26, 2025 17:28

nzmichaelh force-pushed the ch32v-exti branch 2 times, most recently from 5c366e1 to 96cc4c5 Compare June 2, 2025 18:40

nzmichaelh force-pushed the ch32v-exti branch from 96cc4c5 to 13ff65d Compare June 9, 2025 09:35

VynDragon previously approved these changes Jun 9, 2025

View reviewed changes

SoftwareArchitekt suggested changes Jun 10, 2025

View reviewed changes

petejohanson suggested changes Jun 21, 2025

View reviewed changes

nzmichaelh force-pushed the ch32v-exti branch from 13ff65d to 3320e49 Compare June 21, 2025 18:22

nzmichaelh added 3 commits June 21, 2025 18:40

boards: enable the WCH EXTI peripheral on all WCH boards

da14af0

Now that Zephyr has support, enable by default. Signed-off-by: Michael Hope <michaelh@juju.nz>

nzmichaelh dismissed VynDragon’s stale review via da14af0 June 21, 2025 18:45

nzmichaelh force-pushed the ch32v-exti branch from 3320e49 to da14af0 Compare June 21, 2025 18:45

SoftwareArchitekt approved these changes Jun 22, 2025

View reviewed changes

petejohanson approved these changes Jun 22, 2025

View reviewed changes

nzmichaelh requested review from VynDragon and nordicjm June 25, 2025 17:27

VynDragon approved these changes Jun 25, 2025

View reviewed changes

nordicjm approved these changes Jun 26, 2025

View reviewed changes

kartben merged commit badce12 into zephyrproject-rtos:main Jun 26, 2025
27 checks passed

SoftwareArchitekt mentioned this pull request Jul 11, 2025

dts: add exti to riscv wch ch32v30x boards #93030

Open

drivers: gpio: add interrupt support for the CH32V family #89139

drivers: gpio: add interrupt support for the CH32V family #89139

Uh oh!

Conversation

nzmichaelh commented Apr 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

nzmichaelh commented May 10, 2025

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

nzmichaelh commented Jun 2, 2025

Uh oh!

nzmichaelh commented Jun 9, 2025

Uh oh!

nzmichaelh commented Jun 9, 2025

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!

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!

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!

petejohanson left a comment

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!

Choose a reason for hiding this comment

Uh oh!

nzmichaelh commented Jun 21, 2025

Uh oh!

sonarqubecloud bot commented Jun 21, 2025

Quality Gate passed

Uh oh!

petejohanson left a comment

Choose a reason for hiding this comment

Uh oh!

nzmichaelh commented Jun 25, 2025

Uh oh!

VynDragon left a comment • 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!

Uh oh!

Uh oh!

nzmichaelh commented Apr 26, 2025 •

edited

Loading

VynDragon left a comment •

edited

Loading