On an AMP platform, the system may include two processors:
	- An MCU running an RTOS
	- An MPU running Linux

These processors communicate via the RPMSG protocol.
The driver implements the standard GPIO interface, allowing
the Linux side to control GPIO controllers which reside in
the remote processor via RPMSG protocol.

Cc: Bartosz Golaszewski <brgl@bgdev.pl>
Cc: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
---
 drivers/gpio/Kconfig      |  17 ++
 drivers/gpio/Makefile     |   1 +
 drivers/gpio/gpio-rpmsg.c | 596 ++++++++++++++++++++++++++++++++++++++
 3 files changed, 614 insertions(+)
 create mode 100644 drivers/gpio/gpio-rpmsg.c

diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig
index b45fb799e36c..cff0fda8a283 100644
--- a/drivers/gpio/Kconfig
+++ b/drivers/gpio/Kconfig
@@ -1892,6 +1892,23 @@ config GPIO_SODAVILLE
 
 endmenu
 
+menu "RPMSG GPIO drivers"
+	depends on RPMSG
+
+config GPIO_RPMSG
+	tristate "Generic RPMSG GPIO support"
+	depends on OF && REMOTEPROC
+	select GPIOLIB_IRQCHIP
+	default REMOTEPROC
+	help
+	  Say yes here to support the generic GPIO functions over the RPMSG
+	  bus. Currently supported devices: i.MX7ULP, i.MX8ULP, i.MX8x, and
+	  i.MX9x.
+
+	  If unsure, say N.
+
+endmenu
+
 menu "SPI GPIO expanders"
 	depends on SPI_MASTER
 
diff --git a/drivers/gpio/Makefile b/drivers/gpio/Makefile
index c05f7d795c43..501aba56ad68 100644
--- a/drivers/gpio/Makefile
+++ b/drivers/gpio/Makefile
@@ -158,6 +158,7 @@ obj-$(CONFIG_GPIO_RDC321X)		+= gpio-rdc321x.o
 obj-$(CONFIG_GPIO_REALTEK_OTTO)		+= gpio-realtek-otto.o
 obj-$(CONFIG_GPIO_REG)			+= gpio-reg.o
 obj-$(CONFIG_GPIO_ROCKCHIP)	+= gpio-rockchip.o
+obj-$(CONFIG_GPIO_RPMSG)		+= gpio-rpmsg.o
 obj-$(CONFIG_GPIO_RTD)			+= gpio-rtd.o
 obj-$(CONFIG_ARCH_SA1100)		+= gpio-sa1100.o
 obj-$(CONFIG_GPIO_SAMA5D2_PIOBU)	+= gpio-sama5d2-piobu.o
diff --git a/drivers/gpio/gpio-rpmsg.c b/drivers/gpio/gpio-rpmsg.c
new file mode 100644
index 000000000000..9c609b55bc14
--- /dev/null
+++ b/drivers/gpio/gpio-rpmsg.c
@@ -0,0 +1,596 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright 2026 NXP
+ *
+ * The driver exports a standard gpiochip interface to control
+ * the GPIO controllers via RPMSG on a remote processor.
+ */
+
+#include <linux/completion.h>
+#include <linux/device.h>
+#include <linux/err.h>
+#include <linux/gpio/driver.h>
+#include <linux/init.h>
+#include <linux/irqdomain.h>
+#include <linux/mod_devicetable.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/remoteproc.h>
+#include <linux/rpmsg.h>
+#include <linux/virtio_gpio.h>
+
+#define MAX_PORT_PER_CHANNEL    10
+#define GPIOS_PER_PORT_DEFAULT	32
+#define RPMSG_TIMEOUT		1000
+
+/* GPIO RPMSG Type */
+#define GPIO_RPMSG_SEND		0
+#define GPIO_RPMSG_REPLY	1
+#define GPIO_RPMSG_NOTIFY	2
+
+struct rpmsg_gpio_packet {
+	u8 type;	/* Message type */
+	u8 cmd;		/* Command code */
+	u8 port_idx;
+	u8 line;
+	u8 val1;
+	u8 val2;
+};
+
+struct rpmsg_gpio_line {
+	u8 irq_shutdown;
+	u8 irq_unmask;
+	u8 irq_mask;
+	u32 irq_wake_enable;
+	u32 irq_type;
+	struct rpmsg_gpio_packet msg;
+};
+
+struct rpmsg_gpio_info {
+	struct rpmsg_device *rpdev;
+	struct rpmsg_gpio_packet *reply_msg;
+	struct completion cmd_complete;
+	struct mutex lock;
+	void **port_store;
+};
+
+struct rpmsg_gpio_port {
+	struct gpio_chip gc;
+	struct rpmsg_gpio_line lines[GPIOS_PER_PORT_DEFAULT];
+	struct rpmsg_gpio_info info;
+	u32 ngpios;
+	u32 idx;
+};
+
+struct rpmsg_gpio_fixed_up {
+	int (*send_fixed_up)(struct rpmsg_gpio_info *info, struct rpmsg_gpio_packet *msg);
+	struct rpmsg_gpio_packet *(*recv_fixed_up)(struct rpmsg_device *rpdev, void *data);
+};
+
+/*
+ * @rproc_name: the name of the remote proc.
+ * @recv_pkt: a pointer to the received packet for protocol fix up.
+ * @protocol_fixed_up: optional callbacks to handle protocol mismatches.
+ * @channel_devices: an array of the devices related to the rpdev.
+ */
+struct rpdev_drvdata {
+	const char *rproc_name;
+	void *recv_pkt;
+	struct rpmsg_gpio_fixed_up *protocol_fixed_up;
+	void *channel_devices[MAX_PORT_PER_CHANNEL];
+};
+
+static int rpmsg_gpio_send_message(struct rpmsg_gpio_port *port,
+				   struct rpmsg_gpio_packet *msg,
+				   bool sync)
+{
+	struct rpmsg_gpio_info *info = &port->info;
+	struct rpdev_drvdata *drvdata;
+	int ret;
+
+	drvdata = dev_get_drvdata(&info->rpdev->dev);
+	reinit_completion(&info->cmd_complete);
+
+	if (drvdata->protocol_fixed_up)
+		ret = drvdata->protocol_fixed_up->send_fixed_up(info, msg);
+	else
+		ret = rpmsg_send(info->rpdev->ept, msg, sizeof(*msg));
+
+	if (ret) {
+		dev_err(&info->rpdev->dev, "rpmsg_send failed: %d\n", ret);
+		return ret;
+	}
+
+	if (sync) {
+		ret = wait_for_completion_timeout(&info->cmd_complete,
+						  msecs_to_jiffies(RPMSG_TIMEOUT));
+		if (ret == 0) {
+			dev_err(&info->rpdev->dev, "rpmsg_send timeout!\n");
+			return -ETIMEDOUT;
+		}
+
+		if (info->reply_msg->val1 != 0) {
+			dev_err(&info->rpdev->dev, "remote core replies an error: %d!\n",
+				info->reply_msg->val1);
+			return -EINVAL;
+		}
+
+		/* copy the reply message */
+		memcpy(&port->lines[info->reply_msg->line].msg,
+		       info->reply_msg, sizeof(*info->reply_msg));
+	}
+
+	return 0;
+}
+
+static struct rpmsg_gpio_packet *
+rpmsg_gpio_msg_init_common(struct rpmsg_gpio_port *port, unsigned int line, u8 cmd)
+{
+	struct rpmsg_gpio_packet *msg = &port->lines[line].msg;
+
+	memset(msg, 0, sizeof(struct rpmsg_gpio_packet));
+	msg->type = GPIO_RPMSG_SEND;
+	msg->cmd = cmd;
+	msg->port_idx = port->idx;
+	msg->line = line;
+
+	return msg;
+}
+
+static int rpmsg_gpio_get(struct gpio_chip *gc, unsigned int line)
+{
+	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
+	struct rpmsg_gpio_packet *msg;
+	int ret;
+
+	guard(mutex)(&port->info.lock);
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_GET_VALUE);
+
+	ret = rpmsg_gpio_send_message(port, msg, true);
+	if (!ret)
+		ret = !!port->lines[line].msg.val2;
+
+	return ret;
+}
+
+static int rpmsg_gpio_get_direction(struct gpio_chip *gc, unsigned int line)
+{
+	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
+	struct rpmsg_gpio_packet *msg;
+	int ret;
+
+	guard(mutex)(&port->info.lock);
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_GET_DIRECTION);
+
+	ret = rpmsg_gpio_send_message(port, msg, true);
+	if (ret)
+		return ret;
+
+	switch (port->lines[line].msg.val2) {
+	case VIRTIO_GPIO_DIRECTION_IN:
+		return GPIO_LINE_DIRECTION_IN;
+	case VIRTIO_GPIO_DIRECTION_OUT:
+		return GPIO_LINE_DIRECTION_OUT;
+	default:
+		break;
+	}
+
+	return -EINVAL;
+}
+
+static int rpmsg_gpio_direction_input(struct gpio_chip *gc, unsigned int line)
+{
+	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
+	struct rpmsg_gpio_packet *msg;
+
+	guard(mutex)(&port->info.lock);
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_DIRECTION);
+	msg->val1 = VIRTIO_GPIO_DIRECTION_IN;
+
+	return rpmsg_gpio_send_message(port, msg, true);
+}
+
+static int rpmsg_gpio_set(struct gpio_chip *gc, unsigned int line, int val)
+{
+	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
+	struct rpmsg_gpio_packet *msg;
+
+	guard(mutex)(&port->info.lock);
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_VALUE);
+	msg->val1 = val;
+
+	return rpmsg_gpio_send_message(port, msg, true);
+}
+
+static int rpmsg_gpio_direction_output(struct gpio_chip *gc, unsigned int line, int val)
+{
+	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
+	struct rpmsg_gpio_packet *msg;
+	int ret;
+
+	guard(mutex)(&port->info.lock);
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_DIRECTION);
+	msg->val1 = VIRTIO_GPIO_DIRECTION_OUT;
+
+	ret = rpmsg_gpio_send_message(port, msg, true);
+	if (ret)
+		return ret;
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_VALUE);
+	msg->val1 = val;
+
+	return rpmsg_gpio_send_message(port, msg, true);
+}
+
+static int gpio_rpmsg_irq_set_type(struct irq_data *d, u32 type)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+	u32 line = d->hwirq;
+	int ret = 0;
+
+	switch (type) {
+	case IRQ_TYPE_EDGE_RISING:
+		type = VIRTIO_GPIO_IRQ_TYPE_EDGE_RISING;
+		irq_set_handler_locked(d, handle_simple_irq);
+		break;
+	case IRQ_TYPE_EDGE_FALLING:
+		type = VIRTIO_GPIO_IRQ_TYPE_EDGE_FALLING;
+		irq_set_handler_locked(d, handle_simple_irq);
+		break;
+	case IRQ_TYPE_EDGE_BOTH:
+		type = VIRTIO_GPIO_IRQ_TYPE_EDGE_BOTH;
+		irq_set_handler_locked(d, handle_simple_irq);
+		break;
+	case IRQ_TYPE_LEVEL_LOW:
+		type = VIRTIO_GPIO_IRQ_TYPE_LEVEL_LOW;
+		irq_set_handler_locked(d, handle_level_irq);
+		break;
+	case IRQ_TYPE_LEVEL_HIGH:
+		type = VIRTIO_GPIO_IRQ_TYPE_LEVEL_HIGH;
+		irq_set_handler_locked(d, handle_level_irq);
+		break;
+	default:
+		ret = -EINVAL;
+		irq_set_handler_locked(d, handle_bad_irq);
+		break;
+	}
+
+	port->lines[line].irq_type = type;
+
+	return ret;
+}
+
+static int gpio_rpmsg_irq_set_wake(struct irq_data *d, u32 enable)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+	u32 line = d->hwirq;
+
+	port->lines[line].irq_wake_enable = enable;
+
+	return 0;
+}
+
+/*
+ * This unmask/mask function is invoked in two situations:
+ *   - when an interrupt is being set up, and
+ *   - after an interrupt has occurred.
+ *
+ * The GPIO driver does not access hardware registers directly.
+ * Instead, it caches all relevant information locally, and then sends
+ * the accumulated state to the remote system at this stage.
+ */
+static void gpio_rpmsg_unmask_irq(struct irq_data *d)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+	u32 line = d->hwirq;
+
+	port->lines[line].irq_unmask = 1;
+}
+
+static void gpio_rpmsg_mask_irq(struct irq_data *d)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+	u32 line = d->hwirq;
+
+	/*
+	 * When an interrupt occurs, the remote system masks the interrupt
+	 * and then sends a notification to Linux. After Linux processes
+	 * that notification, it sends an RPMsg command back to the remote
+	 * system to unmask the interrupt again.
+	 */
+	port->lines[line].irq_mask = 1;
+}
+
+static void gpio_rpmsg_irq_shutdown(struct irq_data *d)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+	u32 line = d->hwirq;
+
+	port->lines[line].irq_shutdown = 1;
+}
+
+static void gpio_rpmsg_irq_bus_lock(struct irq_data *d)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+
+	mutex_lock(&port->info.lock);
+}
+
+static void gpio_rpmsg_irq_bus_sync_unlock(struct irq_data *d)
+{
+	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
+	struct rpmsg_gpio_packet *msg;
+	u32 line = d->hwirq;
+
+	/*
+	 * For mask irq, do nothing here.
+	 * The remote system will mask interrupt after an interrupt occurs,
+	 * and then send a notification to Linux system. After Linux system
+	 * handles the notification, it sends an rpmsg back to the remote
+	 * system to unmask this interrupt again.
+	 */
+	if (port->lines[line].irq_mask && !port->lines[line].irq_unmask) {
+		port->lines[line].irq_mask = 0;
+		mutex_unlock(&port->info.lock);
+		return;
+	}
+
+	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_IRQ_TYPE);
+
+	if (port->lines[line].irq_shutdown) {
+		port->lines[line].irq_shutdown = 0;
+		msg->val1 = VIRTIO_GPIO_IRQ_TYPE_NONE;
+		msg->val2 = 0;
+	} else {
+		/* if irq type is not set, use low level trigger as default. */
+		msg->val1 = port->lines[line].irq_type;
+		if (!msg->val1)
+			msg->val1 = VIRTIO_GPIO_IRQ_TYPE_LEVEL_LOW;
+		if (port->lines[line].irq_unmask) {
+			msg->val2 = 0;
+			port->lines[line].irq_unmask = 0;
+		} else /* irq set wake */
+			msg->val2 = port->lines[line].irq_wake_enable;
+	}
+
+	rpmsg_gpio_send_message(port, msg, false);
+	mutex_unlock(&port->info.lock);
+}
+
+static const struct irq_chip gpio_rpmsg_irq_chip = {
+	.irq_mask = gpio_rpmsg_mask_irq,
+	.irq_unmask = gpio_rpmsg_unmask_irq,
+	.irq_set_wake = gpio_rpmsg_irq_set_wake,
+	.irq_set_type = gpio_rpmsg_irq_set_type,
+	.irq_shutdown = gpio_rpmsg_irq_shutdown,
+	.irq_bus_lock = gpio_rpmsg_irq_bus_lock,
+	.irq_bus_sync_unlock = gpio_rpmsg_irq_bus_sync_unlock,
+	.flags = IRQCHIP_IMMUTABLE,
+};
+
+static void rpmsg_gpio_remove_action(void *data)
+{
+	struct rpmsg_gpio_port *port = data;
+
+	port->info.port_store[port->idx] = NULL;
+}
+
+static int rpmsg_gpiochip_register(struct rpmsg_device *rpdev, struct device_node *np)
+{
+	struct rpdev_drvdata *drvdata = dev_get_drvdata(&rpdev->dev);
+	struct rpmsg_gpio_port *port;
+	struct gpio_irq_chip *girq;
+	struct gpio_chip *gc;
+	int ret;
+
+	port = devm_kzalloc(&rpdev->dev, sizeof(*port), GFP_KERNEL);
+	if (!port)
+		return -ENOMEM;
+
+	ret = of_property_read_u32(np, "reg", &port->idx);
+	if (ret)
+		return ret;
+
+	if (port->idx >= MAX_PORT_PER_CHANNEL)
+		return -EINVAL;
+
+	ret = devm_mutex_init(&rpdev->dev, &port->info.lock);
+	if (ret)
+		return ret;
+
+	ret = of_property_read_u32(np, "ngpios", &port->ngpios);
+	if (ret || port->ngpios > GPIOS_PER_PORT_DEFAULT)
+		port->ngpios = GPIOS_PER_PORT_DEFAULT;
+
+	port->info.reply_msg = devm_kzalloc(&rpdev->dev,
+					    sizeof(struct rpmsg_gpio_packet),
+					    GFP_KERNEL);
+	if (!port->info.reply_msg)
+		return -ENOMEM;
+
+	init_completion(&port->info.cmd_complete);
+	port->info.port_store = drvdata->channel_devices;
+	port->info.port_store[port->idx] = port;
+	port->info.rpdev = rpdev;
+
+	gc = &port->gc;
+	gc->owner = THIS_MODULE;
+	gc->parent = &rpdev->dev;
+	gc->fwnode = of_fwnode_handle(np);
+	gc->ngpio = port->ngpios;
+	gc->base = -1;
+	gc->label = devm_kasprintf(&rpdev->dev, GFP_KERNEL, "%s-gpio%d",
+				   drvdata->rproc_name, port->idx);
+
+	gc->direction_input = rpmsg_gpio_direction_input;
+	gc->direction_output = rpmsg_gpio_direction_output;
+	gc->get_direction = rpmsg_gpio_get_direction;
+	gc->get = rpmsg_gpio_get;
+	gc->set = rpmsg_gpio_set;
+
+	girq = &gc->irq;
+	gpio_irq_chip_set_chip(girq, &gpio_rpmsg_irq_chip);
+	girq->parent_handler = NULL;
+	girq->num_parents = 0;
+	girq->parents = NULL;
+	girq->chip->name = devm_kasprintf(&rpdev->dev, GFP_KERNEL, "%s-gpio%d",
+					  drvdata->rproc_name, port->idx);
+
+	ret = devm_add_action_or_reset(&rpdev->dev, rpmsg_gpio_remove_action, port);
+	if (ret)
+		return ret;
+
+	return devm_gpiochip_add_data(&rpdev->dev, gc, port);
+}
+
+static const char *rpmsg_get_rproc_node_name(struct rpmsg_device *rpdev)
+{
+	const char *name = NULL;
+	struct device_node *np;
+	struct rproc *rproc;
+
+	rproc = rproc_get_by_child(&rpdev->dev);
+	if (!rproc)
+		return NULL;
+
+	np = of_node_get(rproc->dev.of_node);
+	if (!np && rproc->dev.parent)
+		np = of_node_get(rproc->dev.parent->of_node);
+
+	if (np) {
+		name = devm_kstrdup(&rpdev->dev, np->name, GFP_KERNEL);
+		of_node_put(np);
+	}
+
+	return name;
+}
+
+static struct device_node *
+rpmsg_get_channel_ofnode(struct rpmsg_device *rpdev, char *chan_name)
+{
+	struct device_node *np_chan = NULL, *np;
+	struct rproc *rproc;
+
+	rproc = rproc_get_by_child(&rpdev->dev);
+	if (!rproc)
+		return NULL;
+
+	np = of_node_get(rproc->dev.of_node);
+	if (!np && rproc->dev.parent)
+		np = of_node_get(rproc->dev.parent->of_node);
+
+	/* The of_node_put() is performed by of_find_node_by_name(). */
+	if (np)
+		np_chan = of_find_node_by_name(np, chan_name);
+
+	return np_chan;
+}
+
+static int rpmsg_gpio_channel_callback(struct rpmsg_device *rpdev, void *data,
+				       int len, void *priv, u32 src)
+{
+	struct rpmsg_gpio_packet *msg = data;
+	struct rpmsg_gpio_port *port = NULL;
+	struct rpdev_drvdata *drvdata;
+
+	drvdata = dev_get_drvdata(&rpdev->dev);
+	if (drvdata && drvdata->protocol_fixed_up)
+		msg = drvdata->protocol_fixed_up->recv_fixed_up(rpdev, data);
+
+	if (!msg || !drvdata)
+		return -EINVAL;
+
+	if (msg->port_idx < MAX_PORT_PER_CHANNEL)
+		port = drvdata->channel_devices[msg->port_idx];
+
+	if (!port || msg->line >= port->ngpios) {
+		dev_err(&rpdev->dev, "wrong port index or line number. port:%d line:%d\n",
+			msg->port_idx, msg->line);
+		return -EINVAL;
+	}
+
+	if (msg->type == GPIO_RPMSG_REPLY) {
+		*port->info.reply_msg = *msg;
+		complete(&port->info.cmd_complete);
+	} else if (msg->type == GPIO_RPMSG_NOTIFY) {
+		generic_handle_domain_irq_safe(port->gc.irq.domain, msg->line);
+	} else {
+		dev_err(&rpdev->dev, "wrong command type (0x%x)\n", msg->type);
+	}
+
+	return 0;
+}
+
+static int rpmsg_gpio_channel_probe(struct rpmsg_device *rpdev)
+{
+	struct device *dev = &rpdev->dev;
+	struct rpdev_drvdata *drvdata;
+	struct device_node *np;
+	int ret = -ENODEV;
+
+	if (!dev->of_node) {
+		np = rpmsg_get_channel_ofnode(rpdev, rpdev->id.name);
+		if (np) {
+			dev->of_node = np;
+			set_primary_fwnode(dev, of_fwnode_handle(np));
+		}
+		return -EPROBE_DEFER;
+	}
+
+	drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL);
+	if (!drvdata)
+		return -ENOMEM;
+
+	drvdata->rproc_name = rpmsg_get_rproc_node_name(rpdev);
+	drvdata->protocol_fixed_up = (struct rpmsg_gpio_fixed_up *)rpdev->id.driver_data;
+	dev_set_drvdata(dev, drvdata);
+
+	for_each_child_of_node_scoped(dev->of_node, child) {
+		if (!of_device_is_available(child))
+			continue;
+
+		if (!of_match_node(dev->driver->of_match_table, child))
+			continue;
+
+		ret = rpmsg_gpiochip_register(rpdev, child);
+		if (ret < 0)
+			break;
+	}
+
+	return ret;
+}
+
+static const struct of_device_id rpmsg_gpio_dt_ids[] = {
+	{ .compatible = "rpmsg-gpio" },
+	{ /* sentinel */ }
+};
+
+static struct rpmsg_device_id rpmsg_gpio_channel_id_table[] = {
+	{ .name = "rpmsg-io" },
+	{ },
+};
+MODULE_DEVICE_TABLE(rpmsg, rpmsg_gpio_channel_id_table);
+
+static struct rpmsg_driver rpmsg_gpio_channel_client = {
+	.callback	= rpmsg_gpio_channel_callback,
+	.id_table	= rpmsg_gpio_channel_id_table,
+	.probe		= rpmsg_gpio_channel_probe,
+	.drv		= {
+		.name	= KBUILD_MODNAME,
+		.of_match_table = rpmsg_gpio_dt_ids,
+	},
+};
+module_rpmsg_driver(rpmsg_gpio_channel_client);
+
+MODULE_AUTHOR("Shenwei Wang <shenwei.wang@nxp.com>");
+MODULE_DESCRIPTION("generic rpmsg gpio driver");
+MODULE_LICENSE("GPL");
-- 
2.43.0

Hello,

On 3/13/26 20:57, Shenwei Wang wrote:
> On an AMP platform, the system may include two processors:
> 	- An MCU running an RTOS
> 	- An MPU running Linux
> 
> These processors communicate via the RPMSG protocol.
> The driver implements the standard GPIO interface, allowing
> the Linux side to control GPIO controllers which reside in
> the remote processor via RPMSG protocol.
> 
> Cc: Bartosz Golaszewski <brgl@bgdev.pl>
> Cc: Andrew Lunn <andrew@lunn.ch>
> Signed-off-by: Shenwei Wang <shenwei.wang@nxp.com>
> ---
>   drivers/gpio/Kconfig      |  17 ++
>   drivers/gpio/Makefile     |   1 +
>   drivers/gpio/gpio-rpmsg.c | 596 ++++++++++++++++++++++++++++++++++++++
>   3 files changed, 614 insertions(+)
>   create mode 100644 drivers/gpio/gpio-rpmsg.c
> 
> diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig
> index b45fb799e36c..cff0fda8a283 100644
> --- a/drivers/gpio/Kconfig
> +++ b/drivers/gpio/Kconfig
> @@ -1892,6 +1892,23 @@ config GPIO_SODAVILLE
>   
>   endmenu
>   
> +menu "RPMSG GPIO drivers"
> +	depends on RPMSG
> +
> +config GPIO_RPMSG
> +	tristate "Generic RPMSG GPIO support"
> +	depends on OF && REMOTEPROC
> +	select GPIOLIB_IRQCHIP
> +	default REMOTEPROC
> +	help
> +	  Say yes here to support the generic GPIO functions over the RPMSG
> +	  bus. Currently supported devices: i.MX7ULP, i.MX8ULP, i.MX8x, and
> +	  i.MX9x.
> +
> +	  If unsure, say N.
> +
> +endmenu
> +
>   menu "SPI GPIO expanders"
>   	depends on SPI_MASTER
>   
> diff --git a/drivers/gpio/Makefile b/drivers/gpio/Makefile
> index c05f7d795c43..501aba56ad68 100644
> --- a/drivers/gpio/Makefile
> +++ b/drivers/gpio/Makefile
> @@ -158,6 +158,7 @@ obj-$(CONFIG_GPIO_RDC321X)		+= gpio-rdc321x.o
>   obj-$(CONFIG_GPIO_REALTEK_OTTO)		+= gpio-realtek-otto.o
>   obj-$(CONFIG_GPIO_REG)			+= gpio-reg.o
>   obj-$(CONFIG_GPIO_ROCKCHIP)	+= gpio-rockchip.o
> +obj-$(CONFIG_GPIO_RPMSG)		+= gpio-rpmsg.o
>   obj-$(CONFIG_GPIO_RTD)			+= gpio-rtd.o
>   obj-$(CONFIG_ARCH_SA1100)		+= gpio-sa1100.o
>   obj-$(CONFIG_GPIO_SAMA5D2_PIOBU)	+= gpio-sama5d2-piobu.o
> diff --git a/drivers/gpio/gpio-rpmsg.c b/drivers/gpio/gpio-rpmsg.c
> new file mode 100644
> index 000000000000..9c609b55bc14
> --- /dev/null
> +++ b/drivers/gpio/gpio-rpmsg.c
> @@ -0,0 +1,596 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright 2026 NXP
> + *
> + * The driver exports a standard gpiochip interface to control
> + * the GPIO controllers via RPMSG on a remote processor.
> + */
> +
> +#include <linux/completion.h>
> +#include <linux/device.h>
> +#include <linux/err.h>
> +#include <linux/gpio/driver.h>
> +#include <linux/init.h>
> +#include <linux/irqdomain.h>
> +#include <linux/mod_devicetable.h>
> +#include <linux/module.h>
> +#include <linux/mutex.h>
> +#include <linux/of.h>
> +#include <linux/of_device.h>
> +#include <linux/of_platform.h>
> +#include <linux/platform_device.h>
> +#include <linux/remoteproc.h>
> +#include <linux/rpmsg.h>
> +#include <linux/virtio_gpio.h>

you still needed to include virtio_gpio.h?

> +
> +#define MAX_PORT_PER_CHANNEL    10
> +#define GPIOS_PER_PORT_DEFAULT	32
> +#define RPMSG_TIMEOUT		1000

I wonder if the timeout is not too high. but
not blocking.

> +
> +/* GPIO RPMSG Type */
> +#define GPIO_RPMSG_SEND		0
> +#define GPIO_RPMSG_REPLY	1
> +#define GPIO_RPMSG_NOTIFY	2
> +
> +struct rpmsg_gpio_packet {
> +	u8 type;	/* Message type */
> +	u8 cmd;		/* Command code */
> +	u8 port_idx;
> +	u8 line;
> +	u8 val1;
> +	u8 val2;
> +};
> +
> +struct rpmsg_gpio_line {
> +	u8 irq_shutdown;
> +	u8 irq_unmask;
> +	u8 irq_mask;
> +	u32 irq_wake_enable;
> +	u32 irq_type;
> +	struct rpmsg_gpio_packet msg;
> +};
> +
> +struct rpmsg_gpio_info {
> +	struct rpmsg_device *rpdev;
> +	struct rpmsg_gpio_packet *reply_msg;
> +	struct completion cmd_complete;
> +	struct mutex lock;
> +	void **port_store;
> +};

Except if I missunderstood Mathieu and Bjorn's request:
"reuse all the design-work done in the gpio-virtio"
We should find similar structures here to those defined
in virtio_gpio.h.
struct rpmsg_gpio_config {
	__le16 ngpio;
	__u8 padding[2];
	__le32 gpio_names_size;
};

/* Virtio GPIO Request / Response */
struct virtio_gpio_request {
	__le16 type;
	__le16 gpio;
	__le32 value;
};

struct rpmsg_gpio_response {
	__u8 status;
	__u8 value;
};

struct rpmsg_gpio_response_get_names {
	__u8 status;
	__u8 value[];
};

/* Virtio GPIO IRQ Request / Response */
struct rpmsg_gpio_irq_request {
	__le16 gpio;
};

struct rpmsg_gpio_irq_response {
	__u8 status;
};

> +
> +struct rpmsg_gpio_port {
> +	struct gpio_chip gc;
> +	struct rpmsg_gpio_line lines[GPIOS_PER_PORT_DEFAULT];
> +	struct rpmsg_gpio_info info;
> +	u32 ngpios;
> +	u32 idx;
> +};
> +
> +struct rpmsg_gpio_fixed_up {
> +	int (*send_fixed_up)(struct rpmsg_gpio_info *info, struct rpmsg_gpio_packet *msg);
> +	struct rpmsg_gpio_packet *(*recv_fixed_up)(struct rpmsg_device *rpdev, void *data);
> +};
> +
> +/*
> + * @rproc_name: the name of the remote proc.
> + * @recv_pkt: a pointer to the received packet for protocol fix up.
> + * @protocol_fixed_up: optional callbacks to handle protocol mismatches.
> + * @channel_devices: an array of the devices related to the rpdev.
> + */
> +struct rpdev_drvdata {
> +	const char *rproc_name;
> +	void *recv_pkt;
> +	struct rpmsg_gpio_fixed_up *protocol_fixed_up;
> +	void *channel_devices[MAX_PORT_PER_CHANNEL];
> +};
> +
> +static int rpmsg_gpio_send_message(struct rpmsg_gpio_port *port,
> +				   struct rpmsg_gpio_packet *msg,
> +				   bool sync)
> +{
> +	struct rpmsg_gpio_info *info = &port->info;
> +	struct rpdev_drvdata *drvdata;
> +	int ret;
> +
> +	drvdata = dev_get_drvdata(&info->rpdev->dev);
> +	reinit_completion(&info->cmd_complete);
> +
> +	if (drvdata->protocol_fixed_up)
> +		ret = drvdata->protocol_fixed_up->send_fixed_up(info, msg);

Seems not part of a generic implementation

> +	else
> +		ret = rpmsg_send(info->rpdev->ept, msg, sizeof(*msg));
> +
> +	if (ret) {
> +		dev_err(&info->rpdev->dev, "rpmsg_send failed: %d\n", ret);
> +		return ret;
> +	}
> +
> +	if (sync) {
> +		ret = wait_for_completion_timeout(&info->cmd_complete,
> +						  msecs_to_jiffies(RPMSG_TIMEOUT));
> +		if (ret == 0) {
> +			dev_err(&info->rpdev->dev, "rpmsg_send timeout!\n");
> +			return -ETIMEDOUT;
> +		}
> +
> +		if (info->reply_msg->val1 != 0) {
> +			dev_err(&info->rpdev->dev, "remote core replies an error: %d!\n",
> +				info->reply_msg->val1);
> +			return -EINVAL;
> +		}
> +
> +		/* copy the reply message */
> +		memcpy(&port->lines[info->reply_msg->line].msg,
> +		       info->reply_msg, sizeof(*info->reply_msg));
> +	}
> +
> +	return 0;
> +}
> +
> +static struct rpmsg_gpio_packet *
> +rpmsg_gpio_msg_init_common(struct rpmsg_gpio_port *port, unsigned int line, u8 cmd)
> +{
> +	struct rpmsg_gpio_packet *msg = &port->lines[line].msg;
> +
> +	memset(msg, 0, sizeof(struct rpmsg_gpio_packet));
> +	msg->type = GPIO_RPMSG_SEND;
> +	msg->cmd = cmd;
> +	msg->port_idx = port->idx;
> +	msg->line = line;
> +
> +	return msg;
> +}
> +
> +static int rpmsg_gpio_get(struct gpio_chip *gc, unsigned int line)
> +{
> +	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
> +	struct rpmsg_gpio_packet *msg;
> +	int ret;
> +
> +	guard(mutex)(&port->info.lock);
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_GET_VALUE);
> +
> +	ret = rpmsg_gpio_send_message(port, msg, true);
> +	if (!ret)
> +		ret = !!port->lines[line].msg.val2;
> +
> +	return ret;
> +}
> +
> +static int rpmsg_gpio_get_direction(struct gpio_chip *gc, unsigned int line)
> +{
> +	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
> +	struct rpmsg_gpio_packet *msg;
> +	int ret;
> +
> +	guard(mutex)(&port->info.lock);
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_GET_DIRECTION);
> +
> +	ret = rpmsg_gpio_send_message(port, msg, true);
> +	if (ret)
> +		return ret;
> +
> +	switch (port->lines[line].msg.val2) {
> +	case VIRTIO_GPIO_DIRECTION_IN:
> +		return GPIO_LINE_DIRECTION_IN;
> +	case VIRTIO_GPIO_DIRECTION_OUT:
> +		return GPIO_LINE_DIRECTION_OUT;
> +	default:
> +		break;
> +	}
> +
> +	return -EINVAL;
> +}
> +
> +static int rpmsg_gpio_direction_input(struct gpio_chip *gc, unsigned int line)
> +{
> +	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
> +	struct rpmsg_gpio_packet *msg;
> +
> +	guard(mutex)(&port->info.lock);
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_DIRECTION);
> +	msg->val1 = VIRTIO_GPIO_DIRECTION_IN;
> +
> +	return rpmsg_gpio_send_message(port, msg, true);
> +}
> +
> +static int rpmsg_gpio_set(struct gpio_chip *gc, unsigned int line, int val)
> +{
> +	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
> +	struct rpmsg_gpio_packet *msg;
> +
> +	guard(mutex)(&port->info.lock);
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_VALUE);
> +	msg->val1 = val;
> +
> +	return rpmsg_gpio_send_message(port, msg, true);
> +}
> +
> +static int rpmsg_gpio_direction_output(struct gpio_chip *gc, unsigned int line, int val)
> +{
> +	struct rpmsg_gpio_port *port = gpiochip_get_data(gc);
> +	struct rpmsg_gpio_packet *msg;
> +	int ret;
> +
> +	guard(mutex)(&port->info.lock);
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_DIRECTION);
> +	msg->val1 = VIRTIO_GPIO_DIRECTION_OUT;
> +
> +	ret = rpmsg_gpio_send_message(port, msg, true);
> +	if (ret)
> +		return ret;
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_SET_VALUE);
> +	msg->val1 = val;
> +
> +	return rpmsg_gpio_send_message(port, msg, true);
> +}
> +
> +static int gpio_rpmsg_irq_set_type(struct irq_data *d, u32 type)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +	u32 line = d->hwirq;
> +	int ret = 0;
> +
> +	switch (type) {
> +	case IRQ_TYPE_EDGE_RISING:
> +		type = VIRTIO_GPIO_IRQ_TYPE_EDGE_RISING;
> +		irq_set_handler_locked(d, handle_simple_irq);
> +		break;
> +	case IRQ_TYPE_EDGE_FALLING:
> +		type = VIRTIO_GPIO_IRQ_TYPE_EDGE_FALLING;
> +		irq_set_handler_locked(d, handle_simple_irq);
> +		break;
> +	case IRQ_TYPE_EDGE_BOTH:
> +		type = VIRTIO_GPIO_IRQ_TYPE_EDGE_BOTH;
> +		irq_set_handler_locked(d, handle_simple_irq);
> +		break;
> +	case IRQ_TYPE_LEVEL_LOW:
> +		type = VIRTIO_GPIO_IRQ_TYPE_LEVEL_LOW;
> +		irq_set_handler_locked(d, handle_level_irq);
> +		break;
> +	case IRQ_TYPE_LEVEL_HIGH:
> +		type = VIRTIO_GPIO_IRQ_TYPE_LEVEL_HIGH;
> +		irq_set_handler_locked(d, handle_level_irq);
> +		break;
> +	default:
> +		ret = -EINVAL;
> +		irq_set_handler_locked(d, handle_bad_irq);
> +		break;
> +	}
> +
> +	port->lines[line].irq_type = type;
> +
> +	return ret;
> +}
> +
> +static int gpio_rpmsg_irq_set_wake(struct irq_data *d, u32 enable)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +	u32 line = d->hwirq;
> +
> +	port->lines[line].irq_wake_enable = enable;
> +
> +	return 0;
> +}
> +
> +/*
> + * This unmask/mask function is invoked in two situations:
> + *   - when an interrupt is being set up, and
> + *   - after an interrupt has occurred.
> + *
> + * The GPIO driver does not access hardware registers directly.
> + * Instead, it caches all relevant information locally, and then sends
> + * the accumulated state to the remote system at this stage.
> + */
> +static void gpio_rpmsg_unmask_irq(struct irq_data *d)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +	u32 line = d->hwirq;
> +
> +	port->lines[line].irq_unmask = 1;
> +}
> +
> +static void gpio_rpmsg_mask_irq(struct irq_data *d)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +	u32 line = d->hwirq;
> +
> +	/*
> +	 * When an interrupt occurs, the remote system masks the interrupt
> +	 * and then sends a notification to Linux. After Linux processes
> +	 * that notification, it sends an RPMsg command back to the remote
> +	 * system to unmask the interrupt again.
> +	 */
> +	port->lines[line].irq_mask = 1;
> +}
> +
> +static void gpio_rpmsg_irq_shutdown(struct irq_data *d)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +	u32 line = d->hwirq;
> +
> +	port->lines[line].irq_shutdown = 1;
> +}
> +
> +static void gpio_rpmsg_irq_bus_lock(struct irq_data *d)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +
> +	mutex_lock(&port->info.lock);
> +}
> +
> +static void gpio_rpmsg_irq_bus_sync_unlock(struct irq_data *d)
> +{
> +	struct rpmsg_gpio_port *port = irq_data_get_irq_chip_data(d);
> +	struct rpmsg_gpio_packet *msg;
> +	u32 line = d->hwirq;
> +
> +	/*
> +	 * For mask irq, do nothing here.
> +	 * The remote system will mask interrupt after an interrupt occurs,
> +	 * and then send a notification to Linux system. After Linux system
> +	 * handles the notification, it sends an rpmsg back to the remote
> +	 * system to unmask this interrupt again.
> +	 */
> +	if (port->lines[line].irq_mask && !port->lines[line].irq_unmask) {
> +		port->lines[line].irq_mask = 0;
> +		mutex_unlock(&port->info.lock);
> +		return;
> +	}
> +
> +	msg = rpmsg_gpio_msg_init_common(port, line, VIRTIO_GPIO_MSG_IRQ_TYPE);
> +
> +	if (port->lines[line].irq_shutdown) {
> +		port->lines[line].irq_shutdown = 0;
> +		msg->val1 = VIRTIO_GPIO_IRQ_TYPE_NONE;
> +		msg->val2 = 0;
> +	} else {
> +		/* if irq type is not set, use low level trigger as default. */
> +		msg->val1 = port->lines[line].irq_type;
> +		if (!msg->val1)
> +			msg->val1 = VIRTIO_GPIO_IRQ_TYPE_LEVEL_LOW;
> +		if (port->lines[line].irq_unmask) {
> +			msg->val2 = 0;
> +			port->lines[line].irq_unmask = 0;
> +		} else /* irq set wake */
> +			msg->val2 = port->lines[line].irq_wake_enable;
> +	}
> +
> +	rpmsg_gpio_send_message(port, msg, false);
> +	mutex_unlock(&port->info.lock);
> +}
> +
> +static const struct irq_chip gpio_rpmsg_irq_chip = {
> +	.irq_mask = gpio_rpmsg_mask_irq,
> +	.irq_unmask = gpio_rpmsg_unmask_irq,
> +	.irq_set_wake = gpio_rpmsg_irq_set_wake,
> +	.irq_set_type = gpio_rpmsg_irq_set_type,
> +	.irq_shutdown = gpio_rpmsg_irq_shutdown,
> +	.irq_bus_lock = gpio_rpmsg_irq_bus_lock,
> +	.irq_bus_sync_unlock = gpio_rpmsg_irq_bus_sync_unlock,
> +	.flags = IRQCHIP_IMMUTABLE,
> +};
> +
> +static void rpmsg_gpio_remove_action(void *data)
> +{
> +	struct rpmsg_gpio_port *port = data;
> +
> +	port->info.port_store[port->idx] = NULL;
> +}
> +
> +static int rpmsg_gpiochip_register(struct rpmsg_device *rpdev, struct device_node *np)
> +{
> +	struct rpdev_drvdata *drvdata = dev_get_drvdata(&rpdev->dev);
> +	struct rpmsg_gpio_port *port;
> +	struct gpio_irq_chip *girq;
> +	struct gpio_chip *gc;
> +	int ret;
> +
> +	port = devm_kzalloc(&rpdev->dev, sizeof(*port), GFP_KERNEL);
> +	if (!port)
> +		return -ENOMEM;
> +
> +	ret = of_property_read_u32(np, "reg", &port->idx);
> +	if (ret)
> +		return ret;
> +
> +	if (port->idx >= MAX_PORT_PER_CHANNEL)
> +		return -EINVAL;
> +
> +	ret = devm_mutex_init(&rpdev->dev, &port->info.lock);
> +	if (ret)
> +		return ret;
> +
> +	ret = of_property_read_u32(np, "ngpios", &port->ngpios);

The number of GPIOs should be obtained from the remote side, as done in 
virtio_gpio. In virtio_gpio, this is retrieved via a get_config 
operation. Here, you could implement a specific RPMsg to retrieve the 
remote topology.

> +	if (ret || port->ngpios > GPIOS_PER_PORT_DEFAULT)
> +		port->ngpios = GPIOS_PER_PORT_DEFAULT;
> +
> +	port->info.reply_msg = devm_kzalloc(&rpdev->dev,
> +					    sizeof(struct rpmsg_gpio_packet),
> +					    GFP_KERNEL);
> +	if (!port->info.reply_msg)
> +		return -ENOMEM;
> +
> +	init_completion(&port->info.cmd_complete);
> +	port->info.port_store = drvdata->channel_devices;
> +	port->info.port_store[port->idx] = port;
> +	port->info.rpdev = rpdev;
> +
> +	gc = &port->gc;
> +	gc->owner = THIS_MODULE;
> +	gc->parent = &rpdev->dev;
> +	gc->fwnode = of_fwnode_handle(np);
> +	gc->ngpio = port->ngpios;
> +	gc->base = -1;
> +	gc->label = devm_kasprintf(&rpdev->dev, GFP_KERNEL, "%s-gpio%d",
> +				   drvdata->rproc_name, port->idx);
> +
> +	gc->direction_input = rpmsg_gpio_direction_input;
> +	gc->direction_output = rpmsg_gpio_direction_output;
> +	gc->get_direction = rpmsg_gpio_get_direction;
> +	gc->get = rpmsg_gpio_get;
> +	gc->set = rpmsg_gpio_set;
> +
> +	girq = &gc->irq;
> +	gpio_irq_chip_set_chip(girq, &gpio_rpmsg_irq_chip);
> +	girq->parent_handler = NULL;
> +	girq->num_parents = 0;
> +	girq->parents = NULL;
> +	girq->chip->name = devm_kasprintf(&rpdev->dev, GFP_KERNEL, "%s-gpio%d",
> +					  drvdata->rproc_name, port->idx);
> +
> +	ret = devm_add_action_or_reset(&rpdev->dev, rpmsg_gpio_remove_action, port);
> +	if (ret)
> +		return ret;
> +
> +	return devm_gpiochip_add_data(&rpdev->dev, gc, port);
> +}
> +
> +static const char *rpmsg_get_rproc_node_name(struct rpmsg_device *rpdev)
> +{
> +	const char *name = NULL;
> +	struct device_node *np;
> +	struct rproc *rproc;
> +
> +	rproc = rproc_get_by_child(&rpdev->dev);
> +	if (!rproc)
> +		return NULL;
> +
> +	np = of_node_get(rproc->dev.of_node);
> +	if (!np && rproc->dev.parent)
> +		np = of_node_get(rproc->dev.parent->of_node);
> +
> +	if (np) {
> +		name = devm_kstrdup(&rpdev->dev, np->name, GFP_KERNEL);
> +		of_node_put(np);
> +	}
> +
> +	return name;
> +}
> +
> +static struct device_node *
> +rpmsg_get_channel_ofnode(struct rpmsg_device *rpdev, char *chan_name)
> +{
> +	struct device_node *np_chan = NULL, *np;
> +	struct rproc *rproc;
> +
> +	rproc = rproc_get_by_child(&rpdev->dev);
> +	if (!rproc)
> +		return NULL;
> +
> +	np = of_node_get(rproc->dev.of_node);
> +	if (!np && rproc->dev.parent)
> +		np = of_node_get(rproc->dev.parent->of_node);
> +
> +	/* The of_node_put() is performed by of_find_node_by_name(). */
> +	if (np)
> +		np_chan = of_find_node_by_name(np, chan_name);
> +
> +	return np_chan;
> +}
> +
> +static int rpmsg_gpio_channel_callback(struct rpmsg_device *rpdev, void *data,
> +				       int len, void *priv, u32 src)
> +{
> +	struct rpmsg_gpio_packet *msg = data;
> +	struct rpmsg_gpio_port *port = NULL;
> +	struct rpdev_drvdata *drvdata;
> +
> +	drvdata = dev_get_drvdata(&rpdev->dev);
> +	if (drvdata && drvdata->protocol_fixed_up)
> +		msg = drvdata->protocol_fixed_up->recv_fixed_up(rpdev, data);
> +
> +	if (!msg || !drvdata)
> +		return -EINVAL;
> +
> +	if (msg->port_idx < MAX_PORT_PER_CHANNEL)
> +		port = drvdata->channel_devices[msg->port_idx];
> +
> +	if (!port || msg->line >= port->ngpios) {
> +		dev_err(&rpdev->dev, "wrong port index or line number. port:%d line:%d\n",
> +			msg->port_idx, msg->line);
> +		return -EINVAL;
> +	}
> +
> +	if (msg->type == GPIO_RPMSG_REPLY) {
> +		*port->info.reply_msg = *msg;
> +		complete(&port->info.cmd_complete);
> +	} else if (msg->type == GPIO_RPMSG_NOTIFY) {
> +		generic_handle_domain_irq_safe(port->gc.irq.domain, msg->line);
> +	} else {
> +		dev_err(&rpdev->dev, "wrong command type (0x%x)\n", msg->type);
> +	}
> +
> +	return 0;
> +}
> +
> +static int rpmsg_gpio_channel_probe(struct rpmsg_device *rpdev)
> +{
> +	struct device *dev = &rpdev->dev;
> +	struct rpdev_drvdata *drvdata;
> +	struct device_node *np;
> +	int ret = -ENODEV;
> +
> +	if (!dev->of_node) {
> +		np = rpmsg_get_channel_ofnode(rpdev, rpdev->id.name);
> +		if (np) {
> +			dev->of_node = np;
> +			set_primary_fwnode(dev, of_fwnode_handle(np));
> +		}
> +		return -EPROBE_DEFER;

Here is it a bug ? else you should explain in a comment why you perform 
some actions when np != 0 but return -EPROBE_DEFER

Regards,
Arnaud

> +	}
> +
> +	drvdata = devm_kzalloc(dev, sizeof(*drvdata), GFP_KERNEL);
> +	if (!drvdata)
> +		return -ENOMEM;
> +
> +	drvdata->rproc_name = rpmsg_get_rproc_node_name(rpdev);
> +	drvdata->protocol_fixed_up = (struct rpmsg_gpio_fixed_up *)rpdev->id.driver_data;
> +	dev_set_drvdata(dev, drvdata);
> +
> +	for_each_child_of_node_scoped(dev->of_node, child) {
> +		if (!of_device_is_available(child))
> +			continue;
> +
> +		if (!of_match_node(dev->driver->of_match_table, child))
> +			continue;
> +
> +		ret = rpmsg_gpiochip_register(rpdev, child);
> +		if (ret < 0)
> +			break;
> +	}
> +
> +	return ret;
> +}
> +
> +static const struct of_device_id rpmsg_gpio_dt_ids[] = {
> +	{ .compatible = "rpmsg-gpio" },
> +	{ /* sentinel */ }
> +};
> +
> +static struct rpmsg_device_id rpmsg_gpio_channel_id_table[] = {
> +	{ .name = "rpmsg-io" },
> +	{ },
> +};
> +MODULE_DEVICE_TABLE(rpmsg, rpmsg_gpio_channel_id_table);
> +
> +static struct rpmsg_driver rpmsg_gpio_channel_client = {
> +	.callback	= rpmsg_gpio_channel_callback,
> +	.id_table	= rpmsg_gpio_channel_id_table,
> +	.probe		= rpmsg_gpio_channel_probe,
> +	.drv		= {
> +		.name	= KBUILD_MODNAME,
> +		.of_match_table = rpmsg_gpio_dt_ids,
> +	},
> +};
> +module_rpmsg_driver(rpmsg_gpio_channel_client);
> +
> +MODULE_AUTHOR("Shenwei Wang <shenwei.wang@nxp.com>");
> +MODULE_DESCRIPTION("generic rpmsg gpio driver");
> +MODULE_LICENSE("GPL");

> > +static int rpmsg_gpio_send_message(struct rpmsg_gpio_port *port,
> > +				   struct rpmsg_gpio_packet *msg,
> > +				   bool sync)
> > +{
> > +	struct rpmsg_gpio_info *info = &port->info;
> > +	struct rpdev_drvdata *drvdata;
> > +	int ret;
> > +
> > +	drvdata = dev_get_drvdata(&info->rpdev->dev);
> > +	reinit_completion(&info->cmd_complete);
> > +
> > +	if (drvdata->protocol_fixed_up)
> > +		ret = drvdata->protocol_fixed_up->send_fixed_up(info, msg);
> 
> Seems not part of a generic implementation

Agreed. Lets have a clean generic implementation first, and then
patches on top for legacy stuff.

> > +	ret = of_property_read_u32(np, "ngpios", &port->ngpios);
> 
> The number of GPIOs should be obtained from the remote side, as done in
> virtio_gpio. In virtio_gpio, this is retrieved via a get_config operation.
> Here, you could implement a specific RPMsg to retrieve the remote topology.

The DT property ngpios is pretty standard, so i think it makes a lot
of sense to have. But i also agree that asking the other side makes
sense. What is being implemented here with rpmsg is not that different
to greybus, and the greybus GPIO driver also ask the other side how
many GPIO lines it has. So yes, it should be part of the protocol, but
maybe optional?

	  Andrew

> > +struct rpmsg_gpio_info {
> > +	struct rpmsg_device *rpdev;
> > +	struct rpmsg_gpio_packet *reply_msg;
> > +	struct completion cmd_complete;
> > +	struct mutex lock;
> > +	void **port_store;
> > +};
> 
> Except if I missunderstood Mathieu and Bjorn's request:
> "reuse all the design-work done in the gpio-virtio"
> We should find similar structures here to those defined
> in virtio_gpio.h.
> struct rpmsg_gpio_config {
> 	__le16 ngpio;
> 	__u8 padding[2];
> 	__le32 gpio_names_size;
> };
> 
> /* Virtio GPIO Request / Response */
> struct virtio_gpio_request {
> 	__le16 type;
> 	__le16 gpio;
> 	__le32 value;
> };

The core of the issue is that Shenwei is stone walling any change
which makes it hard to keep the legacy firmware. It is possible to use
these structures, but it makes the extra code Shenwei needs to
translate this protocol to the legacy protocol more difficult. It
might need to keep state, etc. 

Two points...

The firmware implements more than GPIO. There is definitely I2C as
well, the first version of the patch has bits of I2C code. Looking at:

https://lwn.net/ml/all/20250922200413.309707-3-shenwei.wang@nxp.com/

There is also RTC, and a few other things which don't directly map to
Linux subsystems, but maybe do have Linux drivers?

Give how much pushback there has been on the existing protocol for
GPIO, it would be wise to assume that I2C, and RTC is going to get the
same amount of pushback. If any of these three, GPIO, I2C, or RTC
decide that only a new, clean protocol will be accepted, no legacy
shims, the firmware has to change, breaking compatibility to legacy
protocols, and the accepted shims become pointless Maintenance burden.

Point two is that the customers who are pushing for these drivers to
be added to Mainline probably know that nearly nothing gets into
Mainline without some changes. There is some short term pain to
swapping to Mainline because of these changes, in this case, firmware
upgrades. But in the long run, it is worth the pain to be able to use
Mainline. And those customers who don't want to upgrade the firmware
can keep with the out of tree drives.

So, what are our choices?

1) We accept the code as it is now, with the shim?

2) We keep pushing for the virtio protocol, with the shim?

3) We keep pushing for the virtio protocol, no shim, firmware changes

4) We pause GPIO where it is today, and restart all the arguments with
   the I2C driver. We can come back to the GPIO driver in a few months
   time once we have a better idea how I2C is going. And maybe we also
   need to see the watchdog driver, and argue about its protocol.

I also understand ST has a generic I2C driver nearly ready, if that
gets merged first, that probably kills the NXP I2C protocol, and maybe
the NXP GPIO and RTC protocols.

My vote is for 3. If not 3, then 4.

     Andrew

On Tue, 17 Mar 2026 at 08:11, Andrew Lunn <andrew@lunn.ch> wrote:
>
> > > +struct rpmsg_gpio_info {
> > > +   struct rpmsg_device *rpdev;
> > > +   struct rpmsg_gpio_packet *reply_msg;
> > > +   struct completion cmd_complete;
> > > +   struct mutex lock;
> > > +   void **port_store;
> > > +};
> >
> > Except if I missunderstood Mathieu and Bjorn's request:
> > "reuse all the design-work done in the gpio-virtio"
> > We should find similar structures here to those defined
> > in virtio_gpio.h.
> > struct rpmsg_gpio_config {
> >       __le16 ngpio;
> >       __u8 padding[2];
> >       __le32 gpio_names_size;
> > };
> >
> > /* Virtio GPIO Request / Response */
> > struct virtio_gpio_request {
> >       __le16 type;
> >       __le16 gpio;
> >       __le32 value;
> > };
>
> The core of the issue is that Shenwei is stone walling any change
> which makes it hard to keep the legacy firmware. It is possible to use
> these structures, but it makes the extra code Shenwei needs to
> translate this protocol to the legacy protocol more difficult. It
> might need to keep state, etc.
>

I agree with everything Andrew points out above.

> Two points...
>
> The firmware implements more than GPIO. There is definitely I2C as
> well, the first version of the patch has bits of I2C code. Looking at:
>
> https://lwn.net/ml/all/20250922200413.309707-3-shenwei.wang@nxp.com/
>
> There is also RTC, and a few other things which don't directly map to
> Linux subsystems, but maybe do have Linux drivers?
>
> Give how much pushback there has been on the existing protocol for
> GPIO, it would be wise to assume that I2C, and RTC is going to get the
> same amount of pushback. If any of these three, GPIO, I2C, or RTC
> decide that only a new, clean protocol will be accepted, no legacy
> shims, the firmware has to change, breaking compatibility to legacy
> protocols, and the accepted shims become pointless Maintenance burden.
>

I have made this point clear before: modeling legacy protocols in
mainline doesn't scale.  Mainline uses a single generic protocol, and
yes, it means breaking legacy protocols.  This is the cost of moving
to a mainline kernel.  If people want to use the legacy firmware, they
must stick with a legacy kernel.

> Point two is that the customers who are pushing for these drivers to
> be added to Mainline probably know that nearly nothing gets into
> Mainline without some changes. There is some short term pain to
> swapping to Mainline because of these changes, in this case, firmware
> upgrades. But in the long run, it is worth the pain to be able to use
> Mainline. And those customers who don't want to upgrade the firmware
> can keep with the out of tree drives.
>
> So, what are our choices?
>
> 1) We accept the code as it is now, with the shim?
>

NAK

> 2) We keep pushing for the virtio protocol, with the shim?
>

NAK

> 3) We keep pushing for the virtio protocol, no shim, firmware changes
>

Nothing will get merged in the RPMSG subsystem that includes support
for the legacy protocol.  Not today, not in a month, not in 5 years.

> 4) We pause GPIO where it is today, and restart all the arguments with
>    the I2C driver. We can come back to the GPIO driver in a few months
>    time once we have a better idea how I2C is going. And maybe we also
>    need to see the watchdog driver, and argue about its protocol.
>
> I also understand ST has a generic I2C driver nearly ready, if that
> gets merged first, that probably kills the NXP I2C protocol, and maybe
> the NXP GPIO and RTC protocols.
>
> My vote is for 3. If not 3, then 4.
>

Strong vote for 3.

>      Andrew
>

On Wed, Mar 18, 2026 at 5:03 PM Mathieu Poirier
<mathieu.poirier@linaro.org> wrote:

> I have made this point clear before: modeling legacy protocols in
> mainline doesn't scale.  Mainline uses a single generic protocol, and
> yes, it means breaking legacy protocols.  This is the cost of moving
> to a mainline kernel.  If people want to use the legacy firmware, they
> must stick with a legacy kernel.

I mostly agree with this stance.

But it is under the assumption that the contributor is coming from the
same legal body that can define and change the firmware in question.

For example: the mainline Linux kernel supports a whole slew of
funky Apple rpmsg-like protocols. c.f. drivers/soc/apple/rtkit.c

We cannot go and tell the Asahi contributors to change the Apple
firmware to use rpmsg like everyone else, because they are not Apple,
they just want to run Linux on someone else's hardware.

In this case, the contributor is coming from the same legal body as
the one doing the firmware. I know and sympathize with the fact
that sometimes working inside a company to make changes happen
can be as hard as working on the outside, and internal structures
can be as resistant to pressure change as Microsoft, or Apple.
Additionally they hit the contributor on the head with "just get this
done, now, fast".

So it is pretty important in this situation that it is NXP that we address.
The contributor is just a representative of that legal body
in this case.

If someone *outside* of NXP, say an OpenWrt hobbyist contributor
was pushing the same patches based on code drops and reverse
engineering, the response would be *different*.

In a way the situation is a bit icky. As the Linux community we often
see all contributions as personal, individual. And the discussion here
is that of standards committee, which we are not.

So if we are addressing NXP, then we need to be explicit about that,
and careful not to put their representative in the crosshairs. It's unfair,
he's just the messenger.

Yours,
Linus Walleij