From: Chen Wang <unicorn_wang@outlook.com>

The original MSI interrupt type was hard-coded, which was not a good idea.
Now it is changed to read the device tree configuration and then set the
interrupt type.

Signed-off-by: Chen Wang <unicorn_wang@outlook.com>
---
 drivers/irqchip/irq-sg2042-msi.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/drivers/irqchip/irq-sg2042-msi.c b/drivers/irqchip/irq-sg2042-msi.c
index 3b13dbbfdb51..f7cf0dc72eab 100644
--- a/drivers/irqchip/irq-sg2042-msi.c
+++ b/drivers/irqchip/irq-sg2042-msi.c
@@ -30,6 +30,7 @@ struct sg204x_msi_chip_info {
  * @doorbell_addr:	see TRM, 10.1.32, GP_INTR0_SET
  * @irq_first:		First vectors number that MSIs starts
  * @num_irqs:		Number of vectors for MSIs
+ * @irq_type:		IRQ type for MSIs
  * @msi_map:		mapping for allocated MSI vectors.
  * @msi_map_lock:	Lock for msi_map
  * @chip_info:		chip specific infomations
@@ -41,6 +42,7 @@ struct sg204x_msi_chipdata {
 
 	u32					irq_first;
 	u32					num_irqs;
+	unsigned int				irq_type;
 
 	unsigned long				*msi_map;
 	struct mutex				msi_map_lock;
@@ -137,14 +139,14 @@ static int sg204x_msi_parent_domain_alloc(struct irq_domain *domain, unsigned in
 	fwspec.fwnode = domain->parent->fwnode;
 	fwspec.param_count = 2;
 	fwspec.param[0] = data->irq_first + hwirq;
-	fwspec.param[1] = IRQ_TYPE_EDGE_RISING;
+	fwspec.param[1] = data->irq_type;
 
 	ret = irq_domain_alloc_irqs_parent(domain, virq, 1, &fwspec);
 	if (ret)
 		return ret;
 
 	d = irq_domain_get_irq_data(domain->parent, virq);
-	return d->chip->irq_set_type(d, IRQ_TYPE_EDGE_RISING);
+	return d->chip->irq_set_type(d, data->irq_type);
 }
 
 static int sg204x_msi_middle_domain_alloc(struct irq_domain *domain, unsigned int virq,
@@ -298,6 +300,7 @@ static int sg2042_msi_probe(struct platform_device *pdev)
 	}
 
 	data->irq_first = (u32)args.args[0];
+	data->irq_type = (unsigned int)args.args[1];
 	data->num_irqs = (u32)args.args[args.nargs - 1];
 
 	mutex_init(&data->msi_map_lock);
-- 
2.34.1

On Tue, Aug 26 2025 at 09:09, Chen Wang wrote:
> From: Chen Wang <unicorn_wang@outlook.com>
>
> The original MSI interrupt type was hard-coded, which was not a good idea.

That's not really helpful, unless you explain WHY it's not a good
idea...

Also for correctness sake, you want to change the DTs first and not
after you changed the driver to read it from the stale device tree with
the wrong type.

Thanks,

        tglx