On Thu, Feb 26, 2026 at 07:10:22PM +0000, Russell King (Oracle) wrote:

On Thu, Feb 26, 2026 at 09:54:07AM +0100, Jan Petrous via B4 Relay wrote:

quoted

From: "Jan Petrous (OSS)" <jan.petrous@oss.nxp.com>

Read IRQ resources for all rx/tx channels, to allow Multi-IRQ mode
for platform glue drivers.

Reviewed-by: Matthias Brugger <mbrugger@suse.com>
Signed-off-by: Jan Petrous (OSS) <jan.petrous@oss.nxp.com>
---
 .../net/ethernet/stmicro/stmmac/stmmac_platform.c  | 46 +++++++++++++++++++++-
 1 file changed, 45 insertions(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
index 5c9fd91a1db9..93bd915ab6eb 100644
--- a/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c
+++ b/drivers/net/ethernet/stmicro/stmmac/stmmac_platform.c

@@ -697,9 +697,40 @@ struct clk *stmmac_pltfr_find_clk(struct plat_stmmacenet_data *plat_dat,
 }
 EXPORT_SYMBOL_GPL(stmmac_pltfr_find_clk);
 
+static int stmmac_pltfr_get_queue_irqs(struct platform_device *pdev,
+				       struct stmmac_resources *stmmac_res,
+				       bool tx)
+{
+	int *irqs = tx ? &stmmac_res->tx_irq[0] : &stmmac_res->rx_irq[0];
+	char name[16];
+	int i;
+
+	/* RX channels irq */
+	STMMAC_FOREACH_MTL_QUEUE(i, MTL_MAX_RX_QUEUES) {

You've missed that there are two separate definitions for tx and rx
queues - while they are currently the same number, code shouldn't
make that assumption.

Oh, yes. My fault. Next time I shall review my code better.

quoted

+		scnprintf(name, sizeof(name), "%cx-queue-%d",
+			  tx ? 't' : 'r', i);

I'm not happy with this method of combining the two loops.

Maybe instead:

static int stmmac_pltfr_get_irq_array(struct platform_device *pdev,
				      const char *fmt, int *irqs,
				      size_t num)
{
	char name[16];
	size_t i;

	for (i = 0; i < num; i++) {
		if (snprintf(name, sizeof(name), fmt, i) >= sizeof(name))
			return -EINVAL;

		irqs[i] = platform_get_irq_byname_optional(pdev, name);
		if (irqs[i] == -EPROBE_DEFER) {
			return irqs[i];
		} else if (irqs[i] <= 0) {
			dev_dbg(&pdev->dev, "IRQ %s not found\n", name);

			irqs[i] = 0;
			break;
		}
	}

	return 0;
}

which has the advantage that it becomes a generic helper for getting an
any array of IRQs.

quoted

 int stmmac_get_platform_resources(struct platform_device *pdev,
 				  struct stmmac_resources *stmmac_res)
 {
+	int ret;
+
 	memset(stmmac_res, 0, sizeof(*stmmac_res));
 
 	/* Get IRQ information early to have an ability to ask for deferred

@@ -735,7 +766,20 @@ int stmmac_get_platform_resources(struct platform_device *pdev,
 
 	stmmac_res->addr = devm_platform_ioremap_resource(pdev, 0);
 
-	return PTR_ERR_OR_ZERO(stmmac_res->addr);
+	if (IS_ERR(stmmac_res->addr))
+		return PTR_ERR(stmmac_res->addr);
+
+	/* TX channels irq */
+	ret = stmmac_pltfr_get_queue_irqs(pdev, stmmac_res, true);
+	if (ret)
+		return ret;
+
+	/* RX channels irq */
+	ret = stmmac_pltfr_get_queue_irqs(pdev, stmmac_res, false);
+	if (ret)
+		return ret;

These then become:

	ret = stmmac_pltfr_get_irq_array(pdev, "tx-queue-%d", 
					 stmmac_res->tx_irq,
					 MTL_MAX_TX_QUEUES);
	if (ret)
		return ret;

	ret = stmmac_pltfr_get_irq_array(pdev, "rx-queue-%d", 
					 stmmac_res->rx_irq,
					 MTL_MAX_RX_QUEUES);
	if (ret)
		return ret;

This has the advantage that one can grep for rx-queue to find it,
and we also use the correct limit for each queue type.

Agree, your code looks better. Applied in v8 :) Thanks.

BTW, I would prefer to use sizeof() instead of constant for array size,
but this is the style used in stmmac, so I reuse the same approach.

BR.
/Jan