There seems to be a mistake here. There's a num_rx_qs variable that is not
being passed to the allocation function. The mq variant just calls mqs
with the last parameter of the former duplicated to the last parameter
of the latter. That's fine if they match. Not sure they do.

Also avoids manual free_netdev

Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/net/ethernet/freescale/gianfar.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/net/ethernet/freescale/gianfar.c b/drivers/net/ethernet/freescale/gianfar.c
index ecb1703ea150..b0f65cdf4872 100644
--- a/drivers/net/ethernet/freescale/gianfar.c
+++ b/drivers/net/ethernet/freescale/gianfar.c
@@ -475,8 +475,6 @@ static void free_gfar_dev(struct gfar_private *priv)
 			kfree(priv->gfargrp[i].irqinfo[j]);
 			priv->gfargrp[i].irqinfo[j] = NULL;
 		}
-
-	free_netdev(priv->ndev);
 }
 
 static void disable_napi(struct gfar_private *priv)
@@ -681,7 +679,8 @@ static int gfar_of_init(struct platform_device *ofdev, struct net_device **pdev)
 		return -EINVAL;
 	}
 
-	*pdev = alloc_etherdev_mq(sizeof(*priv), num_tx_qs);
+	*pdev = devm_alloc_etherdev_mqs(&ofdev->dev, sizeof(*priv), num_tx_qs,
+					num_rx_qs);
 	dev = *pdev;
 	if (NULL == dev)
 		return -ENOMEM;
-- 
2.46.2

On Tue,  1 Oct 2024 14:21:59 -0700
Rosen Penev <rosenp@gmail.com> wrote:

> There seems to be a mistake here. There's a num_rx_qs variable that is not
> being passed to the allocation function. The mq variant just calls mqs
> with the last parameter of the former duplicated to the last parameter
> of the latter. That's fine if they match. Not sure they do.

As far as I can tell, both queue numbers used during the netdev alloc
are always the same, so this looks good to me.

Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>

Thanks,

Maxime