Thread (5 messages) 5 messages, 5 authors, 2024-11-13

Re: [PATCHv3 net-next] net: modernize IRQ resource acquisition

From: Julian Calaby <hidden>
Date: 2024-11-13 09:48:37
Also in: linux-arm-kernel, linux-can, linux-sunxi, lkml, netdev

Hi Rosen,

On Wed, Nov 13, 2024 at 8:14 AM Rosen Penev [off-list ref] wrote:
In probe, np == pdev->dev.of_node. It's easier to pass pdev directly.

Replace irq_of_parse_and_map() by platform_get_irq() to do so. Requires
removing the error message as well as fixing the return type.

Replace of_address_to_resource() with platform_get_resource() for the
same reason.
Sorry for the drive-by review, but I have to question the utility of
this conversion.
quoted hunk ↗ jump to hunk
diff --git a/drivers/net/can/grcan.c b/drivers/net/can/grcan.c
index cdf0ec9fa7f3..48e93c3445e7 100644
--- a/drivers/net/can/grcan.c
+++ b/drivers/net/can/grcan.c
@@ -1673,9 +1673,8 @@ static int grcan_probe(struct platform_device *ofdev)
                goto exit_error;
        }

-       irq = irq_of_parse_and_map(np, GRCAN_IRQIX_IRQ);
-       if (!irq) {
-               dev_err(&ofdev->dev, "no irq found\n");
+       irq = platform_get_irq(ofdev, GRCAN_IRQIX_IRQ);
+       if (irq < 0) {
In this change and a lot of the others, you're not removing the "np"
variable, so you're basically replacing one wrapper with another.
quoted hunk ↗ jump to hunk
                err = -ENODEV;
                goto exit_error;
        }
diff --git a/drivers/net/ethernet/freescale/fs_enet/mac-fcc.c b/drivers/net/ethernet/freescale/fs_enet/mac-fcc.c
index be63293511d9..9006137e3a55 100644
--- a/drivers/net/ethernet/freescale/fs_enet/mac-fcc.c
+++ b/drivers/net/ethernet/freescale/fs_enet/mac-fcc.c
@@ -83,8 +83,8 @@ static int do_pd_setup(struct fs_enet_private *fep)
        struct fs_platform_info *fpi = fep->fpi;
        int ret = -EINVAL;

-       fep->interrupt = irq_of_parse_and_map(ofdev->dev.of_node, 0);
-       if (!fep->interrupt)
+       fep->interrupt = platform_get_irq(ofdev, 0);
+       if (fep->interrupt < 0)
This one and others like it are fine: it's much cleaner to use the
"platform_irq()" function instead of reaching deep into the structure
to grab the "of_node" property.

That said, in this case and probably a few others this is a driver for
an OF device so I'm still not sure this actually makes sense.

Thanks,

-- 
Julian Calaby

Email: julian.calaby@gmail.com
Profile: http://www.google.com/profiles/julian.calaby/
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help