[PATCH net] net: fec_mpc52xx: Fix resource size format

[PATCH net] net: fec_mpc52xx: Fix resource size format

[Rosen Penev]

resource_size() is printed after casting to unsigned long, but
sizeof() returns size_t.

Use %zx for the sizeof() value so the format string matches the
argument type so that 64-bit builds compile.

Assisted-by: Codex:GPT-5.5
Fixes: 31b7720c8259 ("fec: Convert printks to netdev_<level>")
Signed-off-by: Rosen Penev <rosenp@gmail.com>
---
 drivers/net/ethernet/freescale/fec_mpc52xx.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/net/ethernet/freescale/fec_mpc52xx.c b/drivers/net/ethernet/freescale/fec_mpc52xx.c
index 556f646073af..a770e7edb1dc 100644
--- a/drivers/net/ethernet/freescale/fec_mpc52xx.c
+++ b/drivers/net/ethernet/freescale/fec_mpc52xx.c
@@ -834,7 +834,7 @@ static int mpc52xx_fec_probe(struct platform_device *op)
 		goto err_netdev;
 	}
 	if (resource_size(&mem) < sizeof(struct mpc52xx_fec)) {
-		pr_err("invalid resource size (%lx < %x), check mpc52xx_devices.c\n",
+		pr_err("invalid resource size (%lx < %zx), check mpc52xx_devices.c\n",
 		       (unsigned long)resource_size(&mem),
 		       sizeof(struct mpc52xx_fec));
 		rv = -EINVAL;
-- 
2.54.0

Re: [PATCH net] net: fec_mpc52xx: Fix resource size format

[Joe Perches]

On Sun, 2026-05-17 at 21:40 -0700, Rosen Penev wrote:
> resource_size() is printed after casting to unsigned long, but
> sizeof() returns size_t.

trivia and I wonder if this code path is even possible

> diff --git a/drivers/net/ethernet/freescale/fec_mpc52xx.c b/drivers/net/ethernet/freescale/fec_mpc52xx.c
[]
> @@ -834,7 +834,7 @@ static int mpc52xx_fec_probe(struct platform_device *op)
>  		goto err_netdev;
>  	}
>  	if (resource_size(&mem) < sizeof(struct mpc52xx_fec)) {
> -		pr_err("invalid resource size (%lx < %x), check mpc52xx_devices.c\n",
> +		pr_err("invalid resource size (%lx < %zx), check mpc52xx_devices.c\n",
>  		       (unsigned long)resource_size(&mem),
>  		       sizeof(struct mpc52xx_fec));

mpc52xx_devices.c does not exist.
It was deleted nearly 20 years ago by:

commit 917f0af9e5a9ceecf9e72537fabb501254ba321d
Author: Paul Mackerras <paulus@ozlabs.org>
Date:   Mon Jun 9 14:01:46 2008 +1000

and

It's possible to remove the cast to unsigned long and use %pa
and %zx could be %#zx to match output prefixes.

Re: [PATCH net] net: fec_mpc52xx: Fix resource size format

[Rosen Penev]

On Tue, May 19, 2026 at 10:38 AM Joe Perches <joe@perches.com> wrote:
>
> On Sun, 2026-05-17 at 21:40 -0700, Rosen Penev wrote:
> > resource_size() is printed after casting to unsigned long, but
> > sizeof() returns size_t.
>
> trivia and I wonder if this code path is even possible
Not sure what you mean. probe has a corresponding of_match_table in
the platform_driver struct.
>
> > diff --git a/drivers/net/ethernet/freescale/fec_mpc52xx.c b/drivers/net/ethernet/freescale/fec_mpc52xx.c
> []
> > @@ -834,7 +834,7 @@ static int mpc52xx_fec_probe(struct platform_device *op)
> >               goto err_netdev;
> >       }
> >       if (resource_size(&mem) < sizeof(struct mpc52xx_fec)) {
> > -             pr_err("invalid resource size (%lx < %x), check mpc52xx_devices.c\n",
> > +             pr_err("invalid resource size (%lx < %zx), check mpc52xx_devices.c\n",
> >                      (unsigned long)resource_size(&mem),
> >                      sizeof(struct mpc52xx_fec));
>
> mpc52xx_devices.c does not exist.
right. the message is wrong. this is only meant to be a compilation fix.
> It was deleted nearly 20 years ago by:
>
> commit 917f0af9e5a9ceecf9e72537fabb501254ba321d
> Author: Paul Mackerras <paulus@ozlabs.org>
> Date:   Mon Jun 9 14:01:46 2008 +1000
>
> and
>
> It's possible to remove the cast to unsigned long and use %pa
> and %zx could be %#zx to match output prefixes.
Yeah I haven't had luck getting cleanups like this merged. Compilation
fix has a higher chance of getting in.

Re: [PATCH net] net: fec_mpc52xx: Fix resource size format

[Jakub Kicinski]

On Sun, 17 May 2026 21:40:47 -0700 Rosen Penev wrote:
> resource_size() is printed after casting to unsigned long,

This portion is confusing and irrelevant, you're touching the second
arg.

> but sizeof() returns size_t.
> 
> Use %zx for the sizeof() value so the format string matches the
> argument type so that 64-bit builds compile.
 
Not on 64-bit x86 builds apparently. Please provide more details
on which platform and compiler you see a warning and the warning.

> Assisted-by: Codex:GPT-5.5
> Fixes: 31b7720c8259 ("fec: Convert printks to netdev_<level>")

Not a fix.

> Signed-off-by: Rosen Penev <rosenp@gmail.com>
> ---
>  drivers/net/ethernet/freescale/fec_mpc52xx.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/net/ethernet/freescale/fec_mpc52xx.c b/drivers/net/ethernet/freescale/fec_mpc52xx.c
> index 556f646073af..a770e7edb1dc 100644
> --- a/drivers/net/ethernet/freescale/fec_mpc52xx.c
> +++ b/drivers/net/ethernet/freescale/fec_mpc52xx.c
> @@ -834,7 +834,7 @@ static int mpc52xx_fec_probe(struct platform_device *op)
>  		goto err_netdev;
>  	}
>  	if (resource_size(&mem) < sizeof(struct mpc52xx_fec)) {
> -		pr_err("invalid resource size (%lx < %x), check mpc52xx_devices.c\n",
> +		pr_err("invalid resource size (%lx < %zx), check mpc52xx_devices.c\n",
>  		       (unsigned long)resource_size(&mem),
>  		       sizeof(struct mpc52xx_fec));
>  		rv = -EINVAL;

Re: [PATCH net] net: fec_mpc52xx: Fix resource size format

[Rosen Penev]

On Wed, May 20, 2026 at 5:06 PM Jakub Kicinski <kuba@kernel.org> wrote:
>
> On Sun, 17 May 2026 21:40:47 -0700 Rosen Penev wrote:
> > resource_size() is printed after casting to unsigned long,
>
> This portion is confusing and irrelevant, you're touching the second
> arg.
>
> > but sizeof() returns size_t.
> >
> > Use %zx for the sizeof() value so the format string matches the
> > argument type so that 64-bit builds compile.
>
> Not on 64-bit x86 builds apparently. Please provide more details
> on which platform and compiler you see a warning and the warning.
64-bit powerpc. How would this driver be built on x86? There are a lot
of PPC APIs used here.

drivers/net/ethernet/freescale/fec_mpc52xx.c:46:10: fatal error:
'asm/mpc52xx.h' file not found
   46 | #include <asm/mpc52xx.h>
      |          ^~~~~~~~~~~~~~~
>
> > Assisted-by: Codex:GPT-5.5
> > Fixes: 31b7720c8259 ("fec: Convert printks to netdev_<level>")
>
> Not a fix.
>
> > Signed-off-by: Rosen Penev <rosenp@gmail.com>
> > ---
> >  drivers/net/ethernet/freescale/fec_mpc52xx.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/net/ethernet/freescale/fec_mpc52xx.c b/drivers/net/ethernet/freescale/fec_mpc52xx.c
> > index 556f646073af..a770e7edb1dc 100644
> > --- a/drivers/net/ethernet/freescale/fec_mpc52xx.c
> > +++ b/drivers/net/ethernet/freescale/fec_mpc52xx.c
> > @@ -834,7 +834,7 @@ static int mpc52xx_fec_probe(struct platform_device *op)
> >               goto err_netdev;
> >       }
> >       if (resource_size(&mem) < sizeof(struct mpc52xx_fec)) {
> > -             pr_err("invalid resource size (%lx < %x), check mpc52xx_devices.c\n",
> > +             pr_err("invalid resource size (%lx < %zx), check mpc52xx_devices.c\n",
> >                      (unsigned long)resource_size(&mem),
> >                      sizeof(struct mpc52xx_fec));
> >               rv = -EINVAL;
>

Re: [PATCH net] net: fec_mpc52xx: Fix resource size format

[Jakub Kicinski]

On Wed, 20 May 2026 17:42:37 -0700 Rosen Penev wrote:
> > Not on 64-bit x86 builds apparently. Please provide more details
> > on which platform and compiler you see a warning and the warning.  
> 64-bit powerpc. How would this driver be built on x86? 

Hm, CI did rebuild a bunch of fec objects when this patch was applied.
But looking closer indeed not this exact file.