[PATCH] drivers: virtio: Use PTR_RET function

[PATCH] drivers: virtio: Use PTR_RET function

[Alexandru Gheorghiu]

Used PTR_RET function instead of IS_ERR and PTR_ERR.
Patch found using coccinelle.

Signed-off-by: Alexandru Gheorghiu <gheorghiuandru@gmail.com>
---
 drivers/virtio/virtio_mmio.c |    5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/virtio/virtio_mmio.c b/drivers/virtio/virtio_mmio.c
index 1ba0d68..d1e664f 100644
--- a/drivers/virtio/virtio_mmio.c
+++ b/drivers/virtio/virtio_mmio.c
@@ -567,10 +567,7 @@ static int vm_cmdline_set(const char *device,
 	pdev = platform_device_register_resndata(&vm_cmdline_parent,
 			"virtio-mmio", vm_cmdline_id++,
 			resources, ARRAY_SIZE(resources), NULL, 0);
-	if (IS_ERR(pdev))
-		return PTR_ERR(pdev);
-
-	return 0;
+	return PTR_RET(pdev);
 }
 
 static int vm_cmdline_get_device(struct device *dev, void *data)
-- 
1.7.9.5

Re: [PATCH] drivers: virtio: Use PTR_RET function

[Rusty Russell]

Alexandru Gheorghiu <gheorghiuandru@gmail.com> writes:

> Used PTR_RET function instead of IS_ERR and PTR_ERR.
> Patch found using coccinelle.

WTF is PTR_RET?  PTR_RET doesn't return anything.  Why is it called
that?  It doesn't even make sense.

ZERO_OR_PTR_ERR() maybe.

But what problem are we solving?  Insufficient churn in the tree?  Code
being too readable?  This isn't some hard-to-get right corner case, or a
missed optimization.

Andrew, what am I missing here?

Grumpy,
Rusty.

> Signed-off-by: Alexandru Gheorghiu <gheorghiuandru@gmail.com>
> ---
>  drivers/virtio/virtio_mmio.c |    5 +----
>  1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/drivers/virtio/virtio_mmio.c b/drivers/virtio/virtio_mmio.c
> index 1ba0d68..d1e664f 100644
> --- a/drivers/virtio/virtio_mmio.c
> +++ b/drivers/virtio/virtio_mmio.c
> @@ -567,10 +567,7 @@ static int vm_cmdline_set(const char *device,
>  	pdev = platform_device_register_resndata(&vm_cmdline_parent,
>  			"virtio-mmio", vm_cmdline_id++,
>  			resources, ARRAY_SIZE(resources), NULL, 0);
> -	if (IS_ERR(pdev))
> -		return PTR_ERR(pdev);
> -
> -	return 0;
> +	return PTR_RET(pdev);
>  }
>  
>  static int vm_cmdline_get_device(struct device *dev, void *data)
> -- 
> 1.7.9.5

Re: [PATCH] drivers: virtio: Use PTR_RET function

[Andrew Morton]

On Tue, 26 Mar 2013 13:57:09 +1030 Rusty Russell <rusty@rustcorp.com.au> wrote:

> Alexandru Gheorghiu <gheorghiuandru@gmail.com> writes:
> 
> > Used PTR_RET function instead of IS_ERR and PTR_ERR.
> > Patch found using coccinelle.
> 
> WTF is PTR_RET?  PTR_RET doesn't return anything.  Why is it called
> that?  It doesn't even make sense.
> 
> ZERO_OR_PTR_ERR() maybe.
> 
> But what problem are we solving?  Insufficient churn in the tree?  Code
> being too readable?  This isn't some hard-to-get right corner case, or a
> missed optimization.
> 
> Andrew, what am I missing here?

It seemed like a good idea at the time.  Merged it two years ago and
have since been keenly awaiting an opportunity to use it.

It seems that people _have_ been using it, but mainly netfilter people
and we know they're all crazy ;)

Re: [PATCH] drivers: virtio: Use PTR_RET function

[Rusty Russell]

Andru Gheorghiu <gheorghiuandru@gmail.com> writes:
> PTR_RET does return. It's perfectly equivalent to using IS_ERR and the
> returning PTR_ERR. The implementation is here [1].

Um, I read the implementation, thanks.

> The reason for using it is this: if you have a function that does
> something why not call it instead of reproducing it's behavior by
> explicitly writing what it does.

Because clarity matters, and this function makes callers less clear.
It's the most breathtakingly bad name since BUILD_BUG_ON_ZERO().

Why not change PTR_ERR to return 0 if !IS_ERR()?  Noone breaks, gcc
probably produces the same code, and noone needs to learn your weird
new kernel meme.

In fact, as gcc will produce the same code for "if (PTR_ERR(p))" as it
does for "if (IS_ERR(p))", you get to be one of the very, very few
people who have ever *reduced* the complexity of a kernel interface.

Cheers,
Rusty.