2024-03-01 07:47:03

by Markus Elfring

Subject: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

From: Markus Elfring <[email protected]>
Date: Fri, 1 Mar 2024 08:23:24 +0100

Avoid the specification of a duplicate fwnode_handle_put() call
in this function implementation.

Signed-off-by: Markus Elfring <[email protected]>
---
drivers/media/i2c/ds90ub960.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)

diff --git a/drivers/media/i2c/ds90ub960.c b/drivers/media/i2c/ds90ub960.c
index ffe5f25f8647..eb708ed7b56e 100644
--- a/drivers/media/i2c/ds90ub960.c
+++ b/drivers/media/i2c/ds90ub960.c
@@ -3486,10 +3486,7 @@ static int ub960_parse_dt_rxports(struct ub960_data *priv)
}
}

- fwnode_handle_put(links_fwnode);
-
- return 0;
-
+ ret = 0;
err_put_links:
fwnode_handle_put(links_fwnode);

--
2.44.0

2024-03-01 08:49:35

by Tomi Valkeinen

Subject: Re: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

Hi,

On 01/03/2024 10:46, Sakari Ailus wrote:
> Hi Markus,
>
> On Fri, Mar 01, 2024 at 08:46:25AM +0100, Markus Elfring wrote:
>> From: Markus Elfring <[email protected]>
>> Date: Fri, 1 Mar 2024 08:23:24 +0100
>>
>> Avoid the specification of a duplicate fwnode_handle_put() call
>> in this function implementation.
>>
>> Signed-off-by: Markus Elfring <[email protected]>
>> ---
>> drivers/media/i2c/ds90ub960.c | 5 +----
>> 1 file changed, 1 insertion(+), 4 deletions(-)
>>
>> diff --git a/drivers/media/i2c/ds90ub960.c b/drivers/media/i2c/ds90ub960.c
>> index ffe5f25f8647..eb708ed7b56e 100644
>> --- a/drivers/media/i2c/ds90ub960.c
>> +++ b/drivers/media/i2c/ds90ub960.c
>> @@ -3486,10 +3486,7 @@ static int ub960_parse_dt_rxports(struct ub960_data *priv)
>> }
>> }
>>
>> - fwnode_handle_put(links_fwnode);
>> -
>> - return 0;
>> -
>> + ret = 0;
>
> I think it'd be nicer to initialise ret as zero, then you can just drop the
> assignment above.

I don't like successful execution entering error paths. That's why
there's the return 0.

Tomi

>> err_put_links:
>> fwnode_handle_put(links_fwnode);
>>
>

2024-03-01 12:41:12

by Sakari Ailus

Subject: Re: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

Huomenta,

On Fri, Mar 01, 2024 at 10:49:19AM +0200, Tomi Valkeinen wrote:
> Hi,
>
> On 01/03/2024 10:46, Sakari Ailus wrote:
> > Hi Markus,
> >
> > On Fri, Mar 01, 2024 at 08:46:25AM +0100, Markus Elfring wrote:
> > > From: Markus Elfring <[email protected]>
> > > Date: Fri, 1 Mar 2024 08:23:24 +0100
> > >
> > > Avoid the specification of a duplicate fwnode_handle_put() call
> > > in this function implementation.
> > >
> > > Signed-off-by: Markus Elfring <[email protected]>
> > > ---
> > > drivers/media/i2c/ds90ub960.c | 5 +----
> > > 1 file changed, 1 insertion(+), 4 deletions(-)
> > >
> > > diff --git a/drivers/media/i2c/ds90ub960.c b/drivers/media/i2c/ds90ub960.c
> > > index ffe5f25f8647..eb708ed7b56e 100644
> > > --- a/drivers/media/i2c/ds90ub960.c
> > > +++ b/drivers/media/i2c/ds90ub960.c
> > > @@ -3486,10 +3486,7 @@ static int ub960_parse_dt_rxports(struct ub960_data *priv)
> > > }
> > > }
> > >
> > > - fwnode_handle_put(links_fwnode);
> > > -
> > > - return 0;
> > > -
> > > + ret = 0;
> >
> > I think it'd be nicer to initialise ret as zero, then you can just drop the
> > assignment above.
>
> I don't like successful execution entering error paths. That's why there's
> the return 0.

It could be called a common cleanup path as what you really want to do here
is to put the fwnode handle, independently of whether there was an error.
I think the current code is of course fine, too.

Soon you can do

struct fwnode_handle *links_fwnode __free(fwnode_handle);

and forget about putting it (but you must need putting it).

--
Terveisin,

Sakari Ailus

2024-03-01 15:05:13

by Sakari Ailus

Subject: Re: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

Hi Markus,

On Fri, Mar 01, 2024 at 08:46:25AM +0100, Markus Elfring wrote:
> From: Markus Elfring <[email protected]>
> Date: Fri, 1 Mar 2024 08:23:24 +0100
>
> Avoid the specification of a duplicate fwnode_handle_put() call
> in this function implementation.
>
> Signed-off-by: Markus Elfring <[email protected]>
> ---
> drivers/media/i2c/ds90ub960.c | 5 +----
> 1 file changed, 1 insertion(+), 4 deletions(-)
>
> diff --git a/drivers/media/i2c/ds90ub960.c b/drivers/media/i2c/ds90ub960.c
> index ffe5f25f8647..eb708ed7b56e 100644
> --- a/drivers/media/i2c/ds90ub960.c
> +++ b/drivers/media/i2c/ds90ub960.c
> @@ -3486,10 +3486,7 @@ static int ub960_parse_dt_rxports(struct ub960_data *priv)
> }
> }
>
> - fwnode_handle_put(links_fwnode);
> -
> - return 0;
> -
> + ret = 0;

I think it'd be nicer to initialise ret as zero, then you can just drop the
assignment above.

> err_put_links:
> fwnode_handle_put(links_fwnode);
>

--
Regards,

Sakari Ailus

2024-03-01 23:36:12

by Andy Shevchenko

Subject: Re: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

On Fri, Mar 01, 2024 at 09:02:41AM +0000, Sakari Ailus wrote:
> On Fri, Mar 01, 2024 at 10:49:19AM +0200, Tomi Valkeinen wrote:
> > On 01/03/2024 10:46, Sakari Ailus wrote:
> > > On Fri, Mar 01, 2024 at 08:46:25AM +0100, Markus Elfring wrote:
> > > > From: Markus Elfring <[email protected]>
> > > > Date: Fri, 1 Mar 2024 08:23:24 +0100
> > > >
> > > > Avoid the specification of a duplicate fwnode_handle_put() call
> > > > in this function implementation.
> > > >
> > > > Signed-off-by: Markus Elfring <[email protected]>
> > > > ---
> > > > drivers/media/i2c/ds90ub960.c | 5 +----
> > > > 1 file changed, 1 insertion(+), 4 deletions(-)
> > > >
> > > > diff --git a/drivers/media/i2c/ds90ub960.c b/drivers/media/i2c/ds90ub960.c
> > > > index ffe5f25f8647..eb708ed7b56e 100644
> > > > --- a/drivers/media/i2c/ds90ub960.c
> > > > +++ b/drivers/media/i2c/ds90ub960.c
> > > > @@ -3486,10 +3486,7 @@ static int ub960_parse_dt_rxports(struct ub960_data *priv)
> > > > }
> > > > }
> > > >
> > > > - fwnode_handle_put(links_fwnode);
> > > > -
> > > > - return 0;
> > > > -
> > > > + ret = 0;
> > >
> > > I think it'd be nicer to initialise ret as zero, then you can just drop the
> > > assignment above.

I think tearing apart the assignment and its actual user is not good.

> > I don't like successful execution entering error paths. That's why there's
> > the return 0.
>
> It could be called a common cleanup path as what you really want to do here
> is to put the fwnode handle, independently of whether there was an error.
> I think the current code is of course fine, too.
>
> Soon you can do
>
> struct fwnode_handle *links_fwnode __free(fwnode_handle);
>
> and forget about putting it (but you must need putting it).

Let's wait for the Jonathan's patches to land (v6.9-rc1 I hope) and then
we may modify drivers if needed.

--
With Best Regards,
Andy Shevchenko

2024-03-02 09:14:16

by Dan Carpenter

Subject: Re: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

On Fri, Mar 01, 2024 at 07:36:17PM +0200, Andy Shevchenko wrote:
> On Fri, Mar 01, 2024 at 09:02:41AM +0000, Sakari Ailus wrote:
> > On Fri, Mar 01, 2024 at 10:49:19AM +0200, Tomi Valkeinen wrote:
> > > On 01/03/2024 10:46, Sakari Ailus wrote:
> > > > On Fri, Mar 01, 2024 at 08:46:25AM +0100, Markus Elfring wrote:
> > > > > From: Markus Elfring <[email protected]>
> > > > > Date: Fri, 1 Mar 2024 08:23:24 +0100
> > > > >
> > > > > Avoid the specification of a duplicate fwnode_handle_put() call
> > > > > in this function implementation.
> > > > >
> > > > > Signed-off-by: Markus Elfring <[email protected]>
> > > > > ---
> > > > > drivers/media/i2c/ds90ub960.c | 5 +----
> > > > > 1 file changed, 1 insertion(+), 4 deletions(-)
> > > > >
> > > > > diff --git a/drivers/media/i2c/ds90ub960.c b/drivers/media/i2c/ds90ub960.c
> > > > > index ffe5f25f8647..eb708ed7b56e 100644
> > > > > --- a/drivers/media/i2c/ds90ub960.c
> > > > > +++ b/drivers/media/i2c/ds90ub960.c
> > > > > @@ -3486,10 +3486,7 @@ static int ub960_parse_dt_rxports(struct ub960_data *priv)
> > > > > }
> > > > > }
> > > > >
> > > > > - fwnode_handle_put(links_fwnode);
> > > > > -
> > > > > - return 0;
> > > > > -
> > > > > + ret = 0;
> > > >
> > > > I think it'd be nicer to initialise ret as zero, then you can just drop the
> > > > assignment above.
>
> I think tearing apart the assignment and its actual user is not good.
>
> > > I don't like successful execution entering error paths. That's why there's
> > > the return 0.
> >
> > It could be called a common cleanup path as what you really want to do here
> > is to put the fwnode handle, independently of whether there was an error.
> > I think the current code is of course fine, too.
> >
> > Soon you can do
> >
> > struct fwnode_handle *links_fwnode __free(fwnode_handle);
> >
> > and forget about putting it (but you must need putting it).
>
> Let's wait for the Jonathan's patches to land (v6.9-rc1 I hope) and then
> we may modify drivers if needed.

The __free(fwnode_handle) stuff has already been merged.

We could do some additional work to make a _scoped() macro for
fwnode_handles but here it's function wide so we already have what we
need.

regards,
dan carpenter

2024-03-02 11:26:19

by Markus Elfring

Subject: Re: [PATCH] media: i2c: ds90ub960: Delete duplicate source code in ub960_parse_dt_rxports()

> The __free(fwnode_handle) stuff has already been merged.

Would you like to point a corresponding commit out?

Regards,
Markus