diff mbox series

driver core: Fix an error handling path in fw_devlink_create_devlink()

Message ID 858b140b276ceab52a84ce0d4f924b0c411560d1.1689618373.git.christophe.jaillet@wanadoo.fr
State New
Headers
Series driver core: Fix an error handling path in fw_devlink_create_devlink() |

Commit Message

Christophe JAILLET July 17, 2023, 6:26 p.m. UTC 
  All error handling paths go to 'out', except this one. Be consistent and
also branch to 'out' here.

Fixes: 74c782cff77b ("driver core: fw_devlink: Handle suppliers that don't use driver core")
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
---
/!\ Speculative /!\

   This patch is based on analysis of the surrounding code and should be
   reviewed with care !

   If the patch is wrong, maybe a comment in the code could explain why.

/!\ Speculative /!\
---
 drivers/base/core.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Saravana Kannan July 17, 2023, 6:57 p.m. UTC | #1 
On Mon, Jul 17, 2023 at 11:26 AM Christophe JAILLET
<christophe.jaillet@wanadoo.fr> wrote:
>
> All error handling paths go to 'out', except this one. Be consistent and
> also branch to 'out' here.
>
> Fixes: 74c782cff77b ("driver core: fw_devlink: Handle suppliers that don't use driver core")
> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
> ---
> /!\ Speculative /!\
>
>    This patch is based on analysis of the surrounding code and should be
>    reviewed with care !
>
>    If the patch is wrong, maybe a comment in the code could explain why.
>
> /!\ Speculative /!\

Looks okay to me, but I'd rather not much around the code for a
subjective preference. The existing code also avoids the pointless
put_device() and personally is easier to read (ok, nothing more of
relevance is done after this return) than trying to follow the
execution to the out and then realizing the put_device() is a NOP,
etc. It also means future changes to "out" will need to accommodate
for this path that doesn't really need the "out" path.

-Saravana

> ---
>  drivers/base/core.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/base/core.c b/drivers/base/core.c
> index 3dff5037943e..854c1fab742c 100644
> --- a/drivers/base/core.c
> +++ b/drivers/base/core.c
> @@ -2107,7 +2107,8 @@ static int fw_devlink_create_devlink(struct device *con,
>             fwnode_ancestor_init_without_drv(sup_handle)) {
>                 dev_dbg(con, "Not linking %pfwf - might never become dev\n",
>                         sup_handle);
> -               return -EINVAL;
> +               ret = -EINVAL;
> +               goto out;
>         }
>
>         ret = -EAGAIN;
> --
> 2.34.1
>
Christophe JAILLET July 17, 2023, 7:32 p.m. UTC | #2 
Le 17/07/2023 à 20:57, Saravana Kannan a écrit :
> On Mon, Jul 17, 2023 at 11:26 AM Christophe JAILLET
> <christophe.jaillet@wanadoo.fr> wrote:
>>
>> All error handling paths go to 'out', except this one. Be consistent and
>> also branch to 'out' here.
>>
>> Fixes: 74c782cff77b ("driver core: fw_devlink: Handle suppliers that don't use driver core")
>> Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
>> ---
>> /!\ Speculative /!\
>>
>>     This patch is based on analysis of the surrounding code and should be
>>     reviewed with care !
>>
>>     If the patch is wrong, maybe a comment in the code could explain why.
>>
>> /!\ Speculative /!\
> 
> Looks okay to me, but I'd rather not much around the code for a
> subjective preference. The existing code also avoids the pointless
> put_device() and personally is easier to read (ok, nothing more of
> relevance is done after this return) than trying to follow the
> execution to the out and then realizing the put_device() is a NOP,
> etc. It also means future changes to "out" will need to accommodate
> for this path that doesn't really need the "out" path.
> 
> -Saravana

Agreed,

I didn't realized that this code was executed only if "if (sup_dev)" 
fails, so only if sup_dev is NULL.

Sorry for the noise.

CJ



> 
>> ---
>>   drivers/base/core.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/base/core.c b/drivers/base/core.c
>> index 3dff5037943e..854c1fab742c 100644
>> --- a/drivers/base/core.c
>> +++ b/drivers/base/core.c
>> @@ -2107,7 +2107,8 @@ static int fw_devlink_create_devlink(struct device *con,
>>              fwnode_ancestor_init_without_drv(sup_handle)) {
>>                  dev_dbg(con, "Not linking %pfwf - might never become dev\n",
>>                          sup_handle);
>> -               return -EINVAL;
>> +               ret = -EINVAL;
>> +               goto out;
>>          }
>>
>>          ret = -EAGAIN;
>> --
>> 2.34.1
>>
>
Dan Carpenter July 18, 2023, 6:12 a.m. UTC | #3 
On Mon, Jul 17, 2023 at 11:57:51AM -0700, Saravana Kannan wrote:
> > ---
> >  drivers/base/core.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> >
> > diff --git a/drivers/base/core.c b/drivers/base/core.c
> > index 3dff5037943e..854c1fab742c 100644
> > --- a/drivers/base/core.c
> > +++ b/drivers/base/core.c
> > @@ -2107,7 +2107,8 @@ static int fw_devlink_create_devlink(struct device *con,
> >             fwnode_ancestor_init_without_drv(sup_handle)) {
> >                 dev_dbg(con, "Not linking %pfwf - might never become dev\n",
> >                         sup_handle);
> > -               return -EINVAL;
> > +               ret = -EINVAL;
> > +               goto out;
> >         }
> >
> >         ret = -EAGAIN;

It would have been more obvious that put_device() is a no-op if this
ret = -EAGAIN were changed to "return -EAGAIN;".  That would probably
silence some sort of static checker warning...

regards,
dan carpenter
diff mbox series

Patch

diff --git a/drivers/base/core.c b/drivers/base/core.c
index 3dff5037943e..854c1fab742c 100644
--- a/drivers/base/core.c
+++ b/drivers/base/core.c
@@ -2107,7 +2107,8 @@  static int fw_devlink_create_devlink(struct device *con,
 	    fwnode_ancestor_init_without_drv(sup_handle)) {
 		dev_dbg(con, "Not linking %pfwf - might never become dev\n",
 			sup_handle);
-		return -EINVAL;
+		ret = -EINVAL;
+		goto out;
 	}
 
 	ret = -EAGAIN;