mtd: core: Fix resource leak in a for_each_child_of_node() loop

[PATCH v3] mtd: core: Fix resource leak in a for_each_child_of_node() loop

Posted by Yang Rong 2 years, 7 months ago

Ensure child node references are properly decremented in the error path
  of this function implementation.

Fixes: 2df11f00100d ("mtd: core: try to find OF node for every MTD partition")
Signed-off-by: Yang Rong <yangrong@vivo.com>
---
v2: https://lore.kernel.org/all/20230710061412.2272-1-yangrong@vivo.com/
The modifications made copared to the previous version are as follows:
1.Optimized the expression of the Subject, making it more accurate.
2.Optimized the content of the commit message.
3.Add the tag "Fixes".

 drivers/mtd/mtdcore.c | 1 +
 1 file changed, 1 insertion(+)
 mode change 100644 => 100755 drivers/mtd/mtdcore.c

diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c
index e00b12aa5ec9..b02cb8b1a38d
--- a/drivers/mtd/mtdcore.c
+++ b/drivers/mtd/mtdcore.c
@@ -591,6 +591,7 @@ static void mtd_check_of_node(struct mtd_info *mtd)
 		if (plen == mtd_name_len &&
 		    !strncmp(mtd->name, pname + offset, plen)) {
 			mtd_set_of_node(mtd, mtd_dn);
+			of_node_put(mtd_dn);
 			break;
 		}
 	}
-- 
2.35.3

Re: [PATCH v3] mtd: core: Fix resource leak in a for_each_child_of_node() loop

Posted by Rafał Miłecki 2 years, 7 months ago

On 2023-07-11 04:03, Yang Rong wrote:
> Ensure child node references are properly decremented in the error path
>   of this function implementation.

Nitpicking: it's actually a /success/ code path that misses
of_node_put().


> Fixes: 2df11f00100d ("mtd: core: try to find OF node for every MTD 
> partition")
> Signed-off-by: Yang Rong <yangrong@vivo.com>

Acked-by: Rafał Miłecki <rafal@milecki.pl>


> ---
> v2: 
> https://lore.kernel.org/all/20230710061412.2272-1-yangrong@vivo.com/
> The modifications made copared to the previous version are as follows:
> 1.Optimized the expression of the Subject, making it more accurate.
> 2.Optimized the content of the commit message.
> 3.Add the tag "Fixes".
> 
>  drivers/mtd/mtdcore.c | 1 +
>  1 file changed, 1 insertion(+)
>  mode change 100644 => 100755 drivers/mtd/mtdcore.c
> 
> diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c
> index e00b12aa5ec9..b02cb8b1a38d
> --- a/drivers/mtd/mtdcore.c
> +++ b/drivers/mtd/mtdcore.c
> @@ -591,6 +591,7 @@ static void mtd_check_of_node(struct mtd_info *mtd)
>  		if (plen == mtd_name_len &&
>  		    !strncmp(mtd->name, pname + offset, plen)) {
>  			mtd_set_of_node(mtd, mtd_dn);
> +			of_node_put(mtd_dn);
>  			break;
>  		}
>  	}

Re: [PATCH v3] mtd: core: Fix resource leak in a for_each_child_of_node() loop

Posted by Miquel Raynal 2 years, 6 months ago

Hi Rafał,

rafal@milecki.pl wrote on Tue, 11 Jul 2023 18:37:42 +0200:

> On 2023-07-11 04:03, Yang Rong wrote:
> > Ensure child node references are properly decremented in the error path
> >   of this function implementation.  
> 
> Nitpicking: it's actually a /success/ code path that misses
> of_node_put().

Please also add:

Cc: stable...

> 
> 
> > Fixes: 2df11f00100d ("mtd: core: try to find OF node for every MTD > partition")
> > Signed-off-by: Yang Rong <yangrong@vivo.com>  
> 
> Acked-by: Rafał Miłecki <rafal@milecki.pl>
> 
> 
> > ---
> > v2: > https://lore.kernel.org/all/20230710061412.2272-1-yangrong@vivo.com/
> > The modifications made copared to the previous version are as follows:
> > 1.Optimized the expression of the Subject, making it more accurate.
> > 2.Optimized the content of the commit message.
> > 3.Add the tag "Fixes".
> > 
> >  drivers/mtd/mtdcore.c | 1 +
> >  1 file changed, 1 insertion(+)
> >  mode change 100644 => 100755 drivers/mtd/mtdcore.c
> > 
> > diff --git a/drivers/mtd/mtdcore.c b/drivers/mtd/mtdcore.c
> > index e00b12aa5ec9..b02cb8b1a38d
> > --- a/drivers/mtd/mtdcore.c
> > +++ b/drivers/mtd/mtdcore.c
> > @@ -591,6 +591,7 @@ static void mtd_check_of_node(struct mtd_info *mtd)
> >  		if (plen == mtd_name_len &&
> >  		    !strncmp(mtd->name, pname + offset, plen)) {
> >  			mtd_set_of_node(mtd, mtd_dn);
> > +			of_node_put(mtd_dn);
> >  			break;
> >  		}
> >  	}  


Thanks,
Miquèl