diff mbox series

um: vector: Fix memory leak in vector_config

Message ID 20221229075323.1394010-1-linmq006@gmail.com
State New
Headers
Series um: vector: Fix memory leak in vector_config |

Commit Message

Miaoqian Lin Dec. 29, 2022, 7:53 a.m. UTC 
  kstrdup() return newly allocated copy of the string.
Call kfree() to release the memory when after use.

Fixes: 49da7e64f33e ("High Performance UML Vector Network Driver")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
---
 arch/um/drivers/vector_kern.c | 1 +
 1 file changed, 1 insertion(+)

Comments

Geert Uytterhoeven Jan. 3, 2023, 8 a.m. UTC | #1 
Hi Miaoqian,

On Thu, Dec 29, 2022 at 8:53 AM Miaoqian Lin <linmq006@gmail.com> wrote:
> kstrdup() return newly allocated copy of the string.
> Call kfree() to release the memory when after use.
>
> Fixes: 49da7e64f33e ("High Performance UML Vector Network Driver")
> Signed-off-by: Miaoqian Lin <linmq006@gmail.com>

Thanks for your patch!

> --- a/arch/um/drivers/vector_kern.c
> +++ b/arch/um/drivers/vector_kern.c
> @@ -765,6 +765,7 @@ static int vector_config(char *str, char **error_out)
>
>         parsed = uml_parse_vector_ifspec(params);
>
> +       kfree(params);

Are you sure the memory pointed to by "params" is no longer used?
"parsed" seems to contain pointers pointing to (parts of) the string
pointed to by "params", so it cannot be freed.

>         if (parsed == NULL) {
>                 *error_out = "vector_config failed to parse parameters";
>                 return -EINVAL;

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
Anton Ivanov Jan. 3, 2023, 10:28 a.m. UTC | #2 
On 03/01/2023 08:00, Geert Uytterhoeven wrote:
> Hi Miaoqian,
>
> On Thu, Dec 29, 2022 at 8:53 AM Miaoqian Lin <linmq006@gmail.com> wrote:
>> kstrdup() return newly allocated copy of the string.
>> Call kfree() to release the memory when after use.
>>
>> Fixes: 49da7e64f33e ("High Performance UML Vector Network Driver")
>> Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
> Thanks for your patch!
>
>> --- a/arch/um/drivers/vector_kern.c
>> +++ b/arch/um/drivers/vector_kern.c
>> @@ -765,6 +765,7 @@ static int vector_config(char *str, char **error_out)
>>
>>          parsed = uml_parse_vector_ifspec(params);
>>
>> +       kfree(params);
> Are you sure the memory pointed to by "params" is no longer used?
> "parsed" seems to contain pointers pointing to (parts of) the string
> pointed to by "params", so it cannot be freed.

+1.

I was just about to send the same comment.

>
>>          if (parsed == NULL) {
>>                  *error_out = "vector_config failed to parse parameters";
>>                  return -EINVAL;
> Gr{oetje,eeting}s,
>
>                          Geert
>
> --
> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
>
> In personal conversations with technical people, I call myself a hacker. But
> when I'm talking to journalists I just say "programmer" or something like that.
>                                  -- Linus Torvalds
>
Miaoqian Lin Jan. 3, 2023, 12:16 p.m. UTC | #3 
On 2023/1/3 18:28, Anton Ivanov wrote:
>
> On 03/01/2023 08:00, Geert Uytterhoeven wrote:
>> Hi Miaoqian,
>>
>> On Thu, Dec 29, 2022 at 8:53 AM Miaoqian Lin <linmq006@gmail.com> wrote:
>>> kstrdup() return newly allocated copy of the string.
>>> Call kfree() to release the memory when after use.
>>>
>>> Fixes: 49da7e64f33e ("High Performance UML Vector Network Driver")
>>> Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
>> Thanks for your patch!
>>
>>> --- a/arch/um/drivers/vector_kern.c
>>> +++ b/arch/um/drivers/vector_kern.c
>>> @@ -765,6 +765,7 @@ static int vector_config(char *str, char **error_out)
>>>
>>>          parsed = uml_parse_vector_ifspec(params);
>>>
>>> +       kfree(params);
>> Are you sure the memory pointed to by "params" is no longer used?
>> "parsed" seems to contain pointers pointing to (parts of) the string
>> pointed to by "params", so it cannot be freed.
>
> +1.
>
> I was just about to send the same comment.
>
Oh yes, thanks for spotting this. We should only perform release when uml_parse_vector_ifspec() fails (returns NULL). In this situation, 'params' is no longer used. Do you agree?

Thanks,

>>
>>>          if (parsed == NULL) {
>>>                  *error_out = "vector_config failed to parse parameters";
>>>                  return -EINVAL;
>> Gr{oetje,eeting}s,
>>
>>                          Geert
>>
>> -- 
>> Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org
>>
>> In personal conversations with technical people, I call myself a hacker. But
>> when I'm talking to journalists I just say "programmer" or something like that.
>>                                  -- Linus Torvalds
>>
Geert Uytterhoeven Jan. 3, 2023, 1:06 p.m. UTC | #4 
Hi Miaoqian,

On Tue, Jan 3, 2023 at 1:17 PM Miaoqian Lin <linmq006@gmail.com> wrote:
> On 2023/1/3 18:28, Anton Ivanov wrote:
> > On 03/01/2023 08:00, Geert Uytterhoeven wrote:
> >> On Thu, Dec 29, 2022 at 8:53 AM Miaoqian Lin <linmq006@gmail.com> wrote:
> >>> kstrdup() return newly allocated copy of the string.
> >>> Call kfree() to release the memory when after use.
> >>>
> >>> Fixes: 49da7e64f33e ("High Performance UML Vector Network Driver")
> >>> Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
> >> Thanks for your patch!
> >>
> >>> --- a/arch/um/drivers/vector_kern.c
> >>> +++ b/arch/um/drivers/vector_kern.c
> >>> @@ -765,6 +765,7 @@ static int vector_config(char *str, char **error_out)
> >>>
> >>>          parsed = uml_parse_vector_ifspec(params);
> >>>
> >>> +       kfree(params);
> >> Are you sure the memory pointed to by "params" is no longer used?
> >> "parsed" seems to contain pointers pointing to (parts of) the string
> >> pointed to by "params", so it cannot be freed.
> >
> > +1.
> >
> > I was just about to send the same comment.
> >
> Oh yes, thanks for spotting this. We should only perform release when uml_parse_vector_ifspec() fails (returns NULL). In this situation, 'params' is no longer used. Do you agree?

Yes, that sounds fine to me.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds
diff mbox series

Patch

diff --git a/arch/um/drivers/vector_kern.c b/arch/um/drivers/vector_kern.c
index ded7c47d2fbe..78f32005dd90 100644
--- a/arch/um/drivers/vector_kern.c
+++ b/arch/um/drivers/vector_kern.c
@@ -765,6 +765,7 @@  static int vector_config(char *str, char **error_out)
 
 	parsed = uml_parse_vector_ifspec(params);
 
+	kfree(params);
 	if (parsed == NULL) {
 		*error_out = "vector_config failed to parse parameters";
 		return -EINVAL;