[spp] [PATCH] spp_vf: fix initializing bug of forwarder

[spp] [PATCH] spp_vf: fix initializing bug of forwarder

[x-fn-spp]

From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>

clear_forward_info() always initializes path[0] of g_forward_info[id].
It is a bug and causes stopping forwarding accidentally if path[0] is
referenced by spp_forward().

This patch changes initializing as path[info->upd_index] instead of 0.
It also includes refactoring to be simple by using local vars.

Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
---
 src/vf/spp_forward.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
index 0a43608..ee1254c 100644
--- a/src/vf/spp_forward.c
+++ b/src/vf/spp_forward.c
@@ -51,9 +51,9 @@ spp_forward_init(void)
 
 /* Clear info for one element. */
 static void
-clear_forward_info(int id)
+clear_forward_info(struct forward_path *path)
 {
-	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
+	memset(path, 0x00, sizeof(struct forward_path));
 }
 
 /* Update forward info */
@@ -84,7 +84,7 @@ spp_forward_update(struct spp_component_info *component)
 		return -1;
 	}
 
-	clear_forward_info(component->component_id);
+	clear_forward_info(path);
 
 	RTE_LOG(INFO, FORWARD,
 			"Component[%d] Start update component. (name = %s, type = %d)\n",
-- 
2.18.0

Re: [spp] [PATCH] spp_vf: fix initializing bug of forwarder

[Yasufumi Ogawa]

On 2018/08/02 17:31, x-fn-spp@sl.ntt-tx.co.jp wrote:
> From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> 
> clear_forward_info() always initializes path[0] of g_forward_info[id].
> It is a bug and causes stopping forwarding accidentally if path[0] is
> referenced by spp_forward().
> 
> This patch changes initializing as path[info->upd_index] instead of 0.
> It also includes refactoring to be simple by using local vars.
Hideyuki,

I do not know why claer_forward_info() is requierd because it is called only one time and simply just doing memset(). I think it 
is not needed to be a function without called from several points. In addition, the name of function is not so good. It does not 
clear 'forward_info' but 'forward_path'. I think clear_forward_path() is appropriate if you want to make it as a function.

Thanks,
Yasufumi
> 
> Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
> ---
>   src/vf/spp_forward.c | 6 +++---
>   1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
> index 0a43608..ee1254c 100644
> --- a/src/vf/spp_forward.c
> +++ b/src/vf/spp_forward.c
> @@ -51,9 +51,9 @@ spp_forward_init(void)
>   
>   /* Clear info for one element. */
>   static void
> -clear_forward_info(int id)
> +clear_forward_info(struct forward_path *path)
>   {
> -	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
> +	memset(path, 0x00, sizeof(struct forward_path));
>   }
>   
>   /* Update forward info */
> @@ -84,7 +84,7 @@ spp_forward_update(struct spp_component_info *component)
>   		return -1;
>   	}
>   
> -	clear_forward_info(component->component_id);
> +	clear_forward_info(path);
>   
>   	RTE_LOG(INFO, FORWARD,
>   			"Component[%d] Start update component. (name = %s, type = %d)\n",
>

Re: [spp] [spp 03176] Re: [PATCH] spp_vf: fix initializing bug of forwarder

[Hideyuki Yamashita]

Hello Yasufumi-san,

Thanks for your comments. 
About your first point, I agree that there is only one point which calls this function. But
I still believe it is good to keep this as a function to keep
readability of calling side function.
So, I will change function name from clear_forward_info() to clear_forward_path().

What do you think?

BR,
Hideyuki Yamashita

> On 2018/08/02 17:31, x-fn-spp@sl.ntt-tx.co.jp wrote:
> > From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> >
> > clear_forward_info() always initializes path[0] of g_forward_info[id].
> > It is a bug and causes stopping forwarding accidentally if path[0] is
> > referenced by spp_forward().
> >
> > This patch changes initializing as path[info->upd_index] instead of 0.
> > It also includes refactoring to be simple by using local vars.
> Hideyuki,
> 
> I do not know why claer_forward_info() is requierd because it is called only one time and simply just doing memset(). I think it is not needed to be a function without called from several points. In addition, the name of function is not so good. It does not clear 'forward_info' but 'forward_path'. I think clear_forward_path() is appropriate if you want to make it as a function.
> 
> Thanks,
> Yasufumi
> >
> > Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> > Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
> > ---
> >   src/vf/spp_forward.c | 6 +++---
> >   1 file changed, 3 insertions(+), 3 deletions(-)
> >
> > diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
> > index 0a43608..ee1254c 100644
> > --- a/src/vf/spp_forward.c
> > +++ b/src/vf/spp_forward.c
> > @@ -51,9 +51,9 @@ spp_forward_init(void)
> >  >   /* Clear info for one element. */
> >   static void
> > -clear_forward_info(int id)
> > +clear_forward_info(struct forward_path *path)
> >   {
> > -	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
> > +	memset(path, 0x00, sizeof(struct forward_path));
> >   }
> >  >   /* Update forward info */
> > @@ -84,7 +84,7 @@ spp_forward_update(struct spp_component_info *component)
> >   		return -1;
> >   	}
> >  > -	clear_forward_info(component->component_id);
> > +	clear_forward_info(path);
> >  >   	RTE_LOG(INFO, FORWARD,
> >   			"Component[%d] Start update component. (name = %s, type = %d)\n",
> >
> 

Re: [spp] [spp 03176] Re: [PATCH] spp_vf: fix initializing bug of forwarder

[Yasufumi Ogawa]

On 2018/08/07 15:31, Hideyuki Yamashita wrote:
> Hello Yasufumi-san,
> 
> Thanks for your comments.
> About your first point, I agree that there is only one point which calls this function. But
> I still believe it is good to keep this as a function to keep
> readability of calling side function.
I do not think this function is more understandable than doing memset. It is just one line and no complexity. If you have any 
other reasons for, I think it it better to keep the code by removing the meaningless function.

> So, I will change function name from clear_forward_info() to clear_forward_path().
> 
> What do you think?
> 
> BR,
> Hideyuki Yamashita
> 
>> On 2018/08/02 17:31, x-fn-spp@sl.ntt-tx.co.jp wrote:
>>> From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
>>>
>>> clear_forward_info() always initializes path[0] of g_forward_info[id].
>>> It is a bug and causes stopping forwarding accidentally if path[0] is
>>> referenced by spp_forward().
>>>
>>> This patch changes initializing as path[info->upd_index] instead of 0.
>>> It also includes refactoring to be simple by using local vars.
>> Hideyuki,
>>
>> I do not know why claer_forward_info() is requierd because it is called only one time and simply just doing memset(). I think it is not needed to be a function without called from several points. In addition, the name of function is not so good. It does not clear 'forward_info' but 'forward_path'. I think clear_forward_path() is appropriate if you want to make it as a function.
>>
>> Thanks,
>> Yasufumi
>>>
>>> Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
>>> Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
>>> ---
>>>    src/vf/spp_forward.c | 6 +++---
>>>    1 file changed, 3 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
>>> index 0a43608..ee1254c 100644
>>> --- a/src/vf/spp_forward.c
>>> +++ b/src/vf/spp_forward.c
>>> @@ -51,9 +51,9 @@ spp_forward_init(void)
>>>   >   /* Clear info for one element. */
>>>    static void
>>> -clear_forward_info(int id)
>>> +clear_forward_info(struct forward_path *path)
>>>    {
>>> -	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
>>> +	memset(path, 0x00, sizeof(struct forward_path));
>>>    }
>>>   >   /* Update forward info */
>>> @@ -84,7 +84,7 @@ spp_forward_update(struct spp_component_info *component)
>>>    		return -1;
>>>    	}
>>>   > -	clear_forward_info(component->component_id);
>>> +	clear_forward_info(path);
>>>   >   	RTE_LOG(INFO, FORWARD,
>>>    			"Component[%d] Start update component. (name = %s, type = %d)\n",
>>>
>>
> 
> 
> 
> 


-- 
Yasufumi Ogawa
NTT Network Service Systems Labs

Re: [spp] [spp 03180] Re: [PATCH] spp_vf: fix initializing bug of forwarder

[Hideyuki Yamashita]

Hello Yasufumi-san,

Thanks for your comment again.
OK, then I will follow your suggestion and remove the function.
I am preparing revised version patch now, so please wait a while.

Thanks.

BR,
Hideyuki Yamashita

> On 2018/08/07 15:31, Hideyuki Yamashita wrote:
> > Hello Yasufumi-san,
> >
> > Thanks for your comments.
> > About your first point, I agree that there is only one point which calls this function. But
> > I still believe it is good to keep this as a function to keep
> > readability of calling side function.
> I do not think this function is more understandable than doing memset. It is just one line and no complexity. If you have any other reasons for, I think it it better to keep the code by removing the meaningless function.
> 
> > So, I will change function name from clear_forward_info() to clear_forward_path().
> >
> > What do you think?
> >
> > BR,
> > Hideyuki Yamashita
> >
> >> On 2018/08/02 17:31, x-fn-spp@sl.ntt-tx.co.jp wrote:
> >>> From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> >>>
> >>> clear_forward_info() always initializes path[0] of g_forward_info[id].
> >>> It is a bug and causes stopping forwarding accidentally if path[0] is
> >>> referenced by spp_forward().
> >>>
> >>> This patch changes initializing as path[info->upd_index] instead of 0.
> >>> It also includes refactoring to be simple by using local vars.
> >> Hideyuki,
> >>
> >> I do not know why claer_forward_info() is requierd because it is called only one time and simply just doing memset(). I think it is not needed to be a function without called from several points. In addition, the name of function is not so good. It does not clear 'forward_info' but 'forward_path'. I think clear_forward_path() is appropriate if you want to make it as a function.
> >>
> >> Thanks,
> >> Yasufumi
> >>>
> >>> Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> >>> Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
> >>> ---
> >>>    src/vf/spp_forward.c | 6 +++---
> >>>    1 file changed, 3 insertions(+), 3 deletions(-)
> >>>
> >>> diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
> >>> index 0a43608..ee1254c 100644
> >>> --- a/src/vf/spp_forward.c
> >>> +++ b/src/vf/spp_forward.c
> >>> @@ -51,9 +51,9 @@ spp_forward_init(void)
> >>>   >   /* Clear info for one element. */
> >>>    static void
> >>> -clear_forward_info(int id)
> >>> +clear_forward_info(struct forward_path *path)
> >>>    {
> >>> -	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
> >>> +	memset(path, 0x00, sizeof(struct forward_path));
> >>>    }
> >>>   >   /* Update forward info */
> >>> @@ -84,7 +84,7 @@ spp_forward_update(struct spp_component_info *component)
> >>>    		return -1;
> >>>    	}
> >>>   > -	clear_forward_info(component->component_id);
> >>> +	clear_forward_info(path);
> >>>   >   	RTE_LOG(INFO, FORWARD,
> >>>    			"Component[%d] Start update component. (name = %s, type = %d)\n",
> >>>
> >>
> >
> >
> >
> > 
> 
> -- Yasufumi Ogawa
> NTT Network Service Systems Labs
> 

[spp] [PATCH v2 1/1] spp_vf: fix initializing bug of forwarder

[x-fn-spp]

From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>

clear_forward_info() always initializes path[0] of g_forward_info[id].
It is a bug and causes stopping forwarding accidentally if path[0] is
referenced by spp_forward().

This patch changes initializing as path[info->upd_index] instead of 0.
It also includes refactoring to be simple by using local vars.

Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
---
 src/vf/spp_forward.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)

diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
index 0a43608..03bda10 100644
--- a/src/vf/spp_forward.c
+++ b/src/vf/spp_forward.c
@@ -49,13 +49,6 @@ spp_forward_init(void)
 	}
 }
 
-/* Clear info for one element. */
-static void
-clear_forward_info(int id)
-{
-	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
-}
-
 /* Update forward info */
 int
 spp_forward_update(struct spp_component_info *component)
@@ -84,7 +77,7 @@ spp_forward_update(struct spp_component_info *component)
 		return -1;
 	}
 
-	clear_forward_info(component->component_id);
+	memset(path, 0x00, sizeof(struct forward_path));
 
 	RTE_LOG(INFO, FORWARD,
 			"Component[%d] Start update component. (name = %s, type = %d)\n",
-- 
2.18.0

Re: [spp] [PATCH v2 1/1] spp_vf: fix initializing bug of forwarder

[Yasufumi Ogawa]

On 2018/08/09 9:38, x-fn-spp@sl.ntt-tx.co.jp wrote:
> From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> 
> clear_forward_info() always initializes path[0] of g_forward_info[id].
> It is a bug and causes stopping forwarding accidentally if path[0] is
> referenced by spp_forward().
> 
> This patch changes initializing as path[info->upd_index] instead of 0.
> It also includes refactoring to be simple by using local vars.
Hi Hideyuki,

Thanks for revising. Could you also update commit message for the change.

Thanks,
Yasufumi
> 
> Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
> ---
>   src/vf/spp_forward.c | 9 +--------
>   1 file changed, 1 insertion(+), 8 deletions(-)
> 
> diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
> index 0a43608..03bda10 100644
> --- a/src/vf/spp_forward.c
> +++ b/src/vf/spp_forward.c
> @@ -49,13 +49,6 @@ spp_forward_init(void)
>   	}
>   }
>   
> -/* Clear info for one element. */
> -static void
> -clear_forward_info(int id)
> -{
> -	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
> -}
> -
>   /* Update forward info */
>   int
>   spp_forward_update(struct spp_component_info *component)
> @@ -84,7 +77,7 @@ spp_forward_update(struct spp_component_info *component)
>   		return -1;
>   	}
>   
> -	clear_forward_info(component->component_id);
> +	memset(path, 0x00, sizeof(struct forward_path));
>   
>   	RTE_LOG(INFO, FORWARD,
>   			"Component[%d] Start update component. (name = %s, type = %d)\n",
> 


-- 
Yasufumi Ogawa
NTT Network Service Systems Labs

[spp] [PATCH v3] spp_vf: fix initializing bug of forwarder

[x-fn-spp]

From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>

clear_forward_info() always initializes path[0] of g_forward_info[id].
It is a bug and causes stopping forwarding accidentally if path[0] is
referenced by spp_forward().

This patch changes initializing as path[info->upd_index] instead of 0.
It also includes following refactoring
- using local vars to be simple
- replace clear_forward_info() by memset

Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
---
 src/vf/spp_forward.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)

diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
index 0a43608..03bda10 100644
--- a/src/vf/spp_forward.c
+++ b/src/vf/spp_forward.c
@@ -49,13 +49,6 @@ spp_forward_init(void)
 	}
 }
 
-/* Clear info for one element. */
-static void
-clear_forward_info(int id)
-{
-	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
-}
-
 /* Update forward info */
 int
 spp_forward_update(struct spp_component_info *component)
@@ -84,7 +77,7 @@ spp_forward_update(struct spp_component_info *component)
 		return -1;
 	}
 
-	clear_forward_info(component->component_id);
+	memset(path, 0x00, sizeof(struct forward_path));
 
 	RTE_LOG(INFO, FORWARD,
 			"Component[%d] Start update component. (name = %s, type = %d)\n",
-- 
2.18.0

Re: [spp] [PATCH v3] spp_vf: fix initializing bug of forwarder

[Yasufumi Ogawa]

> From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> 
> clear_forward_info() always initializes path[0] of g_forward_info[id].
> It is a bug and causes stopping forwarding accidentally if path[0] is
> referenced by spp_forward().
> 
> This patch changes initializing as path[info->upd_index] instead of 0.
> It also includes following refactoring
> - using local vars to be simple
> - replace clear_forward_info() by memset
Thanks!

Acked-by: Yasufumi Ogawa <ogawa.yasufumi@lab.ntt.co.jp>
> 
> Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
> Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>
> ---
>   src/vf/spp_forward.c | 9 +--------
>   1 file changed, 1 insertion(+), 8 deletions(-)
> 
> diff --git a/src/vf/spp_forward.c b/src/vf/spp_forward.c
> index 0a43608..03bda10 100644
> --- a/src/vf/spp_forward.c
> +++ b/src/vf/spp_forward.c
> @@ -49,13 +49,6 @@ spp_forward_init(void)
>   	}
>   }
>   
> -/* Clear info for one element. */
> -static void
> -clear_forward_info(int id)
> -{
> -	memset(&g_forward_info[id].path, 0x00, sizeof(struct forward_path));
> -}
> -
>   /* Update forward info */
>   int
>   spp_forward_update(struct spp_component_info *component)
> @@ -84,7 +77,7 @@ spp_forward_update(struct spp_component_info *component)
>   		return -1;
>   	}
>   
> -	clear_forward_info(component->component_id);
> +	memset(path, 0x00, sizeof(struct forward_path));
>   
>   	RTE_LOG(INFO, FORWARD,
>   			"Component[%d] Start update component. (name = %s, type = %d)\n",
> 

Re: [spp] [PATCH v3] spp_vf: fix initializing bug of forwarder

[Ferruh Yigit]

On 8/9/2018 5:25 AM, Yasufumi Ogawa wrote:
>> From: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
>>
>> clear_forward_info() always initializes path[0] of g_forward_info[id].
>> It is a bug and causes stopping forwarding accidentally if path[0] is
>> referenced by spp_forward().
>>
>> This patch changes initializing as path[info->upd_index] instead of 0.
>> It also includes following refactoring
>> - using local vars to be simple
>> - replace clear_forward_info() by memset
> Thanks!
> 
> Acked-by: Yasufumi Ogawa <ogawa.yasufumi@lab.ntt.co.jp>
>>
>> Signed-off-by: Hideyuki Yamashita <yamashita.hideyuki@po.ntt-tx.co.jp>
>> Signed-off-by: Naoki Takada <takada.naoki@lab.ntt.co.jp>

Applied, thanks.