Patchwork [RFC,v7,11/19] Use callback to deal with skb_release_data() specially.

login
register
mail settings
Submitter Eric Dumazet
Date June 5, 2010, 2:56 p.m.
Message ID <1275749773.5238.244.camel@edumazet-laptop>
Download mbox | patch
Permalink /patch/54777/
State RFC
Delegated to: David Miller
Headers show

Comments

Eric Dumazet - June 5, 2010, 2:56 p.m.
Le samedi 05 juin 2010 à 18:14 +0800, xiaohui.xin@intel.com a écrit :
> From: Xin Xiaohui <xiaohui.xin@intel.com>
> 
> If buffer is external, then use the callback to destruct
> buffers.
> 
> Signed-off-by: Xin Xiaohui <xiaohui.xin@intel.com>
> Signed-off-by: Zhao Yu <yzhao81new@gmail.com>
> Reviewed-by: Jeff Dike <jdike@linux.intel.com>
> ---
>  net/core/skbuff.c |   11 +++++++++++
>  1 files changed, 11 insertions(+), 0 deletions(-)
> 
> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
> index 37587f0..418457c 100644
> --- a/net/core/skbuff.c
> +++ b/net/core/skbuff.c
> @@ -385,6 +385,11 @@ static void skb_clone_fraglist(struct sk_buff *skb)
>  
>  static void skb_release_data(struct sk_buff *skb)
>  {
> +	/* check if the skb has external buffers, we have use destructor_arg
> +	 * here to indicate
> +	 */
> +	struct skb_external_page *ext_page = skb_shinfo(skb)->destructor_arg;
> +

Oh well. This is v7 of your series, and nobody complained yet ?

This is a new cache miss on a _critical_ path.


>  	if (!skb->cloned ||
>  	    !atomic_sub_return(skb->nohdr ? (1 << SKB_DATAREF_SHIFT) + 1 : 1,
>  			       &skb_shinfo(skb)->dataref)) {
> @@ -397,6 +402,12 @@ static void skb_release_data(struct sk_buff *skb)
>  		if (skb_has_frags(skb))
>  			skb_drop_fraglist(skb);
>  
> +		/* if the skb has external buffers, use destructor here,
> +		 * since after that skb->head will be kfree, in case skb->head
> +		 * from external buffer cannot use kfree to destroy.
> +		 */

Why not deferring here the access to skb_shinfo(skb)->destructor_arg ?

> +		if (dev_is_mpassthru(skb->dev) && ext_page && ext_page->dtor)
> +			ext_page->dtor(ext_page);
>  		kfree(skb->head);
>  	}
>  }

if (dev_is_mpassthru(skb->dev)) {
	struct skb_external_page *ext_page =
				skb_shinfo(skb)->destructor_arg;
	if (ext_page && ext_page->dtor)
		ext_page->dtor(ext_page);
}

destructor_arg should me moved before frags[] if you really want to use it.




--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Xin, Xiaohui - June 9, 2010, 7:30 a.m.
>-----Original Message-----
>From: Eric Dumazet [mailto:eric.dumazet@gmail.com]
>Sent: Saturday, June 05, 2010 10:56 PM
>To: Xin, Xiaohui
>Cc: netdev@vger.kernel.org; kvm@vger.kernel.org; linux-kernel@vger.kernel.org;
>mst@redhat.com; mingo@elte.hu; davem@davemloft.net; herbert@gondor.apana.org.au;
>jdike@linux.intel.com
>Subject: Re: [RFC PATCH v7 11/19] Use callback to deal with skb_release_data() specially.
>
>Le samedi 05 juin 2010 à 18:14 +0800, xiaohui.xin@intel.com a écrit :
>> From: Xin Xiaohui <xiaohui.xin@intel.com>
>>
>> If buffer is external, then use the callback to destruct
>> buffers.
>>
>> Signed-off-by: Xin Xiaohui <xiaohui.xin@intel.com>
>> Signed-off-by: Zhao Yu <yzhao81new@gmail.com>
>> Reviewed-by: Jeff Dike <jdike@linux.intel.com>
>> ---
>>  net/core/skbuff.c |   11 +++++++++++
>>  1 files changed, 11 insertions(+), 0 deletions(-)
>>
>> diff --git a/net/core/skbuff.c b/net/core/skbuff.c
>> index 37587f0..418457c 100644
>> --- a/net/core/skbuff.c
>> +++ b/net/core/skbuff.c
>> @@ -385,6 +385,11 @@ static void skb_clone_fraglist(struct sk_buff *skb)
>>
>>  static void skb_release_data(struct sk_buff *skb)
>>  {
>> +	/* check if the skb has external buffers, we have use destructor_arg
>> +	 * here to indicate
>> +	 */
>> +	struct skb_external_page *ext_page = skb_shinfo(skb)->destructor_arg;
>> +
>
>Oh well. This is v7 of your series, and nobody complained yet ?
>
>This is a new cache miss on a _critical_ path.

Ok, I would remove the declaration here to avoid the new cache miss.

>
>
>>  	if (!skb->cloned ||
>>  	    !atomic_sub_return(skb->nohdr ? (1 << SKB_DATAREF_SHIFT) + 1 : 1,
>>  			       &skb_shinfo(skb)->dataref)) {
>> @@ -397,6 +402,12 @@ static void skb_release_data(struct sk_buff *skb)
>>  		if (skb_has_frags(skb))
>>  			skb_drop_fraglist(skb);
>>
>> +		/* if the skb has external buffers, use destructor here,
>> +		 * since after that skb->head will be kfree, in case skb->head
>> +		 * from external buffer cannot use kfree to destroy.
>> +		 */
>
>Why not deferring here the access to skb_shinfo(skb)->destructor_arg ?

And references skb_shinfo(skb)->destructor_arg here.

>
>> +		if (dev_is_mpassthru(skb->dev) && ext_page && ext_page->dtor)
>> +			ext_page->dtor(ext_page);
>>  		kfree(skb->head);
>>  	}
>>  }
>
>if (dev_is_mpassthru(skb->dev)) {
>	struct skb_external_page *ext_page =
>				skb_shinfo(skb)->destructor_arg;
>	if (ext_page && ext_page->dtor)
>		ext_page->dtor(ext_page);
>}
>
>destructor_arg should me moved before frags[] if you really want to use it.

Thanks for the patch. But why destructor_arg before frags[] is better than after frags[]?
skb_release_data() will reference both of them........

>
>diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
>index bf243fc..b136d90 100644
>--- a/include/linux/skbuff.h
>+++ b/include/linux/skbuff.h
>@@ -202,10 +202,11 @@ struct skb_shared_info {
>         */
>        atomic_t        dataref;
>
>-       skb_frag_t      frags[MAX_SKB_FRAGS];
>        /* Intermediate layers must ensure that destructor_arg
>         * remains valid until skb destructor */
>        void *          destructor_arg;
>+
>+       skb_frag_t      frags[MAX_SKB_FRAGS];
> };
>
> /* We divide dataref into two halves.  The higher 16 bits hold references
>
>

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Patch

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index bf243fc..b136d90 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -202,10 +202,11 @@  struct skb_shared_info {
         */
        atomic_t        dataref;
 
-       skb_frag_t      frags[MAX_SKB_FRAGS];
        /* Intermediate layers must ensure that destructor_arg
         * remains valid until skb destructor */
        void *          destructor_arg;
+
+       skb_frag_t      frags[MAX_SKB_FRAGS];
 };
 
 /* We divide dataref into two halves.  The higher 16 bits hold references