[PATCH net-next 13/15] net: Add and use skb_list_del_init().

[PATCH net-next 13/15] net: Add and use skb_list_del_init().

[David Miller]

It documents what is happening, and eliminates the spurious list
pointer poisoning.

In the long term, in order to get proper list head debugging, we
might want to use the list poinson value as the indicator that
an SKB is a singleton and not on a list.

Signed-off-by: David S. Miller <davem@davemloft.net>
---
 include/linux/skbuff.h | 6 ++++++
 net/core/dev.c         | 6 ++----
 net/ipv4/ip_input.c    | 6 +-----
 3 files changed, 9 insertions(+), 9 deletions(-)

diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index c4c9e3f5cd9a..e3a53ca4a9b5 100644
--- a/include/linux/skbuff.h
+++ b/include/linux/skbuff.h
@@ -1344,6 +1344,12 @@ static inline void skb_mark_not_on_list(struct sk_buff *skb)
 	skb->next = NULL;
 }
 
+static inline void skb_list_del_init(struct sk_buff *skb)
+{
+	__list_del_entry(&skb->list);
+	skb_mark_not_on_list(skb);
+}
+
 /**
  *	skb_queue_empty - check if a queue is empty
  *	@list: queue head
diff --git a/net/core/dev.c b/net/core/dev.c
index f76dd7e14dd6..0b2d777e5b9e 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -5295,8 +5295,7 @@ static void __napi_gro_flush_chain(struct napi_struct *napi, u32 index,
 	list_for_each_entry_safe_reverse(skb, p, head, list) {
 		if (flush_old && NAPI_GRO_CB(skb)->age == jiffies)
 			return;
-		list_del(&skb->list);
-		skb_mark_not_on_list(skb);
+		skb_list_del_init(skb);
 		napi_gro_complete(skb);
 		napi->gro_hash[index].count--;
 	}
@@ -5481,8 +5480,7 @@ static enum gro_result dev_gro_receive(struct napi_struct *napi, struct sk_buff
 	ret = NAPI_GRO_CB(skb)->free ? GRO_MERGED_FREE : GRO_MERGED;
 
 	if (pp) {
-		list_del(&pp->list);
-		skb_mark_not_on_list(pp);
+		skb_list_del_init(pp);
 		napi_gro_complete(pp);
 		napi->gro_hash[hash].count--;
 	}
diff --git a/net/ipv4/ip_input.c b/net/ipv4/ip_input.c
index eba7f3883230..35a786c0aaa0 100644
--- a/net/ipv4/ip_input.c
+++ b/net/ipv4/ip_input.c
@@ -531,11 +531,7 @@ static void ip_sublist_rcv_finish(struct list_head *head)
 	struct sk_buff *skb, *next;
 
 	list_for_each_entry_safe(skb, next, head, list) {
-		list_del(&skb->list);
-		/* Handle ip{6}_forward case, as sch_direct_xmit have
-		 * another kind of SKB-list usage (see validate_xmit_skb_list)
-		 */
-		skb_mark_not_on_list(skb);
+		skb_list_del_init(skb);
 		dst_input(skb);
 	}
 }
-- 
2.17.1

Re: [PATCH net-next 13/15] net: Add and use skb_list_del_init().

[Sergei Shtylyov]

Hello!

On 9/8/2018 11:11 PM, David Miller wrote:

> It documents what is happening, and eliminates the spurious list
> pointer poisoning.
> 
> In the long term, in order to get proper list head debugging, we
> might want to use the list poinson value as the indicator that

    Poison?

> an SKB is a singleton and not on a list.
> 
> Signed-off-by: David S. Miller <davem@davemloft.net>

[...]

MBR, Sergei

Re: [PATCH net-next 13/15] net: Add and use skb_list_del_init().

[David Miller]

From: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
Date: Sun, 9 Sep 2018 12:45:50 +0300

> Hello!
> 
> On 9/8/2018 11:11 PM, David Miller wrote:
> 
>> It documents what is happening, and eliminates the spurious list
>> pointer poisoning.
>> In the long term, in order to get proper list head debugging, we
>> might want to use the list poinson value as the indicator that
> 
>    Poison?

I knew you would find something Sergei :-)

I'll fix that up, thanks.