summaryrefslogtreecommitdiffstats
path: root/net/core
diff options
context:
space:
mode:
authorEric Dumazet <edumazet@google.com>2014-12-03 17:04:39 -0800
committerDavid S. Miller <davem@davemloft.net>2014-12-09 13:40:20 -0500
commit6ffe75eb53564953e75c051e1c28676e1e56f385 (patch)
treedf3ded93cb68f3e739d1c490e8377dfee2fc028a /net/core
parent395eea6ccf2b253f81b4718ffbcae67d36fe2e69 (diff)
downloadop-kernel-dev-6ffe75eb53564953e75c051e1c28676e1e56f385.zip
op-kernel-dev-6ffe75eb53564953e75c051e1c28676e1e56f385.tar.gz
net: avoid two atomic operations in fast clones
Commit ce1a4ea3f125 ("net: avoid one atomic operation in skb_clone()") took the wrong way to save one atomic operation. It is actually possible to avoid two atomic operations, if we do not change skb->fclone values, and only rely on clone_ref content to signal if the clone is available or not. skb_clone() can simply use the fast clone if clone_ref is 1. kfree_skbmem() can avoid the atomic_dec_and_test() if clone_ref is 1. Note that because we usually free the clone before the original skb, this particular attempt is only done for the original skb to have better branch prediction. SKB_FCLONE_FREE is removed. Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Chris Mason <clm@fb.com> Cc: Sabrina Dubroca <sd@queasysnail.net> Cc: Vijay Subramanian <subramanian.vijay@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'net/core')
-rw-r--r--net/core/skbuff.c35
1 files changed, 18 insertions, 17 deletions
diff --git a/net/core/skbuff.c b/net/core/skbuff.c
index 92116df..7a338fb 100644
--- a/net/core/skbuff.c
+++ b/net/core/skbuff.c
@@ -265,7 +265,7 @@ struct sk_buff *__alloc_skb(unsigned int size, gfp_t gfp_mask,
skb->fclone = SKB_FCLONE_ORIG;
atomic_set(&fclones->fclone_ref, 1);
- fclones->skb2.fclone = SKB_FCLONE_FREE;
+ fclones->skb2.fclone = SKB_FCLONE_CLONE;
fclones->skb2.pfmemalloc = pfmemalloc;
}
out:
@@ -541,26 +541,27 @@ static void kfree_skbmem(struct sk_buff *skb)
switch (skb->fclone) {
case SKB_FCLONE_UNAVAILABLE:
kmem_cache_free(skbuff_head_cache, skb);
- break;
+ return;
case SKB_FCLONE_ORIG:
fclones = container_of(skb, struct sk_buff_fclones, skb1);
- if (atomic_dec_and_test(&fclones->fclone_ref))
- kmem_cache_free(skbuff_fclone_cache, fclones);
- break;
-
- case SKB_FCLONE_CLONE:
- fclones = container_of(skb, struct sk_buff_fclones, skb2);
- /* The clone portion is available for
- * fast-cloning again.
+ /* We usually free the clone (TX completion) before original skb
+ * This test would have no chance to be true for the clone,
+ * while here, branch prediction will be good.
*/
- skb->fclone = SKB_FCLONE_FREE;
+ if (atomic_read(&fclones->fclone_ref) == 1)
+ goto fastpath;
+ break;
- if (atomic_dec_and_test(&fclones->fclone_ref))
- kmem_cache_free(skbuff_fclone_cache, fclones);
+ default: /* SKB_FCLONE_CLONE */
+ fclones = container_of(skb, struct sk_buff_fclones, skb2);
break;
}
+ if (!atomic_dec_and_test(&fclones->fclone_ref))
+ return;
+fastpath:
+ kmem_cache_free(skbuff_fclone_cache, fclones);
}
static void skb_release_head_state(struct sk_buff *skb)
@@ -872,15 +873,15 @@ struct sk_buff *skb_clone(struct sk_buff *skb, gfp_t gfp_mask)
struct sk_buff_fclones *fclones = container_of(skb,
struct sk_buff_fclones,
skb1);
- struct sk_buff *n = &fclones->skb2;
+ struct sk_buff *n;
if (skb_orphan_frags(skb, gfp_mask))
return NULL;
if (skb->fclone == SKB_FCLONE_ORIG &&
- n->fclone == SKB_FCLONE_FREE) {
- n->fclone = SKB_FCLONE_CLONE;
- atomic_inc(&fclones->fclone_ref);
+ atomic_read(&fclones->fclone_ref) == 1) {
+ n = &fclones->skb2;
+ atomic_set(&fclones->fclone_ref, 2);
} else {
if (skb_pfmemalloc(skb))
gfp_mask |= __GFP_MEMALLOC;
OpenPOWER on IntegriCloud