@@ -1356,9 +1356,12 @@ static inline int skb_network_offset(const struct sk_buff *skb)
*
* Various parts of the networking layer expect at least 32 bytes of
* headroom, you should not reduce this.
+ * With RPS, we raised NET_SKB_PAD to 64 so that get_rps_cpus() fetches span
+ * a 64 bytes aligned block to fit modern (>= 64 bytes) cache line sizes
+ * NET_IP_ALIGN(2) + ethernet_header(14) + IP_header(20/40) + ports(8)
*/
#ifndef NET_SKB_PAD
-#define NET_SKB_PAD 32
+#define NET_SKB_PAD 64
#endif
extern int ___pskb_trim(struct sk_buff *skb, unsigned int len);
eth_type_trans() & get_rps_cpus() currently need two 64bytes cache lines in packet to compute rxhash. Increasing NET_SKB_PAD from 32 to 64 reduces the need to one cache line only, and makes RPS faster. NET_IP_ALIGN(2) + ethernet_header(14) + IP_header(20/40) + ports(8) Signed-off-by: Eric Dumazet <eric.dumazet@gmail.com> --- -- 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