Hi, I'm trying to sent packets through different interfaces handled by the same network driver (Intel e1000e). The interfaces are located on the same card as well as on different cards. Each interface has its own instance of my additional custom driver designed to just send packets. The custom driver just allocs the packets by: skb = dev_alloc_skb(packet->len); and then sends them by: result = dev->hard_start_xmit(skb,dev); The problem is that I cant obtain full bandwidth offered by the 1GB interfaces. For one interface I get: 120435948 bytes/sec. For two interfaces I get: 61080233 bytes/sec and 60515294 bytes/sec. For three interfaces I get: 28564020 bytes/sec, 27111184 bytes/sec, 27118907 bytes/sec. And, if I use raw sockets (not the custom pktgen-like driver) to send packets, I'm able to utilize full bandwidth (120*10^6bps) on all four interfaces. What can be the problem? It looks like some bottleneck inside network subsystem, right under hard_start_xmit? IRQs? Locking? -- Regards, Gallus -- To unsubscribe from this list: send the line "unsubscribe linux-net" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
