LRO will just stop accumulating when out-of-sequence data arrives. Nothing complicated at all.
Unless the NIC keeps state, it is not always able to know if data is out of sequence.
The LRO timing is not complicated, the packet limit is simply a linearly increasing value that just makes sure that it's always less than or equal to whatever the congestion window happens to be at that moment. It cares not about the exact value.
Yes that is true. However, the congestion window is not known on the receiving end.
LRO will work, and it's the negative attitude of the TOE folks that inspires me to want to help out the LRO folks and ignore the TOE mania altogether.
Yes, LRO will need serious help to be more than a benchmarking tool. What consititutes negative attitude? The point here is that we need an objective approach to the two options. There is no reason that they both can't go in, especially that LSO/LRO are not as simple and non-intruisive as it may appear from their basic description, and so far cannot match the TOE performance. Let the users decide.
- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
