| One thing is the pointer in the request_sock case that I have to | figure out what is best and possibly rework your patches to provide what | I think is best. | I am looking forward to that - if you want me to change single patches, let me know.
I had one more thought regarding timestamping: currently CCID2 uses the old RFC2988 RTT mechanism from Jacobson's old 1988 paper. RFC 4341 says that the sampling should be done at most once per cwnd, but RFC 1323 says that this is insufficient for high-speed networking. And today it is almost impossible to buy a computer /not/ with a Gbps ethernet card. In short - if you change the architecture, is is possible to keep a door open to add a timestamping-based RTT algorithm to CCID2? That has been your suggestion anyway some emails ago and I like the idea. There is further stuff that can be shared - I have some ideas for using the ICSK_RETRANS to replace a lot of code duplication for the RTO timer in CCID2. Would make sense, since in OPEN/PARTOPEN state the retransmit timer is never used (only for Request/Close/CloseReq). - To unsubscribe from this list: send the line "unsubscribe dccp" in the body of a message to [EMAIL PROTECTED] More majordomo info at http://vger.kernel.org/majordomo-info.html
