On 10/27/05, Ian McDonald <[EMAIL PROTECTED]> wrote: > On 27/10/05, Yoshifumi Nishida <[EMAIL PROTECTED]> wrote: > > > > Hi, > > > > I'm trying to do interoperability tests between netbsd and linux > > implementation. > > Great!!
Indeed :-) > > I found that two issues so far. One issue is that linux implementation > > doesn't > > have feature negotiation function yet, but netbsd implementation presumes it > > and wait for a confirm packet forever. I will work on this issue later. > > Yes we are well aware of this. I started work on a userland library for feature negotiation that later will be just plugged in the kernel, idea is to be able to write stress tests without the need of reboots/module loading etc, but I've been sidetracked by way too many things lately, hope to get back to this soon. > > Second issue is that CCID3 receiver on linux sends ACKs with time stamp > > option > > and does not send ACKs with elapsed time option. (As far as I tested..) > > It seems to me that CCID3 receiver must use elapsed time option or can use > > time > > stamp echo option in case that CCID3 sender uses time stamp option. > > (According to section 8.2.in draft-ietf-dccp-ccid3-11.txt) > > I'm aware that our time options aren't perfect as well and will > investigate or leave to somebody else to investigate. There were a > couple of problems with time options - this being one and the other > that the timestamps were actually incorrect. I believe we have fixed > the latter but I would be interested to see if you agree on that. I've talked with Thomas Gleixner about using the still out of the tree ktimers infrastructure to solve most of the problems in our current timer based code, like poor granularity, "epoch" based struct timeval timestamps getting out of sync when the admin changes the machine clock, etc, interested parties can take a look at all the goodies and issues related to merging the ktimers patch in these nice articles: "A new approach to kernel timers" http://lwn.net/Articles/152436/ "On the merging of ktimers" http://lwn.net/Articles/156325/ - Arnaldo - 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
