| To: | jp.martin-flatin@xxxxxxxx |
|---|---|
| Subject: | Re: TCP_RTO_MAX: 120 --> 240 |
| From: | kuznet@xxxxxxxxxxxxx |
| Date: | Fri, 5 May 2000 21:45:27 +0400 (MSK DST) |
| Cc: | davem@xxxxxxxxxx, ak@xxxxxx, netdev@xxxxxxxxxxx |
| In-reply-to: | <200005042229.AAA09286@xxxxxxxxxxxxxxx> from "J.P. Martin-Flatin" at May 5, 0 00:29:14 am |
| Sender: | owner-netdev@xxxxxxxxxxx |
Hello!
> In Linux 2.3.99pre6, TCP_RTO_MAX is set to (120*HZ) in include/net/tcp.h
> (line 297). I think it should be set to (240*HZ).
If you plan to talk to Mars, it is not enough. 8)
If you do not plan to leave Earth, 60 sec is much more than enough. 8)
Another place from the same RFC:
The TCP specification [TCP:1] arbitrarily
assumes a value of 2 minutes for MSL.
So, MSL is an arbitrary value selected to be realistic.
Nowadays, MSL is 60 seconds. RFC1122 is too conservative in most
of places, wrong in another ones and real implemntation fix it,
when it is necessary.
Alexey
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: tx_timeout and timer serialisation, Donald Becker |
|---|---|
| Next by Date: | Q: 2.2.15 default behavior for IPv4 source address selection ICMP/UDP on multi-alias host, Peter Bieringer |
| Previous by Thread: | TCP_RTO_MAX: 120 --> 240, J.P. Martin-Flatin |
| Next by Thread: | Re: TCP_RTO_MAX: 120 --> 240, jamal |
| Indexes: | [Date] [Thread] [Top] [All Lists] |