| To: | andrewm@xxxxxxxxxx |
|---|---|
| Subject: | Re: [timers] net/ipv4 |
| From: | "David S. Miller" <davem@xxxxxxxxxx> |
| Date: | Tue, 30 May 2000 21:47:03 -0700 |
| Cc: | kuznet@xxxxxxxxxxxxx, netdev@xxxxxxxxxxx |
| In-reply-to: | <393478D7.7D131923@uow.edu.au> (message from Andrew Morton on Wed, 31 May 2000 02:28:39 +0000) |
| References: | <3933B0B2.50AB5EA1@uow.edu.au> from "Andrew Morton" at May 30, 0 10:14:42 pm <200005301659.UAA12445@ms2.inr.ac.ru> <393478D7.7D131923@uow.edu.au> |
| Sender: | owner-netdev@xxxxxxxxxxx |
Date: Wed, 31 May 2000 02:28:39 +0000 From: Andrew Morton <andrewm@xxxxxxxxxx> > > Of course. I repeat again and again, TCP use _reference_ _counting_. Bear with me Alexey, I'm Australian. Where are the refcounts held? For TCP, they are held in the socket. See all the sock_get/sock_put calls around adding/removing the TCP retransmit timer. Only the last sock_put really free's up the socket in the end, and this can be a timer :-) Later, David S. Miller davem@xxxxxxxxxx |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [timers] net/ipv4, Andrew Morton |
|---|---|
| Next by Date: | recent TCP changes adversive on slow links, Aki M Laukkanen |
| Previous by Thread: | Re: [timers] net/ipv4, Andrew Morton |
| Next by Thread: | Re: [timers] net/ipv4, kuznet |
| Indexes: | [Date] [Thread] [Top] [All Lists] |