| To: | Andi Kleen <ak@xxxxxxx> |
|---|---|
| Subject: | Re: [RFC] Use RCU for tcp_ehash lookup |
| From: | "David S. Miller" <davem@xxxxxxxxxxxxx> |
| Date: | Wed, 1 Sep 2004 22:43:06 -0700 |
| Cc: | vatsa@xxxxxxxxxx, davem@xxxxxxxxxx, netdev@xxxxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, dipankar@xxxxxxxxxx, paulmck@xxxxxxxxxx |
| In-reply-to: | <20040831135419.GA17642@xxxxxxxxxxxxx> |
| Organization: | DaveM Loft Enterprises |
| References: | <20040831125941.GA5534@xxxxxxxxxx> <20040831135419.GA17642@xxxxxxxxxxxxx> |
| Sender: | netdev-bounce@xxxxxxxxxxx |
On Tue, 31 Aug 2004 15:54:20 +0200 Andi Kleen <ak@xxxxxxx> wrote: > And it should also fix the performance problems with > cat /proc/net/tcp on ppc64/ia64 for large hash tables because the rw locks > are gone. Time to convert netstat et al. over the netlink too. > > - I presume that one of the reasons for keeping the hash table so big is to > > keep lock contention low (& to reduce the size of hash chains). If the > > lookup > > is made lock-free, then could the size of the hash table be reduced > > (without > > adversely impacting performance)? > > Definitely worth trying IMHO. The current hash tables are far > too big. I would do that as followon patches though. The hashes are big to make the hash effective, not to help the locking contention. |
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| ||
| Previous by Date: | Re: [RFC] Use RCU for tcp_ehash lookup, David S. Miller |
|---|---|
| Next by Date: | Re: [RFC] Use RCU for tcp_ehash lookup, David S. Miller |
| Previous by Thread: | Re: [RFC] Use RCU for tcp_ehash lookup, Srivatsa Vaddagiri |
| Next by Thread: | Fw: Re: [RFC] Use RCU for tcp_ehash lookup, Srivatsa Vaddagiri |
| Indexes: | [Date] [Thread] [Top] [All Lists] |