| To: | cfriesen@xxxxxxxxxxxxxxxxxx |
|---|---|
| Subject: | Re: anyone ever done multicast AF_UNIX sockets? |
| From: | "David S. Miller" <davem@xxxxxxxxxx> |
| Date: | Mon, 03 Mar 2003 09:56:41 -0800 (PST) |
| Cc: | terje.eggestad@xxxxxxxxx, linux-kernel@xxxxxxxxxxxxxxx, netdev@xxxxxxxxxxx, linux-net@xxxxxxxxxxxxxxx |
| In-reply-to: | <3E6399F1.10303@xxxxxxxxxxxxxxxxxx> |
| References: | <3E638C51.2000904@xxxxxxxxxxxxxxxxxx> <20030303.085504.105424448.davem@xxxxxxxxxx> <3E6399F1.10303@xxxxxxxxxxxxxxxxxx> |
| Sender: | netdev-bounce@xxxxxxxxxxx |
From: Chris Friesen <cfriesen@xxxxxxxxxxxxxxxxxx> Date: Mon, 03 Mar 2003 13:07:45 -0500 Suppose I have a process that waits on UDP packets, the unified local IPC that we're discussing, other unix sockets, and stdin. It's awfully nice if the local IPC can be handled using the same select/poll mechanism as all the other messaging. So use UDP, you still haven't backed up your performance claims. Experiment, set the SO_NO_CHECK socket option to "1" and see if that makes a difference performance wise for local clients. But if performance is "so important", then you shouldn't really be shying away from the shared memory suggestion and nothing is going to top that (it eliminates all the copies, using flat out AF_UNIX over UDP only truly eliminates some header processing, nothing more, the copies are still there with AF_UNIX). |
| Previous by Date: | Re: anyone ever done multicast AF_UNIX sockets?, Chris Friesen |
|---|---|
| Next by Date: | Re: anyone ever done multicast AF_UNIX sockets?, Andi Kleen |
| Previous by Thread: | Re: anyone ever done multicast AF_UNIX sockets?, Chris Friesen |
| Next by Thread: | Re: anyone ever done multicast AF_UNIX sockets?, Chris Friesen |
| Indexes: | [Date] [Thread] [Top] [All Lists] |