Re: Ethernet over Kernel Sockets

[Lukas Razik <linux@xxxxxxxxxx>]

Hello Mulyadi!

Many Thanks for your fast answer!!!

> Sounds like it's due to context switch latency..or at least the
> interval between your workqueue getting queued into runnable list and
> actually running. Thus, it leads to ~1 ms...a.k.a the HZ
> interval...assuming you're using 1000 MHz

Yes, HZ is 1000.

> May we ask? what preemption model do you use now? try full preemption
> instead of voluntary preemption.

The preemption is off, because we have a Kerrighed (kerrighed.org) 
patched kernel and I believe it doesn't work correctly with preemption 
(at present) but I'll try that.

> Other than that....maybe you need to think about better asynchronous
> communication. Something like AIO or callback based trigger.

I don't know AIO but the hint with callbacks I also got somewhere else 
and I think that would be a nice solution.
The problem is that I know nothing about the functionality of the socket 
callbacks.
Yesterday I only set sock->sk->sk_data_ready to an own function to see 
if it will be called when a message comes in but I don't know how to 
write my own callback at all... :-(
I've many questions - for example:
How can I read the received data from the socket, what should I do after 
reading, what must be locked and when...?
And what has to be done if I want send through a socket without the 
sock_sendmsg() function which can't be called in interrupt context?

Do you know any sources (books or something else) where I can read more 
about socket callbacks? Because I've looked into the kernel sources 
where these callbacks will be set but there are completely no comments...

Regards,
Lukas

--
To unsubscribe from this list: send an email with
"unsubscribe kernelnewbies" to ecartis@xxxxxxxxxxxx
Please read the FAQ at http://kernelnewbies.org/FAQ