Re: Measured overhead of timer interrupts

Andi Kleen (ak@muc.de)
Thu, 22 Jul 1999 22:36:28 +0200


On Thu, Jul 22, 1999 at 06:32:49PM +0200, kuznet@ms2.inr.ac.ru wrote:
> Hello!
>
> > Actually sysctl allows converting everything, if you just put in a pointer to
> > your own handling function [of course for ms it should be put into sysctl.c
> > as a generic mechanism]. proc_dointvec is just a default.
>
> He-he... Looking at do_proc_dointvec was enough for me to forget about
> idea to create our own private functions 8)8)

Hehe:)

It is actually not that bad, you can simply wrap it, like Pedro did in
net/ipv4/sysctl_net_ipv4.c:ipv4_sysctl_forward()

>
> BTW do you know that we used sysctl absolutely incorrectly?
> No sys/net variables define "strategy" method, hence sysctl() does not work
> at all. To be more exact, it does not any checks or transformations.
> Grr... What sad discovery... :( I make such discoveries not every day 8)

Ooops.

Seems nobody uses the system call...

Ok, so proc_dointvec_jiffies is broken by design and the proposal above
is wrong.

Fortunately most net variables don't need a transformation/state setup,
so only proc_dointvec_jiffies and rtcache_flush needs fixing I think, or
did I miss something?

-Andi

-- 
This is like TV. I don't like TV.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/