svn commit: r230207 - in head/sys: netinet sys

Gleb Kurtsou gleb.kurtsou at gmail.com
Thu Jan 19 16:46:27 UTC 2012


On (19/01/2012 18:51), Gleb Smirnoff wrote:
> On Thu, Jan 19, 2012 at 04:38:38PM +0200, Gleb Kurtsou wrote:
> G> On (16/01/2012 09:53), Gleb Smirnoff wrote:
> G> > Author: glebius
> G> > Date: Mon Jan 16 09:53:24 2012
> G> > New Revision: 230207
> G> > URL: http://svn.freebsd.org/changeset/base/230207
> G> > 
> G> > Log:
> G> >   Drop support for SIOCSIFADDR, SIOCSIFNETMASK, SIOCSIFBRDADDR, SIOCSIFDSTADDR
> G> >   ioctl commands.
> G> 
> G> What was the reason for dropping them? "80-ish ioctl" doesn't justify
> G> reducing compatibility with other unix-like OS'es (namely linux).
> 
> The reason is to get code more readable and thus maintainable. You can compare
> in_control() + in_addprefix() in the stable/9 with what we have in head now.
> Which one would you prefer to hack on?

Your point is valid and I'm all for it.

> I wouldn't claim compatibility for the commands that didn't work very well.
> I won't also name Linux, since these commands predate the Linux itself.

I meant that SIOCSIFADDR is default (if not the only) way to set
interface address on linux.

> Do you use them? Or do you know software that use them?

I do and I've seen other examples of using SIOCSIF*ADDR with BSD
specific tweaks. Although I must admit that nowadays the most common way
of configuring interface is to call /sbin/ifconfig.

It's not a big deal for me, I have no problem with replacing them on
FreeBSD. FreeBSD is not even officially supported platform for the
product and I build/test on FreeBSD solely for the purpose of avoiding
linuxisms and platform specific behaviour. Let's hope it won't break for
somebody else :)

Thanks,
Gleb.

> 
> -- 
> Totus tuus, Glebius.


More information about the svn-src-all mailing list