Howard,

The package uses the iBCS/Unix98 major device number that is already
reserved for that purpose.  It doesn't consume a major device number
because that device number is already consumed for iBCS/Unix98.
Solaris uses this same major number (both Sparc and Intel) and the
Linux kernel built for sparc architectures uses this device number
for iBCS and Solaris compatibility as well.  A device number is not
consumed, because without a STREAMS socksys behind the major, the
major should still not be reused for other things...

--brian

On Wed, 26 Mar 2003, Howard Selover wrote:

> 
>    Hello Brian,
>    Thank  you  for  your  reply.  Since it sits on top of TCP providing a
>    streams based interface it does not really matter if it is optional or
>    not other than the major number consumption.
>    Howard
>    Brian F. G. Bidulock wrote:
> 
> Howard,
> 
> The strinet package provides a STREAMS (TLI XTI/XNS) interface to the
> underlying Linux sockets implementation.  That is, opening up a TCP/IP
> stream opens a TCP/IP socket internal to the kernel.  The Linux native
> TCP implementation does not know the difference between the strinet
> package opening a socket and the user program opening a socket.
> 
> So, in short, the STREAMS implementation is completely compatible with
> the underlying Linux NET4 implementation which it uses rather than
> replaces.
> 
> As to whether strinet compiles as part of the LiS base package (i.e.
> LiS registered drivers) or not is a different question.  It is currently
> written to compile as a loadable module (streams-inet.o of about 28k
> in size) that is somewhat smaller than streams-ldl.o (~38k) and much
> smaller than the streams base object streams.o (~180k).  As the iBCS
> device numbers (major and minor) are already reserved for socksys in
> Linux, there would be no problem compiling it as a loadable module.
> Then if one never opens the devices, the module would never load. So,
> I suppose that would make it "optional".
> 
> --brian
> 
> 
> On Wed, 26 Mar 2003, Howard Selover wrote:
> 
>   
> 
> Hello David,
> 
> I saw the TCP package mails in the past few days.  I am curious if this 
> will be an optional package.  If the TCP package is not optional, is the 
> TCP stack provided in the Linux kernel still useable or does everything 
> have to use the streams implementation?
> 
> Howard
> 
> -- 
> Howard Selover III
> Principal Engineer
> Chief Architect's Office
> Ulticom, Inc.
> 1020 Briggs Road
> Mount Laurel, NJ 08054
> 
> Direct: +1-856-787-2739
> Mobile: +1-856-495-4181
> Fax:    +1-856-866-2033
> Email:  [EMAIL PROTECTED]
> Web:    [2]www.ulticom.com
> 
> 
> 
> _______________________________________________
> Linux-streams mailing list
> [EMAIL PROTECTED]
> [4]http://gsyc.escet.urjc.es/mailman/listinfo/linux-streams
> 
> 
> 
> 
> --
> Howard Selover III
> Principal Engineer
> Chief Architect's Office
> Ulticom, Inc.
> 1020 Briggs Road
> Mount Laurel, NJ 08054
> 
> Direct: +1-856-787-2739
> Mobile: +1-856-495-4181
> Fax:    +1-856-866-2033
> Email:  [EMAIL PROTECTED]
> Web:    [6]www.ulticom.com
> 
> References
> 
>    1. mailto:[EMAIL PROTECTED]
>    2. http://www.ulticom.com/
>    3. mailto:[EMAIL PROTECTED]
>    4. http://gsyc.escet.urjc.es/mailman/listinfo/linux-streams
>    5. mailto:[EMAIL PROTECTED]
>    6. http://www.ulticom.com/

-- 
Brian F. G. Bidulock    � The reasonable man adapts himself to the �
[EMAIL PROTECTED]    � world; the unreasonable one persists in  �
http://www.openss7.org/ � trying  to adapt the  world  to himself. �
                        � Therefore  all  progress  depends on the �
                        � unreasonable man. -- George Bernard Shaw �

_______________________________________________
Linux-streams mailing list
[EMAIL PROTECTED]
http://gsyc.escet.urjc.es/mailman/listinfo/linux-streams

Reply via email to