Subject: Re: where do system calls from userland start
To: Travis H. <solinym@gmail.com>
From: Hubert Feyrer <hubert@feyrer.de>
List: tech-kern
Date: 08/25/2006 03:22:25
  This message is in MIME format.  The first part should be readable text,
  while the remaining parts are likely unreadable without MIME-aware tools.

---1756158181-13040931-1156468945=:30937
Content-Type: TEXT/PLAIN; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: QUOTED-PRINTABLE

On Thu, 24 Aug 2006, Travis H. wrote:
> I was going to teach someone about the kernel, and I wanted to start
> them at the system call entry point, but I've forgotten what file that
> is located in.  I assume it's in the x86-specific code (for the x86
> platform, anyway); anyone know offhand?

See "Ablauf Systemaufruf - von der libc in den Kernel und zur=FCck" at=20
http://www.feyrer.de/BT/ for an example call of fork(2) on a MIPS system -=
=20
this won't be too different on i386. Sorry this is in German language, but=
=20
I guess you'll get the idea. If not, ask.


> Also, I recall reading about NT internals versus Unix, and someone
> mentioned that if you wanted to add a new system call or change the
> meaning of an argument or add an argument, that you got a lot of
> resistance in Unix.  I'm sure there was a good reason, so why exactly
> is that the case?  Does NT have a technical advantage in the way it
> handles system calls?

Well what will all the applications do that expect the old behaviour?


  - Hubert
---1756158181-13040931-1156468945=:30937--