On Thu, 17 Jan 2008, Joerg Schilling wrote:

> "Garrett D'Amore" <[EMAIL PROTECTED]> wrote:
>
>
>> These days, the interface guarantees are made at the app/library
>> interface boundary, rather than the user/kernel boundary.  And
>> compatibility with Solaris is probably more interesting to most people
>> than SYSV compatibility.  I'm not even sure we are still SYSV ABI
>> compliant.  Anyone else know?
>
> most (if not all) programs compiled on Solaris do not run on SCO UnixWare and 
> vice versa.
> I am not sure which side (or maybe both) do no longer follow the ABI.

Even the SysV ABI (for x86) did not specify the user/kernel boundary, 
apart from a _single_ exception, on which the ABI is explicit about:

_exit() was defined to be "lcall $0x7,0".

That's in:

http://www.sco.com/developers/devspecs/abi386-4.pdf

chapter 3-24, pg. 50. - to quote:

        "To ensure a process has a way to terminate itself, the system
         treats exit() as a special case. The ABI does not specify the
         implementation of other system services."

FrankH.
_______________________________________________
opensolaris-code mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/opensolaris-code

Reply via email to