Re: what is kdeinit for?

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Sunday 25 July 2004 14:58, Luke Kenneth Casson Leighton wrote:
> the solution is to hack some code into kdeinit to "fake" the
> transition that occurs when one program executes another.
>
> i.e. whenever kdeinit pseudo-executes a program, you "pretend" that
> it has transitioned into that program's security context.
>
> the trouble is that it needs to be special-cased, just in case
> someone does this KDE_IS_PRELINKED thing.

All the tests for KDE_IS_PRELINKED are in kinit.cpp so if you patch kinit, 
it's easy enough to force it one way or the other.

It should also be noted that there are some programs that are only available 
as kdeinit module, I think this is limited to the IO-slaves, but I'm not 
entirely sure.

> first impressions are that the virtual memory used by applications
> fired up from kdeinit is LESS than that used by the application
> on its own.  in some cases by like... 3mbytes.

Loading via kdeinit improves the number of pages that can be shared between 
the processes somewhat. It also used to reduce the time needed for 
relocations drastically, but that issue has been reduced substantially with 
the prelinking support that has gone into the gcc toolchain.

Maybe someone has some recent data on the difference between letting kdeinit 
loads libs vs. using exec, recent patches take some more avantage of kdeinit 
by doing the XFT initialization from the kdeinit master process.

Cheers,
Waldo
- -- 
[email protected]  |   KDE Community World Summit 2004  |  [email protected]
[email protected]  | 21-29 August, Ludwigsburg, Germany |  [email protected]
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBA8CwN4pvrENfboIRAsFwAJ0UXtAUyxZaS9A7K0Kza2bNY87UuACfcGsw
ziCGG243hA0b5RsmcccG55w=
=1c+l
-----END PGP SIGNATURE-----
 
>> Visit http://mail.kde.org/mailman/listinfo/kde-devel#unsub to unsubscribe <<