[PATCH] [RFC] Emulate "lwsync" to run standard user land on e500 cores

Wolfgang Denk wd at denx.de
Sun Oct 27 21:25:53 EST 2013

Dear Kumar,

In message <6BFC8EB0-1A75-41C3-985A-E3ED14846710 at kernel.crashing.org> you wrote:
> Fair enough
> > 
> > I'm not too worried as long as we warn and account them.
> Than, I'd ask this be under a Kconfig option that is disabled by
> default.  Users should have to explicitly enable this so they know what
> they are doing.

Is this really worth the effort?  Under normal situations (users are
using a user space environment that has been properly buiult for the
processor variant they are using) nobody should ever run into this
situation.  It happens only if you are already doing something wrong -
like using user space that has not been built for an E500 core on such
a machine.

In this situation, it seems more useful to me if a "standard" kernel
just works with a "standard" user space environment, even if this
includes some performance penalty - which actually should be
neglibale.  In my tests (when running standard Debian for PPC on a
E500 only very few programs actually triggered this situation, and
none of them in a time-critical way.  I doubt I would even be able to
measure the performance impact.

Best regards,

Wolfgang Denk

DENX Software Engineering GmbH,     MD: Wolfgang Denk & Detlev Zundel
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd at denx.de
"...all the  good  computer  designs  are  bootlegged;  the  formally
planned  products,  if  they  are built at all, are dogs!" - David E.
Lundstrom, "A Few Good Men From Univac", MIT Press, 1987

More information about the Linuxppc-dev mailing list