[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Powerpc IRQ handling breaks strict EABI compliance



Am Don, 2003-02-13 um 16.39 schrieb Sergei Organov:
> Ralf Corsepius <corsepiu at faw.uni-ulm.de> writes:
> [...]
> > > My reading of the output of powerpc-rtems-gcc --print-multi-lib shows
> > > that
> > > there are no libraries for -meabi which means there is no valid way to
> > > link an EABI program.
> > Exactly.
> > 
> > Further _all_ -mno-eabi compiled programs try to link against __eabi.
> > 
> > This is why I think eabi.h should be removed from the powerpc-rtems-clause
> > in config.gcc.
> 
> But __eabi() surprisingly has almost nothing to do with EABI itself. I.e., it
> does a lot of things and only part of them are EABI-specific :-( I begin to
> believe that always calling __eabi() is intentional decision of gcc
> developers.
I am to little knowledgeable on the powerpc to feel able to judge.

>  However, please take a look at my separate post about the issue to
> see why __eabi() is useless for RTEMS anyway.
> 
> [...]
> 
> > However, I don't know if this would be reasonable (To me, the powerpc
> > multilib-ing scheme already is insane ;) )
> 
> I'm second on this. Or I was first, then you are second :) 
I am the author of this powerpc multilib-ing scheme ;)

However, this only is a reflection of what is really used inside RTEMS
powerpc-port - I had tried to minimize it, but ..

The origin of this insanity is the cascades/plethora of defines being
using in the powerpc's cpu.h and the split into old/new-exception
processing. If some of these defines could be eliminated (esp.
old/new-exception-handling) this amount of multilibs could be minimized.

Ralf