Re: [Qemu-devel] [4188] Fix compiler warnings

[Stefan Weil]

In this special case, this is not possible without some dirty tricks:

The attribute must be assigned to cpu_loop_exit.
cpu_loop_exit would no longer end with an assembler return, so dyngen 
will fail.

Maybe my dummy return is the better solution for the moment
Several architectures used this solution before my patch.

Stefan

Fabrice Bellard schrieb:
> It would be better to use the gcc noreturn function attribute !
>
> Fabrice.
>
> Aurelien Jarno wrote:
> > Revision: 4188
> >           http://svn.sv.gnu.org/viewvc/?view=rev&root=qemu&revision=4188
> > Author:   aurel32
> > Date:     2008-04-11 04:55:17 +0000 (Fri, 11 Apr 2008)
> >
> > Log Message:
> > -----------
> > Fix compiler warnings
> >
> > (Stefan Weil)
> >
> > Modified Paths:
> > --------------
> >     trunk/cpu-exec.c
> >
> > Modified: trunk/cpu-exec.c
> > ===================================================================
> > --- trunk/cpu-exec.c    2008-04-11 04:55:07 UTC (rev 4187)
> > +++ trunk/cpu-exec.c    2008-04-11 04:55:17 UTC (rev 4188)
> > @@ -882,6 +882,8 @@
> >         do it (XXX: use sigsetjmp) */
> >      sigprocmask(SIG_SETMASK, old_set, NULL);
> >      cpu_loop_exit();
> > +    /* never comes here */
> > +    return 1;
> >  }
> >  #elif defined(TARGET_SPARC)
> >  static inline int handle_cpu_signal(unsigned long pc, unsigned long 
> > address,
> > @@ -918,6 +920,8 @@
> >         do it (XXX: use sigsetjmp) */
> >      sigprocmask(SIG_SETMASK, old_set, NULL);
> >      cpu_loop_exit();
> > +    /* never comes here */
> > +    return 1;
> >  }
> >  #elif defined (TARGET_PPC)
> >  static inline int handle_cpu_signal(unsigned long pc, unsigned long 
> > address,