On Tuesday, August 24, 2010 19:09:43 Gavin Lambert wrote:
> Quoth David Howells:
> > Stub out vm_get_page_prot() if there's no MMU.
>
> [...]
>
> > in the fbmem video driver, but the function doesn't exist on NOMMU,
> > resulting in an undefined symbol at link time.
>
> [...]
>
> > +#ifdef CONFIG_MMU
> >
> > pgprot_t vm_get_page_prot(unsigned long vm_flags);
> >
> > +#else
> > +static inline pgprot_t vm_get_page_prot(unsigned long vm_flags)
> > +{
> > + return __pgprot(0);
> > +}
> > +#endif
>
> Wouldn't it be better to define this in the .c for !CONFIG_MMU, rather than
> stub it out in the headers? After all, it's possible that some arches
> (eg. Blackfin, since it has an MPU) might want to actually implement it.even in the MPU case, i think we'd still want a header change. #if defined(CONFIG_MMU) || defined(CONFIG_MPU) pgprot_t vm_get_page_prot(unsigned long vm_flags); #else /* stub */ #endif none of the MPU code today uses this vm field, so there'd be a bit more work before we could handle it and worry about the function. which is why i think we should go with this change today, and worry about the rest tomorrow. -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ uClinux-dev mailing list uClinux-dev@uclinux.org http://mailman.uclinux.org/mailman/listinfo/uclinux-dev This message was resent by uclinux-dev@uclinux.org To unsubscribe see: http://mailman.uclinux.org/mailman/options/uclinux-dev
