Dear Albert ARIBAUD,
In message <20120901182841.3ae8b35f@lilith> you wrote:
>
> Each architecture has its own capabilites regarding native misaligned
> accesses... This is why I consider that as a general rule U-Boot should
> always align its data properly, because (hopefully) all architectures
> can do aligned native accesses; OTOH, if we accept misaligned code on
> the grounds that 'it works on such and suh arches' or that 'any normal
> arch should be able to handle misaligned accesses some way' or 'no one
> in their right mind would physically forbit misaligned accesses', then
> we're just giving Murphy a chance to kick us at some point.
>
> Consider this an application of Postel's principle: we liberally
> accept architectures that maybe allow misaligned accesses and maybe
> handle them well; and we conservatively do not do such accesses unless
> we have no other choice.
Full agreement.
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: w...@denx.de
Even historians fail to learn from history -- they repeat the same
mistakes.
-- John Gill, "Patterns of Force", stardate 2534.7
_______________________________________________
U-Boot mailing list
U-Boot@lists.denx.de
http://lists.denx.de/mailman/listinfo/u-boot
