On Fri, Jul 23, 2010 at 4:24 AM, Richard Purdie <[email protected]> wrote: > On Fri, 2010-07-23 at 11:23 +0100, Phil Blundell wrote: >> On Fri, 2010-07-23 at 11:35 +0200, Esben Haabendal wrote: >> > config.sub will generally not accept something like armv7a-* >> >> That's probably true, but it ought to accept armv7a-*-* (and has done >> for about fifteen years, so I expect almost all packages should now have >> picked up that change). I think OE always uses the triple so there >> oughtn't to be any problem in practice. > > With tune-atom.inc active you'd have core2-*-* which I suspect > config.sub will not accept. The point is the names we choose for sets of > optimisation flags are not necessarily going to map well to config.sub > and we shouldn't be crossing what are currently two independent > namespaces...
if we want to be able to build per machine then we have to add put the tools under MACHINE dir in native sysroot but subfamilies can be differentiated like arm5t, armv5teb, armv7 etc thats how far compiler/toolchains can differentiate. There is no problem with two different architectures e.g. mips, arm > > Cheers, > > Richard > > > _______________________________________________ > Openembedded-devel mailing list > [email protected] > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel > _______________________________________________ Openembedded-devel mailing list [email protected] http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
