This is the mail archive of the gcc-help@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: AW: EABI


On 10/17/2013 09:23 AM, Warlich, Christof wrote:

> As I tried to point out in my post, I'm not so much interested in
> (E)ABI details themselves, but in the effect it has when passing
> different ABI specifications for _different_ target architectures to
> the configure scripts of binutils, the C library being used,
> e.g. glibc and most notable and relevant for this mailing list, gcc,
> i.e.  the differences for
> 
> ./configure --target=xxx-mytoolchain-linux-gnu
> ./configure --target=xxx-mytoolchain-linux-gnueabi
> ./configure --target=xxx-mytoolchain-linux-eabi
> 
> As much as I could find out so far, it seems like -gnueabi only
> makes a difference for xxx=arm and -eabi makes a difference for
> xxx=powerpc and xxx=arm, but this is more a guess that I concluded
> from looking at the related configure and confif.sub scripts for gcc
> (and binutils and glibc).
> 
> So I'd like to confirm if I'm right so far, being interested in
> i*86, mips*, powerpc and arm, and what might be good naming
> convention for those architectures where no differences are to be
> expected for the three cases mentioned above.

There aren't any hard and fast rules because system ABIs aren't
something that is determined by GCC: they are designed by system
architects.  They might choose to define all manner of variants in all
manner of ways.  All we in GCC can do is implement them.  So, if you
want to ask a specific question about an ABI we'll try to answer it,
but that's all we can do.

Andrew.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]