Re: make existing mips files multi-ABI

[Alexandre Oliva <aoliva at redhat dot com>]

On Mar 25, 2003, Andreas Jaeger <aj at suse dot de> wrote:

> So, o32 is broken regarding POSIX - let's fix it for n32/n64.

Yeah, I'm negotiating some more pervasive changes for n32 and n64,
that will take some kernel changes as well.  I understand we can't go
with 64-bit offsets by default on n32, since off_t should be 32-bits,
and I believe o32 and n32 are supposed to have the same data
structures where possible, but I'm considering the possibility of
switching to 64-bit off_t on n32.  If that's not possible, I'd like to
arrange for at least the stat64 data structure in n32 to match exactly
the 64-bit stat structure used by the kernel, and, on n64, arrange for
both stat and stat64 data structures to match exactly what the kernel
uses, obviating the conversion code in n64, and in LFS n32.

>Alexandre Oliva <aoliva at redhat dot com> writes:
>> I'm yet to revisit the need for the fcntl.h changes.

Turns out there are still some kernel issues to be sorted out for
us to be able to define struct flock and flock64 properly.  Currently,
the data structures used by the 32-bit mips kernel seem to be
different from those of the mips64 kernel, breaking o32 programs.

-- 
Alexandre Oliva   Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer                 aoliva at {redhat dot com, gcc.gnu.org}
CS PhD student at IC-Unicamp        oliva at {lsd dot ic dot unicamp dot br, gnu.org}
Free Software Evangelist                Professional serial bug killer