[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [patches] eglibc & powerpc-linuxspe-gcc (problem configuring/compiling eglibc svn trunk for mpc8540)
- To: "Joseph S. Myers" <joseph@xxxxxxxxxxxxxxxx>
- Subject: Re: [patches] eglibc & powerpc-linuxspe-gcc (problem configuring/compiling eglibc svn trunk for mpc8540)
- From: Steven Munroe <munroesj@xxxxxxxxxx>
- Date: Mon, 10 Dec 2007 10:24:32 -0600
Joseph S. Myers wrote:
> On Mon, 10 Dec 2007, Steven Munroe wrote:
>
>
>> e500 is booke and is not powerpc. The common subset is --without-fp. So
>> if you want e500 you have to configure --without-fp.
>>
>
> No, as explained a year ago when the E500 support was added to EGLIBC you
> should use --with-fp for full support. And as explained there
> <http://www.eglibc.org/archives/patches/msg00017.html> the ideal
> arrangement would have been to move the existing directories to 6xx/fpu or
> classic/fpu to reflect the existence of two different hardware
> floating-point solutions on Power, and this was not done only to simplify
> merges from FSF glibc. (This ideal arrangement mirrors what was done for
> m68k/ColdFire.)
>
>
Well obviously this did not happen because this is failing to compile
FPSCR instructions from powerpc on e500 which does not have one.
>> I though I told these guys to define their own platform target?
>>
>
> EGLIBC should move towards less and less dependence on the nominal target
> triplet name. This includes, for example, allowing 32-bit hard-float,
> 64-bit hard-float, 32-bit soft-float and 32-bit E500 all to be built with
> a configuration as powerpc-linux-gnu, with the compiler properties tested
> automatically to determine what multilib is in use and glibc configured
> accordingly. Likewise i686-linux-gnu and x86_64-linux-gnu should be
> interchangable. MIPS and m68k/ColdFire are good examples of preconfigure
> scripts doing the right sort of tests for EGLIBC.
>
> For EGLIBC we *did* put conditionals in common files rather than
> artificially duplicate them because of the libc/ports division. In
> particular, installed headers need to be suitable for all multilibs for an
> architecture (and therefore contain appropriate conditionals) so that one
> set of installed headers can be used with a compiler supporting several
> multilibs (e.g. the four listed above).
>
>
Over use of conditionals will just make a hard job imposible. put all
the e500 specific stuff in /powerpc/powerpc32/e500/fpu/