This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Status of long-double-128 soft-fp for powerpc32
- From: Steven Munroe <munroesj at us dot ibm dot com>
- To: "Joseph S. Myers" <joseph at codesourcery dot com>, David Edelsohn <dje at watson dot ibm dot com>
- Cc: libc-alpha at sources dot redhat dot com, libc-ports at sources dot redhat dot com
- Date: Wed, 14 Jun 2006 13:38:05 -0500
- Subject: Re: Status of long-double-128 soft-fp for powerpc32
- References: <448F44FE.1060306@us.ibm.com> <Pine.LNX.4.64.0606141121590.1610@digraph.polyomino.org.uk>
Joseph S. Myers wrote:
>On Tue, 13 Jun 2006, Steven Munroe wrote:
>
>>I opened a glibc bugzilla
>><http://sources.redhat.com/bugzilla/show_bug.cgi?id=2749> to track
>>progress on this topic. This bugzilla contains patches which to libc,
>>ports, and gcc which allow powerpc32 to configure --without-fp and
>>-mlong-double-128, make and make check. A number of make check failures
>>remain to be resolved.
>>
>>Hopefully this will allow interrested parties to build gcc-4.1/glibc-2.4
>>soft-fp for themselves and pitch in with the debugging that remains :)
>>
>
>I think we should get the patches to ports and GCC checked in more or less
>as is, as being clearly an improvement on the present situation even
>though more work is needed.
>
I'll submit the ports patch and ask David to submit the gcc patch
>The libc patch may be more risky; perhaps you could create an fenv_libc.h
>for powerpc/nofpu in order to reduce the size of patch required?
>
>
This is a bit tricky because of Ulrich Drepper's concerns stated in
#2423. So eliminating fenv_libc.h where possible may be a better
solution even if it is a bigger patch.