AMD64 version of GNAT Ada compiler broken due to libthr

[John Marino]

On 10/30/2011 5:36 PM, Daniel Eischen wrote:
> On Sun, 30 Oct 2011, Gerald Pfeifer wrote:
> 
>> On the last day of 2010, Daniel Eischen wrote:
>>> I've got FSF paperwork on file, specifically to submit my original
>>> FreeBSD and VxWorks GNAT ports to AdaCore (which they then upstreamed
>>> to GCC).  It's been a few years since I submitted the paperwork,
>>> however, and I'm not sure if they require resubmittal at periodic
>>> intervals.  It may be possible for you to explain your changes to
>>> me, without me looking at your original code or changes.
>>
>> This is generally not necessary.  Your paperwork with the FSF holds
>> without a need to renew, except that if you changed employers and
>> require an employer disclaimer that might need to be sent for the
>> new employer.
>>
>> Not sure whether the concrete question here is still open of course,
>> but I'll also be happy to see how I can help in case.
> 
> I lost the original thread; someone had, or was trying to, port
> lang/gnat to AMD64 (that someone not being Mark Murray).
> 

No, I launched the initial thread.  The problem was actually stack
permissions.  It's been solved and the resultant lang/gnat-aux passes
all tests perfectly on x86 and AMD64 now.

The issue is that the patch that allowed this should go in gcc codebase
so all versions of gcc get it.

John