Re: continuing egcs-1.1 problems

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

At 17:24 02.11.98 , bacon@aerodyne.com wrote:
>On  2 Nov, Franz Sirl wrote:
>> At 02:56 02.11.98 , David Edelsohn wrote:
>>>>>>>> Elgin Lee writes:
>>>
>>>Elgin> I'm hazy on the workings of the code, but on the surface this looks
>>>Elgin> like a comparison-reversal bug that is fixed by the patch.  (But why
>>>Elgin> does this work on the RS6000/AIX?  Perhaps because r4 and r8 are
equal
>>>Elgin> there on the RS6000/AIX?)
>>>
>>>	AIX does not use this code whatsoever.  This trampoline code only
>>>is used by SVR4, not AIX PowerOpen ABI.
>> 
>> David, he's talking the appended patch. I wonder why it still is in. The
>> first part is definitely wrong (I thought Gary Thomas already agreed to
>> that?) and the second part breaks the SVR-ABI spec AFAIK (it is meant to be
>> a speedup patch). I assume Gary is somekind of dead-locked here, removing
>> the ABI patch will produce a correct egcs, but break older installations
>> compiled with the patch then.
>> Btw, on glibc-2.0.99 the test-program compiles and runs just fine, I don't
>> think it is necessary to spend more thoughts on this.
>> 
>
>Are you then saying that we need to recompile glibc using a compiler
>without the patch?  The reason I ask is that the patch comes from the
>offical distributions of mklinux DR3 and LinuxPPC R4.  This would mean
>that we need to back out this patch, and rebuild glibc and possibly
>everything else which may have been built with a version of egcs-1.0.2
>that contained the patch.
>
>I have no quarrel with how this problem gets solved.  I have no
>convictions about where the problem is located.  But the problem exists
>and needs to be corrected.  If that means rebuilding all of user space
>in order to keep the libraries and compiler consistent...fine.  Let's
>do it.
>
>I hope I don't sound too harsh.  I simply want this problem solved, and
>I'm willing to do whatever it takes to see it done.  Let's identify
>the correct problem and apply the correct solution.  
>
>At this juncture, it looks as though glibc is broken because it was
>built with a broken compiler.  The proper solution seems to be that we
>should build a new compiler, without the patch from Gary, and use that
>to rebuild glibc.
>
>Is that the general consensus?  Should we rebuild the libraries with
>the unpatched compiler?  How extensive are the incompatibilities likely
>to be?  Should everything be rebuilt, or will rebuilding glibc be
>sufficient?

I think you should really talk to Gary Thomas <gdt@opengroup.org> about
that. I'm on glibc-2.1 snapshots since begin of the year and no longer
involved in the glibc-0.961212 stuff. I checked ftp.linuxppc.org and found
a newer SRPM (from Gary?) in the contrib dir which no longer contains the
trampoline patch, so eventually he has made a decision on this already.

Franz.