🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: gnu.gcc.help
Article #41366

Re: Argh, Big Endian I860

#41366
From: Mike Collison
Date: Mon, 01 Apr 1996 00:00
55 lines
2639 bytes
In article <4je9gs$1mv@minerva.worldbank.org>, ARAJKUMAR@worldbank.org 
says...
>
>tnelson@cherokee.cs.mci.com (Tony Nelson) wrote:
>>Hi All,
>>   After reading through ./config/i860/i860.h I see that gcc 
>> will 'probably' never support a Big Endian I860.  Our Stratus FTX 
>> box happens to be just on of those processors.  I know GCC will 
>> compile on it (Stratus has offered to sell it to us) but I want to > 
compile it on my own.  I've tried several things
>> but no matter what I do when enquire runs it just hangs.
>>
>> enquire -f > tmp-float.h
>>
>> It never returns.  I let it use up about 30sec of cpu time and it 
>> just sat there.  Compile enquire with /usr/ccs/bin/cc works just 
>> fine so I know it's something I'm doing wrong.
>>Any help would be greatly appreciated.
>>Thanks
>>Tony Nelson
>Hi
>
>    I have an Okistation-7300 Model-20 which also uses the 
>i860 CPU running OKIX-1.1 a port of the SVR4 for the i-860
>class of CPU and have been trying to get gcc compiled on
>it and ran exactly into the same problem with 'enquire -f'
>not RETURNING at all.
>
>    For some reason, GNU folks have decided to stop supporting
>the i860 Class of CPUS and for those of us, lost in oblivion
>since the manufacture went belly-up, we have not way to
>keep up with the newer version of gcc.
>
>    I would greatly appreciate if any one could shed any light
>on this. Since OKIDATA still manufactures and sells these products
>in Japan, I would presume that somewhere somebody knows how to
>fix this, but just wont let the world know about it.

I believe the problem is that although GCC can be configured to be
big or little-endian it is an all or nothing proposition. If I remember
correctly, when you put the i860 in big-endian mode it only refers to
values in memory. In other words values in registers are still little-
endian. GCC has an implicit assumption that the "endianess" of values
in memory and in registers is the same. I suspect this is the problem
you were seeing.

>
>Regards
>Joseph Rajkumar
>
>PS:- Please email any information to
><A HREF="mailto:ARAJKUMAR@worldbank.org"> me </A>
>
>
>
Technical Details
Message-ID: <4jphin$ffe@shore.shore.net>
Path: arc1-ca.pugleaf.net!unknown!not-for-mail
References: <u6d95y2m9s.fsf@cherokee.cs.mci.com> <4je9gs$1mv@minerva.worldbank.org>

Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 11318c,ea83813eb7063c75
X-Google-Attributes: gid11318c,public
From: collison@iced.com (Mike Collison)
Subject: Re: Argh, Big Endian I860
Date: 1996/04/01
Message-ID: <4jphin$ffe@shore.shore.net>
X-Deja-AN: 145293112
references: <u6d95y2m9s.fsf@cherokee.cs.mci.com> <4je9gs$1mv@minerva.worldbank.org>
organization: Integrated Computing Engines, Inc.
mime-version: 1.0
newsgroups: gnu.gcc.help