bug-glibc
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Question regarding glibc-2.3.2 & sparc


From: Kumba
Subject: Question regarding glibc-2.3.2 & sparc
Date: Fri, 09 May 2003 00:03:43 -0400
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.3) Gecko/20030312


        Greetings,

I am trying to investigate a problem compiling glibc-2.3.2 on sparc64 architecture (Sun Blade 100). The machine currently runs Gentoo Linux, and when I try to compile glibc-2.3.2, the configure stage fails because it is unable to determine the size of "long double".

The "config.log" states the reason is an "Architecture mismatch" on several lines of assembly, stating "Requires v9|v9a|v9b; requested architecture is sparclite.". In Gentoo, we've normally discovered this bug appears when building glibc with "-mcpu=ultrasparc/v9" flags. The general solution is to replace ultrasparc/v9 with "v8", and this resolves the main issues.

Something looks to have changed in glibc-2.3.2 that causes the configure script to fail because the test for long double does not like the -mcpu=v8 option, but it works and properly detects long double when -mcpu=v9 is used. The problem with this is glibc fails shortly after on crti.o with more Assembler Mismatches (from crti.S). Thus, we are stuck between a rock and a hard place on this issue.

I am wondering if anyone within the glibc team is aware of this oddity and may have any insight, or recommendations on how best to tackle this. If it's a genuine bug, I can submit a bug report for it, but I thought it better to first seek an inquiry on the matter.

I have attached a tarball containing the "config.log" file from the failed configure script, and a second text file containing the resultant output when -mcpu=ultrasparc (or v9) is used (omitting -mcpu/-mtune doesn't work either, still fails to detect size of long double).


Thanks in advance,

--Kumba

Attachment: glibc-errors.tar.gz
Description: GNU Zip compressed data


reply via email to

[Prev in Thread] Current Thread [Next in Thread]