Re: missing linker symbols w/ binutils-2.17

[jbbachky at aim dot com]

Nick,
Thank you for making the time to investigate my problem I have rebuilt 
with the 12/06 snapshot of binutils and do not see any difference - the 
problem is still there. So I will work on creating a subset of my code 
to demonstrate the problem. Assuming I get there... how should I 
send/submit it? Attached tarball?

Thanks and TIA, John.

-----Original Message-----
From: nickc@redhat.com
To: jbbachky@aim.com
Cc: binutils@sourceware.org
Sent: Tue, 5 Dec 2006 11:03 AM
Subject: Re: missing linker symbols w/ binutils-2.17

Hi John,

> The linker script which follows has been working fine with 
binutils-2.15 > on a number of architectures. Now that I'm attempting 
to upgrade to > 2.17, I find that one of the linker-script-defined 
symbols does does NOT > get added to the symbol table.

> The
> "lexunix_prefetch_1_start" symbol simply doesn't make it into the > 
resulting symbol file.

Can you provide a way to reproduce this problem ?

I tried to reproduce it using the script you supplied, but the symbol 
appeared in the symbol table:

 % cp hello.o lexthread.o
 % cp hello.o lexunix.o
 % arm-ld hello.o -T johns_script.t
  warning: cannot find entry symbol _start; defaulting to 
0000000000008000
 % nm a.out
 00000000005a0054 T lexunix_prefetch1_end
 00000000005a0054 A lexunix_prefetch1_start
 00000000005a0054 T lexunix_prefetch2_end
 00000000005a0054 T lexunix_prefetch2_start
 00000000005a0054 T lexunix_prefetch_end
 00000000005a0054 A lexunix_prefetch_start
 00000000005a0054 t my_variable
 00000000005a0078 t my_variable
 00000000005a009c t my_variable

This was using the current binutils mainline sources, rather than the 
2.17 release sources, so it is possible that there was a bug that has 
now been fixed.

Cheers
 Nick
________________________________________________________________________
Check Out the new free AIM(R) Mail -- 2 GB of storage and 
industry-leading spam and email virus protection.