To at least give someone a start to track down this bug... I played a bit with objdump today, trying to find out what the hell is going wrong.
According to gdb, this is what happens when I try to run an application compiled with the broken binutils: #0 0x800002e0 in ?? () (gdb) I then did 'objdump -D <binary>', and got the following at that address (section <.plt>): 800002e0: 4efb 0171 8000 jmp %pc@(28b6 <_init-0x7fffd9fa>)@(0000000000000000) 800002e6: 25d4 whereas the exact same source compiled with working binutils produces this (different address, but it's also the sixth instruction in the <.plt> section): 80000380: 4efb 0171 0000 jmp %pc@(8000267c <_GLOBAL_OFFSET_TABLE_+0xc>)@(0000000000000000) 80000386: 22fa I think it's clear that the first is totally wrong in that it tries to jump to an address outside the code section. If required, I can send the binaries (both the broken and the working one) and a core file. I'll try to further identify the bug myself, but I don't know much about binutils' internals or the ELF file format, so don't expect much luck... -- The amount of time between slipping on the peel and landing on the pavement is precisely one bananosecond -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
