[OE-core] Problem with new image-prelink

[Mark Hatle]

Please try the new version of the prelinker.  I just sent a pull request for it,
but there was a missing typecast on a printf.  This was causing problems on
various systems, and likely could be causing the issue that you observed on ARM
as well.

I did build the latest version for the ARM target, and successfully ran it there
under QEMU.

--Mark

On 9/13/11 4:16 AM, Phil Blundell wrote:
> On Tue, 2011-09-13 at 02:58 +0000, James Limbouris wrote:
>> root at 192:~# gdb prelink
>> <...>
>> Reading symbols from /usr/sbin/prelink...Reading symbols from /usr/sbin/.debug/prelink...done.
>> done.
>> (gdb) set follow-fork-mode child
>> (gdb) run -a
>> Starting program: /usr/sbin/prelink -a
>> [New process 1712]
>> process 1712 is executing new program: /usr/sbin/prelink-rtld
>> [ 2777.370000] Alignment trap: prelink-rtld (1712) PC=0x410f9990 Instr=0xe5922024 Address=0x00000025 FSR 0x001
>>
>> Program received signal SIGBUS, Bus error.
>> [Switching to process 1712]
>> 0x410f9990 in __ctype_b_loc () at ../include/ctype.h:30
>> 30          *tablep = (const uint16_t *) _NL_CURRENT (LC_CTYPE, _NL_CTYPE_CLASS) + 128;
> 
> Just in case it's not obvious, the fact that this address is unaligned
> is the least of your worries.  Even if alignment didn't matter, you
> would just get a segfault instead since 0x25 is never going to be a
> valid pointer.
> 
> I guess you need to investigate where the value in r2 is coming from and
> figure out why it has this bogus value.
> 
> p.
> 
> 
> 
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core