[uClibc] ARM shared libs segfault

Wen Yang weny at promise.com
Tue May 25 21:03:59 UTC 2004


Hi,

I got the same segfault problem. Any luck of trying the binutil 2.15 or
any other fix?

Thank you.

Wen

-----Original Message-----
From: uclibc-bounces at uclibc.org [mailto:uclibc-bounces at uclibc.org] On
Behalf Of Erik Andersen
Sent: Friday, May 21, 2004 11:32 PM
To: David Poole
Cc: uclibc at uclibc.org
Subject: Re: Re[2]: [uClibc] ARM shared libs segfault

On Fri May 21, 2004 at 07:46:44AM -0600, David Poole wrote:
> Sorry. I should have made the horrible nature of my hack clear.  That 
> was code I wrote to discover where segfault occurred but it doesn't 
> address the linker problem that's causing the bad executable that 
> precipitates the segfault.
> 
> >No, it is not in the latest snapshot and it will not be in any
> >future snapshots.  This is a hack, demonstrating the nature of
> >the problem, not a solution fixing the problem.  The Right
> >Thing(tm) would be to find the linker bug and fix it.  It may
> >very well be this bug is fixed already, by someone, somewhere.
> 
> This bug only exists in toolchain's binutils-uclibc-001-debian.patch. 
> Where did this patch come from? I couldn't find any other binutils 
> snapshots or releases that had both the count and pc_count fields in 
> the elf32_arm_relocs_copied structure.

binutils-uclibc-001-debian.patch was taken from binutils in Debian
testing.

> >Adding random hacks working around compiler or linker bugs is
> >generally a bad idea.
> 
> Would it be worthwhile for me to take a whack at putting binutils 2.15

> (latest release) into the toolchain script?

Sure, worth a shot,

 -Erik

--
Erik B. Andersen             http://codepoet-consulting.com/
--This message was written using 73% post-consumer electrons--



More information about the uClibc mailing list