PowerPC target with gcc-4.4.x build failure

[McConnell, Christopher (GE Global Research)]

Andy,

That seemed to be the trick! You are a life saver. I left the
'-fno-stack-protector' in there, not sure that it was needed, but rather
be safe. 

I'll attach the newest, working configuration incase Yann wants to add
another example to the PowerPC configurations, with updated Linux kernel
and GCC, as well as threading, FP and shared cores. 

Thanks again Andy,

Chris

-----Original Message-----
From: Andy Gibbs [mailto:andyg1001@hotmail.co.uk] 
Sent: Thursday, June 17, 2010 3:46 AM
To: McConnell, Christopher (GE Global Research)
Cc: crossgcc@sourceware.org
Subject: Re: PowerPC target with gcc-4.4.x build failure

Chris,

I had the same or a similar problem.  You may want to try setting the
following flags in your crosstool configuration:

CT_LIBC_GLIBC_EXTRA_CFLAGS="-U_FORTIFY_SOURCE -fno-stack-protector"

The first part, "-U_FORTIFY_SOURCE" gets around the an unimplemented
function inside eglibc.  The second part "-fno-stack-protector" I needed
because eglibc/glibc was giving me segmentation faults on the target (an
MPC5200B).  You may not need this of course...

Hope this helps!

Andy



----- Original Message -----
From: "McConnell, Christopher (GE Global Research)" <mcconnel@ge.com>
To: <crossgcc@sourceware.org>
Sent: Wednesday, June 16, 2010 8:19 PM
Subject: RE: PowerPC target with gcc-4.4.x build failure



Sorry, it could help if I remembered the configuration file...

Chris
-----Original Message-----
From: McConnell, Christopher (GE Global Research)
Sent: Wednesday, June 16, 2010 2:17 PM
To: crossgcc@sourceware.org
Cc: McConnell, Christopher (GE Global Research)
Subject: PowerPC target with gcc-4.4.x build failure

All,

I'm looking to get a toolchain going for a target PowerPC 8548 e500v2,
running Linux 2.6.33.2 and a compiler with at least a gcc-4.4.x
(preferably 4.4.3).

[...]

I've been researching for a while now and found other users trying to
get gcc-4.4.x to work for other target architectures, however they seem
to hit a wall as well (some similar to the one above, others are not).

I was hoping someone has come across this problem before and I've some
how missed a patch, or configuration option that should get over this
hurdle.

Any help/thoughts/suggestions/comments will be very appreciated.

Thanks in advance,

Chris McConnell
Edison Engineering Development Program - Software
K1-3A42



------------------------------------------------------------------------
--------


> --
> For unsubscribe information see http://sourceware.org/lists.html#faq 

-------------- next part --------------
A non-text attachment was scrubbed...
Name: powerpc_64_4.4.x.config
Type: application/octet-stream
Size: 9291 bytes
Desc: powerpc_64_4.4.x.config
URL: <http://sourceware.org/pipermail/crossgcc/attachments/20100617/668969a0/attachment.obj>
-------------- next part --------------
--
For unsubscribe information see http://sourceware.org/lists.html#faq