Re: linux-next: Tree for August 5

Re: linux-next: Tree for August 5

[Stephen Rothwell]

[-- Attachment #1: Type: text/plain, Size: 832 bytes --]

Hi Boaz,

On Wed, 05 Aug 2009 11:11:20 +0300 Boaz Harrosh <bharrosh@panasas.com> wrote:
>
> On 08/05/2009 10:23 AM, Stephen Rothwell wrote:
> > 
> > This tree fails to build for powerpc allyesconfig (final link problem).
> 
> Above is reported for a long time. Is it related to this:
> http://www.spinics.net/lists/kernel/msg921978.html

No, it is this:

powerpc-linux-ld: TOC section size exceeds 64k

It is powerpc specific and the only way we can think of fixing it
involves stopping building the built-in.o files and linking the entire
kernel in one go at the end.  We just haven't had time (or the energy) to
try to fix it properly while it only affects the final link of the
allyesconfig kernel. 

-- 
Cheers,
Stephen Rothwell                    sfr@canb.auug.org.au
http://www.canb.auug.org.au/~sfr/

[-- Attachment #2: Type: application/pgp-signature, Size: 197 bytes --]

Re: linux-next: Tree for August 5

[Boaz Harrosh]

On 08/05/2009 01:13 PM, Stephen Rothwell wrote:
> Hi Boaz,
> 
> On Wed, 05 Aug 2009 11:11:20 +0300 Boaz Harrosh <bharrosh@panasas.com> wrote:
>>
>> On 08/05/2009 10:23 AM, Stephen Rothwell wrote:
>>>
>>> This tree fails to build for powerpc allyesconfig (final link problem).
>>
>> Above is reported for a long time. Is it related to this:
>> http://www.spinics.net/lists/kernel/msg921978.html
> 
> No, it is this:
> 
> powerpc-linux-ld: TOC section size exceeds 64k
> 
> It is powerpc specific and the only way we can think of fixing it
> involves stopping building the built-in.o files and linking the entire
> kernel in one go at the end.  We just haven't had time (or the energy) to
> try to fix it properly while it only affects the final link of the
> allyesconfig kernel. 
> 

Ha, OK allyesconfig. Last time I attempted an allyesconfig for i386 I got
me the OOM killer attack on my KDE, half way into the final link.

That's why I use allmodconfig for when I need "try to compile everything ++"
I get just as much compilation coverage if not more. And is faster.
It'll not show symbol conflicts if any exist, though.

Thanks
Boaz