[ELDK] installing 5.6

K Richard Pixley rpixley at graphitesystems.com
Tue Jan 27 22:47:54 CET 2015

On 20150127 09:01, K Richard Pixley wrote:
> On 1/26/15 23:20 , Jeroen Hofstee wrote:
>> Hi,
>> On 01/27/15 06:42, Sergei Nikulov wrote:
>>> 27 янв. 2015 г. 8:23 пользователь "K Richard Pixley" <
>>> rpixley at graphitesystems.com> написал:
>>>> I have a question about how installation is supposed to work.
>>>> I've installed for powerpc-e500v2, using:
>>>>      sudo ./install.sh -d /opt/eldk-5.6 -a i686 -s toolchain -r
>>>> "lsb-sdk
>>>>      minimal-mtdutils" powerpc-e500v2
>>>> (I've also tried toolchain-qte, they seem to behave the same in this
>>>> regard.)
>>>> And I'm trying to build a kernel using:
>>>>      export PATH :=
>>> /opt/eldk-5.6/powerpc-e500v2/sysroots/i686-eldk-linux/usr/bin:/opt/eldk-5.6/powerpc-e500v2/sysroots/i686-eldk-linux/usr/bin/powerpc-linux:$(PATH)
>>> Have you sourceing environment script?
>>> source /opt/eldk-5.6/<you target>/environment-setup-*
>> Yeah, he likely did, which is actually part of the problem.
>> If I recall correctly you need to unset LDFLAGS thereafter.
>> (perhaps CFLAGS and CPPFLAGS as well)
> I'm not sourcing the env file.  I'll try that today but I don't expect
> it to matter.  I recognize most of those env variables and I don't
> believe they're relevant to building the kernel.  I'll try it though,
> just to be sure.
> But I have little hope.  I'm duplicating a process that I already have
> working with powerpc-4xx.  But powerpc-e500v2 seems to work
> differently.  For example, in powerpc-4xx, in the env file,
> /opt/eldk-5.6/powerpc-4xx/environment-setup-ppc44e-linux I find:
>     export
>     SDKTARGETSYSROOT=/opt/eldk-5.6/powerpc-4xx/sysroots/ppc440e-linux
> whereas with powerpc-e500v2, in the env file,
> /opt/eldk-5.6/powerpc-e500v2/environment-setup-ppce500v2-linux I find:
> That looks to me like a string that is intended for replacement that
> has failed to be replaced.  In the environment file, it probably
> doesn't mean much.  But if there's a similar unreplaced trigger in the
> compiler or linker specs, that could be a blocker.
> So... I /expect/ that sourcing the env file will cause breakage,
> although it might be a different sort of breakage since the LD
> definition depends on the value of SDKTARGETSYSROOT.

Here's the problem.  One way or another, this is a bug.  Is this the
right place to report bugs in eldk?  Or is there a better place?

    rpixley at ural>

    rpixley at victory>


More information about the eldk mailing list