RE: [Ltib] Problem compiling latest ARM kernel

[Stuart Hughes]

Hi Colin,

I'm not sure why, but currently the imx21 is on the older toolchain.  If
you reconfigure for one of the others you'll pick up the later
toolchain.

Regards, Stuart

On Tue, 2008-12-16 at 14:14 +0000, Colin Helliwell wrote:
> I'm just checking everything is up to date from cvs, but I only checked out
> just last week.
> I can see the Toolchain line you mention, in toolchain.lkc, however the
> config menu show the toolchain as "arm gcc-3.3.2/glibc-2.3.2 hard-float". It
> seems this comes from the IMX21 default config, whereas the IMX27 and IMX31
> call up the 4.1.2 toolchain.
> I'll try re-configuring for one of these other platforms - at this stage I
> am just checking that I have (or have not!) got things basically installed
> and building ok, prior to defining my own platform anyway.
> 
> > -----Original Message-----
> > From: Stuart Hughes [mailto:address@hidden
> > Sent: 16 December 2008 11:43
> > To: address@hidden
> > Cc: address@hidden
> > Subject: Re: [Ltib] Problem compiling latest ARM kernel
> > 
> > Hi Colin,
> > 
> > I'm not sure about the problem you're hitting, but in the latest
> > Savannah CVS ltib source, the default toolchain is now:
> > 
> >  Toolchain (ARMv5te gcc-4.1.2,Multi-lib,gnueabi/glibc-2.5-nptl-3)
> > 
> > If you pull the update and run ./ltib --configure
> > you should see this.  It will force a full rebuild.
> > 
> > Regards, Stuart
> > 
> > On Tue, 2008-12-16 at 11:23 +0000, Colin Helliwell wrote:
> > > I am building an up-to-date (2.6.27.6) ARM kernel, locally installed,
> > via
> > > ltib and hitting a few problems. I 'solved' the first by disabling "Use
> > the
> > > ARM EABI", but I still have a compiler/inline assembler error in
> > > page_alloc.c.  From what scant info I can find, I *think* this is a
> > problem
> > > due to an outdated toolchain - ltib is using gcc 3.3.2 cross toolchain.
> > >
> > > Before I go diving into hooking in a newer toolchain, has anyone else
> > seen
> > > the problem and maybe have an alternative solution?
> > >
> > 
> 
>