On Mon, 2006-07-17 at 18:17 +0200, Gilles Chanteperdrix wrote:
> Detlef Vollmann wrote:
> > Gilles Chanteperdrix wrote:
> > > Danilo Levantesi wrote:
> > > > So what is the status of the port? Is anything working on a pxa arch?
> > >
> > > More recently, Detlef Vollmann seem also to have begun a port to
> > > PXA.
> > Unfortunately my priorities were shifted, so I'm not able to
> > work on it for the next few weeks.
> > What I've done so far is attached.
> > IT'S UNTESTED!!!
> > So it's nothing you can just use, but something that you can build
> > on (possibly).
> > I'd be happy to hear about any experience with this.
> > What is missing is a look at entry-macro.S.
> > Stelian Pop has done something for the Integrator that I don't
> > really understand and therefore I can't say whether the PXA needs
> > something similar.
> > Detlef
> Starting from Detlef patch, I made some tests on an SA1100 based ARM,
> attached is a patch that works for me, which I also adapted to PXA
> without testing it. It would be nice if someone could test it on PXA.
> In order to do these tests, I had to adapt the ARM ipipe 2.6.15 patch to
> linux 2.6.16, so the 2.6.16 ipipe patch is attached too for further
> testing. Note that I had to make a small modification in
> include/asm-arm/system.h, because the SA1100 version of __xchg uses
> local_irq_restore, which use PSR_I_BIT, which is defined elsewhere, so I
> replaced PSR_I_BIT with its value. I now wonder if it is the proper fix,
> or if __xchg should be fixed to use local_irq_save_hw and
> local_irq_restore_hw instead, as is the case for the atomic operations
> defined in atomic.h and bitops.h.
Definitely, it should be fixed to use hw masking ops.
Xenomai-core mailing list