Re: Familiar 0.9: target kernel

On Sat, 2005-06-11 at 13:08, Patrick Ohly wrote:
> On Thu, 2005-06-09 at 23:24 +0200, Rene Wagner wrote:
> > On Thu, 2005-06-09 at 22:52, Patrick Ohly wrote:
> > > as I mentioned last Sunday on the OE mailing list,
> > > currently building the pxa -h39 kernel fails
> > > because of -3.3.3 aus compiler suffix
> >
> > bk pull/svn up
>
> Only the handhelds-sa*.bb files have been fixed,
> but not handhelds-pxa_2.4.19-rmk6-pxa1-hh39.bb.

http://oe-devel.bkbits.net:8080/openembedded/cset@1.3499.1.10

> So let me ask the question which I originally
> wanted to ask in my email and then forgot:
> which kernel is Familiar 0.9 (now 0.8.3)
> going to use for the iPAQ 39xx series?

The latest tagged and tested code from handhelds.org CVS I guess.
Obviously, that requires someone to have the time and hardware
to do the testing of such an OE-built kernel.

> If it is the latest handhelds-sa kernel (-hh40),
> shouldn't h3900.conf set
> PREFERRED_PROVIDER_virtual/kernel = "handhelds-sa"
> ?

What makes you bevieve an sa kernel would be any good for
pxa machines?

> The reason I am asking is that I'd like to see
> the following features in that kernel and would
> provide patches for OE to include them:
> - USB gadget modules (not used by default, but
> available for those who need them)
> - updated MMC/SD card driver (patch is currently being
> discussed on kernel-discuss_at_handhelds.org and
> seems to address some of the problems reported
> by Familiar users)

Personally, I'm not going to start adding random kernel patches in OE.

Development of the handhelds.org kernels is done in handhelds.org
CVS. Patches hit OE/familiar once a new version has been tagged
in CVS.

As you've mentioned already, kernel-discuss@ is the right place to
discuss patches.

> > > and (worse)
> > > after fixing that bitbake doesn't interpret
> > > h3900.conf as it used to, so that certain
> > > kernel modules are not added to the image.
> >
> > File a bug with detailled information at bugs.openembedded.org
> > please.
>
> I'm still not sure whether it is a bitbake or OE
> problem, but okay, I'll file it as a bug in OE.

I believe Koen has fixed that this afternoon. Maybe you want to try
a new build with a fresh pull.

Rene
Received on Sat Jun 11 2005 - 15:55:58 EDT