Date: Mon, 22 Feb 2016 18:12:44 -0700 From: Ian Lepore <ian@freebsd.org> To: =?ISO-8859-1?Q?Mika=EBl?= Urankar <mikael.urankar@gmail.com> Cc: "freebsd-arm@freebsd.org" <freebsd-arm@freebsd.org> Subject: Re: Latest 11.0 snapshot won't boot on CUBOX either Message-ID: <1456189964.1294.69.camel@freebsd.org> In-Reply-To: <CAJwjRmRD23vECYLaOrjVKevN0Hh0CxqW==CtLUqCgfBCnXTVig@mail.gmail.com> References: <201602220109.SAA20866@mail.lariat.net> <1456113068.1294.46.camel@freebsd.org> <CAJwjRmSH_6EN5pVDrZzazBxzscY4cz3szKL%2B1d9BzNTmojMydA@mail.gmail.com> <CAJwjRmRD23vECYLaOrjVKevN0Hh0CxqW==CtLUqCgfBCnXTVig@mail.gmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Mon, 2016-02-22 at 11:30 +0100, Mikaël Urankar wrote:
> 2016-02-22 10:30 GMT+01:00 Mikaël Urankar <mikael.urankar@gmail.com>:
> > 2016-02-22 4:51 GMT+01:00 Ian Lepore <ian@freebsd.org>:
> > > On Sun, 2016-02-21 at 18:09 -0700, Brett Glass wrote:
> > > > After the experience described in my last message, I tried the
> > > > mid-February 11.0 snapshot (r295683). This one actually began
> > > > to
> > > > use the video display, but did not boot whether or not I
> > > > attached a
> > > > USB keyboard. A serial terminal showed the following:
> > > >
> > > >
[...]
> > > > --Brett Glass
> > >
> > > My wandboards have hung up the same way for about a week now, but
> > > I
> > > haven't had time to investigate why.
> > >
> > > -- Ian
> >
> > The update of the Linux dts breaks the imx6 (r295436)
>
> They have changed the interrupt-parent to gpc in the 'soc' structure:
> /usr/src/sys/gnu/dts/arm/imx6qdl.dtsi
> soc {
> interrupt-parent = <&gpc>;
>
> My imx6q boots fine if I put back &intc
> I don't know what is the correct fix though
Thanks for diagnosing this, it saved me a bunch of time. The real fix
some day will be to have a driver for the second interrupt controller
and hand off control between that driver and the GIC driver as we
switch in/out of deep sleep modes. First we would have to support deep
sleep modes. :)
For now, I've committed a workaround in r295904 that just rewrites the
FDT in kernel memory during early kernel init, pointing the soc node
interrupt parent back to the GIC controller, to make everything work
the way it always has for us.
-- Ian
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1456189964.1294.69.camel>