Re: Crashing 'kzm' target in next-20160913 due to 'gpio: mxc: shift gpio_mxc_init() to subsys_initcall level'

[Vladimir Zapolskiy]

Hi Linus,

On 09/15/2016 04:18 PM, Vladimir Zapolskiy wrote:
> Hi Linus,
>
> On 09/15/2016 03:19 PM, Linus Walleij wrote:
> > On Wed, Sep 14, 2016 at 4:33 PM, Guenter Roeck <linux@xxxxxxxxxxxx> wrote:
> > > On 09/14/2016 12:19 AM, Linus Walleij wrote:
> > > > On Wed, Sep 14, 2016 at 5:20 AM, Guenter Roeck <linux@xxxxxxxxxxxx> wrote:
> > > >
> > > > > So, in other words, lots of bugs here. Nevertheless, I would suggest to
> > > > > keep
> > > > > using postcore_initcall(), at least until it is sure that all gpio
> > > > > clients handle
> > > > > -EPROBE_DEFER
> > > > > correctly.
> > > >
> > > >
> > > > So can I just revert this patch in isolation? None of the other patches
> > > > depend on it?
> > > Good question, if this was part of a series.
> >
> > No response from author, so betting on it and reverting this in isolation
> > with your Reported-by.
>
> please revert the revert :)

ok, the problem is recognized, but as I said it is not sufficient to revert
only that change, if it is selected to go by path of reverts commit
("gpio: mxc: add generic gpio request/free callbacks to pinctrl") from
your for-next branch should be reverted as well, the latter commit depends
on the reverted commit.

I have one i.MX31 powered board on hand, I'll start preparing a proper
fix for the issue.

--
With best wishes,
Vladimir