Re: [PATCH] gpio: tegra: don't auto-enable for COMPILE_TEST

From: Arnd Bergmann
Date: Wed Jul 20 2016 - 07:07:42 EST


On Wednesday, July 20, 2016 12:41:48 PM CEST Ingo Molnar wrote:
> * Linus Walleij <linus.walleij@xxxxxxxxxx> wrote:
>
> > On Wed, Jul 6, 2016 at 2:54 PM, Arnd Bergmann <arnd@xxxxxxxx> wrote:
> >
> > > I stumbled over a build error with COMPILE_TEST and CONFIG_OF
> > > disabled:
> > >
> > > drivers/gpio/gpio-tegra.c: In function 'tegra_gpio_probe':
> > > drivers/gpio/gpio-tegra.c:603:9: error: 'struct gpio_chip' has no member named 'of_node'
> > >
> > > The problem is that the newly added GPIO_TEGRA Kconfig symbol
> > > does not have a dependency on CONFIG_OF. However, there is another
> > > problem here as the driver gets enabled unconditionally whenever
> > > COMPILE_TEST is set.
> > >
> > > This fixes both problems, by making the symbol user-visible
> > > when COMPILE_TEST is set and default-enabled for ARCH_TEGRA=y.
> > >
> > > As a side-effect, it is now possible to compile-test a Tegra
> > > kernel with GPIO support disabled, which is harmless.
> > >
> > > Signed-off-by: Arnd Bergmann <arnd@xxxxxxxx>
> > > Fixes: 4dd4dd1d2120 ("gpio: tegra: Allow compile test")
> >
> > Patch applied.
>
> This is still not upstream and this build failure is very prominent in the
> upstream kernel right now.

Note that Linus Walleij is currently on vacation. I don't know
if Alexandre could forward this patch in the meantime.

I also have another patch for a similar regression in linux-next
that prevents an x86 allmodconfig kernel with CONFIG_OF=n
from building (see "pinctrl: bcm: add OF dependencies") and that
one hasn't made it into -next yet (unlike this one that is
fixed in -next but broken in v4.7).

> > I'm starting to get a bit tired about anything related to compile testing.
>
> They are very useful in finding real problems as well, and some people
> (like me! rely on being able to generate random kernel images and
> boot them.

Agreed, I send many patches for build-time regressions every week. Most
of them are for harmless problems, but the other ones that find actual
bugs make it absolutely worthwhile.

I don't mind doing this, and most folks are happy to get an early
feedback about when things break. What's annoying is when I send
a patch within hours after a linux-next release that breaks something
and then the bug makes it into mainline anyway, weeks later.

Arnd