Re: [PATCH v7 07/20] gpiolib: cdev: support GPIO_V2_GET_LINE_IOCTL and GPIO_V2_LINE_GET_VALUES_IOCTL

[Bartosz Golaszewski <bgolaszewski@baylibre.com>]

On Wed, Sep 9, 2020 at 11:35 AM Kent Gibson <warthog618@gmail.com> wrote:
>
> On Wed, Sep 09, 2020 at 11:26:00AM +0200, Bartosz Golaszewski wrote:
> > On Wed, Sep 9, 2020 at 11:24 AM Kent Gibson <warthog618@gmail.com> wrote:
> > >
> > > On Sat, Sep 05, 2020 at 09:35:36PM +0800, Kent Gibson wrote:
> > > > Add support for requesting lines using the GPIO_V2_GET_LINE_IOCTL, and
> > > > returning their current values using GPIO_V2_LINE_GET_VALUES_IOCTL.
> > > >
> > > > The struct linereq implementation is based on the v1 struct linehandle
> > > > implementation.
> > > >
> > > > Signed-off-by: Kent Gibson <warthog618@gmail.com>
> > > > ---
> > > >
> > >
> > > [snip]
> > >
> > > >               if (copy_from_user(&offset, ip, sizeof(offset)))
> > > >                       return -EFAULT;
> > > > @@ -1104,6 +1505,25 @@ int gpiolib_cdev_register(struct gpio_device *gdev, dev_t devt)
> > > >                MAJOR(devt), gdev->id);
> > > >
> > > >       return 0;
> > > > +     /*
> > > > +      * array sizes must ensure 64-bit alignment and not create holes in
> > > > +      * the struct packing.
> > > > +      */
> > > > +     BUILD_BUG_ON(IS_ALIGNED(GPIO_V2_LINES_MAX, 2));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(GPIO_MAX_NAME_SIZE, 8));
> > > > +
> > > > +     /*
> > > > +      * check that uAPI structs are 64-bit aligned for 32/64-bit
> > > > +      * compatibility
> > > > +      */
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_attribute), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_config_attribute), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_config), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_request), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_info), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_info_changed), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_event), 8));
> > > > +     BUILD_BUG_ON(IS_ALIGNED(sizeof(struct gpio_v2_line_values), 8));
> > > >  }
> > > >
> > >
> > > A couple of things here - these should all be !IS_ALIGNED.
> > > And the BUILD_BUG_ON gets compiled out, and so doesn't fail, if they are
> > > after the return.
> > >
> > > How would you like that fixed - v8 or a patch once v7 is in?
> > >
> > > Cheers,
> > > Kent.
> >
> > v8 please. Why is it compiled out though? Does it need some config option?
> >
>
> Not sure - haven't looked into it.  I only noticed the condition was
> inverted in passing, and when I flipped it it still compiled.
> Moving the return to the end of the function made them all fail,
> as they should if IS_ALIGNED is true.
>
> Having BUILD_BUG_ON being able to compile out quietly is a problem in
> itself.  Nothing special in my setup that I am aware of.
>
> Cheers,
> Kent.

From include/linux/compiler_types.h:

295 #ifdef __OPTIMIZE__
296 # define __compiletime_assert(condition, msg, prefix, suffix)           \
297         do {                                                            \
298                 extern void prefix ## suffix(void)
__compiletime_error(msg); \
299                 if (!(condition))                                       \
300                         prefix ## suffix();                             \
301         } while (0)
302 #else
303 # define __compiletime_assert(condition, msg, prefix, suffix) do {
} while (0)
304 #endif

__OPTIMIZE__ is a predefined macro. I'm not sure about your setup but
it it's defined for me in all my yocto SDK builds and BUILD_BUG_ON(1)
fails as expected.

Bartosz