Re: [PATCH v5 09/14] iio: magnetometer: yas530: Introduce "chip_info" structure

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Mon, Aug 08, 2022 at 08:04:20PM +0200, Andy Shevchenko wrote:
> On Mon, Aug 8, 2022 at 5:59 PM Nathan Chancellor <nathan@xxxxxxxxxx> wrote:
> > On Mon, Aug 08, 2022 at 01:18:06PM, +0200, Andy Shevchenko wrote:
> > > On Mon, Aug 8, 2022 at 7:40 AM kernel test robot <lkp@xxxxxxxxx> wrote:
> > >
> > > ...
> > >
> > > > All errors (new ones prefixed by >>):
> > > >
> > > > >> drivers/iio/magnetometer/yamaha-yas530.c:933:19: error: initializer element is not a compile-time constant
> > > >                    .product_name = yas5xx_product_name[yas530],
> > > >                                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~
> > > >    1 error generated.
> > >
> > > What?!
> > >
> > > The yas530 is a part of the enum, how come that compiler can't see
> > > this? Looks like a Clang bug.
> >
> > That is not what clang is complaining about here, you'll see the same
> > error even if you used '0', '1', or '2' here:
> >
> >   drivers/iio/magnetometer/yamaha-yas530.c:933:19: error: initializer element is not a compile-time constant
> >                   .product_name = yas5xx_product_name[0],
> >                                   ^~~~~~~~~~~~~~~~~~~~~~
> >   1 error generated.
> >
> > It is complaining that the initializer element
> > ('yas5xx_product_name[yas530]', rather than just 'yas530') is not
> > constant, which is a true complaint if I am reading C11 standard 6.6.7
> > correctly.
> >
> > GCC 8+ has chosen to accept const structures as constant expressions in
> > designated initializers, which it is allowed to do per 6.6.10. Nick did
> > have a patch to try and match this behavior in clang but the work that
> > was requested doesn't seem to be trivial so it was never finalized:
> > https://reviews.llvm.org/D76096
> >
> > You'll see the same error with GCC 7:
> >
> >   drivers/iio/magnetometer/yamaha-yas530.c:933:19: error: initializer element is not constant
> >      .product_name = yas5xx_product_name[yas530],
> >                      ^~~~~~~~~~~~~~~~~~~
> >   drivers/iio/magnetometer/yamaha-yas530.c:933:19: note: (near initialization for ‘yas5xx_chip_info_tbl[0].product_name’)
> >   drivers/iio/magnetometer/yamaha-yas530.c:938:19: error: initializer element is not constant
> >      .product_name = yas5xx_product_name[yas532],
> >                      ^~~~~~~~~~~~~~~~~~~
> >   drivers/iio/magnetometer/yamaha-yas530.c:938:19: note: (near initialization for ‘yas5xx_chip_info_tbl[1].product_name’)
> >   drivers/iio/magnetometer/yamaha-yas530.c:943:19: error: initializer element is not constant
> >      .product_name = yas5xx_product_name[yas533],
> >                      ^~~~~~~~~~~~~~~~~~~
> >   drivers/iio/magnetometer/yamaha-yas530.c:943:19: note: (near initialization for ‘yas5xx_chip_info_tbl[2].product_name’)
> 
> > > >    930  static const struct yas5xx_chip_info yas5xx_chip_info_tbl[] = {
> > > >    931          [yas530] = {
> > > >    932                  .devid = YAS530_DEVICE_ID,
> > > >  > 933                  .product_name = yas5xx_product_name[yas530],
> > > >    934                  .version_name = yas5xx_version_names[yas530],
> 
> Would then
> 
>   .product_name = "YAS530 MS-3E",
>   .version_names = { "A", "B" },
> 
> work?

I haven't tested it but there is no reason that shouldn't work.

> Jakob, note 's' in the field name as well.
> 
> > > >    935          },
> > > >    936          [yas532] = {
> > > >    937                  .devid = YAS532_DEVICE_ID,
> > > >    938                  .product_name = yas5xx_product_name[yas532],
> > > >    939                  .version_name = yas5xx_version_names[yas532],
> > > >    940          },
> > > >    941          [yas533] = {
> > > >    942                  .devid = YAS532_DEVICE_ID,
> > > >    943                  .product_name = yas5xx_product_name[yas533],
> > > >    944                  .version_name = yas5xx_version_names[yas533],
> > > >    945          },
> > > >    946  };
> 
> -- 
> With Best Regards,
> Andy Shevchenko



[Index of Archives]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Input]     [Linux Kernel]     [Linux SCSI]     [X.org]

  Powered by Linux