On Mon, Feb 27, 2023, at 04:54, Benjamin Gray wrote: > The ppc64le_allmodconfig sets I2C_PASEMI=y and leaves COMPILE_TEST to > default to y and I2C_APPLE to default to m, running into a known > incompatible configuration that breaks the build [1]. > > Disable the I2C_APPLE build during PPC compile tests to prevent this. > > [1]: https://lore.kernel.org/all/202112061809.XT99aPrf-lkp@xxxxxxxxx > > Suggested-by: Michael Ellerman <mpe@xxxxxxxxxxxxxx> > Signed-off-by: Benjamin Gray <bgray@xxxxxxxxxxxxx> This should work, but I find the added dependency is a bit confusing to readers. > drivers/i2c/busses/Kconfig | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig > index d4975444a32d..3db56e446902 100644 > --- a/drivers/i2c/busses/Kconfig > +++ b/drivers/i2c/busses/Kconfig > @@ -896,7 +896,7 @@ config I2C_PASEMI > > config I2C_APPLE > tristate "Apple SMBus platform driver" > - depends on ARCH_APPLE || COMPILE_TEST > + depends on ARCH_APPLE || (COMPILE_TEST && !PPC) > default ARCH_APPLE I would add a line such as depends on !I2C_PASEMI which actually still allows building both drivers as loadable modules but prevents I2C_APPLE=y when I2C_PASEMI=m or vice versa. For completeness, it might be good to extend the I2C_PASEMI dependency to include '|| COMPILE_TEST', which would ensure that any future problems show up on x86 builds as well, rather than just powerpc. Arnd
