On 16 December 2013 11:05, Maxime Ripard <maxime.ripard@xxxxxxxxxxxxxxxxxx> wrote: > On Sun, Dec 15, 2013 at 11:01:17PM +0100, Michal Suchanek wrote: >> On 15 December 2013 14:44, Maxime Ripard >> <maxime.ripard@xxxxxxxxxxxxxxxxxx> wrote: >> >> --- a/drivers/mmc/host/Kconfig >> >> +++ b/drivers/mmc/host/Kconfig >> >> @@ -665,3 +665,11 @@ config MMC_REALTEK_PCI >> >> help >> >> Say Y here to include driver code to support SD/MMC card interface >> >> of Realtek PCI-E card reader >> >> + >> >> +config MMC_SUNXI >> >> + tristate "Allwinner sunxi SD/MMC Host Controller support" >> >> + depends on ARCH_SUNXI >> >> + default y >> > >> > I'm not that fond of these "default y" patterns. It forces the driver >> > down to every user of the multiplatform kernels. I'd suggest removing >> > the default and adding the driver to the defconfigs we have. >> >> And why would you build support for sunxi without building support for >> sunxi mmc? > > Because you can? But does that need to be the default? And is that sane default? > > And where am I actually suggesting to do so? Just a few lines above. You suggest it should not be built by default. > >> Like 90%+ sunxi devices actually have a SD slot or an eMMC or a SDIO >> card connected so what's the point of supporting sunxi without sunxi >> mmc? > > And does your assumption still stand if you take into account all the > other ARM boards? Sure. It depends on sunxi platform so if you compile in support for that 90%+ boards you just added have something connected to a mmc controller so the sane thing is to also add the mmc controller. Thanks Michal -- To unsubscribe from this list: send the line "unsubscribe devicetree" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html