Re: [PATCH v2] arm64: defconfig: Enable Rockchip UFS host driver

[Heiko Stuebner <heiko@sntech.de>]

Am Mittwoch, 14. Mai 2025, 14:48:46 Mitteleuropäische Sommerzeit schrieb Diederik de Haas:
> Hi,
> 
> I'm updating my own kernel config and have a question about this change.
> 
> On Wed Mar 5, 2025 at 8:54 AM CET, Shawn Lin wrote:
> > Enable Rockchip UFS host driver.
> >
> > This driver is first introduced on Rockchip RK3576 SoC, and will
> > be used for more SoCs in near future, so add it to defconfig
> > will make these platforms work better.
> >
> > Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
> > Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
> > ---
> >
> >  arch/arm64/configs/defconfig | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/arch/arm64/configs/defconfig b/arch/arm64/configs/defconfig
> > index 3a3706d..907e92d 100644
> > --- a/arch/arm64/configs/defconfig
> > +++ b/arch/arm64/configs/defconfig
> > @@ -1191,6 +1191,7 @@ CONFIG_SCSI_UFS_HISI=y
> >  CONFIG_SCSI_UFS_RENESAS=m
> >  CONFIG_SCSI_UFS_TI_J721E=m
> >  CONFIG_SCSI_UFS_EXYNOS=y
> > +CONFIG_SCSI_UFS_ROCKCHIP=y
> 
> Is there a particular reason why this is `=y` and not `=m`?
> 
> For SCSI_UFS_EXYNOS there is an explicit reason for `=y` in
> 5be91fe8d91b ("arm64: defconfig: Enable Exynos UFS driver")
> (it won't detect (some?) UFS devices otherwise)
> 
> But this patch/commit doesn't say if/why it must be built-in.
> Is there such a reason or is `=m` fine as well?

Most things are supposed to be modules. Though the on-soc mass-
storage devices get a bit of leniency ;-) .

I.e. defconfig is supposed to get you started, so not having to jump
through an initramfs to load the mmc of ufs module is helpful.


Heiko