Re: [PATCH] arm64: Kconfig.platforms: fix warning unmet direct dependencies

[Lucas Stach]

Am Freitag, den 25.01.2019, 14:32 +0000 schrieb Catalin Marinas:
> On Tue, Jan 15, 2019 at 08:18:39PM +0100, Anders Roxell wrote:
> > When ARCH_MXC get enabled, ARM64_ERRATUM_845719 will be selected and
> > this warning will happen when COMPAT isn't set.
> > 
> > WARNING: unmet direct dependencies detected for ARM64_ERRATUM_845719
> > Â Depends on [n]: COMPAT [=n]
> > Â Selected by [y]:
> > Â - ARCH_MXC [=y]
> > 
> > Rework to add 'if COMPAT' before ARM64_ERRATUM_845719 gets selected,
> > since ARM64_ERRATUM_845719 depends on COMPAT.
> > 
> > > > Signed-off-by: Anders Roxell <anders.roxell@xxxxxxxxxx>
> > ---
> > Âarch/arm64/Kconfig.platforms | 2 +-
> > Â1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/arch/arm64/Kconfig.platforms b/arch/arm64/Kconfig.platforms
> > index 251ecf34cb02..d4faca775d9c 100644
> > --- a/arch/arm64/Kconfig.platforms
> > +++ b/arch/arm64/Kconfig.platforms
> > @@ -145,7 +145,7 @@ config ARCH_MVEBU
> > Âconfig ARCH_MXC
> > > > Â	bool "ARMv8 based NXP i.MX SoC family"
> > > > Â	select ARM64_ERRATUM_843419
> > > > -	select ARM64_ERRATUM_845719
> > > > +	select ARM64_ERRATUM_845719 if COMPAT
> > > > Â	help
> > > > Â	ÂÂThis enables support for the ARMv8 based SoCs in the
> > Â	ÂÂNXP i.MX family.
> 
> Actually, do we need to select the errata workarounds explicitly? That
> seems to be the only case where we do it (commit 930507c18304, "arm64:
> add basic Kconfig symbols for i.MX8"). They are default y, so we
> shouldn't need to force them on.

This is based on past experience. We've had a lot of cases were people
did not enable the necessary CPU errata workaround, which then usually
lead to very hard to debug system failures. It is on our list of things
to look out for now, but I would feel much better if there is just no
chance for a user to misconfigure the kernel in this way.

Regards,
Lucas