[PATCH v8 01/69] arm64: Add ARM64_HAS_NESTED_VIRT cpufeature

Marc Zyngier maz at kernel.org
Tue Jan 31 06:00:44 PST 2023


Hi Suzuki,

On Tue, 31 Jan 2023 13:47:31 +0000,
Suzuki K Poulose <suzuki.poulose at arm.com> wrote:
> 
> Hi Marc
> 
> On 31/01/2023 09:23, Marc Zyngier wrote:
> > From: Jintack Lim <jintack.lim at linaro.org>
> > 
> > Add a new ARM64_HAS_NESTED_VIRT feature to indicate that the
> > CPU has the ARMv8.3 nested virtualization capability, together
> > with the 'kvm-arm.mode=nested' command line option.
> > 
> > This will be used to support nested virtualization in KVM.
> > 
> > Reviewed-by: Russell King (Oracle) <rmk+kernel at armlinux.org.uk>
> > Signed-off-by: Jintack Lim <jintack.lim at linaro.org>
> > Signed-off-by: Andre Przywara <andre.przywara at arm.com>
> > Signed-off-by: Christoffer Dall <christoffer.dall at arm.com>
> > [maz: moved the command-line option to kvm-arm.mode]
> 
> Should this be separate kvm-arm mode ? Or can this be tied to
> is_kernel_in_hyp_mode() ? Given this mode (from my limited
> review) doesn't conflict with normal VHE mode (and RME support),
> adding this explicit mode could confuse the user.

What is exactly the objection here? NV is more or less a VHE++ mode,
but is also completely experimental and incomplete.

> In case we need a command line to turn the NV mode on/off,
> we could always use the id-override and simply leave this out ?

I really want an explicit user buy-in. There is absolutely no way this
can be enabled by default, the risks are way too high. Just look at
the x86 story: it took them 10 years to enable NV by default. I don't
expect to do any better.

Thanks,

	M.

-- 
Without deviation from the norm, progress is not possible.



More information about the linux-arm-kernel mailing list