[PATCH v2 10/21] arm64: KVM: VHE: Split save/restore of sysregs shared between EL1 and EL2

On Tue, Feb 02, 2016 at 04:19:44PM +0000, Marc Zyngier wrote:

On 02/02/16 15:46, Christoffer Dall wrote:

quoted

On Tue, Feb 02, 2016 at 09:46:05AM +0000, Marc Zyngier wrote:

quoted

On 01/02/16 13:54, Christoffer Dall wrote:

quoted

On Mon, Jan 25, 2016 at 03:53:44PM +0000, Marc Zyngier wrote:

quoted

A handful of system registers are still shared between EL1 and EL2,
even while using VHE. These are tpidr*_el[01], actlr_el1, sp0, elr,
and spsr.

So by shared registers you mean registers that do both have an EL0/1
version as well as an EL2 version, but where accesses aren't rewritten
transparently?

No, I mean that these registers do *not* have a separate banked version.
There is only a single set of registers, which have to be save/restored
the old way.

huh, ARMv8 clearly specifies the existence of TPIDR_EL0, TPIDR_EL1, and
TPIDR_EL2, for example.

I cannot seem to find anywhere in the VHE spec that says that the
TPIDR_EL2 goes away.  I'm confused now.

Nothing goes away, but these registers do not get renamed either. For
example, TPIDR_EL1 doesn't magically access TPIDR_EL2 when running at
EL2+VHE, and there is no TPIDR_EL12 accessor either.

So TPIDR_EL1 is effectively "shared" between host and guest, and must be
save/restored (note that the host kernel still uses TIPDR_EL1 even when
running with VHE, and that KVM still uses TPIDR_EL2 to cache the current
vcpu).

ok, I can understand as long as we're saying a register is shared
between the host and the guest, but it was the "registers are shared
between EL1 and EL2" that threw me off.

quoted

quoted

quoted

also, by sp0 do you mean sp_el0, and by elr you mean elr_el1, and by
spsr you mean spsr_el1 ?

sp0 -> sp_el0 indeed. elr and spsr really are the guest PC and PSTATE,
so I should really reword this commit message, it is utterly confusing.

I guess I don't understand the definition of a 'shared' register given
your comments here...

Does this make it clearer?

yes.  You could change the host to path it when using VHE to use
TPIDR_EL2 if you wanted and store the vcpu pointer on the stack while
running the guest, but there's probably no real benefit of doing so.

I'll be shutting up now...

Thanks,
-Christoffer