Re: [PATCH v6 11/25] KVM: arm64: Add remaining ID registers to id_reg_desc_table

[Oliver Upton <oupton@google.com>]

On Wed, Mar 23, 2022 at 10:25:35PM +0000, Oliver Upton wrote:
> On Wed, Mar 23, 2022 at 03:22:43PM -0700, Ricardo Koller wrote:
> > On Wed, Mar 23, 2022 at 08:44:26PM +0000, Oliver Upton wrote:
> > > On Wed, Mar 23, 2022 at 01:13:32PM -0700, Ricardo Koller wrote:
> > > > On Wed, Mar 23, 2022 at 07:53:14PM +0000, Oliver Upton wrote:
> > > > > Hi Reiji,
> > > > > 
> > > > > On Thu, Mar 10, 2022 at 08:47:57PM -0800, Reiji Watanabe wrote:
> > > > > > Add hidden or reserved ID registers, and remaining ID registers,
> > > > > > which don't require special handling, to id_reg_desc_table.
> > > > > > Add 'flags' field to id_reg_desc, which is used to indicates hiddden
> > > > > > or reserved registers. Since now id_reg_desc_init() is called even
> > > > > > for hidden/reserved registers, change it to not do anything for them.
> > > > > > 
> > > > > > Signed-off-by: Reiji Watanabe <reijiw@google.com>
> > > > > 
> > > > > I think there is a very important detail of the series that probably
> > > > > should be highlighted. We are only allowing AArch64 feature registers to
> > > > > be configurable, right? AArch32 feature registers remain visible with
> > > > > their default values passed through to the guest. If you've already
> > > > > stated this as a precondition elsewhere then my apologies for the noise.
> > > > 
> > > > Aren't AArch64 ID regs architecturally mapped to their AArch32
> > > > counterparts?  They should show the same values.  I'm not sure if it's a
> > > > problem (and if KVM is faithful to that rule),
> > > 
> > > I believe it's a bit more subtle than that. The AArch32 feature registers
> > > are architecturally mapped to certain encodings accessible from AArch64.
> > > For example, ID_PFR0_EL1 is actually a 64 bit register where bits [31:0]
> > > map to the ID_PFR0 AArch32 register. ID_PFR0_EL1 is only accessible from
> > > AArch64 with the MRS instruction, and ID_PFR0 is only accessible from
> > > AArch32 with the MRC instruction. KVM just so happens to handle both of
> > > these reads from the same sys_reg_desc.

Ughhhhh.

We actually clear HCR_EL2.TID3 for AArch32 guests, so AArch32 EL1 reads
straight from hardware. Considering the work we put in to make sure
feature registers are consistent system-wide and the limitations on
certain features, this is plain wrong.

I have a series that addresses this but need to go find some 32 bit
hardware to test with :)

--
Thanks,
Oliver
_______________________________________________
kvmarm mailing list
kvmarm@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm