On Fri, Oct 19, 2018 at 03:42:23PM +0100, Kristina Martsenko wrote:

On 19/10/2018 12:35, Catalin Marinas wrote:

quoted

On Tue, Oct 16, 2018 at 05:14:39PM +0100, Kristina Martsenko wrote:

quoted

On 05/10/2018 10:04, Ramana Radhakrishnan wrote:

quoted

On 05/10/2018 09:47, Kristina Martsenko wrote:

quoted

+Virtualization
+--------------
+
+Pointer authentication is not currently supported in KVM guests. KVM
+will mask the feature bits from ID_AA64ISAR1_EL1, and attempted use of
+the feature will result in an UNDEFINED exception being injected into
+the guest.

However applications using instructions from the hint space will
continue to work albeit without any protection (as they would just be
nops) ?

Mostly, yes. If the guest leaves SCTLR_EL1.EnIA unset (and
EnIB/EnDA/EnDB), then PAC* and AUT* instructions in the HINT space will
execute as NOPs. If the guest sets EnIA, then PAC*/AUT* instructions
will trap and KVM will inject an "Unknown reason" exception into the
guest (which will cause a Linux guest to send a SIGILL to the application).

I think that part is fine. If KVM (a fairly recent version with CPUID
sanitisation) does not enable ptr auth, the CPUID should not advertise
this feature either so the guest kernel should not enable it. For the
above instructions in the HINT space, they will just be NOPs. If the
guest kernel enables the feature regardless of the CPUID information, it
deserves to get an "Unknown reason" exception.

quoted

In the latter case we could instead pretend the instruction was a NOP
and not inject an exception, but trapping twice per every function would
probably be terrible for performance. The guest shouldn't be setting
EnIA anyway if ID_AA64ISAR1_EL1 reports that pointer authentication is
not present (because KVM has hidden it).

I don't think we should. The SCTLR_EL1 bits are RES0 unless you know
that the feature is present via CPUID.

quoted

The other special case is the XPACLRI instruction, which is also in the
HINT space. Currently it will trap and KVM will inject an exception into
the guest. We should probably change this to NOP instead, as that's what
applications will expect. Unfortunately there is no EnIA-like control to
make it NOP.

Very good catch. Basically if EL2 doesn't know about ptr auth (older
distro), EL1 may or may not know but leaves SCTLR_EL1 disabled (based on
CPUID), the default HCR_EL2 is to trap (I'm ignoring EL3 as that's like
to have ptr auth enabled, being built for the specific HW). So a user
app considering XPACLRI a NOP (or inoffensive) will get a SIGILL
(injected by the guest kernel following the injection of "Unknown
reason" exception by KVM).

Ramana, is XPACLRI commonly generated by gcc and expects it to be a NOP?
Could we restrict it to only being used at run-time if the corresponding
HWCAP is set? This means redefining this instruction as no longer in the
NOP space.

I think an alternative solution is to just disable trapping of pointer
auth instructions in KVM. This will mean that the instructions will
behave the same in the guest as they do in the host. HINT-space
instructions (including XPACLRI) will behave as NOPs (or perform their
function, if enabled by the guest), and will not trap.

OK, so this means disabling the trap (during early EL2 setup) but still
sanitizing the CPUID not to report the feature to EL1 unless fully
supported on all CPUs.

A side effect of disabling trapping is that keys may effectively leak
from one guest to another, since one guest may set a key and another
guest may use an instruction that uses that key. But this can be fixed
by zeroing the keys every time we enter a guest. We can additionally
trap key accesses (which is separate from instruction trapping), to have
guests fail more reliably and avoid restoring host keys on guest exit.

Actually, if the CPUID doesn't report the feature present, do we care
that a stupid guest writes some register and expects it to be preserved
and not leaked?

Things still won't work well on big.LITTLE systems with mismatched
pointer auth support between CPUs, but as Marc pointed out in the other
email, we can just disable KVM on such systems when we detect a pointer
auth mismatch.

If we ever see such system, we could follow the approach above - disable
trapping at EL2 if the feature is present but do not report it to EL1
via CPUID. A guest should not expect it to work properly.

If we want current stable kernels to support guests that use HINT-space
pointer auth instructions, we'll need to backport the above changes to
stable kernels as well.

I agree.

Even if we restricted userspace to only use XPACLRI if the HWCAP is set,
current stable kernels would still not be able to handle the HINT-space
PAC/AUT instructions that GCC generates, if the guest is pointer auth
aware. None of the stable kernels have the CPUID sanitisation patches,
so the guest would enable pointer auth, which would cause the PAC/AUT
instructions to trap.

Ah, CPUID sanitisation would have to be backported as well. We can
probably get away with something simpler which does not sanitise
big.LITTLE (and I would not expect "enterprise" big.LITTLE machines) but
simply cap/mask the CPUID to what is known to the respective kernel
version. That shouldn't be too intrusive.

-- 
Catalin