Re: [Xen-devel] [PATCH v2] SVM: Add union intstat_t for offset 68h in vmcb struct

[Andrew Cooper <andrew.cooper3@xxxxxxxxxx>]

On 24/03/2020 10:37, Pu Wen wrote:
> According to chapter "Appendix B Layout of VMCB" in the new version
> (v3.32) AMD64 APM[1], bit 1 of the VMCB offset 68h is defined as
> GUEST_INTERRUPT_MASK.
>
> In current xen codes, it use whole u64 interrupt_shadow to setup
> interrupt shadow, which will misuse other bit in VMCB offset 68h
> as part of interrupt_shadow.
>
> Add union intstat_t for VMCB offset 68h and fix codes to only use
> bit 0 as intr_shadow according to the new APM description.
>
> Reference:
> [1] https://www.amd.com/system/files/TechDocs/24593.pdf
>
> Signed-off-by: Pu Wen <puwen@xxxxxxxx>

Hmm - this field doesn't appear to be part of AVIC, which makes me
wonder what we're doing without it.

It appears to be a shadow copy of EFLAGS.IF which is only written on
vmexit, and never consumed, but this is based on Appendix B which is the
only reference I can find to the field at all.  Neither the
VMRUN/#VMEXIT descriptions discuss it at all.

Given its position next to the (ambiguous) INTERRUPT_SHADOW, it just
might actually distinguish the STI shadow from the MovSS shadow, but it
could only do that by not behaving as described, and being asymmetric
with EFLAGS.  I don't have time to investigate this right now.

We need the field described in Xen to set it appropriately for virtual
vmexit, but I think that is the extent of what we need to do.

~Andrew