On 11/09/19 21:53, Sean Christopherson wrote:
> On Wed, Sep 11, 2019 at 05:51:05PM +0200, Paolo Bonzini wrote:
>> On 27/08/19 15:07, Jan Dakinevich wrote:
>>> This series intended to fix (again) a bug that was a subject of the
>>> following change:
>>>
>>> 6ea6e84 ("KVM: x86: inject exceptions produced by x86_decode_insn")
>>>
>>> Suddenly, that fix had a couple mistakes. First, ctxt->have_exception was
>>> not set if fault happened during instruction decoding. Second, returning
>>> value of inject_emulated_instruction was used to make the decision to
>>> reenter guest, but this could happen iff on nested page fault, that is not
>>> the scope where this bug could occur.
>>>
>>> However, I have still deep doubts about 3rd commit in the series. Could
>>> you please, make me an advise if it is the correct handling of guest page
>>> fault?
>>>
>>> Jan Dakinevich (3):
>>> KVM: x86: fix wrong return code
>>> KVM: x86: set ctxt->have_exception in x86_decode_insn()
>>> KVM: x86: always stop emulation on page fault
>>>
>>> arch/x86/kvm/emulate.c | 4 +++-
>>> arch/x86/kvm/x86.c | 4 +++-
>>> 2 files changed, 6 insertions(+), 2 deletions(-)
>>>
>>
>> Queued, thanks. I added the WARN_ON_ONCE that Sean suggested.
>
> Which version did you queue? It sounds like you queued v1, which breaks
> VMware backdoor emulation due to incorrect patch ordering. v3[*] fixes
> the ordering issue and adds the WARN_ON_ONCE.
I applied v1 with all the fixes, then found out v3 existed and replaced
with it (but still added a comment).
Paolo
