OSDN Git Service

KVM: X86: Fix operand/address-size during instruction decoding
authorWanpeng Li <wanpeng.li@hotmail.com>
Mon, 6 Nov 2017 00:54:47 +0000 (16:54 -0800)
committerPaolo Bonzini <pbonzini@redhat.com>
Fri, 17 Nov 2017 12:20:12 +0000 (13:20 +0100)
commit3853be2603191829b442b64dac6ae8ba0c027bf9
tree27838e10a5ee8df4c108c28c964849ddd64d7dbb
parent9b8ae63798cb97e785a667ff27e43fa6220cb734
KVM: X86: Fix operand/address-size during instruction decoding

Pedro reported:
  During tests that we conducted on KVM, we noticed that executing a "PUSH %ES"
  instruction under KVM produces different results on both memory and the SP
  register depending on whether EPT support is enabled. With EPT the SP is
  reduced by 4 bytes (and the written value is 0-padded) but without EPT support
  it is only reduced by 2 bytes. The difference can be observed when the CS.DB
  field is 1 (32-bit) but not when it's 0 (16-bit).

The internal segment descriptor cache exist even in real/vm8096 mode. The CS.D
also should be respected instead of just default operand/address-size/66H
prefix/67H prefix during instruction decoding. This patch fixes it by also
adjusting operand/address-size according to CS.D.

Reported-by: Pedro Fonseca <pfonseca@cs.washington.edu>
Tested-by: Pedro Fonseca <pfonseca@cs.washington.edu>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <rkrcmar@redhat.com>
Cc: Nadav Amit <nadav.amit@gmail.com>
Cc: Pedro Fonseca <pfonseca@cs.washington.edu>
Signed-off-by: Wanpeng Li <wanpeng.li@hotmail.com>
Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Radim Krčmář <rkrcmar@redhat.com>
arch/x86/kvm/emulate.c