| Message ID | 1460365435-12325-1-git-send-email-andrew.cooper3@citrix.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, Apr 11, 2016 at 10:03:55AM +0100, Andrew Cooper wrote: > c/s f71ecb6 "x86: introduce a new VMASSIST for architectural behaviour of > iopl" shifted the vcpu iopl field by 12, but didn't update the logic which > reconstructs the guests eflags for migration. > > Existing guest kernels set a vIOPL of 1, to prevent them from faulting when > accessing IO ports. This bug manifests as a crash after migrate, as the vIOPL > reverts back to the default of 0, and the guest suffers an unexpected #GP > fault. > > Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> > --- > CC: Jan Beulich <JBeulich@suse.com> > CC: Wei Liu <wei.liu2@citrix.com> FWIW: Reviewed-by: Wei Liu <wei.liu2@citrix.com> This is fine for 4.7. Wei. > --- > xen/arch/x86/domctl.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/xen/arch/x86/domctl.c b/xen/arch/x86/domctl.c > index cba1e37..6e709cb 100644 > --- a/xen/arch/x86/domctl.c > +++ b/xen/arch/x86/domctl.c > @@ -1512,7 +1512,7 @@ void arch_get_info_guest(struct vcpu *v, vcpu_guest_context_u c) > > /* IOPL privileges are virtualised: merge back into returned eflags. */ > BUG_ON((c(user_regs.eflags) & X86_EFLAGS_IOPL) != 0); > - c(user_regs.eflags |= v->arch.pv_vcpu.iopl << 12); > + c(user_regs.eflags |= v->arch.pv_vcpu.iopl); > > if ( !compat ) > { > -- > 2.1.4 >
On 04/11/2016 05:03 AM, Andrew Cooper wrote: > c/s f71ecb6 "x86: introduce a new VMASSIST for architectural behaviour of > iopl" shifted the vcpu iopl field by 12, but didn't update the logic which > reconstructs the guests eflags for migration. > > Existing guest kernels set a vIOPL of 1, to prevent them from faulting when > accessing IO ports. This bug manifests as a crash after migrate, as the vIOPL > reverts back to the default of 0, and the guest suffers an unexpected #GP > fault. > > Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Tested-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> (save/restore was part of brokenness as well, obviously) > --- > CC: Jan Beulich <JBeulich@suse.com> > CC: Wei Liu <wei.liu2@citrix.com> > --- > xen/arch/x86/domctl.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > diff --git a/xen/arch/x86/domctl.c b/xen/arch/x86/domctl.c > index cba1e37..6e709cb 100644 > --- a/xen/arch/x86/domctl.c > +++ b/xen/arch/x86/domctl.c > @@ -1512,7 +1512,7 @@ void arch_get_info_guest(struct vcpu *v, vcpu_guest_context_u c) > > /* IOPL privileges are virtualised: merge back into returned eflags. */ > BUG_ON((c(user_regs.eflags) & X86_EFLAGS_IOPL) != 0); > - c(user_regs.eflags |= v->arch.pv_vcpu.iopl << 12); > + c(user_regs.eflags |= v->arch.pv_vcpu.iopl); > > if ( !compat ) > {
On Mon, Apr 11, 2016 at 10:08:58AM -0400, Boris Ostrovsky wrote: > On 04/11/2016 05:03 AM, Andrew Cooper wrote: > >c/s f71ecb6 "x86: introduce a new VMASSIST for architectural behaviour of > >iopl" shifted the vcpu iopl field by 12, but didn't update the logic which > >reconstructs the guests eflags for migration. > > > >Existing guest kernels set a vIOPL of 1, to prevent them from faulting when > >accessing IO ports. This bug manifests as a crash after migrate, as the vIOPL > >reverts back to the default of 0, and the guest suffers an unexpected #GP > >fault. > > > >Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> > > Tested-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> > > (save/restore was part of brokenness as well, obviously) Applied.
diff --git a/xen/arch/x86/domctl.c b/xen/arch/x86/domctl.c index cba1e37..6e709cb 100644 --- a/xen/arch/x86/domctl.c +++ b/xen/arch/x86/domctl.c @@ -1512,7 +1512,7 @@ void arch_get_info_guest(struct vcpu *v, vcpu_guest_context_u c) /* IOPL privileges are virtualised: merge back into returned eflags. */ BUG_ON((c(user_regs.eflags) & X86_EFLAGS_IOPL) != 0); - c(user_regs.eflags |= v->arch.pv_vcpu.iopl << 12); + c(user_regs.eflags |= v->arch.pv_vcpu.iopl); if ( !compat ) {
c/s f71ecb6 "x86: introduce a new VMASSIST for architectural behaviour of iopl" shifted the vcpu iopl field by 12, but didn't update the logic which reconstructs the guests eflags for migration. Existing guest kernels set a vIOPL of 1, to prevent them from faulting when accessing IO ports. This bug manifests as a crash after migrate, as the vIOPL reverts back to the default of 0, and the guest suffers an unexpected #GP fault. Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> --- CC: Jan Beulich <JBeulich@suse.com> CC: Wei Liu <wei.liu2@citrix.com> --- xen/arch/x86/domctl.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)