Re: [Xen-devel] [PATCH 1/2] x86/vtx: Fix fault semantics for early task switch failures

[Jan Beulich <jbeulich@xxxxxxxx>]

On 22.11.2019 14:12, Andrew Cooper wrote:
> On 22/11/2019 13:08, Jan Beulich wrote:
>> On 22.11.2019 13:37, Roger Pau Monné  wrote:
>>> On Thu, Nov 21, 2019 at 10:15:50PM +0000, Andrew Cooper wrote:
>>>> The VT-x task switch handler adds inst_len to rip before calling
>>>> hvm_task_switch().  This causes early faults to be delivered to the guest 
>>>> with
>>>> trap semantics, and break restartibility.
>>>>
>>>> Instead, pass the instruction length into hvm_task_switch() and write it 
>>>> into
>>>> the outgoing tss only, leaving rip in its original location.
>>>>
>>>> For now, pass 0 on the SVM side.  This highlights a separate preexisting 
>>>> bug
>>>> which will be addressed in the following patch.
>>>>
>>>> While adjusting call sites, drop the unnecessary uint16_t cast.
>>>>
>>>> Signed-off-by: Andrew Cooper <andrew.cooper3@xxxxxxxxxx>
>>> Code LGTM:
>>>
>>> Reviewed-by: Roger Pau Monné <roger.pau@xxxxxxxxxx>
>> Acked-by: Jan Beulich <jbeulich@xxxxxxxx>
> 
> It occurs to me that this also fixes a vmentry failure in the corner
> case that an instruction, which crosses the 4G=>0 boundary takes a
> fault.  %rip will be adjusted without being truncated.

I was about to say so in my earlier reply, until I paid attention
to this

@@ -2987,7 +2987,7 @@ void hvm_task_switch(
     if ( taskswitch_reason == TSW_iret )
         eflags &= ~X86_EFLAGS_NT;
 
-    tss.eip    = regs->eip;
+    tss.eip    = regs->eip + insn_len;

together with the subsequent

    regs->rip    = tss.eip;

already having taken care of this aspect before, afaict.

Jan

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxxx
https://lists.xenproject.org/mailman/listinfo/xen-devel