Re: [Xen-devel] [PATCH v8 12/17] x86/entry: Organise the use of MSR_SPEC_CTRL at each entry/exit point

[Andrew Cooper <Andrew.Cooper3@xxxxxxxxxx>]

No, not really. Omitting it on the grounds of "we don't expect to take a double 
fault" don't beat uniformally altering all the entrypoints in a consistent 
manor. 

The only thing which can go wrong is that we forget to do it when it is needed.

~Andrew 
________________________________________
From: Jan Beulich [JBeulich@xxxxxxxx]
Sent: 17 January 2018 08:47
To: Andrew Cooper
Cc: Xen-devel
Subject: Re: [Xen-devel] [PATCH v8 12/17] x86/entry: Organise the use of 
MSR_SPEC_CTRL at each entry/exit point

>>> On 16.01.18 at 22:24, <andrew.cooper3@xxxxxxxxxx> wrote:
> On 15/01/18 12:09, Jan Beulich wrote:
>>>>> On 12.01.18 at 19:01, <andrew.cooper3@xxxxxxxxxx> wrote:
>>> @@ -586,6 +611,10 @@ ENTRY(double_fault)
>>>          movl  $TRAP_double_fault,4(%rsp)
>>>          /* Set AC to reduce chance of further SMAP faults */
>>>          SAVE_ALL STAC
>>> +
>>> +        SPEC_CTRL_ENTRY_FROM_INTR /* Req: %rsp=regs Clob: acd */
>>> +        /* WARNING! `ret`, `call *`, `jmp *` not safe before this point. */
>> Is it actually useful to do _anything_ in the double fault handler?
>
> Typically no, but then again we hope never to execute this code.
>
> OTOH, we would need to do this if we ever get around to doing espfix64.

Could I get you to omit the change to the handler until then, to keep
it as straightforward as possible?

Jan


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