Re: [patch RFC 5/5] x86/speculation: Add basic speculation control code

[Tim Chen]

On 01/10/2018 05:53 AM, Van De Ven, Arjan wrote:
>> ibrs_enabled 2:
>>
>> sets IBRS always in host
> 
> this is not secure
> 
>> This matches the semantics described here by Tim patchset on lkml:
>>
>> https://marc.info/?l=linux-kernel&m=151520606320646
> 
> I will talk to Tim, it's not right.
> 
> 

Yes, there's a misunderstanding on my part.  Leaving IBRS=1 all the
time may not provide full protection on all cpu models.

Tim