On 01/02/18 08:41, Jan Beulich wrote:
>>>> On 24.01.18 at 14:12, <andrew.coop...@citrix.com> wrote:
>>  void __init init_speculation_mitigations(void)
>>  {
>>      enum ind_thunk thunk = THUNK_DEFAULT;
>> +    bool ibrs = false;
>>  
>>      /*
>>       * Has the user specified any custom BTI mitigations?  If so, follow 
>> their
>>       * instructions exactly and disable all heuristics.
>>       */
>> -    if ( opt_thunk != THUNK_DEFAULT )
>> +    if ( opt_thunk != THUNK_DEFAULT || opt_ibrs != -1 )
>>      {
>>          thunk = opt_thunk;
>> +        ibrs  = !!opt_ibrs;
>>      }
> The comment above has become stale with later additions. Question
> is what the intentions are, i.e. whether it is the comment or the code
> that should be changed.

Hmm true.  The comment used to be true, but I changed it based on
internal feedback.  The new logic means that explicit choices only
affect the defaults of their related area.

e.g. the choice of whether to use IBPB or not (which really is an
isolated decision here) shouldn't override the default choice for
whether to use IBRS/retpoline/etc.

~Andrew

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

Reply via email to