Re: [Patch v2 4/4] x86/speculation: Add prctl to control indirect branch speculation per process

[Tim Chen]

On 10/02/2018 02:35 AM, Ingo Molnar wrote:
> 
> * Tim Chen <tim.c.chen@xxxxxxxxxxxxxxx> wrote:
> 
>> To migitgate possible app to app attack from branch target buffer poisoning,
>> a new prctl is provided to control branch speculation for applications in
>> user app.  The following interfaces are provided:
> 
> s/migitgate
>  /mitigate
> 
>>
>> prctl(PR_SET_SPECULATION_CTRL, PR_INDIR_BRANCH, PR_SPEC_DISABLE, 0, 0);
>> - Disable branch target speculation to protect against app to app
>> style attack using IBPB and STIBP
>>
>> prctl(PR_SET_SPECULATION_CTRL, PR_INDIR_BRANCH, PR_SPEC_ENABLE, 0, 0);
>> - Allow branch target speculation, no mitigation for Spectre V2
>>
>> prctl(PR_GET_SPECULATION_CTRL, PR_INDIR_BRANCH, 0, 0, 0)
>> - Query the indirect branch speculation restriction on a process
> 
> Well 'a process' is always 'the current process' in this case, right?

Right.

> 
>> -			lite   - only turn on mitigation for non-dumpable processes
>> +			lite   - turn on mitigation for non-dumpable processes
>> +				 or processes that has indirect branch restricted
>> +				 via prctl's PR_SET_SPECULATION_CTRL option
> 
> s/or processes that has indirect
>  /or processes that have been indirect
> 
> ?
> 
>> +	/*
>> +	 * If being set on non-current task, delay setting the CPU
>> +	 * mitigation until it is next scheduled.
>> +	 * Use speculative_store_bypass_update will update SPEC_CTRL MSR
>> +	 */
>> +	if (task == current && update)
>> +		speculative_store_bypass_update_current();
> 
> Did you mean:
> 
> 	Call to speculative_store_bypass_update_current() will update SPEC_CTRL MSR

Yes.

> 
> ?
> 
> 
>> -	 * For lite protection mode, we only protect the non-dumpable
>> -	 * processes.
>> +	 * For lite protection mode, we protect processes  
>> +	 * where the user explicitly disable indirect branch
>> +	 * speculation or mark the process as non-dumpable. 
> 
> s/where the user explicitly disable
>  /where the user explicitly disables
> 
> ?
> 
> Thanks,
> 
> 	Ingo
> 

Thanks for the corrections.  I'll update the patchset.

Tim