Re: [v6 1/1] x86/bugs: Implement mitigation for Predictive Store Forwarding

[Saripalli, RK]

On 5/17/2021 9:55 PM, Randy Dunlap wrote:
> Hi again,
> 
> On 5/17/21 3:00 PM, Ramakrishna Saripalli wrote:
>> From: Ramakrishna Saripalli <rk.saripalli@xxxxxxx>
>>
>> Certain AMD processors feature a new technology called Predictive Store
>> Forwarding (PSF).
>>
>> PSF is a micro-architectural optimization designed to improve the
>> performance of code execution by predicting dependencies between
>> loads and stores.
>>
>> Incorrect PSF predictions can occur due to two reasons.
>>
> ...
> 
>>
>> Kernel parameter predictive_store_fwd_disable has the following values
>>
>> - on. Disable PSF on all CPUs.
>>
>> - off. Enable PSF on all CPUs.
>>        This is also the default setting.
>>
>> Signed-off-by: Ramakrishna Saripalli<rk.saripalli@xxxxxxx>
>> ---
>>  .../admin-guide/kernel-parameters.txt         |  5 +
>>  arch/x86/include/asm/cpufeatures.h            |  1 +
>>  arch/x86/include/asm/msr-index.h              |  2 +
>>  arch/x86/include/asm/nospec-branch.h          |  6 ++
>>  arch/x86/kernel/cpu/bugs.c                    | 94 +++++++++++++++++++
>>  5 files changed, 108 insertions(+)
>>
>> diff --git a/Documentation/admin-guide/kernel-parameters.txt b/Documentation/admin-guide/kernel-parameters.txt
>> index 04545725f187..a5f694dccb24 100644
>> --- a/Documentation/admin-guide/kernel-parameters.txt
>> +++ b/Documentation/admin-guide/kernel-parameters.txt
>> @@ -3940,6 +3940,11 @@
>>  			Format: {"off"}
>>  			Disable Hardware Transactional Memory
>>  
>> +	predictive_store_fwd_disable=	[X86] This option controls PSF.
>> +			off - Turns on PSF.
>> +			on  - Turns off PSF.
>> +			default : off.
> 
> 
> and as I did earlier, I still object to "off" meaning PSF is on
> and "on" meaning that PSF is off.
> 
> It's not at all user friendly.
> 
> If it's done this way because that's how the h/w bit is defined/used,
> that's not a good excuse IMHO.
> 
> Hm, it sorta seems to be a common "theme" when dealing with mitigations.
> And too late to fix that.

Based on previous feedback from Thomas Gleixner, I reworded this as a mitigation instead of as a "feature".
In that vein, all the mitigation code moved into bugs.c like other mitigations, similar to
spec_control_bypass_disable with an ON/OFF but no prctl/seccomp/auto.


> 
> I look forward to h/w that doesn't need mitigations.  ;)
>