Re: [PATCH V4.1] mcheck, vmce: Allow vmce_amd_* functions to handle AMD thresolding MSRs

[Aravind Gopalakrishnan <aravind.gopalakrishnan@amd.com>]

On 2/13/2014 11:27 AM, Aravind Gopalakrishnan wrote:
> On 2/13/2014 2:38 AM, Jan Beulich wrote:
>>>     *val = 0;
>>>   -    switch ( msr & (MSR_IA32_MC0_CTL | 3) )
>>> +    /* Allow only first 3 MC banks into switch() */
>>> +    switch ( msr & (-MSR_IA32_MC0_CTL | 3) )
>>>       {
>>>       case MSR_IA32_MC0_CTL:
>>>           /* stick all 1's to MCi_CTL */
>> I'm confused: You now add a comment as if the mask was including
>> bit 4, which it doesn't. What am I missing?
>
> Darn. Sorry about that. Will fix..

Jan,

Do let me know if the following wording is fine:

/*
  * Apply mask to allow bits[0:1] (necessary to uniquely identify MC0)
  * MC1 is handled by virtue of 'bank' value.
  */

If not, I'm open to suggestions:)

Thanks,
-Aravind.