lkml.org 
[lkml]   [2015]   [May]   [18]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 4/7] x86/intel_rdt: Implement scheduling support for Intel RDT


On Fri, 15 May 2015, Thomas Gleixner wrote:

> On Mon, 11 May 2015, Vikas Shivappa wrote:
>> struct rdt_subsys_info {
>> /* Clos Bitmap to keep track of available CLOSids.*/
>> @@ -24,6 +30,11 @@ struct clos_cbm_map {
>> unsigned int clos_refcnt;
>> };
>>
>> +static inline bool rdt_enabled(void)
>> +{
>> + return static_key_false(&rdt_enable_key);
>
> So again, why this useless helper function for a single call site?
>
>> +static inline void intel_rdt_sched_in(void)
>> +{
>> + if (rdt_enabled())
>> + __rdt_sched_in();
>> +}
>
>> +void __rdt_sched_in(void)
>> +{
>> + struct task_struct *task = current;
>> + struct intel_rdt *ir;
>> + unsigned int clos;
>> +
>> + /*
>> + * This needs to be fixed
>> + * to cache the whole PQR instead of just CLOSid.
>> + * PQR has closid in high 32 bits and CQM-RMID in low 10 bits.
>> + * Should not write a 0 to the low 10 bits of PQR
>> + * and corrupt RMID.
>
> And why is this not fixed __BEFORE__ this patch? You can do the
> changes to struct intel_cqm_state in a seperate patch and then do the
> proper implementation from the beginning instead of providing a half
> broken variant which gets replaced in the next patch.

Ok , can fix both items in your comments. Reason I had it seperately is that the
cache affects both cmt and cache allocation patches.

>
> Thanks,
>
> tglx
>


\
 
 \ /
  Last update: 2015-05-18 20:41    [W:0.481 / U:0.236 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site