Re: [PATCH 3/6] x86/mbm: Intel Memory B/W Monitoring enumeration and init

[Vikas Shivappa]

On Tue, 8 Mar 2016, Thomas Gleixner wrote:

> On Thu, 3 Mar 2016, Thomas Gleixner wrote:
> > On Thu, 3 Mar 2016, Vikas Shivappa wrote:
> >
> > > On Wed, 2 Mar 2016, Thomas Gleixner wrote:
> > >
> > > > On Wed, 2 Mar 2016, Vikas Shivappa wrote:
> > > > > +	if (cqm_enabled && mbm_enabled)
> > > > > +		intel_cqm_events_group.attrs = intel_cmt_mbm_events_attr;
> > > > > +	else if (!cqm_enabled && mbm_enabled)
> > > > > +		intel_cqm_events_group.attrs = intel_mbm_events_attr;
> > > > > +	else if (cqm_enabled && !mbm_enabled)
> > > > > +		intel_cqm_events_group.attrs = intel_cqm_events_attr;
> > > > > +
> > > > >  	ret = perf_pmu_register(&intel_cqm_pmu, "intel_cqm", -1);
> > > > >  	if (ret) {
> > > > >  		pr_err("Intel CQM perf registration failed: %d\n", ret);
> > > > >  		goto out;
> > > >
> > > > So what cleans up mbm_local and mbm_total in that case?
> > >
> > > I put all the cleanup code in the cqm_cleanup .. - please see copy below
> >
> > Ok, missed that.
> >
> > > @@ -1331,6 +1427,39 @@ static void cqm_cleanup(void)
> > >                 kfree(cqm_rmid_ptrs[r]);
> > >
> > >         kfree(cqm_rmid_ptrs);
> > > +       kfree(mbm_local);
> > > +       kfree(mbm_total);
> > > +       mbm_enabled = false;
> > > +       cqm_enabled = false;
> > > +}
>
> After looking at it closely, you really need a separate mbm_cleanup()
> function, so you can utilize it from mbm_init() and in the exit path.

Will fix and seperate the mbm and cqm cleanups..

Thanks,
Vikas

> Thanks,
>
> 	tglx