Re: [PATCH AUTOSEL 4.9 4/4] thermal: intel_powerclamp: Use get_cpu() instead of smp_processor_id() to avoid crash

[Rafael J. Wysocki]

On Thu, Oct 13, 2022 at 5:06 AM srinivas pandruvada
<srinivas.pandruvada@xxxxxxxxxxxxxxx> wrote:
>
> On Wed, 2022-10-12 at 18:58 +0200, Rafael J. Wysocki wrote:
> > On Tue, Oct 11, 2022 at 3:23 PM Chen Yu <yu.c.chen@xxxxxxxxx> wrote:
> > >
> > > Hi Pavel,
> > > On 2022-10-11 at 13:36:46 +0200, Pavel Machek wrote:
> > > > Hi!
> > > >
> > > > > From: Srinivas Pandruvada <srinivas.pandruvada@xxxxxxxxxxxxxxx>
> > > > >
> > > > > [ Upstream commit 68b99e94a4a2db6ba9b31fe0485e057b9354a640 ]
> > > > >
> > > > > When CPU 0 is offline and intel_powerclamp is used to inject
> > > > > idle, it generates kernel BUG:
> > > > >
> > > > > BUG: using smp_processor_id() in preemptible [00000000] code:
> > > > > bash/15687
> > > > > caller is debug_smp_processor_id+0x17/0x20
> > > > > CPU: 4 PID: 15687 Comm: bash Not tainted 5.19.0-rc7+ #57
> > > > > Call Trace:
> > > > > <TASK>
> > > > > dump_stack_lvl+0x49/0x63
> > > > > dump_stack+0x10/0x16
> > > > > check_preemption_disabled+0xdd/0xe0
> > > > > debug_smp_processor_id+0x17/0x20
> > > > > powerclamp_set_cur_state+0x7f/0xf9 [intel_powerclamp]
> > > > > ...
> > > > > ...
> > > > >
> > > > > Here CPU 0 is the control CPU by default and changed to the
> > > > > current CPU,
> > > > > if CPU 0 offlined. This check has to be performed under
> > > > > cpus_read_lock(),
> > > > > hence the above warning.
> > > > >
> > > > > Use get_cpu() instead of smp_processor_id() to avoid this BUG.
> > > >
> > > > This has exactly the same problem as smp_processor_id(), you just
> > > > worked around the warning. If it is okay that control_cpu
> > > > contains
> > > > stale value, could we have a comment explaining why?
> > > >
> > > May I know why does control_cpu have stale value? The control_cpu
> > > is a random picked online CPU which will be used later to collect
> > > statistics.
> > > As long as the control_cpu is online, it is valid IMO.
> >
> > So this is confusing, because the code makes the impression that
> > getting the number of the CPU running the code matters in some way,
> > which isn't the case.
> >
> > Something like cpumask_first(cpu_online_mask) should work as well if
> > I'm not mistaken and it would be less confusing to use this instead
> > IMO.
> That should work as we are under hotplug lock anyway here.

Well, that's my point.

I guess I'll send a patch with this change.