Re: [PATCH] perf/smmuv3: Remove the leftover put_cpu() in error path

[Will Deacon]

On Tue, Dec 10, 2019 at 09:55:28PM +0800, Hanjun Guo wrote:
> On 2019/12/10 21:24, Will Deacon wrote:
> > On Tue, Dec 10, 2019 at 06:46:24PM +0800, Hanjun Guo wrote:
> >> In smmu_pmu_probe(), there is put_cpu() in the error path,
> >> which is wrong because we use raw_smp_processor_id() to
> >> get the cpu ID, not get_cpu(), remove it.
> >>
> >> Signed-off-by: Hanjun Guo <guohanjun@xxxxxxxxxx>
> >> ---
> >>  drivers/perf/arm_smmuv3_pmu.c | 1 -
> >>  1 file changed, 1 deletion(-)
> >>
> >> diff --git a/drivers/perf/arm_smmuv3_pmu.c b/drivers/perf/arm_smmuv3_pmu.c
> >> index 773128f..fd1d46a 100644
> >> --- a/drivers/perf/arm_smmuv3_pmu.c
> >> +++ b/drivers/perf/arm_smmuv3_pmu.c
> >> @@ -834,7 +834,6 @@ static int smmu_pmu_probe(struct platform_device *pdev)
> >>  out_unregister:
> >>  	cpuhp_state_remove_instance_nocalls(cpuhp_state_num, &smmu_pmu->node);
> >>  out_cpuhp_err:
> >> -	put_cpu();
> >>  	return err;
> > 
> > Can we kill 'out_cpuhp_err' altogether then and just return err if we fail
> > to add the hotplug instance?
> 
> Makes sense, but I think we can go further to kill both 'out_cpuhp_err' and
> 'out_register' as below [1], what do you think?

Although that's functionally correct, I'd prefer to keep out_unregister(),
since it acts as good reminder to anybody extending this function in future
that they need to unregister the hotplug instance on failure.

Will