On Tue, May 31, 2016 at 05:28:34PM +0100, Will Deacon wrote:
> On Tue, May 31, 2016 at 12:41:22PM +0100, Julien Grall wrote:
> > The global variable __oprofile_cpu_pmu is set before the PMU is fully
> > initialized. If an error occurs before the end of the initialization,
> > the PMU will be freed and the variable will contain an invalid pointer.
> >
> > This will result in a kernel crash when perf will be used.
> >
> > Fix it by moving the setting of __oprofile_cpu_pmu when the PMU is fully
> > initialized (i.e when it is no longer possible to fail).
> >
> > Signed-off-by: Julien Grall <[email protected]>
> > Acked-by: Mark Rutland <[email protected]>
> > ---
> > drivers/perf/arm_pmu.c | 6 +++---
> > 1 file changed, 3 insertions(+), 3 deletions(-)
>
> Should this one go to -stable too?
I think so.
The bug has been there at least since 76b8a0e4c8bda5f0 ("ARM: perf:
handle armpmu_register failing"), in v3.8...
Prior to that we wouldn't free the PMU, but it might not have been
initialised correctly.
Thanks,
Mark.
> Will
>
> > diff --git a/drivers/perf/arm_pmu.c b/drivers/perf/arm_pmu.c
> > index 6401f0c..95614d2 100644
> > --- a/drivers/perf/arm_pmu.c
> > +++ b/drivers/perf/arm_pmu.c
> > @@ -992,9 +992,6 @@ int arm_pmu_device_probe(struct platform_device *pdev,
> >
> > armpmu_init(pmu);
> >
> > - if (!__oprofile_cpu_pmu)
> > - __oprofile_cpu_pmu = pmu;
> > -
> > pmu->plat_device = pdev;
> >
> > if (node && (of_id = of_match_node(of_table, pdev->dev.of_node))) {
> > @@ -1030,6 +1027,9 @@ int arm_pmu_device_probe(struct platform_device *pdev,
> > if (ret)
> > goto out_destroy;
> >
> > + if (!__oprofile_cpu_pmu)
> > + __oprofile_cpu_pmu = pmu;
> > +
> > pr_info("enabled with %s PMU driver, %d counters available\n",
> > pmu->name, pmu->num_events);
> >
> > --
> > 1.9.1
> >
>
