Re: [PATCH v2] perf/core: Avoid removing shared pmu_context on unregister

[Chris Wilson]

On Fri, May 12, 2017 at 01:40:37PM -0700, David Carrillo-Cisneros wrote:
> On Fri, May 12, 2017 at 4:45 AM, Chris Wilson <chris@xxxxxxxxxxxxxxxxxx> wrote:
> > In commit 1fd7e4169954 ("perf/core: Remove perf_cpu_context::unique_pmu"),
> > the search for another user of the pmu_cpu_context was removed, and so
> > we unconditionally free it during perf_pmu_unregister. This leads to
> > random corruption later and a BUG at mm/percpu.c:689.
> >
> > v2: Check for shared pmu_contexts under the mutex.
> >
> > Fixes: 1fd7e4169954 ("perf/core: Remove perf_cpu_context::unique_pmu")
> > Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> > Cc: David Carrillo-Cisneros <davidcc@xxxxxxxxxx>
> > Cc: Peter Zijlstra (Intel) <peterz@xxxxxxxxxxxxx>
> > Cc: Ingo Molnar <mingo@xxxxxxxxxx>
> > Cc: <stable@xxxxxxxxxxxxxxx> # v4.11+
> > ---
> >  kernel/events/core.c | 5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> >
> > diff --git a/kernel/events/core.c b/kernel/events/core.c
> > index aaefaa27e1a6..4f60f66b35ad 100644
> > --- a/kernel/events/core.c
> > +++ b/kernel/events/core.c
> > @@ -8983,10 +8983,12 @@ EXPORT_SYMBOL_GPL(perf_pmu_register);
> >  void perf_pmu_unregister(struct pmu *pmu)
> >  {
> >         int remove_device;
> > +       int remove_context;
> >
> >         mutex_lock(&pmus_lock);
> >         remove_device = pmu_bus_running;
> >         list_del_rcu(&pmu->entry);
> > +       remove_context = !find_pmu_context(pmu->task_ctx_nr);
> >         mutex_unlock(&pmus_lock);
> >
> >         /*
> > @@ -9005,7 +9007,8 @@ void perf_pmu_unregister(struct pmu *pmu)
> >                 device_del(pmu->dev);
> >                 put_device(pmu->dev);
> >         }
> > -       free_pmu_context(pmu);
> > +       if (remove_context)
> > +               free_pmu_context(pmu);
> >  }
> 
> Shouldn't be cleaner to keep the check in find_pmu_context, just as it
> was before commit 1fd7e4169954 ("perf/core: Remove
> perf_cpu_context::unique_pmu")?
> 
> (Code below untested)
> 
> diff --git a/kernel/events/core.c b/kernel/events/core.c
> index 6e75a5c9412d..50d90cbf8418 100644
> --- a/kernel/events/core.c
> +++ b/kernel/events/core.c
> @@ -8857,7 +8857,8 @@ static struct perf_cpu_context __percpu
> *find_pmu_context(int ctxn)
>  static void free_pmu_context(struct pmu *pmu)
>  {
>         mutex_lock(&pmus_lock);
> -       free_percpu(pmu->pmu_cpu_context);
> +       if (!find_pmu_context(pmu->task_ctx_nr))
> +               free_percpu(pmu->pmu_cpu_context);
>         mutex_unlock(&pmus_lock);

We have the problem that find_pmu_context looks for a matching
task_ctx_nr, but if a second pmu was registered since our list_del and
before our search, we would wrongly conclude that it was using our pmu
context, but it had actually allocated a new one for itself.

We could do a search by pmu_cpu_context instead, but seems overkill
compared to the remove_context approach.
-Chris

-- 
Chris Wilson, Intel Open Source Technology Centre