On Tue, Jun 11, 2019 at 03:04:46PM -0300, Jason Gunthorpe wrote:
> On Thu, Jun 06, 2019 at 01:53:41PM +0300, Leon Romanovsky wrote:
> > @@ -302,6 +318,46 @@ int rdma_counter_query_stats(struct rdma_counter *counter)
> > return ret;
> > }
> >
> > +static u64 get_running_counters_hwstat_sum(struct ib_device *dev,
> > + u8 port, u32 index)
> > +{
> > + struct rdma_restrack_entry *res;
> > + struct rdma_restrack_root *rt;
> > + struct rdma_counter *counter;
> > + unsigned long id = 0;
> > + u64 sum = 0;
> > +
> > + rt = &dev->res[RDMA_RESTRACK_COUNTER];
> > + xa_lock(&rt->xa);
> > + xa_for_each(&rt->xa, id, res) {
> > + counter = container_of(res, struct rdma_counter, res);
> > + if ((counter->device != dev) || (counter->port != port) ||
> > + rdma_counter_query_stats(counter))
> > + continue;
>
> rdma_counter_query_stats has:
>
> int rdma_counter_query_stats(struct rdma_counter *counter)
> +{
> + struct ib_device *dev = counter->device;
> + int ret;
> +
> + if (!dev->ops.counter_update_stats)
> + return -EINVAL;
> +
> + mutex_lock(&counter->lock);
> + ret = dev->ops.counter_update_stats(counter);
> + mutex_unlock(&counter->lock);
>
> so again, this can't work and would blow up with lockdep if it was
> ever tested. xa_lock is a spinlock, you can't nest mutex's inside
> spinlocks.
>
> Check all the xa_lock for this. No idea why you are not catching this
> during testing. Maybe some might_sleeps() are needed.
I will check it, but on my system LOCKDEP is enabled, /proc/lockdep
exists and full of information, so it is working.
Thanks
>
> Jason
