On Fri, Aug 10, 2018 at 09:25:45PM +0000, Bart Van Assche wrote:
> Hello,
>
> While running the srp tests from the blktests suite I encountered the below
> lockdep splat with kernel v4.18-rc8. Does this mean that downgrade_write()
> was called on a reader lock?
That does seem to be what that WARN is saying, but the code in my tree
is obivously not doing that:
down_write(&lists_rwsem);
list_del(&device->core_list);
spin_lock_irqsave(&device->client_data_lock, flags);
list_for_each_entry_safe(context, tmp, &device->client_data_list, list)
context->going_down = true;
spin_unlock_irqrestore(&device->client_data_lock, flags);
downgrade_write(&lists_rwsem);
No idea what is wrong there - memory corruption? lockdep bug?
Jason
