On Mon, Oct 23, 2017 at 09:24:26AM +0100, Mark Brown wrote:
On Sat, Oct 21, 2017 at 05:05:13PM +0530, Vinod Koul wrote:
On Sat, Oct 21, 2017 at 10:12:30AM +0100, Mark Brown wrote:
On Thu, Oct 19, 2017 at 08:33:19AM +0530, Vinod Koul wrote:
- mutex_lock(&bus->bus_lock);
- if (!list_empty(&bus->slaves))
list_del(&slave->node);Shouldn't that be a while? Or at least warn if there's anything extra there. The code just looks very wrong as is.
I think you missed that it is called from sdw_delete_bus_master() which does the loop by invoking device_for_each_child(), so this ones is supposed to ensure one Slave is removed cleaned.
Let me know if i misread your comment.
My point is that this code just looks so obviously wrong it doesn't matter if it actually works, we should refactor so people don't look at the code and immediately think they've spotted a bug.
Okay will check and try to refactor this part.