On Tue, Oct 08, 2019 at 02:51:20PM +0200, Pavel Machek wrote: > For example this did not have any locking, and is now protected by > > get_device(&chip->dev); > > down_read(&chip->ops_sem); > > . Is that intended? Is this known to fix any bugs? It is, sysfs code can otherwise race when ops is set NULL in class shutdown. /Jarkko
