On Wed 2018-10-17 13:22:54, Josh Poimboeuf wrote: > On Mon, Oct 15, 2018 at 02:37:05PM +0200, Petr Mladek wrote: > > @@ -637,6 +647,7 @@ static int klp_init_patch(struct klp_patch *patch) > > mutex_lock(&klp_mutex); > > > > patch->enabled = false; > > + INIT_LIST_HEAD(&patch->list); > > Is this a bug fix? If so, it should go in a separate patch. I'm > wondering how this worked before. It is not a bug fix. It was not needed before because klp_register_patch() did not checked this field when failed. But this patch started to use the generic klp_free_patch_start()/klp_free_patch_finish() in both klp_unregister_patch() and in klp_init_patch() error patch. Therefore we need to have a valid value from the beginning. Best Regards, Petr
