29.08.2013, 22:28, "Sebastian Andrzej Siewior" <bigeasy@xxxxxxxxxxxxx>: > On 08/29/2013 07:26 PM, Alexander Fyodorov wrote: >> Still this is only half of the solution because the patch doesn't solve the race in wait_task_inactive() (and all other places which test both state and saved_state without holding pi_lock). > > So you are concerned that missing pi_lock in wait_task_inactive(). This > is a problem if the task wakes up from sleeping on the lock while its > state is beeing checked. Hmm it indeed looks legal. > I keep that patch in queue but disabled and take another look once I > get back. > Does this missing pi_lock() affects you or is just a precaution? Yes, my test application failed because of it, a detailed description is in the first message in the thread. I've also looked at other places that test saved_state but I don't understand the code enough to decide whether there are similar bugs there or not. -- To unsubscribe from this list: send the line "unsubscribe linux-rt-users" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
