Messages in this thread | | | Date | Fri, 26 Aug 2016 14:44:53 +0200 | From | Oleg Nesterov <> | Subject | [PATCH 0/2] sched/wait: abort_exclusive_wait() should pass TASK_NORMAL to wake_up() |
| |
To clarify, this can't fix the lock_page() hang reported by Bart, __wait_on_bit_lock() is fine even if abort_exclusive_wait() is buggy.
Still should be fixed I think. And perhaps we should simply kill abort_exclusive_wait() ? See 2/2.
And this reminds me... I still fail to understand why/how the commit 68985633bccb60 "sched/wait: Fix signal handling in bit wait helpers" can help. Apart from s/return 1/return -EINTR/. Peter, do you have any theory?
Oleg.
| |