On 11/28, Oleg Nesterov wrote: > > On a related note... Neil, Al, et al, can you look at > > [PATCH 1/3] fput: don't abuse task_work_add() when possible > https://lore.kernel.org/all/20150908171446.GA14589@xxxxxxxxxx/ Cough... Now that I look at this 8 years old patch again I think it is wrong, fput() can race with task_work_cancel() so it is not safe to dereference the first pending work in theory :/ Oleg.
