Re: [BUG] Use of probe_kernel_address() in task_rcu_dereference() without checking return value
From: Linus Torvalds
Date: Fri Aug 30 2019 - 11:43:46 EST
On Fri, Aug 30, 2019 at 8:40 AM Russell King - ARM Linux admin
> Ah, ok. Might be worth some comments - I find the comments in that
> function particularly unhelpful (even after Oleg mentions this is
> case 2.)
Yeah, a comment like "This might fault if we race with the task
scheduling away and being destroyed, but we check that below".
But that code has some performance issues too, as mentioned. Which is
all kinds of sad since it clearly _tries_ to perform well with RCU
locking and optimistic accesses etc.