On Thu, Dec 08, 2022 at 09:27:05AM +0800, Hou Tao wrote:
> Hi,
>
> On 12/8/2022 6:28 AM, Paul E. McKenney wrote:
> > On Wed, Dec 07, 2022 at 10:24:55AM +0800, Hou Tao wrote:
> >> Forget to cc Paul and RCU maillist for more comments.
> >>
> >> On 12/6/2022 12:29 PM, Hou Tao wrote:
> >>> From: Hou Tao <houtao1@xxxxxxxxxx>
> >>>
> >>> If there are pending rcu callback, free_mem_alloc() will use
> >>> rcu_barrier_tasks_trace() and rcu_barrier() to wait for the pending
> >>> __free_rcu_tasks_trace() and __free_rcu() callback.
> >>>
> >>> If rcu_trace_implies_rcu_gp() is true, there will be no pending
> >>> __free_rcu(), so it will be OK to skip rcu_barrier() as well.
> > The bit about there being no pending __free_rcu() is guaranteed by
> > your algorithm, correct? As in you have something like this somewhere
> > else in the code?
> >
> > if (!rcu_trace_implies_rcu_gp())
> > call_rcu(...);
> >
> > Or am I missing something more subtle?
> Yes. It is guaranteed by the implementation of bpf mem allocator: if
> rcu_trace_implies_rcu_gp() is true, there will be no call_rcu() in bpf memory
> allocator.
Very well, then from an RCU perspective:
Acked-by: Paul E. McKenney <paulmck@xxxxxxxxxx>
Thanx, Paul
> >>> Signed-off-by: Hou Tao <houtao1@xxxxxxxxxx>
> >>> ---
> >>> kernel/bpf/memalloc.c | 10 +++++++++-
> >>> 1 file changed, 9 insertions(+), 1 deletion(-)
> >>>
> >>> diff --git a/kernel/bpf/memalloc.c b/kernel/bpf/memalloc.c
> >>> index 7daf147bc8f6..d43991fafc4f 100644
> >>> --- a/kernel/bpf/memalloc.c
> >>> +++ b/kernel/bpf/memalloc.c
> >>> @@ -464,9 +464,17 @@ static void free_mem_alloc(struct bpf_mem_alloc *ma)
> >>> {
> >>> /* waiting_for_gp lists was drained, but __free_rcu might
> >>> * still execute. Wait for it now before we freeing percpu caches.
> >>> + *
> >>> + * rcu_barrier_tasks_trace() doesn't imply synchronize_rcu_tasks_trace(),
> >>> + * but rcu_barrier_tasks_trace() and rcu_barrier() below are only used
> >>> + * to wait for the pending __free_rcu_tasks_trace() and __free_rcu(),
> >>> + * so if call_rcu(head, __free_rcu) is skipped due to
> >>> + * rcu_trace_implies_rcu_gp(), it will be OK to skip rcu_barrier() by
> >>> + * using rcu_trace_implies_rcu_gp() as well.
> >>> */
> >>> rcu_barrier_tasks_trace();
> >>> - rcu_barrier();
> >>> + if (!rcu_trace_implies_rcu_gp())
> >>> + rcu_barrier();
> >>> free_mem_alloc_no_barrier(ma);
> >>> }
> >>>
>
