bpf: Use rcu_trace_implies_rcu_gp() in bpf memory allocator
authorHou Tao <houtao1@huawei.com>
Fri, 14 Oct 2022 11:39:44 +0000 (19:39 +0800)
committerAlexei Starovoitov <ast@kernel.org>
Tue, 18 Oct 2022 17:27:02 +0000 (10:27 -0700)
commit59be91e5e70a1aa91dfee8088b071f6d05c8a1a3
tree676fe0a850775286ed89bf9a6d6b8801cdb33e75
parente6c86c513f440bec5f1046539c7e3c6c653842da
bpf: Use rcu_trace_implies_rcu_gp() in bpf memory allocator

The memory free logic in bpf memory allocator chains a RCU Tasks Trace
grace period and a normal RCU grace period one after the other, so it
can ensure that both sleepable and non-sleepable programs have finished.

With the introduction of rcu_trace_implies_rcu_gp(),
__free_rcu_tasks_trace() can check whether or not a normal RCU grace
period has also passed after a RCU Tasks Trace grace period has passed.
If it is true, freeing these elements directly, else freeing through
call_rcu().

Signed-off-by: Hou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/r/20221014113946.965131-3-houtao@huaweicloud.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
kernel/bpf/memalloc.c