Commit d114dde2 authored by Alexei Starovoitov's avatar Alexei Starovoitov Committed by Daniel Borkmann
Browse files

bpf: Change bpf_mem_cache draining process. 



The next patch will introduce cross-cpu llist access and existing
irq_work_sync() + drain_mem_cache() + rcu_barrier_tasks_trace() mechanism will
not be enough, since irq_work_sync() + drain_mem_cache() on cpu A won't
guarantee that llist on cpu A are empty. The free_bulk() on cpu B might add
objects back to llist of cpu A. Add 'bool draining' flag.
The modified sequence looks like:
for_each_cpu:
  WRITE_ONCE(c->draining, true); // do_call_rcu_ttrace() won't be doing call_rcu() any more
  irq_work_sync(); // wait for irq_work callback (free_bulk) to finish
  drain_mem_cache(); // free all objects
rcu_barrier_tasks_trace(); // wait for RCU callbacks to execute

Signed-off-by: default avatarAlexei Starovoitov <ast@kernel.org>
Signed-off-by: default avatarDaniel Borkmann <daniel@iogearbox.net>
Acked-by: default avatarHou Tao <houtao1@huawei.com>
Link: https://lore.kernel.org/bpf/20230706033447.54696-8-alexei.starovoitov@gmail.com
parent 74680482

kernel/bpf/memalloc.c

+9 −9
Original line number Diff line number Diff line
@@ -98,6 +98,7 @@ struct bpf_mem_cache { 
	int free_cnt;

	int low_watermark, high_watermark, batch;

	int percpu_size;

	bool draining;



	/* list of objects to be freed after RCU tasks trace GP */

	struct llist_head free_by_rcu_ttrace;
@@ -301,6 +302,12 @@ static void do_call_rcu_ttrace(struct bpf_mem_cache *c) 
		 * from __free_rcu() and from drain_mem_cache().

		 */

		__llist_add(llnode, &c->waiting_for_gp_ttrace);



	if (unlikely(READ_ONCE(c->draining))) {

		__free_rcu(&c->rcu_ttrace);

		return;

	}



	/* Use call_rcu_tasks_trace() to wait for sleepable progs to finish.

	 * If RCU Tasks Trace grace period implies RCU grace period, free

	 * these elements directly, else use call_rcu() to wait for normal
@@ -544,15 +551,7 @@ void bpf_mem_alloc_destroy(struct bpf_mem_alloc *ma) 
		rcu_in_progress = 0;

		for_each_possible_cpu(cpu) {

			c = per_cpu_ptr(ma->cache, cpu);

			/*

			 * refill_work may be unfinished for PREEMPT_RT kernel

			 * in which irq work is invoked in a per-CPU RT thread.

			 * It is also possible for kernel with

			 * arch_irq_work_has_interrupt() being false and irq

			 * work is invoked in timer interrupt. So waiting for

			 * the completion of irq work to ease the handling of

			 * concurrency.

			 */

			WRITE_ONCE(c->draining, true);

			irq_work_sync(&c->refill_work);

			drain_mem_cache(c);

			rcu_in_progress += atomic_read(&c->call_rcu_ttrace_in_progress);
@@ -568,6 +567,7 @@ void bpf_mem_alloc_destroy(struct bpf_mem_alloc *ma) 
			cc = per_cpu_ptr(ma->caches, cpu);

			for (i = 0; i < NUM_CACHES; i++) {

				c = &cc->cache[i];

				WRITE_ONCE(c->draining, true);

				irq_work_sync(&c->refill_work);

				drain_mem_cache(c);

				rcu_in_progress += atomic_read(&c->call_rcu_ttrace_in_progress);