From: Mykyta Yatsenko <mykyta.yatsenko5@gmail.com>
To: bpf@vger.kernel.org, ast@kernel.org, andrii@kernel.org,
daniel@iogearbox.net, kafai@meta.com, kernel-team@meta.com,
memxor@gmail.com, eddyz87@gmail.com
Cc: Mykyta Yatsenko <yatsenko@meta.com>
Subject: [PATCH RFC v2 4/5] bpf: add refcnt into struct bpf_async_cb
Date: Wed, 05 Nov 2025 15:59:06 +0000 [thread overview]
Message-ID: <20251105-timer_nolock-v2-4-32698db08bfa@meta.com> (raw)
In-Reply-To: <20251105-timer_nolock-v2-0-32698db08bfa@meta.com>
From: Mykyta Yatsenko <yatsenko@meta.com>
To manage lifetime guarantees of the struct bpf_async_cb, when
no lock serializes mutations, introduce refcnt field into the struct.
Implement bpf_async_tryget() and bpf_async_put() to handle the refcnt.
Signed-off-by: Mykyta Yatsenko <yatsenko@meta.com>
---
kernel/bpf/helpers.c | 39 ++++++++++++++++++++++++++++++++-------
1 file changed, 32 insertions(+), 7 deletions(-)
diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index 2eb2369cae3ad34fd218387aa237140003cc1853..1cd4011faca519809264b2152c7c446269bee5de 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -1102,6 +1102,7 @@ struct bpf_async_cb {
struct work_struct delete_work;
};
u64 flags;
+ refcount_t refcnt;
};
/* BPF map elements can contain 'struct bpf_timer'.
@@ -1155,6 +1156,33 @@ static DEFINE_PER_CPU(struct bpf_hrtimer *, hrtimer_running);
static void bpf_timer_delete(struct bpf_hrtimer *t);
+static bool bpf_async_tryget(struct bpf_async_cb *cb)
+{
+ return refcount_inc_not_zero(&cb->refcnt);
+}
+
+static void bpf_async_put(struct bpf_async_cb *cb, enum bpf_async_type type)
+{
+ if (!refcount_dec_and_test(&cb->refcnt))
+ return;
+
+ switch (type) {
+ case BPF_ASYNC_TYPE_TIMER:
+ bpf_timer_delete((struct bpf_hrtimer *)cb);
+ break;
+ case BPF_ASYNC_TYPE_WQ: {
+ struct bpf_work *work = (void *)cb;
+ /* Trigger cancel of the sleepable work, but *do not* wait for
+ * it to finish if it was running as we might not be in a
+ * sleepable context.
+ * kfree will be called once the work has finished.
+ */
+ schedule_work(&work->delete_work);
+ break;
+ }
+ }
+}
+
static enum hrtimer_restart bpf_timer_cb(struct hrtimer *hrtimer)
{
struct bpf_hrtimer *t = container_of(hrtimer, struct bpf_hrtimer, timer);
@@ -1304,6 +1332,7 @@ static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u
cb->prog = NULL;
cb->flags = flags;
rcu_assign_pointer(cb->callback_fn, NULL);
+ refcount_set(&cb->refcnt, 1); /* map's own ref */
WRITE_ONCE(async->cb, cb);
/* Guarantee the order between async->cb and map->usercnt. So
@@ -1642,7 +1671,7 @@ void bpf_timer_cancel_and_free(void *val)
if (!t)
return;
- bpf_timer_delete(t);
+ bpf_async_put(&t->cb, BPF_ASYNC_TYPE_TIMER); /* Put map's own reference */
}
/* This function is called by map_delete/update_elem for individual element and
@@ -1657,12 +1686,8 @@ void bpf_wq_cancel_and_free(void *val)
work = (struct bpf_work *)__bpf_async_cancel_and_free(val);
if (!work)
return;
- /* Trigger cancel of the sleepable work, but *do not* wait for
- * it to finish if it was running as we might not be in a
- * sleepable context.
- * kfree will be called once the work has finished.
- */
- schedule_work(&work->delete_work);
+
+ bpf_async_put(&work->cb, BPF_ASYNC_TYPE_WQ); /* Put map's own reference */
}
BPF_CALL_2(bpf_kptr_xchg, void *, dst, void *, ptr)
--
2.51.1
next prev parent reply other threads:[~2025-11-05 15:59 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-11-05 15:59 [PATCH RFC v2 0/5] bpf: avoid locks in bpf_timer and bpf_wq Mykyta Yatsenko
2025-11-05 15:59 ` [PATCH RFC v2 1/5] bpf: refactor bpf_async_cb callback update Mykyta Yatsenko
2025-11-06 17:03 ` Kumar Kartikeya Dwivedi
2025-11-05 15:59 ` [PATCH RFC v2 2/5] bpf: refactor bpf_async_cb prog swap Mykyta Yatsenko
2025-11-06 16:57 ` Kumar Kartikeya Dwivedi
2025-11-05 15:59 ` [PATCH RFC v2 3/5] bpf: factor out timer deletion helper Mykyta Yatsenko
2025-11-06 16:58 ` Kumar Kartikeya Dwivedi
2025-11-05 15:59 ` Mykyta Yatsenko [this message]
2025-11-06 17:48 ` [PATCH RFC v2 4/5] bpf: add refcnt into struct bpf_async_cb Kumar Kartikeya Dwivedi
2025-11-06 21:41 ` Yonghong Song
2025-11-06 22:36 ` Mykyta Yatsenko
2025-11-05 15:59 ` [PATCH RFC v2 5/5] bpf: remove lock from bpf_async_cb Mykyta Yatsenko
2025-11-06 19:25 ` Kumar Kartikeya Dwivedi
2025-11-07 3:15 ` Alexei Starovoitov
2025-11-07 15:58 ` Mykyta Yatsenko
2025-11-07 20:21 ` Alexei Starovoitov
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20251105-timer_nolock-v2-4-32698db08bfa@meta.com \
--to=mykyta.yatsenko5@gmail.com \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=eddyz87@gmail.com \
--cc=kafai@meta.com \
--cc=kernel-team@meta.com \
--cc=memxor@gmail.com \
--cc=yatsenko@meta.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).