From: "luwei (O)" <luwei32@huawei.com>
To: Alexei Starovoitov <alexei.starovoitov@gmail.com>, <davem@davemloft.net>
Cc: <daniel@iogearbox.net>, <andrii@kernel.org>,
<netdev@vger.kernel.org>, <bpf@vger.kernel.org>,
<kernel-team@fb.com>
Subject: Re: [PATCH v4 bpf-next 1/9] bpf: Introduce bpf timers.
Date: Tue, 6 Jul 2021 21:57:52 +0800 [thread overview]
Message-ID: <bcdccd37-6372-859f-824e-c96250361904@huawei.com> (raw)
In-Reply-To: <20210701192044.78034-2-alexei.starovoitov@gmail.com>
在 2021/7/2 3:20 AM, Alexei Starovoitov 写道:
> From: Alexei Starovoitov <ast@kernel.org>
>
> Introduce 'struct bpf_timer { __u64 :64; __u64 :64; };' that can be embedded
> in hash/array/lru maps as a regular field and helpers to operate on it:
>
> // Initialize the timer.
> // First 4 bits of 'flags' specify clockid.
> // Only CLOCK_MONOTONIC, CLOCK_REALTIME, CLOCK_BOOTTIME are allowed.
> long bpf_timer_init(struct bpf_timer *timer, struct bpf_map *map, int flags);
>
> // Configure the timer to call 'callback_fn' static function.
> long bpf_timer_set_callback(struct bpf_timer *timer, void *callback_fn);
>
> // Arm the timer to expire 'nsec' nanoseconds from the current time.
> long bpf_timer_start(struct bpf_timer *timer, u64 nsec, u64 flags);
>
> // Cancel the timer and wait for callback_fn to finish if it was running.
> long bpf_timer_cancel(struct bpf_timer *timer);
>
[...]
> +static enum hrtimer_restart bpf_timer_cb(struct hrtimer *hrtimer)
> +{
> + struct bpf_hrtimer *t = container_of(hrtimer, struct bpf_hrtimer, timer);
> + struct bpf_map *map = t->map;
> + void *value = t->value;
> + struct bpf_timer_kern *timer = value + map->timer_off;
> + struct bpf_prog *prog;
> + void *callback_fn;
> + void *key;
> + u32 idx;
> + int ret;
> +
> + /* The triple underscore bpf_spin_lock is a direct call
> + * to BPF_CALL_1(bpf_spin_lock) which does irqsave.
> + */
> + ____bpf_spin_lock(&timer->lock);
> + /* callback_fn and prog need to match. They're updated together
> + * and have to be read under lock.
> + */
> + prog = t->prog;
> + callback_fn = t->callback_fn;
> +
> + /* wrap bpf subprog invocation with prog->refcnt++ and -- to make
> + * sure that refcnt doesn't become zero when subprog is executing.
> + * Do it under lock to make sure that bpf_timer_start doesn't drop
> + * prev prog refcnt to zero before timer_cb has a chance to bump it.
> + */
> + bpf_prog_inc(prog);
> + ____bpf_spin_unlock(&timer->lock);
> +
> + /* bpf_timer_cb() runs in hrtimer_run_softirq. It doesn't migrate and
> + * cannot be preempted by another bpf_timer_cb() on the same cpu.
> + * Remember the timer this callback is servicing to prevent
> + * deadlock if callback_fn() calls bpf_timer_cancel() on the same timer.
> + */
> + this_cpu_write(hrtimer_running, t);
> + if (map->map_type == BPF_MAP_TYPE_ARRAY) {
> + struct bpf_array *array = container_of(map, struct bpf_array, map);
> +
> + /* compute the key */
> + idx = ((char *)value - array->value) / array->elem_size;
> + key = &idx;
> + } else { /* hash or lru */
> + key = value - round_up(map->key_size, 8);
> + }
> +
> + ret = BPF_CAST_CALL(callback_fn)((u64)(long)map,
> + (u64)(long)key,
> + (u64)(long)value, 0, 0);
> + WARN_ON(ret != 0); /* Next patch moves this check into the verifier */
> + bpf_prog_put(prog);
> +
> + this_cpu_write(hrtimer_running, NULL);
> + return HRTIMER_NORESTART;
> +}
> +
Your patch is awesome, and I tried your example and it works, my
call_back() is called after a certained time. But I am confued that
the bpf_timer_cb() function only returns HRTIMER_NORESTART,
how to trigger a periodic task, that means I want to call my call_back()
periodically, not just for one time ?
[...]
--
Best Regards,
Lu Wei
next prev parent reply other threads:[~2021-07-06 14:55 UTC|newest]
Thread overview: 17+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-07-01 19:20 [PATCH v4 bpf-next 0/9] bpf: Introduce BPF timers Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 1/9] bpf: Introduce bpf timers Alexei Starovoitov
2021-07-02 1:04 ` Martin KaFai Lau
2021-07-04 14:19 ` Alexei Starovoitov
2021-07-07 1:38 ` Alexei Starovoitov
2021-07-06 13:57 ` luwei (O) [this message]
2021-07-06 15:44 ` Toke Høiland-Jørgensen
2021-07-01 19:20 ` [PATCH v4 bpf-next 2/9] bpf: Add map side support for " Alexei Starovoitov
2021-07-02 6:23 ` Martin KaFai Lau
2021-07-04 14:23 ` Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 3/9] bpf: Prevent pointer mismatch in bpf_timer_init Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 4/9] bpf: Remember BTF of inner maps Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 5/9] bpf: Relax verifier recursion check Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 6/9] bpf: Implement verifier support for validation of async callbacks Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 7/9] bpf: Teach stack depth check about " Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 8/9] selftests/bpf: Add bpf_timer test Alexei Starovoitov
2021-07-01 19:20 ` [PATCH v4 bpf-next 9/9] selftests/bpf: Add a test with bpf_timer in inner map 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=bcdccd37-6372-859f-824e-c96250361904@huawei.com \
--to=luwei32@huawei.com \
--cc=alexei.starovoitov@gmail.com \
--cc=andrii@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=davem@davemloft.net \
--cc=kernel-team@fb.com \
--cc=netdev@vger.kernel.org \
/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).