From: Joel Fernandes <joelaf@google.com>
To: Daniel Colascione <dancol@google.com>
Cc: linux-kernel@vger.kernel.org, timmurray@google.com,
netdev@vger.kernel.org,
Alexei Starovoitov <alexei.starovoitov@gmail.com>,
Lorenzo Colitti <lorenzo@google.com>,
Chenbo Feng <fengc@google.com>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
Alexei Starovoitov <ast@fb.com>,
Daniel Borkmann <daniel@iogearbox.net>,
kernel-team@android.com
Subject: Re: [PATCH v4] Wait for running BPF programs when updating map-in-map
Date: Fri, 12 Oct 2018 13:54:28 -0700 [thread overview]
Message-ID: <20181012205428.GA32209@joelaf.mtv.corp.google.com> (raw)
In-Reply-To: <20181012105427.243779-1-dancol@google.com>
On Fri, Oct 12, 2018 at 03:54:27AM -0700, Daniel Colascione wrote:
> The map-in-map frequently serves as a mechanism for atomic
> snapshotting of state that a BPF program might record. The current
> implementation is dangerous to use in this way, however, since
> userspace has no way of knowing when all programs that might have
> retrieved the "old" value of the map may have completed.
>
> This change ensures that map update operations on map-in-map map types
> always wait for all references to the old map to drop before returning
> to userspace.
>
> Signed-off-by: Daniel Colascione <dancol@google.com>
> ---
> kernel/bpf/syscall.c | 14 ++++++++++++++
> 1 file changed, 14 insertions(+)
>
> diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
> index 8339d81cba1d..d7c16ae1e85a 100644
> --- a/kernel/bpf/syscall.c
> +++ b/kernel/bpf/syscall.c
> @@ -741,6 +741,18 @@ static int map_lookup_elem(union bpf_attr *attr)
> return err;
> }
>
> +static void maybe_wait_bpf_programs(struct bpf_map *map)
> +{
> + /* Wait for any running BPF programs to complete so that
> + * userspace, when we return to it, knows that all programs
> + * that could be running use the new map value.
> + */
> + if (map->map_type == BPF_MAP_TYPE_HASH_OF_MAPS ||
> + map->map_type == BPF_MAP_TYPE_ARRAY_OF_MAPS) {
> + synchronize_rcu();
> + }
> +}
> +
> #define BPF_MAP_UPDATE_ELEM_LAST_FIELD flags
>
> static int map_update_elem(union bpf_attr *attr)
> @@ -831,6 +843,7 @@ static int map_update_elem(union bpf_attr *attr)
> }
> __this_cpu_dec(bpf_prog_active);
> preempt_enable();
> + maybe_wait_bpf_programs(map);
> out:
> free_value:
> kfree(value);
> @@ -883,6 +896,7 @@ static int map_delete_elem(union bpf_attr *attr)
> rcu_read_unlock();
> __this_cpu_dec(bpf_prog_active);
> preempt_enable();
> + maybe_wait_bpf_programs(map);
Looks good to me,
Reviewed-by: Joel Fernandes (Google) <joel@joelfernandes.org>
Also I believe that those rcu_read_lock() and rcu_read_unlock() calls in the
existing code are useless. preempt_disable()d code is already an RCU
read-side section, and synchronize_rcu and friends work on those type of
read-side sections as well (as of recent kernel releases) however removing it
may make lockdep unhappy, unless we also replace all rcu_dereference() usages
with rcu_dereference_sched(), so lets leave that alone for now I guess.
thanks,
- Joel
next prev parent reply other threads:[~2018-10-12 20:54 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <CAKOZuesQ6GdNTGDFsNi4o8LYzxLBtYZ=Cz4=aZbqqCNia+QFnQ@mail.gmail.com>
2018-07-29 20:58 ` [PATCH v3] Add BPF_SYNCHRONIZE_MAP_TO_MAP_REFERENCES bpf(2) command Daniel Colascione
2018-07-30 10:04 ` Daniel Borkmann
2018-07-30 10:25 ` Daniel Colascione
2018-07-31 0:26 ` Jakub Kicinski
2018-07-31 0:33 ` Daniel Colascione
2018-07-31 0:45 ` Jakub Kicinski
2018-07-31 0:50 ` Daniel Colascione
2018-07-31 1:14 ` Jakub Kicinski
2018-07-31 8:34 ` Daniel Borkmann
2018-07-31 9:36 ` Daniel Colascione
2018-08-10 22:52 ` Alexei Starovoitov
2018-08-14 20:37 ` Daniel Colascione
2018-08-16 4:01 ` Alexei Starovoitov
2018-10-12 10:54 ` [PATCH v4] Wait for running BPF programs when updating map-in-map Daniel Colascione
2018-10-12 20:54 ` Joel Fernandes [this message]
2018-10-13 2:31 ` Alexei Starovoitov
2018-10-16 17:39 ` Joel Fernandes
2018-10-18 15:46 ` Alexei Starovoitov
2018-10-18 23:36 ` Joel Fernandes
2018-11-10 2:01 ` Chenbo Feng
2018-11-10 15:22 ` Greg KH
2018-11-10 18:58 ` Chenbo Feng
[not found] ` <CAKOZues6SE_c=ix7ap6QaJHqd1TmYpWWMJiu3=TtuqgKuqOUCA@mail.gmail.com>
2018-08-10 22:29 ` [PATCH v3] Add BPF_SYNCHRONIZE_MAP_TO_MAP_REFERENCES bpf(2) command 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=20181012205428.GA32209@joelaf.mtv.corp.google.com \
--to=joelaf@google.com \
--cc=alexei.starovoitov@gmail.com \
--cc=ast@fb.com \
--cc=dancol@google.com \
--cc=daniel@iogearbox.net \
--cc=fengc@google.com \
--cc=kernel-team@android.com \
--cc=linux-kernel@vger.kernel.org \
--cc=lorenzo@google.com \
--cc=mathieu.desnoyers@efficios.com \
--cc=netdev@vger.kernel.org \
--cc=timmurray@google.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).