Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752834AbdHAUwT (ORCPT ); Tue, 1 Aug 2017 16:52:19 -0400 Received: from mail-wr0-f196.google.com ([209.85.128.196]:37406 "EHLO mail-wr0-f196.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752549AbdHAUwR (ORCPT ); Tue, 1 Aug 2017 16:52:17 -0400 Message-ID: <1501620734.25002.10.camel@edumazet-glaptop3.roam.corp.google.com> Subject: Re: [PATCH 2/3] jump_labels: do not use unserialized static_key_enabled From: Eric Dumazet To: Paolo Bonzini Cc: linux-kernel@vger.kernel.org, Peter Zijlstra Date: Tue, 01 Aug 2017 13:52:14 -0700 In-Reply-To: <1501601046-35683-3-git-send-email-pbonzini@redhat.com> References: <1501601046-35683-1-git-send-email-pbonzini@redhat.com> <1501601046-35683-3-git-send-email-pbonzini@redhat.com> Content-Type: text/plain; charset="UTF-8" X-Mailer: Evolution 3.10.4-0ubuntu2 Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2136 Lines: 51 On Tue, 2017-08-01 at 17:24 +0200, Paolo Bonzini wrote: > Any use of key->enabled (that is static_key_enabled and static_key_count) > outside jump_label_lock should handle its own serialization. The only > two that are not doing so are the UDP encapsulation static keys. Change > them to use static_key_enable, which now correctly tests key->enabled under > the jump label lock. > > Cc: Peter Zijlstra > Cc: Eric Dumazet > Signed-off-by: Paolo Bonzini > --- > Documentation/static-keys.txt | 5 +++++ > net/ipv4/udp.c | 3 +-- > net/ipv6/udp.c | 3 +-- > 3 files changed, 7 insertions(+), 4 deletions(-) > > diff --git a/Documentation/static-keys.txt b/Documentation/static-keys.txt > index ef419fd0897f..181998852961 100644 > --- a/Documentation/static-keys.txt > +++ b/Documentation/static-keys.txt > @@ -142,6 +142,11 @@ static_branch_inc(), will change the branch back to true. Likewise, if the > key is initialized false, a 'static_branch_inc()', will change the branch to > true. And then a 'static_branch_dec()', will again make the branch false. > > +The state and the reference count can be retrieved with 'static_key_enabled()' > +and 'static_key_count()'. In general, if you use these functions, they > +should be protected with the same mutex used around the enable/disable > +or increment/decrement function. > + > Where an array of keys is required, it can be defined as: > > DEFINE_STATIC_KEY_ARRAY_TRUE(keys, count); > diff --git a/net/ipv4/udp.c b/net/ipv4/udp.c > index 1d6219bf2d6b..74b7810df9fc 100644 > --- a/net/ipv4/udp.c > +++ b/net/ipv4/udp.c > @@ -1644,8 +1644,7 @@ static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb) > static struct static_key udp_encap_needed __read_mostly; > void udp_encap_enable(void) > { > - if (!static_key_enabled(&udp_encap_needed)) > - static_key_slow_inc(&udp_encap_needed); > + static_key_enable(&udp_encap_needed); > } Looks good to me, but static_key_enable() is not serialized either ? I suspect you should have CCed me on patch 1/3 :)