From mboxrd@z Thu Jan 1 00:00:00 1970 From: =?iso-8859-1?Q?Fran=E7ois-Fr=E9d=E9ric_Ozog?= Subject: Re: Bit spinlocks in DPDK Date: Fri, 20 Dec 2013 17:00:43 +0100 Message-ID: <04be01cefd9c$a3c36460$eb4a2d20$@com> References: <6895EAE0CA8DEE40B92D7CA88BB521F332BA572E6B@HQ1-EXCH02.corp.brocade.com> <4656219.tgqzelRNOJ@x220> <01d001cef375$7167e300$5437a900$@com> <201312201639.05277.thomas.monjalon@6wind.com> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Cc: dev-VfR2kkLFssw@public.gmane.org To: "'Thomas Monjalon'" Return-path: In-Reply-To: <201312201639.05277.thomas.monjalon-pdR9zngts4EAvxtiuMwx3w@public.gmane.org> Content-Language: fr List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces-VfR2kkLFssw@public.gmane.org Sender: "dev" > -----Message d'origine----- > De=A0: Thomas Monjalon [mailto:thomas.monjalon-pdR9zngts4EAvxtiuMwx3w@public.gmane.org] > Envoy=E9=A0: vendredi 20 d=E9cembre 2013 16:39 > =C0=A0: Fran=E7ois-Fr=E9d=E9ric Ozog > Cc=A0: dev-VfR2kkLFssw@public.gmane.org > Objet=A0: Re: [dpdk-dev] Bit spinlocks in DPDK >=20 > Hello, >=20 > 07/12/2013 18:54, Fran=E7ois-Fr=E9d=E9ric Ozog : > > 1) If the critical section deals with weakly ordered loads then > > explicit fencing MUST be used: if not, out of order execution will > > just kill your idea of critical section. > [...] > > So use rte_mb() or rte_wmb() or rte_rmb() where appropriate. I > > recommend the rte_unlock code and documentation explains the out of > > order execution issues and the conditions they have to be mitigated > > with rte*mb(). I wonder if having an explicit mfence in > > rte_sinlock_unlock wouldn't be just necessary to avoid "hairy" bugs. > > In addition, we would have rte_sinlock_unlock_no_mb used internally > > for performance reasons, and usable externally by advanced users. >=20 > Using lock prefix is lighter than using memory barrier and have the = same > effects. Well, in general yes BUT Intel states "../.. locked operations serialize = all outstanding load and store operations ../.. with one exception. Load operations that reference weakly ordered memory types (such as the WC = memory type) may *not* be serialized" in 8.1.2.2 Software Controlled Bus = Locking; particularly if streaming loads are used (may happen on certain devices memory mapped I/O accesses and the compiler generating streaming loads). So this comment is essentially for the PMD writers: use the fencing = where appropriate, even if the lock prefix is there. As I will be the one forgetting the rule, I like to have that in the documentation/comments = as reminders to keep things neat. Fran=E7ois-Fr=E9d=E9ric > But you're right about the bug in spinlocks. > I am going to send a patch for this. >=20 > -- > Thomas