From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Paul E. McKenney" Subject: Re: [PATCH 1/2] remove rcu_assign_pointer(NULL) penalty with type/macro safety Date: Wed, 13 Feb 2008 15:37:44 -0800 Message-ID: <20080213233744.GO12393@linux.vnet.ibm.com> References: <20080213220024.GA30729@linux.vnet.ibm.com> <20080213143537.1b806790@extreme> <20080213224134.GK12393@linux.vnet.ibm.com> <20080213144233.05e860cb@extreme> Reply-To: paulmck@linux.vnet.ibm.com Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: linux-kernel@vger.kernel.org, shemminger@vyatta.com, davem@davemloft.net, netdev@vger.kernel.org, dipankar@in.ibm.com, ego@in.ibm.com, herbert@gondor.apana.org.au, akpm@linux-foundation.org To: Stephen Hemminger <"stephen.hemminger@vyatta.com"@mail.vyatta.com> Return-path: Received: from e35.co.us.ibm.com ([32.97.110.153]:56795 "EHLO e35.co.us.ibm.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756175AbYBMXh4 (ORCPT ); Wed, 13 Feb 2008 18:37:56 -0500 Content-Disposition: inline In-Reply-To: <20080213144233.05e860cb@extreme> Sender: netdev-owner@vger.kernel.org List-ID: On Wed, Feb 13, 2008 at 02:42:33PM -0800, Stephen Hemminger wrote: > On Wed, 13 Feb 2008 14:41:34 -0800 > "Paul E. McKenney" wrote: >=20 > > On Wed, Feb 13, 2008 at 02:35:37PM -0800, Stephen Hemminger wrote: > > > On Wed, 13 Feb 2008 14:00:24 -0800 > > > "Paul E. McKenney" wrote: > > >=20 > > > > Hello! > > > >=20 > > > > This is an updated version of the patch posted last November: > > > >=20 > > > > http://archives.free.net.ph/message/20071201.003721.cd6ff17c.e= n.html > > > >=20 > > > > This new version permits arguments with side effects, for examp= le: > > > >=20 > > > > rcu_assign_pointer(global_p, p++); > > > >=20 > > > > and also verifies that the arguments are pointers, while still = avoiding > > > > the unnecessary memory barrier when assigning NULL to a pointer= =2E > > > > This memory-barrier avoidance means that rcu_assign_pointer() i= s now only > > > > permitted for pointers (not array indexes), and so this version= emits a > > > > compiler warning if the first argument is not a pointer. I bui= lt a "make > > > > allyesconfig" version on an x86 system, and received no such wa= rnings. > > > > If RCU is ever applied to array indexes, then the second patch = in this > > > > series should be applied, and the resulting rcu_assign_index() = be used. > > > >=20 > > > > Given the rather surprising history of subtlely broken implemen= tations of > > > > rcu_assign_pointer(), I took the precaution of generating a ful= l set of > > > > test cases and verified that memory barriers and compiler warni= ngs were > > > > emitted when required. I guess it is the simple things that ge= t you... > > > >=20 > > > > Signed-off-by: Paul E. McKenney > > > > --- > > > >=20 > > > > rcupdate.h | 16 ++++++++++++---- > > > > 1 file changed, 12 insertions(+), 4 deletions(-) > > > >=20 > > > > diff -urpNa -X dontdiff linux-2.6.24/include/linux/rcupdate.h l= inux-2.6.24-rap/include/linux/rcupdate.h > > > > --- linux-2.6.24/include/linux/rcupdate.h 2008-01-24 14:58:37.0= 00000000 -0800 > > > > +++ linux-2.6.24-rap/include/linux/rcupdate.h 2008-02-13 13:36:= 47.000000000 -0800 > > > > @@ -270,12 +270,20 @@ extern struct lockdep_map rcu_lock_map; > > > > * structure after the pointer assignment. More importantly, = this > > > > * call documents which pointers will be dereferenced by RCU r= ead-side > > > > * code. > > > > + * > > > > + * Throws a compiler warning for non-pointer arguments. > > > > + * > > > > + * Does not insert a memory barrier for a NULL pointer. > > > > */ > > > > =20 > > > > -#define rcu_assign_pointer(p, v) ({ \ > > > > - smp_wmb(); \ > > > > - (p) =3D (v); \ > > > > - }) > > > > +#define rcu_assign_pointer(p, v) \ > > > > + ({ \ > > > > + typeof(*p) *_________p1 =3D (v); \ > > > > + \ > > > > + if (!__builtin_constant_p(v) || (_________p1 !=3D NULL)) \ > > > > + smp_wmb(); \ > > > > + (p) =3D _________p1; \ > > > > + }) > > > > =20 > > > > /** > > > > * synchronize_sched - block until all CPUs have exited any no= n-preemptive > > >=20 > > > Will this still work if p is unsigned long? > >=20 > > Hello, Steve, > >=20 > > If p is unsigned long, then use rcu_assign_index() from the next pa= tch in > > the set. Looks like Andrew has applied it to -mm -- so please make= sure > > that he is aware if you do use it. >=20 > Make sure fib_trie still works and doesn't get warnings. Ah. It does take a bit to get fib_trie into one's build -- allyesconfi= g doesn't cut it. Please accept my apologies for my confusion!!! Once fib_trie is configured, I do indeed get: net/ipv4/fib_trie.c: In function =E2=80=98node_set_parent=E2=80=99: net/ipv4/fib_trie.c:182: warning: comparison between pointer and in= teger So, given that node->parent is an unsigned long, I changed node_set_par= ent() to the following: static inline void node_set_parent(struct node *node, struct tnode *ptr= ) { rcu_assign_index(node->parent, (unsigned long)ptr | NODE_TYPE(node)); } This removes the warnings. I am a little ambivalent about this, as this is really a pointer in disguise rather than an array index, but patch below. I suppose that another option would be to make node->pare= nt be a void* and provide appropriate accessor functions/macros. Thoughts? Signed-off-by: Paul E. McKenney --- fib_trie.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff -urpNa -X dontdiff linux-2.6.25-rc1/net/ipv4/fib_trie.c linux-2.6.= 25-rc1-fib_trie-warn.compile/net/ipv4/fib_trie.c --- linux-2.6.25-rc1/net/ipv4/fib_trie.c 2008-02-13 14:38:12.000000000 = -0800 +++ linux-2.6.25-rc1-fib_trie-warn.compile/net/ipv4/fib_trie.c 2008-02-= 13 15:22:02.000000000 -0800 @@ -179,8 +179,7 @@ static inline struct tnode *node_parent_ =20 static inline void node_set_parent(struct node *node, struct tnode *pt= r) { - rcu_assign_pointer(node->parent, - (unsigned long)ptr | NODE_TYPE(node)); + rcu_assign_index(node->parent, (unsigned long)ptr | NODE_TYPE(node)); } =20 static inline struct node *tnode_get_child(struct tnode *tn, unsigned = int i)