linux-kernel - Re: [PATCH v9 2/7] rcu/segcblist: Add counters to segcblist datastructure

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20201107001847.GD1397669@google.com>
Date:   Fri, 6 Nov 2020 19:18:47 -0500
From:   Joel Fernandes <joel@...lfernandes.org>
To:     "Paul E. McKenney" <paulmck@...nel.org>
Cc:     linux-kernel@...r.kernel.org,
        Frederic Weisbecker <frederic@...nel.org>,
        Josh Triplett <josh@...htriplett.org>,
        Lai Jiangshan <jiangshanlai@...il.com>,
        Marco Elver <elver@...gle.com>,
        Mathieu Desnoyers <mathieu.desnoyers@...icios.com>,
        rcu@...r.kernel.org, Steven Rostedt <rostedt@...dmis.org>,
        "Uladzislau Rezki (Sony)" <urezki@...il.com>, fweisbec@...il.com,
        neeraj.iitr10@...il.com
Subject: Re: [PATCH v9 2/7] rcu/segcblist: Add counters to segcblist
 datastructure

On Fri, Nov 06, 2020 at 07:01:57PM -0500, Joel Fernandes wrote:
> On Wed, Nov 04, 2020 at 09:01:33AM -0800, Paul E. McKenney wrote:
> 
> > A casual reader might be forgiven for being confused by the combination
> > of "Return" in the above comment and the "void" function type below.
> > So shouldn't this comment be something like "Add the specified number
> > of callbacks to the specified segment..."?
> 
> You are right, sorry and will fix it.
> 
> > > @@ -330,11 +342,16 @@ void rcu_segcblist_extract_pend_cbs(struct rcu_segcblist *rsclp,
> > >  
> > >  	if (!rcu_segcblist_pend_cbs(rsclp))
> > >  		return; /* Nothing to do. */
> > > +	rclp->len = rcu_segcblist_get_seglen(rsclp, RCU_WAIT_TAIL) +
> > > +		    rcu_segcblist_get_seglen(rsclp, RCU_NEXT_READY_TAIL) +
> > > +		    rcu_segcblist_get_seglen(rsclp, RCU_NEXT_TAIL);
> > 
> > This should be a "for" loop.  Yes, the number and names of the segments
> > hasn't changed for a good long time, but nothing like code as above to
> > inspire Murphy to more mischief.  :-/
> > 
> > Actually, why not put the summation in the existing "for" loop below?
> > That would save a line of code in addition to providing less inspiration
> > for Mr. Murphy.
> 
> I can do that. Actually Frederic suggested the same thing but I was reluctant
> as I felt it did not give much LOC benefit. Will revisit it.

It reduces 1 line of code :) I changed it to the below, will update the patch:

---8<-----------------------

diff --git a/kernel/rcu/rcu_segcblist.c b/kernel/rcu/rcu_segcblist.c
index 9b43d686b1f3..bff9b2253e50 100644
--- a/kernel/rcu/rcu_segcblist.c
+++ b/kernel/rcu/rcu_segcblist.c
@@ -101,7 +101,7 @@ static void rcu_segcblist_set_seglen(struct rcu_segcblist *rsclp, int seg, long
 	WRITE_ONCE(rsclp->seglen[seg], v);
 }
 
-/* Return number of callbacks in a segment of the segmented callback list. */
+/* Increase the numeric length of a segment by a specified amount. */
 static void rcu_segcblist_add_seglen(struct rcu_segcblist *rsclp, int seg, long v)
 {
 	WRITE_ONCE(rsclp->seglen[seg], rsclp->seglen[seg] + v);
@@ -406,13 +406,12 @@ void rcu_segcblist_extract_pend_cbs(struct rcu_segcblist *rsclp,
 
 	if (!rcu_segcblist_pend_cbs(rsclp))
 		return; /* Nothing to do. */
-	rclp->len = rcu_segcblist_get_seglen(rsclp, RCU_WAIT_TAIL) +
-		    rcu_segcblist_get_seglen(rsclp, RCU_NEXT_READY_TAIL) +
-		    rcu_segcblist_get_seglen(rsclp, RCU_NEXT_TAIL);
+	rclp->len = 0;
 	*rclp->tail = *rsclp->tails[RCU_DONE_TAIL];
 	rclp->tail = rsclp->tails[RCU_NEXT_TAIL];
 	WRITE_ONCE(*rsclp->tails[RCU_DONE_TAIL], NULL);
 	for (i = RCU_DONE_TAIL + 1; i < RCU_CBLIST_NSEGS; i++) {
+		rclp->len += rcu_segcblist_get_seglen(rsclp, i);
 		WRITE_ONCE(rsclp->tails[i], rsclp->tails[RCU_DONE_TAIL]);
 		rcu_segcblist_set_seglen(rsclp, i, 0);
 	}