LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Andrew Morton <akpm@osdl.org>
To: dipankar@in.ibm.com
Cc: Ingo Molnar <mingo@elte.hu>,
	Paul E McKenney <paulmck@linux.vnet.ibm.com>,
	linux-kernel@vger.kernel.org
Subject: Re: [mm PATCH 4/6] RCU: preemptible RCU
Date: Fri, 26 Jan 2007 13:25:45 -0800	[thread overview]
Message-ID: <20070126132545.722c787c.akpm@osdl.org> (raw)
In-Reply-To: <20070126210017.GC17134@in.ibm.com>

On Sat, 27 Jan 2007 02:30:17 +0530
Dipankar Sarma <dipankar@in.ibm.com> wrote:

> > > As a consequence of keeping track of RCU readers, the readers
> > > have a slight overhead (optimizations in the paper).
> > > This implementation co-exists with the "classic" RCU
> > > implementations and can be switched to at compiler.
> > 
> > That's yet another question we need to ask people when their kernel dies,
> > and yet another deviation between the kernels which we all test, causing
> > more dilution of testing efforts.  It would be much better if we could
> > remove classic RCU.  You say this would incur extra cost, but the magnitude
> > of that cost is not clear.  Please help us make that decision.
> 
> See the Table 2, page 10 of the paper mentioned above.

argh.

Seems I have to wade through half the paper to understand Table 2.

> There is a
> ~100ns cost per read-side critical section involved in the preemptible
> version of RCU at the moment. Until, we are sure that we don't have
> an impact on common workloads, we need to keep the "classic" 
> implementation around.

Ratios, please..  that 100ns appears to be a 100% increase.  ie 100ns -> 200ns.

There are a couple of ways of working out how much that really matters: a)
run a workload or b) instrument a kernel, work out how many times/sec the
kernel runs rcu_read_lock().  I suspect b) would be more useful and
informative.

Either way, please always prepare such info up-front and summarise in the
changelog?  It's kinda important...

  reply	other threads:[~2007-01-26 21:44 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-01-15 19:19 [mm PATCH] RCU: various patches Dipankar Sarma
2007-01-15 19:21 ` [mm PATCH 1/6] RCU: split classic rcu Dipankar Sarma
2007-01-16 17:49   ` Paul E. McKenney
2007-01-15 19:22 ` [mm PATCH 2/6] RCU: softirq for RCU Dipankar Sarma
2007-01-16 17:49   ` Paul E. McKenney
2007-01-15 19:24 ` [mm PATCH 3/6] RCU: Fix barriers Dipankar Sarma
2007-01-16 17:53   ` Paul E. McKenney
2007-01-15 19:28 ` [mm PATCH 4/6] RCU: preemptible RCU Dipankar Sarma
2007-01-24  0:32   ` Andrew Morton
2007-01-26 21:00     ` Dipankar Sarma
2007-01-26 21:25       ` Andrew Morton [this message]
2007-01-24  0:38   ` Andrew Morton
2007-01-15 19:30 ` [mm PATCH 5/6] RCU: debug trace for RCU Dipankar Sarma
2007-01-15 19:31 ` [mm PATCH 6/6] RCU: trivial fixes Dipankar Sarma
2007-01-16 17:56   ` Paul E. McKenney

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=20070126132545.722c787c.akpm@osdl.org \
    --to=akpm@osdl.org \
    --cc=dipankar@in.ibm.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=paulmck@linux.vnet.ibm.com \
    --subject='Re: [mm PATCH 4/6] RCU: preemptible RCU' \
    /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

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).