mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: Valdis.Kletnieks@vt.edu
Cc: Steven Rostedt <rostedt@goodmis.org>,
	Christoph Lameter <cl@gentwo.org>,
	linux-kernel@vger.kernel.org, Ingo Molnar <mingo@redhat.com>,
	Andrew Morton <akpm@linux-foundation.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Subject: Re: [RFC][PATCH 0/5] Introduce checks for preemptable code for this_cpu_read/write()
Date: Tue, 20 Sep 2011 15:51:32 +0200 (CEST)	[thread overview]
Message-ID: <alpine.LFD.2.02.1109201549520.2723@ionos> (raw)
In-Reply-To: <27409.1316522696@turing-police.cc.vt.edu>

On Tue, 20 Sep 2011, Valdis.Kletnieks@vt.edu wrote:

> On Mon, 19 Sep 2011 23:06:17 EDT, Steven Rostedt said:
> 
> > It is really confusing to know which version to use. I'm confused by the
> > this_cpu_*() compared with __this_cpu_*(). I'm guessing that most places
> > should use __this_cpu*(). But really this_cpu() should be the default,
> > and the places that can have it outside of preemption should have
> > another name. Maybe use the raw_this_cpu() or safe_this_cpu(), as there
> > is an irqsafe_this_cpu(). Maybe make a preemptsafe_cpu_*(). There should
> > only be a very few locations that are OK to have preemption enabled when
> > calling the this_cpu() code. Lets have those have the funny names and
> > not be the default "this_cpu_*()".
> 
> What's the latency hit on those very few locations if we simply put our
> collective foot down and not support a preemptable version of this_cpu_*()?
> "Yes, you *could* preempt here, but for our collective sanity that's not
> supported"...

Full ack.
 
> > All this_cpu*() code, except the funny named ones, should make sure
> > preemption is disabled, otherwise give a nasty warning. As that is
> > usually a bug if you are using a per cpu variable and can migrate away.
> > The next reference to that value may be incorrect.
> 
> You get a much prettier diffstat if you just nuke the funny named ones. ;)

Along with the maze of completely unused incarnations.
 
> But of course it's early morning and I'm still caffeine-deficient and probably
> overlooking some crucial use case. ;)

I doubt that.

Thanks,

	tglx

  reply	other threads:[~2011-09-20 13:51 UTC|newest]

Thread overview: 73+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-09-19 21:20 Steven Rostedt
2011-09-19 21:20 ` [RFC][PATCH 1/5] x86: Remove const_udelay() caring about which cpu var it uses Steven Rostedt
2011-09-19 21:51   ` Christoph Lameter
2011-09-19 23:31     ` Steven Rostedt
2011-09-19 21:20 ` [RFC][PATCH 2/5] mm: Switch mod_state() to __this_cpu_read() Steven Rostedt
2011-09-19 22:02   ` Christoph Lameter
2011-09-19 23:48     ` Steven Rostedt
2011-09-20 14:46       ` Christoph Lameter
2011-09-20 15:16         ` Steven Rostedt
2011-09-20 15:54           ` Christoph Lameter
2011-09-20 16:07             ` Steven Rostedt
2011-09-20 22:19             ` Valdis.Kletnieks
2011-09-20 13:49     ` Thomas Gleixner
2011-09-20 14:01       ` Steven Rostedt
2011-09-20 14:51       ` Christoph Lameter
2011-09-20 15:11         ` Steven Rostedt
2011-09-20 15:59           ` Christoph Lameter
2011-09-20 16:03             ` Steven Rostedt
2011-09-20 16:07               ` Christoph Lameter
2011-09-20 15:27         ` Thomas Gleixner
2011-09-20 16:02           ` Christoph Lameter
2011-09-20 16:51             ` Thomas Gleixner
2011-09-20 17:08               ` Steven Rostedt
2011-09-19 21:20 ` [RFC][PATCH 3/5] memcg: Disable preemption in memcg_check_events() Steven Rostedt
2011-09-20 14:20   ` Johannes Weiner
2011-09-20 14:24     ` Johannes Weiner
2011-09-20 14:33       ` Steven Rostedt
2011-09-24  0:46   ` Steven Rostedt
2011-09-19 21:20 ` [RFC][PATCH 4/5] printk: Have wake_up_klogd() use __this_cpu_write() Steven Rostedt
2011-09-19 21:54   ` Christoph Lameter
2011-09-19 23:33     ` Steven Rostedt
2011-09-20 14:54       ` Christoph Lameter
2011-09-20 14:55         ` Peter Zijlstra
2011-09-19 21:20 ` [RFC][PATCH 5/5] percpu: Add preempt checks back into this_cpu_read/write() Steven Rostedt
2011-09-19 21:49 ` [RFC][PATCH 0/5] Introduce checks for preemptable code for this_cpu_read/write() Christoph Lameter
2011-09-20  3:06   ` Steven Rostedt
2011-09-20 12:44     ` Valdis.Kletnieks
2011-09-20 13:51       ` Thomas Gleixner [this message]
2011-09-20 14:58         ` Christoph Lameter
2011-09-20 15:17           ` Steven Rostedt
2011-09-20 14:57       ` Christoph Lameter
2011-09-20 15:19         ` Steven Rostedt
2011-09-20 16:08           ` Christoph Lameter
2011-09-20 16:31             ` Steven Rostedt
2011-09-20 16:56               ` Steven Rostedt
2011-09-20 17:09                 ` Peter Zijlstra
2011-09-20 17:15                   ` Steven Rostedt
2011-09-20 17:25                     ` Mathieu Desnoyers
2011-09-20 18:03                       ` Steven Rostedt
2011-09-20 18:12                         ` Mathieu Desnoyers
2011-09-20 18:27                           ` Steven Rostedt
2011-09-20 18:34                             ` Mathieu Desnoyers
2011-09-20 22:32             ` Valdis.Kletnieks
2011-09-20 22:17           ` Valdis.Kletnieks
2011-09-21  1:33             ` Steven Rostedt
2011-09-20 15:46     ` Mathieu Desnoyers
2011-09-20 16:00       ` Steven Rostedt
2011-09-20 16:10         ` Christoph Lameter
2011-09-20 16:50           ` Peter Zijlstra
2011-09-20 18:54           ` Steven Rostedt
2011-09-21 15:16             ` Christoph Lameter
2011-09-21 15:31               ` Steven Rostedt
2011-09-21 15:59                 ` Christoph Lameter
2011-09-21 16:12                   ` Steven Rostedt
2011-09-21 16:32               ` Thomas Gleixner
2011-09-20  2:20 ` Andi Kleen
2011-09-20  3:12   ` Steven Rostedt
2011-09-20  3:17     ` Steven Rostedt
2011-09-20  8:32     ` Thomas Gleixner
2011-09-20 12:10       ` Steven Rostedt
2011-09-20 15:03       ` Christoph Lameter
2011-09-20 15:07         ` Peter Zijlstra
2011-09-20 16:05           ` Christoph Lameter

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=alpine.LFD.2.02.1109201549520.2723@ionos \
    --to=tglx@linutronix.de \
    --cc=Valdis.Kletnieks@vt.edu \
    --cc=akpm@linux-foundation.org \
    --cc=cl@gentwo.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=mingo@redhat.com \
    --cc=peterz@infradead.org \
    --cc=rostedt@goodmis.org \
    /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
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome