LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: Borislav Petkov <bp@alien8.de>
To: "Zheng, Lv" <lv.zheng@intel.com>
Cc: linux-edac <linux-edac@vger.kernel.org>,
	Jiri Kosina <jkosina@suse.cz>, Borislav Petkov <bp@suse.de>,
	"Rafael J. Wysocki" <rjw@rjwysocki.net>,
	Len Brown <lenb@kernel.org>, "Luck, Tony" <tony.luck@intel.com>,
	Tomasz Nowicki <tomasz.nowicki@linaro.org>,
	"Chen, Gong" <gong.chen@linux.intel.com>,
	Wolfram Sang <wsa@the-dreams.de>,
	Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>,
	"linux-acpi@vger.kernel.org" <linux-acpi@vger.kernel.org>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Subject: Re: [RFC PATCH 5/5] GHES: Make NMI handler have a single reader
Date: Wed, 29 Apr 2015 10:13:55 +0200	[thread overview]
Message-ID: <20150429081355.GA5498@pd.tnic> (raw)
In-Reply-To: <1AE640813FDE7649BE1B193DEA596E880270FB3B@SHSMSX101.ccr.corp.intel.com>

On Wed, Apr 29, 2015 at 12:49:59AM +0000, Zheng, Lv wrote:
> > > We absolutely want to use atomic_add_unless() because we get to save us
> > > the expensive
> > >
> > > 	LOCK; CMPXCHG
> > >
> > > if the value was already 1. Which is exactly what this patch is trying
> > > to avoid - a thundering herd of cores CMPXCHGing a global variable.
> > 
> > IMO, on most architectures, the "cmp" part should work just like what you've done with "if".
> > And on some architectures, if the "xchg" doesn't happen, the "cmp" part even won't cause a pipe line hazard.

Even if CMPXCHG is being split into several microops, they all still
need to flow down the pipe and require resources and tracking. And you
only know at retire time what the CMP result is and can "discard" the
XCHG part. Provided the uarch is smart enough to do that.

This is probably why CMPXCHG needs 5,6,7,10,22,... cycles depending on
uarch and vendor, if I can trust Agner Fog's tables. And I bet those
numbers are best-case only and in real-life they probably tend to fall
out even worse.

CMP needs only 1. On almost every uarch and vendor. And even that cycle
probably gets hidden with a good branch predictor.

> If you man the LOCK prefix, I understand now.

And that makes several times worse: 22, 40, 80, ... cycles.

-- 
Regards/Gruss,
    Boris.

ECO tip #101: Trim your mails when you reply.
--

  reply	other threads:[~2015-04-29  8:14 UTC|newest]

Thread overview: 32+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-03-27  9:22 [RFC PATCH 0/5] GHES NMI handler cleanup Borislav Petkov
2015-03-27  9:22 ` [RFC PATCH 1/5] GHES: Carve out error queueing in a separate function Borislav Petkov
2015-03-27  9:22 ` [RFC PATCH 2/5] GHES: Carve out the panic functionality Borislav Petkov
2015-03-27  9:22 ` [RFC PATCH 3/5] GHES: Panic right after detection Borislav Petkov
2015-03-27  9:22 ` [RFC PATCH 4/5] GHES: Elliminate double-loop in the NMI handler Borislav Petkov
2015-03-27  9:22 ` [RFC PATCH 5/5] GHES: Make NMI handler have a single reader Borislav Petkov
2015-04-01  7:45   ` Jiri Kosina
2015-04-01 13:49     ` Borislav Petkov
2015-04-23  8:39       ` Jiri Kosina
2015-04-23  8:59         ` Borislav Petkov
2015-04-23 18:00           ` Luck, Tony
2015-04-27 20:23             ` Borislav Petkov
2015-04-28 14:30     ` Don Zickus
2015-04-28 14:42       ` Don Zickus
2015-04-28 14:55       ` Borislav Petkov
2015-04-28 15:35         ` Don Zickus
2015-04-28 16:22           ` Borislav Petkov
2015-04-28 18:44             ` Don Zickus
2015-05-04 15:40               ` Borislav Petkov
2015-04-27  3:16   ` Zheng, Lv
2015-04-27  8:46     ` Borislav Petkov
2015-04-28  0:44       ` Zheng, Lv
2015-04-28  2:24       ` Zheng, Lv
2015-04-28  7:38         ` Borislav Petkov
2015-04-28 13:38   ` Zheng, Lv
2015-04-28 13:59     ` Borislav Petkov
2015-04-29  0:24       ` Zheng, Lv
2015-04-29  0:49       ` Zheng, Lv
2015-04-29  8:13         ` Borislav Petkov [this message]
2015-04-30  8:05           ` Zheng, Lv
2015-04-30  8:48             ` Borislav Petkov
2015-05-02  0:34               ` Zheng, Lv

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=20150429081355.GA5498@pd.tnic \
    --to=bp@alien8.de \
    --cc=bp@suse.de \
    --cc=gong.chen@linux.intel.com \
    --cc=jkosina@suse.cz \
    --cc=lenb@kernel.org \
    --cc=linux-acpi@vger.kernel.org \
    --cc=linux-edac@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=lv.zheng@intel.com \
    --cc=n-horiguchi@ah.jp.nec.com \
    --cc=rjw@rjwysocki.net \
    --cc=tomasz.nowicki@linaro.org \
    --cc=tony.luck@intel.com \
    --cc=wsa@the-dreams.de \
    --subject='Re: [RFC PATCH 5/5] GHES: Make NMI handler have a single reader' \
    /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).