From: William Lee Irwin III <wli@holomorphy.com>
To: Andrea Arcangeli <andrea@novell.com>
Cc: Andrew Morton <akpm@osdl.org>, Ray Bryant <raybry@sgi.com>,
Jesse Barnes <jbarnes@engr.sgi.com>,
linux-kernel@vger.kernel.org
Subject: Re: [profile] amortize atomic hit count increments
Date: Tue, 14 Sep 2004 08:51:03 -0700 [thread overview]
Message-ID: <20040914155103.GR9106@holomorphy.com> (raw)
In-Reply-To: <20040914113419.GH4180@dualathlon.random>
On Mon, Sep 13, 2004 at 09:47:48PM -0700, William Lee Irwin III wrote:
>> timer interrupt, usually at boot. The following patch attempts to
>> amortize the atomic operations done on the profile buffer to address
>> this stability concern. This patch has nothing to do with performance;
On Tue, Sep 14, 2004 at 01:34:19PM +0200, Andrea Arcangeli wrote:
> isn't it *much* simpler and much more efficient to just have a per-cpu
> idle function? I seriously doubt you'll get simultaneous collisions on
> anything but the 'halt' instruction in the idle function.
Sampling the profile buffer at regular intervals shows far less than
256 distinct functions hit in 1s intervals even with all cpus busy. As
for whether that would be sufficient, that will have to be answered by
those who reported the bug. I suppose to test whether things besides
idling do cause this problem, one would boot with a restricted
prof_cpu_mask, load all cpus on the machine, set the prof_cpu_mask to
unrestricted, and see if it livelocks before the load terminates.
-- wli
next prev parent reply other threads:[~2004-09-14 15:55 UTC|newest]
Thread overview: 96+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-09-13 8:50 2.6.9-rc1-mm5 Andrew Morton
2004-09-13 9:22 ` 2.6.9-rc1-mm5 Nick Piggin
2004-09-13 17:24 ` 2.6.9-rc1-mm5 Jesse Barnes
2004-09-13 18:06 ` 2.6.9-rc1-mm5 Paul Jackson
2004-09-13 18:10 ` 2.6.9-rc1-mm5 Jesse Barnes
2004-09-13 21:30 ` 2.6.9-rc1-mm5 Jesse Barnes
2004-09-14 2:02 ` 2.6.9-rc1-mm5 Nick Piggin
2004-09-14 2:12 ` 2.6.9-rc1-mm5 Jesse Barnes
2004-09-13 10:20 ` 2.6.9-rc1-mm5 Christoph Hellwig
2004-09-13 10:48 ` 2.6.9-rc1-mm5 Rafael J. Wysocki
2004-09-13 11:13 ` 2.6.9-rc1-mm5 Nikita Danilov
2004-09-13 13:40 ` 2.6.9-rc1-mm5 Christoph Hellwig
2004-09-13 11:16 ` 2.6.9-rc1-mm5 Rafael J. Wysocki
2004-09-13 11:01 ` 2.6.9-rc1-mm5 William Lee Irwin III
2004-09-13 15:09 ` 2.6.9-rc1-mm5 Martin J. Bligh
2004-09-13 15:18 ` 2.6.9-rc1-mm5 Paul Jackson
2004-09-13 16:11 ` 2.6.9-rc1-mm5 Martin J. Bligh
2004-09-13 16:22 ` 2.6.9-rc1-mm5 Paul Jackson
2004-09-13 15:20 ` 2.6.9-rc1-mm5 Kirill Korotaev
2004-09-13 20:01 ` 2.6.9-rc1-mm5 Andrew Morton
2004-09-14 6:39 ` 2.6.9-rc1-mm5 Kirill Korotaev
2004-09-13 20:30 ` 2.6.9-rc1-mm5 Pasi Savolainen
2004-09-13 21:06 ` 2.6.9-rc1-mm5 Rafael J. Wysocki
2004-09-14 9:07 ` 2.6.9-rc1-mm5 Nikita Danilov
2004-09-14 9:12 ` 2.6.9-rc1-mm5 Andrew Morton
2004-09-14 13:21 ` 2.6.9-rc1-mm5 David Howells
2004-09-14 14:24 ` 2.6.9-rc1-mm5 James Morris
2004-09-14 15:36 ` 2.6.9-rc1-mm5 David Howells
2004-09-13 21:47 ` 2.6.9-rc1-mm5 scheduling while atomic Jesse Barnes
2004-09-13 22:56 ` Paul Jackson
2004-09-13 21:56 ` 2.6.9-rc1-mm5 bug in tcp_recvmsg? Jesse Barnes
2004-09-13 22:36 ` David S. Miller
2004-09-13 22:44 ` Jesse Barnes
2004-09-13 22:47 ` David S. Miller
2004-09-13 23:54 ` Jesse Barnes
2004-09-13 23:55 ` David S. Miller
2004-09-14 0:03 ` Jesse Barnes
2004-09-14 0:21 ` David S. Miller
2004-09-14 17:09 ` Jesse Barnes
2004-09-14 0:25 ` 2.6.9-rc1-mm5: TCP oopses James Morris
2004-09-14 2:08 ` David S. Miller
2004-09-14 3:04 ` James Morris
2004-09-14 3:34 ` Herbert Xu
2004-09-14 4:53 ` David S. Miller
2004-09-14 4:55 ` David S. Miller
2004-09-14 5:07 ` James Morris
2004-09-14 2:25 ` [pidhashing] [0/3] pid allocator updates William Lee Irwin III
2004-09-14 2:28 ` [pidhashing] [1/3] retain older vendor copyright William Lee Irwin III
2004-09-14 2:31 ` [pidhashing] [2/3] lower PID_MAX_LIMIT for 32-bit machines William Lee Irwin III
2004-09-14 2:36 ` [pidhashing] [3/3] enforce PID_MAX_LIMIT in sysctls William Lee Irwin III
2004-09-14 2:38 ` [pidhashing] [2/3] lower PID_MAX_LIMIT for 32-bit machines William Lee Irwin III
2004-09-14 10:55 ` Roger Luethi
2004-09-14 11:10 ` Lars Marowsky-Bree
2004-09-14 12:06 ` Lars Marowsky-Bree
2004-09-14 12:08 ` Roger Luethi
2004-09-14 15:41 ` William Lee Irwin III
2004-09-14 15:47 ` Roger Leuthi
2004-09-14 16:41 ` William Lee Irwin III
2004-09-14 17:16 ` Roger Luethi
2004-09-14 2:53 ` [procfs] [1/1] fix task_mmu.c text size reporting William Lee Irwin III
2004-09-14 2:54 ` William Lee Irwin III
2004-09-15 10:51 ` [procfs] [2/1] report per-process pagetable usage William Lee Irwin III
2004-09-14 4:47 ` [profile] amortize atomic hit count increments William Lee Irwin III
2004-09-14 5:05 ` David S. Miller
2004-09-14 5:32 ` William Lee Irwin III
2004-09-14 5:49 ` David S. Miller
2004-09-14 6:10 ` William Lee Irwin III
2004-09-14 6:18 ` William Lee Irwin III
2004-09-14 5:05 ` Andrew Morton
2004-09-14 5:21 ` William Lee Irwin III
2004-09-14 6:43 ` William Lee Irwin III
2004-09-14 6:52 ` Andrew Morton
2004-09-14 7:55 ` William Lee Irwin III
2004-09-14 8:48 ` William Lee Irwin III
2004-09-14 11:34 ` Andrea Arcangeli
2004-09-14 15:51 ` William Lee Irwin III [this message]
2004-09-14 16:05 ` Andrea Arcangeli
2004-09-14 16:16 ` Jesse Barnes
2004-09-14 16:31 ` Andrea Arcangeli
2004-09-14 16:45 ` William Lee Irwin III
2004-09-14 19:00 ` William Lee Irwin III
2004-09-14 19:23 ` William Lee Irwin III
2004-09-14 20:02 ` William Lee Irwin III
2004-09-14 20:04 ` William Lee Irwin III
2004-09-14 21:04 ` William Lee Irwin III
2004-09-14 21:11 ` William Lee Irwin III
2004-09-14 10:00 ` 2.6.9-rc1-mm5 Lorenzo Allegrucci
2004-09-15 11:36 ` 2.6.9-rc1-mm5 William Lee Irwin III
2004-09-15 11:38 ` 2.6.9-rc1-mm5 Jens Axboe
2004-09-15 12:28 ` 2.6.9-rc1-mm5 William Lee Irwin III
2004-09-15 12:41 ` 2.6.9-rc1-mm5 Jens Axboe
2004-09-15 12:50 ` 2.6.9-rc1-mm5 Jens Axboe
2004-09-15 12:53 ` 2.6.9-rc1-mm5 William Lee Irwin III
2004-09-16 0:38 ` 2.6.9-rc1-mm5 William Lee Irwin III
2004-09-16 5:44 ` 2.6.9-rc1-mm5 William Lee Irwin III
2004-09-16 5:45 ` 2.6.9-rc1-mm5 Jens Axboe
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=20040914155103.GR9106@holomorphy.com \
--to=wli@holomorphy.com \
--cc=akpm@osdl.org \
--cc=andrea@novell.com \
--cc=jbarnes@engr.sgi.com \
--cc=linux-kernel@vger.kernel.org \
--cc=raybry@sgi.com \
/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