From: Andries Brouwer <aebr@win.tue.nl>
To: Ingo Molnar <mingo@elte.hu>
Cc: Linus Torvalds <torvalds@transmeta.com>,
William Lee Irwin III <wli@holomorphy.com>,
<linux-kernel@vger.kernel.org>
Subject: Re: [patch] lockless, scalable get_pid(), for_each_process() elimination, 2.5.35-BK
Date: Wed, 18 Sep 2002 14:32:06 +0200 [thread overview]
Message-ID: <20020918123206.GA14595@win.tue.nl> (raw)
In-Reply-To: <Pine.LNX.4.44.0209180024090.30913-100000@localhost.localdomain>
On Wed, Sep 18, 2002 at 01:06:11AM +0200, Ingo Molnar wrote:
> the underlying problem is very hard - since get_pid() not only has to take
> PIDs into account, but TGIDs, session IDs and process groups as well.
>
> Ie. even in the most hopeless situation, if there are 999,999 PIDs
> allocated already, it takes less than 10 usecs to find and allocate the
> remaining one PID. The common fastpath is a couple of instructions only.
> The overhead of skipping over continuous regions of allocated PIDs scales
> gracefully with the number of bits to be skipped, from 0 to 10 usecs.
>
> memory footprint of the new PID allocator scales dynamically with
> /proc/sys/kernel/pid_max: the default 32K PIDs cause a 4K allocation, a
> pid_max of 1 million causes a 128K footprint. The current absolute limit
> for pid_max is 4 million PIDs - this does not cause any allocation in the
> kernel, the bitmaps are demand-allocated runtime. The pidmap table takes
> up 512 bytes.
I still don't understand the current obsession with this stuff.
It is easy to have pid_max 2^30 and a fast algorithm that does not
take any more kernel space.
It seems to me you are first creating an unrealistic and unfavorable
situation (put pid_max at some artificially low value, starting a
lot of tasks and saying: look! the algorithm is quadratic!) and
then solve the problem that you thus invented yourself.
Please leave pid_max large.
Andries
next prev parent reply other threads:[~2002-09-18 12:27 UTC|newest]
Thread overview: 82+ messages / expand[flat|nested] mbox.gz Atom feed top
2002-09-17 23:06 [patch] lockless, scalable get_pid(), for_each_process() elimination, 2.5.35-BK Ingo Molnar
2002-09-18 0:22 ` William Lee Irwin III
2002-09-18 1:27 ` David S. Miller
2002-09-18 2:22 ` William Lee Irwin III
2002-09-18 10:11 ` Ingo Molnar
2002-09-18 14:28 ` William Lee Irwin III
2002-09-18 12:32 ` Andries Brouwer [this message]
2002-09-18 12:56 ` Ingo Molnar
2002-09-18 16:17 ` Linus Torvalds
2002-09-18 16:46 ` Ingo Molnar
2002-09-18 16:57 ` Ingo Molnar
2002-09-18 18:28 ` Andries Brouwer
2002-09-18 18:27 ` William Lee Irwin III
2002-09-18 19:00 ` Ingo Molnar
2002-09-18 20:11 ` Andries Brouwer
2002-09-18 20:29 ` William Lee Irwin III
2002-09-18 21:15 ` Andries Brouwer
2002-09-19 3:05 ` Ingo Molnar
2002-09-19 13:11 ` Andries Brouwer
2002-09-21 12:56 ` quadratic behaviour Andries Brouwer
2002-09-21 17:06 ` Linus Torvalds
2002-09-21 17:35 ` William Lee Irwin III
2002-09-21 17:49 ` Ingo Molnar
2002-09-21 17:46 ` William Lee Irwin III
2002-09-18 14:49 ` [patch] lockless, scalable get_pid(), for_each_process() elimination, 2.5.35-BK William Lee Irwin III
2002-09-18 15:01 ` Cort Dougan
2002-09-18 15:33 ` Ingo Molnar
2002-09-18 15:31 ` yodaiken
2002-09-18 15:35 ` Cort Dougan
2002-09-18 15:37 ` Rik van Riel
2002-09-18 15:50 ` Alan Cox
2002-09-18 18:31 ` Andries Brouwer
2002-09-18 18:32 ` William Lee Irwin III
2002-09-18 16:15 ` Linus Torvalds
2002-09-18 16:32 ` Rik van Riel
2002-09-18 16:48 ` Linus Torvalds
2002-09-18 17:31 ` Ingo Molnar
2002-09-18 21:46 ` Rogier Wolff
2002-09-18 16:41 ` Ingo Molnar
2002-09-18 16:45 ` William Lee Irwin III
2002-09-18 16:59 ` Linus Torvalds
2002-09-18 17:36 ` Ingo Molnar
2002-09-18 17:36 ` William Lee Irwin III
2002-09-18 17:46 ` Linus Torvalds
2002-09-18 17:47 ` William Lee Irwin III
2002-09-18 17:59 ` Ingo Molnar
2002-09-18 17:50 ` William Lee Irwin III
2002-09-18 18:07 ` Ingo Molnar
2002-09-19 19:27 ` Martin Mares
2002-09-17 15:31 ` Pavel Machek
2002-09-19 20:07 ` Richard B. Johnson
2002-09-18 16:46 ` Rik van Riel
2002-09-18 16:53 ` Linus Torvalds
2002-09-18 17:00 ` Rik van Riel
2002-09-18 17:16 ` Linus Torvalds
2002-09-18 17:28 ` Ingo Molnar
2002-09-18 17:38 ` Ingo Molnar
2002-09-18 17:24 ` Ingo Molnar
2002-09-18 17:30 ` Linus Torvalds
2002-09-18 17:35 ` Cort Dougan
2002-09-18 17:43 ` Rik van Riel
2002-09-18 17:48 ` Martin J. Bligh
2002-09-18 17:57 ` Cort Dougan
2002-09-18 18:08 ` Martin J. Bligh
2002-09-18 18:28 ` Ingo Molnar
2002-09-18 18:33 ` Linus Torvalds
2002-09-18 18:51 ` Ingo Molnar
2002-09-18 18:56 ` interruptible_sleep_on_timeout() and signals Imran Badr
2002-09-19 7:16 ` David Woodhouse
2002-09-18 19:53 ` [patch] lockless, scalable get_pid(), for_each_process() elimination, 2.5.35-BK Martin J. Bligh
2002-09-22 0:34 ` Andrea Arcangeli
2002-09-22 10:02 ` Ingo Molnar
2002-09-22 15:27 ` Andrea Arcangeli
2002-09-18 18:02 ` Ingo Molnar
2002-09-18 17:58 ` Cort Dougan
2002-09-18 18:14 ` Ingo Molnar
2002-09-18 17:54 ` Alan Cox
2002-09-18 17:54 ` Ingo Molnar
2002-09-18 18:00 ` yodaiken
2002-09-18 18:21 ` Ingo Molnar
2002-09-18 18:33 ` yodaiken
2002-09-18 16:55 ` Alan Cox
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=20020918123206.GA14595@win.tue.nl \
--to=aebr@win.tue.nl \
--cc=linux-kernel@vger.kernel.org \
--cc=mingo@elte.hu \
--cc=torvalds@transmeta.com \
--cc=wli@holomorphy.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;
as well as URLs for NNTP newsgroup(s).