public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: David Miller <davem@davemloft.net>
To: segher@kernel.crashing.org
Cc: benh@kernel.crashing.org, linux-kernel@vger.kernel.org,
	devel@laptop.org, dmk@flex.com, wmb@firmworks.com, jg@laptop.org
Subject: Re: [PATCH] Open Firmware device tree virtual filesystem
Date: Tue, 02 Jan 2007 14:05:04 -0800 (PST)	[thread overview]
Message-ID: <20070102.140504.71092282.davem@davemloft.net> (raw)
In-Reply-To: <bb0d56f649449edb8b5cc0e1c12ede29@kernel.crashing.org>

From: Segher Boessenkool <segher@kernel.crashing.org>
Date: Tue, 2 Jan 2007 22:28:21 +0100

> >> Not single thread -- but a "global OF lock" yes.  Not that
> >> it matters too much, (almost) all property accesses are init
> >> time anyway (which is effectively single threaded).
> >
> > Not that true anymore. A lot of driver probe is being threaded 
> > nowadays,
> > either bcs of the new multithread probing bits, or because they get
> > loaded by userland from some initramfs etc..
> 
> The kernel doesn't care if one CPU is in OF land while the others
> are doing other stuff -- well you have to make sure the OF won't
> try to use a hardware device at the same time as the kernel, true.

True, but at the very least you have to prevent multiple cpus
from enterring OFW.  In fact this is very important.

OFW is not multi-threaded therefore you can't let multiple CPUs call
into OFW at one time.  You must use some kind of locking mechanism,
and that locking mechanism is not simple because it has to not just
stop the other cpus, it has to be able to stop the other cpus yet
still allow them to receive SMP cross-calls from the firmware if the
OFW call is 'stop' or similar.

> I'm a bit concerned about the 100kB or so of data duplication
> (on a *quite big* device tree), and the extra code you need
> (all changes have to be done to both tree copies).  Maybe
> I shouldn't be worried; still, it's obviously not a great
> idea to *require* any arch to get and keep a full copy of
> the tree -- it's wasteful and unnecessary.

The largest amount of memory I've ever seen consumed on sparc64
was 76K and this is 1) 64-bit and 2) an ENORMOUS machine with
lots of cpus and devices.  And I know because sparc64 prints
a kernel message at boot which states how much memory was
consumed by the in-kernel device tree copy.

That ENORMOUS machine also had a lot of ram.

76K, that's just over 8 FREAKIN PAGES.  It's nothing!

I think all this worry is overblown, terribly so in fact.

Please let's get over this memory consumption non-issue and move
on to more productive talk.

  parent reply	other threads:[~2007-01-02 22:05 UTC|newest]

Thread overview: 89+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2006-12-31  1:38 [PATCH] Open Firmware device tree virtual filesystem Mitch Bradley
2006-12-31  5:19 ` David Miller
2006-12-31  9:36   ` Mitch Bradley
2006-12-31  9:52     ` David Miller
2006-12-31 10:11     ` David Kahn
2006-12-31 10:49       ` David Miller
2006-12-31 11:47         ` Rene Rebe
2006-12-31 11:53         ` David Kahn
2007-01-01  3:48           ` Segher Boessenkool
2007-01-02  3:56           ` Benjamin Herrenschmidt
2007-01-02 18:43             ` Richard Smith
2006-12-31 15:41         ` Christoph Hellwig
2006-12-31 20:46           ` David Miller
2007-01-01  3:37             ` David Kahn
2007-01-01  8:54               ` David Miller
2007-01-02  4:02               ` Benjamin Herrenschmidt
2007-01-02 12:28                 ` Segher Boessenkool
2007-01-01  3:33           ` Segher Boessenkool
2007-01-01  8:57             ` David Miller
2007-01-01 17:48               ` Segher Boessenkool
2007-01-01 23:08                 ` David Miller
2007-01-01 23:52                   ` Segher Boessenkool
2007-01-02  3:31                     ` David Miller
2007-01-02 11:26                       ` Segher Boessenkool
2007-01-02  1:40                   ` David Kahn
2007-01-02  3:36                     ` David Miller
2007-01-01 18:10               ` Mitch Bradley
2007-01-01 19:21                 ` Jan Engelhardt
2007-01-02  4:05               ` Benjamin Herrenschmidt
2007-01-02  4:30                 ` David Miller
2007-01-02  4:57                   ` Benjamin Herrenschmidt
2007-01-02  5:01                     ` David Miller
2007-01-02  5:09                       ` Benjamin Herrenschmidt
2007-01-02  5:44                         ` David Miller
2007-01-02 12:36                     ` Segher Boessenkool
2007-01-02 11:03                   ` Segher Boessenkool
2007-01-02  3:53     ` Benjamin Herrenschmidt
2007-01-02 12:22       ` Segher Boessenkool
2007-01-02 20:12         ` Benjamin Herrenschmidt
2007-01-02 21:28           ` Segher Boessenkool
2007-01-02 21:32             ` Benjamin Herrenschmidt
2007-01-02 21:40               ` Segher Boessenkool
2007-01-02 22:10                 ` David Miller
2007-01-02 22:05             ` David Miller [this message]
2007-01-03  0:48               ` Segher Boessenkool
2007-01-03  4:34                 ` David Miller
2007-01-03 15:23                   ` Segher Boessenkool
2007-01-04  2:15                     ` David Miller
2007-01-02  3:45   ` Benjamin Herrenschmidt
2007-01-02  3:49     ` David Miller
2007-01-02 11:45     ` Segher Boessenkool
2007-01-02 20:07       ` Benjamin Herrenschmidt
2006-12-31 13:24 ` Pekka Enberg
2006-12-31 18:55   ` Mitch Bradley
2006-12-31 14:12 ` Jan Engelhardt
2006-12-31 20:45   ` David Miller
2006-12-31 21:30     ` Jan Engelhardt
2007-01-02  3:43     ` Benjamin Herrenschmidt
2007-01-02 11:37       ` Segher Boessenkool
2007-01-02 13:22         ` Stefan Reinauer
2007-01-02 20:08         ` Benjamin Herrenschmidt
2007-01-02 20:11           ` Mitch Bradley
2007-01-02 20:48             ` Benjamin Herrenschmidt
2007-01-02 21:37               ` Segher Boessenkool
2007-01-02 21:47                 ` Benjamin Herrenschmidt
2007-01-03  0:35                   ` Segher Boessenkool
2007-01-03  0:44                     ` Benjamin Herrenschmidt
2007-01-03  1:14                       ` Segher Boessenkool
2007-01-03  4:35                         ` David Miller
2007-01-02 22:07                 ` David Miller
2007-01-03  0:52                   ` Segher Boessenkool
2007-01-03  1:13                     ` Jan Engelhardt
2007-01-03  4:38                       ` David Miller
2007-01-03  5:05                         ` Benjamin Herrenschmidt
2007-01-03 15:59                           ` Segher Boessenkool
2007-01-03 15:31                         ` Segher Boessenkool
2007-01-03  4:34                     ` David Miller
2007-01-02 21:15           ` Segher Boessenkool
2007-01-02 21:59             ` David Miller
2007-01-01  3:40   ` Segher Boessenkool
2007-01-01  4:21 ` Segher Boessenkool
  -- strict thread matches above, loose matches on Subject: below --
2007-01-11 17:39 ron minnich
2007-01-11 17:53 ` Mitch Bradley
2007-01-11 17:55   ` ron minnich
2007-01-11 18:36     ` Segher Boessenkool
2007-01-11 18:20 ` Stefan Reinauer
2007-01-11 18:47   ` Segher Boessenkool
2007-01-11 19:12     ` ron minnich
2007-01-11 19:11   ` ron minnich

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=20070102.140504.71092282.davem@davemloft.net \
    --to=davem@davemloft.net \
    --cc=benh@kernel.crashing.org \
    --cc=devel@laptop.org \
    --cc=dmk@flex.com \
    --cc=jg@laptop.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=segher@kernel.crashing.org \
    --cc=wmb@firmworks.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