mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Alan Cox <alan@lxorguk.ukuu.org.uk>
To: Kay Sievers <kay.sievers@vrfy.org>
Cc: "Eric W. Biederman" <ebiederm@xmission.com>,
	Arjan van de Ven <arjan@infradead.org>,
	Peter Zijlstra <peterz@infradead.org>, Greg KH <gregkh@suse.de>,
	Andrew Morton <akpm@linux-foundation.org>,
	Fabio Comolli <fabio.comolli@gmail.com>, Greg KH <greg@kroah.com>,
	linux-kernel@vger.kernel.org
Subject: Re: [patch 00/13] devtmpfs patches
Date: Mon, 11 May 2009 14:10:57 +0100	[thread overview]
Message-ID: <20090511141057.24a5f8ed@lxorguk.ukuu.org.uk> (raw)
In-Reply-To: <ac3eb2510905110434nb123423m134e848c17c3c736@mail.gmail.com>

> And I did, and it's obvious that creating a single file along with the
> ~10 we already create in /sys, instead of running through /sys or
> /proc later and reconstruct what we missed, is always, and in every
> case faster and simpler. It gets rid of a bunch of things we need to

Arjan's numbers for sysfs are 0.06 seconds if I remember the mail
correctly. That doesn't account for any meaningful speedup.

> > - That dumping more crap in kernel without policy and management is a
> >  good idea
> 
> What exactly do you find is "crap"? What do you mean with "management"?

Device spaces have user controlled naming rules, user controlled
permissions, user controlled labelling and the like. That is policy, and
the administering of that is management.

That was one of the things that killed devfs eventually, and it's not a
problem your proposal or devfs solved.

> 
> > - That the other proposals are worse than yours.
> 
> I also did, in exactly this thread.

Sorry, but I can't find any convincing evidence in the mails. I see
unhelpful things like you calling Arjan's numbers "made up", despite the
fact he has spent months working on speeding up boot and has probably the
most detailed analysis and data sets anyone does.

In fact I see a litany of comments in response to technical queries and
asking you to justify the code (as opposed to you trying to ask anyone to
justify why it shouldn't get dumped in tree)

Most of those comments are not helpful

> > Arjan produced detailed boot graphs and boots a box very fast indeed. I
> > see no boot graphs, no detailed analysis here.
> 
> I did in a reply to your earlier mail. And also mentioned the
> init=/bin/sh case, which I think, would be reason enough alone,
> without all the other gains, to do that.

I think you actually hit the nail on the head a lot earlier when you
asked Arjan  "How will you solve the dynamic device numbers". Which is
that we could fix them ....

What is the real underlying concern and what problem should we fix.

We have at least six takes on this

1.	What problem ?
2.	The udev userspace setup you are using sucks so fix it
3.	Udev sucks because it can't get the early boot events so has to
do it the hard way - so queue them
4.	Udev sucks because it can't get the early boot events so create
a single table for it to read
5.	Make the new big block numbers stable
6.	Recreate devfs in a new but more clean form

I don't buy take #1 so I would like to understand

- Why Arjan's systems boot very fast and yours don't

- Why you think sysfs changes will help when the stats say its 0.06 of a
  second and udev is not it appears taking much time anyway

- How you think you've solved the devfs problems about persistency and
  the like and what performance cost that has. That killed devfs in the
  end.

But even more I'd like to know wtf we don't just fix the stability of the
big block device numbering by attaching them to the existing block device
nodes using minors > 256 which are currently unused in this space.
Historically it didn't happen because Al Viro jumped up and down about
the initial proposal but its cleaner than a new devfs and it solves the
problem for everyone.

Given most distros use udev sorting udev out is obviously useful but its
not the 0.06 seconds - its the rest of the time that matters. If a sysfs
table of devices or buffering the events until it starts sorts that then
we have a far better model.

Udev tackles all the hard stuff - policy in user space, event triggering
for management etc, that your new devfs simply doesn't address. Fixing
udev isn't so much fun but it does actually get us something featureful
and useful that does what people want.



  parent reply	other threads:[~2009-05-11 13:11 UTC|newest]

Thread overview: 95+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <20090509142601.874865281@blue.kroah.org>
2009-05-09 14:37 ` Greg KH
2009-05-09 14:26   ` [patch 01/13] Driver Core: add nodename callbacks Greg KH
2009-05-10 12:52     ` Stephen Rothwell
2009-05-10 13:19       ` Kay Sievers
2009-05-11 20:51         ` Greg KH
2009-05-09 14:26   ` [patch 02/13] Driver Core: misc: add nodename support for misc devices Greg KH
2009-05-15 19:58     ` Pavel Machek
2009-05-18 14:34       ` Greg KH
2009-05-18 19:59         ` Pavel Machek
2009-05-18 20:28       ` Alan Cox
2009-05-09 14:26   ` [patch 03/13] Driver Core: usb: add nodename support for usb drivers Greg KH
2009-05-09 14:26   ` [patch 04/13] Driver Core: block: add nodename support for block drivers Greg KH
2009-05-09 14:26   ` [patch 05/13] Driver Core: x86: add nodename for cpuid and msr drivers Greg KH
2009-05-09 14:26   ` [patch 06/13] Driver Core: dvb: add nodename for dvb drivers Greg KH
2009-05-09 14:26   ` [patch 07/13] Driver Core: input: add nodename for input drivers Greg KH
2009-05-09 14:26   ` [patch 08/13] Driver Core: sound: add nodename for sound drivers Greg KH
2009-05-09 14:26   ` [patch 09/13] Driver Core: raw: add nodename for raw devices Greg KH
2009-05-09 14:26   ` [patch 10/13] Driver Core: drm: add nodename for drm devices Greg KH
2009-05-09 14:26   ` [patch 11/13] Driver Core: aoe: add nodename for aoe devices Greg KH
2009-05-09 14:26   ` [patch 12/13] Driver Core: bsg: add nodename for bsg driver Greg KH
2009-05-09 14:26   ` [patch 13/13] Driver Core: devtmpfs - driver core maintained /dev tmpfs Greg KH
2009-05-09 15:10   ` [patch 00/13] devtmpfs patches Fabio Comolli
2009-05-09 15:08     ` Greg KH
2009-05-09 15:22       ` Arjan van de Ven
2009-05-09 16:19         ` Greg KH
2009-05-09 19:09           ` Arjan van de Ven
2009-05-10  4:34           ` Arjan van de Ven
2009-05-10  7:48             ` Eric W. Biederman
2009-05-10 14:56               ` Eric W. Biederman
2009-05-10  5:34           ` Andrew Morton
2009-05-10 15:20             ` Greg KH
2009-05-10 15:59               ` Arjan van de Ven
2009-05-10 18:31               ` Peter Zijlstra
2009-05-10 21:19                 ` Alan Cox
2009-05-10 23:47                   ` Kay Sievers
2009-05-11  0:00                     ` Arjan van de Ven
     [not found]                       ` <ac3eb2510905101822t7fde14b3nf2c689621f69c925@mail.gmail.com>
2009-05-11  2:36                         ` Eric W. Biederman
2009-05-11 10:46                           ` Kay Sievers
2009-05-11 10:55                             ` Alan Cox
2009-05-11 11:34                               ` Kay Sievers
2009-05-11 13:05                                 ` [patch 00/13] devtmpfs Arjan van de Ven
2009-05-11 13:28                                   ` Kay Sievers
2009-05-11 13:49                                     ` Arjan van de Ven
2009-05-11 14:59                                       ` Kay Sievers
2009-05-11 13:10                                 ` Alan Cox [this message]
2009-05-11 14:14                                   ` [patch 00/13] devtmpfs patches Kay Sievers
2009-05-11 14:30                                     ` Arjan van de Ven
2009-05-11 14:42                                       ` Kay Sievers
2009-05-11 15:53                                     ` Alan Cox
2009-05-11 16:28                                       ` Kay Sievers
2009-05-11 16:41                                         ` Arjan van de Ven
2009-05-11 17:32                                           ` Kay Sievers
2009-05-11 17:55                                             ` Alan Cox
2009-05-11 18:04                                               ` Kay Sievers
2009-05-11 18:40                                                 ` Alan Cox
2009-05-11 16:56                                         ` Alan Cox
2009-05-11 18:13                             ` Eric W. Biederman
2009-05-11  3:55                         ` Arjan van de Ven
2009-05-11 11:49                         ` Fabio Comolli
2009-05-11 17:47                           ` Greg KH
2009-05-11 16:40                         ` Eric W. Biederman
2009-05-11 17:16                           ` Kay Sievers
2009-05-11 21:13                             ` Eric W. Biederman
2009-05-11  1:00                     ` Andrew Morton
2009-05-11  3:58                       ` Arjan van de Ven
2009-05-11 17:45                       ` Greg KH
2009-05-09 16:46         ` Kay Sievers
2009-05-09 17:11           ` Alan Cox
2009-05-09 18:09             ` Kay Sievers
2009-05-11 17:40   ` David P. Quigley
2009-05-11 17:56     ` Greg KH
2009-05-11 20:41       ` David P. Quigley
2009-05-11 21:05         ` Kay Sievers
2009-05-11 21:19           ` Alan Cox
2009-05-11 21:27             ` Kay Sievers
2009-05-12 12:45           ` Stephen Smalley
2009-05-12 15:10             ` Kay Sievers
2009-05-12 15:35               ` Stephen Smalley
2009-05-12 15:54                 ` Kay Sievers
2009-05-12 22:55                   ` Kay Sievers
2009-05-12 23:22                     ` David P. Quigley
2009-05-12 23:34                       ` Kay Sievers
2009-05-12 23:50                         ` Greg KH
2009-05-13 12:22                     ` Stephen Smalley
2009-05-13 12:58                       ` Kay Sievers
2009-05-13 12:57                         ` Stephen Smalley
2009-05-13 13:09                           ` Kay Sievers
2009-05-13 12:59                       ` Alan Cox
2009-05-13 13:20                   ` David Howells
2009-05-13 13:34                     ` Kay Sievers
2009-05-13 14:20                       ` Kay Sievers
2009-05-13 14:35                         ` Stephen Smalley
2009-05-13 16:45                           ` Kay Sievers
2009-05-13 22:43                             ` Eric W. Biederman
2009-05-13 23:10                               ` Greg KH

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=20090511141057.24a5f8ed@lxorguk.ukuu.org.uk \
    --to=alan@lxorguk.ukuu.org.uk \
    --cc=akpm@linux-foundation.org \
    --cc=arjan@infradead.org \
    --cc=ebiederm@xmission.com \
    --cc=fabio.comolli@gmail.com \
    --cc=greg@kroah.com \
    --cc=gregkh@suse.de \
    --cc=kay.sievers@vrfy.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=peterz@infradead.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

all inboxes | Powered by JetHome®