From: David Zeuthen <david@fubar.dk>
To: linux-hotplug@vger.kernel.org
Subject: Re: about split the udev
Date: Mon, 12 Jan 2004 17:51:09 +0000 [thread overview]
Message-ID: <1073929869.6637.9.camel@tux.fubar.dk> (raw)
In-Reply-To: <3ACA40606221794F80A5670F0AF15F8402D4EE96@PDSMSX403.ccr.corp.intel.com>
On Mon, 2004-01-12 at 17:28, Robert Love wrote:
> On Mon, 2004-01-12 at 08:54, Kay Sievers wrote:
>
> > Huhh, what a magic :)
> >
> > udev is a part of hotplug.
> > udev _sends_ DBUS-messages.
> > HAL receives these messages with the node name _from_ udev.
>
> HAL _also_ installs a hotplug script to listen for events and emit a
> D-BUS signal, because HAL wants to know about all hotplug events, not
> just ones that udev handles. So I think Martin's idea is actually
> pretty good.
>
> It would mean we end up with a really clean progression for hardware
> integration:
>
> kernel -> hotplug (via /sbin/hotplug) -> hal.hotplug ->
> udev daemon (via D-BUS) -> HAL daemon (via D-BUS) ->
> upper levels of the desktop (via D-BUS)
>
> Which I like.
Yeah, but udev and hal.hotplug works in different ways with respect to
d-bus integration.
hal.hotplug explicitly sends the hotplug events as d-bus messages
addressed to the org.freedesktop.Hal service - no one can piggy back on
these.
In contrast udev, when invoked, claims the org.kernel.udev service and
emits a signal on the org.kernel.udev.NodeMonitor interface - any
application can pick up this signal if they subscribe to it on the
system messagebus.
Due to the nice security infrastructure of D-BUS one can actually trust
that the event originated from udev.
(BUT... note that we do in udev today is fundamentally racey as multiple
copies of udev run at the same time. So in reality a copy of udev
doesn't always claim the org.kernel.udev service; it becomes a
"secondary owner" of the service instead.
This means that signals stemming from such a copy doesn't seem to
originate from the org.kernel.udev service. This is an issue I think
need to be solved in d-bus; I've already raised it on message-bus-list
but I haven't got a reply yet.)
What we want to do with hotplug is probably similar to what we do with
udev rather than how hal.hotplug works. If people are interested, I'll
be happy to submit code for this. Of course we'll run into the same
issues as we have with udev d-bus integration today.
Cheers,
David
-------------------------------------------------------
This SF.net email is sponsored by: Perforce Software.
Perforce is the Fast Software Configuration Management System offering
advanced branching capabilities and atomic changes on 50+ platforms.
Free Eval! http://www.perforce.com/perforce/loadprog.html
_______________________________________________
Linux-hotplug-devel mailing list http://linux-hotplug.sourceforge.net
Linux-hotplug-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/linux-hotplug-devel
next prev parent reply other threads:[~2004-01-12 17:51 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2004-01-12 7:53 about split the udev Ling, Xiaofeng
2004-01-12 9:36 ` Martin Waitz
2004-01-12 13:54 ` Kay Sievers
2004-01-12 16:23 ` Martin Waitz
2004-01-12 16:28 ` Robert Love
2004-01-12 17:51 ` David Zeuthen [this message]
2004-01-13 1:12 ` Greg KH
2004-01-13 1:29 ` Greg KH
2004-01-13 19:05 ` Tristan Wibberley
2004-01-13 20:36 ` Greg KH
2004-01-15 14:36 ` Ling, Xiaofeng
2004-01-15 21:45 ` Greg KH
2004-01-21 13:38 ` Kay Sievers
2004-01-22 0:27 ` Kay Sievers
2004-01-22 1:08 ` Kay Sievers
2004-01-22 23:01 ` Kay Sievers
2004-01-23 0:34 ` Greg KH
2004-01-23 1:00 ` Kay Sievers
2004-01-23 3:37 ` Kay Sievers
2004-01-23 3:46 ` Greg KH
2004-01-23 3:58 ` Kay Sievers
2004-01-23 4:04 ` Greg KH
2004-01-24 16:06 ` Ling, Xiaofeng
2004-01-24 16:57 ` Kay Sievers
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=1073929869.6637.9.camel@tux.fubar.dk \
--to=david@fubar.dk \
--cc=linux-hotplug@vger.kernel.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;
as well as URLs for NNTP newsgroup(s).