Plans for hal 0.5.x
kay.sievers at vrfy.org
Mon Dec 13 09:03:17 PST 2004
On Mon, 2004-12-13 at 11:34 -0500, David Zeuthen wrote:
> On Mon, 2004-12-13 at 17:11 +0100, Kay Sievers wrote:
> > HAL will only need to use _one_ notification program in /etc/hotplug.d/.
> > Multiple instances are run at the same time, yes. This is intentional,
> > for performance reasons (think of SMP boxes with a lot of disks).
> > But hal doen't need to care about that I think, cause every device chain
> > sequence is still serialized. Only events for independent devices (every
> > single one is still serialized) may be mixed up:
> > /block/sda/sda1 will wait for /block/sda but not for /block/sdb
> Ah, so, in other words, you guarantee that all hotplug.d notifiers
> for /block/sda are done before you invoke the hotplug.d notifiers
> for /block/sda/sda1?
Yes, exactly. All childs will wait for the parent udev process to exit
(exit == handled all notifiers). The whole event sequence is serialized
by the DEVPATH (including the PHYSDEVPATH device).
> > So you may get /block/sdb before you get /block/sda Do you think that's
> > a problem and we need to reorder that?
> No, that would be just fine. In fact it would be perfect - what I want
> to do is actually handle sdb* in parallel to sda* in hald (which is
> possible as the major work will be invoking volume_id processes). How
> does that sound?
That's what I wanted to hear :).
hal mailing list
hal at lists.freedesktop.org
More information about the Hal