I would like to know what distributions not relying on systemd are
supposed to get bluetoothd autostarted now that udev rule was dropped:
http://git.kernel.org/?p=bluetooth/bluez.git;a=commit;h=2ea98a6a043710ad4958355b62c682b4767f292e
Looks like we go back from udev rule to manually starting bluetoothd all
the time, even if bluetooth device is not connected? Why are we going
back to old behavior?
Thanks for the info
El vie, 29-06-2012 a las 22:23 +0200, Pacho Ramos escribió:
> I would like to know what distributions not relying on systemd are
> supposed to get bluetoothd autostarted now that udev rule was dropped:
> http://git.kernel.org/?p=bluetooth/bluez.git;a=commit;h=2ea98a6a043710ad4958355b62c682b4767f292e
>
> Looks like we go back from udev rule to manually starting bluetoothd all
> the time, even if bluetooth device is not connected? Why are we going
> back to old behavior?
>
> Thanks for the info
Should we revert that change downstream if we are not going to rely on
systemd and until udev is changed to break that setup (forcing us to go
to old "always running" behavior)?
El mar, 03-07-2012 a las 10:21 -0300, Lucas De Marchi escribió:
> On Sat, Jun 30, 2012 at 3:15 PM, Pacho Ramos <[email protected]> wrote:
> > El vie, 29-06-2012 a las 22:23 +0200, Pacho Ramos escribió:
> >> I would like to know what distributions not relying on systemd are
> >> supposed to get bluetoothd autostarted now that udev rule was dropped:
> >> http://git.kernel.org/?p=bluetooth/bluez.git;a=commit;h=2ea98a6a043710ad4958355b62c682b4767f292e
> >>
> >> Looks like we go back from udev rule to manually starting bluetoothd all
> >> the time, even if bluetooth device is not connected? Why are we going
> >> back to old behavior?
> >>
> >> Thanks for the info
> >
> > Should we revert that change downstream if we are not going to rely on
> > systemd and until udev is changed to break that setup (forcing us to go
> > to old "always running" behavior)?
>
> Udev rules are not supposed to spawn daemons. It never was. And now if
> you do that udev will kill your daemon after a certain timeout. You'd
> need to a) maintain that rule by yourself and b) patch udev to not do
> that.
>
>
> Lucas De Marchi
>
Will go with the old way of starting it with an init.d script then.
Thanks for the info.
On Sat, Jun 30, 2012 at 3:15 PM, Pacho Ramos <[email protected]> wrote:
> El vie, 29-06-2012 a las 22:23 +0200, Pacho Ramos escribi?:
>> I would like to know what distributions not relying on systemd are
>> supposed to get bluetoothd autostarted now that udev rule was dropped:
>> http://git.kernel.org/?p=bluetooth/bluez.git;a=commit;h=2ea98a6a043710ad4958355b62c682b4767f292e
>>
>> Looks like we go back from udev rule to manually starting bluetoothd all
>> the time, even if bluetooth device is not connected? Why are we going
>> back to old behavior?
>>
>> Thanks for the info
>
> Should we revert that change downstream if we are not going to rely on
> systemd and until udev is changed to break that setup (forcing us to go
> to old "always running" behavior)?
Udev rules are not supposed to spawn daemons. It never was. And now if
you do that udev will kill your daemon after a certain timeout. You'd
need to a) maintain that rule by yourself and b) patch udev to not do
that.
Lucas De Marchi