:: Re: [DNG] vdev
Top Page
Delete this message
Reply to this message
Author: aitor_czr
Date:  
To: Joel Roth, dng
Subject: Re: [DNG] vdev

Hi,

On 08/05/2016 09:09 PM, Joel Roth <joelz@???> wrote:
> Didier Kryn wrote:
>> >     No news from Jude for almost one year:-(
> Commits to vdev on github are as recent as April this year.
> Also, I think his continued work depends on feedback from testers.

>
>> >     I'm partial time reading documentation and sources and considering
>> >rolling my own hotplugger, simpler than vdev - Jude has done an enormous
>> >work and vdev is a complex software with hundreds of scripts, too complex
>> >for me.
> sysvinit also has lots of scripts!

>
> But vdev has 33 not in helpers/ and all of those are fairly
> simple.
>
>> >     Reading uevents from netlink and parsing them is easy, but I'm not sure
>> >to know what to do with these events. My first conclusion is that
>> >     when there is MODALIAS=something, then I just have to modprobe a
>> >module's alias
>> >     when there is FIRMWARE=something, I just need to transfer the firmware
>> >by some well defined API
>> >     otherwise it is more complicated because there are many different cases
>> >(not all drivers provide the same uevent contents); but the job is to
>> >create/delete a device file, and set owner/permissions and symlinks; but
>> >some devices are neither char nor block (eg network devices) and I don't
>> >know what to do with them.

>> >
>> >     For this reason, I'm tempted to clone mdev and only change the way it is
>> >invoked. mdev is spawned for every uevent through the
>> >/proc/sys/kernel/hotplug mechanism; instead I would prefer one long-lasting
>> >daemon reading uevents from netlink. The design of mdev makes full sense for
>> >embeded devices, but not for others.

>> >
>> >     I'm sure mdev is very well written, because it has been written,
>> >maintained, and reviewed by highly skilled professionals. I have also found
>> >a netlink reader at Skarnet, which is certainly written with great care but
>> >does not seem very usefull since it just forwards the uevents to standard
>> >output as they come; but It might be instructive to look if it uses any
>> >special trick to read the netlink.

>> >
>> >     Any suggestions are welcome - there are still a lot of devices I don't
>> >know anything about.
> It's your decision to write your own. However Jude's work is
> technically interesting, and can be tested without changing
> out udev.

>
> Perhaps it just needs some additional user help in the form
> of documentation, tutorials, etc.
>
> have fun,
>
> Joel
>


As i announced yersterday in the IRC Channel, i just uploaded the
repository of gnuinos containing (in addition to linux-libre-4.6.2 and
simple-netaid) the packages of vdev:

deb http://packages.gnuinos.org/ jessie main
deb-src http://packages.gnuinos.org/ jessie main

Linux-libre has been built using libudev-dev. Now, i'll try it again
using Jude Nelson's *libudev-compat* (which should be recalled to
libudev-compat-dev ?? ), instead of libudev-dev.

I also should add the following line to debian/control:

Provides: libudev-dev

Cheers,

Aitor.