Re: [PATCH 0/5] isdn: deprecate non-mISDN drivers

Arnd Bergmann <[email protected]> Wed, 24 Apr 2019 20:54:04 +0200
Newsgroups gmane.linux.isdn.i4l.user
Message-ID <CAK8P3a08+bU_DxAQYi9MVCeZRjFwHz=g6yyiDLjF3RhfjAO7wA@mail.gmail.com>
On Wed, Apr 24, 2019 at 4:49 PM Birger Harzenetter <[email protected]> wrote:

> >We currently have three ISDN stacks in the kernel: the original
> >isdn4linux, the newer CAPI, and finally mISDN stack.
>
> Where mISDN and CAPI can be used in combination.

Can you elaborate on how those are combined? The supported hardware
for mISDN (HFC, fritz!card PCI, Seldbaur Speedfax, ...) and CAPI (AVM b1,
hysdn, gigaset, bluetooth cmtp) does not seem to have any overlap.

Did you mean that isdn4linux can be run on top of CAPI with capidrv?

> >The older stacks support additional features for data networks, but those
> >typically make no sense any more if there is no network to connect to.
>
> They are still there.
> And even if they weren't there are always industrial installations that rely
> on things that have been obsoleted otherwise.
>
> The main culprit here it, off course, that mISDN has never been a full
> replacement for I4L, unfortunately. But that's history.

Can you point to a specific installation that uses those obsolete drivers
on modern kernels? Generally I would assume that any installation
that is conservative enough to still rely on those drivers would have
long ago stopped upgrading major kernel versions.

> >My proposal for this time is to kill off isdn4linux entirely, as it seems
> >to have been unusable for quite a while. This code has been abandoned
>
> Since when?
> I just dug out my old laptop running 4.12 (I fried my current one last week)
> to check.
> While the card config tools for Icn and Eicon didn't want to compile and
> isdnlog was missing some pearl module, the rest compiles and seems to run
> fine.
> I didn not investigate the compile errors.

The last time that anyone seems to have actually done any isdn4linux
development appears to be in 2012 when Jiri Slaby sent a series of
cleanups for the the tty code, but it's not entirely clear if he tested it.

in 2010, Tilman Schmidt sent the last bugfix that was clearly tested.

Anyything else in the last 10 years was treewide cleanups or build fixes.

> Anyway I certeinly use the divertctrl application from time to time.
> Much better to carry a USB dongle than a whole telephone that needs setup.
> I remember Karsten Keil told once that he started an replacement for mISDN,
> but AFAIK that has never seen the light of day.
> Also using minicom on an ttyI to quickly make a test call can be quite usefull.
>
> And maybe that idea isn't mainline any more, but I think that the ipppd stuff
> and certainly the ttyIs have some educational value.

No educational value is lost by removing the code, one can still read the
sources for linux-5.1 in the git history.

> >for many years and it does get in the way of maintenance.
>
> That is a valid point, off course.
> On the other hand having to keep old versions for tasks no longer supported
> has become quite a task in itself for some users.
>
> So I think in the end the workload is just shifted elsewhere and the real
> issue is coordination of efforts. Which seems be be a huge topic in it's own
> right.
>
> Indeed, after I posted that pach recently the question has come up again:
> How amy versions of mISDN do actually exist?

There is exactly one version of mISDN in the mainline kernel. Any other version
is meaningless to this discussion.

> And from the technical aspect, I4L and mISDN seem to have quite a bit of
> duplicated code. So the ideal situation would be to share those parts.
> But off course someone would have to do that.

This is easily solved by removing I4L ;-)

> >CAPI in turn is not quite as obsolete, but as I have been unable to find
> >any users, I would suggest moving it to drivers/staging while we
>
> I've never been a fan of CAPI, but AFAIK, the combination of CAPI and hylafax
> if the only available implementation of G4 Fax which makes a good argument
> for me.
> It also seems to be a (partial?) replacement for data connections with mISDN,
> but I have never looked in to that.

Could you be more specific which hardware you still see being used with
CAPI+hylafax? Is it all the avmb1 driver, or also the other ones (gigaset,
hysdn, cmtp)?

> >If anyone is still using isdn4linux and CAPI with modern kernels, please
> >speak up now.
>
> That's mean. You should have asked for OR, not AND. :-)
>
> Not in production, but as stated above I4L is still a valuable tool for me.
> OTOH running production machines on older kernels is probably easier than on
> a laptop.

How much longer do you expect to need it? SLES15 and RHEL8 will have
4.x kernels for many years to come, and even Ubuntu 18.04 will still
get updates until 2023.

      Arnd