Re: [PATCH] pktcdvd: remove driver.

From: Pali Rohár
Date: Sat Dec 24 2022 - 11:36:14 EST


On Saturday 24 December 2022 17:18:51 Borislav Petkov wrote:
> On Sat, Dec 24, 2022 at 05:00:55PM +0100, Pali Rohár wrote:
> > Maybe it would be a wise to read a documentation which is in the kernel
> > source tree?
>
> I guess you mean udftools...

Yes, pktcdvd tools (pktsetup, pktcdvd-check, cdrwtool) are (for
historical reasons) in udftools project. Ok, maybe it is unintuitive to
find them here but it is documented in kernel tree.

> > Or at least read the deletion patch itself as it is linked from there?
>
> You mean the documentation file is pointed at there?

Yes. In Kconfig option which is being removed, is a documentation link

- See the file <file:Documentation/cdrom/packet-writing.rst>
- for further information on the use of this driver.

> > Or what else could be easier than this?
>
> Well, apparently it ain't as easy because people do not necessarily see
> it how you see it. That's why I'm asking.

Yes, in more cases it is not easy. But in this case, when kernel in-tree
documentation about this driver was updated in the last year, it is lot
of easier than in other cases to get more information about it.

I sent this patch to keep links up-to-date for packet-writing.rst file:
https://lore.kernel.org/linux-doc/20220210192200.30828-1-pali@xxxxxxxxxx/
(and it was merged)

In documentation are mentioned 3 projects which are related to packet
writing and all have valid homepages with contact information.

> I have removed ancient stuff in the past myself and it is not always
> easy to go dig out who uses it and whether it is used at all in the
> first place.
>
> And people do not always reply and projects are dead and they maybe use
> it but the machine which has this hw hasn't been booted for a decade and
> it ain't worth the enegry to power it back on and so so on and so on...
>
> So you don't have to get all worked up about it - if it is really used,
> I'm sure the maintainers involved will do the right decision. The point
> is, finding out whether something still has users and with the latest
> kernel is not always trivial.

I agree that finding out such information is hard. But do not take me
wrong, but if people are lazy and do not look into in-tree kernel
documentation and check it, then I'm loosing motivation to keep in-tree
kernel documentation up-to-date...