As for making PCMCIA part of the standard kernel... it is something
that I've obviously thought about before. My (possibly outdated,
probably exaggerated) impression was that based on the difficulty I've
had in getting PCMCIA-friendly changes into the kernel over the years,
that Linus was, if not hostile, at best never particularly interested
in it. Also, given the size and complexity of the PCMCIA package, it
seems more easily maintainable as a separate entity, even if it means
I have to deal with the kernel version skew issue from time to time.
I'm not sure it was/is the best way to go and may change my mind in
the future: I'm sure there's a cost to separate distribution, in that
regular kernel developers are less likely to look at the code. But as
I'm sure Linus would agree, having final editorial control also has
its advantages.
I always thought a good reason for keeping PCMCIA separate was that new
PCMCIA cards come out faster than we can keep up with changes to the
stable kernel base. By keeping PCMCIA separate, David can support the
latest PCMCIA cards as they come out, and users who are using Linux 2.0
don't have to wait months for the next stable release to come out so
that their PCMCIA card can be supported.
That's one of the reasons that for a long time, I didn't bother getting
the Rocketport driver into the kernel --- most of the users were using
Linux 1.2 and 2.0 kernels (I still check to make sure the latest
Rocketport driver works under Linux 1.2, because there are a few
die-hard customers who don't quite trust the stability of the 2.0
kernel, and the 1.2 kernel works well enough for them).
Of course, it means that adapting my drivers to advancing kernel
revisions is completely my responsibility, but I accepted that when I
choose that particular deployment path. Now that the Rocketport driver
is part of the kernel, it means that when interfaces changes and people
modify the driver to reflect those changes, I have to take those changes
and migrate them to my "kernel-independent" version of the driver, which
I still support for the sake of the Linux 1.2 and 2.0 users of the
driver.
I imagine that the PCMCIA driver has similar issues, and there are a
large number of Linux users who have benefited greatly because it hasn't
been part of the standard kernel. (Or at least, that there was a
version of the PCMCIA driver which wasn't tied to a kernel release.)
- Ted
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/