[Coco] Linux RBF filesystem support
Chuck Youse
cyouse at serialtechnologies.com
Mon Oct 20 20:52:51 EDT 2008
On Mon, 2008-10-20 at 11:34 -0400, John W. Linville wrote:
>
> > Whichever ass-hat wrote that document isn't writing it for me:
> >
> > Executive Summary
> > 18 -----------------
> > 19 You think you want a stable kernel interface, but you really do not, and
> > 20 you don't even know it. What you want is a stable running driver, and
> > 21 you get that only if your driver is in the main kernel tree. You also
> > 22 get lots of other good benefits if your driver is in the main kernel
> > 23 tree, all of which has made Linux into such a strong, stable, and mature
> > 24 operating system which is the reason you are using it in the first
> > 25 place.
> >
> > Total horseshit. And for the record, the main reason why so many people use Linux is because so many other people use Linux. It's as simple as that.
>
> Well, I'll just vaguely smile and pretend that you've said something
> meainingful, then drop the discussion... :-)
My point was
1. I do want a stable kernel interface, despite his implication that I
don't and I'm too stupid to know better.
2. I do not want a stable kernel interface because I want some stupid
driver to work.
3. I do not use Linux because it's a strong, stable and mature operating
system. I use Linux because I'm forced to for various reasons under
various circumstances. The reasons my clients force me to use Linux are
varied and often based on incomplete information and bad assumptions.
To simplify the unstable API problem as "I want my driver to work" is
really quite narrow, and the author is on shaky ground at best. Some of
the reasons the author of the above "white paper" cites as justification
for an unstable API (e.g., alignment of structures caused by different
versions of the compiler) are absolutely ridiculous. The fact that I
have to compile a kernel module against _exactly_ the kernel I am
running - thousands of compile-time options included - is ludicrous.
That Linux HAS thousands of compile-time options is similarly ludicrous.
Linux often gets the job done, but it doesn't mean it's not a big piece
of shit. It simply has a lot of market share, which means it will
continue to have a lot of market share. That's the way of things. But
let's be honest and recognize that Linux gained all that initial
traction not on technical merit, but on "religious" fervor, the
rebellion against Microsoft, and the perception that Linux was "cool" by
a generation of kids who'd never seen anything but Windows.
Ugh, I have to stop here before I start bleeding out my orifices.
C.
More information about the Coco
mailing list