On Fri, 5 May 2000, James Sutherland wrote:
> On Thu, 4 May 2000, Alexander Viro wrote:
> > On Thu, 4 May 2000, James Sutherland wrote:
> >
> > > On the contrary. Look at the issues with the NTFS driver now, for example:
> > > if the Windows [NT] source were available (with the restriction on patent
> > > usage) we could just read the source, and make the Linux driver work
> > > perfectly (well, as well as their version does, anyway :P)
> >
> > BS. Different model, different kernel API, different code practices
> > to the degree that their code doesn't look like C. Besides, we _have_
> > free <RMS-bait excuse="sorry, couldn't resist"> as in _really_ free, BSDL
> > rather than GPL</RMS-bait> NTFS driver that works. For UNIX kernel, BTW.
> > And it helped us which way?
>
> I wasn't suggesting COPYING the driver from NT! My point was that the FS
> is not well documented. Where is this working NTFS driver, by the way, and
> why didn't you suggest it to the user who was unable to read his NTFS
> partition with the one in the kernel ATM?
Because it's a FreeBSD driver, not a Linux one. And porting is not a
trivial exercise. Difference between NT and Linux is much more serious
than between FreeBSD and Linux, BTW.
Where? In their tree, indeed, where else. In /usr/src/sys/ntfs/*.[ch], or
on http://www.FreeBSD.org/cgi/cvsweb.cgi/src/sys/ntfs/?cvsroot=freebsd
if you want to look at the history and don't want to bother with CVS.
If we had working _Linux_ driver there would be no problems, right?
Unfortunately, it's from another UNIX - there are different kernels, you
know...
> > >From the pile of crappy code? Have you _ever_ read the code from project
> > that went hypercritical several years ago? As in, fixing the bug produces
> > more than one new bug... No?
>
> I wasn't suggesting trying to FIX NT or '9x - just use the source code to
> supplement documentation on all those undocumented function calls.
Bwahahahaha... If you have spare time and can produce documentation by
code - go ahead, there was a lot of people complaining about the lack of
documentation on _our_ VFS. Here you don't have to wait for court orders,
code is there and there is a list where you can ask about details.
> Kernel32, for example, exports a whole category of functions by ordinal
> only - we don't even have the function names, let alone any description of
For function names - read Lovecraft.
> what they do! The source would certainly help there...
I doubt it. With Linux it takes several months to understand the whole
core infrastructure, add more for any particular subsystem. And here you deal
with specimen from well-described family. Not true for NT.
> > BTW, what does it do on l-k?
>
> The thread started here; also, documentation of Windows [NT]'s innards
> could be useful in parts (like the NTFS driver).
Source != documentation. Producing the latter is _not_ a trivial task,
especially when the codebase is a spagettish pile of tapeworms.
-
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/
This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:16 EST