Juergen Leising wrote (in part):
> > I have a similar system R.H.6.0, kernel 2.2.5-15, CPU is 166MHz Pentium, and
> > I cannot get Ftape to work either. The version that is supplied is quite
> > obviously 3.04d, because I can make it run enough to get entries in the
> > /var/log/messages file and it so identifies itself. The main trouble I have
> > with it is that it insists on running the floppy disk controller even though
> > I have a dedicated CTC-2Mb controller just for the tape drive. This worked
> > for R.H.5.0, but not for 6.0. I stuck parameters into the /etc/conf.modules> file
>as
> follows:
> >
> > # Load ftape/zftape automatically
> >
> > alias char-major-27 zftape
> > pre-install ftape /sbin/swapout 32
> > # Try to set CTC-2Mb floppy controller
> > # and
> > # Try to reduce overruns: can be [1-16]: higher: less overruns
> > options ftape ft-fdc-base=0x370 ft-fdc-irq=6 ft_fdc_dma=2 ft-fdc-threshold=16
>
> ...under 4.0x and under those unstable drivers the hyphens have to be replaced by
> underlines, i.e. ft_fdc_base instead of ft-fdc-base.
You are absolutely correct. After many hours work, hyphens and underlines look the
same.
It may be that if I noticed that before I got a working version of 4.x-1999_04_25, that
the 3.04d version that came with Red Hat 6.0 (2.2.5-15) would have worked. But I am
just
as glad I have the newer version working.
> > # Try harder to reduce overruns: can be [1000, 500, 250]: lower: less overruns
> > #options ftape ft_fdc_rate_limit=250
>
> ...don't know if this option is essential with your drive, but in case it is
> the comment sign "#" has to be removed.
Not absolutely necessary. With the new driver, I do need it to say:
options ftape ft_fdc_rate_limit=1000
Of course it will work without that, but it stumbles for a while until it sets itself
to
1000 by itself.
My CTC-2Mb controller claims to run at 2Megabits per second, and it seems to under
Windows
95, but not with ftape. Before condemning ftape, I notice that the Windows version does
some "shoe shining" at that speed, as does ftape. But the windows version cannot slow
itself down as ftape can. With it set to 1000, it does none (unless I am using the
machine
heavily for other things at the same time).
> (...)
> > make[2]: Leaving directory `/usr/src/ftape-4.x-1999_04_25/ftape/compressor'
> > set -e; for i in setup lowlevel internal parport zftape compressor; do make
> > -C $i modules; done
> > make[2]: Entering directory `/usr/src/ftape-4.x-1999_04_25/ftape/setup'
> > make[2]: *** No rule to make target `/usr/src/linux/.config', needed by
> > `.ftape-setup.d'. Stop.
> > make[2]: Leaving directory `/usr/src/ftape-4.x-1999_04_25/ftape/setup'
> > make[1]: *** [modules] Error 2
> > make[1]: Leaving directory `/usr/src/ftape-4.x-1999_04_25/ftape'
> > make: *** [all] Error 2
> >
> > It obviously insists on a .config file, which I do not have in the
> > /usr/src/linux directory, since I have never configured or built a kernel,
> > nor do I wish to. I did not need this when building 3.04d on Red Hat Linux
> > 5.0. All I needed to do was fill in the MCONFIG file correctly.
> > I could [and did]
> > install the kernel source, make a .config file, and perhaps even compile the
> > OS. Since the help messages take over 86 pages, and I do not understand them
> > well enough to configure my kernel, I hesitate[d] to do this. As far as I can
>
> well, you don't really need to perform the whole compilation process. Just
> to produce .config you can restrict yourself to
>
> install the kernel sources
> make xconfig,
> have a glance at the section "ftape" to ensure, the defaults don't
>have
> ftape compiled into the kernel;
> change nothing (else) in the setup
> save and exit,
> and stop at this point, i. e. no make dep clean and so on.
I guessed this procedure. I did not know the implications, but since I was not going to
compile (and, more to the point, install) the new kernel, I figured I had nothing to
lose.
And it worked.
> But I must admit I don't know why "make" insists on [using]/usr/src/linux/.config
> Does anybody else know better?
I suppose it is a "bug" somewhere in the makefile tree. I assume this is one of the
reasons why this version is called "unstable" and that Claus will get to it when he has
time.
--
Windows is not the answer. Windows is the question. The answer is no.
Jean-David Beyer
Shrewsbury, New Jersey
S/MIME Cryptographic Signature
