AIC7xxx in Kernel 2.4.3

AIC7xxx in Kernel 2.4.3

[Peter Rottengatter]

Hi

This driver seems to be pretty broken, the way it is. It does not compile.
The new author, Justin T. Gibbs, has been careful in avoiding to mention
his e-mail address in his code :-( Hence the post to this list.

As the first problem, the compile stops in aicasm_gram.c because in
aicasm_gram.y the author forgot a function prototype. The compiler assumed
an implicit declaration which was incompatible with the final definition
of the function. The fix was easy of course.

Next, the code #includes a file called db1/db.h. To me this seems to be a
header file for the Berkeley database version 1. Debian does not provide
development packages for this rather old version, in fact, not even my
old cds from earlier linux distributions had this file. db2/db.h does not
work.

Possibly people by accident had this old file on their machine when the
new kernel was test-compiled. So far, all kernels I compiled by myself did
not seem to depend on any even remotely unusual header files. So this
surprises me a lot. Does not look good for a release that's meant to be
"stable", does it ?

Cheers  Peter

Re: AIC7xxx in Kernel 2.4.3

[Justin T. Gibbs]

>
>Hi
>
>This driver seems to be pretty broken, the way it is. It does not compile.
>The new author, Justin T. Gibbs, has been careful in avoiding to mention
>his e-mail address in his code :-(

I actually don't believe in putting email address in code.  They become
stale far too easily.  If you ever want to find me, type my name
into a Yahoo search.  I did this today and was a little surprised at
the number of acurate hits. ;-)

>Hence the post to this list.

You should really check the archives before posting to LK.  This has
been discussed *a lot*.

The version that was released in 2.4.3 was stale weeks prior to
that final kernel cut.  I'm working on getting revised versions into
2.4.4.  If you want to upgrade to something newer, try the 6.1.9
release from here:

http://people.FreeBSD.org/~gibbs/linux/

Just be sure to configure the bus settle delay to 5000ms as the default
in that release causes a timeout.

You can also configure the aic7xxx_old driver.

--
Justin