Re: [PATCH] de4x5.c [was Re: pre-2.1.116-1 breaks Cyrix 6x86 support]

Gabriel Paubert (paubert@iram.es)
Tue, 11 Aug 1998 19:57:02 +0200 (METDST)


On 11 Aug 1998, Zlatko Calusic wrote:

> Linus, could you please apply appended patch to de4x5.c, unless David
> has even newer version (even though this one works perfectly).
>
> Patch allows the kernel to be compiled with egcs/pgcc/gcc 2.8 (without
> it kernel would Oops in the early booting stage) and helps to get rid
> of the annoying messages also during boot stage (see below). David
> wrote the patch, I'm just trying to get it in before the code freezes. :)

As author of the *kludge* that allows the driver to be succesfully
compiled with egcs/pgcc/gcc-2.8.x, I have to shout as loudly as I can that
it is intrinsically UNSAFE. It is a workaround that happens to work by
sheer luck, I came to it after quite a large number of experiments
compiling the driver and reading the generated assembly.

Any newer version of these compilers might break the driver again (or for
that matter many things using string functions in asm-i386/string.h).
Something has to be done about string.h before 2.2 anyway.

Ok, you've been warned :-)

> without the patch:
>
> eth1: region already allocated at 0xec00.
> eth1: region already allocated at 0xec00.
> eth2: region already allocated at 0xec00.
> eth3: region already allocated at 0xec00.
> ...
> eth7: region already allocated at 0xec00.
>

Zlatko, please take the code freeze for what it is ;-) Don't forget what
Linus said when 2.1.115 came out:
`Bug-fixes are still (and will always be) accepted'

The fix to the multiple eth found is _obviously_ a bug fix.

Gabriel.

-
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.altern.org/andrebalsa/doc/lkml-faq.html