> Well, I doubt that the MTRR code is to blame, since it's pretty
> self-contained. More likely it changes the kernel footprint slightly
> and in the joystick driver there is a bug that is triggered by the
> change.
Quite unlikely - the major is registered quite soon in the init_module
function, so, there would need to be a major memory-overwriting bug
in the timer calibrating routine, which I doubt.
> I also looked at your kernel logs, and I note this:
> Sound initialization complete
> js: Version 1.2.13 using 333 MHz RDTSC timer.
> js: no joysticks found
> [...]
> js: Version 1.2.13 using 333 MHz RDTSC timer.
> js: unable to get major 15 for joystick
>
> This implies that you're trying to load the driver twice, which is
> wrong. You can easily check this by looking at /proc/devices and
> seeing what is allocating major=15.
>From the logs the cause of this is clear - the joystick is both compiled
into the kernel, and attempted to load as a module after the sound card
is initialized. This, of course, can't work, since the joystick driver
will try to register major 15 in both cases, and in the second it'll fail.
Btw, some of the logs also show that the sound module is compiled into
the kernel, and some only have it inserted as a module ...
I think the answer is to check the configuration, and re-compile and
re-install the kernel.
On the other hand, all the config files show CONFIG_JOYSTICK = m, which
is odd.
Have fun,
Vojtech
-
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/