Jonathan Franks <[EMAIL PROTECTED]> writes:
> not sure if this will help or not, it seems that you might have a
> different issue here but this worked for me and I figured I'd at
> least point it out:
>
> http://cvs.openbsd.org/cgi-bin/query-pr-wrapper?full=yes&numbers=5064
>
>
Thank you Jonathan for the link but I've already read this thread and
the bug report yesterday. It's apparently a different problem for me.
I don't know if I must initialize a bug report for my specific kernel
crash (panic: can't deal with not-all-lapics interrupt yet!). Some
changes have been made on /usr/src/sys/arch/i386/i386/mpbios.c since
3.9 was released, but it's only for the amd64 architecture.
Here an excerpt from mpbios.c:
[...]
if (entry->type == MPS_MCT_IOINT) {
sc = ioapic_find(id);
if (sc == NULL) {
printf("mpbios: can't find ioapic %d\n", id);
return;
}
mpi->ioapic = sc;
mpi->ioapic_pin = pin;
altmpi = sc->sc_pins[pin].ip_map;
if (altmpi != NULL) {
if ((altmpi->type != type) ||
(altmpi->flags != flags)) {
printf(
"%s: conflicting map entries for pin %d\n",
sc->sc_dev.dv_xname, pin);
}
} else {
sc->sc_pins[pin].ip_map = mpi;
}
} else {
if (id != MPS_ALL_APICS)
panic("can't deal with not-all-lapics interrupt yet!");
if (pin >= 2)
printf("pin %d of local apic doesn't exist!\n", pin);
else {
mpi->ioapic = NULL;
mpi->ioapic_pin = pin;
lapic_ints[pin] = mpi;
}
}
[...]
Francois
--
http://www.chambaud.org
