I thought I was being nice to myself when I upgraded from a P5-100 to
a dual-PII-266 setup, but have instead suffered nothing but hard
lockups since. :-( I doubt it's SMP related, as my new
motherboard/PII started locking up under windoze-95 just as it does
under linux-2.[01]. However, after running with your NMI watchdog on
2.1.88-SMP, I finally got an OOPS. Actually, the first few lockups
with your patch did not display any OOPSen; the machine evidently
locked up hard enough that even the NMI vectoring was somehow locked
out. But last night I got one, and wrote it down (below). What's
funny is that, while booting the machine this morning (to map
addresses via system.map), the machine locked up again (with an NMI
oops) before it had even initialized its disks! I wrote that one
down, too; the addresses of the most recent stack frames are identical
in both of the OOPSes.
Sorry, I did a 'make clean' after installing my kernel, so I don't
have any vmlinux to disassemble. :-/ Maybe this will be helpful to
people, however.
Kris
-------------
NMI Watchdog detected LOCKUP on CPU1, forcing oops
Call Trace: [<c010b104>(0)] [<c010a0c2>(20)] [<c01a4cd3>(56)] [<c010b6e2>(20)] [<40013d38>(52)]
<E>
>From System.map:
[T do_nmi+168] (0)
[t error_code+50] (20)
[T stext_lock+435] (56)
[t apic_timer_interrupt+34] (20)
[???] (52)
NMI Watchdog detected LOCKUP on CPU1, forcing oops
Call Trace: [<c010b104>(0)] [<c010a0c2>(20)] [<c01a4cd3>(56)] [<c010b6e2>(20)] [<c01d308d>(52)] [<c01d3101>(16)] [<c010a030>(4293189536)]
[<c0108284>(48)] [<c01d1968>(20)] [<c010029e>(28)]
<E>
>From System.map:
[T do_nmi+168] (0)
[t error_code+50] (20)
[T stext_lock+435] (56)
[t apic_timer_interrupt+34] (20)
[T start_secondary+17] (52)
[t do_boot_cpu+37] (16)
[T ret_from_intr+0] (4293189536)
[T cpu_cycle+68] (48)
[T start_kernel+268] (20)
[t L6+0] (28)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu