Messages in this thread | | | Date | Wed, 15 Nov 2000 13:23:53 +0100 (CET) | From | Ingo Molnar <> | Subject | Re: [prepatch] removal of oops->printk deadlocks |
| |
On Wed, 15 Nov 2000, Andrew Morton wrote:
> [...] Problem is, we're getting some reported lockups in which we need > to know what the other CPUs are doing (the other 1%). If a critical > spinlock is stuck on, we don't get to type `dmesg'.
i know - but we cannot please everyone, so i went for the 99% :-)
> So... In this updated patch the console_silent() call remains > as you designed it, but the nmi_watchdog kernel boot parameter has > been overloaded so that > > nmi_watchdog=2 > > will now cause _all_ NMI oops messages to be displayed on the console. > > Is this OK?
this collides with the UP-IOAPIC path's use of nmi_watchdog == 2 ...
i'd rather suggest a cleaner, "nmi_watchdog=2,verbose" (default: silent) type of boot parameter, it looks like the NMI watchdog needs more parameters.
Ingo
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org Please read the FAQ at http://www.tux.org/lkml/
| |