Messages in this thread | | | Date | Fri, 23 Apr 1999 22:59:00 -0400 (EDT) | From | <> | Subject | Re: How to get usefull debugging info... |
| |
Ok... I recompiled my kernel to include magic sysrequest. I figure I will leave this on with ever kernel I compile so I can try and help any way possiable. Anyhow... I think I may have found my problem... it may not be software. I think my CPU voltage may be fluctuating way out of bounds. I have lm_sensors running and I jsut found out that the hardware monitering script is not running @ boot for some reason... I am going to try and figure this one out tonight. Where is a good source for reading about conf.modules? I need to make it load all of the sensors package and not unload them. To cut to the chase... I started health monitering by hand and within about 2 hours my machine halted itself. I use lm_sensors to halt the machine because I figure and idle processor with fan will cool faster than if the machine jsut shut off and had no cooling fans. but when I booted back up and looked at the error message that the health script spit out before it shutdown it was an undervolt on CPU 0... I have already increased the tolerances quite a bit. It was running a 2 volt core at 1.90 volts. Is this enough to cause errors? I think the machine is on the same power as my fridge... Could this be causing it? Also... what temp should a PII run @ under full load.
On Wed, 21 Apr 1999, Andrea Arcangeli wrote:
> On Wed, 21 Apr 1999 kernel@duct.dhs.org wrote: > > > > >*snip* > >> The first step is to be on console at lockup time (luck ;), and press > >> SYSRQ+ALTRIGHT+P and write on paper all different EIP address, then reboot > >> and resolve them through system.map. > > > >What do you mean by resolve them? Sorry If I don't know enough to even > >attempt what I am talking about but I'd like to try. > > I'll produce an example. You'll see on the console: > > EIP: 0010:[<c0107298>] EFLAGS: 00003246 > EAX: 00000000 EBX: c009c000 ECX: c009c000 EDX: c01ae780 > ESI: c009c000 EDI: c01e48e0 EBP: c01c7fac DS: 0018 ES: 0018 > CR0: 8005003b CR2: 0827908c CR3: 00101000 > > So the interesting number is c0107298. > > Then open the System.map of the running kernel (linux/System.map after a > compile) and browse the symbols: > > c0107248 T disable_hlt > c0107250 T enable_hlt > c0107258 T cpu_idle > ^^^^^^^^ > c01072ac T sys_idle > c01072d0 T machine_restart > c01073c8 T machine_halt > c01073cc T machine_power_off > c01073d0 T show_regs > > The interesting number c0107298 means that the CPU was running in the > kernel in the function cpu_idle(). > > Andrea Arcangeli >
- 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/
| |