Re: Machine check expection panic

[kwijibo]

Out of curiosity I decided to try this on some other Athlon
systems I have.  I tried it on a dual Athlon MP 2400(2GHz)
system with a Tyan 2462 motherboard.  Also I tried it on a
single Athlon XP 1800 with a Asus A7V motherboard.  They
both booted fine with the 2.6.0-test2 kernel and the machine
exception code in it.  So I am thinking either it is something
with the older CPU's or the CPU is actually borked.  Like I said
though I have been using those 1.2GHz processors for a long time
with no problems.

Steve

Andi Kleen wrote:

> > The CPU's aren't overclocked and have worked fine for
> > me under much heavier loads than booting a kernel for
> >    
>
> It could be corrected ECC errors in the cache. If that
> happens I would consider it a hardware problem
>
> (now hidden with the disabled bank).
>
>  
>
> > at least a year. Using the 2.4 kernel that is. Once
> > I remove the exception code from the kernel it boots
> > fine and runs fine under any load I put it under.
> >    
>
> I maintain that such a magic hack needs at least a big fat comment.
>
> I still find the change very suspicious, there isn't any errata that 
> says that bank 0 is bad on Athlon.
>
> Also disabling a whole bank just for some buggy CPUs is quite a sledgehammer,
> it would be probably better to identify the bank 0 sub unit that causes it
> and only turn that off.
>
> -Andi
>
>
>
>  


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/