Messages in this thread | | | From | Andi Kleen <> | Subject | Re: [PATCH 2.6.13] x86_64: Rename KDB_VECTOR to NMI_VECTOR | Date | Fri, 9 Sep 2005 10:06:31 +0200 |
| |
On Thursday 08 September 2005 18:39, Tom Rini wrote: > The existing hook from KDB in the IPI code is really just a hook for the > NMI vector. We rename the vector thusly and then it's up to the > debugger to handle things from do_default_nmi().
Jan Beulich pointed out some problems with this:
- First NMI vector is always 2 in the x86 architecture, so calling another vector NMI_VECTOR is misleading.
- And when an NMI is forced in ICR the x86 architecture specifies that the interrupt is delivered to vector 2.
Intel IA32 vol 3 8-23:
>>
100 (NMI) Delivers an NMI interrupt to the target processor or processors. The vector information is ignored. <<
So how did this ever work? I suspect you just got reentered via the hooks in the NMI handler, but never through this vector.
I think I'll just remove this unless someone can explain how it is supposed to work.
Thanks, -Andi
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/
| |