Re: [PATCH] Add mem_nmi_panic enable system to panic on hard error

[Hidetoshi Seto <seto.hidetoshi@jp.fujitsu.com>]

Alan Cox wrote:
>>I could find your code in the -mm tree. It looks good :-)
>>But just one point, isn't it better to use die_nmi() instead of panic()?
>>
>> >>        if (panic_on_unrecovered_nmi)
>> >>                panic("NMI: Not continuing");
> 
> 
> At the point we halt for an unrecovered NMI the call trace isnt likely
> to be interesting and the message being visible is most important. I'm
> not sure if die_nmi would be better therefore - dunno what other folks
> think ?

Indeed, that's true.
In this case of an unrecovered NMI, the call trace is useless.
Thank you for pointing it out.

H.Seto