Re: [PATCH 0/8] git pull request for tip/tracing/core

[Kenji Kaneshige <kaneshige.kenji@jp.fujitsu.com>]

Steven Rostedt wrote:
> On Thu, 12 Feb 2009, Kenji Kaneshige wrote:
>> Luck, Tony wrote:
>>>>> Before we go and make the change, Peter brought up a good point on IRC. Is
>>>>> there any reason that ia64 needs 1 << 14 IRQs?  That's 16384!
>>>>>
>>>>> Perhaps the better solution wolud be (if possible), to simply lower the
>>>>> number of bits.
>>>> i'm the wrong person to be asked about that. (Cc:-ed the right people)
>>> People build some pretty big systems on ia64.  SGI's largest has 4096
>>> cpus ... so 16384 IRQs is only 4 per cpu.  That doesn't sound like very
>>> many to me.
>>>
>>> Fujitsu added the vector domain support for ia64 to get around the shortage
>>> of IRQs for large machines.  Added them to the Cc: list to see if they have
>>> comments on how many IRQs are needed.
>>>
>> The 1024 IRQs are enough for GSIs on our maximum configuration. But
>> if the devices are MSI/MSI-X capable, it could be more than 1024. 
> 
> But you would never expect more than 1024 nested interrupts all on the 
> same CPU?  That is, to have over 1024 interrupts interrupting each other?
> 

Itanium processor has 240 vectors for external interrupts and
has a mechanism to classifies them to 16 priority classes. The
ia64 linux limits maximum nested interrupts depth to 16 using
this mechanism.

Thanks,
Kenji Kaneshige