Gabe Black wrote:
> Ali Saidi wrote:
>   
>> On Sep 6, 2008, at 4:26 PM, Gabe Black wrote:
>>
>>   
>>     
>>> There are two major problems with using the DmaPort. First, I'd want  
>>> to
>>> send the interrupt -now- not when the DMA queuing latency, etc gets  
>>> used
>>> up. Second, DmaPort will fragment a packet which, while maybe  
>>> necessary
>>> in some cases, will break the atomicity requirements for the interrupt
>>> messages and might even make the parts of it get there out of order.
>>> While it may not -normally- fragment the message, if it does things  
>>> will
>>> be broken. I would rather not use something where such a large portion
>>> of its functionality is either not needed or counterproductive.
>>>     
>>>       
>> The only time the DMA port will break up the packet is if it's larger  
>> than a cache block. I can't imagine an interrupt message being that  
>> large. With MSI, interrupts don't magically get propagated. They  
>> device still has to arbitrate for the bus.
>> Ali
>>   
>>     
>
> I can't either, but I don't just want to hope that doesn't happen. Also, 
> there will still be arbitration on the bus, but the APIC doesn't have to 
> queue up a DMA transfer to send an IPI since it's not even transferring 
> anything.
> _______________________________________________
> m5-dev mailing list
> [email protected]
> http://m5sim.org/mailman/listinfo/m5-dev
>   

Also the DmaPort doesn't send it's packets marked uncachable which we 
established as being a requirement for the interrupt messages. I don't 
-want- to have to write a new port type, but what I need is 
fundamentally not a DMA transfer. Making the DmaPort the solution for 
things other than DMA is misleading, and not having it work for other 
things means I need to find a different solution. But like I said, I 
don't want to have to create a new port type, so I certainly wouldn't 
mind finding a way to use something that already exists.

Gabe
_______________________________________________
m5-dev mailing list
[email protected]
http://m5sim.org/mailman/listinfo/m5-dev

Reply via email to