[openib-general] [PATCH] mthca updates (2.6.8 dependent)

[Michael S. Tsirkin]

Hello!
Quoting r. Roland Dreier (roland at topspin.com) "Re: [openib-general] [PATCH] mthca updates (2.6.8 dependent)":
>     Roland> Hmm... that seems like a really ugly "feature" to allow
>     Roland> interrupts to pass posted writes within the chipset.
> 
>     Michael> No, I was talking about interrupt messages in PCI
>     Michael> Express.
> 
> I haven't looked at this in a while but I remember reading the PCI
> Express spec and deciding that interrupt messages sent after a posted
> write cannot pass the write within the PCI Express world.  I thought
> the issue is that once the PCI Express root complex has received the
> interrupt message, it may process it and raise an interrupt with the
> CPU before earlier writes have made through to memory.
> 

Yes, thats it. So PCI Express interrupts are similiar to PCI - ordering
versus writes may not be guaranteed. No idea if this happends in practice
though.

MST