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

[Grant Grundler]

On Mon, Aug 16, 2004 at 08:33:03PM +0300, Michael S. Tsirkin wrote:
> > It seems this would apply to standard INTx mode as well.  Do you know
> > why Mellanox didn't use this in THCA?  In any case, I'll have to
> > benchmark this approach.
> 
> It does apply to the standard INTx mode. However with standard
> INTx mode there is more of a chance that by the time you peek
> the EQ, the EQ is empty and you end up getting an extra interrupt.

Erm. I wonder if the EQ is empty because the interrupt was delivered 
to the CPU before the DMA hit host mem. MSI would avoid this
race condition.

grant