Thanks to all for the pointers and suggestions. It will probably take me a while to follow up on these and make another proposal.
On Thu, 2006-07-13 at 09:02 -0700, Roland Dreier wrote: > > > A cleaner solution would be to make the dma_ API really use the device > > > it's passed anyway, and allow drivers to override the standard PCI > > > stuff nicely. But that would be major surgery, I guess. > > > Clean but expensive, you should not force the rest of the kernel > > to eat the cost of something you want to do when it's totally > > unnecessary for most other users. > > OK, fair enough. > > > For example, x86 never needs to do anything other than a direct > > virt_to_phys translation to produce a DMA address, no matter what > > bus the device is on. It's a single simple integer adjustment > > that can be done inline in about 2 or 3 instructions at most. > > <pedantic>Except x86 needs to handle systems with IOMMUs now...</pedantic> > > > If you need device level DMA mapping semantics, create them for your > > device type. This is what USB does, btw. > > Makes sense -- Ralph, I would suggest looking at USB as a model. > > - R. _______________________________________________ openib-general mailing list openib-general@openib.org http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general