Re: can device drivers return non-ram via vm_ops->nopage?

[Jeff Garzik]

Russell King wrote:
> I don't think this addresses the scatter-gather case I mentioned above.
> Or if we are, we've rewritten ALSA before hand to use Linux scatterlists
> along side several dma_alloc_coherent mappings and have the ability to
> mmap these as well.
>
> Remember that we're fond of telling driver writers to use scatter gather
> lists rather than grabbing one large contiguous memory chunk...  So
> they did exactly as we told them.  Using pci_alloc_consistent and/or
> dma_alloc_coherent and built their own scatter lists.

Agreed...  though IMO that can handled by considering DMA S/G as just 
one more set of helper functions that the driver writer should not have 
to implement ;)  dma_sg_setup_mmap() could function as a peer alongside 
dma_setup_mmap(), mmio_setup_mmap(), etc.  Providing such to driver 
writers gives them incentive to use S/G lists as well as incentive not 
to invent their own mmap(2) setup and handling code.

	Jeff



-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/