> Once again, I'm asking this question.  isa/dma.c seems to support dma in a
> much better way (using bounce buffers and page-sized dma rather than being
> limited by 64k) but is anyone out there using more than 16M in an ISA machine
> with DMA-using devices who can testify that 1.0A works correctly in that
> configuration?
> 
> Thanks,
> --Michael ``16M just isn't enough'' Graff

I know the feeling.  But this ground has been covered before...
The bounce support in isa/dma.c is only for devices that use the
motherboard DMA controller for their DMA.  This would include things
like the floppy controller and driver (anything that takes a drq and
actually uses it).  Bus-mastering controllers don't use the
DMA controller and don't have a drq(/arq) channel assigned to them, so
they aren't helped by this fix.  Thus, ISA bus-mastering devices, like
IDA IDE controllers, the AHA-154x's, etc have the 16M limit still.

Local/PCI bus controllers don't have this limit, this includes VL Bus
IDE and SCSI controllers (like the BT-445S), EISA controllers, like
the AHA-1742, and PCI controllers like the BT-946C and the NCR 53C810
are okay.

-Mark