> But in no case should Linux BY DEFAULT disable interrupts during IDE
> disk I/O. If FreeBSD does not, why should Linux?
Using safe defaults, in combination with a user-level tool which enables
additional performance features, has the potential to be efficient, while
still supporting low-end hardware.
In any case, current IDE designs support bus mastering DMA transfers, and
the kernel enables DMA automatically in many cases.
When using DMA instead of PIO, the kernel no longer spends much time
handling IDE interrupts, and it therefore doesn't matter whether we will
enable or disable interrupts during IDE I/O.
Gadi