Sure it can be DMA. We're doing a lot of things differently, and it's
entirely conceivable that timings just changed for the worse (or better,
depending on how you look at it).
Easy to test, in fact - run bonnie on 2.0.x and a later 2.1.x, and look at
the numbers. If they're consistently different one way or the other it
means that we're hitting the disk subsystem harder/less hard.
> The only common theme right now seems to be SMP and SCSI, and thats SMP
> machines in the SCSI case (not option), unknown in the IDE case.
Which may well be that there is a problem with cache timings or something.
One thing that is new in 2.1.x is that we're a _lot_ more parallell, for
example. Another difference is that we react to interrupts a lot faster
under SMP.
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html