Messages in this thread | | | Date | Thu, 3 Feb 2000 09:58:27 -0700 | From | Ray L <> | Subject | Re: IDE performance question |
| |
hi mark
On Thu, Feb 03, 2000 at 08:56:47AM -0500, Mark Hahn wrote: > the problem is a lack of fairness/tuning of the request-queue. with > ide, one stream of IO can easily starve requests for other IO; this > is not as much an issue for SCSI since SCSI emasculates the request > queue in favor of letting the disk sort requests (which is NOT a pure win.)
ok, that's pretty much what i was expecting...
> several (io_32big, multcount, pio_mode and unmaskirq) are irrelevant > if using_dma is set, I think.
they don't seem to be used for normal requests. however, i noticed that ide_build_dmatable() can fall back to PIO if the DMA setup fails for a given request, so i figured it couldn't hurt.
thanks
-- ----------------------------------------------------------------------------- Ray Lehtiniemi (rayl@mail.com) (rayl@optitech.com)
- 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.tux.org/lkml/
| |