Paul M Dunphy wrote:
> >it'll work. my only concern is that i'm not certain that the linux
> >triton driver recognizes and supports DMA busmastering with the 440LX
> >chipset's integrated EIDE controller. 2.0.32 and 2.0.33 don't appear
> >to. it's a relatively easy patch to make it work, probably, and
> >there may be one out there. this is purely a performance issue,
> >though, and linux will work with this board.
>
> A good point. I don't expect that performance will be a big deal
> overall as this is a high-end system, but it's things like this that I
> wanted to get some opinions on. There is no point getting a fast
> system and then crippling it because of something like this. Will
> certainly look into the bus mastering . . . this is always a pain,
> even with 95/NT on certain boards.
[snip]
> And, I don't want to have to
> fight with hardware compatibility issues for the next month. I've
> spent too many years doing just this on DOS/WIN-311/95/NT systems
> and I'd like to concentrate on the software. I suspect that with
> the few minor changes you've all suggested, this system will be
> "plug and play" with Linux.
I just want to make sure you are absolutely clear on this point. Not only
will linux in the 2.0.33 and under kernel versions not use UDMA speeds, it
also won't use the CRC checking that goes along with UDMA modes. Therefore,
not only can you not set your system up to do bus mastering at UDMA speeds
(and possibly not at all depending on how different the timings are on this
chipset vs. earlier Intel PCI chipsets), but you must also make sure that
the drive itself does not negotiate with the BIOS for anything above PIO
Mode4.
--
Doug Ledford <[EMAIL PROTECTED]>
Opinions expressed are my own, but
they should be everybody's.
--
PLEASE read the Red Hat FAQ, Tips, Errata and the MAILING LIST ARCHIVES!
http://www.redhat.com/RedHat-FAQ /RedHat-Errata /RedHat-Tips /mailing-lists
To unsubscribe: mail [EMAIL PROTECTED] with
"unsubscribe" as the Subject.
