Re: 2.1.111: IDE DMA disabled?

Nathan Hand (
Mon, 27 Jul 1998 21:47:38 +1000 (EST)

On Sun, 26 Jul 1998, Dave Wreski wrote:

> > For people that don't know about these corruption issues, the _only_ sane
> > default is to leave the DMA code turned off. Unless you have some way of
> > categorically making sure that it is absolutely safe to turn on, and so
> > far I haven't heard of any (a white-list is probably the best answer, but
> > I'd argue for keeping the white-list in user space where it belongs, and
> > just have a script in rc.d to parse the BIOS/PCI information).
> - Is using hdparm something the distribution developers will be able to
> determine upon doing an installation?

Yes. There already exists a RedHat contrib package of hdparm which
places boot scripts in /etc/rc.d with user configurable data via a
GUI control panel in /etc/sysconfig.

> - What do other operating systems do? Surely idiot systems like 95 can't
> rely on user-input for such a decision?

Windows 95 gets drivers from the manufacturers. I've never seen it
enable DMA by default. Windows NT has a control panel (which is in
an obscure directory on the CD) which enables it.

Just to illustrate how NT treats this "feature" of IDE, this panel
has warning signs all over it, and when you try and enable DMA the
panel first warns you "this might cause severe disk corruption".

> - Why is it only IDE DMA, and not the bus-mastering DMA on my SCSI card?

Driver dependant.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
Please read the FAQ at