Re: 2.1.111: IDE DMA disabled...BLAH...BLAH...

Hubert Mantel (
Thu, 30 Jul 1998 11:25:19 +0200 (MEST)


On Wed, 29 Jul 1998, Alan Cox wrote:

> > Damn it Alan, and Mark, how much more proof do you need that you're WRONG.
> The facts don't fit your claim. I've just spent 3 days studying reports
> and chasing after people.
> Two people has 2.0 DMA problems, many people have 2.1 DMA problems.

_Several_ people have 2.0 DMA problems. Not many, but several. Definitely
more than only two. I would guess about 0.05% of all machines. It's the
maximum we can tolerate.

Check your helpdesk at RedHat and you should find several cases, too ;-)
And there are several cases of unexplicable errors that _could_ be DMA



I cannot confirm this. The problem definitely exists in 2.0, too. But it's

> I have only _one_ person who has a drive that fails in both Linux and
> Windows of all the people who are using UDMA/mode 2 DMA on _BOTH_ systems
> I have one person whose drive does UDMA faultlessly in FreeBSD and Linux
> 2.0 but not Linux 2.1

Which means that there is a problem with 2.0 that has become worse with

We had this same discussion here at S.u.S.E. some time ago (only talking
2.0!), but we decided that the number of problem machines is too low to
change the default. If with 2.1 the number of machines that show this
problem is about 1% we _need_ to change the default. 1% means _many_

We'll probably disable IDE DMA in the future and let people activate it
from the installer if they want. It's trivial to add the hdparm commands
to /sbin/init.d/boot.

> Alan
> Very close to being the ex Linux Sound/video/network/etc person

You should make hot line support on Red Hat for two or three days. This
could convince you that sane defaults are a very good thing ;-)

I hope Linus does what he thinks is right.
Hubert /\\

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