Thom May thom at ubuntu.com
Thu May 12 21:45:05 UTC 2005

* Stephen R Laniel (steve at laniels.org) wrote :
> On Fri, May 13, 2005 at 02:28:17AM +0800, Trent Lloyd wrote:
> > Because on lots of hardware it breaks and causes all sorts of issues.
> So there's another place where I have to ask: why can't the
> boot process just be smart and figure out whether the given
> CD-ROM drive that it's working with can handle DMA? Maybe
> there'd be a certification list, and if you're on the list
> you get DMA turned on. Make the cert. list really rigorous,
> if you want to eliminate false positives.
> What's wrong with this approach? Why doesn't Linux use it?
Because we don't have the data to do it; hopefully the hwdb stuff that we're
gradually introducing and integrating will allow us to do stuff like this;
it's certainly one of the aims.
"Please insert a cd. I'm now turning on DMA, and ripping a track. Did this

> A lot of people on here have noted that their CD ripping
> speeds are much faster under Windows. And yet Windows users
> don't need to worry about turning DMA off or on; at least in
> this case, it seems to Just Work for them. So it's more
> convenient and faster. Seems like Linux should strive to
> emulate that.
No-one disagrees with you on this. To some extent, we should be able to do
better out of the box, hwdb or no. But there are a lot of things that we
need to do, and performance optimisations like this tend to fall by the
wayside in the face of real bugs.

More information about the ubuntu-users mailing list