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

Linus Torvalds (torvalds@transmeta.com)
Tue, 28 Jul 1998 17:30:31 -0700 (PDT)


On Tue, 28 Jul 1998, Mark Lord wrote:
>
> The IDE driver already does the Most Sensible things by default,
> leaving the chipset and drive settings as set-up by the BIOS.

This I agree with. Don't mess with the DMA settings.

> If DMA fails, the IDE driver automatically falls back to PIO mode,
> without loss of data or corruption of any sort.

And this I obviously disagree with.

The driver should _start_up_ in PIO mode, and only even try DMA if there
is strong reason to think it will work. And the BIOS telling us it is set
up for DMA is obviously not strong enough reason, as shown by the fact
that apparently others have blacklisted certain WDC drives in their
driver.

Right now the IDE driver always seems to default to using DMA if DMA is
reported to be available, and THAT is what is wrong. Because your "without
loss of data or corruption of any sort" claim is obviously bogus. It only
falls back to PIO mode if the DMA failure is spectacular - it doesn't
gracefulle fall back to PIO mode for the silent and non-spectacular
failures.

Linus

-
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.altern.org/andrebalsa/doc/lkml-faq.html