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

Andre M. Hedrick (hedrick@Astro.Dyer.Vanderbilt.Edu)
Wed, 29 Jul 1998 09:17:39 -0500 (CDT)

On Tue, 28 Jul 1998, Mark Lord wrote:

> > As for the (U)DMA question, would it not be sensible for us to flag
> > the drive capablities.
> No, it would not be.
> We cannot blindly (and dangerously) reprogram the DMA mode
> of the drive without knowing what mode the IDE interface
> chipset is trying to use. The two must match.


> If the chipset is BIOS-configured for UDMA,
> then we could break things big-time by selecting DMA at the drive..


> The IDE driver already does the Most Sensible things by default,
> leaving the chipset and drive settings as set-up by the BIOS.

However, consider the case of a BIOS-less chipset on an offboard
card that must be programmed. The chipset is native to UDMA
and is backwards compatable. In all cases the pci-config data
must me set for any mode regardless of the drive's capablities.
Without take this step, an UDMA would register in PIO 0 strength.

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

This is a good thing, and it works with the case stated above.
Since I am having trouble getting the method in question functional,
it is nice that the breakout of a failed (U)DMA results in a clean
transition back to PIO status.

I am not asking for high risk programming and data coruption, I have to
much to lose. This single chipset may be a special case that does need
to relearn what it can do each time it is booted.

Andre Hedrick

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