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

Mark Lord (
Wed, 29 Jul 1998 06:49:17 -0400

Linus Torvalds wrote:
> If you by "all such cases thus far have been proven bogus" mean that so
> far you have been able to blame a too-long cable or something else like
> that, then sure. I can imagine that all DMA corruption problems can be
> adequately explained by something being out-of-spec, whether it's the
> cable or the chipset being bad.
> But that doesn't change the fact that people see corruption that goes away
> when they disable DMA.

That "fact" has never been established.
Repeat, that "fact" has never been established.

The very few cases I have ever seen reported of "corruption that goes
away" when DMA is turned off, have thus far been totally bogus.

I get a heck of a lot more email on this stuff than anyone else,
your godness included. The pattern is always thus:

1. Somebody complains of "disk corruption".
2. They poke around with "hdparm" and claim that "hdparm -pickanoption"
"fixes" the "corruption".
3. They post this to the world, copying linux-kernel and LinusTorvalds.
4. I email them privately asking a few simple questions.
5. They email me back, sheepishly confessing that:
5.1. They were mistaken -- corruption was *still there* even after
playing with the "hdparm" settings.
5.2 They replaced some SIMMs and the problem went away.
5.3 They were overclocking their CPU, and IDE wasn't the only
thing they admit was broken.
6. Since the final exchanges were in private email (reducing noise
linux-kernel), Linus never sees the resolution, and thus assumes
that Gadi and I are shitty coders.

I have yet to see a case where turning off IDE DMA
actually cures a disk corruption problem.

I *do* see lots of reports claiming this,
and I *do* actually investigate them (rather than
reacting as if they were the gospel truth).

Or rather, I *did* do that once.

The (ex-)Linux IDE guy

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