Re: IDE drives and unmaskirq

From: Andrew Morton (
Date: Mon Apr 24 2000 - 08:20:40 EST

Benno Senoner wrote:
> Andre Hedrick wrote:
> >On Sat, 22 Apr 2000, Khimenko Victor wrote:
> >
> >> Unfortunatelly it's NOT legend :-(( I've seen systems (less then 4 years old
> >> BTW) where enabled unmaskirq will corrupt filesystem badly. That's why it's
> >> in zero state by default: it's MUCH better to "just" slow down system then
> >> to trash filesystem on that system.
> >
> >Nicely clarified! ;-)
> >
> >He speaks the true, I was just waiting to see when someone else answer it
> >correctly.
> I agree that it's good to leave it off by default but for realtime stuff
> (alias low latency applications), it's almost mandatory to turn ON the IRQ
> unmasking, since long stalls due to the disk I/O means
> no realtime audio and video or whatever.

I have done some measurements on this. The worst case interrupt
blockage within the IDE code when using UDMA with DMA is 90 microseconds
(500MHz CPU). When interrupts are unmasked this comes down to 32 usecs
worst case. Compare this with the console code's 3,000 microseconds.

More details at

If you're concerned about latency you must knock off the peaks. With
DMA, IDE is down in the noise floor. And that unmasking code is pretty
scary stuff....

> PS: Andre: funny your E-Mail address :-)

Not funny if you're stuck using dialup :-(


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

This archive was generated by hypermail 2b29 : Sun Apr 30 2000 - 21:00:07 EST