Re: Test Patch: 2.5.69 Interrupt Latency

Alan Stern (stern@rowland.harvard.edu)
Fri, 16 May 2003 13:20:52 -0400 (EDT)


On 16 May 2003, Paul Fulghum wrote:

> Moving the wait out of the ISR and doing the wakeup
> only for RD on non-OC ports are winners.
>
> I can't comment on the 1 second grace period. Was that
> in response to this investigation, or have you actually
> seen false RD indications due to noise?

Well, I have actually seen false indications. Whether they are due to
noise is open to debate. Since they occur just at the time when I turn
the power to my USB peripheral on or off, that's my best guess. It might
even turn out that power on/off generates a temporary OC condition, so
fixing that might render the grace period unnecessary. I haven't had a
chance try it yet.

> There is also the more trivial matter of removing the
> unnecessary setting of the FGR bit on wakeup.

Yes. That can be done in any case.

> I'll check that the global RD interrupt does not
> keep repeating after a false RD by an OC port.

Good.

> So I suggest you build a patch that does all of
> the above (with the grace period at your discretion).
> Then we can both test it, and you can submit it
> for actual inclusion.

I will. Probably won't be ready until some time next week.

Alan Stern

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/