> 1. The latency has gone completely to shit. Its so slow handling IRQ's
that I can no longer run 64Kbit synchronous on 2.1.x kernels. People
with sync cards are trapped in 2.0 or blow $500 on a smart card.
This we have to sort - even NT can do 64K sync cards.
Alarm bells! Fast turnaround for interrupts is fundamentally important to
me. I write Linux drivers for my company's image processing boards, and
in some cases the performance of certain communications is directly
related to interrupt latency.
Linux is chosen over NT by at least one of our customers exactly because
of this sort of thing. (document transport spitting image as a receiver
board, which interrupts the host when the little image finishes so that
the host can do OCR etc.)
Frankly, everyone is better off if interrupts are handled fast and
efficiently.
alan@lxorguk.ukuu.org.uk said:
> 2.0.x is OK with the sync card even when its being at its most
> obnoxious (square waving the irq line).
alan@lxorguk.ukuu.org.uk said:
> Anyway Im now off to look at putting the notion of fast interrupt and
> probably syscall returns back into 2.1.x.
Now that I know that trouble might be brewing, I will attach a scope to
so interrupt lines and get some measurements from my devices. I would like
the interrupt latency to be *even better* then 2.0. It is supposed to be
an improved kernel, no?-)
-- Steve Williams "The woods are lovely, dark and deep. steve@icarus.com But I have promises to keep, steve@picturel.com and lines to code before I sleep, http://www.picturel.com And lines to code before I sleep."
- 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