[Asterisk-Users] PCI front mount chassis?
asterisk at linuxguy.com
Fri Mar 12 06:41:36 MST 2004
On Fri, Mar 12, 2004 at 05:26:13AM -0600, Rich Adamson said:
> > > I too am running 6 cards in my system, although not in a "high traffic
> > > capacity" load environment.
> > >
> > > So far my (limited) high-load simulations have shown no problems.
> > So - is it apocryphal that the Digium cards (drivers) won't share
> > interrupts?
> > If there is a real issue with sharing interrupts then it seems to me
> > to be a bug that needs fixing. PCI bus supports shared interrupts,
> > why doesn't the hardware/driver?
> In most cases, sharing an interrupt is not a problem at all. There have
> been a few cases where _some_ issue was resolved by moving cards around,
> however the majority of those seem to be: a) abrupt system changes with
> no effort to seriously identify the root-cause, b) newbie installations
> where the condition of the underlying system infrastructure is totally
> unknown, or, c) wild recommendations that might have had some basis a
> long time ago but no longer apply.
> Example: 'cat /proc/interrupts'
> 9: 1854652239 XT-PIC ehci-hcd, eth0, wcfxo, Intel ICH4
> works just fine, and I can't imagine a more demanding irq arrangement
> where the only nic shares with an x100p, etc.
> Obviously there are performance limits and expecting multiple quad T1
> cards or some other _specific_ high-volume configuration to share one
> or two interrupts could create a problem. But, engineering a system for
> those conditions is no more difficult then understanding the
> requirements of whatever cards are being used and dealing with them
Looking at my /proc/interrupts:
0: 59709041 XT-PIC timer
5: 597050409 XT-PIC wcfxo
7: 597211339 XT-PIC wcfxo
10: 4538876 XT-PIC eth0
11: 3044608 XT-PIC aic7xxx, eth1
The voice cards generate an order of magnitude more interrupts than
anything else. This "may" be why it's not recommended to share
interrupts on voice cards. Don't know if the T1 cards have a similar
issue. I would hope not. The x100p's are a pretty simplistic device.
They probably generate an interrupt for every byte. The x100p's are also
used for timing in things like MOH and MM conferences AFAIK. It seems
like it would be nice to only put one card in "timer mode" if that is
indeed what is generating all those interrupts. Could someone "in the
know" enlighten us?
More information about the asterisk-users