[asterisk-biz] Recent REMOTE CRASH BUG

[Remco Barendse]

>
> Oh I looked.  I had a thread going the other week! Kevin is assuring me that
> shouldn't be the case, and said in a 2950 they have in the lab this is not
> the case and that someone will contact me from Digium to find out what's
> going on... so I look forward to that to try to resolve these issues.
>
> I agree that the Dell SHOULD allow you to independantly put cards on an IRQ,
> however I will not accept what you say that the Sangoma will share as
> well.     I am looking in the BioS.   The Digium card went right onto the
> same IRQ as all the NICs.. the Sangoma card was allowed to sit on its own
> IRQ.     This is the BioS reporting it, not the OS..... the Sangoma was
> allowed to have its own IRQ.  All these threads you see are about DIGIUM
> cards in 2950s.. do you see any about Sangoma cards in 2950s?

AFAIK (and I'm pretty sure) ;) it's not the PCI that decides whether an 
interrupt will be shared or not but it is the mobo where stuff is 
hardwired to share an IRQ with another device. This is my huge problem 
with Dell machines, how on earth can you hardwire interrupts to such basic 
devices as the onboard 2 nics plus scsi controllers while leaving other 
IRQ's unused. If really needed they could have hardwired some devices to 
IRQ's you are most likely not to need in a rack server (like a second 
serial port or parallell port). This is just poor design.

> This says to me there is an issue with the way the Digium cards are
> reporting something to the BioS... but I'll leave this for Digium to discuss
> with me.. and if we find a resolution I'll bring it to the list.

Better read again, the Sangoma *must* share the same IRQ as a Digium card, 
it's just that Sangoma is advertised to handle irq sharing better. I never 
tried though, I don't like the extra hassle of the driver installation for 
a Sangoma card.