[Asterisk-Users] more then two wildcards in one machine
Shoval Tomer
shoval at softov.co.il
Wed Dec 8 17:04:54 MST 2004
Thanks for clearing this up.
> -----Original Message-----
> From: Kristian Kielhofner [mailto:kris at krisk.org]
> Sent: Thursday, December 09, 2004 1:50 AM
> To: Asterisk Users Mailing List - Non-Commercial Discussion
> Subject: Re: [Asterisk-Users] more then two wildcards in one machine
>
> Matthew Boehm wrote:
> > The problem (from what I know) deals with IRQ interrupts. Each
digium
> card
> > must have a seperate IRQ and must be the only card on that IRQ. I
ran
> into a
> > problem plugging in just 1 digium card into each and every PCI slot
on
> our
> > Dell server and it was still sharing an IRQ with the ethernet card.
It
> wasnt
> > until I dissabled the usb, parallel port, serial port #2 and the
mouse
> that
> > it finally got its own irq.
> >
> > Mac's don't have IRQs. This is probably why it was so easy getting 4
> cards
> > into his mac.
> >
> > -Matthew
>
>
> Matthew (and everyone else),
>
> http://www.microsoft.com/whdc/system/sysperf/apic.mspx
>
> The problem is interrupts, not PCI slots. As this article
explains,
> sharing interrupts adds a signifigant amount of latency to the
> interrupt, causing audio droupouts and other weird stuff to happen.
It
> is nothing specific to Digium hardware or Linux or anything like that.
> On a desktop running Windows (or any other OS) with a lot of devices
try
> to play back an MP3 and generate network, disk, and other I/O activity
> on a machine without APIC - the audio WILL skip. Now imagine you are
> trying to drive 96+ T1 channels, ethernet boards and who know hom much
> else. You can see where there might be a problem.
>
> Apple Mac's have interrupts. They just have a lot of them. A
> modern
> Mac is very similar when compared to a modern x86 with an APIC. Read
> more.
>
> This is being mitigated on the x86 arhitecture by the APIC -
> Advanced
> Processor Interrupt Controller. You can have more than one APIC in a
> machine, and each APIC has 24 IRQs. Most new motherboards have at
least
> one APIC/processor.
>
> Dell PowerEdge 2850 (dual Xeon):
>
> ------------------------------------------------------------------
> ast1 etc # uname -a
> Linux ast1 2.6.9-pe2850a #2 SMP Fri Dec 3 09:23:03 CST 2004 i686
> Intel(R) Xeon(TM) CPU 3.00GHz GenuineIntel GNU/Linux
> ast1 etc # cat /proc/interrupts
> CPU0 CPU1
> 0: 6184515 72 IO-APIC-edge timer
> 1: 8 1 IO-APIC-edge i8042
> 9: 0 0 IO-APIC-level acpi
> 12: 65 1 IO-APIC-edge i8042
> 14: 11 2 IO-APIC-edge ide0
> 46: 19595 1 IO-APIC-level megaraid
> 64: 66366 1 IO-APIC-level eth0
> 65: 77045 1 IO-APIC-level eth1
> 101: 6113521 1 IO-APIC-level wctdm
> NMI: 1 0
> LOC: 6184694 6184698
> ERR: 0
> MIS: 0
> ast1 etc #
>
------------------------------------------------------------------------
-
> Apple Macintosh G4:
>
>
------------------------------------------------------------------------
--
> [root at localhost root]# uname -a
> Linux localhost.localdomain 2.6.8-1.ydl.7 #1 Wed Sep 22 18:46:40 EDT
> 2004 ppc ppc ppc GNU/Linux
> [root at localhost root]# cat /proc/interrupts
> CPU0
> 19: 4540 OpenPIC Level ide0
> 20: 12 OpenPIC Level ide1
> 25: 318 OpenPIC Level VIA-PMU
> 26: 0 OpenPIC Level keywest i2c
> 27: 35 OpenPIC Level ohci_hcd
> 28: 97 OpenPIC Level ohci_hcd
> 41: 1248 OpenPIC Level eth0
> 42: 0 OpenPIC Level keywest i2c
> 47: 0 OpenPIC Level GPIO1/ADB
> 54: 19036943 OpenPIC Level wctdm
> 55: 0 OpenPIC Edge NMI - XMON
> 63: 3 OpenPIC Level ohci1394
> BAD: 0
> [root at localhost root]#
>
------------------------------------------------------------------------
--
> -----
>
> As you can see, on the Dell, the wctdm board has IRQ 101. Waaaaayyyyy
> past good ol' 15. On the Mac, it is at 54. Still way beyond the
> standard x86 15 limit.
>
> Now, if you are using an OS that can't do APIC on x86, or have a
> motherboard that has a mal-functioning APIC, you are out of luck and
> have to try to get them all to play nice. I hope that this clears
> things up some.
>
> --
> Kristian Kielhofner
> _______________________________________________
> Asterisk-Users mailing list
> Asterisk-Users at lists.digium.com
> http://lists.digium.com/mailman/listinfo/asterisk-users
> To UNSUBSCRIBE or update options visit:
> http://lists.digium.com/mailman/listinfo/asterisk-users
>
> --
> This message has been scanned for viruses and
> dangerous content by MailScanner, and is
> believed to be clean.
> MailScanner thanks transtec Computers for their support.
More information about the asterisk-users
mailing list