At 2003-02-22 10:03 -0800, Declan. Moriarty wrote:
>On another mailing list I inhabit, someone is trying to set up a new
>motherboard. It has facilities for 14 peripheral interrupts beside the
>usual ones.
>
>The interrupt table (from a ix86 pc) is _weird_!
>
>(from the Bios)
>
>IRQ Assignments for this motherboard
>(S=shared U=used) [not present components]
>===================================================
>                             A  B  C  D  E  F  G  H
>--------------------------- -- -- -- -- -- -- -- --
>pci slot 1                   -  -  -  -  -  S  -  -
>pci slot 2                   -  -  -  -  - -  S  -
>pci slot 3                   -  -  -  -  -  -  -  S
>pci slot 4                   -  -  -  -  S  -  -  -
>pci slot 5                   -  -  -  -  -  S  -  -
>pci slot 6                   -  -  S  -  -  -  -  -
>AGP slot                     S  -  -  -  -  -  -  -
>onboard USB controller HC0   S  -  -  -  -  -  -  -
>onboard USB controller HC1   -  -  -  U  -  -  -  -
>onboard USB controller HC2   -  -  S  -  -  -  -  -
>onboard USB 2.0 controller   -  -  -  -  -  -  -  S
>onboard LAN                  -  -  -  -  S  -  -  -
>[onboard RAID                -  -  -  -  -  -  -  S]
>[onboard 1394                -  -  -  -  -  -  S  -]
>----------------------------------------------------
>
>Now whan you include rtc, serial & parallel, etc they come up to way
>over 16 irqs. There's an onboard soundcard in there onboard as well. 
>What's the story? 
>Anyone got a good explanation or url for this technology? Conventional
>wisdom would be that irq sharing is the surest way to a crash in a pc.

I suspect that the PCI-bus implements all the interrupts
for all the PCI-slots separately (and why not, silicon is
cheap nowadays). You can see it as every slot having
it's own set of two 8259's. But I'm not sure about this.
It would make sense though, because assigning non-interfering
IRQ's used to be a big problem.

>The owner is a software expert setting up linux on the thing, and with a very 
>recent (alpha or beta) kernel, (to handle the chipset)and the problems are:
>
>1. Conventional devices keep SHARING and being assigned half-assed or
>weird irq numbers by the bios.Is there any utility for farting around with
>these irq assignmentsarting around with these irq assignments?

Can't you change it in the BIOS setup? By the way, the combined
system of BIOS and OS is very 'intelligent' and you may need to
reset the settings quite violently (by going back to the BIOS
defaults or removing the battery or shorting it) so the system
will try out other settings when you are trying to make it do so.
Check out the Microsoft and other sites with Google.

>2. She can't get the concert playing in tune. I have a detailed 
>e-mail (down to the nitty-gritty of C code) which is long, but I'll gladly 
>post it if people want to see it, or send it to those who know the area
>off list. She is trying to figure the hardware from kernel code, surely
>the most painful way to do it!! She has built a 'distro' type kernel,
>with all the modules, so she can fool around with it more once it's
>running
>
>3. It's assigned the hd to irq 4, then (different kernel) to irq 9. 
>Can 9 be safely used?  I've got away with it, but Somebody thinks 9 
>should not be used, as 2 is/was masked over to it to give 16 interrupts.
>He's muttering about the PC/XT 8259A Apic - a little before my time with
>pcs. Is that still an issue?

In the old days, you couldn't use IRQ2, because the second 8259
was linked to it on an AT-board. The PC and PC/XT only had one
8259, so it didn't have to link the second 8259 to IRQ2 of
the first one. There was also a problem with IRQ9 because the
coprocessor (or division by zero) was linked to it, or something
like that.

By the way, IDE harddisk buses usually used IRQ 14 and 15
because IRQ 3 and 4 are traditionally used by the COM ports.

>It is probable she would install windows temporarily if that
>was the only fix and she could wipe it again straight away.

Perhaps she should buy a new motherboard? I bought a
Gigabyte motherboard december 2001 and I have no problems
running Windows98 and Linux in the default BIOS set-up.

It may also help not to use any ISA-boards and if need be,
replace them with PCI-boards.

Greetings,
Jaap

-- 
Author: Jaap van Ganswijk
  INET: [EMAIL PROTECTED]

Fat City Network Services    -- 858-538-5051 http://www.fatcity.com
San Diego, California        -- Mailing list and web hosting services
---------------------------------------------------------------------
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB CHIPDIR-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).

Reply via email to