Text 46480, 88 rader
Skriven 2008-06-06 06:48:04 av James Bradley (1:342/77.0)
  Kommentar till text 46353 av Mark Lewis (1:3634/12.0)
Ärende: Matrox G4+(?) and co
============================
Pre-PS: I took a big axe to hidden "?" device drivers, (W2K) and it looks like
that fixed it. For giggles, I removed the two modem drivers, and reinstalled
them, and *Presto!* I think I beat it. I'll leave the rest of this message
intact, JIC.

On or about: 06-02-08  13:10, Mark Lewis did engage James Bradley regarding,
but not limited to: Matrox G4+(?) and co


 JB> IRQ 16: Rain or shine. It takes out 3, *and* 4, and - I didn't
 JB> exhaust testing - an internal at IRQ 7, but that one wasn't
 JB> even handshaking.
 
 ML> IRQ 16 ?!!? there's no such thing in physical hardware... there are
 ML> only 16 numbered from 0 (zero) to 15 and 9 and 2 are the same so that
 ML> gives us only 15 physical IRQs...
 
 JB> Two CPUs.

 ML> ahhh... that means that you have at least one of what's known as an
 ML> APIC,  Advanced Programable Interrupt Controller... it's job is to 
 ML> route interrupt requests to the assigned CPU for 
 ML> servicing... you also have a PCI buss based box which can 
 ML> share the hardware IRQ lines which could not be done on ISA 
 ML> buss based boxes... that means that we're in the realm of 
 ML> software managed and routed psuedo-IRQs... 

 ML> i think that you are likely running into an address or port 
 ML> range conflict... port range is likely... what com port is 
 ML> your modem on? is this a motherboard mounted com port with 
 ML> an external modem or an internal modem handling it's own 
 ML> com port settings?

I tried both mobo comm ports with the same outcome, with two different modems.
The HW handshake cable, should be a problem in both situations, (Dual head, or
not.) so I'm not considering it as a predicater. The ISA PNP modem (IRQ 7) was
unable to handshake, but did take voice calls and faxes IIRC. For giggles, I
turned on the second head in clone mode *while* I was connected with an
external modem, and it seemed to transfer files fine. I'm about to turn it on
again *before* connecting to see if I can exacerbate the problem again. I'm not
putting much faith in the internal modem, and it just *might* be the one
walking where it shouldn't, but why it takes its cue from the vid card... <?>

For my records:

Single head
F6000000 - F7FFFFFF
FCFFC000 - FCFFFFFF
FC000000 - FC7FFFFF
16
03B0 - 03BB
03C0 - 03DF
000A0000 - 000BFFFF

Clone mode shows the same. Should it be different? (IE: Did I miss something?)
I/Os look OK, no? The internal is at 03E8.

 ML> beating up on google, i was able to turn up some problems 
 ML> but nothing specificly related to the dual head use and the 
 ML> loss of a com port...

I expect, not many use modems any more.

 ML> one report was on a problem with arcsense and if dual head 
 ML> was required, the workaround was to try turning off the 3d 
 ML> Acceleration mode of the video card... i don't think this 
 ML> will affect your problem, though...

Something I didn't think of. I've come this far with it...

 ML> even rummaging thru matrox's site doesn't help and the 
 ML> available pdf user manual is woefully thin on the address 
 ML> and port ranges the card(s) use :?

I don't recall too much available from the site, but I have a second card I
need to setup, so I'll look if I can update the drivers for the two while I'm
there. (I use a community I-net connection.)

Well, thanks no less Mark! It's been an interesting quandary.



... James

___ Blue Wave/QWK v2.20

--- Maximus 3.01
 * Origin: -=-= Calgary Organization CDN (403) 242-3221 (1:342/77)