Re: MCP2 on Intel 945 Mobo advice?
- From: "David T. Johnson" <djohnson@xxxxxxxxxxxx>
- Date: Thu, 06 Apr 2006 07:58:26 -0700
Mike Luther wrote:
David .. would you be kind enough to think a bit more on this by chance?
David T. Johnson wrote:
Where I posted:
Both those on board PCI slots are FORCED to share the same IRQ 11
which is shared by a whole lot of other stuff too!
And you replied:
Maybe try a different audio card would be one suggestion. Probably, though, what you want to do is get the onboard sound working and that will help. Sometimes, also, there is a BIOS option for restricting a particular IRQ. You might try setting that and see if you can get a solitary IRQ assigned to the SBLive so that the other devices will go off and find a different IRQ.
I've done more research here. From other OS/2 work before, with the whole thing 'working', but with all the IRQ's the way 'it works', I thought that I would try the <F6> option during the boot run to let it examine hardware. That totally jammed the whole 945GTP board during the boot run. During the boot run, with ALTF2ON.$$$ in the boot partition root, I couldn't see what the lock point was. I'd already made a set of utility floppy diskettes for this box, and they wouldn't boot at all.
So on a hunch, I thought maybe the PCI snoop and driver work was failing. So, I added ALTF2ON.$$$ to the floppy utility diskette set for the box. And PRESTO, I get the boot run jam with PCIBUS.SNP frozen on that bottom line! For research purposes, I then went into the floppy utility diskette set and remarked out the last snoop line in the SNOOP.LST on it. Which is that PCIBUS.SNP instruction. PRESTO! The whole floppy utility diskette disk set for the Intel 945GTP box research project finished booting to an OS/2 command line!
I've never seen this before. So I tried Veit Kannegieser's PCIIRQ.ARJ 'force' IRQ suite off his web site both to cross research this with the PCI sniff tests as well as possibly 'fix' the IRQ for at least the SBLive 5.1 PCI add-in card. With the use of the SHOW_PIR and SHOW_LNK executables in that suite, the data confirms what the PCI sniff shows. The add-in SBLive 5.1 card is on BUS 7 as DEVICE 1 and is, as booted 'normally', assigned IRQ 11 as a resource.
That SBLive card sounds like it just absolutely will not now way no how nevuh share an IRQ with anything else, at least when running with the OS/2 driver. Probably in Windows it shares through some sort of hardware accomodation. I have run into the same thing with other add-on sound cards. Your problem is that it sounds like the 945GTP motherboard will *always* want to use IRQ11 for both of the PCI slots and probably is also using IRQ11 for the video card and/or something else. This is a problem in OS/2 since as soon as something else tries to use IRQ11, everything freezes. On my system, this isn't a problem because the onboard sound is happy to share the IRQ with other stuff. That's why you should try to get the onboard sound working and ditch the SBLive card if you want to use the 945GTP motherboard. Unfortunately, the uniaud driver doesn't look it supports the audio onboard. One other alternative might be to try the old sblive driver which purportedly will share irqs. Look for the file 'sbliveos2-081b.zip' on Hobbes.
If that doesn't work, your only alternative might be to use one or zero pci cards if even the one card will not share with other resources trying to use the irq11.
I then did move Veit's drivers, snoop list and so on to the directory. I added, as he suggests, the snoop list entry for that to the top of the SNOOP.LST in the 845GTP's boot directory file. With just that much done, the Intel 945GTP would still boot. I tried enabling <F6> hardware detection. BLAM .. locked box! No way to recover it but disable it from an <ALT F1> then <F6> disable hardware detection startup.
Discouraged, I thought I'd at least try to add the two variants of the configuration of his driver to CONFIG.SYS:
DEVICE=C:\OS2\BOOT\SPCIIRQ.SYS LINK $6A 10 (or 5 whatever)
or
DEVICE=C:\OS2\BOOT\SPCIIRQ 7 1 0 1 10 (or 5 whatever)
With the on-board SNOOP.LST for the full box run that has his utility at the top of it, and any time either of the above loader lines are in the CONFIG.SYS file, I get a complete HARD LOCK for the boot run, right there at that line in the full box ALTF2ON.$$$ shows the above SPCIIRQ driver runs.
I haven't tried to 'force' the working Kingston KTC-120 NIC card in the other PCI slot to another IRQ, without the SBlive 5.1 card inserted yet. And, remember here that shuffling the two cards in the two available PCI slots on this motherboard does not produce any difference in how things work.
Do you have any suggestions for me? Or anyone else here have any thoughts that can help me or find what I am missing? For example, why, in the case of the IBM floppy utility diskette boot run, does it lock on the PCIBUS.SNP last line there on them, even though hardware detection isn't chosen during that floppy run. Or, am I just not correctly thinking about that for a floppy diskette boot run?
If you have an IRQ conflict, the system locks as soon as a second device uses an irq that is supposed to be exclusive so maybe the pci snoop triggers that.
--
Posted with OS/2 Warp 4.52
and IBM Web Browser v2.0.5
.
- Follow-Ups:
- Re: MCP2 on Intel 945 Mobo advice? (Another idea...)
- From: David T. Johnson
- Re: MCP2 on Intel 945 Mobo advice? (Another idea...)
- References:
- Re: MCP2 on Intel 945 Mobo advice?
- From: Mike Luther
- Re: MCP2 on Intel 945 Mobo advice?
- Prev by Date: Re: MCP2 on Intel 945 Mobo advice?
- Next by Date: Re: MCP2 on Intel 945 Mobo advice?
- Previous by thread: Re: MCP2 on Intel 945 Mobo advice?
- Next by thread: Re: MCP2 on Intel 945 Mobo advice? (Another idea...)
- Index(es):
Relevant Pages
|
