Half a croissant, on a plate, with a sign in front of it saying '50c'
h a l f b a k e r y
The leaning tower of Piezo

idea: add, search, annotate, link, view, overview, recent, by name, random

meta: news, help, about, links, report a problem

account: browse anonymously, or get an account and write.

user:
pass:
register,


                                                     

Who Wants IRQ 6?

Currently inalterable, but should be alterable
  (+6, -4)
(+6, -4)
  [vote for,
against]

One of the legacy components of the modern personal computer is the Floppy Disk Drive. Recently this particular piece of hardware has become optional. Some computers may have LS120 or ZIP or equivalent drives installed instead of traditional floppy drives. In one sense this can be perfectly OK (LS120 drives are backwards compatible with ordinary floppy drives, for example), but in another sense, today's computers should have one additional feature. That single feature will cause future floppy drives to be left on dust heaps everywhere, by customer request. See, one thing that many modern computers have these days is a shortage if IRQ lines. However, one of the IRQs is number 6, which apparently is hardwired directly to the Floppy Disk Controller. Now, should it happen that a computer be built with an LS120 drive instead of an ordinary floppy, the BIOS/CMOS setting for the Floppy Disk Controller might be set to Disabled. This was a recommended setting at one time, in order to allow the LS120 to become bootable, in the tradition of the floppy it replaced. Even if that isn't the case these days, the point is that IF a system's onboard Floppy Controller is Disabled, it means that IRQ 6 should be made available to some other piece of hardware, exactly like IRQ 12 becomes available if one disables PS/2 Mouse Support. The fact that IRQ 6 does not become available means there is an Opportunity For Improvement. Hardware, BIOS/CMOS, and of course Operating Systems all need modification just to ALLOW IRQ 6 to be used for things other than ancient floppies. Is the goal worth it? Well, how many IRQs have YOU been short?
Vernon, May 24 2002

Happy http://www.uclinux....m/archive/0437.html
hacking [thumbwax, May 25 2002, last modified Oct 21 2004]

MSI-63xx series FAQs http://www.msi.com....hp?classif=63Series
[phoenix]

MSI Technical Guide http://www.msi.com....tech_guide/no.1.htm [phoenix, May 25 2002, last modified Oct 21 2004]

MSI Technical Guide http://www.msi.com....tech_guide/no.1.htm
[phoenix, May 25 2002, last modified Oct 04 2004]

MSI chipset/CPU drivers http://www.msi.com....river/mb_driver.htm
[phoenix, May 25 2002, last modified Oct 21 2004]

MSI-6330 compatibility list http://www.msi.com....rd/6330v50/main.htm
[phoenix, May 25 2002, last modified Oct 21 2004]

MSI-6330 manual http://www.msi.com....manual/socket_a.php
in PDF format. [phoenix, May 25 2002, last modified Oct 21 2004]

MSI-6330 BIOS updates http://www.msi.com....?model=MS-6330+V5.0
[phoenix, May 25 2002, last modified Oct 04 2004]

MicroStar newsgroups http://groups.googl...board.msi-microstar
[phoenix, May 25 2002, last modified Oct 04 2004]

Progress! http://support.microsoft.com/kb/275180
It's been a while since anyone posted something on this page. Lots of computers don't have floppy drives these days; it looks like maybe they finally wised up (only took 'em how many years since this Idea was posted?) [Vernon, Aug 20 2008]

[link]






       I was under the impression that most hardware these days no longer need IRQs (or can share them)...?
yamahito, May 24 2002
  

       I'm with yama. I thought in these glorious days of plug and play (not to mention USB) we didn't have to worry about this any more; but I'm not enough of an expert to be certain.
pottedstu, May 24 2002
  

       I've passed it on to an expert for his opinion, so we'll know for certain in a bit. Unless someone wants to assert themselves sooner?
yamahito, May 24 2002
  

       Never having encountered a shortage of interrupts, I don't see how freeing one up is any improvement.
waugsqueke, May 24 2002
  

       IRQs are fading away. Modern ACPI and PnP motherboards with modern OSes don't really need 'em for many devices. Win2K has no problems talking to my sound card, PCI/USB controller and NIC which all share IRQ 11. And you can have 128 USB devices all sharing IRQ 12.
phoenix, May 24 2002
  

       My mom bought a Comcrap with an LS-120 drive, despite my warnings. The drive died in 2 years, the hard drive needed to be replaced twice in 3.
RayfordSteele, May 24 2002
  

       Folks, PCs even with Plug-and-Play still use IRQs. They merely get automatically assigned by the computer, instead of being manually assigned by a technician. With enough different kinds of hardware in a PC, you can still run into device conflicts that ultimately can be traced to a shortage of IRQs. So the idea here is simply to identify one that should be as available for oddball hardware as are IRQs 9, 10, 11 and 12. IRQ 6 could be available as soon as you plugged in, say, a USB floppy drive. (I have a friend who had a USB scanner and a USB mouse on the same channel, and the mouse went dead whenever a scan was made. So while you can save IRQs in that fashion, you have to watch out for bandwidth conflicts.) I was "set off" to post this idea when I discovered that the brand-new motherboard I just bought, which has six PCI slots, is hardwired/limited to only 4 IRQs for those slots. This means I have to carefully select PCI card types that can coexist using the same IRQ. Among the cards that I have and/or want to have:   

       SCSI controller card, MPEG DVD decoder card, hardware modem card (NOT a WinModem), high-quality sound card, multiple Ethernet cards, TV capture card, high-end video card, and even an extra printer-port card. Every one of those cards needs an IRQ; only some are compatible with others.   

       Currently all these cards are scattered among 3 different computers; if I wanted to combine them (can't do 'em all because not enough slots available, so think in terms of "as many as will fit"), there is going to be a big battle to make them coexist. Having an extra randomly-assignable IRQ would definitely help. And yes, I know that some things are available together; just today I saw a combo modem/ethernet card, and some video cards have TV capture ability (but not the very-highest-end of video cards).
Vernon, May 24 2002
  

       I think the problems you describe are more due to bandwidth through the IRQs, but an extra one would help, I suppose. Well, you earn the bun.   

       May I ask why multiple ethernet cards in one machine? Not the main one, I hope?
yamahito, May 24 2002
  

       "Folks, PCs even with Plug-and-Play still use IRQs. They merely get automatically assigned by the computer, instead of being manually assigned by a technician."
Not entirely true. As I said previously: with a good ACPI mobo and a modern OS, devices can peacefully and painlessly share IRQs. I'm at home now and my Win2K machine here has my SoundBlaster Live!, IBM NIC, Intel NIC and USB controller sharing IRQ 9. This machine is a router/NAT for my home network so I frequently use both NICs simultaneously.
  

       "...I just bought, which has six PCI slots, is hardwired/limited to only 4 IRQs..." (and then you go on to list *9* cards!)
Unless two of those slots are CNR/AMR, I find that hard to believe. Would you care to post the manufacturer and model? What OS do you want to / are you running? Perhaps we could find a work-around....
  

       "Every one of those cards needs an IRQ; only some are compatible with others...there is going to be a big battle to make them coexist."
Welcome to the 1980s!
  

       The reason we still use IRQs (and DMAs and I/O Adxs and Memory Adxs - the other major hardware resources) is for backward compatibility. If that compatibility were broken such that you could reassign them, why not break compatibility completely and get rid of them?
phoenix, May 24 2002
  

       Should I care about the mating habits of dinosaurs?
bristolz, May 25 2002
  

       yamahito, if I wanted to hook up a computer to two different networks at the same time, then I need two network cards.   

       phoenix, yes, I did say that some of that list of cards could coexist on a single IRQ. Part of the problem is that the documentation that accompanies the average card doesn't say how much coexistability it has. I also did say something to the effect that **IF** I wanted to combine nine cards... (but thanks for also answering yamahito's question). The mobo that I described is made by MSI (MicroStar International), and is called "K7T-Turbo2". It is not a latest-and-greatest mobo, but its VIA KT133A chipset has only been out for a year or so. I should mention that I think I actually encountered one mobo some time back that in the BIOS/CMOS setup SEEMED to let me assign IRQ6 to something other than a floppy, but Windows didn't agree. So I don't know if the Setup was lying, or if Windows was just being its usual obnoxious self. And it will not break compatibility to allow a greater usage of IRQ6 -- it should continue to be the default IRQ for floppies.   

       One thing that bothers me is that in the Original PC there was this "8259" chip that offered 8 IRQ lines, and ever since the first AT there have been (the equivalent of) two such chips, connected in a way that offers 15 IRQ lines. I wonder why they never bothered to go to the equivalent of three "8259" chips in this age of Very Large Scale Integration? I'm reasonably sure the answer to that is either "nobody could agree on a standard" or "backwards compatibility" (even though the AT was BC with the PC), and so I alternately can wonder why they never bothered to replace the second "8259" (or both of 'em) with a "super" version that all by itself could accommodate more than 8 IRQs. Such a thing would be quite compatible, and the first one to do it would by default set a standard! (especially if they overkilled the need by letting that "super 8259" handle 32 IRQs or so). --And yes, I am aware that probably the CPU would have to be updated as well. So? Don't they do that regularly anyway?   

       UnaBubba, I assure you that even with Plug-and-Play, I have encountered far too many resource conflicts (mostly IRQ related) for me to be happy with the hype. Having a more friendly IRQ 6 would help; having the problem fixed forever, at the source, would be even better.
Vernon, May 25 2002
  

       Technically, it is available for use by other devices, and some devices will allow you to select IRQ6. Most however do not, realizing that virtually every PC uses at least one floppy disk drive. The most common devices that will let you use IRQ6 are probably tape drive accelerator cards. This is probably because these cards are used for tape drives that run off the floppy interface, and many of them can be set to drive floppy disks themselves. Conflicts on IRQ6 are uncommon and are usually the result of an incorrectly configured peripheral card, since IRQ6 is pretty standardized in its use for the floppy disks. If you use a tape accelerator card along with an integrated floppy disk controller on your motherboard, watch out for the accelerator trying to take over IRQ6; some even do this by default. Windows 2000: If you have a non-ACPI system that uses a Plug and Play BIOS, Ntdetect.com always reports a floppy controller on an x86 system, unless your BIOS is legacy-free. To resolve the issue, update your BIOS to ACPI.
Thus sayeth the lord, Amen.
thumbwax, May 25 2002
  

       <aside>Yes, vernon, I know, but the best way to do that isn't at the main machine: use one of your others as an ethernet bridge</aside>   

       I have to speak up in Vernon's favour - the advantages of this would not really be significant, but they would be there (or so it seems to me). I certainly can't see any backwards compatibility problems. I'm also pretty sure that it's rare for a motherboard to assign more than 3 or 4 IRQs to the pci (/isa if you're retro enough to have 'em) array. Normally, one of the pci slots shares an interrupt with the agp slot too. For joe bloggs, this shouldn't be significant, but several power tweakers I know seem to think that it makes a performance difference, and choose their co-existing cards very very carefully.
yamahito, May 25 2002
  

       "Part of the problem is that the documentation that accompanies the average card doesn't say how much coexistability it has."
Doesn't matter as much as you might think. PnP is a three-way process which involves negotiation between the BIOS, expansion card and the OS (the name of which you neglected to mention).
  

       Anyway, I scanned the manual for the motherboard and you're right in that it spreads 4 IRQs across the 6 (1 AGP, 5 PCI) expansion slots. The soundcard is built in to the motherboard and the manufacturer expects you to use the CNR slot for a NIC, so the average consumer would use the remaining slots for video and modem and not much else. Note that the CNR and built-in sound card are taking IRQs which would otherwise be assigned to the PCI slots.   

       I'll post links (which I suspect you already have) to relevant updates and information at MicroStar's web site and suggest you download and install the latest drivers and updates. Beyond that, I can only suggest that you use a PnP OS and use more caution when purchasing hardware in the future.
phoenix, May 25 2002
  

       [after feedback from pet technogeek:]   

       Apparently, pci *in general* is limited to 4 irqs: that's the real bottleneck. (it's pretty much only pci we're worried about, isn't it?)   

       Trying to do anything about that seems a little pointless: by the time you've implemented the hardware and software needed to support it, there'll be an entirely new system based on XXX bit computing...
yamahito, May 25 2002
  

       yamahito, thank you. A major reason for posting this idea is the dichotomy between what is SAID about Plug and Play, and what actually happens if you fail to pay attention to the availability of IRQs. The motherboard that I described has 6 PCI slots and only 4 IRQs to use among them, so I naturally want the freedom represented by having more IRQs. Which, as you wrote in a separate annotation, may as well be part of the next industy standard.   

       thumbwax, yes, tape drives traditionally used a floppy-controller interface, and so needed IRQ 6. But that is not the same as wanting any PnP card to be able to use it if it was available for something other than a floppy controller. As I mentioned in another annotation, there are USB floppy drives available, so that in theory one should be allowed to, say, assign IRQ 6 to the USB system. (Which means that IRQ 11 or whatever normally gets assigned for USB is now free for something else. Which is effectively the same as if IRQ 6 itself had been freely re-assignable.) Now, ONE way I could get ONE more IRQ is simply to disable the onboard IDE controllers (they use IRQs 14 and 15), and plug in a SCSI controller. I need one IRQ for that controller, of course, but it lets me hook up more than the 4 drives maximum allowed by the two IDE controllers. So effectively one of the two IDE IRQs is now available for something else. BUT the stupid motherboard still limits 4 IRQs to 6 PCI slots!!!! (And I see, from what others have posted here, that that stupidity is an industry-wide phenomenon.)   

       phoenix, I think I mentioned somewhere that I wanted to use a high-end sound card. On-board audio can always (so far) be disabled, to allow an IRQ for that sound card. So the existence of on-board audio (and also on-board video) is irrelevant to the shortage-of-IRQs problem.
Vernon, May 31 2002
  

       Vernon: Other people may not like this idea, but I gave you a croissant for it. I find it extremely maddening that the fourth generation of PC card slots (XT, AT, VLB, PCI) still has not adopted a decent interrupt-handling standard.   

       Actually, if there were standard conventions that allowed interrupt sharing to actually work (level-sensitive interrupt wire combined with standard polling methods) even a single interrupt wire would be sufficient. I don't know why the inventors of PCI didn't do this, nor arrange for interrupt vectors to be supplied by PCI cards themselves [Z80-based systems that do this can have 128 interrupts; 8088-based systems that do this can have 256]. Cards that use these direct-vectored interrupts could only use a limited number within non-protected-mode DOS (since software uses many interrupt vectors for other purposes) but since DOS is almost always run from within protected mode these days, that shouldn't pose a problem.
supercat, May 31 2002
  
      
[annotate]
  


 

back: main index

business  computer  culture  fashion  food  halfbakery  home  other  product  public  science  sport  vehicle