Re: [IPCop-user] IRQs Suck

[Bryan J. S. <b.j...@ie...>]

Richard Lynch wrote:
> So, is the IPCop FAQ perhaps a trifle imprecise, and referring to IRQs 
> when, in fact, it should be referring to the "interrupt pins" that can 
> be set by the manufacturer?

It's not set by the manufacturer.  It's a rather _deep_ discussion.

First off, understand that the PCI bus is platform-_independent_.  If 
you move away from the Intel platfrom, many processors do _not_ share 
the same approach as Intel when it comes to "interrupt pins."  PCI is 
actually very well designed.

But, secondly and unfortunately, we live in a 16-bit PC world.  This 
goes back to the Intel 286 chip and its cascaded interrupt controllers. 
  We're still dealing with this on PCs, including mapping IRQs to those 
legacy interrupt pins on the processor -- even for the PCI bus.

Third, a lot has to do with the fact that you still have ISA on your 
mainboard, even if you don't have slots -- aka the LPC (Legacy PC) 
controller in your chipset.  But even if you didn't, a lot of Windows -- 
even the HAL (Hardware Abstraction Layer) for x86 in NT-based Windows, 
would not work if you didn't.

Lastly, Becker's discussion is regarding PCI, and is 0% applicable to 
ISA, both PnP and legacy.

> On top of that, is it not a little presumptuous that I'll *HAVE* a DOS 
> box around?  I don't.

And the fault lies on IPCop/Linux for not providing firmware/config 
utilities for various vendor products?  Heck, Microsoft cannot even do 
the same for NT-based Windows!

> I don't really have a lot of colleagues who would be keen on my waltzing 
> over to their house, opening up their computer, installing some 
> hardware, and running some DOS utility on it either.  I mean, yeah, I 
> *could* find somebody willing to let me do that, but criminy, that seems 
> like a bit much to ask, don't you think?

Yeah, so go complain to the hardware vendors.  It's not the fault of 
IPCop/Linux any more than Microsoft who has the same issue with NT-based 
Windows.

> Meanwhile, my problem is that I have purchased some $10 NICs 
> manufactured by "Fast Ethernet Controller Provider Inc."
> There is no web-site in the NIC docs.
> They specifically suggest changing the INT (presumably the "interrupt 
> pins" referred to by Mr. Becker) in the BIOS to "INT A".
> So I'm back to the BIOS.
> Alas, I don't really understand the BIOS screens.

The PC BIOS world is largely proprietary, and has hindered hardware 
compatibility on the PC forever.  For those that have worked on Digital 
SRM, Sun OpenBoot and lots of other firmware architectures, it's rather sad.

LinuxBIOS is trying to change that, open up the entire PC config/POST 
world, but there are influencial powers trying to stop that.  Hopefully 
Intel's forthcoming BIOS replacement will address this -- heck, it might 
be Linux-based as well -- as Intel internally, like most semiconductor 
firms, use Linux _heavily_.

> Countless 'net searches are leaving me high and dry, at best.

What's your BIOS vendor?  AMI?


-- 
Bryan J. Smith, E.I.   b.j...@ie...   http://thebs.org
If you want the stupid letters:   http://thebs.org/certs.pdf
------------------------------------------------------------
Running Windows applications under Linux does not reduce any
political/legal "costs" but does increase the overall tech-
nical "costs."  Linux is not a better Windows than Windows.