On Oct 2, 12:47am, Greg A. Woods wrote:
} [ On Fri, October 2, 1998 at 06:04:19 (+0200), Zdenek Salvet wrote: ]
} >
} > But, what is the 1st SCSI controller attached to the system ?
} > SVR4ish IRIX 6.4 on our Origin has changed SCSI controller numbers
} > after applying patches to the OS and screwed us even though we had not
} > moved any hardware before :-(
} 
} You can't blame hardware design failures on the OS!  ;-)

     This sounds like an OS bug to me.

} You also shouldn't compare IRIX to anything sane!  ;-)

     In most aspects, I would consider IRIX to be fairly sane.  The
main thing that gets me is its major lack of security out of the box.

} In theory the hardware bus should be ordered in some way (either by
} physical slots, or by address pins/switches on the cards, ROM serial

     The only way that should be used is physical slot.  Everything
else leads to unnecessary complexity and confusion.

     Consider what would happen if you were to fill up all the slots
in an Apple II with floppy controllers?  It would work perfectly
without any mucking around (ignoring the fact that you better not turn
on too many drives at once, or you would blow the power supply), and
you would be able to easily access all the drives.

     Now, consider what would happen if you were to fill up all the
slots in an original PC with floppy controllers?  The first problem
you will have is I/O-addr conflict.  If you're lucky, one or more of
the cards will have a jumper to select a secondary I/O-addr, but don't
even think about a tertiary one.  The next problem is going to be an
IRQ conflict.  Good luck on this one.  Next is a DRQ conflict.  Oops,
there aren't enough of these to go around even if you could tell the
card to use a different one.  Finally, if you manage to solve all
these problems, there is the issue of how to access all the drives.
You can't do that without writing, or somehow obtaining, a custom
device driver.

     For all its simplicity, I think the Apple II was much better
designed and more elegant then the original IBM PC.  With the
invention of the EISA bus, the PC made a major step in right
direction.  With PCI, they finally got it right; although, there is
still some cruft that complicates things on mixed ISA/PCI designs.  Of
course, with the most common PC OS'es you'd still have the device
driver problem if you tried to fill every empty slot with floppy
controllers, but at least the hardware would work right.

} inspection, and by the kernel.  If some patch to your OS causes the
} kernel to count from the other direction suddenly then that's something
} you do need to be aware of, but it's not generally a problem.

     No, but it blows your whole argument.  If you can't depend on the
things to stay put, then why bother using the location as the
identifier?

} Personally I like devices with little jumpers or thumbwheel switches so
} it doesn't matter which slot you use, or what order they appear on the
} bus cable, etc.  This goes for cards on a bus, external peripherals, and
} everything.

     As I showed above, the only reasonable and reliable way to handle
cards is by the physical slot that it is in.  Having switches of some
sort only leads to limitations of how many items of a given kind you
can put in the machine and various other conflicts.

}-- End of excerpt from Greg A. Woods