[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Howto implement device detection without thrashing

On Thu, Apr 01, 1999 at 08:35:39PM -0500, Randolph Chung wrote:
> > Yes, that will work.  However, Randolph suggested saving only the ID of the
> > upcoming test, implying that crash recovery would involve redoing all
> > tests.  That is what I was responding to:
> I wasn't being clear -- what I originally intended was that we can save the
> ID of the current test (call it n), then if we need to restart, we would
> start at test n+1. For all the tests 1..n-1, we need to write out in a
> separate place config data that we have detected. As Knghtbrd pointed out
> though, we also need to do something about scans causing the machine to hang
> so that on subsequent rescans (say if hardware changes) we don't crash again.
> I do intend to write a prototype of this in a few weeks time. Maybe when we
> have some actual code to test we can work out these problems....

My thoughts on the matter after reading major messages in the thread once
again before replying involve this process:

1. Probe PCI if the machine has PCI bus to probe.
2. If the machine has an ISA bus, present the user with a menu allowing
   them to select what to probe for.  Such a menu:

        [X] ISA Plug'n'Play devices
        [ ] network cards
        [ ] sound cards
        [ ] modems

   The title and stuff on the menu really needs to say something about
   this all being ISA stuff and telling them they probably should have
   the machine probe only for the stuff that's installed
3. Present the user with the list of hardware found from the above and
   let them decide what to do about it..

Note we could in theory put PCI devices at the top of the menu and have
just it and PnP checked by default.  Those two are USUALLY safe to test
for.  I think we really need big hairy warnings on the legacy devices. 
Even if we checkpoint for everything we can cut down the pain by making
sure the user is the one who decides what to check for in the legacy
department.  It's reasonable that we should probably give the option to
add a particular device and probe for just that device, but that'll get
kinda hairy if we are going to cram all of this onto the installation

Joseph Carter <knghtbrd@debian.org>            Debian GNU/Linux developer
PGP: E8D68481E3A8BB77 8EE22996C9445FBE            The Source Comes First!
Mere nonexistence is a feeble excuse for declaring a thing unseeable. You
*can* see dragons.  You just have to look in the right direction.
        -- John Hasler

Attachment: pgpk6Zv39rrd6.pgp
Description: PGP signature

Reply to: