PCI BIOS has not enabled this device!

PCI BIOS has not enabled this device!

Post by Harry Putna » Sun, 20 Jun 1999 04:00:00



This boot message: The PCI BIOS has not enabled this device!
Updating PCI command 0000->0005.
And:  SIOCSIFFLAGS: Resource temporarily unavailable (when calling ifup eth0)

Seems to be about the only clue to help diagnose why networking won't
function on a a new RedHat 5.2 install (updates applied).

This machine is part of a 3 part lan all dual boot linux machines
(dual-booting win95/98)

Machine =  HP Pavilion 6350, the network card is Netgear:
NetGear FA 310TX Fast Ethernet PCI Adaptor

This machine is network enabled when booted into win/98, and can  be
accessed by the other machines.  Not so, when booted into Linux.
The relevant boot messages (date and host kernel pruned)

(network not working on this machine)

  pcibios_init : BIOS32 Service Directory structure at 0x000f7ff0
  pcibios_init : BIOS32 Service Directory entry at 0xfd7d0
  pcibios_init : PCI BIOS revision 2.10 entry at 0xfd9e4
  Probing PCI hardware. :

  [...]

  Appletalk 0.17 for Linux NET3.035
  The PCI BIOS has not enabled this device!  Updating PCI command 0000->0005.

  eth0: Lite-On 82c168 PNIC at 0xf400, 00 a0 cc 3a 83 b7, IRQ 255.
  eth0:  MII transceiver found at MDIO address 1, config 1000 status 782d.

Another machine on the lan is running the same version of Redhat and
the Identical Netgear adaptor.  The concomittant boot messages from
there show a small difference:
(Network working on this machine)

  pcibios_init : BIOS32 Service Directory structure at 0x000faee0
  pcibios_init : BIOS32 Service Directory entry at 0xfb360
  pcibios_init : PCI BIOS revision 2.10 entry at 0xfb390
  Probing PCI hardware.

  [...]

  Appletalk 0.17 for Linux NET3.035

  Lite-On 82c168 PNIC at 0x6000, 00 a0 cc 3a 7f 87, IRQ 10.
  eth0:  MII transceiver found at MDIO address 1, config 1000 status 782d.
  eth0: Changing PNIC configuration to half-duplex, CSR6 816e0000.

I thought this might merit a look in the BIOS, and did take a look.
PhoenixBios setup Utility shows an entry for PCI settings, three
slots, but all show these settings:
Option  Rom scan  [x] enabled  (all three)
Enable master     [x] disabled (all three)

I don't know what enable or disable master is all about but tried
enableing on all three.  Nothing changed, so reset to original
settings.  I don't have a clue what I'm looking for, so didn't learn
much.  And since this machines network adaptor works when booted into
win/98.  I'm guessing that means it is enabled.

Hope this is enough material for someone knowledgeable on networking,
to make some guesses.

 
 
 

PCI BIOS has not enabled this device!

Post by M. Buchenried » Mon, 21 Jun 1999 04:00:00


[...]

>  Appletalk 0.17 for Linux NET3.035
>  The PCI BIOS has not enabled this device!  Updating PCI command 0000->0005.

>  eth0: Lite-On 82c168 PNIC at 0xf400, 00 a0 cc 3a 83 b7, IRQ 255.

                                                          ^^^^^^^^^^
[...]

Indeed. That means that the PCI BIOS did not give an IRQ to that
card. Either the card itself is broken, the driver has a bug, or the card
expects the IRQ to be set by the OS (which would be rather strange, since
the BIOS usually gives the IRQs to the PCI devices). If you do have
a diagnostic program (usually a DOS program) on an accompagnying floppy,
try running it in DOS mode. Otherwise you might perhaps want to ask
the author of the tulip.c driver, the honorable Donald Becker , at the
aforementioned email address.

Michael
--

          Lumber Cartel Unit #456 (TINLC) & Official Netscum
    Note: If you want me to send you email, don't munge your address.

 
 
 

PCI BIOS has not enabled this device!

Post by Harry Putna » Mon, 21 Jun 1999 04:00:00




> [...]

> > Appletalk 0.17 for Linux NET3.035 > The PCI BIOS has not enabled
> this device!  Updating PCI command 0000->0005.

> >  eth0: Lite-On 82c168 PNIC at 0xf400, 00 a0 cc 3a 83
> >  b7, IRQ 255.                                                          
> >     ^^^^^^^^^^
> [...]

> Indeed. That means that the PCI BIOS did not give an IRQ to that
> card. Either the card itself is broken, the driver has a bug, or the
> card expects the IRQ to be set by the OS (which would be rather
> strange, since the BIOS usually gives the IRQs to the PCI
> devices). If you do have a diagnostic program (usually a DOS
> program) on an accompagnying floppy, try running it in DOS
> mode. Otherwise you might perhaps want to ask the author of the
> tulip.c driver, the honorable Donald Becker , at the aforementioned
> email address.

Thanks... Yes it turn out the BIOS on this particular machine:
HP Pavillion 6350 (PhoenixBIOS utility) Has a section called
"Installed OS" with three choices:

Other
win/95
win/98/NT5.0

It was set to "win/98/NT5.0"  I changed it to other and now an IRQ is
being assigned (11) and all is well.  For a detailed explanation of
why this is so, Mr. David Becker has an online site explaining it.

http://cesdis.gsfc.nasa.gov/linux/misc/irq-conflict.html

The wording on my BIOS was different but apparently the same setup as
described there.

For fellow readers here... the above URL has lots of useful info.