Where is this bootup/log message printed?

I'm trying to configure a new kernel for a system for which I have one in
which the network does in fact work perfectly [every time it's booted] so I
know absolutely that there's no hardware problem. But in the new kernels
I've made the network is not working and I'm getting the messages during
bootup (and it's also in the system boot log)
Nov 24 10:42:02 Halkin network: Setting network parameters: succeeded
Nov 24 10:42:02 Halkin network: Bringing up interface lo: succeeded
Nov 24 10:42:02 Halkin ifup: SIOCSIFFLAGS: Device or resource busy
Nov 24 10:42:02 Halkin ifup: Failed to bring up eth0.
Nov 24 10:42:02 Halkin network: Bringing up interface eth0: failed

I've grep'ed the *.c and *.h files in the kernel source tree and the script
files in the directories /etc/rc.d/* and I simply cannot find anyplace where
the message
ifup: SIOCSIFFLAGS: Device or resource busy
would be printed. I want to find this as a clue to what's missing in my
kernel configuration.

Any suggestions, hints, URLs, etc gratefully appreciated.

TIA
Norm

I assume you looked at /var/log/kern.log right?


"Norm Dresner" <(E-Mail Removed)> wrote in message
news:OR2pd.966995$(E-Mail Removed)...
> I'm trying to configure a new kernel for a system for which I have one in
> which the network does in fact work perfectly [every time it's booted] so
I
> know absolutely that there's no hardware problem. But in the new kernels
> I've made the network is not working and I'm getting the messages during
> bootup (and it's also in the system boot log)
> Nov 24 10:42:02 Halkin network: Setting network parameters: succeeded
> Nov 24 10:42:02 Halkin network: Bringing up interface lo: succeeded
> Nov 24 10:42:02 Halkin ifup: SIOCSIFFLAGS: Device or resource busy
> Nov 24 10:42:02 Halkin ifup: Failed to bring up eth0.
> Nov 24 10:42:02 Halkin network: Bringing up interface eth0: failed
>
> I've grep'ed the *.c and *.h files in the kernel source tree and the
script
> files in the directories /etc/rc.d/* and I simply cannot find anyplace
where
> the message
> ifup: SIOCSIFFLAGS: Device or resource busy
> would be printed. I want to find this as a clue to what's missing in my
> kernel configuration.
>
> Any suggestions, hints, URLs, etc gratefully appreciated.
>
> TIA
> Norm
>

On 2004-11-24, Norm Dresner <(E-Mail Removed)> wrote:
> ifup: SIOCSIFFLAGS: Device or resource busy

ifup is the program/script that report the error (in /sbin/ or /usr/sbin),
SIOCSIFFLAGS is the error and
Device or resource busy is probablu the description of the error (probably
from error.h)

> would be printed. I want to find this as a clue to what's missing in my
> kernel configuration.

It looks like a bug in the driver, the 'resource' (aka: the NIC) is
'busy'.

Davide

--
I'm still waiting for the "Honk if you hate Microsoft" [bumper sticker],
but that might get annoying, everyone honking at you.
-- From a Slashdot.org post

Those messages were taken from
/var/log/messages
with identical entries in
/var/log/boot.log

That's all the information I have.

Norm


"HisNameWasRobertPaulson" <kalis_anon_REMOVE_@hotmail.com> wrote in message
news:41a4c0d0$0$91004$(E-Mail Removed). ..
> I assume you looked at /var/log/kern.log right?
>
>
> "Norm Dresner" <(E-Mail Removed)> wrote in message
> news:OR2pd.966995$(E-Mail Removed)...
> > I'm trying to configure a new kernel for a system for which I have one
in
> > which the network does in fact work perfectly [every time it's booted]
so
> I
> > know absolutely that there's no hardware problem. But in the new
kernels
> > I've made the network is not working and I'm getting the messages during
> > bootup (and it's also in the system boot log)
> > Nov 24 10:42:02 Halkin network: Setting network parameters: succeeded
> > Nov 24 10:42:02 Halkin network: Bringing up interface lo: succeeded
> > Nov 24 10:42:02 Halkin ifup: SIOCSIFFLAGS: Device or resource busy
> > Nov 24 10:42:02 Halkin ifup: Failed to bring up eth0.
> > Nov 24 10:42:02 Halkin network: Bringing up interface eth0: failed
> >
> > I've grep'ed the *.c and *.h files in the kernel source tree and the
> script
> > files in the directories /etc/rc.d/* and I simply cannot find anyplace
> where
> > the message
> > ifup: SIOCSIFFLAGS: Device or resource busy
> > would be printed. I want to find this as a clue to what's missing in my
> > kernel configuration.
> >
> > Any suggestions, hints, URLs, etc gratefully appreciated.
> >
> > TIA
> > Norm
> >
>
>

"Davide Bianchi" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) .net...
> On 2004-11-24, Norm Dresner <(E-Mail Removed)> wrote:
> > ifup: SIOCSIFFLAGS: Device or resource busy
>
> ifup is the program/script that report the error (in /sbin/ or /usr/sbin),
> SIOCSIFFLAGS is the error and
> Device or resource busy is probablu the description of the error (probably
> from error.h)
>
> > would be printed. I want to find this as a clue to what's missing in my
> > kernel configuration.
>
> It looks like a bug in the driver, the 'resource' (aka: the NIC) is
> 'busy'.

But ... but ...

The same system .. same hardware .. same driver ... with a different kernel
image works just fine.

It's got to be some difference between the kernels where the network works
and doesn't work. Until this afternoon, I had a choice between either the
network working or the Real-Time Linux extensions working, but all of the
kernels I'd build recently had working RTLinux and non-working networks. I
went back to an older config-file that produced the last good network image
and figured out what I had to do to make the RTLinux play properly.

Now, of course, the question is, What's different between the
configurations? AFAICT the major difference that would impact the
networking is that in the kernel with good networking, I [seem to] have
ISA-PnP turned off and in the kernel with bad networking, it's turned on.
It's possible that this is the difference and it would definitely affect the
way the kernel tried to initialize the NIC, but I was pretty sure that a
3Com 3C509 Etherlink III card was ISA-PnP capable since there are MS-DOS
programs that can be used to alter the configuration.

Okay, I've got a working kernel with a working NIC and I should be happy,
right? No, I still think it took me all too long to finally stumble onto a
solution to a problem that I still haven't been able to completely
characterize.

Norm

On 2004-11-24, Norm Dresner <(E-Mail Removed)> wrote:
> The same system .. same hardware .. same driver ... with a different kernel
> image works just fine.

Where the two kernel came from? If one have been built by you
and the other was pre-built, maybe you have a bugged version of
the kernel sources. I had the same problem with a distribution that
distributed a 'modified' version of the kernel. Recompiling the
kernel showed up the problem.

> networking is that in the kernel with good networking, I [seem to] have
> ISA-PnP turned off and in the kernel with bad networking, it's turned on.

I know that Isa-pnp does inficiates somehow the functioning of PCMCIA
and other similar nics, but if that isn't a laptop...

Maybe you should try to check the kernel developer's mailing list.

Davide

--
"A copy of Windows 95 and 4x4 MB of SIMMs, please."

Responses in-line

"Davide Bianchi" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed) .net...
> On 2004-11-24, Norm Dresner <(E-Mail Removed)> wrote:
> > The same system .. same hardware .. same driver ... with a different
kernel
> > image works just fine.
>
> Where the two kernel came from? If one have been built by you
> and the other was pre-built, maybe you have a bugged version of
> the kernel sources. I had the same problem with a distribution that
> distributed a 'modified' version of the kernel. Recompiling the
> kernel showed up the problem.
>
All versions of the kernel were built by me. Until this afternoon,
there were versions in which Real-Time Linux worked and the network didn't
and versions in which the network worked but I couldn't load the Real-Time
Linux extensions. Finally today I got a version in which "everything"
works.

BTW, AFAICT the main problem with the Real-Time Linux in my older
versions of the kernel in which the network worked was that it was built
with gcc 2.96 instead of 2.95.3.

> > networking is that in the kernel with good networking, I [seem to] have
> > ISA-PnP turned off and in the kernel with bad networking, it's turned
on.
>
> I know that Isa-pnp does inficiates somehow the functioning of PCMCIA
> and other similar nics, but if that isn't a laptop...
>

No, it's not a laptop but a quick look at the 3C509.c module in the
kernel source tree showed that it's used in doing the setup of the NIC.

> Maybe you should try to check the kernel developer's mailing list.
>
Thanks for the mailing-list suggestion.

Norm

Norm Dresner <(E-Mail Removed)> wrote:
> configurations? AFAICT the major difference that would impact the
> networking is that in the kernel with good networking, I [seem to] have
> ISA-PnP turned off and in the kernel with bad networking, it's turned on.

Not sure what you mean, but if you mean PNP is done from within the
kernel at bootup instead of from user space, you never want that.
You should always set the values you want yourself via isapnp, the
utility.

> It's possible that this is the difference and it would definitely affect the
> way the kernel tried to initialize the NIC, but I was pretty sure that a
> 3Com 3C509 Etherlink III card was ISA-PnP capable since there are MS-DOS

You are wrong. It is not. You set a 3c509 via a special utility. It is
a well-known fact (for about ten years!) that you configure a 3c509 by
trning OFF pnp on the card, if you can, and etting the irq and io with
the utility (to one of only a few possible values). The you load the
value with those values specified.

> programs that can be used to alter the configuration.

And they aren't PNP!

> Okay, I've got a working kernel with a working NIC and I should be happy,
> right? No, I still think it took me all too long to finally stumble onto a
> solution to a problem that I still haven't been able to completely
> characterize.

Why?

Peter

"Peter T. Breuer" <(E-Mail Removed)> wrote in message
news:9l1g72-(E-Mail Removed)...
> Norm Dresner <(E-Mail Removed)> wrote:
> > configurations? AFAICT the major difference that would impact the
> > networking is that in the kernel with good networking, I [seem to] have
> > ISA-PnP turned off and in the kernel with bad networking, it's turned
on.
>
> Not sure what you mean, but if you mean PNP is done from within the
> kernel at bootup instead of from user space, you never want that.
> You should always set the values you want yourself via isapnp, the
> utility.
>
> > It's possible that this is the difference and it would definitely affect
the
> > way the kernel tried to initialize the NIC, but I was pretty sure that a
> > 3Com 3C509 Etherlink III card was ISA-PnP capable since there are MS-DOS
>
> You are wrong. It is not. You set a 3c509 via a special utility. It is
> a well-known fact (for about ten years!) that you configure a 3c509 by
> trning OFF pnp on the card, if you can, and etting the irq and io with
> the utility (to one of only a few possible values). The[n] you load the
> value with those values specified.
>
> > programs that can be used to alter the configuration.
>
> And they aren't PNP!
>

Perhaps we have a different conception of what PnP means. I think that
it means that the card can be configured [by the BIOS or system
initialization program] at bootup to a configuration that's convenient given
the other peripherals in the system. If it's possible to determine card
settings and modify them via software, then the card is inherently PnP
whether it's used in that mode or not.
In the same hardware box in which I'm running Linux, I'm also running a
Windows 2000 system. Based on the information in the /proc FS under Linux
and the device properties in Win2K, the card is being set to different io
ports and a different IRQ by the two systems. That's a PnP card!
There is, in fact an "industry standard" definition of PnP. I'm using
the Microsoft "PC97 Hardware Design Guide" as a reference and it says, in
part,
"A Plug and Play-compliant system can identify the Plug and Play ISA
cards and the resources they use, and then program the resources so they do
not conflict."
That's what I mean by PnP and that's exactly what the 3C509 card is and
it's exactly what both Linux and Win2K are doing with the card.


> > Okay, I've got a working kernel with a working NIC and I should be
happy,
> > right? No, I still think it took me all too long to finally stumble
onto a
> > solution to a problem that I still haven't been able to completely
> > characterize.
>
> Why?
>

"Why?" I think that I should have not had to turn off something that
seems to work perfectly in at least one OS (Win2K) to get a piece of
hardware to work in another one? I think that there should be some
centralized facility somewhere that would allow me to look-up the error
message I'm getting and see what it means. And most of all, I',
disappointed in not getting some support from an open-source community that
prides itself on producing the best possible software when in fact all it
does is flame someone who asks for help.

> Peter

Norm Dresner <(E-Mail Removed)> wrote:
> Perhaps we have a different conception of what PnP means. I think that

No we don't. Please read up on how to configure a 3c509. It should be
set to a specific io port and irq by the external utility (in linux
"3c5x9setup"), and then the driver should be loaded using the configured
values as parameters.

See "linux 3c509 configuration notes" in GOOGLE.

I think newer drivers can query the card for the irq given the io port,
or vice versa, but just READ instead of spouting.

> > > right? No, I still think it took me all too long to finally stumble onto a
> > > solution to a problem that I still haven't been able to completely
> > > characterize.
> >
> > Why?
> >
>
> "Why?" I think that I should have not had to turn off something that
> seems to work perfectly in at least one OS (Win2K) to get a piece of
> hardware to work in another one? I think that there should be some

You are asking about something that has been known for all of elapsed time.
Hence "why". The 3c509 is not particularly isa pnp compatible - it predates
isa pnp and isa pnp has always been "iffy" in it (it was added later in
ome versions, perhaps b, likely c) and simply does not work well.
Probably the windows driver ignores pnp when it sees that card and tries
a standard value or looks up in its file registry what ip port has been
selected for the card. But yes, the card behaves just as badly in dos
and windows, and always has.


I'll quote the ethernet HOWTO for you on the subject of the 3c509, ince
you seem to like to argue instead of reading:

==============

This card is fairly inexpensive and has good performance for an ISA
non-bus-master design. The drawbacks are that the original 3c509
requires very low interrupt latency. The 3c509B shouldn't suffer from
the same problem, due to having a larger buffer. (See below.) These
cards use PIO transfers, similar to a ne2000 card, and so a shared
memory card such as a wd8013 will be more efficient in comparison.

The original 3c509 has a small packet buffer (4kB total, 2kB Rx, 2kB
Tx), causing the driver to occasionally drop a packet if interrupts are
masked for too long. To minimize this problem, you can try unmasking
interrupts during IDE disk transfers (see man hdparm) and/or increasing
your ISA bus speed so IDE transfers finish sooner.

The newer model 3c509B has 8kB on board, and the buffer can be split
4/4, 5/3 or 6/2 for Rx/Tx. This setting is changed with the DOS
configuration utility, and is stored on the EEPROM. This should
alleviate the above problem with the original 3c509.

3c509B users should use either the supplied DOS utility to disable the
plug and play support, and to set the output media to what they
require. The linux driver currently does not support the Autodetect
media setting, so you have to select 10Base-T or 10Base-2 or AUI. Note
that to turn off PnP entirely, you should do a 3C5X9CFG /PNPISABLE
and then follow that with a hard reset to ensure that it takes effect.

Some people ask about the ``Server or Workstation'' and ``Highest Modem
Speed'' settings presented in the DOS configuration utility. Donald
writes ``These are only hints to the drivers, and the Linux driver does
not use these parameters: it always optimizes for high throughput
rather than low latency (`Server'). Low latency was critically
important for old, non-windowed, IPX throughput. To reduce the latency
the MS-DOS driver for the 3c509 disables interrupts for some
operations, blocking serial port interrupts. Thus the need for the
`modem speed' setting. The Linux driver avoids the need to disable
interrupts for long periods by operating only on whole packets e.g. by
not starting to transmit a packet until it is completely transferred to
the card.''

Note that the ISA card detection uses a different method than most
cards. Basically, you ask the cards to respond by sending data to an
ID_PORT (port 0x100 to 0x1ff on intervals of 0x10). This detection
method means that a particular card will always get detected first in a
multiple ISA 3c509 configuration. The card with the lowest hardware
ethernet address will always end up being eth0. This shouldn't matter
to anyone, except for those people who want to assign a 6 byte hardware
address to a particular interface. If you have multiple 3c509 cards, it
is best to append ether=0,0,ethN commands without the I/O port
specified (i.e. use I/O=zero) and allow the probe to sort out which
card is first. Using a non-zero I/O value will ensure that it does not
detect all your cards, so don't do it.

If this really bothers you, have a look at Donald's latest driver, as
you may be able to use a 0x3c509 value in the unused mem address fields
to order the detection to suit your needs.

======================

Peter