Very Computer

I set up module.conf tulip.o with options io=fc00 irq=10 (from diag dos
utility) and ran depmod -a. I thought I should be OK, but the interface
won't come up.

The card's readme blithely omits the options, and uses pci-scan.o to set
them at boot. To get the pci-scan.o module, I'd need to build the source,
but I can't build the tulip source included with the NIC, the pci-skeleton.c
unit has compile errors, apparently related to the kernel header.

I need help getting this thing running ASAP, please send any suggestions.

Thanks.

If the de4x5 works and tulip doesn't, please file a bug report.
According to one of the networking guys on the linux kernel list,
the de4x5 driver is going to be phased out.  (IMO it should be
the other way around.)

--
Manfred
---------------------------------------------------------------
ipchainsLogAnalyzer, NetCalc, whois at: <http://logi.cc/linux/>

Quote:> I set up module.conf tulip.o with options io=fc00 irq=10 (from diag dos

I'd really like to use the existing tulip.o installed by redhat, it would
seem so much easier, and I know from other installs, that it works with this
NIC. I just don't know what I'm doing wrong here.

Also followed instructions at
http://www.scyld.com/network/updates.html#release

even with kgcc rpm installed:
http://rpmfind.net/linux/RPM/redhat/7.0/i386/kgcc-1.1.2-40.i386.html

I get the compile errors discussed
tulip.c: In function `tulip_open':
tulip.c:1437: structure has no member named `tbusy'
tulip.c:1438: structure has no member named `start'
...

Quote:> The tulip driver seems to have huge problems of late.
> I switched to the de4x5 driver and it worked after explicitly
> specifying the IRQ.

and get:
/lib/modules/2.216-22/net/de4x5.o: invalid pararameter parm_irq

Am I using wrong syntax, or is there another problem?

Allright,
First of all I have heard a lot of bad critics about this card that it doesn't
really work  although it says on the box when you buy it Linux compatible.

I know that for sure that when it asks you for io the format is 0x0000 and for
irq it should be just an integer value. (You get those from WidoZ)

Another suggestion: have you recompiled your kernel to load the right modules
for the tulip driver, incase you haven't you need to get in the /usr/src/linux
as su and type "make xconfig" load the right modules and see what happens.

I am not absolutely sure, but I think I heard that it works with the 2.4.2
kernel. You will probably be using 2.2.16-20 or 2.2.17 I believe, so you might
want to get that kernel.  You can get it from kernel.org
  ------------------------------------------------------------------------

    options de4x5 args='eth0:BNC eth1:TP 10Mb'

The syntax of supplying options is explained in the source file:

    /usr/src/linux/drivers/net/de4x5.c

What do your startup messages say?  Try:

    dmesg | grep 'eth[0-9]'

--
Manfred
---------------------------------------------------------------
ipchainsLogAnalyzer, NetCalc, whois at: <http://logi.cc/linux/>

alias eth0 tulip
options tulip options=3 debug=1

Option List:
         0 Auto-select (default to the 10baseT link)
         1 10base2
         2 AUI
         3 100baseTx
         4 10baseT-FD
         5 100baseTx-FD
         6 100baseT4
         7 100baseFx
         8 100baseFx-FD
         9 MII 10baseT
        10 MII 10baseT-FD
        11 MII (autoselect)
        12 10baseT (no autoselect), v0.69 and later only
        13 MII 100baseTx
        14 MII 100baseTx-FD
        15 MII 100baseT4

Debug List:
         1 normal output
         2 more verbose
         3 even more verbose
         4 even more verbose
         6 insanely verbose

I got this info from their web site:
http://www.linksys.com
Hope this helps

Scott G.

Quote:> I have a RH7 box that I just switched network cards in. The previous one
> used ne.o, and the new one is a Linksys LNE100TX version 4.2

> I set up module.conf tulip.o with options io=fc00 irq=10 (from diag dos
> utility) and ran depmod -a. I thought I should be OK, but the interface
> won't come up.

> The card's readme blithely omits the options, and uses pci-scan.o to set
> them at boot. To get the pci-scan.o module, I'd need to build the source,
> but I can't build the tulip source included with the NIC, the
pci-skeleton.c
> unit has compile errors, apparently related to the kernel header.

> I need help getting this thing running ASAP, please send any suggestions.

> Thanks.

I am a newbie (obvious from my name) and for the second install I choose
doing a "custom" install.  I did this because I wanted to learn what apps
were available and read the descriptions of them to see if they might be of
interest to me.  After choosing a host of them they were installed -
RedHat7.0 checks for cross-compatibility to make sure any necessary apps
that might be required for some of my choices are included.

The modem worked okay at first, I think, I can't remember trying it out but
I am fairly sure that it worked okay.  However after playing around and
trying out a few apps the modem is no longer available.  When I rebooted,
after getting a fail on trying to use the modem to dial up to the internet,
I got a message on a text based setup screen (like XConfigurator) that said
I had 30 seconds to alter/remove/leave the settings of a hardware device
that was now removed.  I said to ignore the request and see if it worked in
RedHat.  It didn't.  The next bootup the same message appeared.  The modem
no longer works in RedHat now.

I've been playing about with the BIOS to see if I could get RedHat to
auto-detect the modem but to no avail.  I've just checked my other PC, the
one with RedHat7.0 on it.  The modem did work okay after the install the
second time around.  It now fails.  It's a dual booter, Win98SE, and the
modem works fine under Windows.  It is NOT a Winmodem - NOT a Winmodem.  It
used to work fine under RedHat but does not do so any more.

It was detected as dev/ttyS0.  I have tried auto-detecting it but it says
that there is no modem installed.

What is wrong here?  Have I mucked up the Configuration so that it no longer
finds my modem?  Is the Kernel knackered?  I think it is 2.2.16, or similar
to that.  It is the default with the RedHat 7.0 install (Guinness).

Please help this newbie as he is banging his head against a brick wall.

Thanks.

:>Update:
:>> I set up module.conf tulip.o with options io=fc00 irq=10 (from diag dos
:>Changed it to io=0xfc00, didn't help.
:>
:>I'd really like to use the existing tulip.o installed by redhat, it would
:>seem so much easier, and I know from other installs, that it works with
this
:>NIC. I just don't know what I'm doing wrong here.
:>
:>Also followed instructions at
:>http://www.scyld.com/network/updates.html#release
:>
:>even with kgcc rpm installed:
:>http://rpmfind.net/linux/RPM/redhat/7.0/i386/kgcc-1.1.2-40.i386.html
:>
:>I get the compile errors discussed
:>tulip.c: In function `tulip_open':
:>tulip.c:1437: structure has no member named `tbusy'
:>tulip.c:1438: structure has no member named `start'

I have done this several times. First, you will need to follow the
instructions about RH 7:

When updating a single driver you will need to build pci-scan.o along
with the specific driver.o file. This means that in addition to the
driver.c file, you must download pci-scan.c, pci-scan.h, and
kern_compat.h.

I copied the compile instructions out of tulip.c and pci-scan.c and used
them EXACTLY along with adding the -I/usr/src/linux/include to the kgcc
(not gcc) compile command.

One other thing. I found that if I tried doing it after a workstation or
server install as opposed to a custom install, the kgcc compile did not
work. I reinstalled the system using the custom option and selected the
development package for install. My guess is that the non-custom install
does not get everything needed for kgcc.

In any event, after much banging around, it does work. I have done it
enough times now to have made just about every mistake possible.

--
    Thank you, Mike.

 I'm trying to compile tulip, pci-scan, and kern_compat using kgcc and
the -I/usr/src/linux/include".  Problem is, that directory doesn't exsist on
my system.

Any ideas?

Thanks in advance,
Scott

:>Update:
:>> I set up module.conf tulip.o with options io=fc00 irq=10 (from diag dos
:>Changed it to io=0xfc00, didn't help.
:>
:>I'd really like to use the existing tulip.o installed by redhat, it would
:>seem so much easier, and I know from other installs, that it works with
this
:>NIC. I just don't know what I'm doing wrong here.
:>
:>Also followed instructions at
:>http://www.scyld.com/network/updates.html#release
:>
:>even with kgcc rpm installed:
:>http://rpmfind.net/linux/RPM/redhat/7.0/i386/kgcc-1.1.2-40.i386.html
:>
:>I get the compile errors discussed
:>tulip.c: In function `tulip_open':
:>tulip.c:1437: structure has no member named `tbusy'
:>tulip.c:1438: structure has no member named `start'

I have done this several times. First, you will need to follow the
instructions about RH 7:

When updating a single driver you will need to build pci-scan.o along
with the specific driver.o file. This means that in addition to the
driver.c file, you must download pci-scan.c, pci-scan.h, and
kern_compat.h.

I copied the compile instructions out of tulip.c and pci-scan.c and used
them EXACTLY along with adding the -I/usr/src/linux/include to the kgcc
(not gcc) compile command.

One other thing. I found that if I tried doing it after a workstation or
server install as opposed to a custom install, the kgcc compile did not
work. I reinstalled the system using the custom option and selected the
development package for install. My guess is that the non-custom install
does not get everything needed for kgcc.

In any event, after much banging around, it does work. I have done it
enough times now to have made just about every mistake possible.

--
    Thank you, Mike.