[3574] in Release_7.7_team
Re-opening issue of "Wireless card ships wrongly configured on T30
daemon@ATHENA.MIT.EDU (Bill Cattey)
Tue Oct 22 18:06:21 2002
From: Bill Cattey <wdc@MIT.EDU>
To: Pam Huntley <phuntley@us.ibm.com>
Cc: Michael Schmedlen <mschmedl@us.ibm.com>,
Steven Sakata
<ssakata@us.ibm.com>,
Steven Stiles <sastile@us.ibm.com>,
Victor R Rios
<riosvic1@us.ibm.com>,
Caroline Patzer <caro@us.ibm.com>, release-team@mit.edu, sly@mit.edu
In-Reply-To: <OF5DE20B57.C274EED3-ON87256C3F.0004D805@boulder.ibm.com>
Content-Type: text/plain
Content-Transfer-Encoding: 7bit
Date: 22 Oct 2002 18:06:19 -0400
Message-Id: <1035324379.23920.45.camel@tokata.mit.edu>
Mime-Version: 1.0
Steve Sakata and Lisa Trisciani successfully arranged the swap of my IBM
T30 for a T30 SXGA+. It is already starting to be useful.
Case ID 1-1 350 46338
Has been opened on the question: Why does my Windows XP Pro ship with
the Wireless Card requiring Disable/Re-enable before it works?
Just between you and me, I think I understand the issue:
Repeat by:
Open box containing new T30.
Power up, and configure Windows.
Use Hardwired Network.
(Memo to myself: CAREFULLY note state of Wireless here.)
Use Partition Magic to shrink NTFS partition.
Install Red Hat Linux via hardwired network card.
Modify /etc/modules.conf to comment out
alias eth0 eepro100
and activate
alias eth0 orinoco_pci
Reboot Red Hat Linux,
Unplug Hardwire line.
make extensive use of wireless, even across suspend/resume.
Shut down Linux
boot Windows XP Professional.
Note that XP can find NO wireless networks
Note that the Status display showss Not Connected and rescan does not
help.
----
Last time I did this, after being in Windows, the Linux Wireless stuff
went haywire until I went BACK into windows, disabled and re-enabled the
wireless network (from the Network Connections control panel). After
that, everything worked just fine.
----
Theory: Something is subtly mis-configured at the Windows XP end.
The mis-configuration ONLY becomes noticeable when the Linux orinoco_pci
driver is used for a while. Stated a different way: The original XP
Wireless config, clashes with assumptions the orinoco_pci driver makes
under Linux; The orinoco_pci driver sets a configuration from there that
XP VERY MUCH dislikes. Disabling and re-enabling under Windows XP
pushes the configuration into a state where thereafter both drivers are
happy.
----
Why this will happen a lot at MIT:
A lot of network help folks are a LOT more comfortable with Linux and
Mozilla then XP and Explorer. They'll install Red Hat, for the purpose
of registering the wireless card at MIT for DHCP, and then they'll
travel down the path of confusion.
----
Current status: Casey at IBM told me that the back line guy who can
work with me to learn more about this is not in today, and that I'll
have to call back tomorrow.
NOTE! Since Red Hat Linux is not officially supported, I have not YET
shared my theory with the IBM front line support folk. As far as they
know, this is a T30 that shipped mis-configured. The possible Linux
effect has not yet been broached.
Pam: Does this sound like a reasonable tack for me to have taken?
-wdc
