Installing FC6 on an Asus S6F laptop

last updated: $Date: 2007/11/04 22:09:41 $

I bought an Asus S6F at DataVision in New York, since I was in town anyway to see my friend Wes, who had recommended the model to me based on his recent research into new notebooks. It's not the leather-covered model; the only thing I'll pay $500 to cover in leather is myself, when riding.

I was very pleased with DataVision; they allowed me to boot their shop S6F off an FC6 install DVD, and play around with bash commands (lspci, cat /proc/cpuinfo, that sort of thing) based purely on my word that I wouldn't damage the installed Windows. They were also honest with me about the prospects of getting the laptop in time (they were out of stock, but expected a delivery), and they shipped it to me exactly when they said they would. I would happily do business with them again (declaration: I have no connection with them, save as a satisfied customer).

image of laptop, for scale purposes

This shows the laptop with a DVD on the keyboard, for scale purposes. It's small, though not as tiny as my lifebook. Then again, it has a CD drive, which makes installing the OS a whole lot easier.

INSTALL

FC6 install from install DVD went like clockwork, even running in graphical mode. I erased all traces of Windoze from the laptop without ever booting it, afaict.

UPDATES

I had a fairly-recent set of FC6 i386 updates on a DVD. I inserted the DVD into the Asus, which mounted it on /media/FC6-updates.

Then I edited /etc/yum.repos.d/fedora-updates.repo and inserted a section

[dvd]
name=Fedora Core $releasever - $basearch - Updates on DVD
baseurl=file:///media/FC6-updates
enabled=0
gpgcheck=1
gpgkey=file:///etc/pki/rpm-gpg/RPM-GPG-KEY-fedora

Then I did the following commands from a text console (ie, press ctrl-alt-F1 and log in as root). The first time, I tried the upgrade from an xterm window under the GUI; the SELinux rpm update bounces some element of the windowing system and kills the update halfway. Since I didn't want a half-completed (huge) yum update leaving the system unstable, that meant blowing the whole OS away and reinstalling.
yum --disablerepo=* --enablerepo=dvd update

That failed claiming it needed a cyrus library to update gaim, so I removed gaim (I don't use IM anyway), then got success with
yum --disablerepo=* --enablerepo=dvd remove gaim yum --disablerepo=* --enablerepo=dvd update

That ran to completion.

GRAPHICS

I had already tested the FC6 install in the store, and the graphics card was supported almost perfectly. The screen seemed to be trying to be about two pixels wider than it really was, so got a strange slightly-sliding window effect where moving the mouse all the way to the RHS of the screen jumped the image about two pixels left, and vice-versa on the LHS. This seems to have since gone away, but I don't quite know how I got rid of it.

Later note: having finally hooked the machine up to an external monitor, there seems to be no signal on the external monitor port (on the LHS of the machine). I have not yet tried rebooting the box to see if the port gets "brightened" if there's a monitor there at card reset time (reboot, X startup). The keyboard control (Fn-F8) seems to have no effect.

ETHERNET

lspci says the network card is a Realtek Semiconductor Co., Ltd. RTL8111/8168B PCI Express Gigabit Ethernet controller (rev 01). This doesn't seem to be supported by FC6 out-of-the-box (so a network install might well fail).

Googling for the card revealed that I need the r1000 drivers. These are made available in source by Realtek and can be found via Realtek's driver page, which points me to ftp://202.65.194.18/cn/nic/r1000_v1.05.tgz amongst others.

I checked I had kernel source (rpm -qa|grep kernel-devel), then unpacked the distribution and built and installed the module, as root, with

cd /usr/local/src
tar xvzf /path/to/tarball/r1000_v1.05.tgz
cd r1000_v1.05
make clean modules
make install
depmod -a

I can then modprobe r1000 and configure eth0 accordingly. The network seems to work fine, although as I don't have any other gigabit devices, I can't comment on the speed.

CRYPTO FILE SYSTEM

I repeated my earlier work, which is described in an earlier technote and I now have a nice encrypted file system on /home.

SUSPEND-TO-DISC / -MEMORY

This just works out-of-the-box. Without any change to grub.conf, from the "foot" menu, system -> suspend works to suspend to memory, and system -> shut down -> hibernate works to suspend to disc. A nice feature: when it wakes up, even if it's seconds later, the screen comes up locked. Combined with encrypted home directories, this will be nicely secure.

There are also hooks for "Suspend" and "Hibernate" on the new GNOME battery applet. These work fine.

WIRELESS

lspci says the card is a Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02). Getting this working requires running non-free software; I am normally dead against this but I don't have any choice given that the hardware is on the motherboard. I also slightly sympathise with Intel's plight, because 802.11 requires a radio, and radios are tightly controlled by governments. In order not to get blasted into the stone age, Intel has no choice but to lock me out of my own hardware. I don't like it, but it's not entirely Intel's fault. If this annoys you, support the GNU software radio project, and press your vendor to use a very open and stupid transceiver to implement 802.11[abg] instead (fat chance).

I went to http://ipw3945.sourceforge.net/, got the driver, and put the source tarball into /usr/local/src/ .

README says the driver needs ieee80211 v1.1.12 or better. I did modprobe ieee80211, /var/log/messages reports this is v1.1.13 so I don't need to faff around building a new ieee80211 subsystem.

So, I did

cd /usr/local/src/ipw3945-1.1.3
make
make install

Following the README in that directory (hereinafter referred to as 'the README'), I did

cd /usr/local/src/ipw3945d-1.7.22
cp x86/ipw3945d /usr/local/sbin/

The README refers to /etc/hotplug/firmware.agent, which doesn't exist on my system. To find out where to put firmware, I did

modprobe ipw3945

then ifconfig -a showed no new interfaces. That's bad.

tail /var/log/messages showed

Dec 29 12:21:57 anni kernel: ipw3945: probe of 0000:02:00.0 failed with error -2
Dec 29 12:21:57 anni firmware_helper[32746]: Loading of /lib/firmware/ipw3945.ucode for ipw3945 driver failed: No such file or directory

so following the README, I did

tar xvzf ipw3945-ucode-1.13.tgz 
cp ipw3945-ucode-1.13/ipw3945.ucode /lib/firmware
modprobe ipw3945

then /var/log/messages showed

Dec 29 12:39:54 anni kernel: ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.1.3d
Dec 29 12:39:54 anni kernel: ipw3945: Copyright(c) 2003-2006 Intel Corporation
Dec 29 12:39:54 anni kernel: PCI: Enabling device 0000:02:00.0 (0000 -> 0002)
Dec 29 12:39:54 anni kernel: ACPI: PCI Interrupt 0000:02:00.0[A] -> GSI 17 (level, low) -> IRQ 177
Dec 29 12:39:54 anni kernel: ipw3945: Detected Intel PRO/Wireless 3945ABG Network Connection

which is better, but still I see no device with ifconfig -a. I ran the "regulatory daemon" with

/usr/local/sbin/ipw3945d --quiet

Then I saw in /var/log/messages

Dec 29 12:43:41 anni kernel: ipw3945: Detected geography ABG (11 802.11bg channels, 13 802.11a channels)
Dec 29 12:43:42 anni kernel: ADDRCONF(NETDEV_UP): eth1: link is not ready

I brought up my wireless link with ifconfig and iwconfig, and I can ping my AP! For repeatability, following the README, I added the following to /etc/modprobe.d/ipw3945 (note the filename and daemon path differ from the README):

install ipw3945 /sbin/modprobe --ignore-install ipw3945 ; sleep 1 ; /usr/local/sbin/ipw3945d --quiet
remove  ipw3945 /usr/local/sbin/ipw3945d --kill ; /sbin/modprobe -r --ignore-remove ipw3945

And now I can just ifup and ifdown the interface with my old config files.

I have noticed that if I suspend the laptop, either to memory or disc, I have to briefly flip the "wireless ariel" switch (LHS of case, about two-thirds of the way towards the back) off and then on, in order to wake up the transceiver circuitry. If I do this before using the network, the bluetooth and wireless lights come back on and service is restored without interruption.

BLUETOOTH

This just seems to work out the box. I turned on bluetooth on my Nokia 6230i and turned on the adaptor on the Asus. I did an hcitool scan on the Asus, and it found my 6230i.

Doing anything interesting was a bit more challenging. I think my 'phone was a bit confused, because nothing worked until I took the battery out and put it back. After that, I needed to:

Put the bluetooth manager tool into my gnome toolbar

Confirmed the 'phone was listening with

[root@anni bluetooth]# hcitool scan
Scanning ...
        00:13:70:AF:09:35       Tom's 6230

Established a connection to the 'phone with

[root@anni bluetooth]# hcitool cc 00:13:70:AF:09:35

Setup a mutual PIN with
```
[root@anni bluetooth]# hcitool auth 00:13:70:AF:09:35
```
which opened up a gnome dialogue asking me to enter a PIN. Once I'd done that, the 'phone asked me if I wanted to accept an incoming connection, and then prompted me for the same PIN. This sidestepped the whole I have only a few seconds to enter a PIN so I'll pick a crappy one problem I'd
earlier had.

Once the PIN was mutually established, I reconfigured the 'phone not to ask for confirmation before my laptop paired, setup my ppp files as detailed here, and established a GPRS connection (over which I am typing this).

MODEM / FIREWIRE

Nothing to report on these as yet.

LIVNA

I wanted pine, mplayer and a few other good things on the laptop, so I added the livna.org repository with
rpm -ivh http://rpm.livna.org/livna-release-6.rpm
and then installed pine with
yum install pine
and similarly, mplayer etc.

Listed on:

Back to Technotes index
Back to main page