
	HARTZLER.NET

♦ Home
  » Pumpkins!
  » DONC
  » Photographs
  » Yard Sail
  » Spain Tree
  » 1805-S274
♦ About This Site

Linux on a Toshiba Satellite 1805-S274

By Peter Hartzler
December 29, 2001

Disclaimer:

The information on this site is provided AS IS and without any guarantee.
USE THIS INFORMATION AT YOUR OWN RISK.

Summary:

I picked up a Toshiba Satellite 1805-S274 with the intent of setting it up running Linux. My usual distribution is RedHat, and RH 7.2 installed reasonably gracefully, happily blowing away Windows XP Home Edition.

Except for a few things, such as the winmodem and accelerated video support, and the much more distubing issue of occasionally failing to wake after an inactivity sleep, the machine works reasonably well with Linux.

I actually did not keep the machine, due to the nasty resume bug that these machines seem to have. However, since I did do a bit of digging to get as far as I did, and you may not have the luxury of a trade-in, I've journaled my adventures here.

Note: See the epilogue section below for updates as support for this machine develops.

Main Issues:

Resume Failures:

Does not always allow waking after the machine goes to sleep due to inactivity. I read various articles and postings from others with similar troubles, but was unable to find a solution. Playing with the power save settings doesn't seem to help, perhaps short of totally disabling power saving entirely. Since power saving features are key on a notebook, IMHO, I didn't bother trying to entirely disable idle sleep.

Finally, I'm actually putting this system together for someone else. Flakiness of this sort is just not acceptable in this situation. If it were for me, I might put up with this, and remember to write data often, just like when I use certain other platforms... Anyhow, I decided that this defect is a showstopper, and I returned the machine for a refund.

Clock Timer Bug?

Periodic syslog messages that read "probable hardware bug: clock timer configuration lost - probably a VIA686a motherboard." The next message is "probable hardware bug: restoring chip configuration." This is probably not a big deal, but it's bit unsettling.

Reboot:

"shutdown -r now" takes the machine down, but it doesn't come back up. "shutdown -h now" does power off the machine. Workaround: edited /etc/rc.d/init.d to coerce a reboot into a halt. Then you push the power button to bring it back up, of course. This would be a big deal on a server where remote administration is important and you might want to replace a kernel without driving to the NOC, but on a notebook it's really just an annoyance.

Video:

Rather slow in consoles, due to the primitive framebuffer video support. X is reasonable, but not really what I'd like it to be.

Here is a Super Microscopic Howto on getting frame buffer running. However, See comments and information from Chris Worley below, there is now a better way...

Winmodem:

Need I say more? Perhaps it can be made to work... Dunno.

Software License:

The machine has a statement which basically says that Toshiba will not refund one for unused components such as Windows. I guess they got tired of Linux/BSD/BeOS types asking... It would certainly be an uphill battle to get the satisfaction of a refund of the Microsoft Tax. It might be worthwhile to file a complaint all the same.

And now for the gory details:

Video:

The video chipset is not supported by the Trident X driver, but the frame buffer device seems to work fine. This driver isn't as snappy as an accelerated driver, but should serve for anything besides intensive 3d applications such as "video" games. Colossal Cave Adventure should be ok. ;-) I would not be surprised to see the video supported in a future version of X.

You'll need to tweak your /etc/lilo.conf to set up the frame buffer geometry. Then you'll want your /etc/X11/XF86Config-4 to pull up the fbdev driver. (As you might see in the X config file, I tried a few of the Trident chipset configs, but no joy. The file is a lot messier than it needs to be, of course.)

PCMCIA:

Works. Tested OK w/ a couple of different cards.

Modem:

Win-modem. Forget it, probably. A PCMCIA modem or an external serial modem should be fine.

The fscking modem manual weighs in at eighty eight pages and doesn't once mention the chipset. A clue on a screenshot suggests it's: Intr: 3, Addr: 0x2f8, UART: NS 16550AN which some web resources suggest is in fact a Lucent Winmodem. A quick attempt with some winmodem drivers gave no joy, but it is probably possible for the intrepid.

USB:

Works. My digital camera (usb-storage type) seems to work flawlessly. Now if I could just learn how to take decent photographs...

LAN:

Works: It's an Intel EtherExpress Pro 100; eepro100 driver.

CD-RW/DVD-ROM:

CD Read: Works.
CD Write: Works.
DVD Read: *shrug* Maybe after the legal idiocy is settled...

I don't know what a DVD playback would be like, and since I don't own any DVD media (and am planning on never owning a DVD as long as DVDs are being used as a tool to take away my rights), I can't easily play with that part. I read a rumor that there is a firmware hack to disable region encoding on this unit.

PRINTER (PARALLEL):

Works.

SOUND:

Play Audio: Works.
Sample: Tried briefly, but didn't get it going. It probably can be done, though I'm not sure if it'd be worth it.
MIDI: Tried briefly, no joy.

IRLAN:

Works.

I didn't see any sign of this being detected, but I didn't really try, since as far as I know I don't need it, and therefore don't have any easy way to test. However, see below, where I've appended email from someone who got it working and explains how you can too.

APM:

Despite things I've read on the 'net, APM seems to work. It does suspend and resume, and seems to know it's battery status. Apparently Toshiba went through a spell of using someone else's machines, simply putting their name on them. It seems as if this machine is after the end of this period, since it seems to play ball with the toshiba kernel module and the toshiba utilities.

I downloaded and built the Toshiba Linux Utilities from http://www.buzzard.org.uk/toshiba/ and these allow control of the power management stuff, including things like setting it so that closing the lid causes it to suspend.

By the way, the build failed on 'make install' because /usr/local/man and it's usual subdirectories were not present. Created those, then the install went through.

RANDOM POSSIBLY USEFUL STUFF:

You'll want to check out the Linux on Laptops pages, of course.
The output of lspci -vv
A sample dmesg -- in case this sheds any light.
Whenever I mention 'make install' I like to point out Izto's excellent CheckInstall program.

I hope this is of use to somebody. I hope that some day vendors will discover how easy it is to support open source systems, and how much it benefits them. I hope Toshiba sorts out their sleep-resume bug before it bites them like their diskette driver firmware bug (and attendant class action lawsuit) bit them.

Finally, since I no longer have the unit in my posession, I'm not likely to be able to offer any very direct assistance with setting up a similar configuration. However, if you feel so moved, you can find me at pete@hartzler.net .

Epilogue:

Here are some notes from people on the same path. Generally, if you do a significant writeup, or try different distributions, then the best idea is probably to create your own page and add it to the Linux on Laptops pages. This way we are all building a central resource. However, some folks simply wrote to say they'd had very similar experiences to mine, and perhaps supplying a small update item. What follows are the notebook specific parts of this feedback.

Date: Sun, 20 Jan 2002
From: John Carney

...
I got the dual boot [with Win-XP] to work by copying the first sector from
/boot to a floppy as bootsect.lnx, then changing the XP startup to have an
option to boot XP or linux.

Date: Fri, 1 Mar 2002
From: Chris Worley

... Was very helpful in setting up my 1805-S207 laptop.  I have all the same 
problems ;)

Your solutions helped out.  I did lower lilo's fb setting and the X 
configuration to 16 bit color, which sped up the display.  

One more problem to add: while hibernate/resume works, the USB device is 
dead upon resume (power is off to the USB).  Since I'm using a USB mouse 
that X is currently using, I can't reload the USB drivers to see if 
reinitializing them will turn the power back on.

Date: Sat, 6 Apr 2002
From: Karl Martindale-Vale

  I came across your article on your experience with Toshiba laptops.  
Something in there reminded of a story I heard a while ago:

"The machine has a statement which basically says that Toshiba will not
refund one for unused components such as Windows."

  There was a  guy who actually got his money back for the unused M$ OS 
on his Toshiba laptop.  If your interested you'll find his tale at:

http://www.netcraft.com.au/geoffrey/toshiba.html

Date: 07 Apr 2002
From: Chris Worley

An update for your page:

We no longer need the frame buffer driver...

I used the latest Toshiba kernel and X 4.0.2 modules at:

http://www.xfree86.org/~alanh/

Along with the 4.0.2 XFree RPMs from

http://www.silug.org/pub/redhat/linux/rawhide/i386/RedHat/RPMS/ :

XFree86-100dpi-fonts-4.2.0-6.52.i386.rpm
XFree86-ISO8859-15-75dpi-fonts-4.2.0-6.52.i386.rpm
XFree86-ISO8859-2-100dpi-fonts-4.2.0-6.52.i386.rpm
XFree86-ISO8859-2-75dpi-fonts-4.2.0-6.52.i386.rpm
XFree86-ISO8859-9-100dpi-fonts-4.2.0-6.52.i386.rpm
XFree86-ISO8859-9-75dpi-fonts-4.2.0-6.52.i386.rpm
XFree86-libs-4.2.0-6.52.i386.rpm
XFree86-tools-4.2.0-6.52.i386.rpm
XFree86-truetype-fonts-4.2.0-6.52.i386.rpm
XFree86-xf86cfg-4.2.0-6.52.i386.rpm
XFree86-xfs-4.2.0-6.52.i386.rpm
XFree86-Xnest-4.2.0-6.52.i386.rpm
XFree86-xtrap-clients-4.2.0-6.52.i386.rpm
XFree86-Xvfb-4.2.0-6.52.i386.rpm
glut-3.7-4.i386.rpm
glut-devel-3.7-4.i386.rpm
Glide3-20010520-13.i386.rpm
Glide3-devel-20010520-13.i386.rpm

I needed to install these with a "--nodeps" and a "--force", mostly
because I didn't add the kernel drivers, which were from Alan H's web
page.

My XF86Config file is attached.

Still no clue on the Winmodem, or how even to detect it.  I have tried
one 56k modem (3COM combined 56K modem + 10/100 LAN, model 3CCFEM556B),
and it did not work.  The modules nicely loaded at card insertion, and I
can talk to the modem via minicom or kppp, but it plays havoc on the
phone line.  Either the card is bad, the winmodem needs to be disabled
(but that's not a bios option), or it's tough to get a pcmcia modem
working on this laptop.

Date: Thu, 10 Jul 2003
From: Chris (hiszpans at ugcs dot caltech dot edu)

Hi Pete,

I read your tutorial about the Toshiba Satellite 1805-S274 on linux laptops.

Since I am too lazy to write my own page and feel that yours is good in all
other areas, I would appreciate it if you could add experiences with IrDA to
your Epilogue.

I did manage to get IrDA to work. I was able to receive packets from my
Seimens S40 cellphone using IrLAN. I was also able to use the remote control
from walmart (One for All 4060) to control xmms.

The trick is that the IrDA port is not powered on by the BIOS. It needs to be
manually powered on. A program to do this is available at
http://www.csai.unipa.it/peri/toshsat1800-irdasetup/

Using that page, I was able to get IrLAN to work with my cellphone.

Then, using the LIRC Project, I was able to get the tv remote to work to a good
extend (I still have to hold down the button on the remote for as long as 10
seconds sometimes for the signal to register. I compiled the lirc-sir module
using lirc-0.6.6. Then I used the Philips VCR5 remote file and set my
One-for-All 4640 to VCR mode, code 1181 (Philips). Apparently the toshsat1800
program is created for the smc-ircc module because I had a difficult time to
get it to register on start up. Finally, I came up with this script that I
put in /etc/rc.d/rc.irda in Slackware 9:

--- cut here ---
#!/bin/sh
# Start/stop infrared port

irda_start() {
  echo "Starting IrDA..."
  /sbin/setserial /dev/ttyS1 uart none port 0x2e8
  /usr/sbin/irsetup --sirbase=0x2e8 --firbase=0x130 --irq=7 --dma=3 --device=0x2e

  # this is needed -- don't know why
  /sbin/modprobe smc-ircc
  sleep 3
  /usr/sbin/irattach irda0 -s
  sleep 3
  ifconfig irda0 down
  /sbin/rmmod smc-ircc
  /sbin/rmmod irport
  /sbin/rmmod irda

  /usr/sbin/irattach /dev/lirc -s
}

irda_stop() {
  echo "Stopping IrDA..."
  /sbin/rmmod lirc_sir
}

case "$1" in
'start')
  irda_start
  ;;
'stop')
  irda_stop
  ;;
*)
  echo "usage $0 start|stop"
esac

--- cut here ---

I added /etc/rc.d/rc.irda start to /etc/rc.d/rc.M as well.

Feel free to email me or have others email me with questions about this, but I
ask that if you publish my email address, please either display it as
'hiszpans at ugcs dot caltech dot edu' or as a small gif so that I don't
receive spam.

Thank you for your tutorial.

Chris

Linux on a Toshiba Satellite 1805-S274

By Peter Hartzler December 29, 2001

Disclaimer:

Summary:

Main Issues:

Resume Failures:

Clock Timer Bug?

Reboot:

Video:

Winmodem:

Software License:

And now for the gory details:

Video:

PCMCIA:

Modem:

USB:

LAN:

CD-RW/DVD-ROM:

PRINTER (PARALLEL):

SOUND:

IRLAN:

APM:

RANDOM POSSIBLY USEFUL STUFF:

Epilogue:

By Peter Hartzler
December 29, 2001