2007-02-18: Upgraded to a 2.6.18 kernel, with the kernel's own support for wifi and ALSA sound, and a patch to slow down the mouse.
2005-11-19: Upgraded to a 2.6.12.2 kernel. Also tried a 2.6.14 kernel, but suspend to disk failed. Moved my external back-up disk into a new enclosure. Documented my back-up script.
2005-06-14: Suspend to disk linked to lid closure, but occasionally fails. Upgraded to a 2.6.11.11 kernel.
2005-03-07: Suspend to disk and external video are working.
2005-01-17: Installed (Free) ipw2200 wifi driver and upgraded to a 2.6.10 kernel
2004-08-31: Driverloader won't compile under 2.6.8.1 kernel.
2004-08-24: (broken) Fn-F2 wifi key combination disabled.
2004-07-26: ndiswrapper failed, driverloader instead.
2004-07-10: The integrated Ethernet failed intermittently, but works again after a warranty motherboard replacement.
In April 2004, I decided my Gateway Solo notebook was ready for retirement, due to poor battery life and continuing problems with the PC card and internal 56K winmodem. However, I still liked its form factor: 12.1 inch screen and less than 4 pounds.
There are several machines available in the "ultralight" class. On the IBM X40, I didn't like the keyboard layout (I'm an emacs user, and I expect the left control key to be in the corner) or the lack of a touchpad. The Toshiba M205-S810 had a nice 1400x1050 display, but lacked a right hand control key, and I didn't think I would use the touch screen. The Sony V505EX had a small right hand shift key and very short key travel. The HP 4010nc had good specs (including disk drives up to 80 MB), but I couldn't find one to try out, and I was worried about the air inlet on the bottom. The Dell X300 had a nice keyboard, but short battery life according to cnet tests. I actually ordered a D400, but when I realized how wide it was I decided to get an X300 instead. (My wife took the D400 and is very happy with it.)
The X300 came with a 60 GB disk (Hitachi model IC25N060ATMR04-0) which hdparm describes as follows:
# hdparm -I /dev/hda
/dev/hda:
ATA device, with non-removable media
Model Number: IC25N060ATMR04-0
Serial Number: MRG357K3K13L6H
Firmware Revision: MO3OAD0A
Standards:
Used: ATA/ATAPI-6 T13 1410D revision 3a
Supported: 6 5 4 3
Configuration:
Logical max current
cylinders 16383 65535
heads 16 1
sectors/track 63 63
--
CHS current addressable sectors: 4128705
LBA user addressable sectors: 117210240
LBA48 user addressable sectors: 117210240
device size with M = 1024*1024: 57231 MBytes
device size with M = 1000*1000: 60011 MBytes (60 GB)
Capabilities:
LBA, IORDY(can be disabled)
bytes avail on r/w long: 4 Queue depth: 1
Standby timer values: spec'd by Vendor, no device specific minimum
R/W multiple sector transfer: Max = 16 Current = 0
Advanced power management level: 128 (0x80)
Recommended acoustic management value: 128, current value: 254
DMA: mdma0 mdma1 mdma2 udma0 udma1 udma2 udma3 udma4 *udma5
Cycle time: min=120ns recommended=120ns
PIO: pio0 pio1 pio2 pio3 pio4
Cycle time: no flow control=240ns IORDY flow control=120ns
Commands/features:
Enabled Supported:
* NOP cmd
* READ BUFFER cmd
* WRITE BUFFER cmd
* Host Protected Area feature set
* Look-ahead
* Write cache
* Power Management feature set
Security Mode feature set
* SMART feature set
* FLUSH CACHE EXT command
* Mandatory FLUSH CACHE command
* Device Configuration Overlay feature set
* 48-bit Address feature set
Automatic Acoustic Management feature set
* SET MAX security extension
Address Offset Reserved Area Boot
* SET FEATURES subcommand required to spinup after power up
Power-Up In Standby feature set
* Advanced Power Management feature set
* General Purpose Logging feature set
* SMART self-test
* SMART error logging
Security:
Master password revision code = 65534
supported
not enabled
not locked
not frozen
not expired: security count
not supported: enhanced erase
50min for SECURITY ERASE UNIT.
HW reset results:
CBLID- above Vih
Device num = 0 determined by the jumper
Checksum: correct
The disk apparently has a temperature sensor. After I added an
appropriate line to the configuration file, hddtemp can report the
disk temperature:
# grep IC25.*ATM /etc/hddtemp.db
"IC25N0[23468]0ATMR04" 194 C "Hitachi Travelstar 80GN"
# hddtemp /dev/hda
/dev/hda: IC25N060ATMR04-0: 35 C
The disk came with partitions of 31.3 MB (DELLUTILITY) and 57.192 GB with a FAT32 filesystem (per my order) with 4.5 GB used. (Serious bloatware - and I don't even have Office installed.) I installed Partition Magic and shrank the C: partition to 9.0052 GB.
I then booted a Debian 3.0 cdrom using Dell's USB cdrom drive.
cfdisk showed 50536 MB of free space, which I allocated as 98.71 MB unused (since I don't trust Windows to stay within its partition), 48899.29 MB for Linux, and 1538.13 MB for swap.
I initialized the swap partition and installed an ext2 filesystem in the root partition.
At this point the installer failed to recognize the cdrom (which was connected via USB), offering only the floppy, a partition on the hard disk, or an already mounted filesystem. I rebooted into Windows, copied /dists/woody/main/disks-i386/ onto the hard disk, and rebooted the install disk. That let me get a few steps further, but I still didn't have access to either the cdrom or the network. I had previously downloaded a copy of the Linux 2.4 driver module and saved it in c:\tmp. I could access it with the commands
mknod /dev/hda2 b 3 2
mkdir /c
mount -t vfat /dev/hda2 /c
insmod /c/tmp/bcm5700.o
However, this failed with many undefined variables.
Next I downloaded a "daily build" of the sarge Debian netinstall cdrom. Its default boot also failed to recognize either the cdrom or the network interface. Booting with "linux26" brought up 2.6.5 kernel which did recognize and mount the cdrom. However, it failed at the next step, logging a message "(none) user.notice hw-detect: Missing modules 'tg3 (BROADCOM Corporation NetXtreme BCM5705M Gigabit Ethernet)". Under Linux 2.6, the network interface is supported by the tg3 module. I compiled a module tg3.ko using Linux 2.6.6 sources on another machine and copied into c:\tmp. Installing it failed with the message
insmod: error inserting '/c/tmp/tg3.ko': -1 invalid module format
Repeating with tg3.o generated the same message.
Googling, I discovered that the tg3 driver had been intentionally removed from the Debian installation disk because of doubt as to its license.
I discovered an excellent listing of alternative Debian installation tools at http//www.linuxmafia.com/faq/Debian/installers.html, which included a link to Victor van Beekum's netinst image at http://www.beekum.nl/pe26xx.iso which claimed to include the tg3 driver. I downloaded that, burned (yet another) cd, and also wrote the cd files to the c: directory (since I didn't know that the USB disk would be accessible). I installed the base system that way. It turned out to have a 2.2.20 kernel, and had no support for the Ethernet card that I could find.
Next I compiled a 2.6.6 kernel on another machine with support for both the Ethernet and cdrom devices:
CONFIG_TIGON3=y
CONFIG_USB=y
CONFIG_USB_DEBUG=y
CONFIG_USB_DEVICEFS=y
CONFIG_USB_EHCI_HCD=y
CONFIG_USB_OHCI_HCD=y
CONFIG_USB_UHCI_HCD=y
CONFIG_USB_STORAGE=y
CONFIG_USB_STORAGE_DEBUG=y
I copied these to c: and booted the 2.2.20 kernel, copied the
kernel to /boot, added an entry to /etc/lilo.conf, ran lilo, and
tried to boot the new kernel. It displayed:
boot: linux-2.6.6
Loading linux-2.6.6............
Uncompressing Linux... Ok, booting the kernel.
and crashed.
Next I copied all eight Debian 3.0 installation disks to c:/tmp on the theory I could install enough to compile a kernel there. I completed the Woody installation that way.
Then a friend asked whether I had configured the 2.6.6 kernel for the right processor family - which of course I had not. I had started with the configuration for my dual Athlon desktop system, but had neglected to change the processor family. While I was at it, I enabled "Dell laptop support" (CONFIG_I8K=y) and disabled CONFIG_SMP. I installed that kernel as before, and this time it worked.
I set up networking manually with route and ifconfig, set up /etc/apt/sources.conf, did
apt-get update
apt-get upgrade
apt-get dist-upgrade
and had a working Linux system!
I had asked for gdm and X windows client and driver, but X wouldn't come up. After quite a few fruitless attempts at reconfiguration (including several copies of XF86Config posted to the net) I noticed that (in spite of the "apt-get dist-upgrade" mentioned above) the version of the server was still 4.1.something. I did an explicit "apt-get install xserver-xfree86" which got 4.3.0.dfsg.1-1, which came up immediately.
The X-300 has a VGA connector to drive an external CRT, but the Fn-F8 key combinations didn't work. I tried the i810switch program. It did enable the video signal, but the image was corrupted (something like a Moire pattern that crawled slowly up the screen) - certainly nothing I would want to use for a business presentation.
Following a pointer at http://www.cavecanen.org/index.shtml, I tried i855crt by Merello Andrea, which works great! The only anomaly was that the cursor disappears over certain windows (notably those for Word and PowerPoint when they're running under Crossover Office). I tried the "software cursor" option, but that left a white square in one corner of the display. (Morello recommends patching XFree86, and I don't have the appropriate patches for version 4.3 of XFree86, which is what I'm running.)
After using the integrated Ethernet for initial Linux installation (as described above), it started failing intermittently. The RJ-45 port has two lights. If the rear light is orange, the port is working. If it is green or off, it is not working. (The front light flashes yellow to show data traffic.) At first I ascribed this to my home-made cables. Sometimes flexing the cable or moving the machine would start the port working. However, the problem has persisted across four cables (all of which worked with other machines) and under both Linux and Windows. Eventually I found that I could restore connectivity by torquing the case: applying pressure down on the back left corner (just beyond the RJ-45 port) and the front left corner, and up in the center. (I have set up ifplugd and ifupdown-roam, so the port gets configured as soon as it comes up.) I guess there's a flaky solder joint or motherboard trace. I plan to back everything up, wipe the disk, and return the machine on warranty.
I talked to the "Technical Support" people at Dell. After spending a long time on the phone and answering some inane questions like "do you use static IP" and "do you use TCP/IP or some other protocol", they finally admitted I had a hardware problem. They sent a new motherboard to the outfit in my area who had the service contract. The next day they called to set up an appointment, and the day after that they came by and fixed the machine. I didn't have to restore from backups or anything. I'm happy with Dell's service!
I got an Intel PRO/Wireless 2200BG miniPCI 802.11 b/g adaptor. There is a project at http://ipw2200.sourceforge.net to develop a Linux driver for it, but it's apparently just getting started.
In the mean time, one can get a kernel module that can load NDIS (windows network driver API) drivers from ndiswrapper. The sources built fine, but "make install" generated a error message "depmod: QM_MODULES: Function not implemented". Documentation/Changes in the Linux sources advised getting version 0.9.10 or newer of module-init-tools, and to check using "depmod -V". Confusingly, the latter replied 2.4.26. It turns out I didn't have module-init-tools. I installed module-init-tools version 3.0-pre10-4 and ndiswrapper installed fine. When I tried to install the driver I got this complaint:
ndiswrapper: version magic '2.6.6 preempt PENTIUM4 gcc-3.3' should
be2.6.6 preempt PENTIUM4 gcc-2.95'
because my successful kernel was compiled by the Woody's gcc.
After recompiling with gcc 3.3.3, it installed fine. Here's the
installed driver list:
# ndiswrapper -l
Installed ndis drivers:
w22n51 present
Installing the driver seems to work:
# modprobe ndiswrapper
# dmesg|tail -3
ndiswrapper version 0.7 loaded
ndiswrapper adding w22n51.sys
wlan0: ndiswrapper ethernet device xx:xx:xx:xx:xx:xx using driver w22n51.sys
# ifconfig
eth0 Link encap:Ethernet HWaddr yy:yy:yy:yy:yy:yy
inet addr:192.168.1.102 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:18664 errors:0 dropped:0 overruns:0 frame:0
TX packets:14248 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:15156061 (14.4 MiB) TX bytes:2347787 (2.2 MiB)
Interrupt:11
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:4248 errors:0 dropped:0 overruns:0 frame:0
TX packets:4248 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:355848 (347.5 KiB) TX bytes:355848 (347.5 KiB)
wlan0 Link encap:Ethernet HWaddr xx:xx:xx:xx:xx:xx
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:4 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 b) TX bytes:1368 (1.3 KiB)
Interrupt:5 Memory:e0210000-e0210fff
(I've replaced the MAC addresses in all these printouts.)
Setting the mode generates an error message, which I have read is harmless:
# iwconfig wlan0 mode Managed
Error for wireless request "Set Mode" (8B06) :
SET failed on device wlan0 ; Invalid argument.
falcon:/usr/local/src/ndiswrapper#
However, the driver itself does not appear to work:
# iwconfig
eth0 no wireless extensions.
lo no wireless extensions.
wlan0 no wireless extensions.
# iwlist wlan0 scan
wlan0 Interface doesn't support scanning.
Eventually I found a discussion at www.minet.uni-jena.de/~dank/eigene/DELL_LATITUDE_X300.html
which (after translation from German to English, thanks to the Google
option) suggested adding the kernel configuration option
CONFIG_NET_RADIO=y. With that change, the driver does recognize the
adaptor:
# iwconfig wlan0
wlan0 IEEE 802.11b ESSID:""
Mode:Managed Access Point: FF:FF:FF:FF:FF:FF Bit Rate:54Mb/s
RTS thr:1600 B Fragment thr:2304 B
Encryption key:off
Power Management min timeout:0us mode:All packets received
Link Quality:0/100 Signal level:24/154 Noise level:0/154
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0
Then I found Debian packages of ndiswrapper by Erik Rigtorp at http://rigtorp.se/debian/unstable/
and installed version 0.8-1.
After must frustrating experimentation, I have settled on the following process:
# modprobe ndiswrapper
# iwconfig wlan0 power on
<press Fn-F2 twice>
# dhclient wlan0
Pressing the "Fn-F2" key generates this warning message:
atkbd.c: Unknown key pressed (translated set 2, code 0x88 on isa0060/serio0).
atkbd.c: Use 'setkeycodes e008 <keycode>' to make it known.
which appears to be harmless. After the first two keypresses,
"Fn-F2" appears to toggle the wifi radio. For example, if I start
pinging a site, I can use "Fn-F2" to interrupt and restart the
replies. However, I have found no indication of the current state of
the radio, either among the status lights along the front edge or
anywhere under /proc/acpi or /sys. I also have not found any way
to enable the radio from a script.
I'm using the driver that came with XP installed by Dell. The driver at the ndiswrapper web site is different, but I suppose it would work as well. Here are the md5sums for the two drivers:
4fed83668f087ecbe810ea90beceb765 /c/drivers/network/addon/w22n51.sys
4c009d4352849d79bf347846b6e03bfd WLAN_8.0.12.9000/w22n51.sys
Here is how iwconfig and iwlist report the wifi configuration:
# iwconfig wlan0
wlan0 IEEE 802.11g ESSID:"linksys"
Mode:Managed Frequency:2.412GHz Access Point: xx:xx:xx:xx:xx:xx
Bit Rate:54Mb/s
RTS thr:1600 B Fragment thr:2304 B
Power Management min timeout:0us mode:All packets received
Link Quality:0/100 Signal level:-43 dBm Noise level:-256 dBm
Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
Tx excessive retries:0 Invalid misc:0 Missed beacon:0
# iwlist wlan0 scan
wlan0 Scan completed :
Cell 01 - Address: xx:xx:xx:xx:xx:xx
ESSID:"foobar"
Protocol:IEEE 802.11g
Mode:Managed
Frequency:2.412GHz (channel 1)
Quality:0/100 Signal level:-48 dBm Noise level:-256 dBm
Encryption key:off
Bit Rate:1Mb/s
Bit Rate:2Mb/s
Bit Rate:5.5Mb/s
Bit Rate:11Mb/s
Bit Rate:18Mb/s
Bit Rate:24Mb/s
Bit Rate:36Mb/s
Bit Rate:54Mb/s
Bit Rate:6Mb/s
Bit Rate:9Mb/s
Bit Rate:12Mb/s
Bit Rate:48Mb/s
# iwlist wlan0 power
wlan0 Current mode:All packets received
min timeout:0us
# iwlist wlan0 txpower
wlan0 unknown transmit-power information.
# iwlist wlan0 freq
wlan0 14 channels in total; available frequencies :
Channel 01 : 2.412 GHz
Channel 02 : 2.417 GHz
Channel 03 : 2.422 GHz
Channel 04 : 2.427 GHz
Channel 05 : 2.432 GHz
Channel 06 : 2.437 GHz
Channel 07 : 2.442 GHz
Channel 08 : 2.447 GHz
Channel 09 : 2.452 GHz
Channel 10 : 2.457 GHz
Channel 11 : 2.462 GHz
Channel 12 : 2.467 GHz
Channel 13 : 2.472 GHz
Channel 14 : 2.484 GHz
Current Frequency:2.412GHz (channel 01)
I tried "iwconfig wlan0 mode Ad-Hoc", but the reported mode stayed at "Managed".
2004-07-26 update: ndiswrapper quit working. The command "modprobe ndiswrapper" stops just after printing "w22n51.sys", apparently going into a infinite loop. "top" running on another virtual console shows a steadily increasing load. None of the processes listed there is using a significant amount of CPU, so the load must be in kernel tasks. I don't know why it broke. I suppose the driver may be running out of stack space (although I do not have "CONFIG_4KSTACKS" configured in the kernel).
I have since switched to "driverloader" from Linuxant. They offer a .deb file. After installation (using the same Windows driver) the wireless adaptor showed up as eth1. Pressing "Fn-F2" twice is still necessary. "driverloader" requires a $20 license which is keyed (naturally enough) to the MAC address of the wifi adaptor.
2004-08-24 update: I discovered an option in the BIOS setup program to disable the Fn-F2 key. Now I can start up wifi with just "iwconfig eth1 power on ; dhclient eth1".
2004-08-31 update: driverloader won't compile under the 2.6.8.1 kernel due to an undefined symbol "waitpid". That's another reason to switch to the free ipw2200 driver.
2005-01-17 update: I installed version 0.19 of the ipw2200 driver, as follows:
+CONFIG_FW_LOADER=y +CONFIG_CRYPTO=y +CONFIG_CRYPTO_ARC4=y +CONFIG_CRYPTO_CRC32C=y(While I was at it, I also made several other changes.)
The X300 BIOS does not implement APM, so I have to use ACPI with this machine.
With ACPI configured, the /proc/acpi/battery directory was empty, and the boot messages included errors like this:
evregion-0347: *** Error: Handler for [EmbeddedControl] returned AE_BAD_PARAMETER
dswexec-0435 [42] ds_exec_end_op : [LEqual]: Could not resolve operands, AE_BAD_PARAMETER
dswstate-0273 [44] ds_result_pop_from_bot: No result objects! State=e7477828
dsutils-0526 [44] ds_create_operand : Missing or null operand, AE_AML_NO_RETURN_VALUE
psparse-1133: *** Error: Method execution failed [\_TZ_.THRM._TMP] (Node c152dc38), AE_AML_NO_RETURN_VALUE
Following the method of Philipp Matthias Hahn described at http://acpi.sourceforge.net/wiki/index.php/HowToOverrideTable
I copied the BIOS generated DSDT from /proc/acpi/dsdt and
decompiled it with the Intel program iasl. I then compared it
with the X300 dsdt files from http://acpi.sourceforge.net/dsdt/tables/Dell/Latitude_X300/.
Some differences seemed to relate to the amount of memory
installed (I have 640 MB). The fixes can be summarized in this patch against the
disassembled dsdt.dsl file. I assembled and installed it like
this:
iasl -tc revised-dsdt.dsl
cp revised-dsdt.hex /usr/local/linux/linux-2.6/drivers/acpi/dsdt.hex
I then made the changes in this patch to /usr/local/linux/linux-2.6/drivers/acpi/osl.c, then compiled and installed the kernel. After booting, I found this power and thermal information available:
# head ac_adapter/*/* battery/*/* thermal_zone/THRM/*
==> ac_adapter/ADP1/state <==
state: off-line
==> battery/BAT1/alarm <==
alarm: unsupported
==> battery/BAT1/info <==
present: yes
design capacity: 1900 mAh
last full capacity: 1900 mAh
battery technology: rechargeable
design voltage: 14800 mV
design capacity warning: 190 mAh
design capacity low: 57 mAh
capacity granularity 1: 1 mAh
capacity granularity 2: 1 mAh
model number:
==> battery/BAT1/state <==
present: yes
capacity state: ok
charging state: discharging
present rate: 711 mA
remaining capacity: 1655 mAh
present voltage: 15945 mV
==> battery/BAT2/alarm <==
alarm: unsupported
==> battery/BAT2/info <==
present: yes
design capacity: unknown
last full capacity: unknown
battery technology: rechargeable
design voltage: unknown
design capacity warning: 10 mAh
design capacity low: 3 mAh
capacity granularity 1: 1 mAh
capacity granularity 2: 1 mAh
model number:
==> battery/BAT2/state <==
present: yes
capacity state: ok
charging state: charged
present rate: unknown
remaining capacity: unknown
present voltage: unknown
==> thermal_zone/THRM/cooling_mode <==
cooling mode: active
==> thermal_zone/THRM/polling_frequency <==
<polling disabled>
==> thermal_zone/THRM/state <==
state: ok
==> thermal_zone/THRM/temperature <==
temperature: 40 C
==> thermal_zone/THRM/trip_points <==
critical (S5): 84 C
active[0]: 65 C: devices=0xc152deb8
The only problem I see is that battery 2 is marked as "present",
when there is in fact only one battery installed (one of the
lower-capacity ones).
Some of the "function" keys work:
2004-06-26 update: The kernel reports a total of 18 errors (AE_BAD_PARAMETERS or AE_AML_NO_RETURN_VALUE). I downloaded and installed the A07 BIOS update file, R79464.exe, from http://support.dell.com. However, the kernel still reports 18 errors, so they didn't fix the DSDT.
The X300 offers several CPU clock frequencies, as described here:
jrv@falcon:/sys/devices/system/cpu/cpu0/cpufreq$ head * head: cannot open `cpuinfo_cur_freq' for reading: Permission denied ==> cpuinfo_max_freq <== 1200000 ==> cpuinfo_min_freq <== 600000 ==> scaling_available_frequencies <== 1200000 1100000 1000000 900000 800000 600000 ==> scaling_available_governors <== powersave userspace performance ==> scaling_cur_freq <== 600000 ==> scaling_driver <== centrino ==> scaling_governor <== powersave ==> scaling_max_freq <== 1200000 ==> scaling_min_freq <== 600000
I wish the frequencies went far lower, so I could extend my battery life. If I use emacs or vi in a virtual console, I would expect it to keep up with my typing at 100 MHz or lower. On the other hand, according to the discharge current reported in /proc/acpi/battery/BAT1/state, cutting the frequency in half saves only about five percent. I suppose the display backlight, RAM, and hard disk make up the rest. Come to think of it, I'd like to be able to turn the backlight off while I'm just taking notes. The "Fn-CRT/LCD" button appears to be disabled if there is no CRT connected.
For comparison, turning on the wifi radio increases discharge current by about ten percent.
After updating to the 2.6.10 kernel, its implementation of suspend to disk started working. To get the Fn-ESC key combination to work, I installed this script in /etc/acpi/events/sleep:
event=button[/]sleep
action=/etc/acpi/actions/hibernate %e
This is the file /etc/acpi/actions/hibernate:
#!/bin/sh
# record battery state
logger -- `power`
# surrender DHCP lease,
# per http://www.x1000forums.com/index.php?showtopic=5405
dhclient -r eth0
dhclient -r eth1
# stop network interfaces, unload modules
/etc/init.d/networking stop
rmmod ipw2200
# we don't want to keep stale DNS information around
/sbin/resolvconf -d eth0
/sbin/resolvconf -d eth1
chvt 1 # change to a virtual terminal
# mysql daemon doesn't survive suspension
/etc/init.d/mysql stop
# suspend. The first line is essential!
echo -n shutdown > /sys/power/disk
echo -n disk > /sys/power/state
###################################################################
# After the suspend, we restart here...
# reset the system clock from the hardware clock
hwclock --hctosys
# Matlab's license manager dies when the system clock is reset
/etc/init.d/flexlm restart
# restart network interfaces.
modprobe ipw2200 #ifname=eth%d
/etc/init.d/networking start
/etc/init.d/mysql start
# record battery state
logger -- `power`
# start anacron, otherwise cron jobs can be never restarted
/etc/init.d/anacron start
Recent changes:
event=button[ /]lid
action=/etc/acpi/actions/hibernate
Now, I can suspend to disk by just closing the lid. This is a
great convenience. Before, I could use Fn-Esc to start
suspending, but closing the lid seemed to interfere with
suspending.Suspending has failed a few times, with the machine displaying "Freeing memory...done (XX pages freed)" then locking up. I don't believe I have seen that problem since upgrading to the 2.6.11 kernel.
More recently, I have encountered the following symptoms:
The X300 has the Intel 810 audio chipset.
2007-02-18 update: I have configured ALSA
with OSS emulation, like this:
The modem is of course a
Winmodem.
scanmodem from http://linmodems.technion.ac.il/ reports:
I checked Linuxant for
drivers. I had to patch their "listmodem" script to get it to run,
but it wasn't very informative:
CONFIG_SOUND=y
#
# Advanced Linux Sound Architecture
#
CONFIG_SND=m
CONFIG_SND_TIMER=m
CONFIG_SND_PCM=m
CONFIG_SND_RAWMIDI=m
CONFIG_SND_SEQUENCER=m
CONFIG_SND_OSSEMUL=y
CONFIG_SND_MIXER_OSS=m
CONFIG_SND_PCM_OSS=m
CONFIG_SND_PCM_OSS_PLUGINS=y
CONFIG_SND_SEQUENCER_OSS=y
CONFIG_SND_RTCTIMER=m
CONFIG_SND_SEQ_RTCTIMER_DEFAULT=y
CONFIG_SND_SUPPORT_OLD_API=y
CONFIG_SND_VERBOSE_PROCFS=y
CONFIG_SND_MPU401_UART=m
CONFIG_SND_AC97_CODEC=m
CONFIG_SND_AC97_BUS=m
CONFIG_SND_VIRMIDI=m
CONFIG_SND_MTPAV=m
CONFIG_SND_SERIAL_U16550=m
CONFIG_SND_MPU401=m
CONFIG_SND_INTEL8X0=m
Here is the audio configuration as reported by /dev/sdnstat:
$ cat /dev/sndstat
Sound Driver:3.8.1a-980706 (ALSA v1.0.12rc1 emulation code)
Kernel: Linux falcon 2.6.18 #21 PREEMPT Wed Oct 11 14:00:11 EDT 2006 i686
Config options: 0
Installed drivers:
Type 10: ALSA emulation
Card config:
Intel 82801DB-ICH4 with STAC9750,51 at 0xe0100c00, irq 10
Audio devices:
0: Intel 82801DB-ICH4 (DUPLEX)
Synth devices: NOT ENABLED IN CONFIG
Midi devices: NOT ENABLED IN CONFIG
Timers:
7: system timer
Mixers:
0: SigmaTel STAC9750,51
56K Modem
Providing detail for device at PCI_bus 0000:00:1f.6
with vendor-ID:device-ID
----:----
Class 0703: 8086:24c6 Modem: Intel Corp. 82801DB (ICH4) AC'97 Modem Controller (rev 01) (prog-if 00 [Generic])
SubSystem 14f1:5422 Conexant: Unknown device 5422
Flags: medium devsel, IRQ 10
I/O ports at 2400
I/O ports at 2000 [size=128]
The soft modem Subsystem operates under a controller
8086:24c6 82801DB ICH4 with Subsystem chipset POSSIBLY from:
Broadcom
AgereSystems
Conexant
Intel
Smartlink
# diff -u listmodem.sh-orig listmodem.sh
--- listmodem.sh-orig 2004-06-05 07:47:05.000000000 -0400
+++ listmodem.sh 2004-06-05 07:47:30.000000000 -0400
@@ -123,7 +123,7 @@
if ( $1 ~ /^00:0/ )
{
if (/\(rev [a-zA-Z0-9]+\)$/)
- revision = toupper (gsub(")", "", $NF))
+ revision = toupper (gsub("\)", "", $NF))
else
revision = "NONE"
next
# sh listmodem.sh
=============================================================
= RESULT OF MODEM QUERY =
=============================================================
Also device ID 5422 does not appear in their listmodem
guide.
I downloaded their HCF driver
hcfpcimodem_1.01lnxt04051300full_i386.deb, but it reported "ERROR:
no device detected by hcfpci driver". However, their HSF
modem Debian package installed without complaint:
$ su -c 'dpkg -i hsfmodem_6.03.00lnxt04051300full_i386.deb '
Password:
Selecting previously deselected package hsfmodem.
(Reading database ... 96984 files and directories currently installed.)
Unpacking hsfmodem (from hsfmodem_6.03.00lnxt04051300full_i386.deb) ...
Setting up hsfmodem (6.03.00lnxt04051300full) ...
Conexant HSF softmodem driver, version 6.03.00lnxt04051300full
If you need license keys, assistance or more information, please go to:
http://www.linuxant.com/
When reporting a problem for the first time, please send
us the file generated by "hsfconfig --dumpdiag".
No pre-built modules for: Debian-testing/unstable linux-2.6.6 i686
Trying to automatically build the driver modules...
(this requires a C compiler and proper kernel sources to be installed)
Where is the linux source build directory that matches your running kernel?
[/lib/modules/2.6.6/build]
Building modules for kernel 2.6.6, using source directory
/lib/modules/2.6.6/build. Please wait...
done.
error: 'kernel.modprobe' is an unknown key
Automatically guessed region (using timezone): "USA"
Please enter region name for modem unit 0 [USA]:
Setting region for modem unit 0: "USA"
To change, use "hsfconfig --region" or "AT+GCI=<T35code>"
The current region can be displayed by entering "ATI9" in a terminal program.
Note: we respect user privacy. Email addresses are not communicated
nor used for any purpose other than to manage licenses!
Please enter your email address [unknown]: jrvz@NOSPAMcomcast.net
License keys can be obtained from http://www.linuxant.com/
Without one, the modem operates in FREE mode (max 14.4Kbps data only, no fax)
The registration ID for modem unit 0 is: 0837-0D7C-D8C0
Please enter license key [FREE]:
Setting license for modem unit 0: "jrvz@NOSPAMcomcast.net/FREE"
Current parameters: ("hsfconfig --info")
Config for modem unit 0: /dev/ttySHSF0
Device instance: 0-PCI-8086:24c6-14f1:5422
HW revision : CXT22
HW profile name: hsfmc97ich
Registration ID: 0837-0D7C-D8C0
License owner : jrvz@NOSPAMcomcast.net
License key : FREE
License status : FREE (max 14.4kbps data only)
Current region : USA (T.35 code: 00B5)
The /dev/modem alias (symlink) points to /dev/ttySHSF0
To enable full 56K modem and FAX functionality, enter your license information
with "hsfconfig --license".
License owner and key data must EXACTLY match the information respectively
provided to and by Linuxant. Otherwise, license status will remain "FREE"!
jrv@falcon:~/hardware/x300/modem$ lsmod
Module Size Used by
hsfmc97ali 51476 0
hsfmc97via 49304 0
hsfmc97ich 50832 0
hsfpcibasic2 56720 0
hsfserial 20996 4 hsfmc97ali,hsfmc97via,hsfmc97ich,hsfpcibasic2
hsfengine 1218980 1 hsfserial
hsfosspec 64996 9 hsfmc97ali,hsfmc97via,hsfmc97ich,hsfpcibasic2,hsfserial,hsfengine
hsfsoar 52480 4 hsfmc97ali,hsfmc97via,hsfmc97ich,hsfpcibasic2
hcfpcihw 9728 0
hcfpciserial 20868 1 hcfpcihw
hcfpciengine 800052 1 hcfpciserial
hcfpciosspec 56748 2 hcfpciserial,hcfpciengine
driverloader 157572 0
Actually /dev/modem didn't exist. I set it up to point to /dev/ttySHSF0, and it worked!
(All the hotel rooms come with wifi or hardwired Ethernet nowadays, so I have not used the modem for months.)
I downloaded an XFree86 driver for the touchpad from http://w1.894.telia.com/~u89404340/touchpad/, and it now does tap-to-click.
I also plan to investigate Dmitry Torokhov's modifications to gpm to handle the Synaptics touchpad (e.g. tap-to-click in a virtual console). Unfortunately his patches collide with the Debian patches. I have not reconciled them.
I picked up a Targus AMW07 wireless optical mouse which works great, but was far too sensitive for me. To make it usable, I have applied my own slow-mouse-patch which adds two configuration parameters:
I got an external DVD/CD-RW drive with my X300. To use it, I configured my kernel with "CONFIG_BLK_DEV_SR=y", created the symbolic link
$ ls -l /dev/cdrom
lrwxr-xr-x 1 root root 3 2004-07-03 12:44 /dev/cdrom -> sr0
and included this line in /etc/fstab:
/dev/sr0 /cdrom iso9660 ro,user,noauto 0 0
I can mount and read CDROMs and DVDs and write CDRs. However, I can't write DVDs. It's not that kind of drive. dvdrecord reports:
falcon:/debian/images# dvdrecord -blank=fast -v speed=4 dev=0,0,0 test.iso
dvdrtools v0.1.5
Portions (c) 2002-2003 Ark Linux <bero@arklinux.org>
dvdrtools patched by Julien Delange for Debian
Note that you have to reports bug to the debian maintainer
Please ask Debian for support, and not ask the original author
This program is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free Software
Foundation; either version 2, or (at your option) any later version.
This program is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR
A PARTICULAR PURPOSE. See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License along with
this program; see the file COPYING. If not, write to the Free Software
Foundation, 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA.
Based on:
Cdrecord 1.11a15 (i386-pc-linux-gnu) Copyright (C) 1995-2001 Jörg Schilling
TOC Type: 1 = CD-ROM
scsidev: '0,0,0'
scsibus: 0 target: 0 lun: 0
Linux sg driver version: 3.5.32
Using libscg version 'schily-0.7'
atapi: 1
Device type : Removable CD-ROM
Version : 0
Response Format: 1
Vendor_info : 'SAMSUNG '
Identifikation : 'CDRW/DVD SN-324F'
Revision : 'U203'
Device seems to be: Generic mmc2 DVD-ROM.
resid: 118
Using generic SCSI-3/mmc CD-R driver (mmc_cdr).
Driver flags : SWABAUDIO BURNFREE
Supported modes: TAO PACKET SAO SAO/R96P SAO/R96R RAW/R16 RAW/R96P RAW/R96R
Drive buf size : 1959936 = 1914 KB
FIFO size : 4194304 = 4096 KB
Track 01: data 20 MB
Total size: 24 MB (02:22.68) = 10701 sectors
Lout start: 24 MB (02:24/51) = 10701 sectors
dvdrecord: Input/output error. test unit ready: scsi sendcmd: no error
CDB: 00 00 00 00 00 00
status: 0x2 (CHECK CONDITION)
Sense Bytes: F0 00 02 00 00 00 00 0A 00 00 00 00 3A 01 00 00
Sense Key: 0x2 Not Ready, Segment 0
Sense Code: 0x3A Qual 0x01 (medium not present - tray closed) Fru 0x0
Sense flags: Blk 0 (valid)
cmd finished after 0.009s timeout 40s
dvdrecord: No disk / Wrong disk!
Of course, I had loaded a disk and closed the tray. The
disk was the Memorex DVD+RW that came packed with the drive - now
why do you suppose they packed it with a disk it can't write?
-CONFIG_MPENTIUM4=y +CONFIG_MPENTIUMM=y +CONFIG_ACPI_CUSTOM_DSDT=y +CONFIG_ACPI_CUSTOM_DSDT_FILE="/boot/dsdt.hex" +CONFIG_ACPI_VIDEO=y +CONFIG_CPU_FREQ_GOV_ONDEMAND=y +CONFIG_PCCARD=y +CONFIG_PNPACPI=y +CONFIG_CDROM_PKTCDVD=y +CONFIG_CDROM_PKTCDVD_BUFFERS=8 +CONFIG_LEGACY_PTYS=y +CONFIG_LEGACY_PTY_COUNT=256 +CONFIG_USB_STORAGE_RW_DETECT=y +CONFIG_MMC=y +CONFIG_MMC_BLOCK=y +CONFIG_MMC_WBSD=yI don't know why I originally got the CPU type wrong.
The stock kernel now has provisions for a custom DSDT. I put it in /boot along with the other boot-time files.
I made the last three changes in hopes of getting the internal MMC/SD card reader working, using the driver for Winbond's W83L518D and W83L519D SD/MMC card readers in the 2.6.10 kernel. No success there, though. According to the project page http://projects.drzeus.cx/wbsd/, the driver should display a message if it recognizes the chip. However, when I load the wbsd module all I get is:
wbsd: Winbond W83L51xD SD/MMC card interface driver, 1.0 wbsd: Copyright(c) Pierre OssmanAlso, the Windows didn't include the file wbsd.sys. Apparently Dell used some other chip for the SD/MMC slot.
2005-11-19 update: I've upgraded to a 2.6.12.2 kernel. I tried a 2.6.14 kernel, but software suspend failed. I've read there were problems with USB support, and assume that was the culprit.
2007-02-18 update: I've upgraded to a 2.6.18 kernel. Here is a 2.6.18 kernel configuration file and the resulting kernel image. Note that this kernel has the above mentioned DSDT compiled in, so it's customized for 640 MB of system memory. It has drivers compiled in for the Ethernet (CONFIG_TIGON3), video, and USB flash memory and CDROM (CONFIG_USB_STORAGE and CONFIG_BLK_DEV_SR to provide /dev/sr0). This kernel also has swsusp configured and working. I've added support for IEEE 1394 (FireWire), 2200BG miniPCI wifi using the version of ipw2200 now shipped with the kernel, LEDs (though I have not found anything that uses them), ALSA sound, and tunneling (for setting up a Virtual Private Network).
2005-11-19 update: I have been backing this machine up to a Lacie external USB drive, which worked fine but got pretty hot. I just moved the drive to a Creative I/O aluminum enclosure with fan that I got from newegg.com. The first one they sent was DOA (no +5 power). I RMA'd it, and the second one works fine. However, there were two errors in the documentation:
#!/bin/bash
# ----------------------------------------------------------------------
# adapted from Mike Rubel's handy rotating-filesystem-snapshot utility
# http://www.mikerubel.org/computers/rsync_snapshots/
# ----------------------------------------------------------------------
PATH=/bin:/usr/bin; unalias -a; set -e # avoid surprises
# ------------- file locations -----------------------------------------
SELF=$(basename $0)
BACKUP_DIR=/backup;
HOST=`uname -n`
d=$BACKUP_DIR/$HOST/snapshot
EXCLUDES=/usr/local/etc/backup_exclude;
# ------------- rotate the backups -------------------------------------
# The parameters are the suffixes of the directory names to be
# included in the rotation, with the oldest mentioned first.
# "rotate 2 1 0" does $d.1 -> $d.2 then $d.0 -> $d.1.
rotate(){
logger -s -t $SELF "rotating $d.0 through $d.$1"
rm -rf $d.$1
while [ -d $d.$2 ]; do
mv $d.$2 $d.$1
shift
done
}
# ------------- the script itself --------------------------------------
# make sure we're running as root
if (( `id -u` != 0 )); then {
echo "$0: Sorry, must be root.";
exit 1;
} fi
# attempt to (re)mount the backup filesystem as RW; else abort
if grep -q $BACKUP_DIR /etc/fstab; then # directory is mountable
if mount | grep -q $BACKUP_DIR; then # directory is mounted
opt=remount,rw
else opt=rw
fi
if ! mount -o $opt $BACKUP_DIR; then
logger -s -t $SELF "could not (re)mount $BACKUP_DIR readwrite.";
exit 1;
fi
fi
# check whether a backup is in progress
if [ -f /var/lock/$SELF ]; then
logger -s -t $SELF "another backup is already in progress."
exit 1;
fi
echo $$ > /var/lock/$SELF # /var/lock is cleaned out on reboot
START=`date +%s`
logger -s "$SELF starts"
# ensure directories exist, to avoid errors in the following
for x in 1 2 3 4 5; do
if [ ! -d $d.$x ]; then mkdir -p $d.$x; touch -d 1970-01-01 $d.$x; fi
done
# Maintain geometrically spaced backups even if backup script is
# called at irregular times (e.g. for a removable drive). (This also
# makes the backup script idempotent, so calling it twice in succession
# only updates the current backup.)
now=`date +%s`
if [ $((($now-`date -r $d.5 +%s`)/86400)) -gt 243 ]; then rotate 5 4 3 2 1 0;
elif [ $((($now-`date -r $d.4 +%s`)/86400)) -gt 81 ]; then rotate 4 3 2 1 0;
elif [ $((($now-`date -r $d.3 +%s`)/86400)) -gt 27 ]; then rotate 3 2 1 0;
elif [ $((($now-`date -r $d.2 +%s`)/86400)) -gt 9 ]; then rotate 2 1 0;
elif [ $((($now-`date -r $d.1 +%s`)/86400)) -gt 3 ]; then rotate 1 0;
else true;
fi
# rsync from the system into the new snapshot, making hardlinks to
# files in the previous snapshot when unchanged
rsync --verbose --archive --delete --delete-excluded --sparse \
--exclude-from="$EXCLUDES" \
--link-dest=$d.1 \
/c/proj /c/Documents\ and\ Settings /c/tmp /c/pub \
/bin /boot /etc /home /initrd /lib /opt /root /sbin /usr /var \
$d.0 || logger -s -t $SELF "rsync returns $?"
rm -f $d.0/TIMESTAMP
touch $d.0/TIMESTAMP # also updates mtime of $d.0
ls -litd $d.0 > /etc/BACKUP-LOCAL
rm -f /var/lock/$SELF
echo "disk space usage on backup partition:"; df /backup
# remount the snapshot mountpoint as readonly
if grep -q $BACKUP_DIR /etc/fstab; then # directory is mountable
if grep -q "$BACKUP_DIR.*noauto" /etc/fstab; then # dir usually unmounted
umount $BACKUP_DIR || \
logger -s -t $SELF "could not umount $BACKUP_DIR.";
else
mount -o remount,ro $BACKUP_DIR || \
logger -s -t $SELF "could not remount $BACKUP_DIR readonly.";
fi
fi
END=`date +%s`
logger -s -t $SELF "finished in $((($END-$START)/60)) min"
DISCLAIMER: I have no connection to any of the companies mentioned here, other than as a customer. For official information, see the official sites.
|
Last modified: 2008-01-19 |
|
|
|
