Now, with 2.6.9 linux kernel, this problem has been corrected and new ACPI features has been added.
| CPU | Intel Pentium M 735 1.7 Ghz (Dothan) |
| System Memory | 512 MB DDR 333 Mhz |
| Video | Intel 855 GME 64 MB Shared Memory |
| Display | SXGA 1400x1050 |
| Hard disk | Hitachi 60 GB 4200 RPM (IC25N060ATMR04) |
| DVD | DVD+R/+RW 4x |
| Wired network | Intel 100 Mbps |
| Wireless network | Intel 2200 b/g 55 Mbps |
| Modem | Softmodem 56 Kbps v.92 |
To resize the partition I had used QtParted program included on System Rescue CD-Rom . I reduced the second partition to 19GB, then I had to reboot the system, boot with Windows and this correct some errors on it.
Then I boot with Fedora Core 2 DVD and I installed it without any problems (I listen that there is a bug on Fedora Core 2 that change the geometry of the hard disk and this makes that Windows can't boot, but I didn't have this trouble). Onto next sections I will explain the specific problems that I had to solve.
Fortunately, Alain Poirier write a C program, the 855resolution , that allow to modify the memory copy of BIOS changing a graphical that isn't used by the 1400x1050 mode. I have to note that Intel has released BIOS updates with this standard graphical mode, but Dell and most laptop companies (as Asus) has not integrated in his laptop BIOS.
Once I downloaded the program I compiled it and put it on "rc.sysinit" to start it before X system.
A month later, googling, I found a message in a forum about a person that had the same problem and that had correct it changing from the file speedstep-centrino.c of the linux kernel the ".x86mask" assigned for Dothan CPU. In this file the mask assigned is 1, but according to the information of /proc/cpuinfo the mask would be 6. This is my CPU info:
processor : 0
vendor_id : GenuineIntel
cpu family : 6
model : 13
model name : Intel(R) Pentium(R) M processor 1.70GHz
stepping : 6
cpu MHz : 1698.751
cache size : 64 KB
fdiv_bug : no
hlt_bug : no
f00f_bug : no
coma_bug : no
fpu : yes
fpu_exception : yes
cpuid level : 2
wp : yes
flags : fpu vme de pse tsc msr mce cx8 sep mtrr pge mca cmov pat clflush
dts acpi mmx fxsr sse sse2 ss tm pbe tm2 est
bogomips : 3375.10
Using 2.6.7 kernel sources I modified the "speedstep-centrino.c". This is the original code:
static const struct cpu_id cpu_id_dothan_a1 = {
.x86_vendor = X86_VENDOR_INTEL,
.x86 = 6,
.x86_model = 13,
.x86_mask = 1,
};
And the modified code:
static const struct cpu_id cpu_id_dothan_a1 = {
.x86_vendor = X86_VENDOR_INTEL,
.x86 = 6,
.x86_model = 13,
.x86_mask = 6,
};
Then I recompiled the linux kernel and the "speedstep-centrino" recognized my CPU and permits to throttle it's frequency, and therefore saves up battery and reduce temperature.
Now, with 2.6.9 linux kernel, this problem has been corrected and new ACPI features has been added.
First, when no native linux drivers exists I had to use the Windows driver with the NDISWrapper Project , a project that allows to use Windows wireless cards drivers.
First of all I downloaded the NDISWrapper sources, then I unpacked and compiled them. Next I gave the Windows drivers and I installed them with NDISWrapper, using the option parameter "-i" followed by the "inffile.inf" path. Unfortunately, with this driver my laptop freeze some times.
Native Linux Drivers
Recently, Intel has created the Intel PRO/Wireless 2200BG project to code an open source native linux driver. From 16th august of 2004, with the 0.4 driver version, the wireless card began to operate under linux. From that version to current, 0.13, a new firmware has been released and the 802.11g mode has been added.
This new driver works fine on my laptop, it doesn't freeze and, moreover, allow to contact with Access Points that under Windows don't work because an incompatibility with this Intel Wireless card. To install the module it's only needed to copy the latest firware to the /usr/lib/hotplug folder (I had to create because it wasn't by default), and to compile the module with a "make install" call. Next we can access to our wireless by the iw commands (iwconfig, iwlist...).
To install the drivers first I downloaded the source from the last link, then I decompressed them. This drivers have kernel 2.6 support and Alsa support. In order to activate it you must compile with:
make SUPPORT_ALSA=1
Then, in order to load the Alsa modem driver, I had to load the intel8x0m module:
modprobe intel8x0m
Without this module "slmodem" software didn't work because was unable to contact with modem device. Then it is only necessary to load the binary result of the compilation. I load everything I need to use my modem. To run it you have to indicate by parameters the alsa device for the modem, that in general is hw:1. This is the call I used to use:
./slmodem --alsa hw:1 &
This leaves the "slmodem" on background serving on /dev/ttySL0 as a normal hardware modem. The only needed thing to do is to configure the pppd daemon or the Fedora Network Device Control to use it. I have to note that while the modem is dialing up it hears no sound, but modem works well.
For the laptop mode I had to include the laptop-mode configuration file on the
/etc/sysconfig folder. Then I had created the control script
on /sbin folder.
About laptop-mode I had encountered a bit hard disk incompatibility related to the spin down timeout setted
with "hdparm" utility; it is a missfunction with the most Hitachi hard drives and has an easy solution. The
problem is that hard drive ignores spin down timeout and it doesn't work. To make it works it is only
needed to deactivate SMART offline tests with this command:
/usr/sbin/smartctl --smart=on --offlineauto=off /dev/hda
In order to do this every time that laptop-mode starts, I had modified laptop-mode supplied control script to add this line when it start, and other to stop it. Apart of this laptop-mode works fine.
Then I created the standard laptop events and his corresponding scripts on ACPI configuration folder, (/etc/acpi). An event for the "ac_adapter", "battery", "lid", "powerbtn" and "sleep" mode and his corresponding action scripts.
All things seem to go well, but William Hanwoody notice me that S3 mode (mode to suspend to RAM) didn't work well because the screen didn't wake up from sleep. Thanks to him and others the bug had been published on kernel bugzilla and they could get a solution. To solve it, it's only needed to run the VideoPost program, made by Venkatesh Pallipadi, after the computer wake up. To do this automatically you can patch the kernel or add a call on ACPI scripts that sleep your computer. I have used this last solution. My ACPI scripts could be downloaded from here .
Unfortunately, despite this ACPI tunning, by the moment my battery, while under Windows run 2h 30min, under Linux it only run for 1h 45 min.
Jordi Cucurull Juan, November 2004
