Linux on a Toshiba Satellite M50-130
Linux on a Toshiba Satellite M50-130 (sold in UK)

This is a quick note list written for laptop websites, it's not meant to be complete, secure or very detailed. 
Note that I use debian kernel sources and a slightly modified debian .config(setting only the pentium-m processor and scheduler actually) but the idea is (in case you use another distribution or kernel) that it is a very modular kernel (with nearly all modules already compiled to not have to recompile it every time a new module is needed). 
If you don't manually compile your kernels, you're probably using a similarly modular one provided by a distribution.

It turns out at least on that debian system using 2.6,  /etc/mkinitrd/modules needs to have these lines for initrd to properly make the kernel recognize the dvdrw and boot properly (because of a kernel issue at least on 2.6.12)


and then initrd has to be remade with

mkinitrd -o /boot/initrd.img-2.6.12 (yaird could also be used but follows a different procedure)

Alternatively, a kernel could have ide and scsi/sata support built in but it's a weird and blury field at the moment since the sata kernel support is still under heavy development. So, some googling would always be helpful in that sata disk+dvdrw recognition matter.

linux 2.6.14, 2/Jan/06:  At least on current debian's linux 2.6.14, the above initrd fix wouldn't do it. Instead,

libata atapi_enabled=1

..on initrd's modules file as described above did the trick, because atapi_enabled is used to turn on libata's support of atapi devices. It is considered experimental for that kernel, but it seems to work ok for this laptop.

2.6.15, Feb/06: no different treatment, disk is sda and dvdrw is now shown as scd0.

The latest binary driver from and the auto generated conf of fglrxconfig should do the trick but, keep in mind the following:

The ChipID may not be recognized, it wasn't here, so the server crashed for not finding a chipset. So, if that happens you must add..

ChipID 0x3150 the 'device' section of the xserver's conf, which corresponds to the x600 mobility chipset. Keep in mind, other chipIDs (e.g. the one for the radeon x600 for desktop (not mobility)), could run fine and accelerated but you may have to add a modeline in the monitor section and the non-native resolutions may not stretch to fit the screen.

The second thing is, the native resolution of the laptop is 1280x768, so, you have to add that resolution in the 'screen' section.

Those two additions were are the only ones needed to have proper display and accelerated 3d graphics (apart from any keyboard/mouse settings you may need to add).

The ati driver on 2.6.14 would need at the time of writing (31/10/05) a fix found here but that won't be needed on later driver releases so, check that out.

fglrx 8.20.8, Jan/06:'s 8.20.8 driver on current debian's xorg, seems to not require the above ChipID line. The chip is recognized normally. That makes life easier since we only have to include the resolution for a basic proper setup.
Feb/06, fglrx 8.21.7, 8.22.5: as expected, same behaviour as 8.20.8. If you need to kill the xserver often and that crashes the system, try to run gdm/xdm/kdm with it configured to restart the server after every logout. e.g. with "TerminateServer=true" in /etc/kde3/kdm/kdmrc in the case of kdm.

March/06: It appears that 8.23.7 is better for that crashing on x log off  problem.

Apr/06: ati 8.24.8 and  Xorg7 is now in debian unstable. It's recommended to avoid upgrading X before things settle. If you do, an easy way, is to use the ati installer to make latest ubuntu's debs and use make-kpkg to build the module.

Note that there is an open source project for radeon cards but at least currently it isn't very complete at least for this card and 3d accel.

For the gamers, etc: If you need a low latency method to do sound mixing for incomplete proprietary applications like quake 3 and teamspeak that don't go well with alsa's mixing, try jackd in connection with oss2jack. For some reason, this system wouldn't even function with artsd for those purposes, but jackd+oss2jack are luckily much more responsive anyway.
Apr/06: check this thread on the "urban terror" quake3 mod linux forum for a patch for oss2jack and linux 2.6.15/16.
Feb/06: The latest kernel trees include an open source driver that makes this nic work. it's reported to have less features than the proprietary driver currently.
Apr/06: 2.6.16 includes sky2, an open source driver for it. It works, but at least here it appears to hang the connection once in a while (especially in many multiple connections). Syskonnect's driver is not updated for 2.6.16 yet but if you edit its sky2.c somehow (by extracting and remaking the tar.bz2 contained or editing the autogenerated patch from its installer) and add the linux/ip.h header in it, it works fine.
March/06: It appears that a set of omnibook tools and a module found here, support several devices through a /proc/omnibook interface. Currently the module will need to be run with the option "ectype=12" that is for late toshiba firmwares.  The devices I can see reporting status are ac status, battery (level), blank (lcd console blanking), display (external),  lcd (brightness), onetouch (hotkeys), temperature. The hotkeys do seem to generate events on /dev/input/event0 with it.
March/06: The omnibook tools mentioned above also support them but the acpi api is more easily used with generic tools easily available.

.: fs at www0 dot org