Marcin Juszkiewicz

I was at cinema and watched ‘Sunshine‘ movie. It was long time since I watched SF on big screen…

Story is easy — the Sun is dying, Earth is frozen so people create big bomb and plan to detonate it inside of our star. During flight they find previous mission ship etc, etc… Finally they complete their mission and everything is good.

But amount of goofs is big… After explosions in oxygen plant they discovered that amount of oxygen is too small for return trip and can be too small to complete mission. But why they keep air on whole ship? Keeping air in only needed parts will give them extra time probably… And there are few others but I already spoiled too much ;)

Anyway I enjoyed movie but it is one of those which you can see once and do not want to see again.

Today I got few USB gadgets:

• Ethernet card (Damicom 9601 based)
• RS232 cable
• multi-port card-reader with USB 2.0 hub integrated (powered)

So during restructure of my USB network I tried to connect all my USB devices to desktop. With two external hubs I got out of ports…

Effect:

12:00 hrw@home:~$ lsusb
Bus 001 Device 055: ID 0a46:9601 Davicom Semiconductor, Inc.
Bus 001 Device 054: ID 058f:6362 Alcor Micro Corp.
Bus 001 Device 053: ID 0fce:d016 Sony Ericsson Mobile Communications AB
Bus 001 Device 036: ID 058f:6254 Alcor Micro Corp.
Bus 001 Device 001: ID 0000:0000
Bus 003 Device 059: ID 0525:a4a2 Netchip Technology, Inc. Linux-USB Ethernet/RNDIS Gadget
Bus 003 Device 058: ID 0424:223a Standard Microsystems Corp. 8-in-1 Card Reader
Bus 003 Device 057: ID 1457:5122
Bus 003 Device 056: ID 0421:0431 Nokia Mobile Phones
Bus 003 Device 055: ID 058f:9254 Alcor Micro Corp. Hub
Bus 003 Device 053: ID 04bf:0319 TDK Corp.
Bus 003 Device 046: ID 0a81:0205 Chesen Electronics Corp. PS/2 Keyboard+Mouse Adapter
Bus 003 Device 001: ID 0000:0000
Bus 002 Device 011: ID 04b8:082f Seiko Epson Corp.
Bus 002 Device 010: ID 046d:c70a Logitech, Inc.
Bus 002 Device 009: ID 046d:c70e Logitech, Inc.
Bus 002 Device 008: ID 046d:0b02 Logitech, Inc.
Bus 002 Device 001: ID 0000:0000

The list contain:

• USB Cup heater with integrated USB 1.1 Hub
• Nokia 770
• Sharp Zaurus C760
• FIC Neo1973
• Davicom Ethernet card
• EPSON Stylus DX3800
• TDK Bluetooth 1.1 dongle
• Logitech Bluetooth 2.0 EDR dongle
• PS/2 -> USB converter with optical mouse connected
• Hama multi-port card reader
• no-name multi-port card reader
• no-name multi-port card reader with integrated USB 2.0 Hub
• Sony Ericsson k750i phone

As result I got:

• 14 usb-storage devices
• 3 serial ports
• 3 network interfaces
• 3 input devices (Logitech Bluetooth need to be switched from HID to HCI)

And everything was working, needed /dev/ entries were created and HAL managed to show me proper tree with all devices. Both hubs were powered during test.

Now I disconnected most of them and left only needed ones.

W trakcie tegorocznych Pingwinariów miałem okazję wygłosić prelekcję o OpenEmbedded. W porównaniu do poprzedniej prezentacji (w Poznaniu) słuchacze wiedzieli o czym mówię. Padły ciekawe pytania (część znałem wcześniej z dyskusji z wieloma osobami), dyskusja przeniosła się później w różne miejsca.

Impreza była ciekawa i w następnym roku też tam będę (chyba, że coś mi wypadnie). Spotkałem wiele osób znanych mi dotychczas raczej tylko z sieci, kilku znajomych jak i wiele innych interesujących postaci. Jedną z nich był Tomasz Zieliński dzięki któremu mogłem podładować baterię w Neo1973 — spędziliśmy nieco czasu dyskutując o projekcie OpenMoko (Tomek miał prelekcję w tym temacie), przedstawiając telefon(y) i projekt żądnym informacji uczestnikom Pingwinariów.

Warto było być :)

Prezentacja do pobrania

OK, system (Debian ‘etch’) was installed and then upgraded to ‘sid’. Everything is working with Linux (kernel 2.6.18 from Etch):

• ACPI reports battery status, battery is charging without problems.
• Dell Wireless 1450 card works — I only had to install bcm43xx firmware (which I took from Cafuego’s Sarge Backports repository). There is also other way — Debian contain package bcm43xx-fwcutter which extracts firmware from Windows drivers.
• Gigabit Ethernet works with tg3 kernel module (tested with 100Mbps Ethernet only).
• CPU Frequency scaling works with speedstep-centrino module and provides wide range: 1600MHz, 1400MHz, 1200MHz, 1000MHz, 800MHz, 600MHz so it is possible to extend battery life with it.
• Backlight control (via Fn+Up/Down buttons) works — it is handled by hardware/BIOS probably.
• Touchpad was wrongly recognized — it is not Synaptic but ALPS so edit of /etc/X11/xorg.conf was needed. All informations what to change are described in README.alps (part of xserver-xorg-input-synaptic package). Idea found on Ubuntu blog.
• Suspend to disk works, suspend to RAM also works.

In other words — no problems yet — everything works like it should.

Yesterday evening I got Dell D400 laptop from courier. It is 12” portable notebook with 1.6GHz Pentium M cpu and 512M of RAM. Only harddisk as storage as in such small case there is no space for CD/DVD drive (which is available as external accessory).

More detailed specification:

• Intel Pentium M 1.6GHz CPU with 2MB cache
• Intel 855 chipset
• 512MB RAM
• 40GB harddisk (5400 RPM)
• 12” LCD
• touchpad
• trackpoint
• Broadcom BCM5705M Gigabit Ethernet
• Broadcom BCM4309 WiFi 802.11/a/b/g miniPCI card (DELL wireless 1450)
• FireWire
• RS232 port

Windows XP Professional was installed so I used Debian Win32 installer but then I broke something in installer so as result machine stopped booting from HDD. Today after half of hour spent on fighting with PXE Boot I have Debian 4.0 ‘etch’ installer running.

The plan is to have Debian ‘sid’ (as I have on desktop) on crypted LVM and use this machine as x86 build machine (some things in OpenEmbedded does not work properly on my amd64) and for conferences so I will be able to read my mail with normal tools. Due to fact that it has RS232 port it will probably also work as machine for checking kernel logs from Zaurus machines (usable for helping with SD/MMC driver for collie).

I will write more when system will be working — now it is still fetching packages to install.

When I got my Neo1973 it was known that there are problems with charging. Package contained USB wall charger (with US plug) which was useless due to not being conform with USB specification as device can use only 100mA and for more it have to ask — but how to ask dumb charger?

Before Pingwinaria 2007 conference I got into moment when my phone was so discharged that it had problem with charging itself. I started hunt for cheap Nokia phone using BL-5C batteries to use it as charger. I was also thinking about Nokia DT-14 (standalone battery charger). None of GSM shops which I visited had phone in acceptable price, but in last one seller told me that I should look for cheap 3rdparty battery chargers. Few days later I had one on my desk (with extra no-name BL-5C compatible battery).

It is good that I did not got DT-14 — Mickeyl got one and it refuse to charge original Neo1973 battery. Nokia phones also does not charge it. I even used Nokia 770 to test — it booted but no charge.

Now my Neo1973 most of time spends on desk connected via USB cable to powered hub which stay powered even when USB Host (my desktop) is not powered — from what I read many USB hubs switch off power on all ports when host goes down.

Few days ago Thomas Kunze wrote to openzaurus-devel list that he has working SD/MMC driver for Sharp Zaurus SL-5000/5500 (aka Collie). He provided patches and later also Ångström image + kernel so anyone can test.

I asked him for patches to current OpenEmbedded metadata and added them later so now anyone can build 2.6.20 kernel for collie to test SD/MMC. Koen Kooi built images for Ångström so it can be taken as start of support for this machine. Who knows.. maybe there will be no new OpenZaurus release when driver will work OK…

My collie (after charging and removing dust) booted nicely to 2.6.20 powered Ångström console image so I was able to test my cards against driver. Only 16MB MMC card was working so I sent two SD cards (64M and 256M) to Thomas. Now he will have also some not-yet-working cards for testing (all his SD/MMC cards works).

Also worth reminding — there is a bounty for writing SD/MMC driver for collie/2.6 so if you want to appreciate Thomas work you can join ;)

Today Koen gave me link to one discussion on Maemo-devel mailing list. It was about end of support for Nokia770.

Thread is interesting — some guys from Nokia wrote that this does not mean end but in same post write that there will be no new OS releases for 770. Hacker edition of OS2007 for 770 is also mentioned as solution but this is unofficial and unsupported (it works quite good on my device).

I think that Nokia770 is another device which will go into trashland due to fact that no one work on building external images for it. Ok, OpenEmbedded can build software for it but flasher is closed source and I do not know does format of flash images is published somewhere.

And some parts of Nokia rootfs are closed source too..