Author Topic: Usb System Totally Non-functional Kernel 2.6.24  (Read 12916 times)

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« on: January 11, 2009, 04:30:01 pm »
Hi All,

Does anyone know how to get the USB sub-system running on an Akita with the following OS's ? :

Angstrom-x11-image-glibc-ipk-2007.12-r13-akita-installkit.tgz

and

Angstrom-x11-image-glibc-ipk-2007.12-r9-akita-installkit.tgz.gz


All of the main kernel modules are present, for example:

the kernel module:

usbmouse.ko

is present , but when the mouse is attached, no go.
same for USB storage, and probably, just about everything else.

Capn_Fish kindly compiled a Module for me which modprobe loads happily, but when my RT73 Radio Card
is plugged in, same as all the other USB devices, no operation, and the system is not recognizing it at all
except for :

lsusb -v

which shows the manufacturers data for each USB device correctly, at least, the USB system is reading this
data from each device, so thats a good start.

Modprobe picks up and loads all Modules properly and normally.

lsusb list them after modprobe has loaded them.

Any suggestions anyone ?

If not, do you have an Angstrom Kernel 2.6.24 OS file we can all try that is not the same as the 2 OS's above ?
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

utx

  • Full Member
  • ***
  • Posts: 135
    • View Profile
    • http://www.penguin.cz/~utx/zaurus/
Usb System Totally Non-functional Kernel 2.6.24
« Reply #1 on: January 11, 2009, 06:43:45 pm »
Quote from: Allelectrix
lsusb -v

Did you check the current requirements? If it is more than 150mA, you are going out of power budget (and dmesg shows proper error).

Jon_J

  • Hero Member
  • *****
  • Posts: 1853
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #2 on: January 12, 2009, 01:42:37 pm »
I have 5 devices that no longer work because of the electrical restriction of the 2.6 kernels.
Using a device that is under 150mA, you still have to use the commandline to get USB host devices to work.
Not like Cacko, or pdaxii13, where most all USB host devices are plug-n-play.

There is a "hack" in this thread that "bypasses" the power limitation.
https://www.oesf.org/forum/index.php?showtopic=20739
« Last Edit: January 12, 2009, 01:52:00 pm by Jon_J »
C3100 Multiboot-->Angstrom 2007.12-r18 | Cacko 1.23 | ArchLinuxARM
C3200 pdaxii13v2-5.5-alpha4 Akita on NAND

Ambicom WL1100C-CF Wifi - Ambicom CF modem - Ambicom CF GPS - Belkin-F5D5050 USB LAN
Socket CF Bluetooth rev K - Iogear 4 port USB micro hub - pocket CF card reader
Targus mini USB optical mouse - 2 Targus SD card readers

utx

  • Full Member
  • ***
  • Posts: 135
    • View Profile
    • http://www.penguin.cz/~utx/zaurus/
Usb System Totally Non-functional Kernel 2.6.24
« Reply #3 on: January 12, 2009, 04:06:52 pm »
Quote from: Jon_J
I have 5 devices that no longer work because of the electrical restriction of the 2.6 kernels.
Using a device that is under 150mA, you still have to use the commandline to get USB host devices to work.
Not like Cacko, or pdaxii13, where most all USB host devices are plug-n-play.

There is a "hack" in this thread that "bypasses" the power limitation.
https://www.oesf.org/forum/index.php?showtopic=20739

No kernel is not the problem. The real problem are devices programmed to lie about their needs.

I have one device, which tells true (WL54USB-A) and does not work with 180mA, and another one, which lies (A-Data 2G SD Flash Duo with USB), wanting 500mA, running with 180mA.

In such case you need my OHCI Power Budget Hack. You need to recompile kernel, even if USB driver is in module (power budget is stored in platform description of kernel, which is in the kernel itself). Replace the number 680 by anything you need.

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #4 on: January 13, 2009, 04:04:53 pm »
Quote from: Jon_J
I have 5 devices that no longer work because of the electrical restriction of the 2.6 kernels.
Using a device that is under 150mA, you still have to use the commandline to get USB host devices to work.
Not like Cacko, or pdaxii13, where most all USB host devices are plug-n-play.

There is a "hack" in this thread that "bypasses" the power limitation.
https://www.oesf.org/forum/index.php?showtopic=20739

Hi Jon_J , many thanks for your reply, yeh, seen that thread already mate, done that and did not work for me unfortuantely, any more ideas ?
there is nothing wrong with the hardware either as Cacko and PdaX fire up the USB with total ease, - everything works including my RT73
Radio Card (courtesy Capn_Fish and evilJazz) - just in case you might be wondering if there is anything wrong with the Hardware.
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #5 on: January 13, 2009, 04:10:22 pm »
Quote from: utx
Quote from: Jon_J
I have 5 devices that no longer work because of the electrical restriction of the 2.6 kernels.
Using a device that is under 150mA, you still have to use the commandline to get USB host devices to work.
Not like Cacko, or pdaxii13, where most all USB host devices are plug-n-play.

There is a "hack" in this thread that "bypasses" the power limitation.
https://www.oesf.org/forum/index.php?showtopic=20739

No kernel is not the problem. The real problem are devices programmed to lie about their needs.

I have one device, which tells true (WL54USB-A) and does not work with 180mA, and another one, which lies (A-Data 2G SD Flash Duo with USB), wanting 500mA, running with 180mA.

In such case you need my OHCI Power Budget Hack. You need to recompile kernel, even if USB driver is in module (power budget is stored in platform description of kernel, which is in the kernel itself). Replace the number 680 by anything you need.

Hi there utx, this looks like a good solution, as my system is rather limited in resources at the mo, would you be so good as to post the recompiled
kernel here to the forum (the zImage file) like you did the patch file , with the patch you have kindly provided already added  so everyone can share
your good work on this issue ?

That would be fantastic !  

I presume this is Kernel 2.6.24 for Angstrom ?

Many thanks ! We await your 2nd post,  Allelectrix
« Last Edit: January 13, 2009, 04:12:54 pm by Allelectrix »
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #6 on: January 13, 2009, 04:16:06 pm »
Quote from: utx
Quote from: Allelectrix
lsusb -v

Did you check the current requirements? If it is more than 150mA, you are going out of power budget (and dmesg shows proper error).

Oh yes, even with this issue, I connected all devices via a powered 2.0 Hub, and they all still did not work, just like they all still did not
work when I connected them up to the Computer by themselves.
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #7 on: January 25, 2009, 08:36:18 am »
As of the end of Janurary 2009, I am still looking for a solutiuon to this issue, I am happy to send my SL-C1000 to anyone
who can resolve this USB Start issue. Many thanks to you all !  
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #8 on: April 29, 2009, 09:30:50 am »
as of end of April 2009, this issue is still unresolved. does anyone know if USB using kernel 2.6.24 Angstrom can be used on the other machine versions ?

that is, all machines issued after the SL-C1000 which feature the same ARM Processor ?

thanks all !
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

Allelectrix

  • Jr. Member
  • **
  • Posts: 84
    • View Profile
    • http://
Usb System Totally Non-functional Kernel 2.6.24
« Reply #9 on: June 01, 2009, 05:07:51 pm »
June 1st 2009, issue still unresolved. anyone want to exchange this C1000 for a OQO x86 ? - contact me at my email address.
SL-C1000 running PdaXii13
Edimax EW-7318USg rt73 Chipset USB 802.11 g/b Wifi
ASPj Source Code rt73.o 2.4.20 kernel module with monitor + packet injection capability
Aircrack-ng 0.9
Miniature USB 2.0 Hub that attaches to rear of screen comfortably using velcro
2 x 500mA 12v DC - 5v DC Converters powering the USB Hub, WiFi Card and the Computer
12v DC 7Ah Sealed Lead Acid Rechargeable Cell powering the DC to DC Convertors

sdjf

  • Sr. Member
  • ****
  • Posts: 447
    • View Profile
    • http://www.sdjf.wordpress.com and http://www.sdjf.esmartdesign.com
Usb System Totally Non-functional Kernel 2.6.24
« Reply #10 on: June 01, 2009, 06:33:07 pm »
USB is really tricky.  I have different kernel and module...Sharp ROM on sl6000 and the problem for me has been that if I change which CF card I plug in to my USB CF reader, the zaurus assigns it a different device number.

I found it quite annoying as I need to change and it has nothing to do with power requirements.

Have you watched dmesg to see what device names have been assigned each time?  That could be your problem, I don't know how to fix it.  For me, rebooting does it, but I'd rather not have to reboot every time I change what's plugged into USB.

Just some thoughts in case you haven't noticed this.

sdjf
http://www.sdjf.esmartdesign.com
http://www.sdjf.wordpress.com
-----------------
sl5500 running Sharp ROM 2.38 (dead batteries)
sl6000L running Sharp ROM 1.12 (still working)
Opera 7.25 and 7.30
Socket CF 56k modem
3Com USB Ethernet Adapter
Toshiba, Lexar and Kingston SD cards
Lexar, Kingston and Transcend CF cards

koen

  • Hero Member
  • *****
  • Posts: 1008
    • View Profile
    • http://dominion.thruhere.net/koen/cms/
Usb System Totally Non-functional Kernel 2.6.24
« Reply #11 on: June 02, 2009, 10:01:15 am »
Quote from: Allelectrix
June 1st 2009, issue still unresolved. anyone want to exchange this C1000 for a OQO x86 ? - contact me at my email address.

Why don't you read http://www.angstrom-distribution.org/contact and follow the instructions there?
Forums are not bugtrackers!!! Smart questions
Ångström release team
iPAQ h2210, iPAQ h5550, iPAQ hx4700, Zaurus SL-C700, Nokia 770, all running some form of GPE
My blog

craigtyson

  • Sr. Member
  • ****
  • Posts: 462
    • View Profile
Usb System Totally Non-functional Kernel 2.6.24
« Reply #12 on: June 02, 2009, 06:06:51 pm »
Did they work in Angstrom before?

I have / still use several USB devices on my C1000 on Angstrom.  The default however is that the Z is in slave mode and so has to be put into Master mode to host the USB devices.

As Koen says, the instructions how to do so are in the FAQ

Q: How can I get the USB Host working (on machines with USB Host HW)?

    * A: The ohci-hcd module is mandatory for USB connections; load it with "modprobe ohci-hcd". To have it loaded at boot time, create a file called "/etc/modutils/ohci-hcd", with the single line "ohci-hcd", and run "update-modules". Don't forget that some machines (Zaurus, for example) can't deliver much power to usb devices (disks, network adapters, ...); you may need a powered usb hub.
Craig
SL-C1000, Angstrom GPE 2007.12r13
Ambicom WL1100C, SanDisk UltraII 1GB SD
 
Old Faithfull - SL-5500 battery out of the C750 (see below) TKC v1 (New) 1GB SD
Dead - SL-C750, Cacko 1.23 lite (Dropped, very messy!)
Ambicom WL1100C, SanDisk Dead - 1GB SD
Keep your Z on a bungee, you never know.....

semen

  • Newbie
  • *
  • Posts: 18
    • View Profile
Usb System Totally Non-functional Kernel 2.6.24
« Reply #13 on: June 02, 2009, 09:39:49 pm »
Is there a reason why you only want this working on Angstrom and not looking at any other options?

If the full kernel source is available, I don't see why this isn't working right now. When my C1000 arrives, I'll look into this for you.