[Gta04-owner] QtMoko v44

[Dr. H. Nikolaus Schaller]

Am 14.04.2012 um 16:59 schrieb Andreas Kemnade:

> Hi,
> 
> On Sat, 2012-04-14 at 16:18 +0200, Dr. H. Nikolaus Schaller wrote:
>> Hi,
>> 
>> Am 14.04.2012 um 15:32 schrieb Andreas Kemnade:
>> 
>>> Hi,
>>> 
>>> On Thu, 2012-04-12 at 15:40 +0200, Radek Polak wrote:
>>>> Hi,
>>>> QtMoko v44 is now available for download [1]. This should be best QtMoko 
>>>> release for GTA04 and i think i can recommend it for daily use. The GTA04 HW 
>>>> adaption is now nearly finished.
>>>> 
>>> The gpio186 thing works not so well yet.
>>> I get either
>>> 	a dialog with a number block and "please wait" above it (after
>>> 	showing "no modem" for some time
>>>       (probably killing qpe here will help here)
>>> 	a simple message "no network"
>>> That depends on the previous state (was it off or on at reboot?).
>>> The modem also needs some time to boot before
>>> it appears as an usb device.
>> 
>> Yes, ca. 1-2 seconds.
>> 
>>> I propose this check here:
>>> 
>>> if [ -f /sys/class/gpio/gpio186/value ];
>>> then
>>>  for dummy in 1 2 3 4 5 6; do
>>>    if [ ! -e /dev/ttyHS_Application ]; then
>>>      echo 0 >/sys/class/gpio/gpio186/value
>>>      sleep 0.5
>>>      echo 1 >/sys/class/gpio/gpio186/value
>>>      sleep 8
>>>    fi
>>>  done
>>> fi
>> 
>> Hm. That doesn't seem correct as well.
>> Please refer to 6.4.1 in the manual.
>> 
>> The module should be controlled by 010 impulses
>> of >200 ms each.
> 6.4.1.
> "Setting the GPIO to "1" (for at least 200 ms) turns the module on."
> That does not say that I should set it to zero soon.

Ok, you are right. I will improve that page (there is also a typo).

> http://projects.goldelico.com/p/gta04-kernel/page/Wireless/ is a bit
> clearer. But copying from the hw-test-script is of course a very good
> idea. I have just done that now. 

BTW, it also check if the modem has powered up on a device
that has no GPIO186 (i.e. the GTA04A3 boards).

> But I guess that should definitively go into the kernel as a rfkill
> device, I just do not know yet how and where to put the lsusb | grep
> into the kernel in a clean way

Another option could be to check if the HSO driver module has been loaded.
Or check the existence of /dev/ttyHS*

But I don't know if that is easier to handle in kernel space...

And, Neil may comment on this since we have a similar
toogle button control for the GPS module. There, an rfkill driver
should monitor the serial line for (missing) incoming data.

So I think both areas may have similar needs (timed monitoring of
some /dev/tty from some other driver and pulsing a GPIO).

BR,
Nikolaus