← Back to team overview

multi-touch-dev team mailing list archive

Re: MosArt RMTP-MA957: Pointer always at (0,0)

 

Mmm. If you could recompile the hid and hid-mosart modules from these fiels, we would have more certainty. For this, use "make -C `your_kernel_sources` SUBDIRS=`pwd` and another option (-m ?) to tell make where to find files such as Modules.symvers, which are somewhere in /lib/modules".

Before that, it could be useful to make sure that the hid default code is used. In /var/log/messages you should find a line that shows when your device is grabbed by a driver. The first work in that line should be either usb-generic or something like egalax-touch. If the former, the problem is confirmed and recompiling the modules would be useful. If the latter, then the issue is elsewhere.

Cheers,

St.



Le 21 sept. 10 à 17:54, Manuel Schölling a écrit :

Dear Stéphane,
it appears eventually that your device is not managed by the hid-
egalax driver but by the defaut hid-core code. It might be that the
device is not listed in the hid_blacklist structure (in hid-core.c)
or in the lists of supported devices (in hid-egalax.c). Could you get
the sources of your kernel and check these two files in drivers/
hid/ ? you should look for EGALAX in the code, as far as I remember.
My device seems to be blacklisted by hid-core.c:
 static const struct hid_device_id hid_blacklist[] = {
 [...]
 { HID_USB_DEVICE(USB_VENDOR_ID_DWAV,
   USB_DEVICE_ID_DWAV_EGALAX_MULTITOUCH) }, [...]

and supported by hid-egalax.c:
 [...]
 static const struct hid_device_id egalax_devices[] = {
        { HID_USB_DEVICE(USB_VENDOR_ID_DWAV,
                        USB_DEVICE_ID_DWAV_EGALAX_MULTITOUCH) },
        { }
 };
 [...]

Although I'm not 100% sure if the sources I was looking at corresponds
to the version of the ubuntu stock kernel images (at least both is
v2.6.35, but I don't know if the ubuntu kernel images include other
patches)

Le 21 sept. 10 à 16:07, Benjamin Tissoires a écrit :

Forwarding your mail to Stéphane, as it is a kernel related problem.

- output 'unknown class' right at the beginning of 'xinput test 10'
-> totally normal

- The link you gave may be the same problem. I can not manage to
exactly understand, but Stéphane will

Cheers,
Benjamin

-------- Message original --------
Sujet: Re: [Multi-touch-dev] MosArt RMTP-MA957: Pointer always at
(0,0)
Date : Tue, 21 Sep 2010 15:50:48 +0200
De : Manuel Schölling <manuel.schoelling@xxxxxx>
Pour : Benjamin Tissoires <tissoire@xxxxxxx>

Hi Benjamin,
Am Dienstag, den 21.09.2010, 15:26 +0200 schrieb Benjamin Tissoires:
Thanks, however I will need another file to finish diagnose your
problem:
can you give me the output of the same device sequence (press - move
release) given by the command:
$ evtest /dev/input/event5
this one again is attached to this mail.

and finally, are you running the stock xorg and xf86-input-evdev from
ubuntu, or did you recompile any?
No, I'm using the ubuntu packages.

Currently, I can only say that your device is correctly
recognized, but
the events are lost somewhere (between the kernel and X or into the
Xserver itself)
Just two more questions:
- Did you note that 'xinput test 10' printed 'unknown class' right at
the beginning? Is that ok?
- Mohamed wrote that it might be the same problem as
http://www.spinics.net/lists/linux-input/msg11182.html
What do you think about it? Should I give it a try?


Once again: Thank you very much for your effort!

Cheers,
Manuel


<evtest_output>







Follow ups

References