On Thu, 2008-12-11 at 05:41 -0800, Pivotian wrote: > can you put some more light on using touch screen with the help of > "tslib".? > tslib is a user space library which basically hooks to your driver and provides filtering like dejitter , variance It comes with certain tool like ts_print_raw , ts_calibrate which basically are helpful in getting your calibration up and running without the need of X.
This is how you can proceed 1) check /etc/ts.conf, make sure the module_raw entry for your platform is appropriate, and linear, variance, and dejitter are uncommented. A sample ts.conf file is like this module_raw input * Note that i had enabled events in the input layer of the kernel module variance module dejitter module linear 2) export TSLIB_TSDEVICE=/dev/yourtouchscreendevicenode other env settimgs that have to be exported TSLIB_CALIBFILE=/etc/pointercal TSLIB_CONFFILE=/etc/ts.conf TSLIB_PLUGINDIR=/usr/share/ts/plugins TSLIB_FBDEVICE=/dev/fb0 TSLIB_CONSOLEDEVICE=/dev/tty 3) ts_print_raw .. confirm you can get raw unfiltered data from the touchscreen. Tap it a few times while it’s running. 4) ts_calibrate and make sure your /etc/pointercal has been updated with calibrating values 5) ts_print .. confirm you can get filtered data 6) ts_test .. confirm the filters are doing what they should. not too much jitter, etc. Once you have atleast X running with touch succesfully you could use the same pointercal file in the /system/etc/ of android. recompile android after sean's changes for using pointercal file. Should do the trick. > On Dec 11, 2:40 pm, Manav Gautama <[email protected]> wrote: > > On Thu, 2008-12-11 at 01:27 -0800, Pivotian wrote: > > > good.What did you do for Touchscreen? Did you used "tlib"? > > > > tslib : patches from sean > > > > > > > > > On Dec 11, 9:59 am, Manav Gautama <[email protected]> wrote: > > > > On Wed, 2008-12-10 at 20:34 -0800, Pivotian wrote: > > > > > yeah, looks great even I tried the same thing earlier to build android > > > > > for ARMv6 instruction with separate tool chain but I got compiler > > > > > errors that's why I asked? Any ways its good to see when some one gets > > > > > succeeded in building Android for ARMv6. One more thing I would like > > > > > to ask you is, whether your audio is working fine? are you using ALSA > > > > > driver? if yes, what are all the changes you made in android source > > > > > code (outside kernel) to make audio work. Please share your ideas. > > > > > > My existing kernel was a 2.6.21 mobilinux one and all i could forward > > > > port to the 2.6.27 > > > > are NAND, USB, LCD, Touch as of now. Yet to do alsa fixes in the > > > > kernel for my platform. Will keep you posted. > > > > > > > Thanks, > > > > > Sujit > > > > > > > On Dec 11, 7:22 am, Manav Gautama <[email protected]> wrote: > > > > > > On Tue, 2008-12-09 at 20:44 -0800, Pivotian wrote: > > > > > > > hey its great to see that you have compiled Android for ARMv6. > > > > > > > Can you > > > > > > > please tell us the steps required to compile the Android for ARMv6 > > > > > > > instruction set using our own tool chain? > > > > > > > > Thanks Pivotian ! I changed the the build/core/combo/linux-arm.mk > > > > > > file > > > > > > and added > > > > > > up changes pertaining to toolchain by specifying a new > > > > > > TARGET_ARCH_VERSION and > > > > > > added arch_version_cflags to include -march=armv6 . These flags are > > > > > > the > > > > > > ones that make the difference. Am currently trying to rebuild it > > > > > > against > > > > > > a mobilinux toolchain as compared to having just passed the above > > > > > > flags > > > > > > to the existing toolchain even though it > > > > > > compiles it without any errors, i suspect it might not have really > > > > > > made a difference. > > > > > > > > > On Dec 9, 8:08 pm, Manav Gautama <[email protected]> > > > > > > > wrote: > > > > > > > > Dear droids, > > > > > > > > I could get a 2.6.27 kernel ported to > > > > > > > > omap2430 based > > > > > > > > board along with android changes. Also compiled android using > > > > > > > > armv6 > > > > > > > > toolchain and could tazer it up successfully on the machine. > > > > > > > > I faced issues on the touch side of the device. Hacked ads7864.c > > > > > > > > according to > > > > > > > > inputs from Maxime, the READ_X and READ_Y calls had to be > > > > > > > > swapped. > > > > > > > > > > Added sean mcneals tslib Inputdevice integration patch's and > > > > > > > > calibrated > > > > > > > > using tslib and generated a pointercal that works perfectly > > > > > > > > with X11 . > > > > > > > > > > Rebuild android after the changes and populated the resulting > > > > > > > > root > > > > > > > > filesystem. Copied the the same pointercal to system/etc and > > > > > > > > got android > > > > > > > > up with a working touch. > > > > > > > > > > Video and steps used to get a working device > > > > > > > > @www.bandwidthcrunch.com > > > > > > > > > > Keep up the good work ! > > > > > > > > > > regards, > > > > > > > > Manav > > > > > > > > > > bandwidthcrunch at gmail dot com > > > > > > > > > >www.bandwidthcrunch.com > > > > > > > > -- > > > > > > regards, > > > > > > Manav > > > > > > > > bandwidthcrunch at gmail dot com > > > > > > > >www.bandwidthcrunch.com > > > > > > -- > > > > regards, > > > > Manav > > > > > > bandwidthcrunch at gmail dot com > > > > > >www.bandwidthcrunch.com > > > > -- > > regards, > > Manav > > > > bandwidthcrunch at gmail dot com > > > > www.bandwidthcrunch.com > > > -- regards, Manav bandwidthcrunch at gmail dot com www.bandwidthcrunch.com --~--~---------~--~----~------------~-------~--~----~ unsubscribe: [email protected] website: http://groups.google.com/group/android-porting -~----------~----~----~----~------~----~------~--~---
