Re: RFC: Input: edt-ft5x06 - FT6336G Touch Panel

[Oliver Graute <oliver.graute@gmail.com>]

On 04/03/20, Marco Felsch wrote:
> On 20-03-03 20:37, Oliver Graute wrote:
> > On 02/03/20, Marco Felsch wrote:
> > > On 20-03-02 18:15, Oliver Graute wrote:
> > > > On 02/03/20, Andy Shevchenko wrote:
> > > > > On Mon, Mar 02, 2020 at 03:30:35PM +0100, Oliver Graute wrote:
> > > > > > Hello,
> > > > > > 
> > > > > > I have on of these FT6336G Touch Panels here and I try to get it work
> > > > > > with the ft6236 driver. It comes up with model identification "0x11". So
> > > > > > I added that identification to the switch case in
> > > > > > edt_ft5x06_ts_identify(). But it crashes directly after the probe with a
> > > > > > corrupted stack. No clue why. Some ideas?
> > > > > 
> > > > > >  		 * the identification registers.
> > > > > > +		case 0x11:   /* EDT EP0110M09 */
> > > > > 
> > > > > I have a question, how do you know it's M09 flavour?
> > > > 
> > > > This was an assumption after reading this thread on removing the old ft6236
> > > > driver and looking what the old driver did. There they talk about M09. 
> > > > 
> > > > https://patchwork.kernel.org/patch/9264845/
> > > > 
> > > > How can I be sure that that is right?
> > > > 
> > > > > 
> > > > > > [    2.232596] i2c i2c-0: LPI2C adapter registered
> > > > > > [    2.238931] edt_ft5x06 1-0038: probing for EDT FT5x06 I2C
> > > > > > [    2.244369] edt_ft5x06 1-0038: 1-0038 supply vcc not found, using dummy regulator
> > > > > 
> > > > > > [    2.257639] edt_ft5x06 1-0038: Model "EP0110M09", Rev. "�", 0x0 sensors
> > > > > 
> > > > > ...because clearly here something wrong with Revision field.
> > > > 
> > > > yes, this is suspicious.
> > > >  
> > > > > 
> > > > > > [    2.264536] input: EP0110M09 as /devices/platform/bus@5a000000/5a820000.i2c/i2c-1/1-0038/input/input0
> > > > > > [    2.273920] edt_ft5x06 1-0038: EDT FT5x06 initialized: IRQ 35, WAKE pin -1, Reset pin -1.
> > > > > > [    2.282124] Kernel panic - not syncing: stack-protector: Kernel stack is corrupted in: edt_ft5x06_ts_probe+0x9d4/0xa88
> > > > > > [    2.292830] CPU: 0 PID: 5 Comm: kworker/0:0 Not tainted 5.6.0-rc1-next-20200214-00051-g9874f196742d #69
> > > > > > [    2.302224] Hardware name: Advantech iMX8QM DMSSE20 (DT)
> > > > > > [    2.307544] Workqueue: events deferred_probe_work_func
> > > > > > [    2.312686] Call trace:
> > > > > > [    2.315141]  dump_backtrace+0x0/0x1c0
> > > > > > [    2.318802]  show_stack+0x14/0x20
> > > > > > [    2.322124]  dump_stack+0xb4/0xfc
> > > > > > [    2.325440]  panic+0x158/0x320
> > > > > > [    2.328494]  print_tainted+0x0/0xa8
> > > > > > [    2.331978]  edt_ft5x06_ts_probe+0x9d4/0xa88
> > > > 
> > > > The crash is gone after applying Marcos patch from here:
> > > > 
> > > > https://www.spinics.net/lists/linux-input/msg65908.html
> > > 
> > > That shouldn't fix the problem..
> > > 
> > > > Now the driver is loaded, but can't read data with evtest.
> > > 
> > > Did you enable the CONFIG_INPUT_EVDEV option?
> > 
> > yes, it was already enabled. But now its working for me. I see events
> > coming through during touching. My mistake was a wrong configured
> > interrupt line. But I still have this suspicious revision field and I
> > observe sometimes a different value there.
> 
> Did you saw those suspicious behaviour with the GENERIC_FT too?
> Shouldn't make a real difference because the GENERIC_FT acts most the
> time like the EDT_M09.
> 
> The problem is that display suppliers can flash their own firmware on
> the touchcontroller and most the time we don't know the firmware
> behaviour. Can you give us some more information about the used display?

I try to get more information about FT6336G Touch Panel from vendor. I
have a datasheet here and a excel sheet (both under NDA) with some
default register settings. Nothing mention the M06/M09/M12 flavour.
Currently I run it with M09, sometimes I still observe random crashes
after probe. But most of the time it seems to work now.

Best Regards,

Oliver