SanDisk Sansa c200v2, m200v4, clipv1, clipv2, clip+, and fuzev2

[sko]

Hi, I applied kugel´s changes on the fuze button driver to the modified version wich is working with the e200v2, now the DBOP-values on the debug-menu are visible on the e200v2 too.

EDIT:
Hmm... the value it shows for the REC-button seems to depend on the value of the scroll wheel (I have no plan about this dbop-voodoo thing, so I don't know if this behavior is correct or not), this are the values it shows (I turned the wheel clockwise and watched the DBOP_DIN values):

before pressing REC	while pressing REC
FFFF887F	087F
FFFFA87F	287F
FFFFE87F	687F
FFFFC87F	487F

[kugel.]

"REC-button seems to depend on the value of the scroll wheel"
What do you mean by this? Does that mean there's no reaction of dbop when pressing REC without scrollwheel, or just that the values differ?
The latter one is easily explainable. REC seems to be bit 15, which means it subtracts 8 from the 2nd pair of numbers (each pair is 1 byte). The REC button only represents 1bit, so it only changes the total number by a fixed value.

[sko]

Just the values differ, it's also working while scrolling, so REC is bit 15. Thanks for the explanation

[kugel.]

You can implement rec button reading just like it's done with the power button. But, don't forget that bit 15 NOT set indicates a button press (so something like, bool _button_rec = !(_dbop_din & (1<<15)) )

If you got it working, please post your patch to http://www.rockbox.org/tracker/task/9639

[FlynDice]

I've tried this code to get the rec button.  It sees the rec button as pressed during startup and clears my settings sending me back into cabbiev and buttons/wheel don't work anymore.  Is there a reason for making a  bool _button_rec variable?

static void get_rec(void)
{
    if (!(_dbop_din & (1<<15)))
        int_btn |= BUTTON_REC;
}

[kugel.]

I've tried this code to get the rec button.  It sees the rec button as pressed during startup and clears my settings sending me back into cabbiev and buttons/wheel don't work anymore.  Is there a reason for making a  bool _button_rec variable?

static void get_rec(void)
{
    if (!(_dbop_din & (1<<15)))
        int_btn |= BUTTON_REC;
}

That should work. But you need to reset the button state at the beginning of the dbop function (int_btn &= ~BUTTON_REC), and call get_rec at the end of it.

[FlynDice]

Yes that's what I've done I think.  I used this line to reset since get_power could also modify int_btn. 

  int_btn &= ~(BUTTON_REC|
              BUTTON_POWER);

EDIT: I think I got the REC button working by putting in a bypass while rockbox is booting up.  I'm not sure how to check it out real well but when I press it the button light illuminates and the backlight comes back on.  If I can figure out how I'll post a diff to the tracker.

EDIT2:  I've tested it in several plugins also (sudoku and calculator and a few others).  It is seeing the rec button but it seems something else besides the rec button is setting DBOP 15 to 0 also.  In XOBOX the game is paused/unpaused in a regular pattern.  I did post a patch to the tracker if anyone wants to play with it some.

[weltyj]

I found BUTTON_LEFT on GPIOC2

Found nothing else on GPIO's A or B for buttons, tested both high and low states.

The backlight function as implemented in the SVN trunk does not appear to work.  I could see no backlight at all.

Testing in bootloader/sansa_as3525.c:

Buttons work fine via button_read_device(), until a call is made to lcd_init(), at which point I see what appears to be the backlight come on, but the buttons are hosed (I assume because the issue of using GPIOC for both buttons and lcd has not been sorted out yet.)

I tried turning on the backlight with GPIOD 0 thru 6, but no success.  Is there any chance GPIOD might be used for BUTTTON_VOL_UP, BUTTON_VOL_DOWN or BUTTON_REC ?

[kugel.]

I found BUTTON_LEFT on GPIOC2

On 1 or 0?

Found nothing else on GPIO's A or B for buttons, tested both high and low states.

The backlight function as implemented in the SVN trunk does not appear to work.  I could see no backlight at all.

Testing in bootloader/sansa_as3525.c:

Buttons work fine via button_read_device(), until a call is made to lcd_init(), at which point I see what appears to be the backlight come on, but the buttons are hosed (I assume because the issue of using GPIOC for both buttons and lcd has not been sorted out yet.)

This could very well also be your sansa crashing. The lcd driver is copied from the e200v2, seems it's not working at all. Can you try the Fuze lcd-driver? (Copy lcd-fuze.c into lcd-c200v2.c) I don't think it will work but it's worth a try.

I tried turning on the backlight with GPIOD 0 thru 6, but no success.  Is there any chance GPIOD might be used for BUTTTON_VOL_UP, BUTTON_VOL_DOWN or BUTTON_REC ?

Could be, no idea It could also be that you need to switch some other PIN to find other buttons (which could then even be on the same pins as buttons already found). The backlight is certainly not on GPIO though.

[weltyj]

I found BUTTON_LEFT on GPIOC2

On 1 or 0?

On 0, just like all the other buttons.

This could very well also be your sansa crashing. The lcd driver is copied from the e200v2, seems it's not working at all. Can you try the Fuze lcd-driver? (Copy lcd-fuze.c into lcd-c200v2.c) I don't think it will work but it's worth a try.

I'll give it a try.
---
No luck.  I tried both the Fuze and the e200v2 drivers.  The clip and m200v4 drivers aren't drop in replacements so that will take a little more effort to try those if you think it's possible they'll work.

I tried turning on the backlight with GPIOD 0 thru 6, but no success.  Is there any chance GPIOD might be used for BUTTTON_VOL_UP, BUTTON_VOL_DOWN or BUTTON_REC ?

Could be, no idea It could also be that you need to switch some other PIN to find other buttons (which could then even be on the same pins as buttons already found). The backlight is certainly not on GPIO though.

I am thinking that too -- could be a keyscan col/row setup similar to what's been found on the clip/m200 models, we've just found the first column so far...

[FlynDice]

I played around with the cabbiev2 theme settings on my e280v2 and found I could make the scrollwheel work if I change the background color to white and don't use the backdrop.  It worked one bit off of white(EFFFF...) but I haven't been able to test to failure so far.   I did set the top adjustment to the middle of the settings limits and scrolling failed though.  Would this point to an lcd driver issue?

EDIT:  For background colors scrolling only works for RGB: values FFXXXX F7XXXX EFXXXX and E7XXXX.  The green and blue values all work fine it's the red values that seem to be a problem.

[kugel.]

I played around with the cabbiev2 theme settings on my e280v2 and found I could make the scrollwheel work if I change the background color to white and don't use the backdrop.  It worked one bit off of white(EFFFF...) but I haven't been able to test to failure so far.   I did set the top adjustment to the middle of the settings limits and scrolling failed though.  Would this point to an lcd driver issue?

I would think so yes.

[FlynDice]

Back on the e200v2 REC button, I'm trying to figure out if I'm chasing my tail here a bit.  I've got it seemingly working in the main rockbox app.  I say that because the backlight and buttonlight both come back on when I push the rec button just like any other button.  I also enabled recording in the config file and a long rec press takes me into the recording menu.  There do not seem to be any "phantom" rec presses in the main rockbox area aside from on bootup.  On bootup rockbox behaves as if the rec button is pushed, ie it clears all settings and a grey box appears saying settings cleared.  I have managed to bypass the rec button on startup to get into rockbox with a working rec button however.  Playing around with the plugins though has made me wonder what else may be going on.  BlackJack gets stuck in a button switch because it continuously pulls the BUTTON_REC  code off the button queue even though the rec button isn't being pressed.  There's also one of the higher bits set so it gets stuck.  Sudoku behaves properly when I press the rec button but there are phantom rec presses when I turn the scrollwheel.  Because of the startup issue I'm not sure if this is a rec button issue or a plugin issue.  Any thoughts?

[bertrik]

Just letting you know, the ams sansas should now have working radio, except for the c200v2.

Most of them seem to use an si4702 FM chip which is supported by the si4700.c driver, the m200v4 uses a tea5767 FM chip supported by the tea5767.c driver. There is no radio yet for the c200v2 because we don't know yet how to talk to the FM chip.

[FlynDice]

I believe the rec button and scrollwheel issues on the e200v2's are basically the same problem.  I changed the background color in blackjack to LCD_WHITE and the rec button gets read just fine.  This actually makes the bootup issue with the rec button also make sense.  The background color when rockbox boots up is black.  This is the timeframe I have been bypassing.  Once I make it into the menu's, I've set the background color to white and rec and scrollwheel work fine.  I'm thinking that for some reason any background color that doesn't have it's red component maxed out is altering DBOP 15 and 14, and maybe also 13 but I can't tell yet.  Does this bolster the LCD driver as the culprit?