From: Antonius Hellmann <antonius.hellmann_at_gmx.de>
Date: Wed, 9 May 2007 20:11:57 +0200

I just checked your patch. One idea:
Probably the reading/writing of the gpio input generates another interrupt,
which in turn again accesses the gpio register? If I have it right, then the
original software disables the interrupt (GPIOH_INT_CLEAR), before reading
the registers and reenables the irq again at the end of the irq routine.

----- Original Message -----
From: "Michael Sevakis" <jethead71_at_sbcglobal.net>
To: "Barry Wardell" <barry.wardell_at_gmail.com>; "Rockbox development"
<rockbox-dev_at_cool.haxx.se>
Sent: Wednesday, May 09, 2007 5:33 PM
Subject: Re: Sansa (PP502x?) GPIO interrupts in rockbox software


> It doesn't seem to actually enable the GPIO interrupts.
>
> When I do, thinks break badly with rockbox stuck at a black screen when
> the
> logo should be displayed but GPIO IRQs functional. The firmware doesn't
> seem
> to actually be running because the buttons themselves are still being
> polled
> and do work otherwise.
>
> Here the quick tester which I'd expect to have boot, disable the
> scrollwheel
> and blink when turned. It only does the latter two.
>
> You'll get some warnings but that's irrelevant to this check.
>
>
> ----- Original Message -----
> From: "Barry Wardell" <barry.wardell_at_gmail.com>
> To: "Michael Sevakis" <jethead71_at_sbcglobal.net>; "Rockbox development"
> <rockbox-dev_at_cool.haxx.se>
> Sent: Wednesday, May 09, 2007 4:34 AM
> Subject: Re: Sansa (PP502x?) GPIO interrupts in rockbox software
>
>
>> On 5/9/07, Michael Sevakis <jethead71_at_sbcglobal.net> wrote:
>> > I would've whipped up a IRQ button driver real quick and committed it
> but I
>> > see I've been beaten to it while fixing playback bugs. :)
>>
>> Well apparently the one I posted has some problems with it still and I
>> still don't have my Sansa with me for testing so I can't do much more
>> with it for now. Feel free to fix it up and commit ;)
>>
>> One word of caution - it was based on a older version of the button
>> driver, so some of the newer fixes that were added in the past while
>> may not be present.
>
>
>
Received on 2007-05-09