Messages in this thread | | | Date | Mon, 23 Sep 2013 18:46:37 +0200 | From | Matthias Kaehlcke <> | Subject | Re: [PATCH 2/2] iio: ti_am335x_adc: Take touchscreen channels into account for conversion timeout |
| |
Hi Sebastian,
El Mon, Sep 23, 2013 at 03:31:14PM +0200 Sebastian Andrzej Siewior ha dit:
> On 09/10/2013 11:02 PM, Matthias Kaehlcke wrote: > > The calculation of the old conversion timeout value was based on the number of > > channels used by this driver. This doesn't take into account that other channels > > can be used by the touchscreen driver. Adjust the timeout value to the maximum > > if the touchscreen driver is enabled > > What bug / miss behave are you trying to fix?
I ran into timeouts when using the touchscreen driver at the same time as the general purpose ADC and reviewed the timeout calculations. I found that the IDLE_TIMEOUT value is/was wrong (I submitted another patch for this) and that the ADC driver doesn't take into account the steps used by the touchscreen driver
> The difference in timming is minimal and therefore I would prefer to get > rid of this ifdef and assume the max value of those two instead.
Jonathan also expressed his concerns about this, I submitted a follow-up patch without the ifdef (https://lkml.org/lkml/2013/9/16/460). I would appreciate your comments on this patch
note that the timing difference isn't that minimal with the correct IDLE_TIMEOUT (~100us instead of 10us), it sums up to a max timeout of ~1.6ms (16 steps) and we are busy looping (though in the non-error case we will bail out as soon as the conversion cycle is finished)
best regards
-- Matthias Kaehlcke Embedded Linux Developer Amsterdam
In the absence of clearly-defined goals, we become strangely loyal to performing daily trivia until ultimately we become enslaved by it (Robert Heinlein) .''`. using free software / Debian GNU/Linux | http://debian.org : :' : `. `'` gpg --keyserver pgp.mit.edu --recv-keys 47D8E5D4 `-
| |