Re: [alsa-devel] [PATCH v3 3/5] ASoC: codecs: msm8916-wcd-analog: add MBHC support

[Damien Riegel]

On Mon, Aug 14, 2017 at 05:34:53PM +0100, Srinivas Kandagatla wrote:
> 
> 
> On 14/08/17 15:12, Damien Riegel wrote:
> > Hi,
> > 
> > 
> > On Thu, Aug 10, 2017 at 09:33:29AM -0400, Damien Riegel wrote:
> > > Hi,
> > > 
> > > On Thu, Aug 10, 2017 at 11:02:34AM +0100, Srinivas Kandagatla wrote:
> > > > Hi Damien,
> > > > Thanks for testing.
> > > > 
> > > > On 09/08/17 22:10, Damien Riegel wrote:
> > > > > Hi Srinivas,
> > > > > 
> > > > > On Wed, Aug 09, 2017 at 06:49:25PM +0200, srinivas.kandagatla@xxxxxxxxxx wrote:
> > > > > > From: Srinivas Kandagatla <srinivas.kandagatla@xxxxxxxxxx>
> > > > > > 
> > > > > > MBHC (MultiButton Headset Control) support is available in pm8921 in two
> > > > > > blocks, one to detect mechanical headset insertion and removal and other
> > > > > > block to support headset type detection and 5 button detection and othe
> > > > > > features like impedance calculation.
> > > > > > 
> > > > > > This patch adds support to:
> > > > > > 1> Support to NC and NO type of headset Jacks.
> > > > > > 2> Mechanical insertion and detection of headset jack.
> > > > > > 3> Detect a 3 pole Headphone and a 4 pole Headset.
> > > > > > 4> Detect 5 buttons.
> > > > > > 
> > > > > > Tested it on DB410c with Audio Mezz board with 4 pole and 3 pole
> > > > > > headset/headphones.
> > > > > 
> > > > > I have the same issue with this patchset, KEY_MEDIA is being reported
> > > > > when unplugging a headset. I added a few traces and what I observe is
> > > > Am unable to reproduce the same issue, I tried atleast 6 different mix of
> > > > headset/headphones.
> > > > here is my evtest log:  http://paste.ubuntu.com/25282592/
> > > > 
> > > > Could you explain bit more about your setup, so that I can try to reproduce
> > > > the same issue.
> > > > 
> > > > My setup is DB410c + Audio Mezz board
> > > > https://www.arrow.com/en/products/audiomezz/seeed-technology-limited
> > > 
> > > I'm using a hardware based on Intrinsyc Open-Q 410 Development Kit:
> > > https://www.intrinsyc.com/snapdragon-embedded-development-kits/snapdragon-410-development-kit/
> > > 
> > > We use the same SoM with a different carrier board, but connections to
> > > the jack connector are routed the same way on both carrier boards.
> > > 
> > > > > that the "button pressed" irq is fired when unplugging, but no "button
> > > > > released" irq follows. To get a "button released" irq, I need to either
> > > > > plug a headset, or plug-and-unplug headphones.
> > > > > 
> > > > > So basically, detection fails because we don't get the "button pressed"
> > > > > irq prior to the mechanical switch irq. Do you know what could explain
> > > > > the MBHC not firing the "button released" irq when unplugging headset?
> > > > > 
> > > > Can you also share output of evetest and /proc/interrupts so that i can see
> > > > if the extra logic of masking btn0 is creating the issue, we can also try
> > > > removing the btn0 accessory detect logic and see if we get correct button
> > > > press/releases.
> > > 
> > > It's definitely what's causing the headphones to be reported as headset,
> > > but not what's causing the interrupt not to be triggered in the first
> > > place. I don't have access to the hardware right now, so I'll provide
> > > you with logs next week.
> > 
> > I did the test, and addded the content of /proc/interrupts.
> > 
> >      https://paste.ubuntu.com/25312443/
> > 
> > As you can see, after "[INSERT HEADPHONE]", we only get "mbhc switch
> > irq", but not "mbhc btn press irq". Let me know if you need more
> > details.
> > 
> > As I said before, I don't see that 100% of the time, but I don't do
> > anything special to trigger it either (like pulling the cable extra
> > slowly or whatnot), it just happens.
> 
> Looking at the interrupts count its clear that there was a real button press
> event from the hardware, so My suspect is the voltage threshold values
> programmed for the buttons, which might not be very much correct if you are
> using same values as db410c. These values are partly board specific.
> 
> Can you tune those values and give it a try, specially btn0 values.

Correct me if I'm wrong, but if the design is the same, the values
should be more a less the same as well, right? Moreover, the driver
detects the right key press in normal use cases.

Anyway, I tried to tweak the values but ended up with the same results.
Pressing btn0 brings the DC tension to 0V, so I'm not sure changing the
detection threshold of btn0 has any impact.

Another explanation could be the jack connector. I did a few tests on
the Open-Q 410 [1] and could not reproduce the issue of the spurious key
when unplugging headset, it only happens on the in-house device. So I
guess they have slightly different jack connectors and only one kind
triggers that issue. With that conclusion:


Tested-by: Damien Riegel <damien.riegel@xxxxxxxxxxxxxxxxxxxx>

[1] https://www.intrinsyc.com/snapdragon-embedded-development-kits/snapdragon-410-development-kit/

Thank you,
-- 
Damien