Re: [RFC 0/4] Intel Integrated Sensor Hub Support (ISH)

From: Bastien Nocera <hadess@hadess.net>
To: Grant Likely <grant.likely@secretlab.ca>
Cc: Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>,
	Jiri Kosina <jikos@kernel.org>,
	benjamin.tissoires@redhat.com,
	Jonathan Cameron <jic23@kernel.org>,
	"linux-input@vger.kernel.org" <linux-input@vger.kernel.org>,
	linux-iio@vger.kernel.org,
	Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
	chaya.golan@intel.com, daniel.drubin@intel.com,
	Darren Hart <darren@dvhart.com>
Subject: Re: [RFC 0/4] Intel Integrated Sensor Hub Support (ISH)
Date: Tue, 14 Jun 2016 19:01:05 +0200	[thread overview]
Message-ID: <1465923665.606.19.camel@hadess.net> (raw)
In-Reply-To: <CACxGe6siZMA3GX=M8=07GbHsnzRzq0=dKQ8aeC_VN=9iPcc=Ew@mail.gmail.com>

On Tue, 2016-06-14 at 17:08 +0100, Grant Likely wrote:
> On Tue, Jun 14, 2016 at 4:12 PM, Bastien Nocera <hadess@hadess.net>
> wrote:
> > On Tue, 2016-06-14 at 14:44 +0100, Grant Likely wrote:
> > > On Fri, Jun 10, 2016 at 4:27 PM, Bastien Nocera <hadess@hadess.ne
> > > t>
> > > wrote:
> > > > On Fri, 2016-06-10 at 08:23 -0700, Srinivas Pandruvada wrote:
> > > > > On Fri, 2016-06-10 at 17:04 +0200, Bastien Nocera wrote:
> > > > > > > 
> > > > > 
> > > > > [...]
> > > > > 
> > > > > > Are there any errors when setting the triggers?
> > > > > > 
> > > > > Is there any debug option in this service to give more
> > > > > verbose
> > > > > output?
> > > > 
> > > > Add:
> > > > Environment="G_MESSAGES_DEBUG=all"
> > > > 
> > > > To the service file. You should see the debug in systemctl:
> > > > systemctl status iio-sensor-proxy.service
> > > > 
> > > > Or in journalctl if there's too much data:
> > > > journalctl --reverse -u iio-sensor-proxy.service
> > > 
> > > Still no joy on the sensors. The proxy starts up and detects the
> > > sensors, but I don't see any sensor events occurring:
> > 
> > Could you try again with the current master of iio-sensor-proxy?
> > Make
> > sure to install the file in the same location as in your distro
> > provided package.
> > 
> > I've added more debug/warnings to a few cases where it would have
> > silently failed in the past.
> > 
> > In particular, I don't see any messages that would be coming out
> > of:
> > https://github.com/hadess/iio-sensor-proxy/blob/master/src/iio-buff
> > er-utils.c#L512
> > 
> > Could it be that the sub-directory is populated after the device is
> > created in the kernel, causing a race?
> > 
> > Either that, or the contents of the scan_elements/ directory is not
> > the
> > one expected by this code.
> > 
> > The output of this command should us that:
> > ls /sys/devices/pci0000:00/0000:00:13.0/{33AECD58-B679-4E54-9BD9-
> > A04D34F0C226}/001E:8086:22D8.0001/HID-SENSOR-
> > 200073.8.auto/iio:device1/scan_elements
> 
> Hmmm, this is interesting. I fetched, built and ran the latest
> iio-sensor-proxy. It quit with the following error. However,
> immediately afterwards, I ran the Debian packaged version (1.1-1),
> and
> the sensors started to work. Something in the latest master activates
> the sensors enough for the older version to work.

There's no changes in that code other than me introducing a bug that'd
make it throw a warning when the sensor was already enabled. Fixed now.

Could you try again, but make sure to replace the existing iio-sensor-
proxy so it's started on boot, when the device appears. If you see the
warning again, please attach the output out from "ls" as requested in
the earlier mail.

If it doesn't warn on startup, and it fails to work, then we're
probably looking at a bug in the kernel...

Cheers