Re: [BUG] Re: [PATCH] iio: imu: st_lsm6dsx: remove invalid gain value for LSM9DS1

[Jonathan Cameron <jic23@kernel.org>]

On Sat, 31 Aug 2019 11:32:24 +0200
Lorenzo Bianconi <lorenzo@kernel.org> wrote:

> > On 30.08.19 09:23, Lorenzo Bianconi wrote:  
> > >> On 29.08.19 10:37, Lorenzo Bianconi wrote:  
> > >>>> On 27.08.19 22:08, Jonathan Cameron wrote:  
> > >>>>> On Tue, 27 Aug 2019 10:26:35 +0200
> > >>>>> Lorenzo Bianconi <lorenzo@kernel.org> wrote:
> > >>>>>  
> > >>>>>> Get rid of invalid sensitivity value for LSM9DS1 gyro sensor
> > >>>>>>
> > >>>>>> Fixes: 687a60feb9c6 ("iio: imu: st_lsm6dsx: add support for accel/gyro unit of lsm9ds1")
> > >>>>>> Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>  
> > >>>>> The zero degree scale is certainly odd otherwise, so good to tidy
> > >>>>> this up.
> > >>>>>
> > >>>>> Applied to the togreg branch of iio.git.
> > >>>>>  
> > > 
> > > [...]
> > >   
> > >>> struct st_lsm6dsx_fs_table_entry in st_lsm6dsx_settings will always have 4
> > >>> elements for fs_avl array and since the array is defined as static the
> > >>> uninitialized elements are set to 0.
> > >>>
> > >>> Could you please share the ops you are getting?
> > >>>  
> > >>
> > >> How this oops during startup can look like is appended below. I know
> > >> that exactly this change causes it. Can you test this too please?  
> > > 
> > > I did it but I have no issues
> > >   
> > >>
> > >> Given the cleanup nature of this patch, I think it's best to revert it
> > >> in case of any doubt.
> > >>
> > >> thanks,
> > >>
> > >>                           martin
> > >>  
> > > 
> > > is it the full ops? It seems some parts are missing.
> > > Are you running some userspace aps reading in_anglvel_scale_available or reading/writing in in_anglvel_scale?
> > > Could you please double check if the following patch helps? (just compiled)
> > >   
> > 
> > it does not, and you're right, the problem should be somewhere else.
> > I've yet to debug it further.
> > 
> > thanks,  
> 
> Looking at the previous patch I spotted an issue (not related to the one you
> are facing)..actually we can set device gain to 0 forcing to 0 sensor output.
> I will post a formal patch to fix it.
> 

Rather than messing around with a pull request that has already gone
to Greg, I'm going to leave this as it is for now.

The trace doesn't make a lot of sense to me and whilst a bit messy (as fixed
up by Lorenzo's follow up) I can't see why things would crash.

So needs debugging and that isn't a problem given we are only looking
at putting this support into rc1.  Not ideal, but there is time
to work out what is wrong and fix it up!

Thanks,

Jonathan


> Regards,
> Lorenzo
> 
> > 
> >                         martin
> > 
> >