Message-ID: <20240429204027.3e47074a@jic23-huawei>
Date: Mon, 29 Apr 2024 20:40:27 +0100
From: Jonathan Cameron <jic23@...nel.org>
To: "Gradinariu, Ramona" <Ramona.Gradinariu@...log.com>
Cc: Nuno Sá <noname.nuno@...il.com>, Ramona Gradinariu
 <ramona.bolboaca13@...il.com>, "linux-kernel@...r.kernel.org"
 <linux-kernel@...r.kernel.org>, "linux-iio@...r.kernel.org"
 <linux-iio@...r.kernel.org>, "linux-doc@...r.kernel.org"
 <linux-doc@...r.kernel.org>, "devicetree@...r.kernel.org"
 <devicetree@...r.kernel.org>, "corbet@....net" <corbet@....net>,
 "conor+dt@...nel.org" <conor+dt@...nel.org>,
 "krzysztof.kozlowski+dt@...aro.org" <krzysztof.kozlowski+dt@...aro.org>,
 "robh@...nel.org" <robh@...nel.org>, "Sa, Nuno" <Nuno.Sa@...log.com>
Subject: Re: [PATCH 4/5] iio: adis16480: add support for adis16545/7
 families

On Mon, 29 Apr 2024 13:17:42 +0000
"Gradinariu, Ramona" <Ramona.Gradinariu@...log.com> wrote:

> > -----Original Message-----
> > From: Nuno Sá <noname.nuno@...il.com>
> > Sent: Monday, April 29, 2024 10:59 AM
> > To: Jonathan Cameron <jic23@...nel.org>; Ramona Gradinariu
> > <ramona.bolboaca13@...il.com>
> > Cc: linux-kernel@...r.kernel.org; linux-iio@...r.kernel.org; linux-
> > doc@...r.kernel.org; devicetree@...r.kernel.org; corbet@....net;
> > conor+dt@...nel.org; krzysztof.kozlowski+dt@...aro.org; robh@...nel.org;
> > Gradinariu, Ramona <Ramona.Gradinariu@...log.com>; Sa, Nuno
> > <Nuno.Sa@...log.com>
> > Subject: Re: [PATCH 4/5] iio: adis16480: add support for adis16545/7 families
> > 
> > [External]
> > 
> > On Sun, 2024-04-28 at 16:25 +0100, Jonathan Cameron wrote:  
> > > On Tue, 23 Apr 2024 11:42:09 +0300
> > > Ramona Gradinariu <ramona.bolboaca13@...il.com> wrote:
> > >  
> > > > The ADIS16545 and ADIS16547 are a complete inertial system that
> > > > includes a triaxis gyroscope and a triaxis accelerometer.
> > > > The serial peripheral interface (SPI) and register structure provide a
> > > > simple interface for data collection and configuration control.
> > > >
> > > > These devices are similar to the ones already supported in the driver,
> > > > with changes in the scales, timings and the max spi speed in burst
> > > > mode.
> > > > Also, they support delta angle and delta velocity readings in burst
> > > > mode, for which support was added in the trigger handler.
> > > >
> > > > Signed-off-by: Nuno Sá <nuno.sa@...log.com>  
> > >
> > > What is Nuno's relationship to this patch?  You are author and the sender
> > > which is fine, but in that case you need to make Nuno's involvement explicit.
> > > Perhaps a Co-developed-by or similar is appropriate?
> > >  
> > > > Signed-off-by: Ramona Gradinariu <ramona.gradinariu@...log.com>  
> > > A few comments inline.  Biggest one is I'd like a clear statement of why you
> > > can't do a burst of one type, then a burst of other.  My guess is that the
> > > transition is very time consuming?  If so, that is fine, but you should be
> > > able
> > > to let available_scan_masks handle the disjoint channel sets.  
> > 
> > Yeah, the burst message is a special spi transfer that brings you all of the
> > channels data at once but for the accel/gyro you need to explicitly configure
> > the chip to either give you the "normal vs "delta" readings. Re-configuring the
> > chip and then do another burst would destroy performance I think. We could
> > do
> > the manual readings as we do in adis16475 for chips not supporting burst32.
> > But
> > in the burst32 case those manual readings should be minimal while in here we
> > could have to do 6 of them which could also be very time consuming...
> > 
> > Now, why we don't use available_scan_masks is something I can't really
> > remember
> > but this implementation goes in line with what we have in the adis16475
> > driver.
> > 
> > - Nuno Sá
> >   
> 
> Thank you Nuno for all the additional explanations.
> Regarding the use of available_scan_masks, the idea is to have any possible
> combination for accel, gyro, temp and timestamp channels or delta angle, delta 
> velocity, temp and  timestamp channels. There are a lot of combinations for 
> this and it does not seem like a good idea to write them all manually. That is 
> why adis16480_update_scan_mode is used for checking the correct channels 
> selection.

If you are using bursts, the data is getting read anyway - which is the main
cost here. The real question becomes what are you actually saving by supporting all
the combinations of the the two subsets of channels in the pollfunc?
Currently you have to pick the channels out and repack them, if pushing them all
looks to me like a mempcy and a single value being set (unconditionally).

Then it's a question of what the overhead of the channel demux in the core is.
What you pass out of the driver via iio_push_to_buffers*()
is not what ends up in the buffer if you allow the IIO core to do data demuxing
for you - that is enabled by providing available_scan_masks.  At buffer
start up the demux code computes a fairly optimal set of copies to repack
the incoming data to match with what channels the consumer (here probably
the kfifo on the way to userspace) is expecting.

That demux adds a small overhead but it should be small as long
as the channels wanted aren't pathological (i.e. every other one).

Advantage is the driver ends up simpler and in the common case of turn
on all the channels (why else did you buy a device with those measurements
if you didn't want them!) the demux is zerocopy so effectively free which
is not going to be the case for the bitmap walk and element copy in the
driver.

Jonathan

> 
> Ramona G.

Powered by blists - more mailing lists