Message-ID: <9e88fd9640ccd9a8d1e3a45a4e328797a7d3a9bf.camel@analog.com>
Date:   Sat, 16 May 2020 13:08:46 +0000
From:   "Ardelean, Alexandru" <alexandru.Ardelean@...log.com>
To:     "jic23@...nel.org" <jic23@...nel.org>,
        "lars@...afoo.de" <lars@...afoo.de>
CC:     "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
        "linux-iio@...r.kernel.org" <linux-iio@...r.kernel.org>
Subject: Re: [RFC PATCH 00/14] iio: buffer: add support for multiple buffers

On Tue, 2020-05-12 at 06:26 +0000, Ardelean, Alexandru wrote:
> [External]
> 
> On Mon, 2020-05-11 at 21:56 +0200, Lars-Peter Clausen wrote:
> > [External]
> > 
> > On 5/11/20 4:56 PM, Ardelean, Alexandru wrote:
> > > On Mon, 2020-05-11 at 15:58 +0200, Lars-Peter Clausen wrote:
> > > > [External]
> > > > 
> > > > On 5/11/20 3:24 PM, Ardelean, Alexandru wrote:
> > > > > On Mon, 2020-05-11 at 13:03 +0000, Ardelean, Alexandru wrote:
> > > > > > [External]
> > > > > > 
> > > > > > On Mon, 2020-05-11 at 12:37 +0200, Lars-Peter Clausen wrote:
> > > > > > > [External]
> > > > > > > 
> > > > > > > On 5/11/20 12:33 PM, Ardelean, Alexandru wrote:
> > > > > > > > On Sun, 2020-05-10 at 11:09 +0100, Jonathan Cameron wrote:
> > > > > > > > > [External]
> > > > > > > > > 
> > > > > > > > > On Sat, 9 May 2020 10:52:14 +0200
> > > > > > > > > Lars-Peter Clausen <lars@...afoo.de> wrote:
> > > > > > > > > 
> > > > > > > > > > On 5/8/20 3:53 PM, Alexandru Ardelean wrote:
> > > > > > > > > > > [...]
> > > > > > > > > > > What I don't like, is that iio:device3 has iio:buffer3:0
> > > > > > > > > > > (to
> > > > > > > > > > > 3).
> > > > > > > > > > > This is because the 'buffer->dev.parent = &indio_dev-
> > > > > > > > > > > >dev'.
> > > > > > > > > > > But I do feel this is correct.
> > > > > > > > > > > So, now I don't know whether to leave it like that or
> > > > > > > > > > > symlink to
> > > > > > > > > > > shorter
> > > > > > > > > > > versions like 'iio:buffer3:Y' -> 'iio:device3/bufferY'.
> > > > > > > > > > > The reason for naming the IIO buffer devices to
> > > > > > > > > > > 'iio:bufferX:Y'
> > > > > > > > > > > is
> > > > > > > > > > > mostly to make the names unique. It would have looked
> > > > > > > > > > > weird
> > > > > > > > > > > to
> > > > > > > > > > > do
> > > > > > > > > > > '/dev/buffer1' if I would have named the buffer devices
> > > > > > > > > > > 'bufferX'.
> > > > > > > > > > > 
> > > > > > > > > > > So, now I'm thinking of whether all this is acceptable.
> > > > > > > > > > > Or what is acceptable?
> > > > > > > > > > > Should I symlink 'iio:device3/iio:buffer3:0' ->
> > > > > > > > > > > 'iio:device3/buffer0'?
> > > > > > > > > > > What else should I consider moving forward?
> > > > > > > > > > > What means forward?
> > > > > > > > > > > Where did I leave my beer?
> > > > > > > > > > Looking at how the /dev/ devices are named I think we can
> > > > > > > > > > provide
> > > > > > > > > > a
> > > > > > > > > > name
> > > > > > > > > > that is different from the dev_name() of the device. Have a
> > > > > > > > > > look
> > > > > > > > > > at
> > > > > > > > > > device_get_devnode() in drivers/base/core.c. We should be
> > > > > > > > > > able
> > > > > > > > > > to
> > > > > > > > > > provide the name for the chardev through the devnode()
> > > > > > > > > > callback.
> > > > > > > > > > 
> > > > > > > > > > While we are at this, do we want to move the new devices
> > > > > > > > > > into
> > > > > > > > > > an
> > > > > > > > > > iio
> > > > > > > > > > subfolder? So iio/buffer0:0 instead of iio:buffer0:0?
> > > > > > > > > Possibly on the folder.  I can't for the life of me remember
> > > > > > > > > why
> > > > > > > > > I
> > > > > > > > > decided
> > > > > > > > > not to do that the first time around - I'll leave it at the
> > > > > > > > > mysterious "it may turn out to be harder than you'd think..."
> > > > > > > > > Hopefully not ;)
> > > > > > > > I was also thinking about the /dev/iio subfolder while doing
> > > > > > > > this.
> > > > > > > > I can copy that from /dev/input
> > > > > > > > They seem to do it already.
> > > > > > > > I don't know how difficult it would be. But it looks like a good
> > > > > > > > precedent.
> > > > > > > All you have to do is return "iio/..." from the devnode()
> > > > > > > callback.
> > > > > > I admit I did not look closely into drivers/input/input.c before
> > > > > > mentioning
> > > > > > this
> > > > > > as as good precedent.
> > > > > > 
> > > > > > But, I looks like /dev/inpput is a class.
> > > > > > While IIO devices are a bus_type devices.
> > > > > > Should we start implementing an IIO class? or?
> > > > > What I should have highlighted [before] with this, is that there is no
> > > > > devnode()
> > > > > callback for the bus_type [type].
> > > > But there is one in device_type :)
> > > Many thanks :)
> > > That worked nicely.
> > > 
> > > I now have:
> > > 
> > > root@...log:~# ls /dev/iio/*
> > > /dev/iio/iio:device0  /dev/iio/iio:device1
> > > 
> > > /dev/iio/device3:
> > > buffer0  buffer1  buffer2  buffer3
> > > 
> > > /dev/iio/device4:
> > > buffer0
> > > 
> > > 
> > > It looks like I can shift these around as needed.
> > > This is just an experiment.
> > > I managed to move the iio devices under /dev/iio, though probably the IIO
> > > devices will still be around as /dev/iio:deviceX for legacy reasons.
> > > 
> > > Two things remain unresolved.
> > > 1. The name of the IIO buffer device.
> > > 
> > > root@...log:/sys/bus/iio/devices# ls iio\:device3/
> > > buffer          in_voltage0_test_mode           name
> > > events          in_voltage1_test_mode           of_node
> > > iio:buffer:3:0  in_voltage_sampling_frequency   power
> > > iio:buffer:3:1  in_voltage_scale                scan_elements
> > > iio:buffer:3:2  in_voltage_scale_available      subsystem
> > > iio:buffer:3:3  in_voltage_test_mode_available  uevent
> > > 
> > > 
> > > Right now, each buffer device is named 'iio:buffer:X:Y'.
> > > One suggesttion was  'iio:deviceX:bufferY'
> > > I'm suspecting the latter is preferred as when you sort the folders,
> > > buffers
> > > come right after the iio:deviceX folders in /sys/bus/iio/devices.
> > > 
> > > I don't feel it matters much the device name of the IIO buffer if we
> > > symlink
> > > it
> > > to a shorter form.
> > >   
> > > I'm guessing, we symlink these devices to short-hand 'bufferY' folders in
> > > each
> > > 'iio:deviceX'?
> > 
> > I think that would be a bit excessive. Only for the legacy buffer we 
> > need to have a symlink.
> > 
> > > [...]
> > > 2. I know this is [still] stupid now; but any suggestions one how to
> > > symlink
> > > /dev/iio:device3 -> /dev/iio/device3/buffer0 ?
> > > 
> > Does not seem to be possible. Userspace will have to take care of it. 
> > This means we need to keep legacy devices in /dev/ and only new buffers 
> > in /dev/iio/.
> 
> One thought about this, was that we keep the chardev for the IIO device for
> this.
> i.e.  /dev/iio:deviceX and /dev/iio/deviceX/buffer0 open the same buffer.
> This means that for a device with 4 buffers, you get 5 chardevs.
> This also seems a bit much/excessive. Maybe also in terms of source-code.
> It would at least mean not moving the event-only chardev to 'industrialio-
> event.c', OR move it, and have the same chardev in 3 places ['industrialio-
> event.c', 'industrialio-buffer.c' & 'industrialio-buffer.c'
> 
> Maybe this sort-of makes sense to have for a few years/kernel-revisions until
> things clean-up.
> 
> I guess at this point, the maintainer should have the final say about this.

Another 'compromise' idea, is that we make this '/dev/iio/deviceX/bufferY' thing
a feature for new devices, and leave '/dev/iio:deviceX' devices [for buffers] a
thing for current devices.
It would mean adding a 'new' iio_device_attach_buffer(); no idea on a name [for
this yet].

Over time, people can convert existing drivers to the new IIO-buffer format, if
they want to. That also gives them a bit better control over symlinking
'/dev/iio:deviceX' -> '/dev/iio/deviceX/bufferY' [or symlinking in reverse if
they want to].

That may create confusion I guess during a transition period.
And it would [ideally] have a mechanism [preferably at build/compile time] to
notify users to use the new IIO buffer mechanism [vs the old one] when adding
new drivers.
Otherwise, there's the risk of people copying the old IIO buffer mechanism.
This can be brought-up at review, but ¯\_(ツ)_/¯ ; it can be annoying.


> 
> > 

Powered by blists - more mailing lists