[PATCH v13 00/10] mux controller abstraction and iio/i2c muxes

From: Peter Rosin
Date: Thu Apr 13 2017 - 12:41:33 EST

Hi Greg!

Please apply.

This adds a new mux controller subsystem with an interface for accessing
mux controllers, along with two drivers providing the interface (gpio
and adg792) and two consumers (iio and i2c). This is done in such a way
that several consumers can independently access the same mux controller
if one controller controls several multiplexers, thus allowing sharing.
But sharing is by no means required, of course. It is perfectly fine to
have a single consumer of a dedicated mux controller controlling only
one mux for said consumer.

The prediction is that the typical use case will be for gpio-based muxing
(which is also what drove the development), where the below schematics
show the flexibility with one gpio-based mux controller being shared by
the iio-mux and i2c-mux-gpmux drivers.

.----.
|GPO0|----------.
|GPO1|--------. |
| | | |
| | .-------.
| | |dg4052a|
| | | |
|ADC0|-----|X X0|---- signal X0
| | | X1|---- signal X1
| | | X2|---- signal X2
| | | X3|---- signal X3
| | | |
|SDA0|-----|Y Y0|---- i2c segment Y0
|SCL0|--. | Y1|---- i2c segment Y1
'----' | | Y2|---- i2c segment Y2
| | Y3|---- i2c segment Y3
| '-------'
| 0 1 2 3 (feed SCL0 to each of
| | | | | the 4 muxed segments)
'----------------+-+-+-'

GPO0 and GPO1 may also be fed to further parallel muxers, which is perhaps
desired in a real application to minimize digital noise from the i2c Y
channel leaking into the analog X channel. I.e. it might be a good idea
to separate the analog and digital signals...

And the below hypothetical schematics indicate something similar but using
the i2c-based adg792a multiplexer instead.

.----.
|SDA0|----------.
|SCL0|--------. |
| | | |
| | .-------.
| | |adg792a|
| | | |
|ADC0|-----|D1 S1A|---- signal S1A
| | | S1B|---- signal S1B
| | | S1C|---- signal S1C
| | | S1D|---- signal S1D
| | | |
|SDA1|---+-|D2 S2A|---- i2c segment S2A
|SCL1|-. | | S2B|---- i2c segment S2B
'----' | | | S2C|---- i2c segment S2C
| | | S2D|---- i2c segment S2D
| | | |
| '-|D3 S3A|---- i2c segment S3A
| | S3B|---- i2c segment S3B
| | S3C|---- i2c segment S3C
| | S3D|---- i2c segment S3D
| '-------'
| A B C D A B C D (feed SCL1 to each of
| | | | | | | | | the 8 muxed segments)
'-----------------+-+-+-+---+-+-+-'


v12 -> v13 changes
- Philipp Zabel noticed a bad compatible string in the gpio muxer.
I amended patch 3/10 with his patch.
- Fixed reversed sense of the reset bit in the adg792 muxer (patch 10/10).

v11 -> v12 changes
- patches 11 and 12 folded into patches 6 and 3 respectively.

v10 -> v11 changes
- added a fixes-tag and an ack from Jonathan on patch 11
- added a new patch (12) with a fix for messed up error path reported
by Paul Gortmaker.
- fixed some editorial nitpicks in the documentation comments in patch 3.

v9 -> v10 changes
- rebased onto v4.11-rc1
- added reviewed-by tags from Rob on patches 7 and 9
- added a new patch (11) with a fix for an unsigned compare with less than
zero detected by CoverityScan and reported by Colin Ian King
- allowed the mux core to be built as a module, after discussion with Paul
Gortmaker
- added explicit includes of linux/export.h and linux/init.h from the mux
core, also noted by Paul
- fixed trivial whitespace issue in drivers/mux/Makefile
- added trailing '>' after my mail address in MODULE_AUTHOR, which was missing
in all new modules in drivers/mux

v8 -> v9 changes
- dropped the suffix from the compatible string of the i2c-mux-simple
binding (was ,mux-locked or ,parent-locked) and add an optional
mux-locked property instead to change the desired locking behavior
from the default parent-locked
- add description of the difference between mux-locked and parent-locked
- renamed i2c-mux-simple to i2c-mux (bindings for this general purpose
i2c mux are in i2c-mux-gpmux.txt since i2c-mux.txt is already occupied
by the common i2c-mux bindings)
- changed compatible from mux-gpio to gpio-mux
- changed bindings for idle-state back to a single array, but add defines
for as-is and hi-Z thus avoiding magic numbers
- make use of the above defines in the code as well
- make idle-state a common mux property described in mux-controller.txt
instead of repeating the info in individual mux controller bindings
- drop the adi,parallel property from the adg792 bindings and piggy-back
on the #mux-control-cells property
- refrain from using the compatible string as node name
- dropped the simplified bindings for single-user gpio mux
- added acks on patches 2/10 and 5/10 from Rob

v7 -> v8 changes
- cleanup the implementation of the simplified gpio bindings, but still...
- ...reorder patches so that they appear last in the series (patches 11
and 12 were patches 4 and 5 in v7) since Jonathan convinced me that
they were perhaps not such a good idea after all. But I still wanted
to show the last version I had and I'm still a bit undecided...
- added some words to the remaining otherwise empty commit messages
- added various acks/reviews from Jonathan and Wolfram.
- move mux last in drivers/Kconfig and drivers/Makefile
- bump copyright years
- centralize error reporting of common operatinons to the mux core
- add WARN_ON for faulty usage of mux_chip_register
- simplify code for other WARN_ON call sites

v6 -> v7 changes
- move from drivers/misc/mux-* to drivers/mux/
[I will remove Cc:s to Arnd and Greg for v8, when/if that happens]
- add managed versions of mux_chip_alloc and mux_chip_register
- use the above in mux-gpio.c and mux-adg792a.c
- also use them to support a simplified binding of a gpio mux for the common
case where there is a single consumer (and no specific idle requirements)
- new binding for describing idle behaviour of mux-adg792a
- add bindings for the gpo-pins on the mux-adg792g
- use device_property_read_u32 instead of of_property_read_u32 in mux-gpio.c
- rename iio mux compatible to io-channel-mux (was iio-mux)
- remove linuxism in the bindings (it was mentioning a function name)
- add missing quote in the example in the io-channel-mux binding
- factor out preparatory cleanup of devres docs to its own patch
- add blank line in mux_chip_free
- use SIMPLE_{PARENT,MUX}_LOCKED instead of magic numbers {0,1} in
i2c-mux-simple.c
- add some acks and a reviewed-by from Jonathan

v5 -> v6 changes
- fix stupidity in mux_chip_priv, mux_gpio_remove and adg792a_remove.
- change the devicetree bindings for the iio-mux to use a list of strings
(channels property) instead of a list children.

v4 -> v5 changes
- remove support for fancier dt layouts and go back to the phandle
approach from v2 and before, killing the horrible non-working
refcounting crap from v4 and avoiding a bunch of life-time issues
in v3.
- introduce the concept of a mux-chip, that can hold one or more
mux-controllers (inspired by the pwm subsystem).
- add dt #mux-control-cells property needed to get to the desired
mux controller if a mux chip provides more than one.
- take away the option to build the mux-core as a module.
- if the mux controller has an idle state, make sure the mux controller
is set up in the idle state initially (when it should be idle).
- do not use a variable length array on the stack in mux_gpio_set to
temporarily store the gpio state, preallocate space instead.
- fix resource leak on one failure path in mux_gpio_probe.
- driver for Analog Devices ADG792A/G, literally the first mux chip
I found on the Internet with an i2c interface (that was not a
dedicated i2c multiplexer like PCA9547) which I used to verify
that the abstractions in the mux core are up to the task. Untested,
just proof of concept that at least looks pretty and compiles...
- various touch-ups.

v3 -> v4 changes
- rebased onto next-20161122 (depends on recent _available iio changes).
- added support for having the mux-controller in a child node of a
mux-consumer if it is a sole consumer, to hopefully even further satisfy
the complaint from Rob (and later Lars-Peter) about dt complexity.
- the above came at the cost of some rather horrible refcounting code,
please review and suggest how it should be done...
- changed to register a device class instead of a bus.
- pass in the parent device into mux_control_alloc and require less
work from mux-control drivers.
- changed device names from mux:control%d to mux%d
- move kernel-doc from mux-core.c to mux.h (and add some bits).
- give the gpio driver a chance to update all mux pins at once.
- factor out iio ext_info lookup into new helper function. /Lars-Peter
- use an unsigned type for the iio ext_info count. /Lars-Peter
- unified "brag strings" in the file headers.

v2 -> v3 changes
- have the mux-controller in the parent node of any mux-controller consumer,
to hopefully satisfy complaint from Rob about dt complexity.
- improve commit message of the mux subsystem commit, making it more
general, as requested by Jonathan.
- remove priv member from struct mux_control and calculate it on the
fly. /Jonathan
- make the function comments in mux-core.c kernel doc. /Jonathan
- add devm_mux_control_* to Documentation/driver.model/devres.txt. /Jonathan
- add common dt bindings for mux-controllers, refer to them from the
mux-gpio bindings. /Rob
- clarify how the gpio pins map to the mux state. /Rob
- separate CONFIG_ variables for the mux core and the mux gpio driver.
- improve Kconfig help texts.
- make CONFIG_MUX_GPIO depend on CONFIG_GPIOLIB.
- keep track of the number of mux states in the mux core.
- since the iio channel number is used as mux state, it was possible
to drop the state member from the mux_child struct.
- cleanup dt bindings for i2c-mux-simple, it had some of copy-paste
problems from ots origin (i2c-mux-gpio).
- select the mux control subsystem in config for the i2c-mux-simple driver.
- add entries to MAINTAINERS and my sign-off, I'm now satisfied and know
nothing in this to be ashamed of.

v1 -> v2 changes
- fixup export of mux_control_put reported by kbuild
- drop devicetree iio-ext-info property as noted by Lars-Peter,
and replace the functionality by exposing all ext_info
attributes of the parent channel for each of the muxed
channels. A cache on top of that and each muxed channel
gets its own view of the ext_info of the parent channel.
- implement idle-state for muxes
- clear out the cache on failure in order to force a mux
update on the following use
- cleanup the probe of i2c-mux-simple driver
- fix a bug in the i2c-mux-simple driver, where failure in
the selection of the mux caused a deadlock when the mux
was later unconditionally deselected.


Old background:

I have a piece of hardware that is using the same 3 GPIO pins
to control four 8-way muxes. Three of them control ADC lines
to an ADS1015 chip with an iio driver, and the last one
controls the SDA line of an i2c bus. We have some deployed
code to handle this, but you do not want to see it or ever
hear about it. I'm not sure why I even mention it. Anyway,
the situation has nagged me to no end for quite some time.

So, after first getting more intimate with the i2c muxing code
and later discovering the drivers/iio/inkern.c file and
writing a couple of drivers making use of it, I came up with
what I think is an acceptable solution; add a generic mux
controller driver (and subsystem) that is shared between all
instances, and combine that with an iio mux driver and a new
general purpose i2c mux driver that is only hooking the i2c
muxing and the new mux controller.

One thing that I would like to do, but don't see a solution
for, is to move the mux control code that is present in
various drivers in drivers/i2c/muxes to this new minimalistic
muxing subsystem, thus converting all present i2c muxes (but
perhaps not gates and arbitrators).

I'm using an rwsem to lock a mux, but that isn't really a
perfect fit. Is there a better locking primitive that I don't
know about that fits better? I had a mutex at one point, but
that didn't allow any concurrent accesses at all. At least
the rwsem allows concurrent access as long as all users
agree on the mux state, but I suspect that the rwsem will
degrade to the mutex situation pretty quickly if there is
any contention.

Also, the "mux" name feels a bit ambitious, there are many muxes
in the world, and this tiny bit of code is probably not good
enough to be a nice fit for all...

Cheers,
peda

Peter Rosin (10):
devres: trivial whitespace fix
dt-bindings: document devicetree bindings for mux-controllers and
gpio-mux
mux: minimal mux subsystem and gpio-based mux controller
iio: inkern: api for manipulating ext_info of iio channels
dt-bindings: iio: io-channel-mux: document io-channel-mux bindings
iio: multiplexer: new iio category and iio-mux driver
dt-bindings: i2c: i2c-mux: document general purpose i2c-mux bindings
i2c: i2c-mux-gpmux: new driver
dt-bindings: mux-adg792a: document devicetree bindings for ADG792A/G
mux
mux: adg792a: add mux controller driver for ADG792A/G

.../devicetree/bindings/i2c/i2c-mux-gpmux.txt | 99 +++++
.../bindings/iio/multiplexer/io-channel-mux.txt | 39 ++
.../devicetree/bindings/mux/adi,adg792a.txt | 75 ++++
Documentation/devicetree/bindings/mux/gpio-mux.txt | 69 ++++
.../devicetree/bindings/mux/mux-controller.txt | 157 +++++++
Documentation/driver-model/devres.txt | 10 +-
MAINTAINERS | 15 +
drivers/Kconfig | 2 +
drivers/Makefile | 1 +
drivers/i2c/muxes/Kconfig | 13 +
drivers/i2c/muxes/Makefile | 1 +
drivers/i2c/muxes/i2c-mux-gpmux.c | 173 ++++++++
drivers/iio/Kconfig | 1 +
drivers/iio/Makefile | 1 +
drivers/iio/inkern.c | 60 +++
drivers/iio/multiplexer/Kconfig | 18 +
drivers/iio/multiplexer/Makefile | 6 +
drivers/iio/multiplexer/iio-mux.c | 459 +++++++++++++++++++++
drivers/mux/Kconfig | 46 +++
drivers/mux/Makefile | 7 +
drivers/mux/mux-adg792a.c | 141 +++++++
drivers/mux/mux-core.c | 422 +++++++++++++++++++
drivers/mux/mux-gpio.c | 114 +++++
include/dt-bindings/mux/mux.h | 16 +
include/linux/iio/consumer.h | 37 ++
include/linux/mux.h | 252 +++++++++++
26 files changed, 2233 insertions(+), 1 deletion(-)
create mode 100644 Documentation/devicetree/bindings/i2c/i2c-mux-gpmux.txt
create mode 100644 Documentation/devicetree/bindings/iio/multiplexer/io-channel-mux.txt
create mode 100644 Documentation/devicetree/bindings/mux/adi,adg792a.txt
create mode 100644 Documentation/devicetree/bindings/mux/gpio-mux.txt
create mode 100644 Documentation/devicetree/bindings/mux/mux-controller.txt
create mode 100644 drivers/i2c/muxes/i2c-mux-gpmux.c
create mode 100644 drivers/iio/multiplexer/Kconfig
create mode 100644 drivers/iio/multiplexer/Makefile
create mode 100644 drivers/iio/multiplexer/iio-mux.c
create mode 100644 drivers/mux/Kconfig
create mode 100644 drivers/mux/Makefile
create mode 100644 drivers/mux/mux-adg792a.c
create mode 100644 drivers/mux/mux-core.c
create mode 100644 drivers/mux/mux-gpio.c
create mode 100644 include/dt-bindings/mux/mux.h
create mode 100644 include/linux/mux.h

--
2.1.4