On 12/18/23 01:08, Matyas, Daniel wrote:
-----Original Message-----
From: Guenter Roeck <groeck7@xxxxxxxxx> On Behalf Of Guenter Roeck
Sent: Friday, December 15, 2023 6:03 PM
To: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>; Matyas, Daniel
<Daniel.Matyas@xxxxxxxxxx>
Cc: Jean Delvare <jdelvare@xxxxxxxx>; Rob Herring
<robh+dt@xxxxxxxxxx>; Krzysztof Kozlowski
<krzysztof.kozlowski+dt@xxxxxxxxxx>; Conor Dooley
<conor+dt@xxxxxxxxxx>; Jonathan Corbet <corbet@xxxxxxx>; linux-
hwmon@xxxxxxxxxxxxxxx; devicetree@xxxxxxxxxxxxxxx; linux-
kernel@xxxxxxxxxxxxxxx; linux-doc@xxxxxxxxxxxxxxx
Subject: Re: [PATCH 2/3] bindings: hwmon: Add adi,adaq4224_temp as
compatible string
[External]
On 12/15/23 00:49, Krzysztof Kozlowski wrote:
On 14/12/2023 15:36, Daniel Matyas wrote:
In the device ada4224 the max31827 temperature sensor will be used,
so the default values corresponding to adaq4224_temp are the same
for
max31827.
Signed-off-by: Daniel Matyas <daniel.matyas@xxxxxxxxxx>
Please use subject prefixes matching the subsystem. You can get them
for example with `git log --oneline -- DIRECTORY_OR_FILE` on the
directory your patch is touching.
---
Documentation/devicetree/bindings/hwmon/adi,max31827.yaml | 5
++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git
a/Documentation/devicetree/bindings/hwmon/adi,max31827.yaml
b/Documentation/devicetree/bindings/hwmon/adi,max31827.yaml
index f60e06ab7d0a..9f3b0839aa46 100644
--- a/Documentation/devicetree/bindings/hwmon/adi,max31827.yaml
+++
b/Documentation/devicetree/bindings/hwmon/adi,max31827.yaml
@@ -20,6 +20,7 @@ properties:
- const: adi,max31827
- items:
- enum:
+ - adi,adaq4224_temp
Underscores are not allowed
That isn't the main problem with this patch.
https://urldefense.com/v3/__https://github.com/analogdevicesinc/linux/
tree/dev_adaq4224_dts__;!!A3Ni8CS0y2Y!_2D1w1DD5sjJrNyArZYZ3QW9
nS8URmP6X0n6R7q1sBnDB1HPL6jROhD_w9u3fJixt2hDDNtO6VpgLM1Jka
Q$
suggests that it may be a development system which utilizes the
max31827.
From there, we can see that there is a devicetree description of a board
with that name which uses
temperature1: temperature@5f {
compatible = "adi,adaq4224_temp";
reg = <0x5f>;
vref-supply = <&vio>;
adi,comp-int;
adi,alarm-pol = <0>;
adi,fault-q = <1>;
adi,timeout-enable;
};
That doesn't make sense to me. I don't know why they don't just
reference max31827.
I am most definitely not going to accept a driver change just to map
adi,adaq4224_temp (or adi,adaq4224-temp) to the actually used
temperature sensor chip. If we start accepting that, we'd end up with no
end of "<vendor>,<board_name>-{temp,voltage,current,power,...}"
compatibles.
Looking more into the above mentioned repository/branch, an earlier
version of the dts file did reference adi,max31827 for that chip. It also
looks like there may be an adaq4224 ADC (per drivers/iio/adc/ad4630.c),
but that would be a SPI chip. It seems highly unlikely that a SPI ADC would
have a separate I2C interface connected to a temperature sensor.
Confusing.
There is also some indication that this may some IP to be loaded into an
FPGA.
which utilizes an FPGA implementation of max31827 (or maybe connects
to one).
If that is the case, it should still be referenced as max31827.
All that wasted time because of "let's provide as little as possible
information about what we are actually doing" :-(.
Guenter
I asked around to get some more clarification on the matter. There will be a new chip released, named adaq4224. This chip will have the max31827 implemented in the silicon, so the driver used to get temperature information would be max31827.c. The chip will have spi and i2c communication too. The other driver you mentioned, the ad4630.c will communicate through spi.
Because the chip has a different name, I was asked to add a new label for the max31827, so that it will be clear for the user, that the max31827 is part of the chip.
It is still a max31827 core. If there is no difference in programming
to max31827, it can possibly be reflected in devicetree as
"adi,adaq4224-temp", in addition to "adi,max31827". This would not
require a change in the driver code as "adi,max31827" would still
be mandatory.
But I don't think changing the driver or device name for this purpose
would be appropriate. We don't do that for any of the other hwmon
drivers, and I really don't want to open a floodgate. The lm75 driver
supports more than 20 chips from various different vendors,
but it is still the lm75 driver. The jc42 driver supports many chips
from different vendors, but the driver only has a single
"jedec,jc-42.4-temp" binding, and devices are instantiated as "jc42".
I don't see why this one should be handled differently.
Guenter
