[PATCH v2 1/9] dm: core: Document the common error codes
sjg at chromium.org
Tue Mar 23 05:14:19 CET 2021
Driver model uses quite strong conventions on error codes, but these are
currently not clearly documented. Add a description of the commonly used
Signed-off-by: Simon Glass <sjg at chromium.org>
Changes in v2:
- Add a patch to document the common error codes
doc/driver-model/design.rst | 111 ++++++++++++++++++++++++++++++++++++
1 file changed, 111 insertions(+)
diff --git a/doc/driver-model/design.rst b/doc/driver-model/design.rst
index 4e5cecbab6a..30a07bdf768 100644
@@ -900,6 +900,117 @@ Some special flags are used to determine whether to remove the device:
The dm_remove_devices_flags() function can be used to remove devices based on
their driver flags.
+Driver model tries to use errors codes in a consistent way, as follows:
+ Try later, e.g. dependencies not ready
+ Invalid argument, such as `dev_read_...()` failed or any other
+ devicetree-related access. Also used when a driver method is passed an
+ argument it considers invalid or does not support.
+ Failed to perform I/O
+ Do not bind the device. This should not be used to indicate an
+ error probing the device or for any other purpose, lest driver model get
+ confused. Using `-ENODEV` inside a driver method makes no sense, since
+ clearly there is a device.
+ Entry or object not found
+ Out of memory
+ Ran out of space (e.g. in a buffer or limited-size array)
+ Function not implemented. This is returned by uclasses where the driver does
+ not implement a particular method. It can also be returned by drivers when
+ a particular sub-method is not implemented. This is widely checked in the
+ wider code base, where a feature may or may not be compiled into U-Boot. It
+ indicates that the feature is not available, but this is often just normal
+ operation. Please do not use -ENOSUPP. If an incorrect or unknown argument
+ is provided to a method (e.g. an unknown clock ID), return -EINVAL.
+ Couldn't find device/address
+ This is -1 so some older code may use it as a generic error. This indicates
+ that an operation is not permitted, e.g. a security violation or policy
+ constraint. It is returned internally when binding devices before relocation,
+ if the device is not marked for pre-relocation use.
+ Missing uclass. This is deliberately an uncommon error code so that it can
+ easily be distinguished. If you see this very early in U-Boot, it means that
+ a device exists with a particular uclass but the uclass does not (mostly
+ likely because it is not compiled in). Enable DEBUG in uclass.c or lists.c
+ to see which uclass ID or driver is causing the problem.
+ Cannot talk to peripheral, e.g. i2c
+Less common ones:
+ Attempt to remove a device which does not match the removal flags. See
+ Devicetree read failure, specifically trying to read a string index which
+ does not exist, in a string-listg property
+ Attempt to use a uclass method on a device not in that uclass. This is
+ seldom checked at present, since it is generally a programming error and a
+ waste of code space. A DEBUG-only check would be useful here.
+ Devicetree read error, where a property exists but has no data associated
+ with it
+ Devicetree read error, where the property is longer than expected
+ Attempt to remove a non-vital device when the removal flags indicate that
+ only vital devices should be removed
+ Returned by regmap functions when arguments are out of range. This can be
+ useful for disinguishing regmap errors from other errors obtained while
+ probing devices.
+Drivers should use the same conventions so that things function as expected.
+In particular, if a driver fails to probe, or a uclass operation fails, the
+error code is the primary way to indicate what actually happened.
+Printing error messages in drivers is discouraged due to code size bloat and
+since it can result in messages appearing in normal operation. For example, if
+a command tries two different devices and uses whichever one probes correctly,
+we don't want an error message displayed, even if the command itself might show
+a warning or informational message.
+Error messages can be logged using `log_msg_ret()`, so that enabling
+`CONFIG_LOG` and `CONFIG_LOG_ERROR_RETURN` shows a trace of error codes returned
+through the call stack. That can be a handy way of quickly figuring out where
+an error occurred. Get into the habit of return errors with
+`return log_msg_ret("here", ret)` instead of just `return ret`. The string
+just needs to be long enough to find in a single function, since a log record
+stores (and can print with `CONFIG_LOGF_FUNC`) the function where it was
More information about the U-Boot