drivers: dai: intel: dmic: don't use assert for error handling #53394
Conversation
kv2019i
requested review from
marcinszkudlinski and
abonislawski
as code owners
The dai.h interface does not prohibit calling dai_config_get() with different direction values. The dmic driver should handle invalid direction value explicitly and not rely on an assert. Link: thesofproject/sof#6896 Signed-off-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
kv2019i
force-pushed
the
202212-dmic-get-config-fix
branch
from
ba4ffcd to
96366a3
Compare
| @@ -743,7 +743,9 @@ static const struct dai_config *dai_dmic_get_config(const struct device *dev, en | |||
| { | |||
| struct dai_intel_dmic *dmic = (struct dai_intel_dmic *)dev->data; | |||
|
|
|||
| __ASSERT_NO_MSG(dir == DAI_DIR_RX); | |||
| if (dir != DAI_DIR_RX) | |||
There was a problem hiding this comment.
Are you sure the exit value is checked everywhere?
i.e.
dai-zephyr.c, line 159:
dereferenced without checking for null
same in dai.c line 157:
Please add checking for null at every call to the function
EDIT: the references in question are in SOF code, not in zephyr, anyway NULL checking should be introduced prior to removing assert
There was a problem hiding this comment.
@marcinszkudlinski Ack, let me check. I was only fixing cases where the assert is already hit now (meaning you cannot actually enable asserts in builds to catch other errors). But true, need to check all cases.
There was a problem hiding this comment.
@marcinszkudlinski both cases now handled thesofproject/sof#6897 and thesofproject/sof#6899
Given a) there are no Zephyr in-tree inproper usage and b) asserts are not enable by default (yet) in SOF builds, I'd propose we proceed and merge this fix.
The dai.h interface does not prohibit calling dai_config_get() with different direction values. The dmic driver should handle invalid direction value explicitly and not rely on an assert.
Link: thesofproject/sof#6896
Signed-off-by: Kai Vehmanen kai.vehmanen@linux.intel.com