device: Remove z_device_is_ready #72906
Conversation
zephyrbot
requested review from
andyross,
cfriedt,
dcpleung,
gmarull,
nashif,
npitre,
peter-mitsis,
TaiJuWu and
tbursztyka
| @@ -40,13 +40,13 @@ const struct device *z_impl_device_get_binding(const char *name) | |||
| * performed. Reserve string comparisons for a fallback. | |||
| */ | |||
| STRUCT_SECTION_FOREACH(device, dev) { | |||
| if (z_device_is_ready(dev) && (dev->name == name)) { | |||
| if (z_impl_device_is_ready(dev) && (dev->name == name)) { | |||
There was a problem hiding this comment.
why are you calling the implementation directly here and not using device_is_ready instead?
There was a problem hiding this comment.
When we get here, it is already in kernel mode, so there is no need to do the check again whether you are in user mode or kernel mode (as in the syscall wrapper). Note that the wrapper calls the z_impl_* directly if already in kernel mode.
There was a problem hiding this comment.
do we agree that calling implementation (z_impl) in general should not be the norm and that any direct usage of the implementation function need some justification? If yes, this justifcation should be in the commit message, i.e. what @dcpleung wrote above.
There was a problem hiding this comment.
@nashif I have update the commit msg. Please take a look again.
This duplicates the functionality of device_is_ready. Calls for z_device_is_ready are being done in kernel mode, so it is safe to call its implementation directly. Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
This duplicates the functionality of device_is_ready.