devicetree: Add DT_CHILD_BY_UNIT_ADDR #99120
Conversation
zephyrbot
added
area: Devicetree
area: Devicetree Bindings
area: Boards/SoCs
area: Tests
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
9bf2b99 to
e3c70ba
Compare
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
e51dfee to
5eac67f
Compare
Fix dts formatting issues with dts-linter. Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
3 times, most recently
from
fa20529 to
2c51d4f
Compare
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
aaddcc3 to
6d71377
Compare
include/zephyr/devicetree.h
Outdated
| * @brief Get a node identifier for a child node by a reg address | ||
| * | ||
| * Example devicetree fragment: | ||
| * | ||
| * @code{.dts} | ||
| * / { | ||
| * soc-label: soc { | ||
| * serial1: serial@40001000 { | ||
| * status = "okay"; | ||
| * current-speed = <115200>; | ||
| * ... | ||
| * }; | ||
| * }; | ||
| * }; |
There was a problem hiding this comment.
in the description you say it's by reg address, but I don't see a reg property here in the example, unless you mean to refer to the unit address, in which case the name of the macro is also wrong
There was a problem hiding this comment.
I've added a proper _BY_UNIT_ADDR variant, which is needed to fix the adc dt spec. I could drop the _BY_REG_ADDR, but it doesn't harm I guess?
There was a problem hiding this comment.
Not sure it makes sense either given #99120 (comment)
include/zephyr/devicetree.h
Outdated
| * | ||
| * @return node identifier for the child node with the reg address at a specified index | ||
| */ | ||
| #define DT_CHILD_BY_REG_ADDR(node_id, idx, addr) \ |
There was a problem hiding this comment.
if we are going to do this, then i suggest you make a _BY_IDX version , and make this one do index 0. and also could be useful to add a _BY_NAME version.
There was a problem hiding this comment.
Note that resources can be shared between multiple nodes (reg property addresses do not need to be unique across nodes).
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
6d71377 to
21c217a
Compare
Can you elaborate a bit on the use case, please? |
I've added a commit that replaces the |
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
f7ba568 to
3c5f497
Compare
pdgendt
changed the title
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
2 times, most recently
from
bc74e56 to
04106bf
Compare
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
04106bf to
3d2eaf9
Compare
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
3d2eaf9 to
19ab0dd
Compare
Allow fetching child node identifiers by unit address. Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
Introduce test cases to the devicetree API test suite for DT_CHILD_BY_UNIT_ADDR and DT_INST_CHILD_BY_UNIT_ADDR. Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
The macro to initialise a struct adc_dt_spec instance uses a FOREACH macro, which can be problematic when nesting in other FOREACH macros, for example in driver implementations. Replace it with DT_CHILD_BY_UNIT_ADDR. Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
Add a DeviceTree section and entries for the new C macros. Signed-off-by: Pieter De Gendt <pieter.degendt@basalte.be>
pdgendt
force-pushed
the
dt-child-by-reg-addr
branch
from
19ab0dd to
dd22799
Compare
|
7 participants
Allow fetching child node identifiers by unit address.
It also modifies the
ADC_DT_SPEC_STRUCTmacro to not useFOREACHmacro expansion, which doesn't play nice withCOND_CODE_Xand other macro loops.