Pretend unused slots presence in dictionaries from extended manifest #3691
Conversation
ktrzcinx
requested review from
a team,
dbaluta,
jajanusz,
lbetlej,
lgirdwood,
mmaka1 and
plbossart
as code owners
src/init/ext_manifest.c
Outdated
There was a problem hiding this comment.
Just wondering here if we can use a linker variable here for CONFIG_ELEM_COUNT as a way to automate this as it may be error prone for developers i.e. something cmake can pass into the linker pre-processor and have the linker set this based on the FW binary build ???
There was a problem hiding this comment.
@ktrzcinx agreed with Liam, we can try to use some trick with enum like
| sizeof(struct config_elem) * CONFIG_ELEM_CNT, | |
| sizeof(struct config_elem) * (EXT_MAN_CONFIG_LAST_ELEM - 1), |
You do not need to update the EXT_MAN_CONFIG_LAST_ELEM or CONFIG_ELEM_CNT if we add a new type
xiulipan
left a comment
xiulipan
left a comment
There was a problem hiding this comment.
I think we can keep EXT_MAN_*_LAST_ELEM and calculated the CNT with (LAST_ELEM - 1)
No need to update the CNT anymore if we add new enum.
src/init/ext_manifest.c
Outdated
There was a problem hiding this comment.
@ktrzcinx agreed with Liam, we can try to use some trick with enum like
| sizeof(struct config_elem) * CONFIG_ELEM_CNT, | |
| sizeof(struct config_elem) * (EXT_MAN_CONFIG_LAST_ELEM - 1), |
You do not need to update the EXT_MAN_CONFIG_LAST_ELEM or CONFIG_ELEM_CNT if we add a new type
…ig_data This value shouldn't be included in any ABI header, because it may vary between firmware configurations. This change makes header file more similar to version provided by kernel. Signed-off-by: Karol Trzcinski <karolx.trzcinski@linux.intel.com>
…ifest This value should be included to output image with proper value instead of including only for some configurations. It will reduce dictionary element counting complexity. Signed-off-by: Karol Trzcinski <karolx.trzcinski@linux.intel.com>
…_config_data This value shouldn't be included in any ABI header, because it may vary between firmware configurations. This change makes header file more similar to version provided by kernel. Signed-off-by: Karol Trzcinski <karolx.trzcinski@linux.intel.com>
ktrzcinx
force-pushed
the
ext-man-dict-cnt
branch
from
24c7796 to
7e9fd5d
Compare
xiulipan
left a comment
xiulipan
left a comment
There was a problem hiding this comment.
LGTM not. I am OK to have a more clear define for the COUNT.
|
@ktrzcinx can you check internal CI, looks like tests are failing ? |
|
@lgirdwood @ktrzcinx I re-check our discussion in #3527 |
|
@lgirdwood @ktrzcinx Good to go, simple on-device test did not show issue with GCC FW binary. |
|
@lgirdwood in internal CI single test is failing with message |
|
@zrombel good to merge, this error is unrelated ? |
|
It seemed unrelated. I run tests one again to be sure and now it's all green. Good to merge. |
4 participants
Solution extracted from discussion from #3527