soc: ambiq: adjust init process #90909
Conversation
| return 0; | ||
| } | ||
|
|
||
| SYS_INIT(apollo5x_init, PRE_KERNEL_1, 0); |
There was a problem hiding this comment.
This is another attempt to try fix #90777 as we have too much stuff be putting into early init hook. There seems to be no other way as modifying thread.c or simply making the dummy bigger is not the way to go…
There was a problem hiding this comment.
Too much stuff how? This goes against the idea of what soc hooks were introduced for @nashif
There was a problem hiding this comment.
Too much stuff how? This goes against the idea of what soc hooks were introduced for @nashif
Our low power init function and its calling functions occupy too much ROM space, it gets to the point that the literals used by the ldr instruction in thread.c gets pushed too far away from its caller. Then the offset out of range error is emitted.
I think it is also inappropriate to run large amount of code in the early init hook. @nashif Is this correct?
There was a problem hiding this comment.
no, you can do pretty much whatever you want in the hooks as long you have the right dependencies in place.
There was a problem hiding this comment.
no, you can do pretty much whatever you want in the hooks as long you have the right dependencies in place.
okay, then can we split it for now, in order to get around the build failure… We don’t have a lot of choice here.. and it is blocking all our PR for apollo5 platform.
split soc init into two parts as it is too large to be up front at early stages. Signed-off-by: Swift Tian <[email protected]>
github-actions
bot
requested review from
aaronyegx,
AlessandroLuo and
RichardSWheatley
|
Can we get this in? The CI fail #90777 becomes a blocker for all our PRs. |
|
split soc init into two parts as it is too large to be up front at early stages.