imx-secure-enclave: Update to lf-6.12.3-1.0.0 #2252
Conversation
|
Hi, @thochstein @otavio and I think it's best to combine the different recipes in the meta-imx layer into a single one, for better organization. Can you check if everything is ok? |
thochstein
left a comment
thochstein
left a comment
There was a problem hiding this comment.
I don't think it makes sense. For 95, the imx-secure-enclave recipe did depend on the imx-secure-enclave-seco recipe. So, two builds were being done, one for PLAT=ele, and one for PLAT=seco.
| } | ||
|
|
||
| do_install:append:mx95-nxp-bsp() { | ||
| # Remove common content that is to be installed by imx-secure-enclave |
There was a problem hiding this comment.
This doesn't make sense since the recipes are now combined. Combining the recipes suggests that you can just drop the append. Note that this is secondary to my other comment.
There was a problem hiding this comment.
We can certainly put both together. However, we will need to have a variable to control whether we need to remove or not those files.
There was a problem hiding this comment.
With two recipes, the source install duplicated headers, a README, and a script, which caused a build break due to installing the same file multiple times to the rootfs. With one combined recipe, this specific problem doesn't exist, so there is by definition no need to include this specific logic from the dual recipe solution.
|
Ok, I understand what you are explaining, that you have two different builds, but does it make sense to have it build twice? Could you provide more information about why do we need to have both builds and why this is specific of this SOC type? |
|
I don't understand the details, but what I do understand is that they requested both PLAT flavors to be installed for certain SOCs. Building both PLAT flavors in a single recipe seems a good idea. To verify the results, please check that the packages created with the combined recipe do accurately contain the full union of the packages created with the separate recipes. |
|
Okay, but I think that we need more details to understand why this needs to be done. You know that sometimes they envision some fancy architecture for the recipes that are not required, or it's just a misunderstanding between what needs to be done and the proper way of doing that. So could you inquire that internally to understand why this needs to be done that way? And then we see how we can try to support that properly. |
|
I think this answers your question: https://github.com/nxp-imx/imx-secure-enclave/blob/lf-6.12.3_1.0.0/README |
mdrodrigo
force-pushed
the
topic/imx-secure-enclave
branch
from
ee3bc1c to
aae1cef
Compare
Signed-off-by: Rodrigo M. Duarte <[email protected]>
mdrodrigo
force-pushed
the
topic/imx-secure-enclave
branch
from
aae1cef to
9fb2dc1
Compare
|
is this PR still relevant? |
|
I think it's relevant, but unfinished. The commit did an update and a rework, but the rework looks to have problems as I pointed out. |
4 participants
No description provided.