stm32u5 hal dma_ex function not inlined with GCC 11 or 12

[FRASTM]

Fix the Issue linked to compilation of file HAL_DMAEx_List_ReplaceNode_Head.c
using a GCC version comprised between v11.x and v12.x.

and blocking the zephyr CI

Issue is seen only with TF-M build config RelWithDebInfo (With MinSizeRel config, this file is not built).
Issue could be fixed by "adding attribute ((no_inline)) to DMA_List_CheckNodesBaseAddresses allows the build to pass"

Zephyr issue: zephyrproject-rtos/zephyr#80932

[tomi-font]

• Did you submit this fix upstream? I could not find anything with the Change-Id mentioned.
• When this is done, make sure to cherry pick the upstream commit into the Zephyr fork by using the -x option of git cherry-pick to keep track of the cherry-picked commit.
• This PR is not quite solving the issue in Zephyr by itself. The issue will be solved in Zephyr once this PR is merged and that commit is integrated into Zephyr. For that you will need to create a manifest PR in the Zephyr repository (see PRs with the manifest-trusted-firmware-m label for examples), and there you can have the Fixes X.

[FRASTM]

• Did you submit this fix upstream? I could not find anything with the Change-Id mentioned.
• When this is done, make sure to cherry pick the upstream commit into the Zephyr fork by using the -x option of git cherry-pick to keep track of the cherry-picked commit.
• This PR is not quite solving the issue in Zephyr by itself. The issue will be solved in Zephyr once this PR is merged and that commit is integrated into Zephyr. For that you will need to create a manifest PR in the Zephyr repository (see PRs with the manifest-trusted-firmware-m label for examples), and there you can have the Fixes X.

@tomi-font
fix upstream is https://review.trustedfirmware.org/c/TF-M/trusted-firmware-m/+/33420

Even If I push this change on ~/zephyrproject/modules/tee/tf-m/trusted-firmware-m
the upstream main branch is on 8134106
and what I see in the west.yaml is a11cd27

Where Should my change be ? on top of remotes/upstream/main or on a11cd27 or ?

[tomi-font]

First, where in Zephyr would you like this fix to be included? v3.7-branch or main? (Or both?)

And I'm not sure I get what you're saying, but it's normal that the end commit hashes differ even when cherry picking the commit (because the two branches are not identical.)

Anyway, we'll want to wait for your patch to be approved upstream before merging it here.

[erwango]

Anyway, we'll want to wait for your patch to be approved upstream before merging it here.

@tomi-font TF-M upstream team won't integrate the change until PR is merged and delivered in official HAL package (they don't merge HAL fixes). See https://review.trustedfirmware.org/c/TF-M/trusted-firmware-m/+/33420.
I don't think we can afford to wait that long for current PR. Can we consider merging it faster ?

[tomi-font]

Can't you somehow speed up the process on the HAL side?
(after re-having a look at the upstream PR) I just saw the comment in the review which states 3-4 months. Sounds a bit crazy.

Anyway, sure we can integrate that faster into Zephyr. But can you rework the commit to cherry pick the one you submitted upstream with the -x option (so we have Change-Id and cherry picked from lines)?

[erwango]

Can't you somehow speed up the process on the HAL side?

That's not only about merging this fix. It's about integrating it in a newSTM32Cube package version, perform validation and deliver. Then this Cube package should be integrated into TF-M, validated, then upstreamed. That's a long process.
Then of course this is to be done by teams that have their own priorities which don't fit exactly ours, so you should put additional delay on top of that to get some of their free slots.

Anyway, sure we can integrate that faster into Zephyr.

Ok, great. Thanks

But can you rework the commit to cherry pick the one you submitted upstream with the -x option (so we have Change-Id and cherry picked from lines)?

Sure, we'll do.