Update SDK to 2.14

[mmahadevan108]

cc @JiafeiPan @manuargue @Dat-NguyenDuy

[mmahadevan108]

cc @iuliana-prodan @dbaluta

[manuargue]

Changes LGTM, we will run validation with this branch for the S32 devices to see if there's any issue.

[JiafeiPan]

Hi, @mmahadevan108, may I know whether this PR covers CMSIS update from https://github.com/nxp-mcuxpresso/CMSIS_5/tree/mcux_develop? thanks.

[dbaluta]

@mmahadevan108 Can you please also integrate this PR: nxp-mcuxpresso/mcux-sdk#135 into this SDK update?

Later edit:

I noticed that the changes are merged with SDK update (so big thanks!!)

We need to open a discussion (maybe in another thread) about the process of updating the SDK. Ideally the changes should be cherry-picked preserving the commit message that we have in the original mcu-sdk repo.

I assume that this is not possible for some reason. Then at the least the commit subjects should be added in the commit updating the SDK.

Something like this:

 mcux-sdk: Update to SDK 2.14

This pulls in following commits from mcu-sdk:
34bbf137  fsl_common: Add support for xt-clang compiler
6868378c Incorrect Fifo status on Message Lost

[dbaluta]

Ideally there should be no difference between SDK repo and this repo modulo some specific files that are only used in hal_nxp.

Not sure why this can't be merged to SDK repo? It shouldn't bother anyone but I think you know better so I don't have anything against this change.

[mmahadevan108]

I will check with the SDK team if this can be merged.

[mmahadevan108]

@mmahadevan108 Can you please also integrate this PR: nxp-mcuxpresso/mcux-sdk#135 into this SDK update?

Later edit:

I noticed that the changes are merged with SDK update (so big thanks!!)

We need to open a discussion (maybe in another thread) about the process of updating the SDK. Ideally the changes should be cherry-picked preserving the commit message that we have in the original mcu-sdk repo.

I assume that this is not possible for some reason. Then at the least the commit subjects should be added in the commit updating the SDK.

Something like this:

 mcux-sdk: Update to SDK 2.14

This pulls in following commits from mcu-sdk:
34bbf137  fsl_common: Add support for xt-clang compiler
6868378c Incorrect Fifo status on Message Los

@mmahadevan108 Can you please also integrate this PR: nxp-mcuxpresso/mcux-sdk#135 into this SDK update?

Later edit:

I noticed that the changes are merged with SDK update (so big thanks!!)

We need to open a discussion (maybe in another thread) about the process of updating the SDK. Ideally the changes should be cherry-picked preserving the commit message that we have in the original mcu-sdk repo.

I assume that this is not possible for some reason. Then at the least the commit subjects should be added in the commit updating the SDK.

Something like this:

 mcux-sdk: Update to SDK 2.14

This pulls in following commits from mcu-sdk:
34bbf137  fsl_common: Add support for xt-clang compiler
6868378c Incorrect Fifo status on Message Lost

Ideally you should have seen it when you did a git log inside the mcux-sdk folder. However I did a manual merge of a few commits as I was about to push this PR when I got information about this change, hence you do not see it.

[mmahadevan108]

@hakehuang, can you run a full test cycle on the Zephyr PR to update to SDK 2.14

[mmahadevan108]

Hi, @mmahadevan108, may I know whether this PR covers CMSIS update from https://github.com/nxp-mcuxpresso/CMSIS_5/tree/mcux_develop? thanks.

No, I have not updated the CMSIS package. Can you help update the CMSIS pack we maintain for the A series

[JiafeiPan]

Hi, @mmahadevan108, may I know whether this PR covers CMSIS update from https://github.com/nxp-mcuxpresso/CMSIS_5/tree/mcux_develop? thanks.

No, I have not updated the CMSIS package. Can you help update the CMSIS pack we maintain for the A series

Sure, will create PR for that.

[manuargue]

Changes LGTM, we will run validation with this branch for the S32 devices to see if there's any issue.

we have some SPI tests failing with this branch that we need to further investigate, will feedback soon.

[hakehuang]

@hakehuang, can you run a full test cycle on the Zephyr PR to update to SDK 2.14

there are many build fails, need fix, I will forward you details

[JiafeiPan]

Hi, @mmahadevan108, may I know whether this PR covers CMSIS update from https://github.com/nxp-mcuxpresso/CMSIS_5/tree/mcux_develop? thanks.

No, I have not updated the CMSIS package. Can you help update the CMSIS pack we maintain for the A series

Done with PR: #272 please review, thanks.

[Dat-NguyenDuy]

Hi @mmahadevan108, could you please also integrate this PR: nxp-mcuxpresso/mcux-sdk#132 into the update, thanks.

[mmahadevan108]

Hi @mmahadevan108, could you please also integrate this PR: nxp-mcuxpresso/mcux-sdk#132 into the update, thanks.

Done

[mmahadevan108]

@manuargue Kindly send me a patch for the SPI issue.

[mmahadevan108]

cc @LaurentiuM1234

[hakehuang]

@mmahadevan108 all tset done. now all issues fixed

[manuargue]

@manuargue Kindly send me a patch for the SPI issue.

@mmahadevan108 I've sent you the patch from MCUX SDK team. Let me know if further support is needed. I'll remove the -1 once the patch is pushed.

[mmahadevan108]

@manuargue the patch has been added to this PR. Kindly check.

[manuargue]

All my comments addressed

[marc-hb]

Congratulations, you broke west :-)

• Use 'backslashreplace' not to crash on malformed UTF from subprocess west#700

[butok]

Congratulations, you broke west :-)

• Use 'backslashreplace' not to crash on malformed UTF from subprocess west#700

Could you be more specific what & how is broken,?

[marc-hb]

Did you open the link you quoted?

[hakehuang]

Could you be more specific what & how is broken,?

the original code is as below °C is the problem one. @SusanSu I think we need enforce some encode check for source code.

#define FXPQ3115_TEMPERATURE_CONV_FACTOR (256) /* Will give °C */

[mcuxsusan]

I see the linked improvement with west tool at zephyrproject-rtos/west#700, does this issue finally relate to the improvement to west tool? What I mean is that our code comment which has encode issue hit the hole in the west tool.

[hakehuang]

I see the linked improvement with west tool at zephyrproject-rtos/west#700, does this issue finally relate to the improvement to west tool? What I mean is that our code comment which has encode issue hit the hole in the west tool.

per my understand, the fix goes to tool is better. In the meantime, we need somehow standard our hal with charset encodecing.

[marc-hb]

BTW @dbaluta does it really make sense to dump an entire SDK in a "HAL" module? hal_nxp is currently the largest zephyr/modules/ by far (git checkout = 1.2G) and this almost entirely caused by this sort of SDK dump.

There is a constant stream on Discord and Github of users complaining that west update is too big and too slow and there are regular optimization efforts to solve this problem. Such efforts are easily destroyed by PRs like this one.

See

• manifest: move optional modules to a submanifest and make them optional zephyr#61505

+ a lot more from that link.

[fabiobaltieri]

Oh wow, yeah at a quick glance there's a lot that does not really belong to here, besides 3000+ png images, 2000+ html files and 13 pdf files there's also a 7MB sqlite database and 3 ELF binaries, which I imagine are not used but should probably not be redistributed under the org.

@nashif @carlescufi

We should probably think about making the west diff a bit smarter and detect binary files or unreasonably big deltas.