hwmv2: Port NXP RT Family

[decsny]

No description provided.

[tejlmand]

Mostly minor nits.

Please check compliance.

As pointed out by @nordicjm then please consider if using board variant for "hyperflash" and "qspi" isn't more correct.

With hwmv2 we now support a more gerenal variant term.

[decsny]

@nordicjm @tejlmand

I did consider and try using variants but found revisions was easier to implement and in my opinion made more sense. The only issue is that there is no such thing as multidimensional revions. Such as, revision A and B, and revision hyperflash and qspi.

My response to Jamie above:

Not only are revisions the most convenient way to handle this in the board definition, but also, in my view, these really are different board revisions, because switching between hyperflash and qspi requires reworking the board hardware with a soldering iron. And hyperflash and qspi are not really variants of the SOC at all, they are still using the same interface to the SOC, they are just two different flash chips. The fact they are just on the same board is just because it's an EVK... these are really two different board versions that happen to exist on the same PCB.

[tejlmand]

Minor proposal, not blocking for a 👍

[tejlmand]

My response to Jamie above:

Not only are revisions the most convenient way to handle this in the board definition, but also, in my view, these really are different board revisions, because switching between hyperflash and qspi requires reworking the board hardware with a soldering iron.

I can accept this as a valid reason for choosing board revision over variants.
A bit of advice to avoid future questions / wondering related to this, then I propose to add this reasoning in the revision.cmake.

But you still need to address comments like these:
#69264 (comment)
#69264 (comment)

I suggest you try it once more, cause the proposed pattern has worked for any other board.
Alternative push a fixup commit to this PR with the changes you did which was not working, perhaps @nordicjm or I can see what is wrong.

[nordicjm]

Looks good, minor nit.

[danieldegrasse]

@decsny it looks like some of the RT1170 EVKB board files were missed in the rename:

$ find tests -name "mimxrt1170_evkb_cm7*"
tests/drivers/i2c/i2c_target_api/boards/mimxrt1170_evkb_cm7.overlay
tests/drivers/dma/loop_transfer/boards/mimxrt1170_evkb_cm7.overlay
tests/drivers/dma/loop_transfer/boards/mimxrt1170_evkb_cm7.conf
tests/drivers/mbox/mbox_data/boards/mimxrt1170_evkb_cm7.overlay
tests/drivers/mbox/mbox_data/boards/mimxrt1170_evkb_cm7.conf
tests/drivers/adc/adc_api/boards/mimxrt1170_evkb_cm7.overlay

[danieldegrasse]

This block

if FLASH_MCUX_FLEXSPI_XIP
# Avoid RWW hazards by defaulting logging to disabled
choice FLASH_LOG_LEVEL_CHOICE
	default FLASH_LOG_LEVEL_OFF
endchoice
choice MEMC_LOG_LEVEL_CHOICE
	default MEMC_LOG_LEVEL_OFF
endchoice
endif

Currently depends on SOC_SERIES_IMXRT10XX || SOC_SERIES_IMXRT11XX. It should only depend on SOC_FAMILY_NXP_IMXRT, this setting needs to be applied for all IMXRT SOCs