Backport v3.6: Fix IAR endianness detection in alignment.h

[nzzlinh]

Description

This is a backport of Mbed-TLS/TF-PSA-Crypto#838.

This PR fixes a critical bug where the IAR compiler always triggers the Little-Endian code path, even when compiling for Big-Endian CPU. Fix #10381
The IAR C/C++ compiler always defines both __LITTLE_ENDIAN__ and __BIG_ENDIAN__ macros internally, assigning '1' to the active endianness and '0' to the inactive one. The previous check defined(__BIG_ENDIAN__) always evaluated to true on IAR. This commit updates the check to verify the actual value: (__BIG_ENDIAN__ == 1).

References from IAR Development Guide for several MCU families

[1] ARM: https://wwwfiles.iar.com/arm/webic/doc/EWARM_DevelopmentGuide.ENU.pdf - p.494
[2] Renesas RX: https://wwwfiles.iar.com/RX/webic/doc/EWRX_DevelopmentGuide.ENU.pdf - p. 428
[3] Renesas RL78: https://wwwfiles.iar.com/rl78/EWRL78_DevelopmentGuide.ENU.pdf -p.389
[4] STM8: https://wwwfiles.iar.com/stm8/guides/EWSTM8_DevelopmentGuide.pdf - p.334

PR checklist

• changelog provided
• TF-PSA-Crypto development PR provided Fix IAR compiler endianness detection in alignment.h TF-PSA-Crypto#838
• TF-PSA-Crypto 1.1 PR Backport 1.1: Fix IAR endianness detection in alignment.h TF-PSA-Crypto#845
• mbedtls 3.6 PR : This PR
• tests not required because ...

[nzzlinh]

Dear @bensze01 ,

Just a quick question. We will need this fix to be merged into the mainstream before we can pull it back into our work, so we are trying to adjust our plan accordingly.

Could you please let us know which MbedTLS release this MR is currently targeted for, and if possible, the tentative timeline for merging?

Thank you 😄

[gilles-peskine-arm]

@nzzlinh There is no target release. We lack review bandwidth and this pull request is competing with a lot of others.

Applications for becoming a maintainer are open. Serious enquiries only please — in practice, you need to work on features you don't care about in order to make time for the features you don't care about.

[nzzlinh]

@gilles-peskine-arm Thank you for the clarification.

Just to be clear, I was not asking for a target date nor trying to rush the review process. I simply wanted to understand whether there was any plan to include this fix in a future release.

I completely understand that maintainers have limited review bandwidth and many competing priorities. We're happy to wait and appreciate the work that goes into keeping the project healthy.

In any case, thank you for taking the time to respond, and hopefully this fix will find its way into a future release when bandwidth permits.

Many thanks and have a good day 😸